Copies of this publication may be procured fro~ the Genetics Foundation, Department of Zoology, The Uniyers!ty of Texas, . ' Austin 12, Texa~ PROFESSOR OF ZOOLOGY studies in GENETICS II. Research reports on Drosophila Genetics, Taxonomy and Evolution THE UNIVERSITY OF TEXAS AUSTIN The benefits of education and of useful knowledge, generally diffused through a community, are essential to the preservation oi a free government. SAM HOUSTON Cultivated mind is the guardian genius of Democracy, and while guided and controlled by virtue, the noblest attribute of man. It is the only dictator that freemen acknowledge, and the only security which freemen desire. MIRABEAU B. LAMAR PUBLISHED BY THE UNIVERSITY TWICE A MONTH. S)>COND CLASS POSTAGE PAID AT AUSTIN, TEXAS. Contents I. Genetic studies of irradiated natural populations of Drosophila. V. Summary and discussion of tests of populations col­lected in the Pacific Proving Ground from 1955 through 1959 WILSON S. STONE, MARSHALL R. WHEELER, AND FLORENCE D. WILSON 1 II. Fixed heterozygosity in a parthenogenetic species of Drosophila HAMPTON L. C ARSON 55 III. Cytological studies of the repleta group of the genus Dro­sophila: III. The mercatorum subgroup MARVIN WASSERMAN 63 IV. Cytological studies of the repleta group of the genus Dro­sophila: IV. The hydei subgroup MAR\.'.IN WASSERMAN 73 V. Cytological studies of the repleta group of the genus Dro­sophila: V. The mulleri subgroup MARVIN WASSERMAN 85 VI. Cytological studies of the repleta group of the genus Dro­sophila: VI. The fasciola subgroup MARVIN WASSERMAN 119 VII. Notes on the taxonomy, morphology, and distribution of the saltans group of Drosophila, with descriptions of four new species Luiz EDMUNDO DE MAGALHAES 135 VIII. The alagitans-bocainensis complex of the willistoni group of Drosophila . MARSHALL R. WHEELER AND Luiz EDMUNDO DE MAGALHAES 155 IX. Genetic characteristics of island populations WILLIAM B. HEED 173 X. The problem of phylogeny in the genus Drosophila LYNN H. THROCKMORTON 207 XI. Effects of X-ray irradiation in Drosophila virilis at different stages of spermatogenesis FRANCESE. CLAYTON 345 The University of Texas Publication XII. Behavior of mitochondria in living cells during meiosis in Drosophila virilis . 375 E1z1 MoMMA XIII. Further observations on the relation between gas pressure and X-ray damage in Drosophila melanogaster . 385 TsuENG-HSING CHANG XIV. The flavopilosa species group of Drosophila . 395 MARSHALL R. WHEELER, HARUO TAKADA, AND DANKO BRNCIC XV. The use of biochemical characteristics for the study of prob­lems of taxonomy and evolution in the genus Drosophila . 415 LYNN H. THROCKMORTON XVI. Changes with evolution of pteridine accumulations in species of the saltans group of the genus Drosophila . . . 489 LYNN H . THROCKMORTON AND L. E. MAGALHAES XVII. The dominance of natural selection and the reality of super-species (species groups) in the evolution of Drosophila 507 WILSON S. STONE XVIII. Hybridization studies within the cardini species group of the Genus Drosophila 539 DAVID G. FUTCH Acknowledgment of Research Support The studies reported here have been conducted primarily by the staff, students, post-graduate fellows, and visiting investigators of the Genetics Foundation of the University of Texas. Support for certain specific projects has been acknowledged in footnotes accompanying these reports. However, adequate financial support is so important in maintaining a large laboratory and its correlated field work that we wish here to express our special gratitude to the various groups and agencies whose support has made these contributions possible. The basic genetics laboratory has been supported for a number of years by the Rockefeller Foundation and by funds from the University of Texas. The Robert A. Welch Foundation has supported part of the work in biochemical genetics and related fields. The U.S. Atomic Energy Commission, through grant AT-(40-1)-1323, provided funds for the investigation of the Drosophila popula­tions of the Marshall Islands and adjacent areas, as well as other laboratory studies of radiation damage. The National Science Foundation, through grants G-1653 and G-4999, financed extensive field work in the Caribbean region. Finally, we wish to acknowledge the extensive support, of both laboratory and field work, furnished by the National Institutes of Health, U. S. Public l Iealth Service, through grant RG-6492, C1 and C2. I. Genetic Studies of Irradiated Natural Populations of Drosophila. V. Summary and Discussion of Tests of Populations Collected in the Pacific Proving Ground from 1955 through 19591 WILSON S. STONE, MARSHALL R. WHEELER, AND FLORENCE D. WILSON When Muller (1927) demonstrated that X-rays caused mutations and chro­mosome abnormalities, this inevitably raised the question of the amount of effect of natural ·and artificial radiations on the frequency of mutations in natural populations. At that time cosmic radiations, natural radioactive isotopes and some medical X-radiation were the known sources of short wave radiations. Since then the amount of medical and industrial X-radiation and isotope radiation has increased tremendously. Furthermore there has been added the radiations and isotopes from atomic reactors as well as those produced by atomic and ther­monuclear explosions. We have studied the genetic effects of radiations, primarily from fallout, due to weapons testing in the Pacific Proving Ground, on populations of an endemic species, Drosophila ananassae (Stone, Wheeler, Spencer, Wilson, Gregg, Neuen· schwander and Ward, 1957; Stone and Wilson, 1958; Gregg, 1959; Stone and Wilson, 1959). Chart 1 shows the main localities in the Marshall Islands and the eastern Carolines including all islands from which we tested populations. The irradiated natural populations of Drosophila ananassae came from three atolls in the northern Marshall Islands, Bikini, Rongelap and Rongerik; these have less rainfall, around 50 inches per year, than does Majuro, farther to the south, which has between 120 and 150 inches per year. The Ponape lowlands from which we usually collected the population samples have around 220 inches per year. There is a dense rain forest on this island but only the limited atoll island flora on the Marshall Islands. The usual range of temperature in the northern Marshalls is from 77° to 87°F at midday with a mean morning and evening temperature of 78°F. The other islands where collections were made have the same general temperature range, although more extreme temperatures do occur at all these localities. The humidity is generally high. Populations and radiation history.-The natural food sources for Drosophila varied from island to island and from year to year (Stone, et al, 1957). On Bikini the main sources of food were the fallen fruits of Marinda citrifolia, pandanus and papaya, but the latter had almost entirely disappeared in 1958 and 1959. On Rongelap, in addition to these three fruits, which obviously were Drosophila food and breeding sites, there were some breadfruit trees at the Marshallese lThis work was supported by a contract with the Atomic Energy Commission [AT-(4-0-1)­1323] and by a grant from the Rockefeller Foundation. The University of Texas Publication ENIWETOK ATOLL ".~Parry BIKINI ATOLL RONGELAP r.'.' ;Bwn; ATOLL RONGERIK ATOLL . ...:<.?· '·f::!~Y:r·* ·~~ Ro11911 tap 10° KWA.JALEIN ATOLL .. '·.· 0 0 if;.PONAPE , KUSAIE. MAJUROu11ga._ ATOLL t:..• ~· ..... .,··-. village. Majuro had these four fruits plus bananas. In 1955 on Eniwetok island of the Rongerik atoll, the only available food appeared to be the small fruit of Guettarda speciosa; the Drosophila population was very small and had disap­peared by 1956. On the island of Rongerik we found Drosophila ananassae on the fallen fruit of Morinda and pandanus. Ponape, the "high island," had Morinda, pandanus, papaya, breadfruit, and bananas, and there were other types of fruits in the forest which could provide a reserve breeding ground for the lowland population. Occasionally we would find D. ananassae on the open coconuts on Majuro but Drosophila bryani was much more common on that food. We found no population of Drosophila ananassae on any of the islands of the Eniwetok atoll; there was a sparse and scattered population, which had disap­peared in 1959, on Bikini island in the atoll of that name. Similarly, there was a sparse and scattered population on the island of Rongelap, except that in 1956 we found a large local population on the fallen fruit which was temporarily plentiful at the site of the Marshallese village, and there were very thin scattered populations on the island of Rongerik in the Rongerik atoll (and on Eniwetok in 1955, but this disappeared by August, 1956). As controls, we sampled two populations which. had received very little fall­out radiation. The one in the Marshall Islands was located on the island of Uliga in the Majuro atoll. The other control population came from the island of Ponape in the eastern Caroline Islands (1956 through 1959). Drosophila ananassae is widespread in tropical areas and Pipkin (1953) had shown that it was the dominant Drosophila species on the Truk atoll. Only Drosophila ananassae was found on the northern Marshall Islands. At Majuro there are several other species and at Ponape there are still others, but ananassa.e is the dominant Drosophila ~pecies on both islands. On Majuro, Drosophila ananassae forms scattered small to medium-sized populations whereas the Ponape population is usually medium­sized to large. However in 1958 a typhoon had largely destroyed the fruit crop on Ponape and the Drosophila population existed as a series of small remnants in the lowlands that summer. It had recovered in size fairly well by August, 1959. Other natural factors affecting the population size of Drosophila anana.ssae, such as competitors and predators, have been discussed at length in the first publication of this series (Stone, et al 1957). Population fluctuations were ob­served from year to year, the most extreme examples being the apparent extinc­tion of the species on Eniwetok island of Rongerik atoll between 1955 and 1956, and on Bikini island, Bikini atoll, between 1958 and 1959. Weapons tests were carried out on or around the Eniwetok and Bikini atolls. The islands of these two atolls received fallout each of the test years, while the Rongelap and Rongerik atolls received fallout principally in 1954. The atomic tests prior to 1954 undoubtedly caused some genetic damage to the Drosophila population on the island of Bikini, but the amount of radiation from fallout due to the March 1, 1954 thermonuclear detonation was so great compared to earlier tests that we can assign most of the genetic damage due to irradiation of popula­tions sampled in the summers of 1955 through 1959 to that exceptionally heavy fallout. Figure 1, reproduced from page VIII in Some Effects of Ionizing Radia­tion on Human Beings, TID 5358, published by the AEC in 1956, gives some idea of the fallout pattern and radiation intensity. Figures 2 and 3, reproduced from pages 2 and 3 of Radioactive Contamination of Certain Areas in the Pacific Ocean from Nuclear Tests edited by Gordon M. Dunning (1957), also published by the AEC, illustrate still other ways to estimate the amount of gamma and other radiations. Dr. Douglas Grahn, Geneticist, Biology Branch, Division of Biology and Medi- FIGURE I ISOD OS E LINES OF ESTIMATED PATTERN OF RADIOAC TIVE FALLOUT PACIFIC PROVING GROUNDS MARCH I, 1954 166° 168° 170° I ATOJ~ BIKAR dB / KAR ISLAND 12 ° 12() UTIRIK ATO Liif TAKA \] Ur/RIK ISLAND ATOLL 170° The numbers on the above map represent the doses that would have been received over ap­proximately 48 hours without shielding. The dose, above which survival is unlikely, is 800 r and below which survival is probable is 200 r. The University of Texas Publication F IGURE 2 APPROXI MATE GAMMA DOSE RATES AT THREE FE ET ABOVE THE GROUND ON D+l. (One day after Detonation) (Roentgens per hour) 168° !70° BIKAR ATOLL o _ BI KAR {16) 12°--------1----------t----------t------12° LUKUEN (21.0) GEJEN(JS.O) lIKABELl E (19.0) LOMU/LAL /.J5.0/\ /,,LATO BA CK (6.6) BIKIN I a "\ ,/RONGERIK (5.$} AT OLL ENIAETOK UTIRIK ATOLL I R/ I (8.5/ LABAREOJ RO tlGER IK AILING !NA E (l.J.0 } ATOLL ~ RONGELAP j / ~o~~~~~K AT OLL AT OLL ENIWETAK (J.8 ) C::i-UTIRIK (0 3 4) TA K5J ATOLL SIFO ISL ANO (I J) RONG£LAP (:J.5) I 100 150 200 S!alu1' M lt1 1 10"-------1------------t----------+------10° 166° 168° cine, The Atomic Energy Commission, kindly provided the following information obtained from their Radiation Effects Branch: The so-called infinity dose for Rongelap would be about 420 r, for Rongerik about 640 r. An estimate for the island of Bikini is one or more kiloroentgens but a closer estimate is not available (this is not an island open to the Marshallese) . Over 99 %of the fallout on Ronge­lap and Rongerik was due to the March 1, 1954 test. Although no other fallout was as heavy, the island of Bikini received some fallout in each operation. Some additional idea of the amount of external radiation (usually measured as gamma three feet above the ground) can be obtained from the following com­parisons. The so-called infinity dose at Rongelap was 420 r. Table 36 of Radio­active Contamination of Certain Areas in the Pacific Ocean from Nuclear Tests gives the external radiation dose of animals from the living area of Rongelap. They had received 280 r by eight days post detonation (March 1, 1954) which is a period equivalent to almost one generation in the Drosophila ananassae time scale; in 25 days, 330 r had been received by the first two and some immature stages of the third generation; in 33 days, roughly three generations in the ana.nassae time scale, the external radiation dose was. 340 r; while in 51 to 53 days, about five generations, the dose had risen to 360 r. In other words, the Drosophila on Rongelap had received more than two-thirds of the external radia­tion the first generation and about 85.7% in the first five generations. The situa­tion on Bikini island must have resembled that on Rongelap with some additional effect of very short-lived isotopes which increased the relative dose to the first generation. The fallout from the test of March 1,' t954 was 4 to 6 hours reaching Rongelap from Bikini atoll, therefore there must paye been proportionally more early radiation on Bikini island (which had a sh(( time lag from detonation to fallout) due to the very high radiation activity/ m short-lived isotopes. We were informed that there was over a kiloroent · · 'on Bikini island and we can x 1.0 10 10 0 1000 TIME AFTER DETONATION (Days) conclude that there was less than 20 kiloroentgens total radiation (external and internal) in the first generation received by at least some of the flies. Although the LD 50 of several species of Drosophila is around 100 kiloroentgens using X-rays, 20 kiloroentgens will cause complete or nearly complete sterility and make survival of the population doubtful. In view of the position of the islands Bikini and Rongerik (the latter with a calculated infinity dose of 640 r) relative to the point of detonation and time of fallout, we can estimate that the first four generations of flies on Bikini probably :i;eceived between 5,000 and 10,000 r, with many of them receiving closer to the latter amount (Figure 1 and Table 11). The University of Texas Publication Stone, Alexander, Clayton and Dudgeon ( 1954) showed that Drosophila virilis males were fertile when subjected to 1,500 rep of neutrons with some gamma contamination from an atomic explosion at the AEC Nevada Proving Ground. However males receiving double that dose were sterile (although they produced motile sperm which fertilized eggs) because of the excessive genetic damage, mostly chromosome abnormalities. Females are sterilized by lower dosages of radiation than males. Even taking into account the external gamma and beta radiations from the several measured depositions of radioactive substances in terms of their decay rates, the infinity dose to the populations, calculated in terms of gamma radiation measured three feet above the ground, cannot tell us the amount of radiation to the gonads-for example some of the few alpha or the beta radiations might penetrate small larvae. There is also the problem of internal decay from ingestion of radioactive materials in food. The two AEC publications mentioned above present information on isotopes in some plants and animals. We do not know if the microorganisms used by Drosophila also concentrate at least some radioactive isotopes; for example, pandanus fruit concentrates Strontium 90 but it is not clear if this is readily available to the Drosophila in the juice, retained in the hard tissues, or concentrated further by microorganisms growing on the fallen fruit. The fact that Drosophila lives on or near the ground to feed and breed on the fallen fruit makes even a gamma estimate of dose based on measurements three feet above the ground open to some error. The continuing radiation from fallout over many generations (the radiation level was well above normal background on Bikini island in July, 1955) continues to add new mutations at a higher than normal frequency even though natural selection is eliminating adverse muta­tions through time. Background and characterizations of populations.-Certain information is necessary before a population analysis of the effects of radiations will be meaning­ful. Very many people have worked on the qualitative and quantitative genetic effects of radiations since Muller's original discovery. Many of the findings are summarized in such monographs as Radiation Biology edited by Hollaender (1954). Wallace and his collaborators (for example, see Wallace, 1956) have studied the effect of short wave radiation of Drosophila melanogaster popula­tions in the laboratory. Among other important points he has shown something of the upper tolerance limits for continuous irradiation, the accumulation of lethal and other mutations during radiation, and the decay of induced variability through time after the cessation of irradiation. A large number of studies of the genetics of Drosophila populations have been made since Chetverikov (1926, 1927) published on this problem. We will refer to some of them in connection with various phases of the analysis and discussion. With these basic studies available, we needed to know if Drosophila ananassae responded to radiation like other Drosophila. Seecof (in Stone, et al, 1957) ran a test of the effect of X-rays on the production of chromosome abnormalities. The response was similar to that of Drosophila melanogaster, a member of the same species group of the subgenus Sophophora (see Muller, 1954, and Kaufmann, 1954, for detailed discussions of radiation damage to D. melanogaster). No control test of mutation rates was run but Seecof's tests together with Moriwaki's (1935, 1938) studies of ananassae indicated that we could expect a similar in­crease in mutation frequency with radiations. Each year we have investigated variables which influence survival of the populations or species. In addition we tried to characterize the populations of ananassae genetically and cytologically. Drosophila ananassae resembles melano­gaster in that there are known a number of gene arrangements, paracentric in­versions, scattered around the world. However populations and individuals at any one locality show few rearrangements. There are no examples of multiple inversion balanced polymorphism such as exist in Drosophila pseudoobscura although the inversion differences present must be retained by selective advan­tage, at least when present in limited frequency. Seecof found that these Pacific island populations were typical. Two inversions previously described by Kikkawa ( 1938) as present in other populations are found in the islands. In addition there are eight new inversions, four from Bikini, three from Majuro and two from Ponape (one of these is also present on Majuro). Some of those on Bikini may have resulted from radiation. All are infrequent and the differences between the irradiated and control populations are not significant. Certainly if there were many chromosome abnormalities produced in the population on Bikini by the heavy fallout radiations in March 1954, most of them had been eliminated by natural selection in the 16 or 28 months between the thermonuclear detonation and the times the population was sampled. In order to determine something about competition, we ran population cages (which may contain 10,000 adults when going well) using arumassae in compe­tition with other species such as Drosophila funebris, D. busckii, D. novamexi­cana, D. virilis and D. melanogaster. Drosophila ananassae easily displaced all of these except melanogaster (an ebony strain). The latter had the advantage at 71±1°F, was about equal in competitive ability at 74±1°F and was displaced rapidly by ananassae at 77±1°F, the temperature most comparable to that of the Marshall Islands. Furthermore, ananassae larvae are very hardy and developed a much greater percentage of the time on a limited amount of food. Spencer (in Stone et al, 1957) reported his analyses of populations of D. ananassae from Bikini, Majuro and Ponape for visible mutations. The pattern and frequency of visible mutations resembled those of many other species he had studied. There were present some unique types of mutations in addition to the types (such as eye color, bristle shape, etc.) usually found. The frequency of visibles at Bikini (0.566 per fly tested), where the population underwent a large and continuing irradiation was not greater than on the control islands of Majuro (0.795) and Ponape (0.529). On the other hand, the effects of the small population size, a higher level of lethals, and the presence of highly inviable mutants in the Bikini population may have masked the presence of some addi­tional visibles. The main conclusion to be drawn from his study of visible muta­tions and from Seecof's studies on chromosome abnormalities is that ananassae is a fairly typical Drosophila resembling Drosophila melanogaster genetically, al­though it is much more fragile and difficult to test effectively in the laboratory than are many other Drosophila species. Results and comparisons of populations within and between years.-In addi­tion to the characteristics of the populations already discussed, we investigated The University of Texas Publication the differences in viability (percent of eggs laid by inseminated females develop­ing into adults)' fertility (percentage of pairs fertile) and fecundity (average number of eggs per fertile female per day) between populations and between years. Statistical tests measuring the degree of differences were given earlier for the 1955 and 1956 populations (Stone, et al, 195 7). These variables are im­portant components of evolutionary fitness. Each is influenced by environmental and genetic factors. We reduced sources of differential laboratory environmental influence as far as possible by running similar tests of different stocks or crosses at the same time any one year. Comparisons between irradiated (Bikini, Ronge­lap, Rongerik) and control (Majuro, Ponape) populations allowed us to make . some deductions concerning the effects of irradiation as well as changes from year to year. Flies caught on the islands in July or August were shipped by air or brought back to the laboratory personally. Their progeny were tested by in­breeding and crossbreeding in pairs. Records were kept of the number of eggs laid per day by each female and the number of eggs that developed into adults from each lot of eggs. This allowed us to determine the percentage of sterile pairs (no eggs developed), the average number of eggs laid per day by fertile females, and the percentage of eggs that developed into offspring, again only from the fertile females. Only those lots of eggs were included that produced offspring. This grouped all causes of failure to produce viable offspring by either male or female into the one sterile class. Table 1 gives the data obtained with the flies collected during the summer of TABLE 1. On the left of Table 1 are shown the strains and their letter symbols, and the crosses; the remaining columns show the number of flies tested and the results. Crosses. 1-8 are within-population tests, 1, 3, 5, 7 being brother X sister matings of F1 of wild females, and Z, 4, 6, 8 being randomly mated F1 .: 9-11 are between-population crosses which were made, as nearly as possible, at the same time as crosses 1-8. 12-14 are matings of F 1 from the between-population crosses (9-11) . 15-30 show the various breeding tests for each island popu­lation, P, M, K, and A, symbolized as follows: (b-u): F1 brothers from wild females X unrelated females (b-s): F, from (b-u) mated brother X sister (c) : F, from (b-u) mated as cousin-pairs (u): F. from (b-u) crossed at random (unrelated) 31-34 are crosses arranged as in 15-30 but beginning with the F1 from the A X M between­population cross of M brothers X unrelated A females. Results. Column 1 gives the number of females which gave fertile progeny as demonstrated in brother X sister matings (cols. 3-5); if one or more of these progeny pairs had a lethal or lethal equivalent gene complex (scored lethal when less than 80 percent of the eggs laid by a female which produced progeny developed, columns 10-14), their parent pair was scored as having one or more lethals in column Z, "Percent with Lethals". This is the very crude and minimum estimate, for many P1 females had too few F 1 pairs fertile; the arbitrary but necessary level of egg development regarded as normal (from 80 to 100%) influences the number of flies classed as carrying lethals, and furthermore double fertilization of the P 1 females would reduce the lethals detected. It is the estimate given in earlier papers but Table 7 gives a much more accurate estimate using more effective methods. The total number of F1 pairs tested and the percent fertile are given in columns 3-5. The results of tests of fertile pairs (column 4) give egg development and its pattern. Column 6 gives the days eggs. were counted from these pairs and column 7 gives the average number of eggs laid by females known to have been fertilized an:l the percent of these eggs that developed into adults from eggs laid the next 4 to 6 days. The pattern of egg development in the several crosses is sho~n iri columns 10-14. These columns g;ves the percent of females in classes with their egg developni:imt ranging from 1 to 19 percent, etc. Only the last class (80-100%) is regarded as normal develo,Iiment. >~~~ TABLE 1 Results of tests within and between populations, inbred and crossbred in various combinations, to demonstrate the fertility, fecundity and viability of the irrailiated and control populations sampled in 1959 P1 females tested F1 fertility F1 egg prnduction Egg developmcn t pa Lte111 N um ber producing Stock or cross fertile F, ( pai rs) Column (1) 1. Ponape (P ) 57 2. P (random ) 3. M ajuro (M ) 42 4. M (random ) 5. Rongerik (K ) 34 6. K (random ) 7. Rongelao (A 1 48 8. A (rando1ti) 9. M x P Percent with letha ls (2) 94.4 90.5 88 .2 87.5 Pairs tested (3) 242 227 191 201 168 227 191 2.32 151 Pairs fertile Number Percent (4) (5) 11 3 59.1 111 48.9 121 63.4 155 77.1 80 47.6 91 40.1 146 76.4 134 57.3 26 17.3 Days eggs Eggs/ day counted per ~ (6) (71 364 15.4 264 12.0 193 9.6 423 14.3 169 13.0 184 12.6 306 10.2 366 13.5 31 13.3 No. eggs (8) 5609 3167 1850 6051 2200 2320 3116 4934 41 3 Percent developed (9) 67.6 80.0 61.0 74.8 60.4 64.6 62.8 69.1 80.4 Percen t of 'i' 'i' whose eggs devel oped 1-19 20-39 40-59 69-79 80-100% (10) (11) (12) (13) (14) 2. 7 10.6 23.0 33.6 30.1 2.4 7.2 7.2 14.4 68.7 2.7 12.2 29.7 32.4 23.0 0.8 4.7 18.9 25.2 50.4 8.7 20.0 27.0 20.0 24.3 5.6 11.1 2.3. 6 26.4 33 .3 1.8 16.8 16.8 31.9 32.7 3.3 9.2 14.2 39.2 34.2 8.3 50.0 41.7 v:i..... 0 ;::sSti ~ [\I:> "'I ~ ;::s ~ 10. A X M 354 154 43.5 369 13.3 4917 79.4 0.8 6.7 9.2 17.5 65.8 11. P X A 12. MP X MP 13. AM X AM 17 30 88.2 93 .3 154 75 134 80 59 89 52.0 78.7 66.4 233 172. 186 10.8 13.5 16.4 2522 2.324 3055 81.2 62.9 62.2 4.2 4.3 8.0 14.1 5.8 32.0 21.1 29.0 28 .0 31.0 60.9 32.0 29.6 ~ .... §" 14. PA X PA 64 75.0 264 161 61.0 431 19.4 8013 66.1 5.0 8.6 24.3 22.9 39.3 ~ 15. P (b-u ) 16. P (b-s) 17. P (c) 18. P (u ) 19. M (b-u ) 20. M (b-s) 21. M (c) 22. M (u ) 23. K (b-u ) 24. K (b-s) 25 . K (c) 26. K (u ) 27. A ( b-u) 28. A (b-s) 29. A (c) 30. A ( u ) 31. A X M (b-u ) 32. A X (b-s) 33 . A X M (c) 34. A X M (u ) 173 64 63 67 201 92 89 94 227 58 60 56 184 102 98 90 134 68 63 64 89 33 21 23 155 61 58 47 91 21 24 16 94 71 71 62 82 47 50 36 51.4 51.6 33.3 34.3 77.1 66.3 65.2 50.0 40.1 36.3 40.0 28.6 51.1 69.6 72.4 68.9 61 .2 69.1 79.4 56.3 217 100 54 85 383 97 127 103 155 43 64 20 232 118 133 139 207 106 57 53 12..3 16.2 15.3 17.0 14.3 9.7 10.4 10.3 12.5 12.6 11.5 8.3 13 .6 6.4 9.4 8.6 14.0 18.5 '18.4 16.8 2662 1620 824 1449 5492 938 1321 1062 1939 541 739 166 3150 750 1254 1201 2888 1957 1051 892 79.5 65.5 75.0 90.8 75. 8 63.4 79.3 85 .2 64.0 66.7 70.9 57.2 68.7 67.2 71.0 70.9 78.4 61.3 73.1 80.7 3.0 3.4 5.6 0.9 2.9 2.7 6.7 7.7 4.7 2.2 4.0 7.5 6.1 13 .8 5.6 4.4 14.3 2.7 3.1 11 .7 18.5 9.5 10.1 10.5 13.0 6.0 12.0 7.4 12.5 9.4 5.6 7.6 24.1 11.1 9.5 16.7 34.2 13.5 12.5 21.7 27.7 19.0 44.4 9.4 13.0 18.0 16.0 10.3 27 .5 2 1.9 11.1 15.2 24.1 27.8 26.3 28.6 21.6 18.8 26.7 21.5 42.9 22.2 40.0 32.6 28.0 32.0 17.6 20.0 37.5 16.7 68 .2 34.5 50.0 90.5 51.8 20.0 59.4 65.6 33.3 24.6 28 .6 22.2 35.3 39. 1 44.0 40.0 64.7 32.5 31.3 66.7 [ s· ...... \I:> ~ t1 d "'0 "'t;j ~ ....... ~ c.o The University of Texas Publication 1959, and illustrates many of the types of crosses made each year. Tables 2-6 give earlier data and special comparisons discussed later. Figures 4 through 7 summarize the egg development tests from fertile females for all years. In these figures the females of each test are grouped into five catagories determined by the percent of their eggs that developed. The percentages in each range for 1959 are given in the last five columns of Table 1. These graphs give a picture of the effectiveness of the fertile females in producing adult offspring in the several types of tests. Since it is difficult to obtain satisfactory results with ananassae. inlaboratory tests, we have been unable to get a consistent very high egg develop­ment even in random matings and in crosses between populations, where most detrimental factors should be heterozygous. Therefore we regard 80 to 100% egg development as normal, understanding that part of the time the reduced percent development in this range is due to some detrimental genetic effect; the reduction to a level between 60 and 79% egg development we regard as due to the effect of one or two lethals or lethal equivalents (due to the summed effect of two or more detrimental factors which reduce viability either heterozygous or homozygous); 40 to 59% egg development represents the adverse summed TABLE 2 Fecundity: Variation in number of eggs laid per day by fertile females 1955 1956 195i 1958 1959 Average *Ponape (P ) 27.5 34.9 29.9 15.4 26.9 p 21.8 31.0 24.4 12.0 22.3 *Majuro (M ) 24.1 20.9 24.1 23.7 9.6 20.5 M 23.7 29.5 21.2 14.3 22.2 *Bikini (B) 29.9 22.9 30.4 26.0 27.3 B 27.7 30.1 20.0 25.9 *Rongelap (A) 17.9 28.2 26.0 22.4 10.2 20.9 A 24.9 17.2 13.5 18.5 M x P 13.6 15.6 21.8 13.3 16.1 A X M 31.4 22.2 22.6 17.0 13.3 21.3 M X B 29.9 23.9 15.4 20.2 22.4 P X A 26.3 16.6 25.9 10.8 19.9 B X P 33.9 23.2 19.8 25.6 B x A 33.3 30.3 34.9 38.5 34.3 **AM X AM 31.8 21.1 30.1 26.7 16.4 25.2 **BA X BA 22.9 27.8 31.4 22.4 26.1 **MB X MB 34.2 21.0 33.5 23.4 28.2 **BP X BP 18.9 27.5 24.5 23.6 **MP X MP 18.0 28.3 29.0 13.5 22.2 **PA X PA 20.9 31.1 32.2 19.4 25.9 AM X BP 23.6 24.2 23.4 23.7 BA X MP 17.7 24.9 22.4 21.7 MBXPA 16.6 25.2 18.8 20.2 MB x BA 32.8 22.1 31.8 23.2 27.5 AM X MB 31.8 23.2 35.4 23.3 28.4 *Stocks mated brother X sister (18) 23.9 Stocks mated at random ( 14) 22.2 All matings between stocks (24) 23.1 ••All F 1 X F 1 mated brother X sister (24) 25 .3 All F1 three-and four-way matings ( 17) 24.7 TABLE 3 Fertility: Variation in percent of fertile pairs Average 1955 1956 1957 1958 1959 *Ponape (P ) 67.8 56.5 66.9 59.1 62.6 p 71.4 58.6 55 .4 48.9 58.6 *Majuro (M) 53.5 54.9 46.3 47.4 63.4 53.1 M 60.0 51.2 59.3 77.1 61.9 *Bikini (B) 72.8 66.0 72.1 59.4 67.6 B 82.2 89.4 57.4 76.3 *Rongelap (A) 42.6 73.0 57.3 55.5 76.4 61.7 A 66.7 68.6 57.3 64.2 M x P 50.4 37.1 50.3 17.3 38.8 AXM 45.8 24.0 53.8 74.2 43.5 48.3 M x B 48.5 63.4 81.0 55.1 62.0 P xA 37.5 39.4 6.6 52.0 33.9 B x P 56.5 26.7 58.7 47.3 B x A 51.6 58.9 69.4 71.7 62.9 **AM x AM 68.4 55.6 72.6 65.1 66.4 65.6 **BAX BA 89.3 68 .9 80.8 46.3 71.3 **MB X MB 74.7 67.4 79.8 73.6 73.9 **BP X BP 45.8 54.5 43.7 48.0 **MP X MP 62.5 58.5 46.2 78.7 61.5 **PA X PA 68.4 30.5 75.0 61.0 58.7 AM x BP 63.5 61.1 74.1 66.2 BA x MP 44.8 29.7 34.1 36.2 MB X PA 38.6 68.1 68.6 58.4 MB X BA 74.7 87.2 84.5 82.2 82.2 AM XMB 51.0 45 .3 86.9 66.7 62.5 *Stocks mated brother X sister ( 18) 61 .0 Stocks mated at random (14) 65.3 All matings between stocks (24) 48.9 ••All F, X F , mated brother X sister (24) 63 .9 All F1 three-and four-way matings (17) 62.4 TABLE 4 Recessive fertility factors in D. ananassae F 2 ~ertility if F1 mating was: Stock Brother X sister N umber Percent Fertil e Steril e Fertil e Unrelated pairs umber Percent Fertile Sterile Fertile Change in ferti lity on in breeding Majuro 95 61 60.9 126 48 72.4 -11.5 Pomipe 71 91 43.8 84 64 56.8 -13.0 Rongelap Rongerik 90 37 91 41 49.7 47.4 118 73 67 92 63 .8 44.2 -17.1 + 3.2 AM X AM 87 41 68.0 94 32 74.6 - 6.6 All tests 380 325 53.9 495 303 62.0 - 8.1 effects on viability of two or three lethals or lethal equivalents; 20-39% egg development is the result of the summed adverse effects of three to five lethals or lethal equivalents; 1 to 19% egg development results when there are five or more lethals or lethal equivalents which become effective in the several genetic The University of Texas Publication TABLE 5 Gregg's laboratory test of changes in populations Year Stock Type test Type breeding Orig inal test Sample 1 Sample 2 1956 Rongerik pairs Inbred 68.2 64.0 60 .5 1956 Rongerik pairs Crossbred 79.5 75.3 1956 Bikini pairs Inbred 60.3 60.1 64.2 1956 Bikini pairs Crossbred 75.4 77.3 75.0 1957 Rongelap pairs Inbred 64.5 58.6 73.3 1957 Rongelap pairs Crossbred 83.6 70.3 82.7 1957 Bikini pairs Inbred 57.7 48.3 54.0 1957 Bikini pairs Crossbred 81.8 77.5 79.2 1956 Bikini cage 1 Inbred 60.3 57.4 59.7 1956 Bikini cage 1 Crossbred 75.4 72.0 81.7 1956 Bikini cage 2 Inbred 60.3 64.2 60.9 1956 Bikini cage 2 Crossbred 75.4 82.8 70.3 1955 Majuro cage 1 Inbred 66.4 64.9 61.0 1956 Majuro cage 1 Crossbred 80.8 78.0 87.8 1956 Majuro cage 2 Inbred 66.4 64.8 1956 M ajuro cage 2 Crossbred 80.8 89.1 1957 Bikini cag2 1 Inbred 57.7 59.4 1957 Bikini cage 1 Crossbred 81.8 74.5 86.2 1957 Bikini cage 2 Inbred 57.7 56.7 52.2 1957 Bikini cage 2 Crossbred 81.8 73.8 77.2 1957 Majuro cage 1 Inbred 63.7 63.4 58.5 1957 Majuro cage 1 Crossbred 74.9 62.6 76.8 combinations in the eggs. The net reduction by each additional lethal or lethal equivalent becomes less as their number increases-that is, around 25 % of the survivors are eliminated by each homozygous recessive lethal in so far as they are free to form independent combinations, plus some reduction in viability in the heterozygotes. Sometimes the failure of eggs to develop may be due to a factor or factors which influence adversely the ability of eggs or sperm to func­tion properly. Gregg (1959) carefully examined the adverse cumulative effects of more and more lethals and lethal equivalents. His two types of tests, cages and pair matings (the latter never brother x sister matings) , provide measures of the effectiveness of rather stringent selection in large populations in the cages or small population samplings in the vials on these components of natural selection under investiga­tion, viability, fertility and fecundity. Therefore his work (summarized in Table 5 ·and Figures 8, 9 and 10) is a control test measuring the effect of natural selec­tion and the decay of variability due to sampling in the laboratory which may be compared with the changes in the natural populations on the islands. An additional set of tests was run in 1959. A special test was made crossing siblings (brother x sister) , first cousins, and unrelated pairs, all F1 from the same matings of brothers to unrelated females from each locality. The results are given in Table 1, lines 15 through 34, and Figure 1,1 which summarizes the viability tests. In this special test, sets of several brothers, F1 of original island females, were mated in pairs to unrelated females (Table 1, lines 15, 19, 23, 27, TABLE 6 Viability: Variation in percent of egg development on inbreeding or random mating 1955 1956 1957 1958 1959 Average *Ponape (P ) 67.6 68.8 63.4 67.6 66.9 p 88.3 87.8 81 .5 80.0 84.4 •Ma;uro (M ) 56.6 66.4 63.7 50.3 61 .0 59.6 M 80.8 74.9 69.8 74.8 75.1 *Bikini (B) 46.4 60.3 57.7 65.1 57.4 B 75.4 81.8 . 74.7 77.3 *Rongelap (A) 40.8 78.9 64.5 61.2 62.8 61.6 A 83.6 73.5 69.1 75.4 M x P 58.2 70.1 77.1 80.4 71.5 A X M 64.3 84.6 81.8 71.1 79.4 76.2 M x B 67.8 75.4 79.4 73.6 74.1 P X A 84.9 83.8 63.9 81.2 78.5 B x P 72.7 86.1 73.8 77.5 B x A 70.2 76.5 73.6 60.0 70.1 **AM X AM 59.8 71.1 64.0 57.6 62.2 62.9 **BA X BA 54.2 70.3 57.0 42.0 55 .9 **MB X MB 62.5 58.6 64.6 57.4 60.8 **BP X BP 68.4 61.5 57.7 62.5 **MP X MP 67.4 60.1 53.9 62.9 61.1 **PA X PA 74.7 68.7 62.0 66.1 67.9 AM x BP 89.7 85.7 76.2 83.9 BA X MP 87.6 86.8 75 .9 83.4 MB X PA 91 .7 73.0 70.2 78.3 MB X BA 81.1 85.2 72.4 64.8 75.9 AM X MB 89.4 83.7 78.7 71.8 80.9 *Stocks mated brother X sister (18) 61.3 Stocks mated at random (14) 78.3 All matings between stocks (24) 74.6 **All F1 X F 1 matings brother X sister (24) 61.8 All F1 three-and four-way matings ( 17) 80.2 31). Their progeny were mated as sibling pairs (lines 16, ZO, 24, 28, ZZ), cousin pairs (lines 17, Z1, Z5, Z9, 33), and unrelated pairs (lines 18, ZZ, Z6, 30, 34). The estimate of lethals is a minimum one. Some, perhaps most, of the P1 females may have mated two or more times. Insofar as the F1 males used were half sibs, the values for lethal equivalents should be doubled. Therefore the true values are betwe.en the values given and double those values. Since in this article we are trying to summarize the results of the entire five years of study, we have in­cluded much of the data given in the figures and tables for earlier years in ad­dition to tests of material collected in the summer of 1959, together with a discussion of the whole study. Fecundity.-Table Z shows the variations in fecundity (measured a~ eggs laid per day by the fertile females) within the stocks, in crosses between them, and in the F1 tests. If we look at the strains from an island each year, we find some differences, but a major difference in 1959; in fact, in that year the fertile females averaged only about half as many eggs per day as those from the earlier years. There is no appreciable difference in average eggs per day between The University of Texas Publication PONAPE (P) MAJUR O (M) RONGELAP(A) RONG ERI K (K) BIKINI (B) 55 56 5 7 56 59 55 56 57 56 59 55 56 57 56 55 56 57 56 59 90 j 100 ~Jjj JJ,., J 0 60 j.,.j 81.8 j w 0.. g ~J.JJ 75.4 _J w "'...u 73.5 ~ > 70 ~_J j w ·o ~~J 6l6 68·8 J 67.6 ..tj 698 ... 66.4 * 64.5 J 63.4 ...JI 65.1 .I 63.7 63.6 ...... 61.2 62.8 60 ­ _L 61.0 60.4 60.3 ... 57.7 56.6 55.2 50 ­ 50.3 46.4 .. 40 ,__ 40.8 FIG. 4. Egg development within populations. The darkened figures are brother X sister matings; the plain figures are crosses between offspring of different pairs. These were run at the same time, with the progeny of pairs tested both ways when possible. The test years are indi­cated at the top for each population. The histogram shows the relative percentage of females with egg development of 1-19, Z0-39, 40-59, 60-79, and 80-100 percent from left to right. The fre­quency distribution of females which gave different effective egg developments gives a better picture of the effective productivity of the population or cross. The average percent development of all eggs laid by fertile females is given under each population histogram. The letter symbol for the particular island is indicated with the name. 100 90 0 60 w 0.. g ~ 70 w 0 ~ 60 50 40 MXP AXM 56 57 56 59 55 56 57 56 59 J.J j BIB j 794 ~~J J 70.1 aJill 71.1 64.3 5B.2 PXA 56 5 7 56 59 jJ j J Bl 2 63.9 MXB 55 56 57 58 ~J.J oJj] 754 736 6lB BXP BXA '55 56 57 58"J J B6 I j JJ.J 72 7 BB 70.2 J 60.0 FIG. 5. Distribution of classes and average egg development for crosses between flies from different populations. crosses within stocks, crosses between F1's or the average of all crosses between stocks. We would expect no inbreeding effect in these tests. This is true whether the bad year, 1959, is included or not. There were some special differences (see the summary of differences over the five years in Fig. 15). The four crosses between Majuro females and Ponape males averaged 16.1 eggs per day for the fertile females, which is lower than either population, while the four crosses between Bikini females and Rongelap males averaged 34.3 eggs per day, which AMXAM MPX MP PAX PA BA XBA MBX MB BP X B f' 55 56 5 7 58 59 56 57 58 59 56 57 5 8 59 55 56 57 58 55 56 57 58 56 5 7 58 100 a 80 w _. j a. '3 _.4 74.7 _I JI ~70 ~ • 70.3 1 JIw 71.1 J ..t 68.7 ja 68.4 J67.4 j J 66.1 j 640 J 62.9 0 6~2 ~1i~J 62.0 61.5 J 59.8 .i 60.1 .IL 58.6 57.7 57.6 57.4 57.0 53.9 52.4 50 40 42.0 FrG. 6. Distribution of classes and average egg development for brother X sister matings of F1 from crosses between populations. AM X BP BA X MP MB X PA MB X BA AM X MB 55 56 I 57 58 I I 56 I 57 I 58 II 57 1 58 55 56 I 57 I 58 I 56 I 57 I 58 56 100 J . ----" 90 - 91.7 0 JJ JJ w ,,.JJ a. 87.6 86.8 j] JJ J g 857 J] w 80 > 81.1 83.7 78.7 J w 0 76.2 75.9 ~ 70 71.8 72.4 J 70.2 '°J 64.8 ... - 60 FrG. 7. Distribution of classes and average egg development for three-way and four-way crosses between different F 1's. is higher than either parent population. All four years showed these differences so we must conclude that the stimuli for the fertile females to lay eggs in these two tests were different. In the following discussion the standard abbreviations are: B =Bikini population; P = Ponape; M =Majuro; A= Rongelap; K = Rongerik. The other tests fell much closer to the mean, B x P (3 tests) = 25.6, M X B (4 tests)= 22.4, PX A (4 tests)= 19.9 and A x M (5 tests)= 21.3. All other crosses fall within the range of the parent populations. Both inbreeding and random mating values are given (e.g. Bi and Be on Figures 15, 16, 17) . Since the P, Mand B females were tested twice each year and each was normal in the other cross, this adds validity to our claim of specific interaction in the insemi­nated females. In all these averages WP. are weighing each cross equally. Most The University of Texas Publication .....------I 9 5 6 -----...-------I 9 5 7--------.. BIKINI RONGER IK BIKINI RONGELAP Origin al ISa mple 'Sa mple Originad Sample ISam ple Ori gin al ISamp le ISample Original,Sample !Sample Test l 2 Test I l 2 Test l 2 Test 1 2 I II I 4 8.3 FIG. 8. Gregg's laboratory control population tests using pair matings. Histograms as in Figure 4. See text for sampling procedure, and reference. individual tests involved 1,000 to 10,000 eggs, although a few were below a thousand (lowest 375) and some involved more than 10,000 (highest 21 ,464 for Bilcini inbred in 1955) . The variations do not depend on the stocks alone for in the high year, 1957, Bikini averaged 30.3, Rongelap 25.5, Majuro 26.8 and Ponape 33.0 eggs per day, and the average values for all years are similar but slightly lower (Table 2). We conclude that environmental factors influence egg laying by fertile females ( 1959 versus other years), that the stocks differ ge­netically in their egg laying pattern, and that the effectiveness of stimuli to lay may be quite different within stocks and in crosses between the several stocks from different islands. Figure 12 shows the fluctuations of the stocks inbred or random mated during the five year period. Fertility.-The percent fertility of each of the crosses for 1959 is given in Table 1. In addition Table 3 presents the data for tests of fertility in most crosses each year including inbreeding or random mating within the stocks (32 crosses for all years), crosses between the stocks (24) and all F1 tests (both crossbred and inbred, 41). We would expect no inbreeding effect in these tests. The special test for inbreeding effect on fertility is given in Table 4. The lowest average fertility of stocks but not all tests that year was found in 1955 and highest in 1956, so the laboratory environmental factor (s) which reduced fecundity in 1959 did not affect fertility to the same degree. Here again we find that tests between members of the same island population and between F1 heterozygotes from crosses between populations of different islands are about alike, giving an average of approximately 63 percent fertile pairs for these tests. The crosses 100 090 w a.. 0 _J ~80 w Q ~ 0 70 FIG. 9. Gregg's laboratory control tests in cages (as Figure 8). between stocks are quite different with an average of about 49 percent fertile pairs. The several crosses differ widely in percent fertile pairs: BX P (3 tests) = 47.3%; M x B (4) = 62.0%; M x P (4) = 38.8%; Bx A (4) = 62.9%; P x A (4) = 33.9%; AX M (5) = 48.3%. There are again differences between the stocks, with Bikini the most fertile and Ponape the second most fertile, similar to their respective positions in fecundity for the comparable years, 1956, 57 and 58. There is no general heterosis for fertility just as there was none for fecundity. Furthermore, some crosses, especially those involving Ponape, show decided incompatibility in fertility as in fecundity. This is shown in their high average cross sterility, and is represented on Figure 16. The fluctuations in the stocks year by year are shown on Figure 13. Table 4 shows the effect of an additional recessive component which affects fertility. The tests summarized in Table 3 demonstrated the dominant genetic components of fertility and cross fertility, modified by environmental effects. A special test in 1959 measured the fertility of F 2 flies when their F1 parents were mated brother x sister or were matings of unrelated flies. For Majuro, Ponape and Rongelap and Rongelap-Majuro heterozygotes, there was a recessive com­ponent which increased sterility when their parents were inbred. Rongerik, the very small population, showed no such effect but the overall test was clearly The University of Texas Publiccrtion 100 0 w a.. 90 g w > w 0 00 ~ 0 60 Frc. 10. Gregg's laboratory control tests in cages (as Figure 8). significant. Unfortunately, we do not know how much this varied from year to year. It should be pointed out that there are differences between the tests of fecun­dity, fertility and viability. The fecundity tests and the fertility tests summarized in Table 3 measure only the dominant components of the effects of the genes influencing these characters. Table 4 is a different test which allows us to de­termine the dominant effects of factors on fertility and in addition some of the effects of homozygous recessive factors on fertility. On the other hand the viability tests regularly compare dominant gene effects on viability (random mating and crossing) and dominant plus recessive gene effects on viability (inbreeding, brother X sister, or cousins). These differences must be kept in mind when com­paring different measured components of evolutionary fitness. p M A K AM 100 95 a w Cl. '.3 80 w > w a FIG. 11. Special tests to measure the effect of various degrees of inbreeding. Cross lines, brothers X unrelated females; solid, F1 inbred brother X sister; stippled, F1 inbred cousins; plain, F1 unrelated pairs. See Table 1, lines 15-34 and text. Viability.-The tests for viability (egg development) are summarized in Figures 4 through 7, 14 and 17 with aditional special tests in Figures 8 through 11. Details of the tests are to be found in Table 1 for the 1959 collections. Before discussing these tests it is well to examine Gregg's laboratory control tests (Gregg, 1959) as a background to evaluate our tests of populations immediately after collection from the field. He tested samples of the populations collected in 1956 on Rongerik, Bikini and Majuro and those collected in 1957 on Rongelap, Bikini and Majuro. The tests were of two types each year. In 1956 the populations to be tested were built up several generations, then 500 ~ ~ and 500 2 2 of the Ronge­rik population sample were mated in pairs, taking care not to make brother X sister matings. A similar test was run with Bikini flies. The next generation 500 pairs were made, taking care not to use disproportionally high numbers from especially fertile pairs, etc., making sure in each generation that no sib mating occurred. The tests in cages, which were set up in duplicate, were made with several thousand F2 flies from the Bikini (or Majuro) population sample. The population cages were set up after the manner of Wright and Dobzhansky (1946). The 15 food cups were changed 3 at a time so no cup remained over 15 days. These tests were run at 77 ± 1 °F, which is close to the mean morning temperature of the Proving Ground islands. Sixty vials from the pair mating experiments were tested by inbreeding F1 brother X sister, or between males and females taken from different vials (relatively unrelated) taking care to test The University of Texas Publication 1955 1956 1957 1958 1959 Frn. 12. Graphs showing variations in eggs laid per day by fertile females as a measure of fecundity of the several island populations when mated at random or brother X sister for 1955­1959. males and females from each vial to females and males of different other vials for the random matings. For the cages, two food cups were removed after being in the cages 48 hours and the contents were spread into several bottles so larval competition would be at a minimum. The flies that emerged were pair mated and their F1 progeny mated brother x sister or at random. Records were kept of the eggs laid per day from fertile females (fecundity), the percent of pairs fertile (fertility), and the percent of eggs to develop into adults (viability) . The last is illustrated in Figures 8, 9 and 10 which show the usual sets of egg develop­ment classes. The general results of Gregg's tests can be summarized easily. For Bikini the average percent fertility for the 23 tests used by Gregg was 83.4% which was 30 20 10 >­ ­ I­.....J I­ a: w 50 LL I- z w u 80 a: w a... 70 60 50 / ,.0\ ~· / I I \ I \ I \ \ \ I f 0----Ponape 6--Majuro D········· Rongelap 0·-·-·-Bikini e AVERAGES 0 J The University of Texas Publication higher than the tests of populations, and for the 10 tests with Majuro it was 75.7%, which is about the same (see Table 3). The number of eggs laid per day was lower for Bikini than in the general population tests (see Table 2), especially in 1957. The 23 tests with Bikini averaged 22.2 eggs per day; the 10 tests with Majuro, 22.1. The average is not low for Majuro. The viability profiles, based upon the percent of the females falling in the five different egg development categories, and the average egg development in each test are shown in Table 5 and Figures 8 through 10. Table 5 gives good examples of response to random and perhaps laboratory environmental fluctuations. The vials and cages can be discussed together although there was a higher proportion of comparatively large changes in the percent development (maximum 14.7) in the vial series. The large number of changes from the original test to the first pair test or cage sample resulted in a reduction in viability but the average change between Gregg's first and second sample was an increase in viability. The values are: pairs, first sample, -5.1 % ; cages, -2.6%; pairs, second sample, + 4.6%; cages,+ 2.0%; overall pairs, -0.4%; cages, -0.6%. The largest overall change through both samples in the pair tests was+ 8.8% (Rongelap, 1957), and in the cage tests was + 7.0% (Majuro, cage 1, crossbred, 1956). It is conceivable that such fluctuations might have led to the large changes of the irradiated popula­ tions, Bikini and Rongelap, between 1955 and 1956 but it seems improbable. Most of the viability tests with the populations collected in the summers of 1955 through 1959 are included in Figures 4, 5, 6, 7, 11, 14 and 17. The tests in 1959 are given in Table 1 which illustrates the types of information obtained. Most of the tests carried out from 1955 through 1959 are summarized for fecun­dity, fertility and viability in Tables 2, 3 and 6. There were some additional tests in 1955and1956, which were discussed earlier (Stone, et al, 1957). To analyze the viability data for the whole period, it is first necessary to scan the data in Table 6 and Figures 4-7 and 14 to see if there are obvious effects of laboratory conditions. There are the following rather consistent results which must reflect the effect of variations in laboratory conditions. Letting 1956 repre­sent the optimum laboratory conditions because the viability tests were best that year, we see that there was a drop in the averages of the 1957 tests, an appreci­able further drop in viability average in 1958 and some recovery in 1959. This is true not only of the stocks, Figure 14, but of crosses between them and their F1 tests (see Figures 4 through 7). Ponape, which is consistently the most viable on inbreeding or random mating, shows this relation and so does Majuro, which is less viable and seems to give a pronounced response to environmental fluctua­tions. If we use Majuro as a basis for determining adverse laboratory environ­ment, 1955 was not as bad a year as 1958, judged by inbreeding or random matings within the stock. Parenthetically, we are not including the tests of the populations on the Rongerik atoll shown in Figure 4 (Eniwetok island in 1955, Rongerik island in 1956, 57, 58 and 59); these were such small populations that we could not make the other regular tests. They show the same trend as the stocks from other islands. Rongerik is included in the special information given in Figure 11 and Table 1. In 1958, Majuro is the least viable stock on inbreeding and crossbreeding, with the viabilities of the Bikini and Rongelap populations appreciably better. Therefore the very low viabilities of Bikini and Rongelap 1955 1956 1957 1958 1959 90 80 70 0 w 60 a.. 0 _j w > w 0 80 I- z w u 0::: 70 w a.. 60 50 40 0-----­ A v, ' .... .... , '-,',<:>----------(> ··n RANDOM 0----Ponape 8--Majuro 0-········ Rongelap 0--·-·-Bikini --AVERAGES 0 ,,,0 ..... , ................ D _......... ~ ),A ,,,.."'.,.. ···· .(, v \ /' "·O-··""" , "'\ \ \ / \ / 'd' INBRED FIG. 14. Graphs showing differences and variations (from 1955-1959) in egg development between random matings and brother X sister matings. The University of Texas Publication inbred in 1955 are not accounted for by a bad laboratory environment and prob­ably represent the effect of a large accumulation of lethal and detrimental muta­tions, but not major chromosomal abnormalities, due to the very heavy fallout in March, 1954. Greenberg and Crow ( 1960) give a method for estimating the number of lethals and/ or lethal equivalent gene complexes per gamete in Drosophila. We have calculated the number of lethals and/or lethal equivalents per gamete from some of these Drosophila data, Table 7. Most of the estimates were obtained by TABLE 7 Estimate of lethal equivalents per gamete Population Random Inbred (sib matings) Lethal equivalents per gamete 1955 (B X M ) = M 67.8 (43 ) 56.6 (139) 0.716 (A X M ) = A 64.3 ( 48) 40.8 (90) 1.820 (B X A ) = B 70.2 (39) 46.4 (297) 1.657 1956 (M X B) = M 75.4 (76) 66.4 (217) 0.509· (A X M ) = A 84.6 (26) 78.9 (263) 0.279 (B X A ) = B 76.5 (66) 60.3 (256) 0.952 1957 (M X B) = M 79.4 (88) 63.7 (86) 0.881 (A X M ) = A 81.8 (71 ) 64.5 (79 ) 0.951 (B X A ) = B 73.6 (93 ) 57. 7 (218) 0.974 1956 Ponape (P ) 88.3 (47 ) 67.6 (339) 1.069 M 80.8 (55 ) 66.4 (21 7) 0.785 B 75.4 (34) 60.3 (256) 0.894 1957 p 87.8 (50) 68 .8 (66 ) 0.976 M 74.9 (57) 63.7 (86) 0.648 A 83.6 (63 ) 64.5 (79 ) 1.038 B 81.8 (101 ) 57.7 (218) 1.396 1958 p 81.5 (68) 63.4 (71 ) 1.005 M 69.8 (55) 50.3 (29) 1.311 A 73.5 (82) 61.2 (47) 0.733 B 74.7 (84) 65.1 (93 ) 0.550 1959 p 80.0 (111 ) 67.6 (113) 0.674 M 74.8 (155) 61.0 (121) 0.816 A 69.1 (134) 62.8 (146) 0.382 Hongerik (K) 64.6 (91 ) 60.4 (80) 0.269 TABLE 7. These estimates were madz by the method used by Greenberg and Crow (1960). The lethal equivalents per gamete are four times loge Random (% survivors from random mating)-loge Inbred (% survivors from brother X sister matings) as only ¥.i of the loci are made homozygous by sib matings (the value for cousin matings is 16 loge R-log.I, where Rand I are the percent survivors of randomly mated and inbred groups). In 1955 no random matings were made, and none with the Rongelap population in 1956. Therefore we calculated the number comparing egg development of brother X sister matings with that from random mating between members of the same population. The percent egg development, with the number of fertile pairs tested in parenthesis, is given for each cross, along with the estimate of lethals and/ or lethal equivalents. There were no random matings within a population in 1955. Therefore the three crosses indicated were used to estimate B, A, and M random, which were then used with the data on inbreeding to calculate the lethals per gamete. Figure 17 suggests that this procedure gave a good estimate of the random mating values, although the Bikini estimate might be a little more in error than the others. The 1956 and 1957 data were calculated in the same way and these three sets are given first in the table, then the regular comparisons of inbred and random matings are given FECUNDITY 15 20 25 30 35 B,A B,M B,P A,M A,P I I I I ]85 I I f0.9 I I I I 26.l 25.91 27.3 I I p43 I Ac Ai 22.4 20.5, 22 ~i Bel Bi BAxBA ~59 t73 f8,2 Bx A Mi Me l MxB f2 .3 Be Bi MBxMB 25.9 27.3 23.6 25~ i 26.? ! 18.5 I Pc 20.9 20.5 !21.3 22.2 bPxBP iI pj I BxP Be Bi 25.2 Ac 18.5 Mi iAxM I Ai Mc 19.9 20.9 22."3 AMxAM 2~.9 f6 9 ]6.l Ac PxA Ai Pc 22 .3 2p.5 22.21/ I PAxPA Pi f6.9 I I MxP 1 Mi MJ i MPxMP1 Pc I I I I Pi I I I I I M,P Frc. 15. Summary of the fecundity shown as line graphs. The figures are the average values for the four or five years of testing of the several island populations mated at random (e. g. Be) or brother X sister (e.g. Bi), the crosses between them (e.g. B X A), and their F1 inbred brother X sister (e.g. BA X BA). of lethals in 1955 using values for random mating from the crosses indicated~ The values for 1956 and 1957 are given from both types of calculations (using crosses for random versus using random). These agree fairly well. In 1959 there was a special test of egg development from random, cousin and brother X sister matings. The agreement for Majuro was very good, 1.182 lethal equivalents per gamete com­paring random with the brother X sister matings and 1.148 lethal equivalents per gamete com­paring random with cousin matings. The other tests were not satisfactory because of small samples, etc. Letter symbols for island populations (M, A, B, K, P) are the same as in previous tables. The University of Texas Publication FERTILITY FIG. 16. Summary of percentage fertility of pair matings shown as line graphs. The figures are the average values for the four or five years of tests. Values are given for random and brother X sister matings within the island populations, their crosses, and their F1 inbred brother X sister. for the last four years (1956-9). Direct comparisons show that the estimates of lethal equivalents are close together for the calculations made two ways in 1956 and 1957, therefore we have more confidence in the 1955 figures. The 1959 special test fell one laboratory generation after the other tests. The conclusions to be drawn from this method of determining lethals and/ or lethal equivalents per gamete agree with those arrived at in other ways. The data follow the same pattern as those in Table 6 for if we sum all tests each year, then 1955 had the low egg development, implying more lethals which is in agreement with the calculations in Table 7; 1956 and 1957 are high and close together with fewer lethals, 1958 is low in both and 1959 is intermediate. Table 7 indicates a larger number of lethals in 1955 than does Table 6 but the Table 7 measure would be closer to the correct evaluation for the stocks. All stocks and crosses showed a decrease in egg development in 1958 except Bikini, and that same relation shows in Table 7. We ascribe a larger proportion of the reduction in egg development in 1958 to an adverse laboratory environment. If Bikini is affected in the same way as the others by bad lab environment, then the increase in normal egg de­velopment from loss of detrimental factors from the Bikini population must have been appreciable between 1957 and 1958. Another way to compare the changes in value of lethal frequencies year to year is to determine the average mean and see how the values fluctuate. These data show Ponape to have fewer lethals and to vary in lethal frequency less than VIABILITY 50 60 70 80 90 B,A B,M B,P A,M Frc. 17. Summary of percentage viability of eggs from fer;tile females shown as lin_e graphs. The figures are the average values for the four or five years of tests. Percentages of egg develop­ment are given for random and brother X sister matings within the island populations, their crosses, and their F 1 inbred brother X sister. the three atoll island populations. 'In fact the largest fluctuation is in the reduction in number of lethals in the 1957 sample. The means of the three atoll island populations are close to each other but Majuro shows three years better and two worse than its mean. The fluctuations around the mean are uniform, equally up or down. Both Rongelap and Bikini present a different picture. The number of lethals is much higher than the mean in 1955, but the numbers for 1956 through 1959 are all below the mean, despite the fact that in 1958 the laboratory condi­tions were adverse for an.anassae. In fact the number of lethals fell each year at Bikini, although 1956 and 1957 were nearly alike in"these calculations. This suggests that the numbers of lethals were very high in these two populations in 1955, presumably due to the radiation, and that these populations otherwise usually had somewhat fewer lethals than Majuro. However another alternative seems more probable. In addition to the detrimental mutants due to radiation, we find that the smaller atoll populations are more influenced by environmental fluctuations. Ponape with its much larger total population and its local popula­tions fluctuating and moving with the fruit supply has been able to establish a more stable genotype. These two differences, radiation and the net effect of selection on different size populations, would account for the results of these viability tests. Th.e University of Texas Publication 1 / z ' 0 ;::: ' ' Percent O Percent Species Chromosome lethal sterile sterile subvital willistoni second Z8.4--41.2 40.5 64.8 third 25.6-3Z.8 40.5 66.7 pseudoobscura second Zl.3-33.0 10.6 8.3 93.5 third Z5.0 13.6 10.5 41.3 fourth Z5.9 4.3 11.8 95.4 persimilis second Z5.5 18.3 13.2 84.4 third ZZ.7 14.3 15.7 74.Z fourth 28.1 18.3 8.4 98.4 prosaltans second 32.6 9.2 11.0 33.4 third 9.5 6.6 4.2 14.5 pseudoobscura second 21.5-25.6 78.0 (Spassky et al.) third 16.4--2Z.9 61.0 The University of Texas Publication genes it carries. For example, the second chromosomes of willistoni and prosaltans are V's formed by the fusion of two rods, whereas all other chromosomes listed in the table are rods (single, major autosomal elements). The different species, and even different chromosomes of the same size within a species, may differ markedly. Further, the chromosomes of persimilis, those of pseudoobscura, and the prosaltans second chromosome carry as many lethals as does willistoni from some localities, but this is not so for sterility factors. To understand part of the reasons for these differences it must be remembered that willistoni is the most successful Drosophila in the Neotropical region, with tremendous populations across South America extending up into Central America and the Antilles. Drosophila pseudoobscura and persimilis are very successful species with large populations in western North America. On the other hand prosaltans is a much more restricted or specialized species with relatively small scattered populations. We must conclude that many factors, including the net disadvantageous domi­nant component of the effects of the mutation, the genome complex of the species which determines its adaptability and vigor, and the population size and struc­ture, among other things, must all influence the frequencies of disadvantageous mutations which may be present on the chromosomes in populations of different species. Drosophila ananassae is a member of the same species group in the subgenus Sophophora as Drosophila melanogaster. Several investigators have made studies of the incidence of lethal and semilethal mutants in different populations, and the spontaneous mutation rate to lethals. Dobzhansky and Wright ( 1941) and Wright, Dobzhansky and Hovanitz (1942) studied these problems using D. pseu­doobscura, also a member of the subgenus Sophophora. They determined the mutation rate of several stocks and the incidence of mutation for pseudoobscura populations from several localities. They concluded that the frequency of lethals is lower than would be expected if the lethals are recessive and the populations are large and breeding at random. Therefore there must be some inbreeding or there must be a selectively disadvantageous dominant component of the action of heterozygous lethals or both. The situation with respect to rate of mutation and incidence of lethal mutants in populations of melanogaster resembles that in pseudoobscura. The frequencies of lethals (or lethals plus semilethals) and their changes through time at a number of localities are shown in Table 9, together with the investigator whose work is cited. There is a tremendous difference be­tween the frequencies of lethals in different localities, ranging from 5.6% m T ABLE 9 Lethals and semilethals in populations of Drosophila melanogaster P ercent Percent lethals + Source reference Locality Date of test lethals semilethals Dubinin (1946) Gelendzhik, USSR July, 1933 7.9 July, 1934 12.6 Simferopol, USSR July, 1938 20.6 October, 1938 29.2 October, 1939 24.1 Uman, USSR September, 1937 22.6 Wisconsin (Hiraizumi and Crow, 1960) to 44.1 % in Florida (Ives, 1945, 1954), and for lethals plus semilethals, from 7.5% in Korea (Paik, 1960) to 67.0% in Florida (Ives, op. cit.). The three populations in Korea had lethal frequencies July, 1938 21.6 September, 1938 29.8 Kislovodsk, USSR July, 1932 9.8 Mashuk, USSR July, 1932 8.8 Essentuki, USSR July, 1932 10.2 Vladikavkaz, USSR July, 1932 19.9 Batumi, USSR July, 1932 13.9 Piatigorsk, USSR July, 1932 9.8 Erevan, USSR July, 1932 15.7 Armavir, USSR July, 1932 21.4 Sochi, USSR October, 1939 26.0 Kutaisi, USSR November, 1939 38.9 Ives (1945; 1954) So. Amherst, September and Mass., USA October, 1938 31.8 45.0 Sept., Oct., 1941 59.3 Sept., Oct., 1945 45.8 Sept., Oct., 1946 49.3 Sept., Oct., 1947 39.3 Sept., Oct., 1948 37.0 Sept., Oct., 1949 30.3 Sept., Oct., 1950 31.9 Sept., Oct., 1951 37.4 Sept., Oct., 1952 37.9 Winter Park, Florida, USA 1940 44.1 67.0 1942 61.8 195 1 51.1 New York, USA 1950 32.3 Cannonsburg, Pa., USA 1952 28.2 Wooster, Ohio, USA 1951 36.2 Lincoln, Neb., USA 1952 25.6 Pullman, Wash., USA 1951 39.1 Austin, Texas, USA 1952 41.8 Goldschmidt et al. (1955) Qiryat Anavim, Israel Autumn, 1951 30.6 40.1 Spring, 1952 30.3 34.7 Summer, 1952 28.4 35 .1 Autumn, 1952 33.6 42.3 Winter, 1953 37.2 41.5 Spring, 1953 34.8 41.9 Paik (1960) Na;oo, Korea September, 1957 10.5 13.5 August, 1958 6.5 7.9 Quilport, Korea September, 1957 7.5 7.5 Taegoo, Korea October, 1957 9.7 14.9 Hiraizumi and Crow (1960) Madison, Wis., USA Sept. 13, 22, 1957 8.2 24.7 Oct. 2, 13, 14, 1957 5.6 27.0 October 17, 1957 10.8 28.8 October 19, 195 7 8.8 27.4 October 20, 1957 6.5 29.2 October 29, 1957 19. 8 37.2 The University of Texas Publication from 6.5 to 10.5, the lowest range in the group; the largest range was reported from the U.S.A., from 5.6% to 44.1 %. Dubinin (1946) reported a similar range in lethal frequency, from 7.9% to 38.9%, in the U.S.S.R. The lethal plus semi­lethal classes summed ranged from 24.7% to 67.0% in the United States. The changes through time are of marked interest for analysis of our Pacific Proving Ground studies. The most extensive short time changes reported by Dubinin for lethals are at Uman, U.S.S.R., of 8.2% (21.6% to 29.8%) from July to Septem­ber 1938, and of 8.6% (20.6% to 29.2%) from July to October 1938, at Sim­feropol, U.S.S.R. The most extensive change reported from Qiryat Anarim, Israel, by Goldschmidt, Wahrman, Ledemann-Klein and Weiss (1955) was 8.8% in the 6 months from summer, 1952, to winter, 1953. Goldschmidt and her col­leagues stated that the Israel populations fluctuated around their average of 32.2%. Presumably each population from the U.S.S.R. may have fluctuated around the average for that population although the rapidity of change in a few months was quite remarkable. Hiraizumi and Crow (1960) found a change in percentage lethals plus semilethals from 24.7% to 37.2% between September 13, 1957 and October 29, 1957. The largest change was between October 20th and 29th, from 29.2% to 37.2%. We do not know if this change would have led to a change in level of lethals in the population such as that found in Massachu­setts (Ives, 1954) or if it would have varied around a mean as Goldschmidt and her collaborators reported, although this seems more probable. The longest record of fluctuations in percentage of lethals plus semilethals is that from Massachu­setts reported by Ives. The four recorded years between 1938 to 1946 fluctuated aro:m'.1 48% le~hals. There was a drop of 10% in frequency of lethals between the colle::t:on.s in 194·6 and those in 1947. From 1947 to 1952 the six years col­lections fluctuated around 36%. The change from 1946 to 1947 is suggestive of some special adjustment as Ives remarked while the general fluctuations resemble ~ h'.)::;e of Ponape among the populations we studied. One of the important problems concerns the kinds of mutations and chromo­some rearrangements that occur spontaneously and that are produced by radia­tions. Dominant lethals present no problem to the surviving populations unless they could kill such a large portion of the population that the survivors cannot retain reproductive potential and the genetic flexibility necessary to survive. Chromosome abnormalities are usually rapidly eliminated. Some paracentric inversions may persist if they have selective advantage but few were found even on Bikini. There remain both detrimental and beneficial mutations. Completely recessive mutations that are lethal, detrimental or favorable in the genotype as tested against the existing environment will be selected for or against only wmn homozygous. This may occur in these island populations, at least when the popu­lation becomes distributed as very small and local pockets during unfavorable conditions in these limited environments. Favorable or unfavorable mutations that have a dominant component (for example, recessive lethals which when heterozygous have some disadvantage or an advantage) will be selected for or against as long as heterozygosity persists. Stern and Novitski (1948) and Stern, Carson, Kinst, Novitski and Uphoff (1952) tested 36 spontaneous sex-linked lethals and 39 sex-linked lethals in D. melanogaster that occurred in sperm which had received about 50 roentgens of gamma radiation; part of this latter group were of spontaneous origin. None of the 75 lethals produced any phenotypic effect when heterozygous. The group of 75 as a whole reduced the heterozygote viability, compared to a similar heterozygote lacking the lethal, by about 4%, the average viability index being .961, with a range from .602 to 1.312. There­fore the majority were detrimental to the heterozygote but a minority were heterotic and increased the viability of the heterozygote. Muller and Campbell (Muller, 1950) studied autosomal lethals produced by ultraviolet irradiation. These also reduced the viability of the heterozygote. Goldschmidt and Falk ( 1959) compared, as heterzygotes, second chromosomes free of lethals to those carrying spontaneous lethals collected from a wild population. Most heterozygous lethals had a dominant effect reducing viability but not all; one had an excellent combining ability with the Cy L chromosome in such a way that the combina­tion, having at least one lethal in each homologue, gave heterosis. Hiraizumi and Crow (1960) showed that this dominant disadvantageous component of most lethals, even those collected from wild populations, not only affects viability but also fecundity as measured by the eggs laid the first 23 days from emergence (the important period under natural conditions), as well as the rate of develop­ment which was slower with heterozygous lethals; there was also an effect on the longevity of males but not of females. The adaptive disadvantage of heterozygotes for recessive lethals is not re­stricted to D. melanogaster but is also found for lethals present in wild popula­tions of other Drosophila, for example D. willistoni (Cordeiro, 1952; Prout, 1952). Furthermore, da Cunha, Toledo, Pavan, de Souza, Pires de Camargo and de Mello (1958) and da Cunha, de Toledo, Pavan, de Souza, Melara, Gabrusewycz, Gama, Pires de Camargo and de Mello (1959) found that both spontaneous and X-ray induced lethals in willistoni seem to reduce the viability of the hetero­zygote about the same amount. The frequency of lethals is too low in natural populations to be accounted for by the elimination of homozygotes if the lethals were completely recessive and mating was random; da Cunha et al do not believe that inbreeding is frequent enough to account for the reduced incidence of muta­tion and ascribe most of the reduction effect to selection against the dominant detrimental component of the lethal. Dobzhansky and his colleagues (B. Spassky, N. Spassky, 0. Pavlovsky, M. G. Krimbas, C. Krimbas and Th. Dobzhansky, 1960; Th. Dobzhansky, C. Krimbas, M. G. Krimbas, 1960) have recently studied the magnitude and characteristics of the genetic load in D. pseudoobscura. The frequencies of lethals and subvitals (Spassky et al, 1960) is given in Table 8. In contrast to the results of the several independent investigations of melanogaster and willistoni which demonstrated a dominant detrimental component of the majority of recessive lethals (others were without effect or heterotic), Dobzhansky, Krimbas and Krimbas (1960) did not find such a dominant detrimental effect of heterozygous lethals in pseudo­obscura. All classes of lethals and detrimentals and the normal classes were on the average normal or slightly heterotic. The supervitals, the classes where homozygous viability was above normal, were decidedly heterot!c. Insofar as these are beneficial and heterotic in natural populations, they are an important group of mutations. From the viewpoint of chemical genetics such beneficial mutations would most often be expected to produce an active useful enzyme. The University of Texas Publication The relative concentration of the substrate and enzyme-:-which genie balance studies show are often determined by the frequency of the allele--would de­termine the dominance and heterotic nature of such a mutation. This would seem to be the type of mutant most readily modified to dominance by the selec­tion of modifiers (Fisher, 1930, revised ed. 1958), perhaps by selecting among alternative isoalleles present at loci that influence substrate concentration, co­factor concentration or other mechanisms whereby genes determine the rapidity and extent of cellular activity. Dobzhansky, Pavlovsky, B. Spassky and N. Spassky (1955) and Dobzhansky and Levene (1955) have shown that supervital, normal and subvital chromo­somes or gene complexes vary in their response to differences in environm11ntal conditions so as to change their viability markedly, both in homozygous and in different heterozygous combinations with and between classes. In general the heterozygotes are more homeostatic and respond favorably even under adverse conditions, and this is true even with the heterozygotes for subvital factors. Therefore we cannot regard supervitals as necessarily new beneficial mutations, but the fact that they are supervitals in some environments might favor modi­fiers which in turn increased their viability in other environments. Dobzhansky, Krimbas and Krimbas (op. cit.) could not determine from their data whether the genetic load of pseudoobscura populations was mutational (due to recurrent mutation to detrimental alleles which were selected against) or balanced and retained by natural selection (due to alleles which were beneficial heterozygous but detrimental homozygous) or to both types of mutants. Wallace and his coworkers have made the most extensive tests of irradiated Drosophi.la melanogaster populations in the laboratory. Wallace (1956) dis­cussed most of this work which demonstrated many interesting aspects of ir­radiated populations kept under the stringent selection of population cages. The experimental populations as set up contained 14 second chromosomes free of lethals and semilethals derived from the Oregon-R strain. The other chromo­somes of the genome were mixtures from Oregon-Rand the marked stock used to isolate these second chromosomes. All five original stocks and three derived stocks were kept in population cages where the populations were around 10,000 flies except two, numbers 5 and 17 (a derivative of 5) which were kept in cages of about 1,000 adult flies. The original five populations (1, 3, 5, 6, 7) were started from descendants of the stock established with the 14 lethal-free second chromo­somes, and had no (or very few) spontaneous lethals at the beginning of the experiment. Number 3, the large control population, received no radiation. A number of spontaneous lethals were accumulated in about 45 generations to a level where about 25% of the second chromosomes carried lethals and remained at about that level through 125 generations (Wallace, 1956, Fig. 5). Population 1 (large) which was established at the same time was irradiated once, the fe­males receiving 1,000 r, the males 7,000 r; this population started with about 20% of the second chromosomes carrying radiation-induced lethals, and the level of lethals remained at about that level for the 125 generations plotted. Populations 5 (small), 6 (large) and 7 (large) were started about 20 genera­tions after the first two. All three were subjected to chronic gamma irradiation from a radium source, 5 and 6 receiving 5.1 r per ho1<1r and 7 only 0.9 r per hour. This last population, number 7, showed an increase in the number of second chromosomes carrying a lethal from 0 to approximately 30% in about 50 genera­tions, where it remained. It is difficult to be sure that this population was much above the control but both were definitely above number 1 from its 60th genera­tion. Both 5 and 6 showed an increase in frequency of lethal-bearing second chromosomes from 0 to 80% in about 60 generations and remained around ~his level. Although each received 5.1 r per hour, the small population (5) showed much larger fluctuations in percentage of lethals present on the second chromo­somes all through the test. In fact it often fell 20 to 30% below population 6 in frequency of lethal chromosomes. After receiving about 250 kiloroentgens of irradiation, subpopulations of 5 ( 17, small and 18, large) and of 6 ( 19, large) were set up free of radiation and sampled through 25 generations. The second chromosome lethal frequency of population 17 fell from 85.7% to 56.2% in 24 generations, and that of 18 to 28.0% in 25 generations. However, population 19 showed little reduction in frequency of second chromosomes carrying lethals; during 24 generations with­out radiation, the lethal level fell only from 79. 7 to 76.7%. At about the 130th generation Wallace and his collaborators analyzed adult males from population 6 to determine the extent of lethals in both the second and third chromosomes. Of the 140 second chromosomes tested, 87% were lethal when homozygous; 92% of all third chromosomes were lethal. Furthermore, the frequency of combinations of normal and lethal chromosomes fit the possible combinations of the second and third based on the Hardy-Weinburg equation very well, although there is a slight excess of normal/lethal heterozygotes. It is obvious from this work that populations can build up a high frequency of lethal factors (perhaps three times the normal control frequency in each tested chromosome) if they are subjected to radiation over a number of genera­tions. The level of radiation necessary for lethals to reach this high frequency would vary depending on the breeding response characteristics of the organism. Furthermore, organisms can tolerate large amounts of radiation spread over a number of generations compared to their tolerance (ability to reproduce) if the radiation is given to one generation. In fact, Russell, Russell and Kelly (1958) and Russell and Russell (1959) have shown that rate of mutation from chronic radiation of spermatogonia of the mouse is less than that for the same amount of radiation as an acute dose. Oster, Zimmering and Muller (1959) have shown that this also applies to gonia of Drosophila. There does not seem to be a threshold, only a reduced rate for equivalent radiation if given as a chronic dose over a long· period of time. Wallace (1959) has presented his ideas of the role of heterozygosity in Droso­phila populations with evidence from some special tests of material from his irradiated populations of Drosophila melanogaster. We quote two remarkable· paragraphs from his conclusions: As a result of the studies described above, we are adopting the tentative conclusion that heterosis plays a fundamental role in the genetic structure of Drosophila popu­lations. We feel that at every locus there are heterozygous combinations of alleles which, on the average, give rise to individuals of higher viabilities or greater fitnesses than do homozygous combinations of the same alleles. Subject to the limitations; The University of Texas Publication imposed by chance elimination of alleles, by mating of close relatives, and by the finite number of alleles at a locus, we feel that the proportion of heterozygosis among gene loci of representative individuals of a population tends toward 100 per cent. We suspect, though, that the heterotic effect exhibited by two alleles at one locus is not independent of the proportion of heterozygosity at other loci. It is difficult, for example, to imagine under the conditions of our experiments an increase in viability of 1~ per cent per locus (or per few loci) accompanying mutations at each of hun­dreds of additional loci. On the other hand, we do not visualize the change in sign for the viability effects of heterozygosity that the hypothesis of optimal hybridity seems to demand. There remains the necessity of evaluating his evidence to determine if it supports such sweeping conclusions or even if it supports the proposition whether muta­tions, spontaneous or radiation-induced, are often heterotic. It has been known for some time from the work of Gustafson (1946, et seq.) that some X-ray mutants, even if detrimental when homozygous, produce single gene heterosis. Probably much of the genetic component of heterosis that Wallace reported was single gene heterosis. It is necessary to comment on the complexities of a system of total heterosis which is the logical extreme of Wallace's position. If there were only four alleles at a locus, there would be 10 homozygous or heterozygous allele combinations at each locus. There are probably 10,000 loci inDrosophila melano­gaster; this would imply that there are 1010•000 combinations theoretically possible by recombination, a number of combinations far beyond the possibilities of all creatures that have or will live on the earth during the effective lifetime of our sun. A further limitation on this tremendous degree of complexity was established by Fisher (1930) long ago, who showed that the fate of all mutations without selective advantage was elimination, and in fact that most individual mutations, even with selective advantage, were lost. The chemical genetics of enzyme or reaction control of even a much more modest system of heterotic factors of any appreciable number are difficult to envision, particularly as Hal­dane (1957a) has calculated that alleles to be retained because of heterosis must not only have general combining ability to work with all the other combinations but that the cumulative advantage of those which will be retained must exceed the simple sum of their advantages. Wallace's statement that more heterozygous loci might not continue to add further advantage but might add less does not seem to agree with Haldane's restrictions on adding heterotic loci to a population. We therefore do not know how to evaluate Wallace's hypothesis without more and different experimental tests. The dangers inherent in small populations with reduced variability are all too well illustrated by those of Eniwetak and Bikini. The optimum genetic variability for different sizes and kinds of populations is not yet clear. Haldane (1937) pointed out the quantitative relation between mutation rate and amount of genetic damage due to detrimental mutations. [Haldane (1957b) discussed the somewhat different problem raised by beneficial mutations.] He expressed the mutational damage in terms of Darwinian fitness-the ability to survive and leave offspring. Muller ( 1950, and other publications) measured the mutational damage iri terms of genetic deaths, that is, extinction of gene lineage. Crow (1957) discussed this problem, giving several examples. Both Haldane and Muller point out that a mutant reducing the fitness of a heterozyg­ous individual by a fraction swill persist in the population 1/ s generations on the average. They go on to point out that the effect of mutation on the population is proportional to mutation rate. Expressed in Muller's terms each such detri­mental mutation leads to a genetic death. Wright (1950, and other publications) has also considered this problem in terms of continuous irradiation as well as short periods of irradiation, and the changes in fitness through time for dominant, semidominant and recessive mutations. In order to give some impre8sion of the persistence of genes with different selective disadvantages, we are including (Table 10) some calculations by Crow ( 195 7), modified to fit Drosophila ananas­sae in terms of its 15-day generation time. Expressed another way with the loss through time illustrated more realistically, Crow showed that if there were 1,000 new mutant genes with s = .2, it would take about three generations to reduce their number by one half, ten generations to reduce it to a tenth and another ten generations to reduce it to one or two per cent. Ans value of .1, which takes 6.9 generations to reduce the frequency of mutants to one half, takes about 22 gener­ations to reduce it to 10 % and 44 generations to get to 1 per cent. These calculations give us some idea of the net detrimental heterozygous effect of the mutations that could have survived from their origin till we checked the populations the first time, some 16 months after the March 1, 1954 thermonuclear explosion. Table 11-shows the radiation received and the spread through time. As mentioned earlier 99 % of the radiation on Rongelap and Rongerik came from fallout due to the March 1, 1954 thermonuclear explosion. The amount of radiation on Bikini is different in that the 1956 and 1958 series of tests also caused fallout on this island. Furthermore the very short lived radioactive isotopes would have been effective on Bikini from the March 1, 1954 test which deposited the most and first fallout there. The AEC has not indicated to us the infinity dose on that island but it was more than 1,000 r and probably less than 20,000 r, at least as it applies to the breeding Drosophila population. Table 11 shows that the ananassae during the first eight days of generation one received about 67% of all the radiation from the first test or about 75 % during the first generation of 15 days. However there remained about 15% of the radiation from this test after T ABLE 10 Loss of detrimental mutations from a population, for different s values, wheres is the detrimental effect of the mutant in the heterozvgote (although these island populations become small enough that some inbreeding would eliminate part of the mutants which became homo­ zygous). T is the time when mutants occurring in a particular generation will have caused one half their harm to the population (half have ended as genetic deaths). Generation time is taken as 15 clays. (Modified from Crow, 1957). T Genera li ons Days ~ .25 2.8 42 .10 6.9 103.5 .03 13.9 208.5 .01 69.3 1039.5 .001 693.1 10395.5 The University of Texas Publication TABLE 11 Dosage accumulation after detonation and increased mutation rate with time 8 days=66.7% accumulated next 45 days= 19% added remaining time= 14.3% added Island Percent over r normal accmnu­mutation lated rate Percent over r normal accmnu­in ulation lated rate Percent over normal accwnu­mutation lated rate Infinity dose in r Overall percent over normal mutation rate Rongelap 280 70 80 20 60 15 420 100 Rongerik 427 110 121 30 92 20 640 160 .Bikini (minimum estimate) 667 170 190 50 143 40 1000 250 (maximum estimate) 13,340 3340 3800 950 2860 800 20,000 5000 This table shows the accumulated dose through time, in r, and the increase in mutation rate, expressed as the pe':'­centage of the normal rate for one generation of 15 days. For example, within the first 8 days after detonation, ap~rox 1 ­mately two-thirds of the radiation has been received by the organisms, and in the next 45 days (about three generat10n s) an additional 19% has been received. Thus, for Rongelap, the mutation rate increased by 70% over the normal rate during the first 8 days, and increased by an additional ZO% during the following 45 days, etc. 53 days and Bikini had the added fallout radiations from later tests. An idea of the increased radiation on Bikini is given in Figure 18. In Table 11 we have indicated the increase in rate of mutations over and above the spontaneous rate due to this radiation. It has been estimated by Muller (1958) that 400 r acute irradiation would double the mutation rate in Drosophila gonial cells (with more certainty for females than for males). The work of Russell, Russell and Kelly (1958) and Russell and Russell (1959) on the mouse and Oster, Zimmering and Muller (1959) on Drosophila melanogaster have shown that chronic irradiation (which would apply to these fallout radiations except for part of the first gener­ation on Bikini) produces a lower mutation rate at least in gonia. Although the relative values for acute and chronic radiation have not been well established (except that there is a lower rate but not a threshold amount necessary to produce mutations) we have used 400 r to calculate the amount necessary to double the normal mutation rate. The values in Table 11 are calculated on this basis and show the added relative frequency of mutation due to the fallout radiations. The dose of 400 r necessary to double the mutation rate in Drosophila gonia is probably considerably greater than that needed to double the rate of mutation in other stages since the very susceptible spermatid stages of the male occupy perhaps one-third of the 15-day life cycle of ananassae. Nevertheless, 400 r seems the most suitable estimate in view of the lower efficiency of low level continuous gamma radiation, even though ingested radioactive isotopes may be more effec­tive. Ifmost of the mutations are present as mutational load, then we may assume that doubling the spontaneous rate (the effect of 400 r or less) might add a lethal and detrimental gene complement, plus occasional beneficial mutations, equiva­lent to that of Ponape or perhaps Majuro. Insofar as part of the genetic load is a balanced load, carried without inversions, the added increment would be re­duced. Except for late mutations due to residual or new fallout (predominantly on Bikini) we expect that most of the mutations which )Vould have persisted until July, 1955 (when the first population samples were taken) would have 11ad a relatively lows in the heterozygotes; there were perhaps a few withs= .1, more as we go to s = .05 then on to s = .01 when a very large proportion should not have been eliminated by selection against the detrimental allele of itself in the heterozygote (but some in addition would have been lost in combination or homozygous). The considerable fraction of the group represented by recessive lethals with a dominant detrimental component often falling between s = .05 ands= 0.1 would still be present. Wright ( 1955) discussed and illustrated irradiation effects in population genetics. Figure 19 is his Figure 5 in that discussion. This figure shows the mean w GENERATION FIG. 19. Mean selective values (;-) of a population with respect to semidominant mutations with 1, 10, or 100% selective disadvantage, under 20-fold increase in mutation rate, and under return to the normal rate after 1, 10, or ZO generations at the high rate (after Wright, 1955). selective values (;.) of a population with respect to semidominant mutations with 1, 10 or 100% selective disadvantage (solid lines), under a 20 fold increase in mutation rate, and under return to the normal rate after 1, 10 or 20 genera­tions at the high rate (dashed curves). Semidominance is assumed but any muta­tion with a detrimental effect in the heterozygote would give essentially similar results. If the mutation rate is increased 20 fold as in the figure, mutations with a selective disadvantage of s = 0.10 go about half way to equilibrium (with op­posed selection) in seven generations and are nearly to equilibrium in 30 gener­ations, whereas genes with s = 0.01 require 69 generations to go half way to equilibrium (Wright, 1955). In case of dominant lethals, s = 1.0, the population goes all the way to equilibrium at once, since by definition dominant lethals are not transmitted. This implies that situations like Huntington's chorea of humans are special cases; due to their late development, after at least part of the reproduc­ The University of Texas Publication tive period is over, they contribute misery to successive generations despite the lethal character of the disease. As shown in Figure 19 (dashed lines), after removal from irradiation the populations recover slowly, in fact at the same rate as the downhill process under irradiation. In a diploid sexual form such as Drosophila, w under spontaneous mutation is 1-2 v0 and with a 20 fold increase in mutation rate is 1-4-0 Vo. The relations between the rate of mutation, the degree of selective disadvantage, the rate of reaching equilibrium under chronic irradiation, and the rate of recovery after cessation of irradiation are all illustrated in the figure. These factors apply to the effects of fallout radiations on islands such as Bikini. The situation is slightly more complex since radioactive decay reduced irradiation through time (see Figures 3 and 18). Figure 18 also shows that the amount of irradiation was increased on Bikini by the operations of 1956 and 1958, but no subsequent test, even those in 1956, produced fallout comparable to that of March 1, 1954. How­ever, the general situation resulting from the increase in mutation rate and, in our case, in the loss back toward the normal equilibrium as the fallout decayed can easily be understood by observing the relationships shown in Figure 19. The data from the tests of populations sampled in the summers of 1955 through 1959* must be examined carefully to determine if they can reveal the effect of radiation in producing detrimental mutations in these populations, and to see if we can detect the decay of this induced variability. If we cannot, then we have demonstrated the usual fluctuations of these populations through time--that is, we have been studying evolution (without obvious radiation effects) of these small isolated populations. Two points must be made first: ( 1) There is definite isolation between several of these island populations, especially between the Ponape population and those of Bikini, Rongelap and most especially Majuro. This is shown as sexual isolation-lower crossfertility (Figure 16), but also as reduced fecundity, since Majuro females inseminated by Ponape males lay too few eggs (Figure 15) . An effect on viability of fertilized eggs is not obvious although some effect may be present, for there is no heterosis on crossing (Figure 17). These components of the evolution of small isolated populations will be discussed later. (2) There was sufficient radiation to cause a sharp increase in mutation rate (the minimum would add radiation-induced mutations equivalent to those that occur spontaneously and so at least double the mutation rate, Table 11). The radiation was particularly heavy on Bikini and later weapons test series added to the fallout on that island. In fact these estimates of the amount of damage due to radiation as multiplication of the spontaneous mutation rate in Drosophila ananassae is low. Estimating a generation of 15 days, about one-third to one-fourth of the spermatogenesis cycle is spent in the spermatid stages which are two to ten or more times as susceptible to damage as sperm or spermatogonial stages. This would increase the damage from fallout radiation (see Seecof, in Stone et al, 195 7) . The question to be answered is whether the residue of the· induced detrimental mutations could be demonstrated together with its effect on the population in our samples of Drosophila from these islands. The answer is in *Since the Bikini population died out between 1958 and 1959, just as the Eniwetak population did between 1955 and 1956, we doubt if it will be profitabl~ to attempt to get more data from the Pacific Proving Ground. the affirmative and several relations between amount and time of radiation, population structure, and the effect on components of Darwinian fitness in our tests are shown (see Table 7 for example). . With the tests for the five years, we can determine the comparative effects of the laboratory testing conditions on the viability results, particularly from the responses of the Majuro and Ponape control populations (Table 6). In terms of optimal conditions for Drosophila ananassae, the laboratory conditions improved between tests in 1955 and 1956 (the best year), became somewhat worse in 1957, became still less satisfactory in 1958 (the worst year), and improved slightly in 1959. For certain years no tests were available for some stocks or crosses, and the Rongerik atoll sample in 1955 is from the island Eniwetak (this population had disappeared by 1956) while tests from 1956-1959 are from the island of Rongerik. The several tests of egg development given in Figures 4-7 show this factor of Darwinian fitness very well. We must ignore the crosses between Majuro and Ponape because the sexual isolation and egg reduction shown in crosses may also be associated with effects on egg development. Fecundity (eggs laid per day) does not follow the same pattern (Figure 12), but Figures 13 and 14 summarize the changes in laboratory environment for fertility and egg development. In the case of egg development the inbred crosses of 1955 bring out the effect of radiation on Rongelap and Bikini. The percentage egg development for both these stocks was significantly below that of Majuro in 1955; in 1956 only Rongelap differed significantly from Majuro and in this case it was higher than Majuro. Other P1 and F1 tests with Rongelap were high in 1956 so the considerable improvement indicated from brother X sister mating in the stock was real. This is also demon­strated by the improvement shown by Rongelap compared to Bikini in Table 7, as well as the relative frequency of lethals as estimated from the data. Table 11 shows the approximate radiation present to influence mutation rate and the period of time in which most of it would have been effective. Bikini was still radioactive above normal in 1955 with some "hot spots" easily detectable with a radiation counter (Figure 18). The mutations with a detrimental effect in the heterozygotes are eliminated at a rate determined by s, Table 10 and Figure 19, but are also influenced by the fact that these small atoll island popula­tions are affected by considerable random fluctuations. This last statement fits the data on fertility and egg development (see Figures 4, 12 and 13 and Table 3 and 6). Ponape, the only large population with considerable area and diversi­fied habitat, is a very stable population as compared to the three atoll island populations. The number of eggs per day laid by fertile females is a species specific character and the populations, although differing slightly, tend to vary together for this character in relation to environmental variables (Figure 11 and Table 2). If we compare all fertility tests, year to year, we find remarkably little variation. Either fertility was not modified appreciably by environmental and genetic variation or they compensated in this set of tests. Table 1 and earlier publications in this series show that even in the control populations, Majuro and Ponape, there is usually at least one lethal or lethal complex in one parent of the pairs tested. In fact, a comparison of the distribution of egg development classes between the randomly mated individuals and brother x sister matings from the same population suggests that there are often several The University of Texas Publication lethal factors present. Any detrimental factors added by the radiation will have proportionally less further reduction in viability on inbreeding with any increase in the number of detrimental factors already present. Further the new factors with appreciable detrimental effect heterozygous would tend to be eliminated first (Figure 19). With this in mind the changes in percent egg development from year to year, the average egg development (Table 6 and Figures 4-7), and the general trends from year to year allow us to show the effects of radiation in terms of a residue of mutations which are less detrimental in the heterozygote. In the elimination of mutations, dominant lethals would be lost in one genera­tion and serious detrimental semidominants before the 16 months (equalling about 34 to 35 generations) between the March 1, 1954 test and the time of our first collections on Bikini. There was continuing improvement of the Bikini population from 1955 to 1958 (see Table 7)-in fact, of the two samples collected on Bikini, one in late July and one in early September, 1955, two or three gener­ations apart, the first sample, inbred brother x sister at 71°F, gave 40.9% egg development, while the second, tested at 77°F gave 54.1 % egg development. Two samples of the Majuro population collected late in August were tested: at 71°F egg development was 63.1 %, at 77°F, it was 52.5%. Both these pairs of values differ significantly at the .05 level. Unless we consider that these two stocks differed in their response to difference in temperature during testing, even though they were both selected under the same temperature conditions, this suggests that there was improvement in the Bikini stock, due to loss of one or more detrimental factors between about 34 and 37 generations. As we pointed out above (Table 6), while most populations varied with changes in lab environ­ment, Bikini inbred improved between 1955 and 1956, lost slightly (as did most of the other stocks) between 1956 and 1957, and improved 7.4% in egg develop­ment from 1957 to 1958 (the worst year of lab conditions) although each of the 24 other tests that year recorded reduction in egg development (average reduc­tion about 7 .3 % ) . Rongelap inbred made a major improvement from 1955 to 1956 to a peak above any other inbred test, dropped back slightly in 1957 to the usual atoll level and remained there in 1958 and 1959. One of the three stocks of Drosophila melanogaster studied by Wallace (1956) that was removed from radiation after receiving about 250,000 r of gamma radiation recovered very slowly. Rongelap may have had some similar detrimental factors. However, it is probable that the added det:r;imental component of mutations both reduced via­bility and caused more pronounced fluctuations in the population, which oscil­lated back to normal after 1956. Furthermore the most heavily irradiated popula­tion, Bikini, showed the lowest average viability from inbred tests, Bikini X Rongelap had the low average for crosses between strains, and the F1 inbred of that cross (BAX BA) was low for F1 inbred tests. These results are all consistent with our opinion that there was radiation damage from fallout to the populations on the islands of the Northern Marshalls in the Pacific Proving Ground area, followed by the loss of the detrimental factors due to natural selection through time. Those mutations with a dominant component of detrimental effects dropped out rapidly so that most factors that remained by 1956 were in the usual range of lethal and detrimental mutations of the type found by the usual analysis of ordinary Drosophila populations. Even with the rapid elimination of detrimental factors with dominant heterozygous effects, and with the small size of the popula­tions which occur in these islands tending to increase the speed of elimination of detrimentals by inbreeding, we could show an effect of fallout radiations which produced detrimental mutations; these were still being eliminated from the Bikini population between 1957 and 1958, the Bikini population showing improvement each successive year (Table 7). Evolution in small isolated populations: The study provided information on five isolated populations: one on Bikini island; one on Rongelap island; two in the Rongerik atoll, Eniwetak island in 1955 and Rongerik island in 1956-9; one on Majuro atoll, the island of Uliga; and the population on Ponape (see Chart 1). The atolls are sufficiently distant from each other and the local population on any atoll is so small that we could expect a few migrants to move from one to another very infrequently. Ponape, which supports a much larger population, is sufficiently far (hundreds of miles) from the atoll islands we studied that a certain amount of interpopulation isola~ tion has developed between them. Very few genetic studies of such small isolated populations have been made. Stone, Alexander and Clayton ( 1954) studied Drosophila novamexicana which forms very small linear populations along desert streams in the American southwest, and Drosophila hydei which has a thin and scattered population in the desert areas where the samples were collected. Dobz­hansky and Pavlovsky (1957) studied the interactions between drift and selection in large and small populations and Dobzhansky, Levene, B. Spassky and N. Spas sky ( 1959) studied Drosophila paulistorum which has a thin and scattered population in neotropical America. Only novamexicana has populations as small as these isolated ananassae populations and it has a somewhat different popula­tion structure, being restricted to and scattered along desert streams so that prob­ably no local population builds up as did Rongelap locally in 1956. Figure 20 .lJrosophilo novemexicano Drosophila hydei 15•9) . 4 6 616 4xl 416 h.6 (4xl)f,(4x6)F, hG)F, 100 ] j 91.7 J 90.9 86 7 _J ~ 82 -8 z J ~eo 798 w "' ~ _.. 70 -_Jf 731 ulifl 67 7 69 2 _...~ . 660 5 3 5x7 7x 9 7x 3 9 x 3 5.r.9)f, (7x9)F, (7x3)f,{9x3lF, ld7d) 100 JJJJ,, J J.. " j _J B>O 822 82.2 B•.B J, 834 J BO 78.5 _J 770 74.3 70 Fie. ZO . Comparable tests with two other species, showing brother X sister matings, the darkened histograms, and random crosses within a population or crosses between individuals from two populations, the plain histograms. The numbers along the top margin refer to stocks -of Drosophila novamexicana and of Drosophila hydei from different localities; see Stone, Alex­ander and Clayton, 1954. The University of Texas Publication shows part of the data from the studies of novamexicana and hydei, which were tested the same way, for comparison to ananassae. We have mentioned that Ponape was the most stable population of the five studied. Ponape is a high island with large areas of tropical island rain forest. There are numerous fruits in addition to those in the Marshall Islands, and other situations where yeasts and fungus grow. Even the year in which a typhoon destroyed the fruit crops and reduced the ananassae population to very small numbers in the lowland area around the town where we usually collected, there was little change in the results of our tests that year or in 1959 when the population had built up again. On the other hand Majuro, Rongelap and Bikini varied considerably from year to year, . in part with the laboratory environment, but also from the genetic effect of radiation on the last two island populations. The structure of these populations must lead to wider random fluctuation and change in the genotype, which may also be accentuated by inbreeding in localities within the populations as the numbers fluctuate and as flies develop on or migrate between small local fruit concentrations (see Tables 3 and 6, Figures 4-7, 13 and 14) . The populations on the two islands of the Rongerik atoll were never abundant enough to allow ade­quate cross tests. Table 2 and Figure 12 show that the fecundity, measured as the average number of eggs laid per day by the fertile females, is a species specific trait and does not vary much between freshly collected strains of wild popula­tions. This was true of novamexicana and hydei also (Stone, Alexander and Clayton, 1954). However, it is modifiable by adverse laboratory conditions and this is shown in Table 2 and Figure 12 for 1959 tests. Another point of importance about these populations is shown in Table 7 and Figure 11 . The genetic variability is sharply reduced as a result of the popula­tions periodically being reduced to small numbers. This is shown by the loss of recessive detrimental factors as determined by comparing the egg development in brother X sister matings, cousin matings, and random matings from a popula­tion (Table 7). Figure 11 shows a low residual variability in Rongelap and Rongerik in 1959, for there is little difference on inbreeding or crossbreeding. Both tests show low values for cousin and random matings, suggesting fixation of relatively less effective gene combinations with minimum heterozygosity. Also Rongerik in 1959 shows low variability for recessive sterility factors; as is shown in Table 4, the smaller Rongerik population was much less variable ge­netically than was the Rongelap population. With these Pacific Proving Ground populations passing through cycles of very small numbers rather than being very large and with a great reserve of genetic variability, when the fallout. dropped on them there was undoubtedly a still greater reduction in numbers from dominant lethals, sterility and injury especially to immature forms; the radiation may be considered to increase the probability of fixation of slightly detrimental factors and cause a generally adverse reduction in genetic variability. Wright (1931, et seq.) has pointed out that very small populations must usually die out because of lack of genetic variability. Wallace (1956) tested for elimi­nation of lethals from three populations long subjected to gamma radiation which provided them an accumulated dose of some 250,000 r. After removal from the radiation field, one population very rapidly (in 25 generations) eliminated the accumulated second chromosome lethals from a level where 85.7% of the chro­mosomes carried lethals to a level of 28.0%. The second population reduced the number of lethal second chromosomes from 85.7% to 56.2% in 24 generations. However, the third population did not lose lethals from the second chromosome to any appreciable degree, changing only from 79.7% to 76.7% ih 24 genera­tions. All the Pacific Proving Ground populations received much less radiation than the Drosophila melanogaster populations studied by Wallace. By 1959, all of them had recovered or nearly recovered to their normal level of lethal and lethal equivalent mutations which falls between that of the Majuro atoll and the high island, Ponape, population. However, the Pacific Proving Ground popula­tions probably represented a skewed distribution of genotypes compared to the control populations, still retaining some of the detrimental mutations with a small s, but being much less genetically variable and probably having fixed a less favorable series of alleles in addition to lack of heterozygosity (and heterosis). In fact two of the four populations, Bikini and the Eniwetak island population on the Rongerik atoll, died out during the five years we collected. Bikini arid Eniwetak were well established populations and at least Bikini and Rongelap were of long enough duration as separate populations for genetic diversification of isolating factors (see below) . We must attribute the destruction of the Bikini and Eniwetak populations to a series of circumstances, partly en­vironmental but aggravated especially by the decay of genetic variability which has been shown. The detrimental environmental factors that were detected were as follows: 1) A poor year for fruit production aggravated by 2) competition for fruit by land crabs and especially rats. The latter were so plentiful (on Eniwetak in 1955 and on Bikini in 1958) that none of the usual assortment of fermenting fruits was to be found on the ground. In fact the rats were eating the green fruit on the trees, so we found no fallen fruit of Marinda, and practically no fresh pandanus fruits. Human activity had destroyed the source of papaya fruit that year. We also observed that even the large sphinx moths which were exceedingly common and plentiful around a vine, lpomoea, were almost entirely missing although open flowers were plentiful in August, 1959. "'\Ve also noted that the spiders, which depend on insects for food, were noticeably rarer on Bikini in 1959 compared to the other four summers. 3) Both islands had been subjected to heavy insecticide spraying the year before the populations disappeared and Drosophila ananassae were present in small numbers at that time. However, these small populations were still in existence after the cessation of spraying the year before they died out. 4) The radiation had probably reduced the amount and types of food available on Bikini; for example, coconuts are used by ananassae to some degree and they were not producing due to past irradiation (and/or blast dam­age) . Ifwe superimpose on these factors the lowered genetic variability and small population size, at least in part due to the irradiation, we see how the comb;nation of factors could lead to the death of the populations. The size and variability of a population can be of critical importance when the population is subjected to the bottleneck effect of drastic reduction in numbers. whether from sudden disaster or from gradual reduction over longer perior1.s of time. A large poµulation, with its normal content of variability, can survive a disaster (short of actual total extinction) with relative ease since the remnants of the population are exactly the same individuals which were part of the The University of Texas Publication "normal" pre-disaster population; as such they are capable of restoring the population to normal size within a relatively short time. The prompt recovery of the Ponape population after its extreme reduction following the typhoon illus­trates this situation well, and strongly contrasts with the inability of the Bikini and Eniwetak populations to survive the bottleneck produced by irradiation ef­fects and other environmental hazards imposed upon them over a considerable period of time. Small populations, subjected to sudden disaster, would have reduced chance of survival, but more importantly, a slow and gradual reduction, a prolonged fun­nelling process leading to a bottleneck, would not only reduce population size and reduce genetic variability, but would also allow the fixation of deleterious genes and gene complexes so that the population would be unable to respond in a normal manner after the period of the bottleneck had been passed. Extinction, such as that observed to have occurred on Bikini and Eniwetak, would be the most likely result. An example of other combinations of difficulties with which populations may be faced is illustrated by the Rongelap population. In 1955 the population sample from this island gave a very low percentage egg development on inbreeding brother X sister. By 1956 the population was much improved where we sampled it, and in fact these particular samples gave the highest egg development of any inbred test of these populations (Figure 4). This was a beneficial genetic com­bination for it carried over its effects to crosses and F1 tests. This is the only and very fragmentary evidence that beneficial mutations or most probably a bene­ficial combination of genes could occur in irradiated populations. This population was found on fallen fruit at the site of the old native village on Rongelap. In clearing and rebuilding a new village for the Marshallese, the fruit source was drastically reduced, then still further reduced by the Marshallese and their chick­ens and pigs. Therefore this population which was so favorable in 1956 could not raise the general genetic level of the population on the island and by 195 7 it had dropped back to the general Majuro level; in fact in 1959 it showed the low variability in factors influencing egg development already discusssed. It should be pointed out that the number of Drosophila ananassae found by us on Ponape in 1958 was not much greater than that on Eniwetak in 1955 or Bikini in 1958. Fruit production was reduced by a typhoon, and in addition the giant African snail, Achatina fulica, and chickens devour the fallen fruit. There was also some spraying on Ponape; however, Ponape had many local population pockets in the rain forest not reached by the spraying. The total population was able to recover very effectively even though local populations in certain areas of the island died out. Furthermore Ponape did not have the radiation damage from fallout as did the islands in the Northern Marshalls. These Pacific island populations of ananassae may be compared to the desert hydei and desert stream novamexicana populations although the latter were tested only one year. Stone, Alexander and Clayton (1954) made tests of popu­lations of these two species and some of them are given in Figure 20. Only three small novamexicana populations (numbered 1, 4, and 6) were found but one of them, number 4, carried a very interesting factor. It had little effect within the 4 population (see 4 inbred) but acted as a semidominant crosslethal so that part of the fertile male or female F1 (and F2, which are not shown in this table) which were used in crosses of this stock to either number 1 or number 6 gave a very poor egg development (see F 1 in Figure 20). This factor had the effect of an iso­lating factor, reducing the effectiveness of heterozygotes between strain 4 and the other stocks, many of them having 4-0% or less of their eggs develop. The genetic analysis of this factor and its interaction was incomplete although there seemed to be a modifying factor(s) present in 4 which reduced its effect within that population. Stocks 1 and 6 are rather like Ponape but brother X sister matings of 6 show fewer detrimental factors than any of these Paeific island ananassae stocks. The desert hydei populations are somewhat different from and better than the anaruzssae. The hydei strains show few detrimental factors on inbreeding as compared to crossing. In fact only three of the four P1 inbred and none of the four F1 inbred brother X sister gave egg development below 80%. The maximum spread between the lowest inbred (74.3%) and the highest (88.0%) egg develop­ment from a four-way cross which would give a maximum heterosis is only 13.7%. There is some increased egg development on crossing but the best F1 inbred was 87.6% which is just as good as the four-way cross. Their values fall above that for any anaruzssae stock inbred except for the exceptional Rongelap stock in 1956, and all F1 inbred for hydei fall decidedly above those for ananassae. Only stock 6 in novamexicana and 1 x 6 heterozygotes resemble hydei, but stock 1 inbred shows more detrimental factors. We must conclude that the thin mesh­work of hydei scattered widely over the southwestern desert can achieve a much better population structure, at least insofar as the wide range of the hydei insures adequate genetic variability. Certainly these populations of hydei and novamexi­cana (stock 4 excepted because of the semi-isolating factor) have fewer lethals and detrimental factors than most other species. Tables 8 and 9 allow some comparisons but the tests were run in different ways. The special viability tests of the Rongelap and Rongerik populations in 1959 (Figure 11) show that the variability is low (the Rongerik random test is too small to consider significant) . They must differ from hydei in prossessing a fixed but less well adapted genotype, for certainly the Majuro and Ponape populations are demonstrably better in that test. Therefore we cannot assign the poor viability of Rongelap and Rongerik to bad laboratory conditions. Patterson and Stone (1952) reviewed the considerable body of literature on isolating mechanisms in the genus Drosophila. They gave examples of partial isolation between strains of the same species. Additional cases have been reported, for example the work of Santibanez and Waddington ( 1958). The four ananassae populations tested in crosses show that considerable isolation has developed be­tween Ponape and the other three, Majuro, Rongelap and Bikini. The pertinent comparisons are to be found in Tables 2, 3 and 6 and Figures 15, 16 and 17. In the last three figures random mating within a stock is indicated by c and brother X sister mating by i. Table 6 and Figure 17 give the data on viability measured as egg development. These results are as expected for the values in crosses are about the same as those in random matings of the two stocks, while the value of F1 inbred is about like the two P1 inbred. The results from four or five years of tests for crossfertility and cross-fecundity The University of Texas Publication bring out the evidence for isolating factors. In Table 3 and Figure 16, showing fertility, the crosses M x B and Bx A are in the normal range, as are their F1 inbre.d. The crosses between Ponape and the other three showed (sexual) isola­tion since too few of the pairs tested were fertile. The A X M cross was low al­though the F1 inbred (AM x AM) was in the expected range. However, even the F1 inbred of Bikini X Ponape was low. Furthermore, the heterozygous BA females showed strong isolation from the MP males in the four-way crosses so some isolat­ing factors were dominant, (e.g., BP X BP and BAX MP) while others are re­cessive (AM x AM). We have the interesting case of low fecundity in the case of Majuro fem ales inseminated by Ponape males (Table 2 and Figure 15). The fecundity of this cross is significantly below the parent Majuro females or their cross of Bikini males. Furthermore, we have the complementary situation, since Bikini females crossed to Rongelap males lay a much higher average number of eggs per day than any other of these ananassae combinations. The fact that the sperm or semen are important in stimulating laying in the females in some species, as well as the insemination reaction, was already well documented (Patterson and Stone, 1952). It is very interesting to note that these four small isolated island popula­tions evolved several genetic differences from each other which were detectable as isolation mechanisms, and that we were able to detect the special positive compatibility between Bikini and Rongelap populations, which caused a higher number of eggs to be laid than are laid by either stock in crosses within a population. SUMMARY 1. Samples of the circumtropical species, Drosophila ananassae, were collected in July and August for a five-year period, 1955-1959, from islands in the USAEC Pacific Proving Ground area. Populations from two control areas, one from the island of Uliga in the Majuro atoll of the southern Marshall Islands and a second from the high island of Ponape (which has many times the area of all the Marshall Islands combined) in the eastern Caroline Islands, were collected each year except for Ponape in 1955. These two islands received little fallout com­pared to the islands sampled in the northern Marshalls. The experimental popu­lations were collected from the northern Marshall Islands. Rongelap island of Rongelap atoll and Eniwetak island (1955) and Rongerik island (1956-9) of the Rongerik atoll received fallout from the thermonuclear device exploded March 1, 1954 (over 99 % of all fallout) and some slight additional fallout from the test series in 1956 and 1958. Bikini island of Bikini atoll received very heavy fallout from the March 1, 1954 test and received added fallout from the test series of 1956 and 1958 (See Chart 1, Figures 1, 2 and 3, and Table 11). The fallout on Bikini is estimated to have been between 850 and 17000 r to most flies within the first four generations. A closer estimate, considering the fallout on Rongerik which lies at an appreciably greater distance from the point of detonation, Figure 1 and Table 11, would be between 3000 and 8000r, since Bikini was in the zone which received over 800r in a 48-hour period (Figure 1). 2. The ecology of these island populations was studied to determine competi­tors, predators, food, etc., and the effect of man ;~.....Imling destruction of normal habitat, insecticide sprays and radiation and fallout. The general and comparative sizes of the populations were determined as far as possible. 3. Samples of the populations were shipped back or carried by air to the Austin laboratory where they were tested and compared in terms of viability, fertility and fecundity each year. For the 1955 and 1956 samples, Spencer checked certain populations for the kinds and frequencies of visible mutants which could then be compared to similar tests with other species of Drosophila. The populations were much alike and mutation incidence was of similar magnitude to that of other Drosophila. Seecof checked the Bikini and the two control populations for fre­quency of rearrangements. There were a few inversions present including several new ones, but the frequencies were similar to those in other ananassae popula­tions, with no excess in the Bikini population. Seecof also ran an irradiation ex­periment which showed that Drosophila ananassa.e was very similar in its response to irradiation to Drosophila melanogaster, a member of the same species group in the subgenus Sophophora. The inversion frequencies in their natural populations are much alike also. 4. The data on these populations were obtained four or five summers, begin­ning about 16 months after the heavy fallout from the thermonuclear device exploded March 1, 1954. Unfortunately two of the four irradiated populations died out (or dropped to such a low level that we were not able to collect any specimens, either by trapping or sweeping). Eniwetak island had an ananassae population in 1955 but that population, which was very small then, had disap­peared by 1956; we therefore collected samples of the population on the nearby Rongerik island in 1956 through 1959. The Bikini population which was small in 1958 had disappeared by 1959. As Bikini received by far the largest amount of radiation, we terminated the tests after the 1959 collection. We conclude from our studies that these two populations died out for a combination of reasons: (a) a poor season for fruit, (b) general islandwide insecticide spraying which re­duced each population, (c) intense competition for food both for adults and especially larvae, (d) radiation damage to the genetic system. The rats on each of these islands had built up a tremendous population; on Bikini in 1959 they quickly ate all the fallen fruit such as Morinda (where adults feed and larvae develop) and were observed climbing trees and eating the green fruit. Even the spiders which depend on insect food were very scarce on Bikini in 1959. A typhoon destroyed the fruit crop on Ponape in 1958 and ananassa.e was very scarce in the lowland area where we collected (other species normally present were reduced even more as judged by our collections) . Nevertheless ananassae had recovered by August of the following year, probably because there are many reserve areas of different fruits and other situations where yeasts and fungi grow in the rainforest area of this high island which escapes insecticide spraying. Radiation damage must also have contributed to the elimination of these popu­lations. Detrimental factors which act in the heterozygote are eliminated in relation to their detrimental effect (Figure 19). Most of the factors with 1 or 2% disadvantage in the heterozygote would persist many generations; in fact, an appreciable proportion of those with 5% disadvantage might still have been in the Bikini population in 1955 (this population improved each succeeding year, see Table 7). As these populations became very small, chance increase or even The University of Texas Publication fixation of such factors may have reduced their survival probability appreciably. All factors together proved to be more than the Eniwetak and Bikini Drosophila populations could survive. 5. We measured the important components of Darwinian fitness, fecundity, fertility and viability. These tests measured the effects of the dominant com­ponents of the genes influencing these characters. Certain special experiments measured the effect of some genes being homozygous from cousin or brother X sister matings. These are simpler to discuss separately. (a) Fecundity was measured in terms of the number of eggs laid per day by fertilized females within the first week after mating, usually over a four-day period after insemination. In general, fecundity is a species specific character. The several populations differed slightly, but all varied the same way in response to environmental conditions (Table 2 and Figure 12). In general only dominant components of gene action affecting fecundity were measured since the X chro­mosome in the male cannot influence egg laying. A very interesting component of fecundity is shown very effectively by this data. Figure 15 summarizes the four or five years of tests of four populations and six crosses between them. Four of these give the expected results, with the crosses falling within the range of the stocks (which are given for inbreeding, i, and crossbreeding, c). Two are outside the range: Majuro !i' !i' x Ponape J J (M, P) produced too few eggs as compared to either stock (which is reflected in the low fecundity for this same cross, Figure 15); Bikini !i' !i' X Rongelap J J (B, A) produced more eggs than expected. Patterson and Stone (1952) reviewed evidence that sperm and semen stimulate egg laying and that alien sperm were not as effective. These data give both complementary cases: one cross, M !i' x P J showed lack of stimulation or inhibition of laying below the normal level of Majuro (or Ponape) females; the B!i' X A J cross showed a special or added stimulus to egg laying of Rongelap sperm in Bikini females, above that expected from either parental cross. These cases of reduced or stimulated fecundity do not affect the conclusions to be drawn from Figure 12, based on inbreeding and random mating within a strain. The worst year in terms of laboratory conditions which influence laying was 1959; that year less than half as many eggs were laid per day by inseminated females as in the best year, 1957. These tests show both strong genetic effects and en­vironmental effects on this component of evolutionary fitness. (b) Fertility. Table 3 shows mainly dominant genetic and environmental effects on fertility while Table 4 shows additional recessive gene influence, tested only in 1959, by comparing pairs of F 2 tests-when the F1 parents were unrelated versus tests using brother and sister as F1 parents. A greater percentage of pairs of Bikini flies are fertile than from any other population (Table 3, Figures 13 and 16). Except for the very small Rongerik population in 1959, all three other stocks tested and the A, M heterozygote all showed inbreeding depression on fertility; that is, recessive factors which reduce fertility were present. Probably inbreeding and fixation of factors by chance and selection which occurred in the very small Rongerik population reduced or fixed the number of recessive sterility factors present. The data also demonstrate isolating factors between Rongelap ~ !i' and Majuro ~ ~, Ponape !i' !i' and Rongelap ~ ~, Majuro !i' !i' and Ponape J J, and Bikini !i' !i' and Ponape J J. In only this last ca!le were these cross­sterility factors dominant (F1 isolated from each other), the other three cases were the result of recessive cross-sterility factors in the separate populations (the F1 crossed freely). As in fecundity, fertility is influenced by genetic and slightly by environmental factors. Unlike fecundity and viability, fertility did not show a large year to year fluctuation (Table 3). (c) Viability. We studied both the dominant components of gene action which influenced viability and the recessive factors as well since we regularly inbred brother X sister (and cousins in 1959), mated at random within a stock, crossmated between stocks, and tested F1 and in some cases F2 progeny (Tables 5, 6, and 7, Figures 4--11 and summary Figures 14 and 17). Ponape was the most stable stock (Table 6 and Figure 14), with fewer lethals (Table 7) and less fluctuation in response to differences in the environment. Majuro varied with the environment more than the others, as we regard the very low value of Ron­gelap inbred in 1955 as due to persistence of detrimental mutations induced by radiation from fallout. The summary of viability, Figure 17, does not show any unusual values in crosses of F1 X F1 in view of the isolating factors demonstrated for fecundity and fertility (Figures 15 and 16), except the low values of crosses of BX A and BAX BA. We regard these as actually the result of factors which have dominant and recessive detrimental effects, and which resulted from the radiation of these two populations. The general picture we get from the sum­mary of lethal frequencies through the years 1955-1959 ·(Table 7) is one in which the Ponape and Majuro populations fluctuated about a mean lethal fre­quency, but that the number of lethals in the Rongelap and Bikini populations was too high (due to radiation) ; therefore the lethal value fell in both these populations. The Rongelap population adjusted rapidly but that on Bikini re­duced the relative number of lethals present each successive year of the tests (Table 7). Some of the detrimental factors that the Bikini population eliminated over the four-year period were from the fallout from tests in 1956 and 1958. 6. A most consistent and interesting finding is the great variability of the smaller atoll populations as compared to the relatively large population on Pon­ape. Part of this is due to lesser ability to resist fluctuation with environmental change. On the other hand as the populations become very small the genetic variability may become quite low (Table 1 and Figure 11). In some populations the cumulative effect of fluctuations, reduction in size with loss of variability, adverse environmental conditions and radiation damage due to fallout, were suf­ficient to lead to the disappearance (death) of two of the four populations studied from islands in or near the Pacific Proving Ground in the northern Marshalls. Another very interesting finding is that isolating factors that reduced fecundity in crossmatings and that reduced crossfertility had developed between several of these populations (Figures 15 and 16) . ACKNOWLEDGMENT The authors wish to express their indebtedness to the many persons who helped us with this work. This includes the students and associates who carried out part of the work and published on their part in earlier articles. We wish to mention that Drs. Warren Spencer, Calvin Ward and Marvin Wasserman helped us collect the Drosophila on these Pacific islands several summers. Dr. Forbes The University of Texas Publication Robertson of the Institute of Animal Genetics, Edinburg, was kind enough to help us with the mathematical analysis of the 1955 and 1956 data, and Dr. William Hanly and Mr. Robert Rinehart made the calculations for Table 7. Dr. James F. Crow of the Department of Medical Genetics, The University of Wisconsin, was kind enough to read the manuscript and suggest some corrections. We are indebted to Prof. Robert W. Hiatt, Dean of the Graduate School, University of Hawaii, and Director of the Eniwetok Marine Biological Laboratory for his kind­ness and assistance, together with the help of several of his associates and as­sistants at EMBL. Mr. Maynard Neas, Mr. Henry Hedges, and a number of other persons associated with the administration of the Trust Territory of the Pacific Islands were of great assistance to us in obtaining our control populations. from Majuro and Ponape. We are particularly indebted to Mr. Ernest Wynkoop and Mr. Thomas Hardison and to other members of the AEC Pacific Proving Ground staff, who looked after our needs including our necessary travel between islands with the cooperation of Air Force and Navy personnel. We wish to thank many of our students and assistants who have helped in the laboratory work including (in addition to those persons who were joint authors with us in the first study of this series) Dr. N. B. Krishnamurthy, Dr. Tsueng-Hsing Chang, Richard Dicker­man, Robert Rinehart, and Mrs. Viriginia Gerstenberg. This research was made possible by grants from the Rockefeller Foundation and the NIH, RG-6492, which supported part of the laboratory personnel, a grant from the Welch Foundation which provided the X-ray machine used by Dr. Robert Seecof in the control test of the effects of radiations on Drosophila ananassae, and especially by contract AT-(40-1) ~1323 with the Atomic Energy Commission which supported our work in the laboratory and in the field. LITERATURE CITED Chetverikov, S. S. 1926. On certain features of the evolutionary process from the v;iewpoint of modern genetics. J. Exp. Biol. (Russian), 2: 3-54. ----. 1927. Uber die genetische Beschaffenheit wilder Population. Verh. V. internat. Kongr. Vererbungsw., 2: 1499-1500. ' Cunha, A. B. da. J. S. de Toledo, C. Pavan, H . L. de Souza, M . L. Pires de Camargo, and L. C. de Mello. 1958. A comparative analysis of the effects of natural and radiation-induced leth,,ls in heterozvgous individuals and of their frequendes in natural populations of Dro­sophila willistoni. Proc. X Intern. Congr. Genetics, Vol. II.: 63-64. Cunha. A. B. da, J. S. de Toledo, C. Pavan, H. L. de Souza, H. E. Melara, N. Gabrusewycz, M. R. Gama. M . L. Pires de Camargo and L. C. de Mello. 1959. A comparative analysis of the effects of natural and of radiation-induced lethals in heterozygous individuals and of their frequencies in natural populations of Drosophila willistoni. Progress in Nuclear Energy, Series VI, Biological Sciences: 359-363. London: Pergamon Press. Cordeiro, A. R. 1952. Experiments on the effects in heterozygous condition of second chromo­somes from natural populations of Drosophila willistoni. Proc. Nat. Acad. Sci., 38: 471-478. Cronkite, E. P., V. P. Bond and C. L. Dunham (editors). 1956. Some Effects of Ionizing Radia­tion on Human Beings, 106 pp. Washington, D.C.: U.S. Govt. Printing Office, T.l.D. 5358. Crow, James F. 1957. Possible consequences of an increased mutation rate. Eugenics Quarterly 4: 67-80. Dobzhansky, Th. and Sewall Wright. 1941. Genetics of natural populations. V. Relations between mutation rate and accumulation of lethals in populations of Drosophila pseudoobscura. Genetics 26: 23-51. Dobzhansky, Th., 0. Pavlovsky, B. Spassky, and N. Spassky. 1955. Genetics of natural popula­tions. XXIII. Biological role of deleterious recessives in populations of Drosophila pseudo­obscura. Genetics 40: 781-796. Dobzhansky, Th., and H. Levene. 1955. Genetics of natural populations. XXIV. Developmental homeostasis in natural populations of Drosophila pseudoobscura. Genetics 40: 797-808. Dobzhansky, Th., and Olga Pavlovsky. 1957. An experimental study of interaction between genetic drift and natural selection. Evolution 11: 311-319. Dobzhansky, Th., H . Levene, B. Spassky and N. Spassky. 1959. Release of genetic variability through recombination. III. Drosophila prosaltans. Genetics 44: 75-92. Dobzhansky, Th. 1959. Variation and evolution. Proc. Amer. Philosophical Soc., 103: 252--263. Dobzhansky, Th., C. Krimbas, and M. G. Krimbas. 1960. Genetics of natural populations. XXX. Is the genetic load in Drosophila pseudoobscura a mutational or a balanced load? Genetics 45: 741-753. Dubinin, N. P. 1946. On lethal mutations in natural populations. Genetics 31: 21-38. Dunning, Gordon M. 1957. Radioactive Contamination of Certain Areas in the Pacific Ocean from Nuclear Tests. Washington, D.C.: U .S. Atomic Energy Commission. Fisher, Ronald A. 1930, revised ed. 1958. The Genetical Theory of Natural Selection. New York: Dover Publications, Inc. Goldschmidt, E., J. Wahrman, A. Ledermann-Klein and R. Weiss. 1955. A two years' survey of population dynamics in Drosophila melanogaster. Evolution 9: 353-366. Goldschmidt, E., and R. Falk. 1959. On the dominance of "recessive" lethals. The Bulletin of the Research Council of Israel, Vol. BS: 1-8. Greenberg, Rayla and James F. Crow. 1960. A comparison of the effect of lethal and detri­mental chromosomes from Drosophila populations. Genetics 45: 1153-1168. Gregg, Thomas G. 1959. Genetic studies of irradiated natural populations of Drosophila. III. Experimental populations of Drosophila ananassae derived from irradiated natural popu­lations. Univ. of Texas Pub. No. 5914: 207-222. Gustafsson, Ake. 1946. The effect of heterozygosity on variability and vigour. Hereditas 32: 263-286. Haldane, J.B. S. 1937. The effect of variation on fitness. Amer. Nat. 71 : 337-349. -----. 1957a. The conditions for coadaptation in polymorphism for inversions. J. Gen. 55: 218-225. -----. 1957b. The cost of natural selection. J. Gen. 55: 511-524. Hiraizumi, Yuichiro and James F. Crow. 1960. Heterozygous effects on viability, fertility, rate of development, and longevity of Drosophila chromosomes that are lethal when homozygous. Genetics 45: 1071-1083. Hollaender, A. 1954. Radiation Biology, Vol. I, Parts 1 and 2 (and Vol. II, 1955). McGraw Hill: New York. Ives, Philip Truman. 1945. The genetic structure of American populations of Drosophila melanogaster. Genetics 30: 167-196. -----. 1954. Genetic changes in American populations of Drosophila melanogaster. Prue. Nat. Acad. Sci. 40: 87-92. Kaufmann, Berwind P. 1954. Chromosome aberrations induced in animal cells by ionizing radiations. Chapter 9 in Radiation Biology, Vol. I, Part 2, pp. 627-711, edited by A. Hollaen­der. McGraw Hill: New York. Kikkawa, H. 1938. Studies on the genetics and cytology of Drosophila ananassae. Genetica 20: 458-516. Moriwaki, D. 1935. Some mutant characters in Drosophila ananassae. Genetica 17: 32--46. -----. 1938. The genetics of some mutant characters in Drosophila ananassae. Jap. Jour. Genet. 14: 1-23. The University of Texas Publica~ion Muller, H. J. 1927. Artificial transmutation of the gene. Science 66: 84--87. -----. 1950. Our load of mutations. Amer. J. Human Genetics 2: 111-176. -----. 1954. The nature of the genetic effects produced by radiation, and the manner of production of mutations by radiation, Chapters 7 and 8 in Radiation Biology, Vol. I, Part 1, pp. 351-626, edited by A. Hollaender. McGraw-Hill: New York. -----. 1958. General survey of mutational effects of radiation. Chapter 6, pp. 145-177 of Radiation Biology and Medicine, ed. by W. D. Claus. Reading, Mass.: Addison-Wesley Pub~ C~fu~ . Oster, I. I., S. Zimmering and H. J. Muller. 1959. Evidence of the lower mutagenicity of chronic than intense radiation in Drosophila gonia. Science 130: 1423. Paik, Yong Kyun. 1960. Genetic variability in Korean populations of Drosophila melanogaster. Evolution 14: 293-303. Patterson, J. T., and Wilson S. Stone. 1952. Evolution in the Genus Drosophila. New York: The Macmillan Co. Pipkin, Sarah Bedichek. 1953. Fluctuations in Drosophila populations in a tropical area. Amer. Nat. 87: 317-322. Prout, T. 1952. Selection against heterozygotes for autosomal lethals in natural populations of Drosophila willistoni. Proc. Nat. Acad. Sci. 38: 478-481. Russell, W. L., Liana Brauch Russell, and Elizabeth M. Kelly. 1958. Radiation dose rate and mutation frequency. Science 128: 1546-1550. Russell, W. L., and Liane Brauch Russell. 1959. Radiation-induced genetic damage in mice. Progress in Nuclear Energy, Series VI, Vol. 2: 179-188. London: Pergamon Press. Santibanez, S. Koref, and C. H. Waddington. 1958. The origin of sexual isolation between different lines within a species. Evolution 11: 485--493. Spassky, B., N. Spassky, 0. Pavlovsky, M. G. Krimbas, C. Krimbas and Th. Dobzhansky. 1960. Genetics of natural populations. XXIX. The magnitude of the genetic load in populations of Drosophila pseudoobscura. Genetics 45: 723-740. Stern, Curt and Edward Novitski. 1948. The viability of individuals heterozygous for recessive leth11ls. Science 108: 538-539. Stern. C., G. Carson, M. Kinst, E. Novitski, and D. Uphoff. 1952. The viability of heterozygotes for lethals. Genetics 37: 413-449. Stone, Wilson S., Mary L. Alexander and Frances E. Clayton. 1954. Heterosis studies with species of Drosophila living in small populations. Univ. of Texas Pub No. 5422: 272--307. Stone. Wilson S., Mary L. Alexander, Frances E. Clayton, and Edna Dudgeon. 1954. The pro­duction of translocations in Drosophila virilis by fast neutrons from a nuclear detonation. Amer. Nat. 88: 287-293. Stone, Wilson S., Marshall R. Wheeler, Warren P. Spencer, Florence D. Wilson, June T. Neuen­schwander, Thomas G. Gregg, Robert L. Seecof, and Calvin L. Ward. 1957. Genetic studies of irradiated natural populations of Drosophila. Univ. of Texas Pub. No. 5721: 260-316. Stone, Wilson S., and Florence D. Wilson. 1958. Genetic studies of irradiated natural populations of Drosophila. II. 1957 Tests. Proc. Nat~ Acad. Sci. 44: 565-575. -----. 1959. Genetic studies of irradiated natural populations of Drosophila. IV. 1958 Tests. Univ. of Texas Pub. No. 5914: 223-233. Wallace, Bruce. 1956. Studies on irradiated populations of Drosophila melanogaster. J. Genetics 54: 280-293. -----. 1958. The role of heterozygosity in Drosophila populations. Proc. X Int. Congr. Genetics, Vol. I. Wright, Sewall. 1931. Evolution in Mendelian populations. Genetics 16: 97-159. -----. 1950. Discussion on population genetics and radiation. J. Cell. and Comp. Physiol. 35: 187-210. -----. 1955. Population genetics: irradiation effects. AECU-3065: 59-62. Wright, Sewall, Th. Dobzhansky, and W. Hovanitz. 1942. Genetics of natural populations. VIL The allelism of lethals in the third chromosome of Drosophila pseudcobscura. Genetics 27: 363-394. Wright, S., and Th. Dobzhansky. 1946. Genetics of natural populations. XII. Experimental reduction of some of the changes caused by natural selection in certain populations of Drosophila pseudoobscura. Genetics 31: 125-156. II. Fixed Heterozygosity in a Parthenogenetic Species of Drosophila 1 HAMPTON L. CARSON An animal or plant species which is widely distributed geographically and altitudinally inevitably comes in contact with widely diverse environments. Within such a species, there appear to be two possible modes of genetic adjust­ment to such environmental diversity. First, the species may develop a series of ecotypes or geographical subspecies in which natural selection tends to fix, possibly in the homozygous state, the genetic basis of local adaptation. Although gene flow may be retained between locally adapted isolates, selection may be strong enough to prevent disruption of the genetic basis of these adaptations. Continental species showing genetic gradients and local adaptation are exemplary of this type of adjustment. Secondly, a species may develop, perhaps in one original locaJ or marginal site, a key genotype, or possibly a small number of genotypes, which has homeostatic properties. Once achieved, such a genotype may enable its carrier to live in a variety of environments because it has the genetic basis of a general adaptability. This !ability, of course, is relative and can only operate within the genetic restrictions which have been imposed by previous adaptation. Cosmopolitan species and "weeds" appear to show this kind of adjustment to their environments. The purpose of the present paper is to present the facts of a case which exempli­fies the latter type of adjustment in the genus Drosophila. Drosophila mangabeirai Malogolowkin is a member of the willistoni group of the genus (Malogolowkin 1958) and is widely distributed geographically, altitudinally and ecologically in the Neotropical region. Despite this, the species reproduces by obligatory thely­tokous parthenogenesis and is a fixed heterozygote. These conditions appear to impose great genetic rigidity and to preclude local ecotypic adaptation. MATERIAL AND METIIODS From 1957 to 1960, 97 specimens of Drosophila mangabeirai have been cap­tured in Puerto Rico, Costa Rica and Panama. When added to the specimens reported by Carson et al. (1957) the total now stands at 116; 113 of these were females (Table 1). Wherever possible, wild-caught females were placed in isola­tion in individual culture tubes immediately after capture. Every one or two days the individuals were transferred to new food. D. mangabeirai females do not feed well on either Fleishmann's yeast or the special yeast (Y-2; Wagner 1944) used at the University of Texas laboratory. Good results have been obtained however, by yeasting the cultures very lightly with a dilute mixture of clover honey, Fleishmann's yeast and water. When larvae appeared in the vials, at­ 1 Department of Zoology, Washington University, St. Louis, Missouri; Research Associate, Genetics Foundation, The University of Texas. The University of Texas Publication tempts were made to obtain salivary gland smears of the offspring of each wild female, using the usual aceto-lacto-orcein technique. Following this, offspring were reared from each female. These were examined under the binocular micro­scope and the sex of each individual determined. Some females which were damaged at capture were killed and their spermathecae and seminal receptacles examined for the presence of spermatozoa. OBSERVATIONS Each larva examined, from all females tested, has shown the same identical karyotypic configuration described in Carson et al. 1957 from a laboratory stock originating from a single female captured in El Salvador. These data are detailed in Table 2. Comparison with Table 1 will show that this investigator was able to check cytologically the offspring of only a little more than one-third of the wild females caught. A number of wild-caught females were either over-etherized or were damaged beyond recall in the actual netting procedure. Others were senile at capture; still others produced few or no offspring. The very low fecundity of this species (Carson et al. 195 7; Murdy and Carson 1 ~59) makes cytological work on the progeny of individual females difficult. In general, five or six larvae were examined from each female; in some instances, however, the test depended on fewer than this; in three cases only one larva was examined. The salivary gland configuration observed in all 270 smears is figured in Carson et al. 1957. For brevity, it will be called the "El Salvador karyotype". Reference to Table 2 will reveal the geographical distribution of this fixed karyo­type; this is further depicted in Figure 1. There is a complex of three heterozygous inversions, one set manifested by an overlapping inversion configuration in one salivary chromosome arm and another nearly terminal inversion in a different chromosome arm. These two arms break out of the chromocenter joined together and show at their bases a large amount of heterochromatin. Homologizing with D. bocainensis (Carson 1954), D. nebulosa (Pavan 1952) and D. willistoni (Dobzhansky 1950) , this chromosome resembles autosome No. 2 of this species group, which universally has large amounts of heterochromatin at the base of its arms. The X chromosomes of all these species, relatives of D. mangabeirai, are also V-shaped. The salivary chromosome arms are lightly conjoined but there is little conspicuous heterochromatin. This leads one to believe that the other sali­vary gland chromosome of D. mangabeirai which is present in every nucleus as a sectionally homozygous, 2-armed element without heterochromatin at the base is in reality the X-chromosome. The single structurally homozygous element of the D. mangabeirai salivary gland cell nucleus apparently represents chromo­some 3, which is a rod-shaped chromosome throughout the willistoni group. As can be seen from Table 2, all 270 smears were of female larvae. Further­more, all emerging adults in the laboratory, a total actually counted of 4362, were females. In addition, several thousand more individuals have been screened during work with experimental populations of this species, as will be reported in a later publication. Massive doses of X-rays were applied during these popu­lation cage experiments. No males were observed in either control or experi­mental populations. The species has been used to demonstrate parthenogenesis TABLE 1 Recent collections of Drosophila mangabeirai Locality Date Number and sex Habitat Palmar, Costa Rica Changuinola, Rep. of Panama Boquete, Rep. of Panama Bano Colorado Island, Panama Canal Zone Frijoles, Panama Canal Zone Mayaguez, Puerto Rico Total reported in Carson et al., 1957 Total specimens collected June 29-July 8, 1959 July 11, 1959 August 3, 1958 July 21, 1959 May 13, 1960 August 20-23, 1960 October 11-19, 1957 50 females 2 females 1 male 2 females 1 male 39 females 2females 18 females 1 male 113 females 3 males Near sea level; swept in small private orchard with banana, plantain and guava ( Psidium guajava) trees Elevation about 200 feet; swept in cacao finca Elevation about 3800 feet; swept in hillside forest; females were taken over fallen orange flowers Swept just inside rain forest; Snyder Molino Trail Elevation about 70 feet; at edge of Gatun Lake, swept over small palm fruits (Guilielma gasipaes) on ground under a non-fruiting mango tree Near sea level Varying altitudes, from 4000 feet to near sea level. Male was swept at Barro Colorado Island T ABLE 2 Salivary gland chromosome examinations of F 1 and later generation larvae from wild-caught females of Drosophila mangabeirai. Without exception all larvae examined have been cytologically identical. All were female and were heterozygous for a complex of three inversions in one of the chromosomes (probably autosome No. 2). This condition is called the "El Salvador karyotype". No. of larvae showing the "El Salvador" karyotype: Number of No. of wild adult females Locality where wild females tested Ffrom from later emerging in F fema les were captured cytologically wild wild generation and later generations 1 F2 1 El Salvador, San Salvador 50 2365 Palmar, Costa Rica 24 121 45 1761 Changuinola, Panama 1 4 Boquete, Panama 1 6 Frijoles, Panama Canal Zone 16 37 7 236 Total 43 168 7 95 4362 The University of Texas Publication • EL SALVADOR KAR YOTYPE EB NO CYTOLOGICAL DATA FIG. 1. Geographical distribution of Drosophila mangabeirai. in other laboratories but no case of a male being produced has come to the atten­tion of the writer. From collections of wild flies at Palmar, Costa Rica and Frijoles, Panama Canal Zone, fourteen freshly-caught females were found, upon sorting, to have been seriously damaged during collecting. As they were still alive, each was dissected and both spermathecae and the seminal receptacles examined for the presence of spermatozoa. None of these females was found to be inseminated. Although the original El Salvador strain comes from an elevation of 2000 feet on the west coast of Central America, Carson et al. (1957) pointed out that other specimens had come from such diverse regions as 4000 feet in the mountains of Nicaragua, near sea level on the island of Trinidad and in Honduras and Brazil. Ecological sites were, of course, equally varied. The new locations show similar diversity (Table 1; Fig. 1.). Palmar, Costa Rica lies close to sea level on the west coast in an area of exuberant rainforest, not far from the Golfo Dulce. The flies were captured in a single small finca close to a patch of residual forest. Changuin­ola, Panama is likewise in the lowlands, but is on the east side of the Isthmus of Panama. Boquete, Panama represents still another quite different site; the area: supports an extraordinarily complex floristic association on the volcanic soils of the west slope of the mountains at an altitude of close to 3800 feet. Near Frijoles, Panama Canal Zone, a good number of specimens were caught on a point extend­ing into Gatun Lake. This site was a fruit orchard but despite repeated efforts over a period of three days, the species could be captured only in one confined area of not more than a hundred square meters. The extreme localization of the flies at both Palmar and Frijoles is a striking feature. Such foci could certainly be easily missed, especially as the species does not seem to be associated with any Carson: Fixed H eterozygosity in Drosophila particular micro-habitat, fruit, exposure or site. The occurrence of the flies in small private fincas, which are not frequently sprayed with insecticide and often have remaining forest trees and some undergrowth is hardly significant. Such sites have been found to be consistently good for collecting a very large variety of tropical species of this family (see Heed 1957). The possibility of strict localiza­tion of D. mangabeirai for whatever basic reason, may explain the rarity of the species in collections, despite its wide distribution. Although no males have ever been obtained in the laboratory, three specimens have been captured in nature (Table 1). The first of these ( 1955, Barro Colorado Island) was studied in detail by Malogolowkin (1958) who pictured the genitalia. Both of the other males conform closely to this description and are morphologic­ally very similar to the females, including the presence of an unique row of minute spines on the inner apical side of the fore-femur. On May 13, 1960, a single male was captured alive on Barro Colorado Island. This specimen was netted by general sweeping just inside the rain forest on the Snyder Molino Trail by Mr. L. 0 . Powers and was kindly made available for study by Dr. Sarah B. Pipkin. Dr. Pipkin placed this male with 28 females of the thelytokous stock P-29 (Palmar, Costa Rica) and sent these flies in a single culture tube to the writer in St. Louis. Upon arrival on May 16, the flies were in excellent condition. Twenty-five females and the male were active and vigorous and the tube contained healthy larvae. The male was observed repeatedly to make court­ing movements towards a number of females, but no copulations were observed. Until May 20, the male was kept with the original females, with a daily change to fresh food. Following this, 17 of the females were placed individually into cul­ture tubes and the remaining 8 were kept with the male for another week. After this, the remaining live females, which had previously been separated (there were 13 left alive) were put back with the male for another week and the other 8 separated from the male and isolated. On June 3, the male was placed with 7 freshly-emerged P-29 females, with which the male was kept for an additional 15 days. Six of these latter females were dissected on June 18 and none was found to be inseminated although all but one had large ovaries with filamented eggs and the group was producing larvae. Offspring were obtained from females which had been confined in the same culture tube as the male. A total of 1634 adults were reared from eggs laid by these females; all were female. In addition, smears of salivary gland chromosomes of 80 larvae were examined; all of these were female and showed the El Salvador karyotype. Accordingly, there is no evidence that the male ~ither inseminated any female with which he was placed or sired any off spring. The conclusion that this male was sterile seems inescapable. On June 18, more than one month after arrival in the laboratory, the male, which was still active, was sacrificed. One testis was examined in saline solution and showed motile sperm; the other testis was smeared in aceto-orcein but no chromosomal figures could be found. The genitalia, of which a permanent mount was made by Dr. L. E. Magalhaes, were found to conform closely to those of the other Barro Colorado Island male described by Malogolowkin (1958). This speci­men, together with the slide of the genitalia, was deposited in the collection of the Genetic5 Foundation, The University of Texas. The University of Texas Publication Through the kindness of Dr. M. R. Wheeler, the three male specimens in the University of Texas collection were made available to the writer so that an examination of the size of their wing cells could be made. One wing of each male specimen was placed under a stereoscopic microscope giving a magnification of about 115 diameters. A wing of a dried and pinned female specimen from Frijoles, Canal Zone, was brought into the same microscope field at the same focal level so that the hair 'rows in comparable wing cells could be directly compared. No difference between the size of the cells in any of the males and the female was observed. The standard female was then compared to five other females from Frijoles and found to be the same as them. Accordingly, all flies thus examined display what is apparently diploid cell size. DISCUSSION Drosophila mangabeirai appears to be wholly thelytokous in its mode of repro­duction throughout central America. All females which have produced offspring have given rise only to females and have shown the same fixed sectionally­heterozygous condition (The El Salvador karyotype). This condition is known to be made possible by automictic fusion between two of the four haploid nuclear products of meiosis in the egg (Murdy and Carson 1959). The mode of fusion is such that the two nuclei which come together are, in the majority of cases, derived from different secondary oocytes. This event in the absence of crossing over between the inversions and the centromere, assures the reconstitution of the complex sectional heterozygote. One may conclude, therefore, that D. ma.nga­beirai almost universally consists of fixed female heterozygotes resulting from a system of balanced lethals. This condition is strikingly similar to that found by Stalker (1956) in the parthenogenetic fly Lonchoptera dubia, except that, in the latter species, four different fixed sectionally heterozygous karyotypes are known which coexist widely throughout the continental United States. The El Salvador karyotype of D. mangabeirai is now known to be distributed from El Salvador to central Panama, a direct distance of over 700 miles (Fig. 1). It has been found on both sides of the isthmus and from sea level to elevations near 4000 feet. Local genetic adaptation to these very diverse habitats would appear to be precluded by the genetic system observed. One may conclude, there­fore, that at the present time, D. mangabeirai exists widely, if not universally, with a fixed total genotype which confers on its carriers some general adapta­bility through which it is able to exist in this variety of habitats. In this conne.c­tion, its adjustment to the environment resembles that of many weed species with fixed genetic conditions (e.g. Taraxacum or Oenothera). The presence of occasional males in nature requires explanation. The discovery in 1960 of a "brood" of females having the El Salvador karyotype at Frijoles, Canal Zone, speaks against the hypothesis that the males· belong to a different, perhaps sibling, species of which females have not yet been found. The location where these females were collected is just across the lake from Barro Colorado Island, where two of the three known males were collected. One female from Boquete, Panama, the source of the other male, also showed the El Salvador fixed karyotype. The writer believes that most of the evidence indicates that the males are indeed the same species as the females. Under as yet unspecified conditions, Carson: Fixed Heterozygosity in Drosophila it is deemed possible that some sort of non-disjunctional event involving the X chromosome could occur so that a male phenotype might result. Experiments in the laboratory designed to induce such an event have not been done. Simple X­chromosome non-disjunction followed by the proper type of fusion to give a diploid XO male (see Stalker 1954) is a possible explanation and is consistent with observations of wing cell size in the males. The fact that the one apparently sterile male displayed motile spermatozoa in the testis raises problems for this hypothesis, as XO males in D. melanogaster and D. parthenogenetica, for ex­ample, do not have motile sperm. XO males have apparently never been recog­nized in any Drosophila species which has V-shaped X chromosomes. A less likely explanation is that non-disjunction might be followed by fusion of two XA and one A meiotic products. This would produce a 2X3A condition which would be expected to be a male-type intersex. Intersexes of this type in Drosophila americana are sterile yet have motile sperm (Stalker 1942) . The fact that the ploidy of the males appears identical to that of diploid females virtually elimi­nates this possibility. Triploid females, furthermore, apparently do not occur in D. mangabeirai (Carson et al. 1957), making fusion between three pronuclei in the egg an unlikely event as prerequisite to the formation of the observed males. SUMMARY 1. Drosophila mangabeirai, a member of the willistoni group of the genus, appears to reproduce widely throughout its range by thelytokous automictic diploid parthenogenesis. One hundred and thirteen females and three males have been collected in nature. The only male which has been tested was sterile and is chromosomally unknown. The evidence indicates that the males are the same species as the females and have a diploid XO condition. 2. Only females are produced in the laboratory; this statement is based on individual examination of more than 4000 flies over a period of several years. 3. Every salivary gland chromosome preparation examined (total 270) from both larvae from wild females and those from later generations, were female and showed a universally identical sectionally-heterozygous three-inversion differ­ence. This involves both arms of what is apparently chromosome 2. The con­dition is designated as the El Salvador karyotype. . 4. The species is widely distributed geographically, altitudinally and ecologic­ally. The El Salvador karyotype has been recognized from the type locality south through Costa Rica to the Panama Canal Zone, from sea level to an altitude of about 4000 feet. The fact that the species is universally an automictic fixed heterozygote essentially precludes the possibility that the species could show genetic gradients or ecotypes. Rather, it is suggested that the species is able to exist in such widely different habitats because the El Salvador karyotype confers homeostatic properties on its carriers such that they exploit these varied environ­ments by means of a general adaptability. ACKNOWLEDGMENTS Collection of the specimens for this study was made possible by a grant from the National Science Foundation (NSF-G 4999) to Dr. Marshall R. Wheeler, The University of Texas Publication whom I wish to thank for his personal interest and valuable assistance. I wish also to express my appreciation to my colleagues Drs~ William B. Heed and Marvin Wasserman who collected some of the specimens and to Dr. Sarah B. Pipkin for making available the specimen from Barro Colorado Island. Thanks are also due to Dr. M. Monyihan of the Canal Zone Biological Area and to the United Fruit Company for its hospitality while collections were being made at Palmar and Changuinola. The writer would also like to take this opportunity to acknowledge the universally friendly assistance of dozens of Latin American farmers without whose hospitality and kindness collections such as these would have been impossible. Analysis of the data in St. Louis was done under grant No. G-7441 from the National Science Foundation. LITERATURE CITED Carson, H. L. 1954. lnterfertile sibling species in the willistoni group of Drosophila. Evolution 7: 148-165. Carson, H. L., M. R. Wheeler and W. B. Heed. 1957. A parthenogenetic strain of Drosophila mangabeirai Malogolowkin. Univ. Texas. Publ. 5721 : 115-122. Dobzhansky, Th. 1950. The chromosomes of Drosophila willistoni. J. Hered. 41: 156-158. Heed, W . B. 1957. Ecological and distributional notes on the Drosophilidae (Diptera) of El Salvador. Univ. Texas Puhl. 5721 : 62-78. Malogolowkin, C. 1958. Sohre a genitalia dos drosofilideos. V. A genitalia masculina em "D. mangabeirai" (Diptera, Drosophiladae). Rev. Brasil. Biol. 18: 443-445. Murdy, W. H. and H. L. Carson. 1959. Parthenogenesis in Drosophila mangabeirai Malog. Amer. Nat. 93: 355-363. Pavan, C. 1952. Chromosomal variation in Drosophila nebulosa. Genetics 31: 546-557. Stalker, H. D. 1942. Triploid intersexuality in Drosophila americana Spencer. Genetics 27: 504-518. ----. 1954. Parthenogenesis in Drosophila. Genetics 39: 4-34. ----. 1956. On the evolution of parthenogenesis in Lonchoptera (Diptera) . Evolution 10: 345-359. Wagner, R. P. 1944. The nutrition of Drosophila mulleri and D. aldri~~i. Growth of the larvae on cactus extract and the microorganisms found in cactus. Univ. Texas Bull. 4445: 104-128. III. Cytological Studies of the Repleta Group of the Genus Drosophila: III. The Mercatorum Subgroup 1 MARVIN WASSERMAN INTRODUCTION The mercatorum subgroup of the repleta group consists of two species, Dro­sophila mercatorum and Drosophila paranaerzsis. These species have been investi­gated cytologically by Wasserman (1954) and Wasserman and Wilson (1957) . Recently collected strains which sample a new geographical area, Brazil, Bolivia and Chile, have considerably modified our concept of this subgroup. Although data from large areas of the ranges of these species are lacking or at best frag­mentary, it is felt that a summary of all the available information should be presented along with the newer data. Tables 1 and 2 list the strains examined, the new, previously unpublished, strains having an asterisk next to the stock number. The methods used are described in Wasserman ( 1954) . RESULTS All strains of D. paranaensis have a dark, very distinct color pattern. D. mer­catorum is variable, ranging from a diffuse pattern to one similar to that of TABLE 1 Chromosome Configuration of D. paranaensis. Standard sequence: repleta X; 2d, 2e, 2t, 2v3; 3b, 3f; repleta 4; repleta 5 Strain Locality x 2 3 4 2251.5 Gomez Farias, Tamaulipas, Mexico i/+ 2263.13 Tezuitlan, Mexico i 1793.6 Tepatitlan, Mexico i/ ig H 48.5 E. A. P., Zamorano, Honduras ii+ H 29.13 San Salvador, El Salvador ii+ H 34.20 San Salvador, El Salvador H 62.13 La Palma, El Salvador H 56.94 El Recreo, Nicaragua H 73.6 Turrialba, Costa Rica H163.16 Turrialba, Costa Rica H 75.10 Heredia, Costa Rica i/ig H303.19* Las Cruces Trail, Canal Zone i/ ig H360.90* Boquete, Panama H360.93* Boquete, Panama H101.12 Palmira, Colombia ilig H194.44 Villavicencio, Colombia H186.33 Sierra Nevada de Santa Marta, Colombia b/+ H113.5 Trinidad m2 i/+, gl+ bl+ H332.11 * Trinidad m2 il+,g/+ b/+ .. New strains. 1 Present address: Zoology Department, University of Melbourne, Melbourne, Australia. 64 The University of Texas Publication TABLE 2 Chromosome Configuration of D. mercatorum. Standard sequence: repleta X; 2d, 2e, 2t, 2u, 2v; 3b, 3f; repleta 4; repleta 5 Strain Locality x 2 3 4 5 2097.3 Lanikai, Hawaii v3 h 2210.8 Lanikai, Hawaii v3 h/ hg 0-20.3 Poa Moho, Oahu, Hawaii v3 h 2370.5 Oahu, Hawaii v3 h 955.4B S• ~i,1\ll811,....J111tJll11~1~1~i~!A»·nc!7fJn\'" 1ra·1~,l ~~, l!ali'd? ' Iii;~ lD... ii. . . ' . "' ~: ~ I, I ;i,,j ~l-lllWlll lll~JJ/P~""il.I'"' .;~ ~ 6 17 I I I 2 I 3 I 4 I5 ~ 1'~1 2 311 6 I I I 2 I 3 I 4 15 I CD DE F G G H .Q. Fig. 7 ~ ~ r------H/ +-------~ --NI+--~ 3~ I I '.) I~ " I£ Iz , 1, ~1 ' I " I€ I §' ..... n•WllIm~11J1n;u·111 1·mn.~l''1~1 §" 3 1 .._____ + , i.w,1" ..11H1 llll/..... m;:·~l ;1WJU 1 11 FIGs. 6-8. Salivary gland chromosomes of the mercatorum subgroup. See text for explanation. FIG. 6. The standard second chromosome of D. paranaensis with d,e,t,v3. FIG. 7. The standard second chromosome of D. mercatorum with d,e,t,u,v. FIG. 8. The standard third chromosome of the mercatorum subgroup with b,f. Wasserman: The Mercatorum Subgroup D. paranoensis D. merea forum 2m2/t 4b/t 2s'l+ 2v7+ Xi/t . 3g/t 3i/+ 5k/t 3g/t 3h/t 3n/t / ~ / 2-3F 2d 2e 2t MERCATORUM SUBGROUP 3f 3g/t 2l1+ ANCESTOR r ~----rep/efo standard l Xa Xb Xe 2a 2 b 3b PRIMITIVE SEQUENCE Frc. 9. Cytological evolution in the mercatorum subgroup. Heterozygous inversions are indi­cated by being shown over a plus sign. Complete explanation in the text. where it is heterozygous. The standard second chromosome was found at one locality, Montero, Bolivia, where it is present with v3 and v3s3 chromosomes, all other localities having the 2v3 fixed homozygous. The 2s3 arrangement is present heterozygous in Brazil and Chile as well as Bolivia. All mercatorum populations have the 3h sequence. The standard third and 3n are restricted to Brazil and Bolivia, whereas the 3g sequence has a disjunct distri­bution, Brazil, Colombia and Hawaii. The above data demonstrate a marked difference in the cytological structure of the populations from different geographical areas. Primitive chromosomes and newer arrangements are found together in populations from Brazil and Bolivia,. the six strains from Brazil having a total of five heterozygous inversions: 2v3/ 2v3s3; 3 standard/3g/3h/ 3n; five standard/5k; and the Montero, Bolivia strains having four heterozygous inversions: 2 standard/2v3/2v3s3; 3 standard/3h/3n. These Brazilian and Bolivian strains contrast strongly with the populations found in the Andes of Bolivia, Colombia, Peru and north to the United States and Hawaii, a total of eighteen strains from thirteen localities. In the latter thirteen localities the 2v3 and 3h arrangements are homozygous: eleven of these contain no other sequences; one Hawaiian strain is heterozygous for 3g, and one Colom­bian population is homozygous for 3g and Xi. Two subspecies have been described in D. mercatorum, Wharton (1944) con­sidering the United States and Mexican populations as D. m. mercatorum, and The University of Texas Publication the Brazilian populations as D. m. pararepleta. Subsequent tests have shown that these two subspecies, although fully fertile between each other, differ in the type of offspring produced when crossed top. paranaensis. Only sterile offspring were obtained in crosses involving m. mercatorum, while the Brazilian m. pararepleta produced some fertile females in the same type of crosses. In view of the cytological data presented above, we would define the subspecies pararepleta as the polymorphic populations occurring in the Brazilian and Bolivian lowl~nds. The mercatorum subspecies consists of the homozygous, or essentially homozygous, populations in the Andes and north to the United States. The Chilean population, which is isolated geographically from both major popu­lations, is polymorphic for 2s3, a gene sequence which is known elsewhere only from Brazil and Bolivia, and therefore probably represents a population of rxzra­repleta. D. mercatorum probably arose in the Brazilian and Bolivian lowlands, where the origin of new gene sequences allowed for a certain amount of ecological diversification. The spread westward across the Andes and northward has been limited to those flies homozygous for the 2v3 (an old sequence) and the 3h (a new sequence) chromosomes, some of these flies also carrying the ancient 3g arrange­ment. Since it is likely that some gene flow exists between the lowlands and the Andes, the inability of many of the lowland gene sequences to migrate is probably due to their being ill-adapted to the Andean conditions even when present in the heterozygous condition (Wasserman 1960) . The basic sequence of D. paranaensis is the same as the ancestral type except the 2v3 inversion has been fixed (Figs. 3 and 6). Aside from this we have little to add to the cytological picture presented by Wasserman and Wilson ( 195 7). The standard 3, the 3g and the 3i sequences are widespread, occurring from Mexico to Trinidad. Although the standard 3 and/ or the 3g are lacking in several popu­lations, we do not have sufficient d~~a to ascertain whether or not this is due to small collections. The Trinidad strains are homozygous for 2m2 (Fig. 6) and heterozygous for 4b, an inversion which is also present in the Santa Marta, Colombia strain. No information concerning gene sequences is available from Brazil, Bolivia, and Venezuela. Metaphase karyotype variability as to the form of the dot chromosome exists in both species. Large dots or small rods may characterize the pararepleta sub­species of D. mercatorum, whereas small metacentrics are found in the merca­torum subspecies. In paranaensis the geographical distribution of the karyotype variations found in our strains also indicates that each type may represent a geographical race. However the absence of samples from certain critical areas of the known distribution of the species prevents us from accurately defining the limits of these races. The known distributions are: Type (a), characterized by the presence of small dots, in Brazil, Trinidad and eastern Colombia; Type (b), with the dots replaced by small metacentrics, in northeastern Colombia and Costa Rica; Type (c), with dots replaced by large metacentrics, in western Colombia, Nicaragua, El Salvador, Honduras and Mexico. Wasserman: The Mercatorum Subgroup SUMMARY (1) Two species, D. mercatorum and D. paranaensis, the only members of the repleta group of the genus Drosophila which have been placed in the mercatorum subgroup, have been investigated cytologically, genetically and morphologically. Samples of populations have been taken from major parts of most of their known distribution ranges. (2) The evolution of the mercatorum subgroup as reflected in the cytological changes can be summarized as follows: a) The fixation of five inversions in the X and the second chromosomes chang­ing the hypothetical repleta group primitive sequence, Xa, Xb, Xe, 2a, 2b, 3b to one differing from D. repleta by only 3b. b) The splitting of this species, one population giving rise to the melanopalpa subgroup of which D. repleta is a member, and the other evolving into the merca­torum subgroup during which there occurred a fusion between chromosomes 2 and 3; and an acquisition of four homozygous inversions, 2d, Ze, 2t and 3f, and two heterozygous inversions, 3g and 2v3 • c) The splitting of this ancestral species and further divergence resulting in the formation of two daughter species: D. paranaensis in which the 2v3 became homozygous, while the 3g sequence has remained heterozygous; and D. merca­torum in which both the 2v3 and the 3g arrangements are still heterozygous, while two newer inversions, 2u and 2v, have become homozygous. d) The division of D. mercatorum into two geographically defined subspecies which differ in their cytological composition and also in the type of progeny produced when crossed with D. paranaensis. e) The division of D. paranaensis into three chromosomal races differing in the amount of heterochromatin present in the dot element, but whose geographi­cal distributions are not well known. LITERATURE CITED Clayton, Frances E., and Calvin L. Ward. 1954. Chromosomal studies of several species of Drosophilidae. Univ. Texas Pub. 5422: 98-105. Clayton, Frances E., and Marvin Wasserman. 1957. Chromosomal studies of several species of Drosophila. Univ. Texas Pub. 5721: 125-131. Dreyfus, A., and R. de Barros. 1948. Mutations chromosomique chez les hybrides de Drosophila mercatorum pararepleta X D. paranaensis. S. Paulo Medico 21: 11-18. ----. 1949. Sex-ratio chez certains hybrides interspecifiques de Drosophila et son interpre­tation par !'analyse des chromosomes salavaires. Suppl. La Ricerca Scientifica, Anno 19: 1-13. Ward, Calvin L. 1949. Karyotype variation in Drosophila. Univ. Texas Pub. 4920: 70-79. Wasserman, Marvin. 1954. Cytological studies of the repleta group. Univ. Texas Pub. 5422: 130-152. ----. 1960. Cytological and phylogenetic relationships in the repleta group of the genus Drosophila. Proc. Nat. Acad. Sci. 46, no. 6: 842-859. Wasserman, Marvin, and Florence D. Wilson. 1957. Further studies on the repleta group. Univ. Texas Pub. 5721: 132-156. Wharton, Linda. 1942. Analysis of the repleta group of Drosophila. Univ. Texas Pub. 4228: 23-52. ----. 1944. Interspecific hybridization in the repleta group. Univ. Texas Pub. 4445: 175-193. IV. Cytological Studies of the Repleta Group of the Genus Drosophila: IV. The Hydei Subgroup MARVIN WASSERMAN1 INTRODUCTION Wharton (1944), using mainly the genetic results obtained from interspecific crosses among the members of the repleta group of the genus Drosop/Ula, estab­lished the hydei subgroup, including within it four species. A morphological con­cept of the subgroup increased the number of forms to six species (Wheeler, 1949). Recent studies including a salivary gland chromosome analysis of the strains available in the University of Texas laboratory (Table 1), have resulted in more data concerning three of the six species included in the subgroup by Wheeler ( 1949), Drosophila hydei, Drosophila nigrohydei and Drosophila bi/urea, and also resulted in the discovery of two new species, Drosophila eohydei and Drosophila neohydei. Published information concerning Drosophila hydeo­ides (Wharton, 1944) and Drosop/Ula novemaristata (Dobzhansky and Pavan, 1943), two of the species not examined by the author, will be incorporated into this report. The eighth species, Drosop/Ula pachea, has been collected only once, and aside from the original morphological description (Patterson and Wheeler, 1942), no other information is available. MORPHOLOGY The hydei subgroup is morphologically distinguished from the other subgroups of the repleta group in that it consists of species which have extensive coiling in both the testis (22 to 51 coils) and the ventral receptacle (245 to 735 coils). The non-hydei repleta group species exhibit much less coiling, the highest number reported being 16 and 116 coils in the testis and ventral receptacle, respectively, of D. melanopalpa (Patterson and Wheeler, 1942). Specific identification is difficult, several of the species being very similar in external appearance. Table 2 lists some of the morphological traits which are of some use, the best interspecific differences within the subgroup often being found internally and in the male external genitalia. Figures 1through9 show diagram­matically a spermatheca, the penis and the male genitalia of several of the species. Collections in the Neotropical Region have yielded strains of two new species whose descriptions, due to the species' likeness to D. hydei, are presented in some­what abbreviated form, the distinguishing characters being stressed. Drosophila eohydei, new species (Fig. 4;7) Externally this species is very similar to D. hydei. Table 2 lists several dis­tinguishing characters, of which the number of coils present in the ventral recep­ 1 Present address: Zoology Department, University of Melbourne, Melbourne, Australia. The University of Texas Publication TABLE 1 Chromosomal Configuration of Stocks Studied x 2 3 4 5 Species Stock Locality hydei 0-25.9 Oahu, Hawaii a,b,c a,b,z b + + 2310.4 Oahu, Hawaii a,b,c a,b,z b + + 2087.1 Israel a,b,c a,b,z b + + 1732.2 Byblos, Lebanon a,b,c a,b,z b + + 2372.13 Queensland, Australia a,b,c a,b,z,a2/+ b + + 2357.5A Prescott, Arizona a,b,c a,b,z b + + 2357.5B Prescott, Arizona a,b,c a,b,z b + + 2358.9C Patagonia, Ari zona a,b;c a,b,z,a2 /+ b + + 2358.9D Patagonia, Arizona a,b,c a,b,z,a2/+ b + + 2360.3A Cave Creek, Arizona a,b,c a,b,z b I I + 2360.3B Cave Creek, Arizona a,b,c a,b,z,a2/+ b + + 2354.5A Palisade, Colorado a,b,c a,b,z b + + 1385.11 San Juan de la Punte, Veracruz, Mexico a,b,c a,b,z,a2/+ b + + 1797.8 Mexico City, Mexico a,b,c a,b,z b + + 2261.1 Tehuacan, Puebla, Mexico a,b,c a,b,z b + + H 45 .1 San Salvador, El Salvador a,b,c a,b,z b + + H 62.28 La Palma, El Salvador a,b,c a,b,z b + + H 50.1 La Lima, Honduras a,b,c a,b,z b + + H1 85.12 Sierra Nevada de Santa Marta, Colombia a,b,c a,b,z b + + H1 94.23 Villavicencio, Colombia a,b,c a,b,z b + + 2395.8 Pisco, Peru a,b,c a,b,z b + + 23 75.7 Santiago, Chile a,b,c a,b,z b + + eohydei H191 .47 Bucaramanga, Colombia a,b,c a,b,z,h3/+ b I I + H191.67 Bucaramanga, Colombia a,b,c a,b,z b + + H186.58 Sierra Nevada de Santa M arta, Colombia a,b,c a,b,z,h3/+ b + + H 18.9 H 62..57 Zamorano, Honduras La Palma, El Salvador a,b,c a,b,c a,b,z,h3/+ a,b,z,h3g3/+ b b + + + + neohyclei H207.26 Caracas, Venezuela a,b,c,m/+ a,b,z,f3/+ b + + bifurca A2. Patagonia, Arizona a,b,c a,b b + + nigro­ A8 .10 25 10.1 Aravaipa Valley, Arizona Guatemala City,Guatemala a,b,c a,b,c a,b a,b,q3,x/+ b b + + + + hydei 2506.1 Guatemala City, Guatemala a,b,c a,b,q3,x/+ b + + 2259.1 Real de Monte, Hidalgo, Mexico a,b,c a,b,x/+ b + + 2155.4 Coron'lcl o N .F., Arizona a,b,c a,b,x,y b + + 25 19.3 Pachuca, Puebla, Mexico a,b,c a,b,q3/ +,x b + + H376.30 San Cristobal, Chiapas, Mexico a,b,c a,b,q3 b + + H376.6 San Cristobal, Chiapas, Mexico a,b,c a,b,q3 b + + tacle (280 to 295 ) and in the testes (16 inner and 16 outer) are the most useful. Figure 4 diagrams the external male genitalia and Figure 7 illustrates the penis and one of the very distinctive female spermathecae. The metaphase karyotype, Figure 11, consists of one pair of dots, four pairs of autosomal rods, a rod-shaped X and a very short rod-shaped Y. chromosome. The salivary chromosomes will be described below. The type locality is Bucaramanga, Colombia (strain H 191.67). One holotype and nine paratypes are deposited in The University of Texas collection, along with several slides from the type strain for identification Wasserman: The Hydei Subgroup TABLE 2 Morphological Characters Coils in Wing characters Arista Sterno-ventral c. I. 4vn 5x 4c formula index receptacle Coils in testes D. eohydei 3.24 1.73 1.09 0.81 3-4 2 .71 280 to 295 16 inner, 16 outer D. neohrdei 3.28 1.76 0.97 0.86 4 -2 .84 340to 400 24 inner, 27 outer D. hrdei 3.30 1.57 0.97 0.76 3 -2 .81 245 25 inner, 18 outer D. bi/urea 3.89 1.28 1.28 0.56 4 -3 .8 735 23 inner, 28 outer D. nigrohydei 3.29 1.51 0.92 0.74 3 -2 .83 340 10 inner, 12 outer D. hydeoides* 3.5 1.7 1.0 0.78 3 -2 .88 330 20 inner, 25 outer D. pachea* 3.1 1.6 1.21 0.85 3 -2 .66 330 16 inner, 13 outer 4 D. novemaristata+ 3-3.3 1.5-1.6 1.1 -3 .8 340 24 inner, 17 outer •Taken from description in Patterson and ~'heeler (1942) . t Taken from description in Dobzhansky and Pavan ( 1943) . purposes. Other known localities are Sierra Nevada de Santa Marta, Colombia; Zamorano, Honduras; and La Palma, El Salvador. This species had been pre­viously designated as species A (Wasserman, 1960). Drosophila neohydei, new species (Fig. 3;8) Externally this species is very similar to D. hydei. Table 2 gives distinguishing characters such as the number of coils in the ventral receptacle (340 to 400) and in the testes (24 inner and 27 outer). Male genitalia, penis and female sperma­theca are shown in Figures 3 and 8. Metaphase chromosomes, Figure 11, consist of one pair of dots, four pairs of autosomal rods, a J-shaped X, and a large rod­shaped Y. Salivary chromosomes are described below. Type, and only known locality, Carpintaro, Venezuela (near Caracas). One holotype and nine paratypes are deposited in The University of Texas collections along with several slides. This species has been designated as species B (Wasserman, 1960) . A key to the species in the hydei subgroup follows. Wherever possible external characters are used and the known distributions of the forms are indicated, inter­nal traits being kept to a minimum. ( 1) Arista with 3 branches below, excluding the terminal fork------------------------2 Arista with 2 branches below, excluding the terminal fork ------------------------3 (2) Costal Index 3.0-3.3; 340 coils in ventral receptacle; testis with 17 outer and 24 inner coils; Brazil --------------------------------------------·---D. novemaristata Costal Index 3.6-3.9; 735 coils in the ventral receptacle; testis with 28 outer and 23 inner coils; United States and Mexico ----------------D. bi/urea The University of Texas Publication Frcs. 1-9. External male genitalia of species in the hydei subgroup: Fig. 1. D. bifurca. Fig. Z. D. nigrohydei. Fig. 3. D. neohydei. Fig. 4. D. eohydei. Penes and spermathecae: Fig. 5. D. nigro­hydei. Fig. 6. D. bifurca. Fig. 7. D. eohydei. Fig. 8. D. neohydei. Fig. 9. D. hydei. (3) Mesonotum brownish, indistinctly spotted; sterno-index 0.66; 330 coils in ventral receptacle, testis with 12-15 outer and 16 inner coils; Mexico ----··-·--·····--·················-·-·---·--·······-·----·-----·-··---········--·······--· D. pachea Not as above ··············-·············--······--·······-·····--······--····--·-··-····-···--····-····--·····--4 (4) Sterno-index about 0.7; 285 coils in the ventral receptacle; testis with 16 outer and 16 inner coils; penis and spermatheca as shown in Fig. 7; Wasserman: The Hydei Subgroup Central America and Colombia --------------------------------------------------D. eohydei Sterno-index greater than 0.8 ------------------------------------------------------------------------5 (5) Arista with 4 branches above, excluding the terminal fork; about 370 coils in the ventral receptacle; testis with 27 outer and 24 inner coils; penis and spermatheca as diagrammed in Fig. 8; Venezuela __ __D. neohydei Arista with 3 branches above, excluding the terminal fork; other characters not as above--------------------------------------------------------------------------------------------------6 (6) 245 coils in the ventral receptacle; testis with 18 outer and 25 inner coils; penis and spermatheca as shown in Fig. 9; world-wide----------------D. hydei 340 coils in the ventral receptacle; testis with 12 outer and 10 inner coils; penis and spermatheca as in Fig. 5; United States, Mexico and Guate­mala ------------------------------------------------------------------------------------------D. nigrohydei 330 coils in ventral receptade; 25 outer and 20 inner coils in the testis; Mexico --------------------------------------------------------------------------------------D. hydeoides GENETICS Genetic data was obtained by the use of mass matings of ten or twenty pairs of flies per vial. All possible combinations of intraspecific crosses were carried out among the strains of eohydei, nigrohydei and bifurca. Forty-three of the possible 420 intraspecific hydei crosses were performed, each strain being tested against at least two other strains. In all instances, intraspecific crosses produced many highly fertile offspring. TABLE 3 Interspecific Crosses in the Hydei Subgroup Species nigrohydei hydeoides hydei neohydei eohydei bi/urea nigrohydei x F3 • 180 s 100 s 1005 605 hydeoides F • x s• s• 3 hydei 100 s s· x 180: ZO F2 z40 s s• neohydei 140 s 140: few F2 x 100 s 60 s eohydei 80: few F2 180 s 100 s x 405 bi/urea 405 s• s· 60 s 405 x The figures refer to the total number of pairs tested in mass matings. S indicates no offspring produced. •Data taken from Wharton (1944). Table 3 summarizes the results of the interspecific matings attempted including those of Wharton ( 1944) who used nigrohydei, hydeoides, hydei and bifurca. The figures in the table refer to the total number of parental pairs used in our tests; "S" indicates no offspring; figures following a semicolon refer to the total number of progeny produced. These data demonstrate the presence of some residual crossability, each of the six species tested (except the genetically isolated bifurca) producing some hybrids in at least one interspecific cross. The difficulty in obtaining progeny even under the artificial conditions of these tests (as many · as twenty pairs of parental flies per vial with no choice of mates) indicates that there is little or no naturally occurring interbreeding between sympatric forms. The fertile hybrids resulting from the nigrohydei-hydeoides crosses probably reflect a close relationship between these two species. The University of Texas Publication SALIVARY GLAND CHROMOSOMES The gene sequence of the cytologically monomorphic species, Drosophila repleta, has been chosen as the standard sequence of the repleta group (Wasser­man, 1954), the sequence(s) found in every other species being ultimately com­pared to this standard. However, Wasserman (1960) using circumstantial evi­dence, proposed that the most probable primitive sequence is one that differs from this standard by six inversions, Xa, Xb, Xe, 2a, 2b, and 3b. This sequence is the basic gene arrangement of the hydei subgroup, being the standard bi/urea and nigrohydei. Only one inversion, 2z, has been fixed as an interspecific difference during the divergence of the five species studied. The newer sequence, 2a, 2b, 2z, is the standard second chromosome of eohydei, neohydei and hydei and is shown in Figure 10, a salivary chromosome map constructed by rearranging a photostat of the second chromosome of D. repleta (after Wharton, 1942). To arrive at the more primitive 2a, 2b chromosome it is necessary to reinvert the 2z section of Figure 10. Each of the five species studied is polymorphic for its own unique arrange­ments, there being a total of ten heterozygous inversions, two in the X and eight in the second. The breakage points of the heterozygous second chromosome inver­sions to be discussed below are shown in Figure 10. No rearrangements have been found in the third, fourth or fifth chromosomes. D. bi/urea is not known to have any structural rearrangements in the euchro­matin and therefore has the primitive sequence. However, Wharton is reported to have found a pericentric inversion in the heterochromatin of the X chromo­some, resulting in an alteration of the metaphase configuration (Ward, 1949, and Figure 11). This is the only pericentric inversion recovered in the repleta group FIG. 10. The salivary gland second chromosome of the hydei subgroup having 2a,b,z. Wasserman: The Hydei Subgroup and is apparently able to survive due to the fact that only heterochromatin is involved. D. nigrohydei, the other species which has the primitive gene sequence, is heterozygous for three second chromosome inversions, 2x, 2y and 2q3 (Figure \/ ~ii~ D. bifurco D. bifurco D. btfurco \/ '\/. ,11., --~~ ~~-­ ~ii~ ·~iiy~5 ,~r' D. novemarisfofo hypofheticoI D. nigrohydei Xa Xb Xe 2a 2b 3b \/ \/ ,11~ ~··~--~··~ '~1·~ mjl \ ~,,, D. eohydei D. neohydef\ D. hydeoides '\ \ ' ' ' \ / --------\ '/ ~t~ ~i~ D. hydei D. hydei Fm. 11. Metaphase karyotypes of the species in the hydei subgroup. All inversions, except Xm found in D. neohydei, occur in the second chromosome. Arrows indicate phylogenetic evolution. The University of Texas Publication 10). Whereas the 2x inversion is widely distributed, the 2y seems to be limited to the northern part of the distribution of the species while the 2q3 is found in the southern part (Table 1). The latter two are independent inversions, see Figure 10, and both may be present together in the unsampled geographically inter­mediate areas. The remaining three species have the 2z inversion fixed with the result that their standard second chromosome is the 2a, 2b, 2z, shown in Figure 10. D. eohydei is heterozygous for two second chromosome inversions, 2g3 and 2h3 (Table 1 and Figure 10). Only one inversion has been found in D. hydei, the 2a2, Figure 10, which seems to be present in the heterozygous condition wherever this world-wide species has been cytologically analyzed, the United States, Mexico and Australia (Table 1), South Africa and Hawaii (Warters, 1944), and most probably Brazil (Wharton, 1944). The one strain of D. neohydei is polymorphic for two inversions: besides a rearrangement in the second chromosome, 2£3, Figure 10, there is present an inversion in the X chromosome, Xm, whose breakage points on the Xa, Xb, Xe, standard are as follows: Xm/+ METAPHASE KARYOTYPES Figure 11 shows the metaphases of seven of the species. The "hypothetical" is the basic karyotype of the repleta species group. Although the Y chromosome was almost undoubtedly an acrocentric, representing it in the figure as a short rod rather than a longer element may not be correct. There have been a number of changes in the metaphase chromosomes through the addition (and/or deletion) of heterochromatin to the dot elements ( nigrohydei and hydeoides) and to the sex chromosomes ( bifurca, neohydei and hydei) with the result that each of the seven species, with the possible exception of the hydeoides-nigrohydei species pair, can be readily identified by its karyotype. The Y chromosome of novemaris­ tata is tentatively shown as a large rod, the sex of the larvae figured by Dobzhan­ sky and Pavan ( 1943) not being indicated. Intraspecific variation was found in two of the species. One of the strains of hydei from Colombia, H 194.23, differs slightly from the other strains checked in that the X chromosome has terminal constrictions on both arms and the Y appears to have lost most of the heterochromatin in the short arm. Variation in the sex chromosomes was found in bi/urea, where Wharton ( 1942) reported a strain from Wild Rose Canyon, Texas, with a rod-shaped Y being the same length as the X; strains from Mexico (Ward 1949) and Arizona (this study) have a small metacentric Y; and Ward (1949) reported that Whar­ ton found a strain of bifurca with a pericentric inversion in the heterochromatin of the X. These karyotypes, shown in Figure 11, are taken in part from Figures 42 Land M of Patterson and Stone (1952). Wasserman: The Hydei Subgroup D1scuss10N A study of the strains in The University of Texas laboratory, Table 1, has yielded morphological, genetic and cytological information concerning five of the eight species in the hydei subgroup of the repleta species group. These data combined with published accounts of D. hydeoides (Wharton, 1944) and D. novemaristata (Dobzhansky and Pavan, 1943), two of the three species not examined, present a tentative and by no means complete, picture of the phylo­genetic relationships among these species. Unfortunately since D. pachea has never been raised in the laboratory and is therefore known only from adult morphological characters, its relationship to the other species is not known. The eight species represent a closely-knit morphological unit which can be readily distinguished from the rest of the repleta species group by the presence of extensive coiling in the testes and the ventral receptacle. Crosses performed in the laboratory where there was no choice of mates show that gene exchange between some species is still possible (Table 3). Allopatric distributions or other isolating mechanisms in sympatric forms probably keep this gene exchange down to a bare minimum. For example, nigrohydei shows "considerable sexual isolation" to the syrnpatric hydeoides (Wharton, 1944). The five species whose salivary gland chromosomes were analyzed show few structural changes in the euchromatin during the history of the subgroup. Only one inversion has become homozygous during the evolution of these five forms from the ancestral species of the repleta group. Indeed two of the species are believed to have the primitive gene sequence as their standard. Of course it cannot be categorically stated that minor rearrangements have not been over­looked in the analysis, but the overall picture is one of stability in the gene sequences. However, inversional polymorphism does exist and is present in each of the five species, euchromatic rearrangements occurring in four of the forms, whereas in bifurca a pericentric inversion in the heterochromatin of the X was found by Wharton. Some of these inversions may be present throughout the whole range of the species, although not necessarily in every population. Thus the 2x inversion in nigrohydei is found from Arizona to Guatemala and the 2a2 sequence of hydei has been reported from the United States, Mexico, Australia (Table 1), South Africa and Hawaii (Warters, 1944) and probably Brazil (Wharton, 1944) . The major cytological evolution has taken place in the metaphase karyotypes through the addition and/or deletion of heterochromatin to the sex chromosomes and the dot elements (Figure 11). Phylogenetic Relationships The best evidence for the determination of phylogenetic affinities is to be found in the inversional composition of the species, since it is assumed that each inver­sion is a unique event which arose only once in the history of the organism. Unfortunately for this purpose, the hydei subgroup is remarkable in that only one inversion, 2z, has been fixed during the evolution of the five species examined. This sequence divides the species subgroup into two complexes: the cytologically primitive (2z lacking) bifurca complex, containing bifurca and nigrohydei, and The University of Texas Publication the cytologically more advanced (2z present) hydei complex, with eohydei, neohydei and hydei. The two species in the·bi/urea complex, in addition to being cytologically primitive, have bell-shaped spermathecae (Figures 5 and 6), the typica] form found generally throughout the repleta group, whereas those of the hydei com­plex are quite distinctive and specialized, Figures 7, 8, and 9. As for the species whose salivary gland chromosomes have not been analyzed, novemaristata has the primitive bell-shaped spermathecae (Dobzhansky and Pavan, 1943) , while the karyotype of hydeoides differs only slightly from that of nigr_ohydei (Figure 11 ) with which itproduces fertile off spring (Wharton, 1944). On the basis of these data novemaristata and hydeoides are assigned to the bi/urea complex at least until more critical information is obtained. The three species in the hydei complex show a step-wise addition of hetero­chromatin to the X chromosome, from eohydei to neohydei to hydei. Although the addition (or deletion) of heterochromatin is not always a reliable character for demonstrating affinities, in this case it is confirmed by morphological and genetic data in so far as hydei is shown to be closer to neohydei than it is to eohydei; the spermathecae of hydei and neohydei are very similar (Figures 8 and 9) and quite different from those of eohydei (Figure 7); and although eohydei is genetically completely isolated from the other two, crosses between neohydei and hydei will produce some off spring (Table 3). · Evidence from the changes in the Y chromosome, although tending to agree with the eohydei to neohydei to hydei proposal, points out the danger of using heterochromatin by itself. If we assume that the primitive Y was a short rod, there is again a step-wise addition of heterochromatin in the proposed direction of evolution. However, one of the hydei strains from Colombia, H 194.23, appar­ently a derived form, shows a reduction in the length of the Y, and therefore a reversion to more "primitive" condition (Figure 11). Also it should be pointed out again that the primitive Y may have been a long rod rather than the short rod depicted in Figure 11. Figure 11 shows the overall phylogenetic relationships of seven of the eight species in the subgroup. The hypothetical karyotype contains the primitive gene sequences. Arrows indicate probable directions of evolution, newer inversions being shown where found either in the homozygous condition, 2z only, or in the heterozygous condition. SUMMARY 1. Two new repleta group species of the genus Drosophila are described and assigned to the hydei subgroup, increasing the number of forms in the subgroup to eight. 2. Strains of five of these species were studied morphologically, genetically and cytologically. These data, including information obtained from the literature concerning the other species, indicate that the subgroup is a single phylogenetic unit, morphologically distinct from the other elements of the species group, and within which some gene exchange is possible, at least under laboratory conditions. 3. Salivary gland chromosome analysis demonstrated a remarkable stability of the euchromatic sequence during the evolution of the subgroup, two of the Wasserman: The Hydei Subgroup species, D. bifurca and D. nigrohydei still maintaining the proposed ancestral gene sequence as their standard; whereas the standard of three other species, D. eohydei, D. neohydei and D. hydei, differs from this primitive by only a single inversion, 2z. 4. Four of the five species are polymorphic for euchromatic inversions, the fifth, bifurca, being heterozygous for a pericentric inversion in the X chromo­some. This is the only pericentric inversion ever recovered in the repleta group and is probably able to ·survive due to the fact that only heterochromatin is involved. 5. Metaphase karyotypes of seven of the eight species are known, and show intraspecific and interspecific differences due to the additions and/ or deletions of heterochromatin to the dot elements and the sex chromosomes. 6 .. Phylogenetic affinities among the five species studied are shown, mor­phology, genetics and metaphase karyotypes being used to supplement the inver­sional relationships. D. hydeoides and D. novemaristata, although not examined by the author, are tentatively classified on the information present in the liter­ature. The eighth species, D. pachea, is known only from the original morpho­logical description of the adult. BIBLIOGRAPHY Dobzhansky, Th., and C. Pavan. 1943. Studies·on Brazilian species of Drosophila. Bol. Fae. Fil. Cien, Letras, S. Paulo, No. 36: 7-72. Patterson, J. T., and W. S. Stone. 1952. Evolution in the genus Drosophila. MacMillan Co., New York. · Patterson, J. T., and M. R. Wheeler. 1942. Descriptions of new species of the subgenera Hirto­drosophila and Drosophila. Univ. Texas Pub. 4213: 69-109. Ward, Calvin L. 1949. Karyotype variation in Drosophila. Univ. Texas Pub. 4920: 70-79. Warters, Mary. 1944. Chromosomal aberrations in wild populations of Drosophila. Univ. Texas Pub. 4445: 129-174. Wasserman, Marvin. 1954. Cytological studies of the repleta group. Univ. Texas Pub. 5422: 130-152. ----. 1960. Cytological and Phylogenetic relationships in the repleta group of the genus Drosophila. Proc. Nat. Acad. Sci. 46, No. 6: 842-859. Wharton, Linda T. 1942. Analysis of the repleta group of Drosophila. Univ. Texas Pub. 4228: 23-52. ----. 1944. Interspecific hybridization in the repleta group. Univ. Texas Pub. 4445: 175-193. Wheeler, Marshall R. 1949. Taxonomic studies on the Drosophilidae. Univ. Texas Pub. 4920: 157-195. V. Cytological Studies of the Repleta Group of the Genus Drosophila: V. The Mulleri Subgroup 1 MARVIN WASSERMAN INTRODUCTION The mulleri subgroup, the largest in the repleta group, consists of twenty-seven Drosophila species (Patterson and Stone, 1952; Wheeler, 1954 and 1957; Wasser­man and Wilson, 195 7; and this report). A salivary gland chromosome analysis has been made on twenty-two forms, including eight new species. This, together with much of the morphological and genetic information accumulated by various authors, will be included in this paper. The mulleri subgroup, as presented here, contains all of the species which do not fall within the bounds of the other subgroups, and which have arisen from a single cytological ancestor. These species form several evolutionary units: (a) the meridiana complex, which consists of three closely related species, two of which are new; (b) the mulleri complex, defined cytologically by Wasserman (1954), which contains twelve species, four of which are new; and ( c) the miscellaneous forms, which do not fit into the somewhat narrowly defined complexes, or are not sufficiently known to be more accurately placed. Table 1 lists all of the new strains examined together with several strains previously worked on. Other localities are to be found in Wasserman (1954) and Wasserman and Wilson ( 195 7). The inversional composition of each stock as compared to D. repleta, the arbitrary standard for the repleta group, is shown in Table 1. MORPHOLOGY The mulleri subgroup can be characterized morphologically by having a rela­tively low number of coils in the testis, ranging from four to nine, and a short ventral receptacle which may vary from twelve to thirty-five coils. Several of the species, however, fall outside these ranges: D. spenceri and D. subviridis have thirteen and eighteen coils respectively in the testis; and D. pegasa, D. subviridis and D. aldrichi have fifty-five, seventy and six coils in their ventral receptacles, respectively. The exceptionally low number found in aldrichi can be accounted for by the looseness of the coiling of the organ in this species. Tables 2 and 3 list in addition to the number of coils in the ventral receptacle and the testis, the wing vein indices of the twenty-seven species in this subgroup; much of the information is derived from the original descriptions. Figures 1 to 20 show the spermathecae of the twenty species examined in this study. The penis, a very useful character for species identification, and the male 1 Present address: Zoology Department, University of Melbourne, Melbourne, Australia. The University of Texas Publication TABLE 1 List of Strains and Localities 4 5 Species Stock Locality Xa,b,c 2a,b 3b meridionalis 2507.21 Angra dos Reis, R. de Jan., Brazil + + + + + promericliana H318.4 Palmira, Colombia + r s + + + meridiana H 381.8 Aca tlan, Puebla, M exico + + + + + rioensis* H381.1 3 Acatlan, Puebla, M exico + w4/+ + + + H381.1 6 Acatlan, Puebla, Mexico + + + + + H381.28 Acatlan, Puebla, Mexico + w4/+ + + + H381.29 Acatlan, Puebla, M exico + + + + + H381.30 Acatlan, Puebla, Mexico + + + + + 305.5b Georgetown, Texas + + + + + anceps 1808.14b Oaxaca, M exico f w e, j a a,b stalkeri 2213.1 St. Petersburg, Florida + l,m,n + + + hamatofila 1981.1 Fort Davis, T exas + x<,y• + + c A 6.3 Patagonia, Arizona + x4,y4,v4,z•l /+,as!+ + + c A 8.3 Aravaipa Valley, Arizona + x4 ,y4,v4 + + c buzzatii * H345.1 3 Cochabamba, Bolivia + xs,k,w3,i/+ + + g H345.3 H347.9 Cochabamba, Bolivia Cordoba, Argentina + + x3,k,w3, j/ + x3,k,w3,j/+ + + + + g g H347.10 2093 .10 Sn. Rius, Argentina Ain Anub, Lebanon + + x3,k,w3,y3/+,jz3/-J-+ + x3,k,w3 + + + g g pegasa 25 19.14 Pachuca, M exico + x3,k/+,bs;+,cs;+ + + + nigricruriat 2395 .1 Lima, Peru k t2,u 2,v2,w2,x2 + + f H347.3 Azapa, Chile k t2,u2,v2,w 2,x2 + + f H347.4 Camarones, Chile k t2,u 2,v2,w 2,x2 + + f eremophila H381.10 Acatlan, Puebla, M exico + d5 ,es ,£5,gs + + + H381.1 8 Aca tlan, Puebla, M exico + d5,e5,fS ,gs + + + H381.22 Acatlan, Puebla, M exico + d5,e5,fS ,gs + + + H381.22a Acatlan, Puebla, M exico + ds,es,fs,gs + + + H381.23 H381.42 Acatlan, Puebla, M exico Acatlan, Puebla, Mexico + + d5 ,es ,fs ,gs d5,es ,fs ,gs + + + + + + mojavensis Chocolate Mts., California e c,f,g,h,q,r,s a,d + + arizonensis • 2156.4 A 2. 7a A 5.5a A 5.5b A 12.11 Fairbanks, Ari zona Patagonia, Arizona Cape Region, Baja + + + + c,f,g,h,i c,f,g,h,i c,f,g,h,i c,f,g,h,i + + + + + + + + + + + + California + c,f,g,h,i + + + mulleri* aldrichi* H 359.7 251 7.2 2521.ct H 52.12 Roy Farm, Austin, T exas Kingston, Jamaica Humbolt, Nebraska Veracruz, M exico Puerto Triunfo, + + + + c,f,g c,f,g c,f,g c,f,g a,c a,c a,c a,c + -!­+ + + + + + El Salvador Austin, Texas + + c,f,g c,f,g a,c a,c + + + + Wasserman: The Mulleri Subgroup wheeleri 1980.1 Arcadia, California + c,f,g a,c + + longicornis 2513.1 Austin, Texas + c,f,g,t3,u3/+ a,c + + martensis H188.12 Santa Marta, Colombia c,d 2 ,e2,f2,g2/ + a,k + + H208.1 Barquesimeto, Venezuela c,d2,e2,f2,g2/+ a,k + + pachuca 2519.1 Pachuca, Mexico + c,g,h4 ,i4,j4 I+,k4I+ a,c + + 2519.20 Pachuca, Mexico + c,g,h4,i4,j4/+ a,c + + 2519.21 Pachuca, M exico + c,g,h4,i4,j4/+,k4;+ a,c + + propachuca 2519.18 Pachuca, Mexico + c,g,h4/+ a,c + + desertorum 2519.22 Pachuca, Mexico + g,e4 ,f4,g4 I+ a,c,r/+ + + ritae 2360.2 Whitewater, New Mexico + g,p a,c + + A 6.4 Patagonia, Arizona + g,p a,c + + tira 252ld.2 Mexico City, M exico + g,p/+,u4/+ a,c + + •Additional strains reported in Wasserman (1954). t Additional strains reported in Wasserman and Wilson (1957). TABLE 2 Morphological Characters of the mulleri subgroup Species c. I. Wing characters 4vn 5x 4c Heavies* Coils in ven. rec. Coils of testis Outer Inner meridiana 2.6 1.9 1.7 1.1 .4 17 3 2 meridionalis 2.8 1.8 1.4 .93 .4 15 2Yz 2 promeridiana 2.7 1.7 1.4 .97 .33 20 2 2 anceps 3.0 1.7 1.2 .87 .5 28 2 2 stalkeri 3.0 1.8 1.4 .33 16 1Yz 2Yz hamatofila 3. 1 1.6 1.3 .85 .25 15 5 4 buzzatii 2.7 1.6 1.1 .90 .33 14 3 3 pegasa 3.3 2.0 1.3 .88 .25 55 3Yz 3 hexastigma 3.6 1.53 1.0 .72 .43 28 racemova 3.0 1.8 1.32 .88 .45 28 4 4 subviridis 3.8 1.5 1.1 .88 .5 70 9 9 mainlandi 3.4 1.6 1.1 .77 .33 15 3 3 spenceri 3.0 1.7 1.4 1.2 .4 35 (13) nigricruria 3.3 1.9 1.2 .83 .5 21 2Yz 2Yz eremophila 2.5 1.7 1.2 1.0 .3 30 3 2 • Proportion of third costal section with heavy bristles. TABLE 3 Morphological characters of the mulleri complex Species c. I. Wing characters 4vn 5x 4c Heavies* Coils in ven. rec. Coils of testis Outer Inner mojavensis 2.9 1.8 1.3 .94 .33 15 2 2Yz anzonens1s 2.4 1.7 1.4 1.0 .33 17 3 3 mulleri 2.8 2.0 1.3 1.0 .25 20 3 2 aldrichi 2.6 1.7 1.3 1.0 .3 loose 2% 2 wheeleri 2.6 .33 2% 2 longicornis 3.2 1.9 1.3 .90 .25 15 3 2 martensis 2.8 1.9 1.4 2.0 .25 15 2Yz 2 pachuca 3.0 1.6 1.0 .84 .28 15 3 2Yz propachuca 3.2 1.5 1.0 .75 .29 12 2Yz 2 ritae 3.1 1.8 1.2 .87 .39 15 2 1Yz tira 3.3 1.6 1.1 .79 .29 18 3 2 desertorum 3.5 1.6 1.0 .75 .33 • Proportion of third costal section with heavy bristles. The University of Texas Publication 10 Frcs. 1-ZO. Spermathecae of species in the mulleri subgroup. Fig. 1. D. meridionalis. Fig. Z. D. promeridiana. Fig. 3. D. meridiana. Fig. 4. D. anceps. Fig. 5. D. stalkeri. Fig. 6. D. hamatofila. Fig. 7. D. buzzatii. Fig. 8. D. pegasa. Fig. 9. D. nigricruria. Fig. 10. D. eremophila. Fig. 11. D. mojavensis. Fig. JZ. D. arizonensis. Fig. 13. D. mulleri. Fig. 14. D. aldrichi. Fig. 15. D. longicornis. Fig. 16. D. martensis. Fig. 17. D. pachuca. Fig. 18. D. propachuca. Fig. 19. D. ritae. Fig. ZO. D. tira. external genitalia of each of these twenty species are shown diagrammatically in Figures 21 to 60. Eight new species have been found in recent collections. Only four pinned specimens were available for the new species, Drosophila desertorum, one being considered the holotype while the other three are paratypes. For each of the other seven species, the holotype and nine paratypes are deposited in The University of Texas collection along with slides of male genitalia, penes, spermathecae, wings, legs, and mouthparts, which can he used as reference material in future studies. The first two species to be described, Drosophila meridionalis and Drosophila promeridiana are morphologically identical to Drosophila meridiana rioensis Patterson, 1943. However, in view of the sterility encountered in the matings among these three forms (see below), it seems advisable to consider them as distinct, but closely related, species forming the meridiana complex. Drosophila meridionalis, new species Similar to Drosophila meridiana rioensis Patterson, 1943. Camera lucida draw­ings of a spermatheca are shown in Figure 1, the penis, Figure 21, male external genitalia, Figure 41, and the metaphase karyotype, Figure 61. Internal characters and wing vein indices, as in Table 2. Wasserman: The Mulleri Subgroup 30 FIGs. 21-30. Side view of penes, and dorsal view of head of penes of ten species in the mulleri subgroup. Fig. 21. D. meridionalis. Fig. 2Z. D. promeridiana. Fig. 23. D. meridiana. Fig. 24. D. anceps. Fig. 25. D. stalkeri. Fig. 26. D. hamatofila. Fig. 27. D. buzzatii. Fig. 28. D. pegasa. Fig. 29. D. nigricruria. Fig. 30. D. eremophila. The University of Texas Publication FIGs. 31-40. Side view of penes, and dorsal view of the head of penes of ten species in the mulleri complex. Fig. 31. D. mojavensis. Fig. 32. D. arizonensis. Fig. 33. D. mulleri. Fig. 34. D. aldrichi. Fig. 35. D. longicornis. Fig. 36. D. martensis. Fig. 37. D. pachuca. Fig. 38. D. pro­pachuca. Fig. 39. D. ritae. Fig. 40. D. tira. Wasserman: The Mulleri Subgroup 43 Frcs. 41-50. Male external genitalia of ten species in the mulleri subgroup. Fig. 41. D. meri­dionalis. Fig. 42. D. promeridiana. Fig. 43. D. meridiana. Fig. 44. D. anceps. Fig. 45. D. ~talkeri. Fig. 46. D. hamatofila. Fig. 47. D. buzzatii. Fig. 48. D. pegasa. Fig. 49. D. nigricruria. Fig. 50. D. eremophila. The University of Texas Publication FrGs. 51-60. Male external genitalia of ten species in the mulleri complex. Fig. 51. D. moja­vensis. Fig. 52. D. arizonensis. Fig. 53. D. mulleri. Fig. 54. D. aldrichi. Fig. 55. D. longicornis. Fig. 56. D. martensis. Fig. 57. D. pachuca. Fig. 58. D. propachuca. Fig. 59. D. ritae. Fig. 60. D. tira. \ · Wasserman: The Mulleri Subgroup 61. D. m. rioensis D. promeridiana D. meridionalis --.. -62. - r D. eremophila 63. D. anceps 64. D. m. meridiana D. buzzatii D. longicornis D. wheeleri D. aldrichi D. mulleri D. arizonensis D. mojavensis ~ ~ 65. ' , D. martensis ?i~~D. homotofilo 66. D. pachuco 67. stalked ritae ~~ hexasfigma 68. ~··~ D. nigricruria D. fira II D. racemova FIGs. 61-68. Metaphase karyotypes of the species in the mulleri subgroup. The University of Texas Publication Distribution: The strain was sent to this laboratory by Dr. C. Pavan who collected it in Angra dos Reis, Rio de Janeiro, Brazil. This species had been previously designated as Drosophi.la species D by Wasserman ( 1960). Drosophila promeridiana, new species Similar to Drosophi.la meridiana rioensis Patterson, 1943. Camera lucida draw­ings of a spermatheca are shown in Figure 2, the penis, Figure 22, the male external genitalia, Figure 42, and the metaphase karyotype, Figure 61. Internal characters and wing vein indices, as in Table 2. Distribution: Only known locality, Palmira, Colombia, collected by Dr. M. R. Wheeler. This species has been previously designated as Drosophila species C by Wasserman ( 1960). The following two new species are not assigned to any complex and are placed among the miscellaneous species in the mulleri subgroup: Drosophila pegasa, new species Arista with three branches above and two below, excluding the terminal fork. Front bronze, orbits and ocellar triangle pollinose gray; posterior orbital and anterior vertical each with a brown spot at base; one oral bristle. Mesonotum pollinose gray with brown spots. Acrostichal hairs in six rows. Female: abdomen elongated, yellow; each ·tergite with an interrupted wide dark brown band which fades at the angle of the tergite; lateral margins of each tergite with a dark brown band. Male: abdomen yellow; each tergite with an inter­rupted wide diffuse gray band which fades at the angle of the tergite. Camera lucida drawings of a spermatheca are shown in Figure 8, the penis, Figure 28, the male genitalia, Figure 48. Internal characters and wing vein indices are listed in Table 2. Distribution: Known only from Pachuca, Puebla, Mexico, collected by Dr. A. C. Faberge. This species is named after Pegasus, the fabled winged horse of the Muses which was the unwilling mount of Bellerophon, one of the Greek heroes. Females have been observed acting as unwilling mounts of the males for periods ranging up to fourteen hours. This species had been designed as Drosophila species E by Wasserman ( 1960). Drosophila eremophila, new species Arista with three branches above and two branches below excluding the terminal fork. Front chocolate brown, orbits and ocellar triangle pollinose silvery gray; posterior orbital and anterior vertical each with a chocolate brown spot at base; one oral bristle. Mesonotum yellowish gray pollinose, with chocolate brown spots ·fusing to form patches. Legs yellowish gray; no distinct bands, base of the hind tibia slightly darker. Abdomen light yellow; each tergite with a narrowly interrupted dark brown band which expands at the angle of the tergite to form solid lateral areas. Wasserman: The Mulleri Subgroup Camera lucida drawings of a spermatheca are shown in Figure 10, the penis, Figure 30, the male external genitalia, Figure 50, and the metaphase chromo­somes, Figure 62. Internal traits and wing vein indices are shown in Table 2. Distribution: Known only from Acatlan, Puebla, Mexico, collected by the author. This species had previously been called Drosophila species F by Wasser­man (1960). The remaining four new species are members of the mulleri complex which also includes D. mulleri, D. aldrichi, D. wheeleri, D. mojavensis, D. arizonensis, D. longicornis, D. ritae and D. martensis. Drosophila pachuca, new species Arista with three branches abov~ and two below excluding the terminal fork. Front chocolate brown, orbits and ocellar triangle pollinose silvery gray; posterior orbital and anterior vertical each with a chocolate brown spot at base; one oral bristle. Mesonotum silvery gray pollinose with chocolate brown spots; no fusion of spots. Abdomen light yellow, each tergite with a broadly interrupted narrow posterior band, the band of the second tergite stopping at the angle of the tergite. In males, the bands of the third through the sixth tergites stop at the angle of the tergites and have forward extensions to the anterior margins; lateral margins of all tergites with a narrow band separated from the posterior band by a narrow silvery yellow area. In females, the third through sixth tergites with the posterior and lateral bands connected at the anterior margin, producing a. lateral area which is solid dark brown e]'cept for small irregularly shaped posterior dark silvery yellow patches. Camera lucida drawings of a spermatheca, Figure 17, the penis, Figure 37, the male external genitalia, Figure 57, and the metaphase chromosomes, Figure 66. Internal characters and wing vein indices as in Table 3. Distribution: Known only from Pachuca, Puebla, Mexico, collected by Dr. A. C. Faberge. This species h&d been previously designated as Drosophila species G by Wasserman ( 1960). Drosophila propachuca, new species Externally very similar to Drosophila pachuca from which it differs in having the abdominal tergites with posterior bands which are wider than those of D. pachuca. Other differences are found in the male reproductive organs, the penis, Figure 38, and the male external genitalia, Figure 58, being distinct. Drawings of spermatheca, Figure 18, .and .the metaphase karyotype, Figure 66, are given. Internal characters and wing vein indices as in Table 3. Distribution: Known only from Pachuca, Puebla, Mexico, collected by Dr. A. C. Faberge. This species had been previously designated as Drosophila species H by Wasserman ( 1960). Drosophila desertorum, new species Externally similar to D. propachuca. Carina broad, sulcate; cheeks broad, dirty tan; palpi tan. Legs pale, tibiae of legs 2 and 3 with narrow dark bands basally; bristles of first coxa and femur quite prominent. Brown mesonotal spots The University of Texas Publication small, numerous. Wing vein indices as in Table 3. Salivary gland chromosomes very distinct, described below. Distribution: Known only from Pachuca, Puebla, Mexico, collected by Dr. A. C. Faberge. This species had been previously designated as Drosophila species I by Wasserman ( 1960) . Drosophila tira, new species Similar to Drosophila ritae Patterson and Wheeler, 1942. Mesonotum dark tan with brown spots, these irregularly fused into a pair of longitudinal rows in acrostichal region. Scutellum darker. Pleura tan, irregularly darker above, appearing to have two blotched stripes when seen from certain angles. Frons dull brown, orbits grayish. Face grayish tan; carina widely flaring below. Cheeks rather broad; palpi pale; second oral Y3 length first. Legs pale; a tendency for darker bands basally on tibiae; last tarsal joints darker. Male first tarsi with numerous curved short hairs on inner side. Female abdomen with large dull brown bands separated by a narrow, pale median stripe; male with paler lateral areas on tergites 1-4. Internal characters and wing indices are shown in Table 3. The spermatheca is shown in Figure 20, the male genitalia in Figures 40 and 60, and the metaphase plate chromosomes in Figure 68. · Distribution: Known only from a stock originally collected about 25 miles east of Mexico City, Mexico by Don Hunsaker. This species had previously been designated as Drosophil~ species N by Wasserman ( 1960) . GENETICS All crosses were made using mass matings of ten or twenty pairs per vial. Each interstrain intraspecific cross produced fertile offspring. Table 4 gives the results obtained from 203 interspecific crosses involving 19 of the 27 mulleri subgroup species (Patterson and Stone, 1952; and new data). Results of crosses involving species not included in Table 4 are presented by Wharton (1944) and Patterson and Alexander ( 1952) . Most of the species are completely genetically isolated from all other forms and all attempts to obtain progeny have been unsuccessful. Some interspecific crosses, however, do produce offspring, but whereas sterile progeny may sometimes be obtained from sympatric forms, only allopatric species yield fertile hybrids (Table 4). Within the meridiana complex F 1 progeny were obtained in most of the inter­specific crosses, but no F2 larvae were ever observed. Back crosses, using small mass matings, showed many of the 342 hybrid females so tested were fertile; whereas none of the 334 hybrid males back crossed produced progeny. Dissection of thirty-eight 9-to 13-week-old males showed testes which were devoid of sperm, highly pigmented and with the exception of a single male, were small and under­developed. Some of the interspecific crosses involving members of the mulleri complex produce offspring. No attempts were made to cross D. desertorum. No offspring have been obtained in any cross where individuals of D. martensis, D. tira, D. pachuca, D. propachuca (Table 4), or D. ritae (Wharton, 1944) are involved. Among the six other species, cross compatibility in mass cultures where no choice TABLE 4 Interspecific crosses in the mulleri subgroup ~/c:J Mu Al Mo 1\J· L Ma T R p Pp Pe E B St II An Mr Md Pm Mu x s (j> ,s 6' f (j> ,s 6' s 6' s (j> s s s s s s larvae s s (j> ,s 6' s s s s Al s x s (j> s (j> s s s s Mo s s x f (j> ,£ 6' s s s s s s s s s s s Ar s s f (j> ,s 6' x s s s s s larvae s s s s L s s s s x s s s s s s s s s Ma s s s s s x T s s x s s s s s s s R s x s p s s s s s s x s s s s s s s s s Pp s s x s s s s s Pe s s s s s x s s s s s s s E s s s s s x s s s s s B s s s s s s s s s x s s s s s s St s s s s s s s s s s x s s s s H s s s s s s s s s s x s s s An s s s s s s x Mr s s s s s s s s s s s s x f (j> ,s 6' f (j> ,s 6' Md s s s s s s s s f (j> ,s 6' x f (j> ,s 6' Pm s s s s s f (j> ,s 6' f (j> ,s 6' x S: no offspring; s: sterile; f: fertile: Mu: mulleri ; Al: aldrichi; Mo : m ojavensis; Ar: ariw nensis:, L: longicornis; Ma: martensis; T: tira; R: ritae; P: pachuca; Pp: propachuca; Pe : pegasa; E: erem ophila; B: buzzat.ii; St : stalkeri; H: hamatofila ; An: anceps; l'vlr : meridiana rioensis; l\1d: merdidionalis ; Pm : prom.eridiana , The University of Texas Publication of mates is available is closely correlated with phylogenetic affinities based on the inversions. The genetic relationships derived from the data presented in Table 4 can be pictured as follows: longicornis I wheeleri-aldrichi-mulleri-mo;avensis--arizonensis The position of wheeleri is determined from data taken from Patterson and Alexander (1952) where it was observed that wheeleri produces a fairly large number of offspring, only the females being fertile, in reciprocal crosses with aldrichi; the number and fertility of the progeny is reduced if mulleri is the other parent; this reduction continues through mo;avensis and arizonensis, the latter being the least compatible when crossed to wheeleri. This picture indicates the overall genetic relationships since it reflects the ability of genotypes from two species to cooperate in the formation of viable and fertile individuals. However, single genes, or at least relatively small segments of the genotype, may be the cause of incompatibility. For example, Crow (1942) was able to demonstrate the presence of an allele (or a few very closely linked alleles operating as a single locus) on some of the aldrichi X chromosomes which was a dominant lethal in mulleri-aldrichi hybrids. Drosophila hamatofila and Drosophila buzzatii will each yield viable zygotes in crosses with one or two members of the mulleri complex, Table 4. Although we do not include these species in the mulleri complex because of cytological differences, the production of some offspring undoubtedly demonstrates a basic genetic similarity between these species and the mulleri complex. Sexual preference, as exhibited by a refusal to mate with an alien individual or a degree of discrimination when given a choice of mates, is an important iso­lating mechanism. In the mulleri subgroup the only pertinent information con­cerns several species in the mulleri complex and is only fragmentary; a large systematic study involving many strains of each species has never been carried out. The data come from a single paper by Patterson (1947) in which mating preference was tested among strains of mulleri, aldrichi, arizonensis, moiavensis and ritae. D. ritae refused to mate with the others under any circumstance. In crosses involving the other species, there was a marked preference for homogamic matings except for one cross, where arizonensis males did not discriminate be­tween arizonensis and·ma;avensis females. This lack of discrimination on the part of the arizonensis males is of particular interest since fertile offspring are pro­duced in the interspecific cross (Table 4). These forms are considered allopatric species, and each is believed to be homozygous for its own inversions (Table 1). Mettler (1957), who was able to distinguish between the X, 2 and 3 chromosomes of these species due to these inversional differences, studied the outcome of cage populations which were initiated with various combinations of parental and hybrid flies. The results varied, depending for the most part upon the frequency of the different genotypes in the initial population, but heterosis was clearly exhibited by each of these three chromosomes in most if not all of the experi­ments. The cages where males and females of both species were used to start the Wasserman: The M ulleri Subgroup populations resulted in the elimination (X chromosome) or marked reduction in frequency (2 and 3 chromosomes) of the arizonensis genes despite the super­iority of the heterozygote. Mettler was able to. show chromosomal interaction, especially between the Y and the X, and the X and the 3, the final outcome depending upon the relative strengths of the heterosis versus the interaction which in turn depended upon the initial genotypes of the population. METAPHASE KARYOTYPES The metaphase karyotypes of twenty-three of the twenty-seven species are known, there being no information available for D. desertorum, D. mainlandi, D. spenceri and D. subviridis (Wharton, 1943; Patterson and Mainland, 1944; Clayton and Ward, 1954; Clayton and Wasserman, 1957). The known meta­phases fall into eight types (Figures 61 to 68). Closely related species often have identical metaphases: mulleri, aldrichi, wheeleri, and longicornis (Figure 64) ; arizonensis and mojavensis (Figure 64) ; meridiana rioensis, promeridiana and meridionalis (Figure 61). However similarity of karyotypes does not necessarily imply close relationships and distantly related forms may have arrived at similar karyotypes: martensis and hamatofila (Figure 65); stalkeri and ritae (Figure 67); tira and nigricruria (Figure 68). There has been one centric fusion during the evolution of the subgroup: the second and third chromosomes formed a meta­centric in the ancestor of m. rioensis, meridionalis and promeridiana (Figure 61). The subspecies, D. meridiana meridiana is reported to lack this fusion (Wharton, 1942), indicating that the subspecies meridiana is cytologically primitive and that the subspecies rioensis is closer to the two other species which are probably derived forms. SALIVARY GLAND CHROMOSOMES Wasserman ( 1960) presented evidence which indicates that the pnm1t1ve gene sequence of the mulleri subgroup, a sequence which differs from that of D. repleta, the standard chosen for the repleta group (Wasserman, 1954), by six inversions Xa, Xb, Xe, 2a, 2b, 3b, is probably the primitive sequence of the repleta group. Among the twenty-two mulleri subgroup species whose salivary gland chromosomes have been analyzed, a total of 72 inversions has been found X CHROMOSOME: The primitive gene order of the X chromosome of the mulleri subgroup differs from the repleta sequence by three inversions, Xa, Xb, and Xe, whose regions of rearrangement are C4g to D4b; F1a to F3a; and F3a to G2a, respectively (Wasserman, 1954.) Figure 73 shows the salivary map of this chromosome. These figures were made by rearranging photostatic copies of the· repleta map (Wharton, 1942) by simply cutting the map at, or close to, the· various breakage points and pasting the pieces together in the new inverted order. A total of four X chromosomal rearrangements have taken place in this subgroup,. each inversion being homozygous in the species in which it occurs: Xe in moia­vensis; Xf in anceps; Xj in martensis; and Xk in nigricruria. These inversions are shown in Figure 73. 2 CHROMOSOME: The standard second chromosome of the mulleri subgroup differs from that of repleta by two inversions, 2a and 2b, whose breakage points are shown in Figure 69 (a). Each species in the subgroup has these two sequences,. The University of Texas Publication 4 FIG. 69. Evolution of the second chromosome from (A), the primitive sequence, 2a, b, to (B), the standard sequence of D. hamatofila, Za, b, x\ y 4• and in addition may have others. A total of 53 rearrangements have been involved in the evolution of the twenty-two species examined. D. hamatofila is homozygous for two rearrangements, x4 and y4 (Figure 69), the resulting chromosome being the standard for this species (Figure 77). The two Arizona strains (Table 1) are homozygous for an additional inversion, v4, and the Patagonia strain is heterozygous for z4 and as. The standard chromosome and the breakage points of the three supplementary inversions are shown in Figure 77. D. eremophila is homozygous for d5, es, fs and gs. The evolution of the second chromosome is shown in Figure 70. FIG. 70. Evolution of the second chromosome, resulting in the standard sequence of D. eremo­phila, Za, b, d5, es, £5, gs. D. buzzatii is homozygous fork, w3 and x3 ; the latter two inversions, which were overlooked in a previous publication (Wasserman, 1954), have breakage points which are fairly close to each other, the x3 being included within w3, with the result that the standard chromosome superficially appears to lack both of these rearrangements. Figure 71 shows the evolution of this chromosome from the 2a, b primitive. Bolivian and Argentinian strains (Table 1) were found to be polymorphic for an inversion, 2j, previously reported from Lebanon (Wasser­man, 1954) and Australia (Mather, 1957). The strain from Argentina has three types of second chromosomes: the standard; y3, a new inversion; and j, z3, the new inversion, z3, always being associated with the j arrangement with which it shares one breakage point. Figure 78 shows the points of rearrangement of these three inversions on a salivary map of the standard chromosome. D. pegasa is homozygous for x3 and heterozygous fork, bs and cs. The k and x3 sequences are identical to those of buzzatii. Figure 71 shows the evolution of the second chromosome leading to pegasa and buzzatii. The salivary map of buzzatii (Figure 78) can be converted to the pegasa standard by reinverting first w3 which is absent in pegasa, and then k which is heterozygous. When this is done the Wasserman: The M ulleri Subgroup ,.---=a -... b XJ (B) A-AJ41a-AJJc1a-c6Jl1g-c6Jo1g-044Ja-.o4JFJa-H (E) A-AJaFla+-AJ~1a-c6~1g-c6ap1g-D4.fJa-F246a-G2ajD4a-F2~6a +-FJafr2d-H Fw. 71. Evolution of the second chromosome from the primitive, 2a, b, to (B), the standard of D. pegasa, 2a, b, x3; to (C), 2a, b, x3, k/+; to (D ), the standard of D. pegasa showing the breakage points of the heterozygous inversions, k, b5 and cs. (E), the standard chromosome of D. buzzatii, 2a, b, x3, k, w3 is derived from ( C). breakage points of the heterozygous inversions, k, b5 and c5 are seen on the standard pegasa chromosome. D. meridiana rioensis and D. meridionalis have as their standard the primitive sequence of the species group, 2a, b, the salivary map of which is shown in Figure 79. D. m . rioensiswas found to be polymorphic forw4 (Figure 79). D. promeridiana is homozygous for r3 • The salivary map of this species can be obtained by inverting the r3 sequence on the a, b chromosome (Figure 79). D. anceps is homozygous for w (Figure 79). D. stalkeri is homozygous for three inversions, 1, m, n, the evolution of which is shown in Wasserman (1954) . The salivary map of this species can be arrived at by inverting these sequences on the a, b chromosome (Figure 79). D. nigricruria is homozygous for t2, u2, v2, w2 and x2 The evolution of the • second chromosome was published in Wasserman and Wilson (1957). The salivary map is shown in Figure 80 where the breakage points of s2 and y2 , the heterozygous inversions found in Mexico, are indicated. D. hoeckeri is cytologically identical to D. nigricruria with which it produces fully fertile hybrids. The only morphological difference seems to be color pattern; the genitalia, male and female, are indistinguishable. Therefore D. hoeckeri is a synonym of D. nigricruria and should be considered, at most, a Chilean sub­species. The mulleri complex consists of twelve species each of which is homozygous for one or more of six rearrangements. Four of these, 2c, 2f, 2g, and 2h, are found in the second chromosome. Figure 72 shows the breakage points of these inver­ c f g (B) A+-A3af;1a-Bl tjc6a-c1ar.3a-Bl Jn1g-Dl~D3a-n1gp6a-D1~3a-E646a-E6~6a-H FrG. 72. A mulleri complex second chromosome which has the sequence 2a, b, c, f and g. sions. The c, f and g are independent rearrangements; the h inversion overlaps 2£ and occurs only in chromosomes which contain 2f. We have illustrated the The University of Texas Publication salivary maps of chromosome a, b, g (Figure 81 ), chromosome a, b, g, c (Figure 82) and chromosome a, b, g, c, f (Figure 83). The cytological composition of the twelve species is as follows: D. ritae is homozygous for g and p (Figure 81). D. tira has two types of chromosomes: g, p and g, u4, the p and u4 overlapping and therefore mutually exclusive (Figure 81). The standard chromosome, the g, has not been found. D. desertorum is homozygous for g, e4 and f4, and is heterozygous for g4• The points of rearrangement are shown on the a, b, g chromosome (Figure 81). D. propachuca is homozygous for g and c, and heterozygous for h4 (Figure 82). D. pachuca is homozygous for g, c, h4 and i4 and is heterozygous for j4 and k4• Rearrangement of the a, b, g, c chromosome (Figure 82) to include h4 and i4 will give the standard sequence. Once this is done, the breakage points of j4 which overlaps both the h4 and i4 inversions can be easily interpreted. e2 D. martensis is homozygous for c, but lacks g. It is also homozygous for d2, and f2 and heterozygous for g2• The evolution of this chromosome was given in Wasserman and Wilson (1957). The standard salivary map can be obtained from Figure 82 by first reinverting the g sequence to form the c chromosome, and inverting the d2, e2, and f2 sequences, the g2 being heterozygous. D. mulleri, aldrichi and wheeleri are each homozygous for the c, g, f chromo­some shown in Figure 83. D. longicornis is homozygous for g, c, f, and t3 and heterozygous for u3 (Figure 83). D. arizonensis is homozygous for g, c, f, hand i (Figure 83). Inversion i over­laps and follows h. D. mojavensis is homozygous for g, c, f, h, q, r, s. The evolution of this chromo­some was incorrectly given by Wasserman ( 1954) in that the breakage points of the q, r and s inversions are inaccurate. However, these three inversions have occurred, and the approximate lengths of the inverted sequences are correct. 3 CHROMOSOME: The primitive third chromosome of the repleta species group and also of the mulleri subgroup differs from that of repleta by having the 3b arrangement (Figure 74). A total of seven inversions have taken place during the evolution of the mulleri subgroup: (1) 3a and 3k are homozygous in martensis (Figure 74). (2) 3e and 3j, the latter overlapping and following the former, are homozy­gous in anceps (Figure 74) . ( 3) 3a and 3c, two independent inversions, are shown in a salivary map (Figure 76) . This is the standard chromosome of mulleri, aldrichi, wheeleri, longicornis, pachuca, propachuca, ritae, tira and desertorum. ( 4) D. desertorum which is homozygous for a and c is heterozygous for 3r (Figure 76) . ( 5) 3a and 3d are homozygous in mojavensis which lacks the 3c. The salivary map of this species can be derived from Figure 76 by reinverting the c to form a 3a chromosome and then inverting the large d inversion. 4 CHROMOSOME: The primitive fourth chromosome is identical to the fourth of repleta. A total of three inversions have taken place in this chromosome: 4a is homozygous in anceps, region E2a to F4a being inverted (Wasserman, I J~ r--K I II . a:> F E---1-11 '.-L~.I 21 2 . I i.l 1 11 J .... . .... . .. ~~~-·~ /~1 " !£ ,, x ~~,~1er~·b)~\\lalll:~ A A B B c .,-....,.--... .,.-:DIE E G H 11 11 Fig. 73 K ~1llPTh,~· !~,.1i~!!ll ~M~111 J (L):~, ~ ,, J ~ 3 ~.11-.wi.1Hifif11i11llW~•~tmllB1'11;tUIIlm1Iiit LJJJf~;\;.!~/t~alJfiltf1]i {{f1JllJi11 ~ I I I 2I 3I4I5 I I I 2I3I 4 I 5 I I 2I 3IJ 5 I I I2I I I I 2 I3I4I5 I "I I 2I3 I 4 1" I I2 I 3 I A A B Bic C D E E F F G ~ Fig. 74 ~ ~ rr1m1i ,,Jaw ~A l~ 111: "' \'. I r) ~ '~ '{r1114r,1"'' !I} ~I'· ~ IllllJJ) 1 ]~lllllll t~I \Wfl!Ju\Cll::::::: s \li~~JJ. '-i ;~ ll!u11JIUJfJttrnrrill'rZJ~i ,tlLiJ~1fn\l\}.lic (~till1i.J~~,·~i11~1flfJJlSJ ~ .... 1 112 3 4 s l 1 2l 3l 4isl1 i 2l3 l 4l s/ 1l2l3i4ls / 1l2i3l4isEF 1i 2l3FG1 l 2 314GIH Fig. 75A A B B c c D DE t F T B r J ~ ~ a {j ::>i' 01'.) w ~ ~ . '1"Il\~1· 1 I .. I E ll r· 1trraih!\1\ ~).... I L I 9 ff11ll7it!mr111 'fl'~' .n1 ,.,l '~\till\UJJJ1QJ.1; 11rrerurn1~11'~'~•fr1;:1wm.~~:lttIMrJ:11~~KDi1~lf,U1JJIJitt: ~ ..... r~~ , 121 · 1 , 12131 4 15 1 121314 1 5 1· , 121314 15~ 11i1.,, 1314' 1 516I'r 1 2 1314 15 1 ~ A C DE EF ~_:_]FG GH Fig. 79 L----+-----1 ~ ,____M N d '-------+W4/+ -§ 13 13 \( 01:) R3 .;JI) • I .. I E rIzl " . . LI 9 g I I • 1 ··i~mt1JCmrn~-.1Jtutam1il•ruiiui~mc~~, :\" , 121 I 1 1131, 12151 121314151 1121!1415 1 jJ,,1314151611121314151 'A c I [_ DE L FG G H Fig. 80 5 2;+ I y 2;+ FIGS. 77-80. Salivary gland chromosomes of the mulleri subgroup. Fig. 77. The harnatofila second chromosome, a, b, x4, y4. Fig. 78. The buzzatii second chromosome, a, b, k, x3, w3. Fig. 79. The primitive second chromosome, a, b. Fig. 80. The nigricruria second chromosome, a, b, t2, u2, v2, w2. 0 - "" Th£ University of Texas Publication gene sequences found: D. anceps having seven inversions, Xf (Figure 73), 2w (Figure 79), 3e and 3j (Figure 74), 4a, and 5a and 5b (Figure 75); D. stalkeri having three inversions, 21, 2m and 2n (Figure 79); and D. eremophila having four inversions, 2d5, 2e5, 2f5 and 2g5 (Figure 70). Several strains identified as D. hoeckeri, including the strain from which the type specimens were chosen (Brncic, 1957), were cytologically identical to the standard sequence of D. nigricruria (Table 1), being homozygous for Xk (Figure 73), 2t2, 2u2, 2v2, 2w2, 2x2 (Figure 80) and 5f (Figure 75). All crosses among the hoeckeri strains and between hoeckeri and nigricruria resulted in fully fertile offspring. No differences in the genitalia! apparatus were found. The only distingliishing character seems to be the highly contrasting color patterns. D. hoeckeri is therefore a synonym of D. nigricruria, and, at most, a Chilean sub­species. D. nigricruria, therefore, is homozygous for its standard gene order wherever it has been sampled in South and Central America, but becomes poly­morphic for 2s2 and 2y2 (Figure 80) and 4c and 4d in Mexico (Wasserman and Wilson, 195 7) . An examination of several D. hamatofila strains (Table 1), including the Texas· strain previously reported on (Wasserman, 1954), showed that in addition to 5c (Figure 75 ), the standard contains two inversions which had been over­looked, 2x4 and 2y4 Inversion 2v4, lacking in the Texas strain, is homozygous in • both of the Arizona stocks, one of which is polymorphic for 2z4 and 2a5, small rearrangements separated by only about a dozen bands but between which recombination occurs since chromosomes containing either 2z4, or 2a5, or both, or neither, exist in the laboratory strain. The hamatofila standard 2 and the breakage points of the heterozygous inversions are given in Figure 77. The second chromosome of buzzatii is difficult to analyze, but seems to consist of 2k and two inversions overlooked by Wasserman ( 1954) , 2w3 and its included inversion, 2x3, whose breakage points are so similar that these two changes superficially appear not to have taken place. D. buzzatii is heterozygous for three inversions, one of which, 2j, is present wherever the species occurs: Lebanon (Wasserman, 1954), Australia (Mather, 1957) and also South America (Table 1) where it undoubtedly arose. The other two sequences were found in an Argentina strain (Table 1) which contained three different second chromosomes: the standard; 2y3, a new sequence; and 2j, z3 ; the Z3 apparently arose in a j chro­ mosome since it is always associated with the j anB. seems to share one of its breakage points with this sequence. The evolution to the buzzatii 2 chromosome is given in Figure 71 ; the chromosome map showing the heterozygous inversions is given in Figure 78. D. pegasa has eliminated mating courtship. The male attacks the female, rapidly gains the mating position and once established remains on the female for a considerable time usually copulating more than once. Even after a successful copulation, the male refuses to alight despite the obvious attempts by the female to dislodge him. Such pairs have been observed continuously for two hours; spot checks at fifteen minute intervals indicate that couples may remain paired for more than 14 hours. Attempts at interspecific matings have been unsuccessful; the males can discriminate and ignore foreign females. When the proper females are lacking, the males will often pair among thems~ves, remaining coupled for r------U4;+--------. F1"~-:.il ,"1' •1, _ I r:r:-1': 1I G'I+ IPl+T 1' iTir· 111 ~:: ''t\';f.1.1'1'JIJ-'l ll·~r1 rHf!ff "ttf\I"'Jl(fl.~~~vUl\.n1.1~:'l:·rr1(1ff.1; 1nt71m1'1n:1rn'.' :.n·~..,,-· ill)}~1\'Y'~'. nff 'T''J~m111A{: ~'°'1tn ·Jti:rnl(!!t ,~~~~,~Lw ll~J . W1l:.. !ii lll",.1:·I2 IJ" ~\!11 1i....IL ~ ... t~/'!~~ 1li41 • ~::1;u.!~•.\••IWJ,iJilu'''"~.:.-~/MI~1~· • 1 4 2 1111 1111 111 A C DE1 FG G H Fig. 81 i ~ I '"'~·:.;~~ ,. I I H4/+ 1"'14 ~ ,~ 'rI z' I I I >.1 " I£Iz I I '.) I 01'.)LI~ . . ·. I>I " I E n,-r9 ~uallllllfllll1:llman11~~-r1'.~lUllill1JiLl'll%~1(f•m&11t1~JlillCtaJJ(l'.'.\U111llill$· ~ ~' ~2 I ' I I I' '' I , I, A\1 I' I' I • i,,I, I, I,I, I, b J-tlI , I, 1, I,clH ~ 0 ~ '--·-----G2!+--__,· .... ~ . i~ ,,/£1 z, I 'I '1"1£1zl' '.)I. ,al zl. ,., . . ,,,. . t..,~••1 ',1",rn,_;i~,~. 1 " {; ~ ~d\U1Jltn!lfllntt~nrt1111rr;:;NUS~1fJl.1turl~lf.Jll~l~J&)))~~~~I. rA I 2 I I 4AIB L e u' 7c I 4 I5 olE I I 1314 I 5 TI I 3 3 2 LJ. ~ (H) I '------( H) I---------' Frcs. 81-83. Salivary gland chromosomes of the mulleri complex. Fig. 81. The second chromosome, a, b, g. Fig. 82.. The second chromosome, a, b, g, c. Fig. 83. The second chromosome, a, b, g, c, f. 0 - ~ The University of Texas Publication as long as a half hour. A triplet, consisting of three males has been observed on several occasions. The pegasa strain, which originated from a single pair of individuals, has two x3 c5 second chromosomes: and x3, k, b5, ; the x3 and k are identical to those sequences of buzza.tii whereas the b5 and c5 are unique to pegasa. Inversions b5 and c5 share one of their breakage points; the 2k inversion is independent of this complex (Figure 78), but in some twenty-odd larvae, the three inversions have always been together. The presence of 2x3 and 2k in both buzza.tii and pegasa indicates phylogenetic kinship. The cytological evolution of these two species is shown in Figure 84; that of the second chromosome is detailed in Figure 71. The D. buzzatii 2j/t 2jz 3/t 2y 3/t 2w 3 5g 2k D. peqasa f D. promeridiano ~2k_l+ 2_b_f_+_2_c_f_+.....l-l2k/t I ~ 55 I / __ ~ D. meridionalis ~r:-;i j 3_4F,~ D. merid/ana . D. niqricruria ~ . -----~ ~ noens1s 2t2 2u2 2x2 Xk Xf hamatophila 4o 2y4 2v4/+ 2o5/+ FIG. 84. Phylogenetic relationships among ten species in the mulleri subgroup. The standard arrangement of each species is the sum of all homozygous inversions leading out from the primi­tive. Heterozygous inversions which occur with the standard are indicated by the inversion shown over the plus sign. ancestor of the two species was homozygous for 2x3 and heterozygous for 2k. D. pegasa has maintained this situation, adding 2b5 and 2c5 in the heterozygous condition, whereas in the line leading to buzzatii, 2w3, which followed and over­lapped 2k, became homozygous with the result that 2k became fixed. D. buzza.tii is also homozygous for 5g and polymorphic for other gene sequences mentioned above. D. hamatofila, D. buzza.tii and D. pegasa, are closely related to the mulleri Wasserman: The Mulleri Subgroup complex; the first two species produce in interspecific crosses viable zygotes with at least one of the mulleri complex species (Table 4), and the latter shows rela­tionship by cytological similarity with buzzatii. The mulleri complex~ Each of the twelve species in this complex contains one or more of the following six inversions, 2c, 2f, 2g, 2h, 3a and 3c; five of these inversions are independent gene orders among which segregation is possible; the sixth, 2h, includes 2c-although we have found 2h only in the presence of the more widely distributed 2c, it could occur independently-and overlaps 2f which, therefore, must be present in all chromosomes containing 2h. The distribution of these inversions among the species is shown in Table 5: each inversion is homozy­gous in the species in which it occurs; however, since the fasciola subgroup, an offshoot of the mulleri complex (Wassennan, 1962b), consists of species with or without 3c, the ancestral population of this subgroup is considered to have been heterozygous for this sequence. The distribution of these six gene rearrangements among the twelve species cannot be interpreted on the basis of the splitting of populations and subsequent independent divergence. One can assume that there has been a series of cycles in which populations became isolated, diverged as to gene orders, and then hybridized. Wasserman ( 1954) showed that three such secondary hybridizations could account for the cytological constitutions of the six species examined at that time: arizonensis, moiavensis, mulleri, aldrichi, wheeleri, and ritae; data relating to the six additional species further complicate the situation and demand more secondary hybridizations. These difficulties can be avoided by considering the species as segregants from a large-scale hybrid swarm involving two species which differed by a number of inversions, perhaps species which were polymorphic since the presence of a minimum of three second chromosomes is required: those lacking 2f; those with 2f; and those with 2f and 2h. Since 2h overlaps 2f there can be no segregation of these two sequences. How­ever, it is felt that neither of these two interpretations is satisfactory, for although the formation of hybrid swarms and the subsequent introgression and segregation TABLE 5 Distribution of the six inversions segregated among the twelve species in the mulleri complex and the ancestor of the fasciola subgroup: + indicating presence; -indicating absence Inversions Species 2g Zc Zf Zh 3a 3c D. tira + + + D. ritae + + + D. desertorum + + + D. pachuca + + + + D. propachuca + + + + D. mulleri + + + + + D. aldrichi + + + + + D. wheeleri + + + + + D. longicornis D. moiavensis D. arizonensis D. martensis + + + + + + + + + + + + + + + + Fasciola subgroup +!­ The University of Texas Publication could furnish the necessary chromosomal types, the critical problem is not to find a source for the variability-inversions are far from rare in Drosophila-but to unearth a mechanism which can explain the distribution of these inversions among the species. In obligatory bisexual organisms, hybrids or recombinations from matings involving dissimilar parents do not form new species without other, extrinsic factors. Wasserman ( 1954) argued that the most plausible interpretation is that there existed an ancestral species composed of geographically semi-isolated populations which differed in their inversional constitutions. These populations diverged in time to form the species. Newer information was later added along with a brief discussion on the general role of inversions with emphasis on speciation (Wasser­man, 1960) . The evidence is presented in detail here. It is assumed that the ancestor of the mulleri complex occupied a large area, the deserts of Southwestern United States, Mexico south into Central and prob­ably South America: the modern species in the complex are limited to the cactus deserts of the New World; the fasciola subgroup evolved from this ancestor, and although most of its species are inhabitants of mesophyllic forests in Central and South America, this may be a secondary modification because the most primitive species of this subgroup, D. fulvalineata, is a desert form found in the United States and Mexico (Wasserman, 1962b). The deserts of the New World do not occupy a single continuous area, but are divided into a number of subunits of varying sizes. This is especially true in Mexico, where broken, mosaic patterns of habitats seem to be the rule. The result was the presence of semi-isolated populations of the precursor of the mulleri complex between which some gene exchange in the form of migrants took place, but which were sufficiently isolated so that a certain amount of adaptation to local conditions was possible. An adaptive rearrangement would be incorporated into the gene pool of the population in which it arose, and spread to neighboring popu­lations if it were also advantageous there. If not, then the inversion would be stopped near the border of the two populations; more distant populations would never get a chance to test this sequence. The result would be a step-like dine of inversions, each inversion occurring in its own specific area, with a certain amount of overlapping of ranges among inversions. This level of cytological evolution is depicted in Figure 85, where the cytologically distinguishable popu­lations are called subspecies; subspecies A, B, C and D were proposed by Wasser­man ( 1954) ; subspecies E, F and G are required by the new data. Divergence of these seven subspecies yielded the precursor of the fasciola subgroup (from sub­species G as described in Wasserman, 1962b) and six members of the mulleri complex. Further speciation of the more usual type of geographical splitting without the sharing of inversions but with the accumulation of other rearrange­ments specific to the isolated species could account for the twelve members of the complex as it now stands. It is of course possible, and indeed probable, that more subspecies were involved in the original population and that some of the inver­sions limited to one or two of the modern species and therefore shown as having originated at a later date (Figure 85) were in fact in the original population; the seven subspecies are the minimum number needed to explain the data, but any larger number may have been present. W a.sserman: The M ulleri Subgroup D. wheeleri D pachuca h4 i4 D propachuca­ 22I2h4/+ I 2j4!+ 2k4/+ D. longicornis D. desertoru~1 2h!I+ I ~D. mojovensis 2g4/+ 3r/t l 2e4 I2p/t I D. riCe \ W D. tira l ~rfensis ~ I2p/2u4 I 2d2 2fl 2f2 SUBGROUP FASCIOLA 3k Xj 2g2/+ Xa Xb Xe 2a 2b 3b I Fw. 85. Phylogenetic relationships among the twelve members of the mulleri complex, and the ancestor or the fasciola subgroup, which are shown to have evolved from a large, cytologically polytypic species. The genetic information, described above, not only confirms the affinities be­tween B, C and D, but further subdivides B into its four components and relates these to the other two species. One arrives at the linear order of arizonensis, mo;avensis, mulleri, aldrichi, wheeleri, with longicomis as a side branch off from mulleri. The South American martensis, ssp. F, could have been contiguous with either ssps. B, or C, or E. However, since C, moiavensis, is now limited so far as known to the Mojave Desert in California, and E, both pachuca and propachuca, is known only from a single locality in Mexico, martensis is tentatively shown as having been contiguous with ssp. B, which in the form of aldrichi, occurs in Colombia. The original seven species probably evolved in the positions relative to each other shown in Figure 85. A major feature which contributed to the speciation of this complex was the mosaic, island-like distribution pattern of the populations. The University of Texas Publication There is no evidence which might indicate that these species evolved as peripheral isolates off from a highly polymorphic centrally located population as may occur in forms which have an areal, continent-wide distribution. The development of the seven subspecies and the subsequent speciation needs further elaboration. The future of a species with a mosaic distribution depends upon the interaction of a number of factors which beside the ones usually associ­ated with a single population include interdeme competition. Differentiation was very likely enhanced by random processes such as chance differences as to which of several adaptive mutations were selected for in an individual population, and the fixation of alleles through fluctuations in population size, marked fluctuations being a common phenomenon among the present day desert species (Patterson, 1943). Geographical isolation is the essential prerequisite of speciation in obligatory sexually reproducing organisms. The rate of divergence of allopatric populations is influenced by a complex interplay of a number of factors of which natural selection must be considered the directing force. If the selective forces in the two localities tend to stabilize the gene pools or move them along parallel lines of change, population characteristics which tend to promote drift and reduce the efficacy of selection would enhance divergence. However, if selection is disruptive, the more genetic variability exposed to selection, the quicker the response. Cyto­logical polymorphism ties up alleles into epistatically _balanced complexes; homozygosity allows for free recombination and releases this concealed variance. A cytologically homozygous system may be able to respond to selection more rapidly than one where recombination is suppressed through a large part of the genetic material (Carson, 1958). However, for long-time changes and survival, the most effective system may be one of compromise, i.e., limiting the amount of polymorphism to a segment of the genetic material. Total recombination may then be unaltered. Moreover, crossing over within the inverted regio~s of the homozygous individuals will allow for the formation of new combinations which, if advantageous, can be preserved and quickly multiplied in the heterozygous condition (Stone, et al., 1960). Cytological polymorphism is one mechanism by which a population may adapt to a heterogeneous environment since it preserves within the same .population different epistatic gene combinations. Both on the cytological and the genie levels, evolution may proceed from one homozygous state with some genetic variability to another such state, or may be a series of replacements of one balanced heterogeneous genetic system by another, the end result of which is superficially the same: homozygous differences between popu­lations separated by time and space. The distribution of the inversions in the mulleri complex offers a clue to the history of these species. The six shared inversions are each homozygous in those species in which they occur (Table 5). They may have originally been main­tained in the populations because of heterozygote superiority, homozygosity arising by polymorphic replacement or genetic drift; or natural selection may have fixed the homozygous sequences. Polymorphic replacement implies the origin of a third sequence which over­laps the inverted region. The resulting intrapopulation triad may lead to the Wasserman: The Mulleri Subgroup elimination of one of the three sequences (Wallace, 1953). Several possibilities are present: the new sequence may occur in the inverted chromosome and either, (a) the standard, or (b) the single inversion chromosome is eliminated; or the new sequence may occur in the standard chromosome and either ( c) the standard or ( d) the old single inversion is eliminated; the elimination of the new sequence is trivial. Situation (a) which results in the replacement of an old polymorphism by a newer one, and the fixation of the first inversion can account for the follow­ing inversions: ( 1) 2g, in mo;avensis (overlapped by 2r); in ritae and tira (2p); in longi­cornis (2u3); in arizonensis (2i); and in desertorum (2g4). (2) 2f, in longicornis (2t3); and in both mo;avensis and arizonensis (2h) . ( 3) 2c, in martensis (2f2) ; and in mo;avensis (2r). ( 4) 2h, in arizonensis (2i) ; and in moiavensis (2r) . (5) 3a, in moiavensis (3d). (6) 3c,indesertorum (3r) . In the mulleri complex there is no example of (b) which would yield a complex overlapping configuration in the heterozygote. Situation ( c) which also gives a complex figure in the heterozygote has not been involved in the evolution of the six inversions in question, but seems to have occurred in our strain of tira where inversions 2p and 2u4 are present, but the intermediate standard is lacking. Situation ( d) is of particular interest to us since it results in the elimination of the original inversion. Cytologically indistinguishable, but more likely than (d), is the presence of a pre-existing polymorphism between the standard and a sequence whose breakage points overlap the inversion in question. This estab­lished polymorphism may prevent the inversion from entering the population. These two possibilities may account for the exclusion of the following inversions: ( 1) 2g, absent in martensis (replaced by 2e2). (2) 2f, absent in pachuca and propachuca (2h4) , and therefore desertorum, ritae and tira which were more distant populations. (3) 2c, absent in desertorum (2g4). ( 4) 2h, absent in longicornis (2t3 or 2u3). It can be seen from the above that overlapping triads may have been important in fixing inversions or excluding them. However too much emphasis should not be placed upon the triads. Homozygosity of the two third chromosomal inversions can not be satisfactorily explained by competition with other polymorphic sys­tems. In the second chromosome there were nineteen other rearrangements from which to choose in seeking explanations for the homozygosity of four inversions, yet not all of the inversion-species relationships can be thus explained. A notable exception is the mulleri-aldrichi-wheeleri trio of species which is homozygous for 2g, 2c, 2f, 3a and 3c. No other gene orders are known for any of these three species, yet both mulleri and aldrichi have been sampled over their entire known distributions: mulleri from Nebraska, Texas, Mexico and Jamaica; aldrichi from Texas, Mexico, El Salvador and Colombia (Table 1; Wasserman, 1954; and Wasserman and Wilson, 1957). Moreover, even in some of the cited examples where homozygosity or exclusion is presumably due to competition with other The University of Texas Publication polymorphic systems, the supposed superior polymorphic system has been lost: mo;avensis, arizonensis and ritae appear to be homozygous for their more ad­vanced sequences. D. pachuca is interesting cytologically in that it is heterozygous for both a small inversion, 2k4, and a large inversion, 2j4, which includes the smaller one (Figure 82). Three types of chromosomes have been recovered from the natural population: the standard, 2k4, and the 2j4 Although these are not overlapping • sequences, the net result is the same: the coexistence of three gene sequences in the same population. In the laboratory, a product of a double crossover, the j4k4 chromosome, was obtained from a female carrying a j4 and a k4 chromosome, but it is not known whether this fourth type of chromosome also survives in nature. It is conceivable that the homozygosity arose in small populations due to genetic drift fixing one of the sequences despite heterozygote superiority. There is, how­ever, evidence that inversional polymorphism is particularly resistant to drift: small laboratory strains usually maintain two or even more forms of chromo­somes. Some of the naturally occurring inversion polymorphisms are very old. There are four known cases in Drosophila where two "sister species" have the same polymorphism and presumably the polymorphism predates the speciation (Wasserman, 1960) . There are three known cases in the repleta group where an ancestral polymorphism has been preserved in one of the daughter species and lost in the other: mercatorum is heterozygous for 2v3, while paranaensis is homo­zygous for this arrangement (Wasserman, 1962a); fulvimacula is heterozygous for 212, while fulvimaculoides is homozygous for this arrangement and also a subsequent overlapping inversion, 2a3 (Wasserman and Wilson, 1957); pro­pachuca is heterozygous for 2h4, while pachuca is homozygous for 2h4 and 2i4, a contiguous inversion (Figure 82). There are then eleven examples of polymor­phism for inversions being maintained in some populations for a period of time longer than that needed for speciation. The above examples do not eliminate the possibility of having random fixation of a sequence in a small population, but they do indicate that this is not a general phenomenon. The repleta group in general, and the mulleri complex in particular, consists of species which are ecologically restricted. Although most of these species have been found to be cytologically polymorphic, when compared to the other members of the genus, this polymorphism is extremely limited, and the vast majority of the species are effectively homozygous. This homozygosity is most probably corre­lated with the marginal existence of these desert species. The deserts exclude all gene orders except those few which are best adapted to the relatively few eco­logical niches. D. mulleri and D. hydei, the two species which have been studied, have a very high frequency of recombination (Spencer, 1957). If these species are typical, the repleta group is characterized by a high frequency of recombi­nation throughout most of the genetic material. If one accepts the premise that homozygosity for a particular gene sequence is adaptive in marginal areas, it follows that cytological differentiation may play an active role in speciation. Neighboring populations adapted to different aspects of the environment may be able to speciate if migrants are few and their offspring are ill-adapted. Selection would favor the development of sexual isolation between Wasserman: The M ulleri Subgroup the populations if hybrids are eliminated. The existence of homozygous inver­sional differences presents to the populations segments of the genetic material which are particularly resistant to fragmentation by recombination with the foreign genetic systems. Factors promoting sexual isolation, if they were linked to these inversions would be especially effective in separating the populations. It is to be expected that these factors would originate and persist at the borders between allopatric, but contiguous, populations. The presence of such inversion­linked factors has yet to be demonstrated. Itis unlikely that the inversional differ­ences originated for the purpose of isolation, but once present, they may have been the raw materials leading to isolation in a population structure such as existed in the mulleri complex. SUMMARY Eight new Drosophila species are described and D. hoeckeri is made a synonym ·of (and perhaps a subspecies of) D. nigricruria, bringing the total number of repleta group species in the mulleri subgroup to twenty-seven. Twenty of the species were available for morphological study, and camera lucida drawings of the spermathecae, penes and male external genitalia of these species are given. Other useful morphological characters of all the inulleri species are presented in tables. The results of interspecific crosses are reviewed and briefly discussed. Meta­ phase karyotypes of twenty-three species are given. An analysis of the salivary gland chromosomes of twenty-two species, represented by eighty-five strains, was made. This includes salivary gland chromosome maps of most of the species and a description of the heterozygous intraspecific rearrangements and the homozy­ gous interspecific rearrangements. Phylogenetic relationships, as determined by the salivary chromosomes and supplemented by the interspecific hybridization, were determined for twenty-two of the mulleri species, and are presented in two figures. The subgroup consists of two complexes, the meridiana complex, and the mulleri complex, plus a number of species which are considered miscellaneous forms for which there is no cyto­ logical basis for further grouping. The meridiana complex is comprised of two widely separated South American species, D. promeridiana and D. meridiOnalis, each known from single localities, and a third, the North American species, D. meridiana, in which there is an eastern subspecies, D. m. meridiana and a more southern and western form, D. m. rioensis. The meridiana subspecies has the primitive karyotype, five pairs of rods and one pair of dots, whereas the other subspecies and the South American species have a pair of metacentrics formed as the result of a centric fusion between two autosomes, the second and the third chromosomes; the three latter forms must have arisen from a southern population where this centric fusion was fixed. There has been very little chromosomal change in the history of this complex: the standard sequence for rioensis, meridionalis and probably meridiana is identi­cal to that proposed as the primitive sequence of the repleta group; the standard of promeridiana differs from these by only one inversion. Inversion polymor­phism must be rare in this complex: one locality of rioensis was heterozygous for Tlze University of Texas Publication a single inversion, whereas six other localities of rioensis, ten of meridiana and one each of promeridiana and meridionalis were homozygous. The mulleri complex consists of twelve species each of which, besides contain­ing species-specific sequences, is homozygous for one or more of six inversions. These six inversions are distributed among the species in such a manner that they could not have been one of several characters arising in geographically isolated populations. Rather it is proposed that the ancestral species consisted of a number of small, semi-isolated populations among which some gene exchange was pos­sible. The species was polytypic for combinations of inversions: each inversion had its own distribution, overlapping the ranges of one or more of the others. Fragmentation of the ancestor into seven distinct forms, several of which later split into two or more species, resulted in the twelve forms which share the six inversions. These six homozygous inversions may have been in the past part of heterotic polymorphic systems, and may have been fixed because of competition with other overlapping inversion systems. However it is likely that many of the inversions are homozygous because the homozygote was superior to the hetero­zygote in the environment in which the population existed. Allopatric cytological differentiation in the form of homozygous inversions may have actively con­tributed to speciation by preserving locally adapted gene complexes within the inverted regions and may have captured genes which would isolate the popula­tion from ill-adapted migrants. LITERATURE CITED Brncic, D. 1957. Las especies Chilenas de Drosophilidae. Col. Monografias biologicas Univ. Chile, 8. Santiago: Universidad de Chile. Carson, H. L. 1958. Response to selection under different conditions of recombination in Dro­sophila. Cold Spring Harbor Symp. Quant. Biol. 23: 291-306. Clayton, F. E., and C. L. Ward. 1954. Chromosomal studies of several species of Drosophilidae. Univ. Texas Pub. 5422: 98-105. Clayton, F. E., and M . Wasserman. 1957. Chromosomal studies of several species of Drosophila. Univ. Texas Pub. 5721: 125-131. Crow, J. F. 1942. Cross fertility and isolating mechanisms in the Drosophila mulleri group. Univ. Texas Pub. 4228: 53-67. Mather, W. B. 1957. Genetic relationships of four Drosophila species from Australia. Univ. Texas Pub. 5721: 221-225. Mettler, L. E. 1957. Studies on experimental populations of Drosophila arizonensis and Dro­sophilamo;avensis. Univ. Texas Pub. 5721: 157-181. Patterson, J. T. 1943. The Drosophilidae of the southwest. Univ. Texas Pub. 4313: 7-216. ----. 1947. Sexual isolation in the mulleri subgroup. Univ. Texas Pub. 4720: 32-40. Patterson, J. T., and M. L. Alexander. 1952. Drosophila wheeleri, a new member of the mulleri subgroup. Univ. Texas Pub. 5204: 129-136. Patterson, J. T., and G. B. Mainland. 1944. The Drosophilidae of Mexico. Univ. Texas Pub. 4445: 9-101. Patterson, J. T., and W. S. Stone. 1952. Evolution in the genus Drosophila. Macmillan Co. Spencer, W. P. 1957. Genetic studies on Drosophila mulleri. II. Linkage maps of the X and chromosome II with special reference to gene and chromosome homologies. Univ. Texas Pub. 5721: 206-217. Wasserman: The Mulleri Subgroup Stone, W. S., W. C. Guest, and F. D. Wilson. 1960. Evolutionary implications of the cytological polymorphism and phylogeny of the virilis group of Drosophila. Proc. Nat. Acad. Sci. 46: 350-361. Wallace, B. 1953. On coadaptation in Drosophila. Amer. Nat. 87: 343-358. Warters, M. 1944. Chromosomal aberrations in wild populations of Drosophila. Univ. Texas Pub. 4445: 129-174. Wasserman, M. 1954. Cytological studies of the repleta group. Univ. Texas Pub. 5422: 130-152. -----. 1960. Cytological and Phylogenetic relationships in the repleta group of the genus Drosophila. Proc. Nat. Acad. Sci. 46: 842-859. 1962a. Cytological studies of the repleta group of the genus Drosophila: III. The mercatorum subgroup. This bulletin. 1962b. Cytological studies of the repleta group of the genus Drosophila: VI. The fasciola subgroup. This bulletin. Wasserman, M., and F. D. Wilson. 1957. Further studies on the repleta group. Univ. Texas Pub. 5721 : 132-156. Wharton, L. T. 1942. Analysis of the repleta group of Drosophila. Univ. Texas Pub. 4228: 23-52. -----. 1943. Analysis of the metaphase and salivary chromosome morphology within the genus Drosophila. Univ. Texas Pub. 4313: 282-319. -----. 1944. lnterspecific hybridization in the repleta group. Univ. Texas Pub. 4445: 175-193.. Wheeler, M. R. 1954. Taxonomic studies of American Drosophilidae. Univ. Texas Pub. 5422: 47-64. -----. 1957. Taxonomic and distributional studies of Nearctic and Neotropical Drosophili­dae. Univ. Texas Pub. 5721 : 79-114. VI. Cytological Studies of the Repleta Group of the Genus Drosophila: VI. The Fasciola Subgroup 1 MARVIN WASSERMAN INTRODUCTION Nine Drosophila species of the repleta group form a phyletic unit which differs from all other known repleta species by at least three homozygous inversions, 202, 2e3 and 213• These species are therefore combined in a new subgroup, the fasciola subgroup; the type species, Drosophila fasciola Williston, 1896, was the first of these species to be described. Unlike the majority of the desert-inhabiting repleta group species, the fasciola subgroup prefers the wetter areas of the West Indies, and Central and South America, and is the dominant section of the repleta group in the rain and cloud forests and in the cultivated banana and coffee fincas. Table 1 lists the strains of the species studied, four of which are new. MORPHOLOGY Typically, in the repleta group, the mesonotal hairs and bristles arise from spots. In the fasciola subgroup there is a tendency for these spots to be lost, either due to a general reduction of pigment in the integument, or else due to the fusion of the spots to form more or less elaborate patterns, with the result that two of the more extremely modified species, D. fulvalineata and D. paraguttata hereto­fore had not been recognized as repleta group species. Table 2 lists the wing vein indices and the number of coils found in the testis and the ventral receptacle of the nine species. With respect to the coiling, the fasciola subgroup (testis: 8 to 15; ventral receptacle: 20 to 76) appears to be intermediate between the mulleri subgroup (testis: 4 to 9; ventral receptacle: 12 to 35, discounting the few forms with exceptionally high number) and the melanopalpa subgroup (testis: 8 to 16; ventral receptacle: 52 to 116). In the fasciola subgroup, the spermathecae show tendencies towards degeneration, the normal organs of D. fulvalineata being considered here as having the primitive condition (Figures 1-9). The penis and male genitalia of each of the nine species are diagrammed in Figures 10-27. Abbreviated descriptions of four new species are given below, as are brief taxonomic notes on some of the other species. The holotype and nine paratypes of each of the new species along with slides containing wings, legs, mouthparts and genitalia of all species are deposited in The University of Texas collections. 1 Present address: Zoology Department, University of Melbourne, Melbourne, Australia. The University of Texas Publication TABLE 1 List of strains and localities 4 Species Strain Locality x 2a,b,o2,e3,l3 3b fasciola H112.16 St. Lucia, B. W. I. c + + + H 86.9 Balboa, Panama c + + + H231.12 Georgetown, Br. Guiana + c + + fascioloides H 51.5 Lancetilla, Honduras p2,d3 c + + H163.28 Turrialba, Costa Rica p2,d3,k3/+ + + c H181.43 Barro Colorado, Canal Zone p2,d3/+ + + c H182.37 Barro Colorado, Canal Zone p2,d3,k3/+ c + + H186.54 Sierra Nevada de Santa Marta, Colombia p2,d3 c + + H327.11 Santa Domingo de los Colorados, Ecuador p2,d3/+ c + + H336.9 Belem, Brazil p2 c + + coroica H346.43 Coroico, Bolivia + c,p + + pictilis H 27 .2 Lake Pichichuela, El Salvador +/n3,m3 c + + Costa Rica B Costa Rica o3 c + + pictura H109.28 Port of Spain, Trinidad + c + + H344.14 Montero, Santa Cruz, Bolivia n3 c + + fulvalineata A 2.4 Patagonia, Arizona c4,d4 + + + paraguttata H356.18 Hardware Gap, Jamaica p2,r2,i3/+ c,q + h,i,j mojuoides H107.7 Arima Valley, Trinidad p2,r2,n 2,j3 c,m + h,i,j H332.9 Maraval, Trinidad p2,r2,n2,j3 c,m + h,i,j moju H400.30 Palmira, Costa Rica p2,r2,n 2,q2 c,m + h,i,j H182.23 Cerra Compagna, Panama p2,r2,n2,q2 c,m + h,i,j H183.23 Cerra Compagna, Panama p2,r2,n2,q2 c,m + h,i,j H 80. 12 Barro Colorado, Canal Zone p2,r2,n2,q2 c,m + h,i,j H181.33 Barro Colorado, Canal Zone p2,r2,n 2,q2 c,m + h,i,j H408.1 Madden Forest, Canal Zone p2,r2,n2,q2 c,m + h,i,j H303.8 Las Cruces Trail, Canal Zone p2,r2,n2,q2 c,m + h,i,j H194.33 Villavicencio, Colombia p2,r2,n2,q2 c,m + h,i,j H343.15 Montero, Santa Cruz, Bolivia p2,r2,n2,q2/+ c,m,o/+ + h,i,j H336.26 Belem, Brazil p2,r2,n2,+/q2/a4b4 c,m,o/+ + h,i,j TABLE 2 Morphological characters of the species in the fasciola subgroup Species c. I. Wing characters 4vn 5x 4c Heavies• Ventral receptacle Coils of testis Outer Inner ful valineata 3.8 1.66 .8 1 .69 .32 50 7 7 fasciola 1.9 1.75 .68 1.17 .5 30 5 5 corozca 2.5 1.86 .71 1.02 .45 30 4 4 pictilis 1.9 1.86 .80 1.19 .54 65 8 5 pictura 2.2 1.79 .81 1.09 .44 76 8 7 fascioloides 2.2 1.59 .79 1.02 .57 35 7 4 paraguttata 2.4 1.56 .79 .95 .41 20+ 5 4 moju 2.6 1.81 .82 .99 .53 23 5 5 mojuoides 2.4 1.68 .82 .97 .49 28 5 5 • Proportion of third costal segment with heavy bristles. Wasserman: The F asciola Subgroup Q~ 6 7 FIGs. 1-9. Spermathecae of the species in the fasciola subgroup: Fig. 1. D. moiuoides. Fig. 2. D. mo;u. Fig. 3. D. fascioloides. Fig. 4. D. fasciola. Fig. 5. D. pictura. Fig. 6. D. pictilis. Fig. 7. D. coroica. Fig. 8. D. paraguttata. Fig. 9. D. fulvalineata. Drosophila mojuoides, new species Externally similar to Drosophila mo;u Dobzhansky and Pavan, 1943, except for being generally lighter colored and with less distinct pattern. Other distin­guishing characters are as follows: trait D. mo;u D. mo;uoides genital arch: toe elongated and pointed (Figure 20) not so (Figure 25) penis: Figure 18 Figure 12 metaphase: Figure 36 Figure 35 distribution: Panama to Brazil Trinidad Other characters of D. mo;uoides are shown in Figure 1 and Table 2. The salivary gland chromosomes are described below. Type locality, Arima Valley, Trinidad (H 107.7), collected by Dr. W. B. Heed. This species has been previously designated as Drosophila species M (Wasser­man, 1960). Drosophila pictura, new species Posterior orbitals do not arise from a dark spot; one oral bristle; fore coxae and femora black. Mesonotum pollinose yellowish gray with the following light brown pattern: two longitudinal stripes between the second and fourth acrostichal rows extending from the anterior margin to just in front of the anterior dorso­ The University of Texas Publication FrGs. 10-18. Penes of the species in the fasciola subgroup: Fig. 10. D. fulvalineata. Fig. 11. D. fasciola. Fig. 12. D. mojuoides. Fig. 13. D. pictura. Fig. 14. D. coroica. Fig. 15. D. fascioloides. Fig. 16. D. paraguttala. Fig. 17. D. pictilis. Fig. 18. D. moju. centrals; two stripes starting at the level of the anterior dorsocentrals and broad­ening posteriorly to include the posterior dorsocentrals, but leaving a light mid­line; an oblong spot including and lateral to the anterior dorsocentrals. Second to fifth abdominal tergites each with an interrupted black wide band which stops at the angle of the tergite and with forward extensions at the interruption and at the angle of the tergite; lateral margins of the tergite with a wide black border Wasserman: The F asciola Subgroup FIC;i. 19-27. Male external genitalia of species in the fasciola subgroup: Fig. 19. D. fulvalineata. Fig. 1:;0. D. rrwju. Fig. 21. D. fascioloides. Fig. 22. D. pictilis. Fig. 23. D. fasciola. Fig. 24. D. paraguttata. Fig. 25. D. mojuoides. Fig. 26. D. pictura. Fig. 27. D. coroica. The University of Texas Publication v v ,II; '~ '~ -'11' ~r' ;I'' 28. fu/volineoto 29. fulvolineofo 30. pictilis v \.~ \~ '~ ?ji~ ~ii~. ~1'' 31. foscio/o 32. fosciola 33. picfura ,, \~ ~··~ ~··~ '"' II It 'It' 34. coroico 35. mojuoides 36. moju paraguttata ~ >.. J .. ( -J L 1~ II A 37. foscioloides 38. foscioloides 39. foscioloides FIGs. 28-39. M etaphase Karyotypes of the species in the fasciola subgroup. Wasserman: The F asciola Subgroup which may connect to the band at the posterior margin. Wings smoky, other characters as in Table 2. Figures 5, 13, 26 and 33 show the spermatheca, penis, genital arch and claspers, and the metaphase karyotype. Salivary gland chromosomes are described in this paper. Type locality, Port of Spain, Trinidad (H 109.28), collected by Dr. W. B. Heed. Also collected at Montero, Bolivia, by the author. This species had been previously designated as Drosophila species K (Wasserman, 1960). Drosophila pictilis, new species Externally similar to D. pictura, except abdominal bands of pictilis are narrow, less than half the width of the abdominal tergites, compared to pictura where these bands are wider than half this width. Other differences are seen in the penis (Figure 17), external male genitalia (Figure 22), and metaphase karyotype (Figure 30). Spermatheca as shown in Figure 6. Salivary· gland chromosomes are described below. Table 2 gives other characters. · · Type locality, Lake Pichichuela, El Salvador (H 27.2), collected by Dr. W. B. Heed. Also found in Costa Rica, unknown locality. This species had been pre­viously designated as Drosophila species J (Wasserman, 1960). Drosophila coroica, new species Externally similar to D. pictura, except having a lanceolate spot instead of an oblong spot, lateral to the anterior dorsocentrals. Figures 7, 14, 27 and 34 show the spermatheca, penis, external male genitalia and metaphase chromosomes respectively. Salivary gland chromosomes are described in this paper. Table 2 gives other useful characters. Type locality, Coroico, Bolivia (H 346.43) collected by the author. This species had been previously designated as Drosophila species L (Wasserman, 1960). Drosophila fasciola Williston, 1896 D. fasciola can be readily identified by the shape of the carina, which is acutely narrowed compared to the usual flattened repleta type. There is polytypic variation in the color pattern of this species. Three strains of living material showed the following patterns: the Balboa, Panama (H 86.9) and the British Guiana (H 231.12) strains have red eyes and a silvery mesonotum with a fairly distinct pattern similar to that of D. pictura; whereas the St. Lucia, B.W.I., strain (H 112.16) has very dark eyes and a yellow mesonotum with a broad centrally located longitudinal band. A check of the pinned specimens present in The University of Texas collections yielded the following: Martinique (one specimen), like St. Lucia; St. Vincent (five specimens), yellow mesonotum with a Balboa-like pattern but a strong tendency toward a central band; Grenada (eight specimens), mesonotum more or less silvery, central band present but tending to break up into stripes anterior to the suture; Santa Tecla, El Salvador (six specimens) and San Salvador (one specimen) like Balboa strain. The eye color of the pinned specimens could not be accurately judged. The Uni.versity of Texas Publication Drosophila paraguttata Thompson, 195 7, and Drosophila fulvalineata Patterson and Wheeler, 1942 D. paraguttata (in Wheeler, 1957) and D. fulvalineata lack the typical repleta spotting and although recognized as members of the subgenus Drosophila, had not been placed in any species group. Hsu (1949) pointed out that the male genitalia of fulvalineata are similar to those of the repleta group. Cytological evidence presented in this paper demonstrates that these species are members of the fasciola subgroup, having lost the spotting on the mesonotum. The nine species of this subgroup can be identified by the use of the folowing key: ( 1) Fore coxae and femora light colored, mesonotum unspotted or essentially unspotted ------------·---------··--·-···---------------------·········--·-···-----·----··-----------··---------2 Fore coxae and femora dark brown or black ---------------------------------·------····----4 (2) Costal Index 3.8 to 4.0; dark fly----·-·······------------------·---·---·---·----D. fulvalineata Costal Index 2.4 to 2.6; light colored fly ------------·--···---------------·--·-··----------------3 (3) Mesonotum light brown; pair of yellow spots in front of suture, no spots at bases of bristles, Jamaica -------------------------------···-------------·--D. paraguttata Mesonotum yellow brown; no yellow spot present; some bristles and hairs arising from dark spots on mesonotum, Central and South America -·--·-----····---·--·---------··--···----·--·--·-------·--·-·····--------·------·--·-·------D. mo;u As mo;u, but lighter colored; Trinidad ----·-·-------------------------------D. mo;uoides ( 4) Posterior orbital arises from a dark spot; mesonotal pattern distinct or not.. 5 No dark spot at the base of the posterior orbital; distinct mesonotum pattern ···········-----·-·---·······---·---·--·····------------····--------------------------·-···-·----------··-6 (5) Carina acutely narrowed····------······--····---------------------------········-------·-D. fasciola Carina flattened --·-·-----------·-----------------·-------··--·-···---------------·-------D. fascioloides (6) Mesonotal pattern with a lanceolate spot outside of the anterior dorso­centrals -----·-·····---·--·-····-·--·····-···--·--·---···-·····----·-··----·--·----····---·-----·· D. coroica Mesonotal pattern with an oblong spot including and lateral to the anterior dorsocentrals ---·-··--····---------··--------·-----·-----·--·-------------------------·-----·--7 (7) Black bands on the abdominal tergites narrow, less than half the w,idth of the tergite --·--········--·-···-----·-·········----------·-·······--·-··----------·----·-----D:"pictilis Black bands on the abdominal tergites broad, wider than half the··~iath of the tergite ---··-··-·--·-··--·-·------------------·--·--···--···-----·······-----------···---D. ·pictura GENETICS All crosses were made using mass matings of ten pairs per vial. Every possible intraspecific cross between strains was attempted and yielded in all instances fully fertile offspring. The data from the interspecific matings are given in Table 3. A standard strain was chosen for each species and tested against every other strain in the laboratory. The figures in Table 3 refer to the total number of pairs tested; S indicates no offspring; figures following a colon refer to the number of F1 offspring produced. Crosses between mo;u and mo;uoides produced a fair number of male and female F 1 off spring, all of which proved to be sterile. D. fulvalineata and D. paraguttata were not tested against the other members of the subgroup. Wasserman~· The F asciola Subgroup T ABL E 3 Results of interspecific crosses fasciola coroica pictilis pictura fascioloides moju mojuoides Species 6 6 6 6 6 6 6 fasciola <;> x 30*S 40 s 40S 100S 70S 40 s coroica <;> 30 s x 20 s 10 s 70 s 50 s 20S pictilis <;> 40S 20 s x 40: 80 s 60 s 30 s 1 ~ ,1 <;> pictura <;> 40:2 ~ 10 s 40: x 70S 50 s 20 s 1 ~ ,1 <;> fascioloides <;> 110: 70 s 80 s 70 s x 100 s 60 s 1 ~ ,1 <;> moju<;> 60 s 40 s 40 s 40S 90 s x 140: 6 ~ , 10<;>t mojuoides <;> 40 s 20 s 30 s 20 s 60 s 150: x 75 H ,70 <;>t fulvalineata X paraguttata (and reciprocal ) , each, 20 S S No offsprinf.. • wnber re ers to total nmn ber of pairs tested. t Sterile F, offspring. METAPHASE KARYOTYPES All of the observed metaphase karyotypes including those reported in the literature (Wharton, 1943, Dobzhansky and Pavan, 1943, and Clayton and Wasserman, 195 7) are shown in Figures 28 to 39. The sex chromosomes are shown at the 6 o'clock position with the X on the left. SALIVARY GLAND CHROMOSOMES The gene sequence of D. repleta was chosen as the standard for the repleta group (Wasserman, 1954). Subsequent studies indicate that a sequence differing from that of repleta by six inversions, Xa, Xb, Xe, 2a, 2b, 3b, is most probably the primitive euchromatic type for the repleta group (Wasserman, 1960). The sequences found in the fasciola subgroup are as follows: X CHROMOSOME: This chromosome proved to be too difficult to analyze. The a, b, c sequences are readily recognizable in D. fasciola, D. coroica, D. pictura and D. pictilis. These species also are homozygous for another inversion. Subse­quent to this inversion a series of rearrangements have taken place leading to the more advanced forms. Several of these inversions have breaks near the hetero­chromatin at the proximal tip, resulting in changes in the general appearance of the euchromatin and making it impossible to arrive at a correct interpretation. Although the total number of inversions in this chromosome is probably not very large, being on the order of a half dozen, the determination of the changes that took place in this chromosome must be postponed until intermediate forms are found. 2 CHROMOSOME: The primitive sequence of the second chromosome for the repleta group differs from the repleta sequence by containing 2a and 2b. A total of eighteen inversions have occurred during the evolution of the fasciola sub­group. The basic type of this subgroup contains the a, b, and three new inversions, The University of Texas Publica1ion 0 2, ea, and 13, whose evolution is shown in Figure 40. A salivary chromosome map of the 2a, b, 0 2, ea, la basic sequence is shown in Figure 43, which was constructed by rearranging photostatic copies of the repleta map of Wharton (1942), and may be considered a salivary map of the second chromosome of D. fasciola, D. coroica, D. pictura and D. pictilis. D. fasciola and D. coroica are homozygous for this sequence. D. pictura from Trinidad is also homozygous for this gene order; whereas the strain from Bolivia (Table 1) is homozygous for na. D. pictilis is heterozygous for three inversions (Table 1), the Salvador strain having the standard chromosome and a chromo­some with n3, and ma; whereas the Costa Rica strain is homozygous for oa whose breakage points overlap those of na. These rearrangements are shown in Figure 43. The standard sequences of D. fulvalineata and D. fascioloides can be obtained from Figure 43. D. fulvalineata is homozygous for two overlapping inversions, c4 and the subsequent d4, whose breakage points are given in Figure 43. To arrive at the standard second chromosome of this species, first the c4 should be inverted, and then the d4• D. fascioloides is homozygous for p2 ; inverting this gene order in Figure 43 results in a salivary map of the standard. Table 1 lists the distribution of the heterozygous inversions found in fascioloides, each shown in Figure 43; the d3 sequence is widespread, while inversion ka which arose in a d3 chromosome occurs in Costa Rica and Panama. The p2 sequence not only occurs in fascioloides but also with the addition of another rearrangement, r2 , led to paragutta1a, moju and mojuoides (Figure 41). Figure 44 shows the salivary gland chromosome containing the sequences a, b, 02, e8, l3, p2 and r2, and is a useful base from which the chromosomes of the moju­like species can be obtained. This gene order is the standard of paraguttata which is polymorphic for ia, shown in Figure 44. The two strains of mojuoides are homozygous for both the n2 and ja inversions (Figure 44). All strains of D. moju are homozygous for n2 , which is considered an element of the standard gene order of this species. Table 1 lists the gene orders of the moju strains. Costa Rican, a b (A) 2a,b: A-AJ41a-AJJ;1a-c641g-c6~1g-H 2 0 (B) 2a,b,o2 A-AJap1a-a1JE6a-n1gf:6a-n1gf:6a-c1ef.Ja-Bl1k6a~H (C) 2a,b,o2,e3,13: i3 e3 FIG. 40. Evolution of the second chromosome from the primitive, a, b, sequence to the basic sequence, a, b, o2, e3, 13, of the fasciola subgroup. 2 2 r-~~~r~__, p A-AJaf la-Bltf:6a-E5,,P::2d-DJJD2--0Ja IE5a-E2~2--01gp6a-01gp6a-c241--E6~1b-AJaj:;1a-c2i1--H FIG. 41. Chromosome 2a, b, 02, ea, 13, p2, r2, which is an intermediate type leading to the more advanced chromosomes of D. moju and D. mojuoides. Wasserman: The F asciola Subgroup Panamanian and Colombian strains are homozygous for q2 which is heterozygous in the Bolivian strain. The stock from Brazil has three types of second chromo­somes: the standard; q 2; and a4 and b4 (Figure 44). 3 CHROMOSOME: The primitive sequence of the repleta group differs from the repleta sequence by having the 3b arrangement (Wasserman, 1960). This 3b chromosome is the basic type of the fasciola subgroup, D. fulvalineata being homozygous for it. All other species in this subgroup are also homozygous for the 3c sequence which is identical to.that found in the mulleri complex (Wasserman, 1962b). The breakage points of 3c are E4a and Glc. Figure 45 gives the salivary gland chromosome map of the 3b, c chromosome, which is the only one found in fasciola, pictura, pictilis and fascioloides . Changes from this chromosome are shown in Figure 45: coroica is homozygous for p; paraguttata is homozygous for q; mo;uoides and moiu are homozygous form, and strains of mo;u from Bolivia and Brazil are polymorphic for o. 4 CHROMOSOME: The repleta sequence is the standard for all of ·the mem­bers in this subgroup, there being no known rearrangements. 5 CHROMOSOME: The repleta sequence is the standard for the fasciola sub­group. Three inversions, h, i, and j, have been fixed in paraguttata, moiu and mo;uoides. The evolution of this chromosome is shown in Figure 42. 6 CHROMOSOME: No comparisons were attempted among the various dot elements. DISCUSSION The fasciola subgroup is morphologically intermediate between the mulleri and the melanopalpa subgroups. However, the presence in the fasciola subgroup of the 3c gene order, one of the six basic sequences of the mulleri complex, indi­cates that the fasciola subgroup arose from, or at least had a common ancestor with the mulleri complex. This ancestor probably inhabited the deserts of Mexico, Central and South America and consisted of semi-isolated populations. Six inver­ . sions, 2c, 2f, 2g, 2h, 3a, and 3c were each locally distributed, occurring in one or several contiguous populations. Divergence of these semi-isolated populations into full species resulted in a sharing of inversions among the derived forms. One of these populations must have been the precursor of the fasciola subgroup, since (A) repleta sequence: A-----~H (B) 5h: i (C) 5h,i: (D) 5h,i,j: FIG. 42. Evolution of the fifth chromosome in the fasciola subgroup. The University of Texas Publication 3c is common to both sections of the repleta group. Besides being heterozygous for 3c, the precursor gave rise to both D. fulvalineata which lacks 3c and to the other eight species which are homozygous for 3c; the population was, or soon became, homozygous for 202, 2e3 and 213, three inversions which are specific for the subgroup. Figure 46 shows the cytological evolution of the subgroup. D. fulvalineata is shown as being the most primitive cytologically. This is borne out morphologically: its spermathecae are more typical of the repleta group than are those of the others. D. fulvalineata, like the members of the mulleri complex, is a desert dweller, whereas the other species are mesophyllic forms, indicating perhaps that fulvalineata split off before adaptation to wet conditions was initiated. The karyotype of the strain from Patagonia, Arizona (Figure 28) was identical to that found in a strain from Glenwood, New Mexico (Wharton, 1943) but differed from that of Cliff, New Mexico (Clayton and Wasserman, 1957) shown in Figure 29. The Glenwood strain was reported to have six long arms and a dot in the salivaries, leading Wharton ( 1943) to con­clude that the metacentric is the result of a pericentric inversion. Since both the Patagonia strain which has the same karyotype and the slightly different Cliff strain show this metacentric pair but have five arms and a dot, we question Wharton's interpretation of the Glenwood strain. The situation is somewhat complicated by the discovery of another, undescribed, morphologically similar species which was also collected at Patagonia, Arizona, and which has six arms and a dot in the salivaries, but a different karyotype (Clayton and Wasserman, 1957). D. fasciola, which has been shown to have polytypic variation in its color · pattern, appears to lack variation in gene sequences. However, the Balboa and the St. Lucia strains have an acrocentric X chromosome (Figure 31) whereas the British Guiana strain has a J-shaped X (Figure 32). D. pictura and D. pictilis are considered two distinct species despite their allo­patric distributions because the strains show consistent morphological and karyo­typic differences (Figures 30 and 33) and are genetically isolated (Table 3). Table 1 lists the gene orders found in our strains: D. pictura has two gene orders, the standard second chromosome, a, b, 02, e3, 13, homozygous in Trinidad; and a newer sequence, n3, in Bolivia. The Salvador pictilis is polymorphic for two second chromosomes: the standard, identical to the Trinidad pictura chromo­some; and a chromosome containing two associated, but independent inversions, 2m3 and 2n3, the latter being identical to the n3 of the Bolivian pictura. Another sequence, 2o3, is found in Costa Rican pictilis. These two species, therefore, both having identical standard and n3 arrangements arose from a form which was polymorphic for these gene orders, the Salvador strain of pictilis having pre­served the polymorphism. In pictura, our two strains are each homozygous in gene order, but it is most probable that there exist populations geographically intermediate between Bolivia and Trinidad which are polymorphic. D. coroica, homozygous for a unique inversion, 3p, is known from a single strain in which no polymorphism was found. All of the remaining species arose from a single ancestor which was homozy­gous for an inversion in the second chromosome, 2p2 (Figure 46). The first branch from this ancestor led to D. fascioloides, a species which is homozygous for the Wasserman: The F asciola Subgroup p2 and for two centric fusions, between chromosomes 2 and 4 and chromosomes 3 and 5. Three different metaphases were obtained: three pairs of metacentrics and a pair of dots (Figure 39) found in Ecuador (H 327.11) and Brazil (H 336.9 and also reported by Dobzhansky and Pavan, 1943); two pairs of metacentrics, a pair of dots, and a pair of acrocentric sex chromosomes, the X having a visible short arm (Figure 38) found in Central America and Colombia (Table 1); and two pairs of metacentrics, one of which has unequal arms with a terminal second­ary constriction in the longer arm, a pair of dots, a metacentric Y, and an X with a visible short arm (Figure 37), occurring in Honduras (H 51.5). D. fascio­loides is also polymorphic for inversions: 2d3 is widespread, while 2k3 which arose in a d3 chromosome (Figure 43) occurs in Costa Rica and Panama (Table 1). The other branch of the 2p2 ancestor led to the mo;u-Iike species, paraguttata, mo;uoides and mo;u, through the fixation of 2r2 (Figure 44) and a marked modi­fication of the proximal third of the fifth chromosome, interpreted as resulting from three inversions, 5h, 5i, and 5j (Figure 42). These three species appear to be allopatric, paraguttata only found in Jamaica, mo;uoides in Trinidad, and mo;u occurring on the mainland from Costa Rica to Brazil. Besides the cytological similarities, the three species show likenesses both in their penes (Figures 12, 16 and 18) and in their tendency towards the replacement of mesonotal spotting by a yellowish coloration. D. paraguttata split off first (Figure 46), maintaining the primitive metaphase karyotype (Figure 34), and became homozygous for a new gene order, 3q (Figure 45) and heterozygous for 2i3 (Figure 44). The precursor of the other forms obtained a J-shaped Y chromosome (Figure 35) and became homozygous for 3m (Figure 45) and 2n2 (Figure 44); the latter rearrangement has breakage points which overlap the 2i3 of paraguttata. There­fore these two inversions cannot occur within the same chromosome, and if present along with the original standard in a population, form an overlapping triad, which according to Wallace ( 1953), may lead to difficulty because of recombination. D. mo;uoides and D. mo;u arose from this precursor (Figure 46), mo;uoides through the fixation of 2j3 (Figure 44) and moiu through further modification in the metaphase karyotype (Figure 36), and becoming polymor­phic for 2q2, 2a4, 2b4 (Figure 44) and 3o (Figure 45). The 2j3 of mo;uoides and the 2b4 of mo;u, although not overlapping, seem to have one breakage point in common. Polymorphism in D. moiu is limited to the lowlands of Brazil and Bolivia (Table 1) where one finds the primitive sequence along with the newer inver­sions. The eight strains from six localities in Colombia, Panama and Costa Rica (Table 1) are all homozygous for the standard third chromosomes and 2q2, one of the newer second chromosome gene orders. The species probably arose where it is now polymorphic, in the lowlands of South America, the closest region to its sister species, the Trinidad-inhabiting mo;uoides. Migration westward across the Andes and then northward appears to have been limited to individuals homozy­gous for the standard 3 and the 2q2 ; the other genotypes presumably were not adapted to the western environment. The homozygous strain from Villavicencio,. Colombia (Table 1) is on the eastern slope of the Andes and consists of a combi­nation of two localities, one occurring at approximately 1500 feet, and the other.. - (03) K3/ + N3/+ F''· r ~O' I+ ,,. ~ . ~\\illl~nlfi~{rjfj/~f-cllltir.m;1:atti~~i1b~1uJ.~1lt!U~J: ~ 'IWl{j 1. , 1 121 1314 ~11!1 I \ 3l4J tleJ1 l2i 3l4lsleJ 1l 2i3l4 1 sl A CD ' AB EF FGI CH J Fig. 43 . C4 p2 04 (C4) 04 ~ N2~13/+~ J3 ~ 1 0 . •.[,I ', 1I'I'/ lj ' I ' I ' I'I'I I 1. I ' I. ' I'I ' '1~9 / ~ I "~' I' (I,'.• Jr ~ I ' ,, ,, ~­ • • ", , , . . . .r1 . ....~1 , .rn . . .~ lli.i ... . ,., 1 .• •• \~,,, l • ., . •• i .... ·. . ,..,...r l . , .· .. i . , . , 1::1. ~ ~;l~Jii1Ul\\llftt]l) 11Wu ~ltL~: tn:i~rrmir :,~~~ft1ltd1rtll~l\WJ~\\n~~i1:lDDf.1l ;dlfftK ,\. tna UtVnJm,.q•.. · "! .Q. 1:· 1 I 2I I e I1clo 1l ~ J _J /c 1L/ 2 / 314 IsG IH 2 ~ ~ Fig . 44 A4/+ 84/+ ~ 0 21+ "ti ;::: :::-:11§" ii\ 1"·1.11 1-1: !'.~i I1 r.. '(NlLr111~~r!l~ii:.::'1J~1,, ":11r:1.1.1, :~ii '.J[:fJ1r"[,.[ 11 n~~ 1, 1Ifll(~i,'rJ:~:1m:fl!lf;ll111mf"~~~· ''.l~rr1nr. . a 3 \ll\l1ll .ll.;;.~•~•I .. fl ..L ~\ n.tt~1UdttiJ~l'll~\)-\\M.,fiJt.l ....:JH1Ulllilll;i:-llll.L\\'1.Ultic~:j . .:d IiI\IJ;I/u. ·:;·•\ i ' \• I\.. iii -~-!ii~"· '1 I 2 I 3 I4 15 I I I 2 I 3 I 4 'j 5 I I I 2 I 3 u5 I Ii 2 I I I I 2 I3 I 1; :~\. 1 ': I 2 I 3 I 41 '1".' I L J I ' A A B ' Blc c D E G H Fig. 45 . c_ ___. 0 p-~ Fws. 43-45. Salivary gland chromosomes of the fasciola subgroup. See text for explanation. Fig. 43. Basic second chromosome of the fasciola subgroup, a, b, o2, ea, l3. Fig. 44. Second chromosome carrying a, b, o2, es, is, p2, r2, the sequence that gave rise to D. moiu, D. moiuoides and D. paraguttata. Fig. 45. Third chromosome with band c. Wasserman: The F asciola Subgroup D. mojuoides ~ D. moju ............._ D. poroguffofo ~I I3q 2i3/t I 2n2 3m I_............ r 2q 2/t 2 b4/t 204/t 3o/+ D picfura ~I 2r2 5h 5i 5j D. foscioloides j 2n3/tj~ 2d3/t 2d3k3/d3 ---•I D pictilis 3 ~ / j 2n /t I 2-4F 3-5F 33 2n /t 2m/t "' ~ 2031+ o ~~D. foscio/o D. fulvat;neafa ~120' }.,~I ~~ra;ca 2c4 2d4 l~o l ~ · 3c ~ ~ i IXo Xb Xe 2o 2b 3b Fw. 46. Phylogenetic relationships in the fasciola subgroup. The standard arrangement of each species is the sum of all homozygous inversions leading out from the primitive. Heterozygous inversions which occur with the standard are indicated by the inversion sho·wn over a plus sign. 10 kilometers away, at 3000 feet. There is no obvious barrier, except for distance, between this homozygous population and the polymorphic strains to the east. More information is needed before the limits of these inversions can be deter­mined and understood. The distribution of the inversions does, however, mimic those of D. mercatorum (Wasserman, 1962a) where more strains are available. There, the information indicates that a change in altitude, presumably along with corresponding changes in ecology, acts as a barrier which prevents certain gene sequences from moving out of the Brazilian and Bolivian lowlands into the Andes. SUMMARY Nine species of the repleta group in the genus Drosophila, including four new forms, are combined into a new subgroup, the fasciola subgroup. These species differ from their nearest relatives, the mulleri complex of the mulleri subgroup, by at least three homozygous inversions. Phylogenetic relationships among the fasciola species are presented, based mainly on the cytological evolution, although morphological and genetic information is used where possible. A total of twenty-six autosomal inversions has been incorporated during the The University of Texas Publication history of the subgroup: sixteen have been fixed in the homozygous condition and are interspecific differences, while ten are heterozygous within species. Each species was found to be cytologically unique, either having different standard gene orders, or different metaphase karyotypes, or both. LITERATURE CITED Clayton, Frances E., and M. Wasserman. 1957. Chromosomal studies of several species of Dro­sophila. Univ. Texas Pub. 5721: 125-131. Dobzhansky, T., and C. Pavan. 1943. Studies on Brazilian species of Drosophila. Bol. Fae. Fil. Cien. Letras, S. Paulo, No. 36: 7-72. Hsu, T. C. 1949. The external genital apparatus of male Drosophilidae in relation to systematics. Univ. Texas Pub. 4920: 80-142. Patterson, J. T., and M. R. Wheeler. 1942. Description of new species of the subgenera Hirto­drosophila and Drosophila. Univ. Texas Pub. 4213: 67-109. Wallace, B. 1953. On coadaptation in Drosophila. Amer. Nat. 87: 343-358. Wasserman, M. 1954. Cytological studies of the repleta group. Univ. Texas Pub. 5422: 130-152. ----.. 1960. Cytological and phylogenetic relationships in the repleta group of the genus Drosophila. Proc. Nat. Acad. Sci. 46: 842-859~ ----. 1962a. Cytological studies of the repleta group of the genus Drosophila: III. The mercatorum subgroup. This bulletin. ----. 1962b. Cytological studies of the repleta group of the genus Drosophila: V. The mulleri subgroup. This bulletin. ----., and Florence D. Wilson. 1957. Further studies on the repleta group. Univ. Texas Pub. 5721 : 132-156. Wharton, Linda T. 1942. Analysis of the repleta group of Drosophila. Univ. Texas Pub. 4228: 23-52. ----. 1943. Analysis of the metaphase and salivary morphology within the genus Dro­sophila. Univ. Texas Pub. 4313: 282-319. Wheeler, M. R. 1957. Taxonomic and distributional studies of Nearctic and Neotropical Dro­sophilidae. Univ. Texas. Pub. 5721: 79-114. Williston, S. W. 1896. On the Diptera of St. Vincent. Trans. Ent. Soc. London 3: 253-446. ·VII. Notes on the Taxonomy, Morphology, and Distribution of the Saltans Group of Drosophi"la, with Descriptions of Four New Species 1 LUIZ EDMUNDO DE MAGALHAES INTRODUCTION The genus Drosophila includes several cases of sibling species-forms so similar that it is almost impossible to distinguish them on the basis of external mor­phology. Their specific identities, nevertheless, are perfectly well established. In these cases the general method for identification is cytological examination and test crosses. The morphology of the male genitalia has been applied in the identification of certain cases, such as melanogaster-simulans by Sturtevant (1921) and Salles (1947), the species of the annulimana group by Breuer and Pavan (1950) and of the willistoni group by Malogolowkin (1952), Spassky ( 195 7) and Wheeler and Magalhaes (this bulletin). In the saltans species group, Spassky (1957) studied two sibling species, and Magalhaes and Bjornberg ( 195 7) studied 13 species, including several siblings. Five nominal species, at that time, were not available to these authors: saltans, earlei, pilifacies, pulchella and cordata. Except for the latter, the descriptions were not detailed enough to distinguish them from other species, and these species have. not been reported in the collection records from Mexico and Central America although several trips into these areas have been made by the members of the genetics group of The University of Texas. These facts led us to believe that these species have been currently recognized under synonymous names rather than their original ones. In order to clarify this point the author has visited the U.S. National Museum collection (U.S.N.M.) in Washington, D.C., and examined the type series of some of these species. It was also possible to study some type specimens belonging to the American Museum of Natural History (A.M.N.H.), New York City, and some from the collection of the Dresden Museum, through the courtesy of Dr. R. Hertel. The results of the examination of these specimens are reported below. In addition we are including the known geographical distributions of the species of this group, and the descriptions of four new species which we found when we examined the material at The Uni­versity of Texas. The type specimens of the new species have been deposited in The University of Texas collection in Austin. Acknowledgments: We want to thank Dr. W. W. Wirth who made the U.S.N.M. specimens available for study; Mr. Geza Knipfer who helped to pre­pare the maps and made the final copies of most of the drawings; Mr. Nelson Buck for agreeing to include the descriptiqn of D. septentriosaltans in this paper; and Dr. Harry M. Miller, Jr., and the Rockefeller Foundation for arranging the fellowship and financing the visit to Washington. · 1 Fellow of the Rockefeller Foundation. Permanent address: Department of Biology, the University of Sao Paulo, Brazil. The University of Texas Publicat.ion We also want to thank Dr. M. R. Wheeler for helping to arrange the visit to the National Museum, assisting in the examination of the types, permitting us to use his collection records and the specimens in The University of Texas col­lection, and for assisting in the preparation of this manuscript; and to Dr. Wilson S. Stone of the Genetics Foundation for his generous hospitality during our research leave at the University, we wish to express our special gratitude._ THE SALTANS GROUP The saltans group was established by Sturtevant (1942) as a member of the subgenus Sophophora; according to his definition of the group it included only the dark species and later it was extended to include yellow forms by Magalhaes (1956). The group was divided by Sturtevant (1942) into two subgroups: "(a) grayish marking on the mesonotum; minute hairs present below carina; (b) no gray mesonotal pattern; no minute hairs below carina." This subdivision later lost its significance because species without a gray mesonotal pattern but with minute hairs below the carina were found by Pavan and Magalhaes (1950) and by Magalhaes (1956). Magalhaes and Bjornberg (1957) divided the group into 5 subgroups, basing them on the characteristics of the male genitalia and external morphology of the body. Presently we are keeping the same subdivisions but are changing the name of the subgroups; we are-using the name of one of the species instead of letters to name each subgroup. Nineteen species of this group are now known, four of them previously un­described. For quick reference, each species has been assigned a number which is used in the key and in the figures illustrating the key characters, as well as in the following discussion. The species are as follows: 1. saltans Sturtevant 9. milleri Magalhaes n. sp. = sellata Sturtevant 10. rectangularis Sturtevant 2. lusaltans Magalhaes n. sp. 11. parasaltans Magalhaes 3. prosaltans Duda 12. subsaltans Magalhaes 4. nigrosaltans Magalhaes n. sp. 13. pulchella Sturtevant 5. septentriosaltans Magalhaes and 14. elliptica Sturtevant Buckn.sp. 15. emarginata Sturtevant 6. austrosaltans Spassky 16. neoelliptica Pavan and Malgalhaes 7. pseudosaltans Magalhaes 17. neosaltans Pavan and Magalhaes 8. sturtevanti Duda 18. cordata Sturtevant = earlei Sturtevant 19. neocordata Magalhaes = pilifacies Malloch = biopaca Sturtevant These species may be arranged in subgroups as follows: 1. Subgroup saltans. Dark species; pattern on the mesonotum and hairs below carina present; undermargin of the male genital arch with a horn-like process; forcipes small, semi-elliptical, with 20-25 teeth; penis with a pair of pincers. Species: saltans, lusaltans, prosaltans, nigrosaltans, septentriosaltans, austro­saltans, and pseudosaltans. 2. Subgroup sturtevanti. Species with pattern on the mesonotum and hairs below carina; penis without pincers. Species: sturtevanti, miUeri and rectangu­laris. 3. Subgroup parasaltans. Yellow species, hairs below carina present; genital arch with one or two horn-like processes on the anterior region of the under~ margin. Species: parasaltans, subsaltans and pulchella. 4. Subgroup elliptica: Dark species (excepting D. emarginata from Peru); large cylindrical penis, with a pair of lateral pincers. Species: elliptica, emargin­ata, neoelliptica and neosaltans. 5. Subgroup cordata. Dark species without mesonotal pattern; small hairs , below carina absent; penis with a cylindrical process, paired, fused to the apex i of the apodeme. Species: cordata and neocordata. · GEOGRAPHICAL DISTRIBUTION We present here, for each species, a list of the places where it has been collected. Our data come from the identification of pinned flies and of living stocks in The University of Texas collection; the records of collecting trips of members of the Genetics Foundation of The University of Texas from 1942 to 1959; the data present in Patterson and Wagner ( 1943) for the distribution in Mexico and Central America, Dobzhansky and Pavan (1943; 1950) and Pavan (1950; 1952) for South America, as well as the information in our own species descriptions. Almost all points of collection are shown on Plates I and II. Flies from the saltans group are not very frequently found on baits. Usually when one species is present, it is generally a small number, sometimes only one fly. This does not necessarily mean, however, that the population size is very small. The number of flies on this basis depends on the attractiveness of the baits; this probably differs from species to species and it is impossible to get any idea about the true population size. Nevertheless the data show quite well the geo­graphical distribution of these species. They are exclusively found in the Nearctic and Neotropical regions, preponderantly in the latter one. Patterson and Stone (1951) mention that elliptica is present only in the Neartic region and rectangu­laris in both. We are not sure about the identification of rectangularis from the Nearctic region and we prefer to consider this form as belonging only to the Neotropical region. From our data saltans and sturtevanti occur in both regions. All the others have been found only in the Neotropical region. Three species are found only in the Caribbean islands: lusaltans, milleri, and pulchella, suggesting that speciation has occurred ori these islands. The others are found on the continent or on both. The most widespread species in the saltans group is D. sturtevanti, occupying almost all the range of distribution of the group. There are no special correlations among the subgroups and their geographic distribution, on the contrary, very closely related forms sometimes show a big gap in their distribution revealing either the incompleteness of the present data or, perhaps, different geographic distributions in the past. 1. Drosophila saltans Sturtevant D. saltans Sturtevant 1916. Ann. Ent. Soc. Amer. 9: 328. = D. sellata Sturtevant 1942. Univ. Texas Pub. 4213: 39. -----------------------------------------------------------------------------------------------di-;-----------------· --·--­ DROSOPHILA : 1 . saltans 2 . lusaltans 3 . prosaltans 4. nigr osaltans S . septentriosaltans 6 . a u atrosaltan s 7 . pseudosaltans GEOGRAPHIC DISTRIBUTION OF THE SALTANS GROUP -----"­ 7 Plate 1. GEOGRAPHIC DISTRIBUTION OF THE SALTANS GROUP 19 12 DROSOPHILA: 8 . sturtevanti 9 . milleri 10. rectangularis 11. parasaltans 12. subsaltans 13· pulchella 14. elliptica IS. emarginata 16. neoelliptica 17. neosaltans 18. cordata 19 .neocordata Plate 2 . The University of Texas Publication Not= D. prosaltans Duda 1927. Arch. Naturg. 91A12 (1925): 164. Dobzhansky and Pavan ( 1943) considered sellata to be a synonym of pro­saltans, and they have been followed in this by most later authors. The male genitalia of sellata were described by Spassky (1957) who considered it to be a race of prosaltans. A more detailed study of the genitalia by Magalhaes and Bjornberg (1957) revealed some distinct differences between the two, and the authors concluded that sellata and prosaltans were distinct species. Hybridization studies now being carried out by the present author support this conclusion. Thus D. prosaltans remains as a valid species and is distinct from the species named sellata. Four specimens from the type series of D. saltans have been examined: (1) Type male, No. 24129, A.M.N.H., labelled "Stock from Guantanamo, Cuba, V 1914" (i.e., May 1914). (2) Paratype female, No. 50010, U.S.N.M., labelled "bred pineapple, Guantanamo, Cuba, 1.13.14" (i.e., January 13, 1914). (3) Paratype female, A.M.N.H., with the same labels as the holotype male. (4) "Metatype" female, A.M.N.H., Panama, R.P., February-March, 1915. Two of these specimens, paratype No. 50010 and the metatype do not agree with the type of saltans, but they do match D. sturtevanti very well. It is to be noted that the "metatype" came from an entirely different locality, and that the paratype did not come from "stock" but was "bred from pineapple," and also bears a different date. The holotype and the other paratype, both from "stock", are clearly not sturtevanti but represent a different species which we identify as being the same as sellata on the basis of the male genitalia. In his original description of saltans, Sturtevant (1916, op. cit.) did not mention the characteristic median spot of the mesonotum, but in looking at the type one can see that the mesonotum is folded inward precisely where the spot should be, thus making it very difficult to see. No doubt exists that sellata and saltans are the same species, and different from prosaltans. The name saltans has priority, having been published long before the name sellata. Geographical distribution: MEXICO: Hidalgo and Guemes in Tamaulipas; Huichihuayan, Matlapa, and Tamazunchale in San Luis Potosi; Jalapa, Cordoba, Puente Nacional, Veracruz, and Minatitlan in Veracruz; Zumpango, Chilpan­cingo, and Zopilote Canyon in Guerrero. GUATEMALA: Antigua and Guate­mala City. CUBA: Guantanamo and Santiago de Cuba. EL SALVADOR: San Salvador. COST A RICA: San Jose. 2. Drosophila lusaltans Magalhaes, new species External characters of imagines. ~, ~. Arista with 10 branches. Antenna blackish brown. Front velvety reddish brown, the orbits and ocellar triangle black. Anterior orbital bristle nearly as long as the posterior one, middle orbital about 1/ 3 length anterior. Two promi­nent oral bristles, the second nearly as long as the first. Face blackish gray; carina narrow, black at its base and with minute hairs below. Cheeks almost black; their greatest width about 1/ 7 greatest diameter of the eyes, the latter dark red with black pile. Acrostichal hairs in 6 regular rows; no prescutellars. Anterior scutellars diver­ . gent. Mesonotum pollinose dark grayish brown with blackish longitudinal stripes as follows: a short one in the middle of the mesonotum; a long pair just within the dorsocentral rows, and another just outside the dorsocentral rows broken in three spots. Scutellum with the same color as the mesonotum, and lighter lateral edges. Pleura black; anterior sternopleural 1/ 5 length posterior. Legs brown, coxae and femora black; apical bristle on the second tibiae, preapical on all three. Halteres whitish. Abdomen black, shining. Tergites 2-5 with a yellow band on the anterior margin of each, the bands interrupted in the middle and failing to reach the lateral edges of the tergite. Sixth tergite of the female with the same yellow anterior band and with a median rough area. Wings with two prominent bristles at the apex of the first costal section; third · costal section with heavy bristles on its basal 2/ 5. Wing vein indices as follows: Costal index 2.0, 2, 2.0, ~ ; 4th vein index 2.1, 2, 2.1, ~ ; 5x index 1.6, 2, 2.5, ~. Body length, 2, 3.0 mm, ~, 2 mm.; wing length, 2, 2 mm., ~, 1.5 mm. Other characteristics: Two anterior and two posterior branches of the Mal­pighian tubes with free ends. Testes yellow, with 3Yz inner and 3Yz outer coils. Ventral receptacle M-shaped as in other members of the saltans group. Sperma­theca (Figure 1) nearly spherical, without apical invagination; a collar at base. The larvae skip. Puparium brown, the anterior spiracle with 8 branches. Chromosomes: The metaphase plate shows two pairs of V's and one pair of rods . . Male genitalia: Genital arch as in prosaltans; hypandrium as in Figure 2, and penis as in Figure 3. Relationship: Belong to the saltans subgroup of the saltans group. Geographical distribution: HAITI: Petionville. Collected by Drs. W. B. Heed and H. L. Carson; July, 1959. 3. Drosophila prosaltans Duda D. prosaltans Duda 1927. Arch. Naturg. 91A12 (1925): 164. Geographical distribution: COSTA RICA: San Isidro del General and Tur­rialba. PANAMA: Balboa. TRINIDAD: Sangre Grande. COLOMBIA: Bucara­manga. BRAZIL: Ferreira Gomes in T. F. do Amapa; Marajo Island, Fordlandia, and Belem in Para; Carolina and lmperatriz in Maranhao; Salvador in Bahia; Palma in Goias; Angra dos Reis and Rio de Janeiro City in Rio de Janeiro; Piras-· sununga, Mogi das Cruzes, Cantareira, Bertioga, ltanhaen, and Sao Sebastiao in Sao Paulo; El Dorado in R. G. do Sul. PARAGUAY: Hohenau (Type locality). The type female, loaned to us by the Dresden Museum, is somewhat paler, especially the legs and the mesonotal pattern, than described by Duda; however it may have lost some color since he saw the specimen about thirty-six years ago. The velvety tergital area is well preserved, and, except for the paleness, the type seems to represent the same species now being considered as prosaltans. 4. Drosophila nigrosaltans Magalhaes, new species External characters of imagines. . Arista with 8-9 branches. Antenna blackish brown, third segment black. Front The University of Texas Publication velvety reddish brown, the orbits and the ocellar triangle black with silvery pollinosity. Anterior orbital bristle nearly as long as posterior one; middle orbital about 1/2 length anterior. Three prominent oral bristles, the third about 2/3 length of the other two. Face dark gray, the carina narrow and short, with some long hairs below. Cheeks almost black, their greatest width about 1/5 greatest diameter of the eyes. Proboscis blackish. Eyes dark red with short black pile. Acrostichal hairs in 6 regular rows; no prescutellars. Anterior scutellars slightly divergent. Mesonotum pollinose, dark grayish brown with blackish longitudinal stripes as follows: a short one in the middle of the mesonotum; a long pair just inside the dorsocentral rows, and another just outside the dorso­central rows and interrupted at the transverse suture. Scutellum blackish brown, with lighter brown lateral edges. Pleura black. Legs black, the tibiae and tarsi lighter brown. Apical bristle on the second tibia, preapicals on all three. Halteres whitish. . Abdomen shining black; tergites 2-5 with silvery pollinose halfmoon shaped areas on each side of the median line near the anterior margin; sixth tergite all black, that of the female with a large rectangular, rough, median mark. Wings brownish; apex of the first costal section with two bristles; third costal section with heavy bristles on its basal 1/3. Wing indices as follows: Costal index 2.1,~, 2.0,~; 4th vein index 2.2,~, 2.0,~; 5x index L7,~, 2.4,~. Body length, ~, 2.8 mm., ~, 2.0 mm.; wing length, ~, 2.0 mm., ~, 1.8 mm. Other characteristics: The two anterior and two posterior branches of the Malpighian tubes with free ends. Testes yellow, with 3¥2 inner and 3¥2 outer coils. Ventral receptacle M-shaped, as in other members of the group. Sperma­theca (Figure 4) nearly spherical and with a strong apical wrinkled invagination. Egg with two filaments, expanded distally. The larvae skip. Puparium reddish brown, the anterior spiracle with about 7-8 branches. The chromosomes consist of two pairs of V's and one pair of rods. Male genitalia: Genital arch as in prosaltans; hypandrium as in Figure 5 and penis as in Figure 6. Relationship: Belongs to the saltans subgroup. Geographical distribution: The species has been collected at Boquete, PANAMA (Type locality) by Drs. M. Wasserman and W. B. Heed in 1958. Also found in Turrialba, COST A RICA. 5. Drosophila septentriosaltans Magalhaes and Buck, new species Arista with 9-10 branches, usually 9. Antenna grayish brown, third segment darkened. Front velvety brown. Middle orbital 1/3 other two; two prominent and equally long oral bristles; face darkened below carina, the latter narrow and with minute hairs below. Cheeks gray, their greatest width about 1/6 greatest diameter of eye, the latter red with short, black pile. Acrostichal hairs in 6-8 rows; no prescutellars. Anterior scutellars parallel. Mesonotum brownish gray, pollinose, with about the same darkened pattern as described for the two former species. Scutellum dark brown, pollinose, with black spots around the bristle bases. Pleura black; anterior stenopleural 1/2 length posterior. Legs . brown, the coxae and femora more blackish; apical bristles on second tibiae, preapicals on all three. Halteres whitish. Abdomen as described for the two former species, with tergites 2-5 black, shining, each with a yellow band on the anterior margin, the bands interrupted in the middle and failing to reach the lateral edges of the tergites. Female sixth tergite with a median rough area. Wings clear; two prominent bristles at apex of first costal section; third costal section with heavy bristles on its basal 2/5. Wing vein indices as follows: Costal · index 1.8, ~, 1.6, 5 ; 4th vein index 2.2, ~, 2.3, 5 ; 5x index 1.9, ~, 1.8, 5. Body length, 5, 1.8 mm., ~, 2.4 mm.; wing length, 5, 2.0 mm., ~, 1.8 mm. Other characteristics: Two anterior and two posterior branches of the Mal­pighian tubes with free ends. Testes yellow, with 3Yz inner and 3Yz outer coils. Ventral receptacle M-shaped, as in other members of the saltans group. Sperma­theca (Figure 7) nearly spherical, with a lateral invagination; the invagination is curved. Egg with two filaments. The larvae skip. Puparium brown, the anterior spiracle with about 8 branches. The chromosomes consist of two pairs of V's and one pair of rods. Male genitalia: Genital arch as in prosaltans; hypandrium as in Figure 8 and penis as in Figure 9. Relationship: Belongs to the saltans subgroup. Geographical distribution: Collected at Sevilla, north of Aracataca, COLOM­BIA, in December, 1955, by Dr. W. B. Heed. 6. Drosophila austrosaltans Spassky D. austrosaltansSpassky 1957. Univ. Texas Pub. 5721: 57. Geographical distribution: Pirassununga, Sao Paulo (Type locality) ; Carolina and Imperatriz, Maranhao, BRAZIL. 7. Drosophila pseudosaltans Magalhaes D. pseudosaltans Magalhaes 1956. Rev. Brasil. Biol. 16: 273. Geographical distribution: Cantareira, Sao Paulo, BRAZIL. 8. Drosophila sturtevanti Duda D. sturtevanti Duda 1927. Arch. Naturg. 91A12 (1925): 167. = D. earlei Sturtevant 1916. Ann. Ent. Soc. Amer. 9: 329. = D. pilifacies Malloch 1926. Proc. U.S. Nat. Mus. 68 (Art. 21): 29. = D. biopaca Sturtevant 1942. Univ. Texas Pub. 4213: 37. Although the name sturtevanti is a junior ,synonym of both earlei and pili­facies, the best interests of science are served by continuing the use of the name sturtevanti since this is the only name which has been used for this species in numerous articles in the fields of taxonomy, geographic distribution, genetics, and evolutionary cytology. An examination of the type female of D. pilifacies Malloch (Type No. 28464, U.S.N.M., San Mateo, Costa Rica) shows clearly the pair of lateral opaque areas on the sixth tergite which are quite characteristic for this species. In all other characteristics also, this specimen agrees with the species which has been known as sturtevanti for many years. The University of Texas Publication D. biopaca Sturtevant was placed as a synonym of sturtevanti by Dobzhansky and Pavan ( 1943). The type has not been re-examined since there is no reason to doubt the synonymy. Three specimens from the type series of D. earlei have been examined: (1) Type male, No. 24146, A.M.N.H., Herradura, Cuba, January 28, 1915. (2) Para­type female, No. 49999, U.S.N.M., Herradura, Cuba. (3) Metatype male, A.M.N.H., Panama, R.P., February-March, 1915. All of the characteristics of the holotype male agree with the well-known D. sturtevanti. Unfortunately, a complete examination of the male genitalia has not been possible, but the visible structures, especially the large, pale forceps and the large pregenital plate, agree very well with the species called sturtevanti. The paratype female is clearly the same as sturtevanti. The male "metatype", coming from an entirely different locality than the holotype, is, however, the same species. We have also been able to examine the type female of sturtevanti, from Bolivia (from the Dresden Museum), and we have no doubt as to the identity of this species. Thus only one species seems to be involved in the complex of names: earlei, pilifacies, biopaca and sturtevanti. The redescription of D. sturtevanti by Dobzhansky and Pavan ( 1943) is the most complete description of the species; the male genitalia was described by Magalhaes and Bjornberg (1957). Geographical distribution: MEXICO: Saltillo in Coahuila; Hidalgo, Guemes, Llera, C. Mantes, and Tampico in Tamaulipas; Matlapa, Huichihuayan, and Tamazunchale in San Luis Potosi; Jalapa, Puente National, Veracruz, and S. A. Tuxtla in Veracruz; Acapulco and Zumpango in Guerrero. HONDURAS: La Lima. EL SALVADOR: Lago Coatepec, Puerto de La Laguna, La Palma, San Salvador. NICARAGUA: El Recreo. COSTA RICA: San Jose, Turrialba, San Mateo, La Lola and San Isidro del General. PANAMA: Barro Colorado, Canal Zone, and Balboa. CUBA. PUERTO RICO: El Yunque and Mayaguez. JAMAICA: Kingston and Ocho Rios. BARBADOS; ST. LUCIA; ST VIN­CENT; GRENADA; MARTINIQUE; HAITI; GUADELUPE. COLOMBIA: Medellin, Sevilla, Palmira, Villavicencio, and Bucaramanga. VENEZUELA: Merida, Maracay, Rancho Grande and Cumbre de Choroni Aragua. BRITISH GUIANA: Georgetown. BRAZIL: Mucajai in T. F. Rio Branco; Fer­reira Gomes in T. F. do Amapa; Icana, Uaupes, Moura, and Manaus in Amazonas; Marajo I., Fordlandia, and Belem in Para; Cruzeiro do Sul and Pal­mares in T. F. Acre; Carolina and Imperatriz in Maranhao; Anapoliz in Goias; Salitre, Catuni, Salvador and Ilheus in Bahia; Angra dos Reis and Rio de Janeiro City in Rio de Janeiro; Mucamba, Prata, Pirassununga, Cantareira, Agua Funda, Mogi das Cruzes, ltanhaen, Bertioga, Vila Atlantica and S. Sebastiao in Sao Paulo. BOLIVIA: Mapiri (type locality). 9. Drosophila milleri Magalhaes, new species External characters of imagines. Male. Arista with 9-11 branches. Antenna brownish gray. Front brownish gray, darkened on the posterior margin. Orbits with silvery pollinosity and with dark spot on the base of the third orbital bristle. Ocellar triangle black with silvery pollinosity at the angles. The second orbital 1/2 other two. Two oral bristles of the same length. Face and carina light brown; few hairs below carina. Cheeks darker than the face, their greatest width about 1/ 6 the greatest diameter of the eyes. Eyes wine red. Acrostichal hairs in 6 regular rows. No prescutellars. Anterior scutellars diver­gent. Mesonotum tan with almost black longitudinal stripes as follows: one broad pair inside of the dorsocentral rows and fused on the middle like an "H". Another stripe outside of the dorsocentral rows and interrupted at suture, with the upper part large and surrounding the humeri. Scutellum blackish brown with lateral edge tan. Pleura tan with two longitudinal blackish brown stripes, the upper one covering the propleura and part of the meta-and hypopleura, the second one on the upper part of the sternopleura. Anterior sternopleural bristle about 1/ 2 the length of the posterior. Legs yellowish tan. Halteres whitish. Abdomen yellowish. Tergites 2--5 with greyish brown basal bands expanded in the middle and laterally, being lighter in this lateral expansion. Sixth yellow­ish and without band. Female sixth tergite with a pair of lateral opaque areas. Wings clear; two prominent bristles at the apex of the first costal section; third costal section with heavy bristles on its basal 2/ 5. Wing vein indices as follows: Costal index 2.1, 'i?, 1.9, ~; 4th vein index 2.1, 'i?, 1.9, ~; 5x index 1.6, 'i?, 2.0, ~. Body length, 'i?, 3.2 mm., ~, 2.5 mm.; wing length, 'i?, 2.5 mm., ~, 2.0 mm. Other characteristics: Two anterior and two posterior branches of the Mal­pighian tubes with free ends. Testes whitish or pale yellow; 5-6 outer coils. Ventral receptacle M-shaped like in other species of the saltans group. Sperma­theca (Figure 10) nearly spherical, without apical invagination; collar at base. The larvae skip. Puparium brown, the anterior spiracle with 9 branches. Male genitalia very similar to sturtevanti; small differences in the penis (Figure 11). Relationship: Belong to the sturtevanti subgroup of the saltans group. Geographical distribution: EI Yunque, PUERTO RICO; collector Dr. W . B. Heed, January, 1956. 10. Drosophila rectangularis Sturtevant D. rectangularis, Sturtevant 1942. Univ. Texas Pub. 4213: 38. Geographical distribution: Orizaba, Veracruz (type locality); Tixtla, Guerrero, MEXICO. 11. Drosophila parasaltans Magalhaes D. parasaltans Magalhaes 1956. Rev. Brasil. Biol. 16: 276. Geographical distribution: BRAZIL: U aupes in Amazonas (type locality). COLOMBIA: Villavicencio. 12. Drosophila subsaltans Magalhaes D. subsaltans Magalhaes 1956. Rev. Brasil. Biol. 16: 277. Geographical distribution: BRAZIL: Belem~ Para. 13. Drosophila pulchella Sturtevant D. pulchella Sturtevant 1916. Ann. Ent. Soc. Amer. 9: 327, nom. nov. for bellula Williston 1896, not Bergroth 1894. The University of Texas Publication This little known species was placed in the saltans group by Wheeler (1957). Sturtevant ( 1921), who had examined the cotypes of bellula Williston from the island of St. Vincent, also identified as this species a single female from Mont­serrat, Trinidad. This female (U.S.N.M. collection) has been examined; it clearly represents D. sturtevanti and does not agree with a cotype of pulchella (A.M. N.H. collection), now on loan at The University of Texas. Geographical distribution: St. Vincent, B.W.I. 14. Drosophila elliptica Sturtevant D. elliptica Sturtevant 1942. Univ. Texas Pub. 4213: 35. Geographical distribution: MEXICO: Pachuca (type locality) and Chapul­huacan in Hidalgo. 15. Drosophila emarginata Sturtevant D. emarginata Sturtevant 1942. Univ. Texas Pub. 4213: 36. Geographical distribution: MEXICO: Jalapa, Huatusco, Cordoba, San Andres Tuxtla, San Juan, Cotopec and Xico in Veracruz; Tamazunchale in San Luis Potosi. GUATEMALA: Quirigua (type locality) and Guatemala City. EL SAL­VADOR: La Palma. COSTA RICA: San Jose, La Lola, Turrialba, and San Isidro del General. PANAMA: Boquete and Cerro la Campana. COLOMBIA: Medellin, El Recuerdo, Villavicencio, and Bucaramanga. VENEZUELA: Merida. ECUADOR: Santo Domingo de los Colorados. PERU: Urubamba and Tingo Maria (yellow form). 16. Drosophila neoelliptica Pavan and Magalhaes D. neoelliptica Pavan and Magalhaes 1950. Bol. Fae. Fil. Cien. Letr. Univ. S. P. 86: 13. Geographical distribution: BRAZIL: Anapolis in Goias; Jaguariaiva in Parana; Angra dos Reis in Rio de Janeiro; Agua Funda in Sao Paulo. 17. Drosophila neosaltans Pavan and Magalhaes D. neosaltans Pavan and Magalhaes 1950. Bol. Fae. Fil. Cien. Letr. Univ. S. P. 86: 16. Geographical distribution: BRAZIL: Mogi das Cruzes and Cantareira in Sao Paulo. 18. Drosophila cordata Sturtevant D. cordata Sturtevant 1942. Univ. Texas Pub. 4213: 34. Geographical distribution: GUATEMALA: Quirigua (type locality). 19. Drosophila neocordata Magalhaes D. neocordata Magalhaes 1956. Rev. Brasil. Biol. 16: 275. Geographical distribution: MEXICO: Tixtla in Guerrero. BRAZIL: Montes Claros (type locality) in Minas Gerais. NOTES ON COMPARATIVE MORPHOLOGY Wheeler (1960) first described the v~stigial remnants of the true first sternite in Drosophila species and the sensilla of these and other sternites. The sensilla of the first sternite and the sensilla of the seventh sternite of males have been observed in all of the species of the saltans group which were available for this study (Table 2). Vestigial plates of the first sternite are clearly present in some species, highly reduced in others, and seem to be completely absent in the rest. From our study it seems probable that in some species the presence or absence of vestigial first sternite plates may be a good specific character, but in others con­siderable variation occurs. Table 1 illustrates this for D. sturtevanti. The per- TABLF. 1 Presence of vestigial plates of first sternite in strains of Drosophila sturtevanti Females Males N umber Per cent with N umber Per cent \vi th examined plates Ratio exami_ned plates Ratio Mainland strains H 25.5 El Salvador 20 100.0 1:0 20 100.0 1:0 H 66.2 El Salvador 96 47.7 .9: 1 42 76.2 3.2: 1 H 50.3 Honduras 20 100.0 1:0 20 100.0 1 :0 H 73.3 Costa Rica 20 65 .0 1.8: 1 20 85.0· 5.7: 1 H1 58. 1 Costa Rica 38 7.9 .09:1 22 29.4 .42:1 H166.6 Costa Rica 50 4.0 .04: 1 20 11.1 .12: 1 H 79. 1 Panama 40 14.2 .17: 1 20 17.6 .21: 1 H101.2 Colombia 20 42.1 .73: 1 24 62.5 1.67: 1 H103.6 Colombia 50 62.0 1.6: 1 50 86.0 6.1: 1 2374.4 Brazil 40 100.0 1: 0 30 100.0 1:0 F1 of 23 7 4 'i' X 166 c! 25 16.0 1.9: 1 20 20.0 .25: 1 Island strains H352.13 Jamaica 21 90.4 9.4:1 14 100.0 1:0 H254.8 Puerto Rico 12 0 0: 1 21 9.5 .10: 1 H252.16 Guadeloupe 63 85.7 5.9: 1 37 86.0 6. 1: 1 H119.1 Barbados 0 0: 1 0 0: 1 H1 21.8 St. Lucia 100.0 1 :0 100.0 1:0 centage presence in differentgeographic strains ranged from 100 through various intermediate values to complete absence, although the latter was observed only in strains from Puerto Rico (females only) and from Barbados (both sexes). There does not seem to be any geographic gradient for this character among the strains tested. The stock of emarginata from Peru differs from the typical in several ways (Table 2) , especially in body color, subcarinal hairs, and presence of vestigial plates of the first sternite. Hybrids between the two types were intermediate in body color, possessed subcarinal hairs and lacked vestigial stemite plates. In this intraspecific cross, then, the presence of subcarinal hairs appeared to be dominant to their absence, while the absence of vestigial first sternite plates seemed to be dominant to their presence. This relationship was not so clear, however, in a cross of sturtevanti females from Brazil (in which both sexes have well-developed sternite plates; see Table 1) to males from Costa Rica (stock H166.6, in which The University of Texas Publication TABLE 2 Comparison of certain morphological characters in species of the saltans group Characters• Species 2 3 4 6 7 8 9 10 11 12 saltans prosaltans + + + + + + + + + + 0 0 0 0 153 155 59.4 50.3 38.9 32.5 30.8 31.8 20.1 20.7 austrosaltans + + + + + R R 168 57 33.9 30 17.8 pseudosaltans + + + + + R R 180 60 33.3 0 0 septentriosal tans + + + + + 0 0 150 37 24.6 20 13.3 lusaltans + + + + + R R 256 88 34.3 54 21.0 nigrosaltans + + + + + 0 0 182 69 37.9 40 21.9 sturtevanti + + + + + (Table 1) 157 231 14.61 milleri + + + + + 0 0 280 361 12.81 rectangularis + + + pulchella + I-r parasaltans + + + + 0 0 subsaltans + + + + 0 0 187 65 34.7 neosaltans + + elliptica + + + 190 42 22.1 neoelliptica + + + + + 192 32 16.6 24 12.5 emarginata + + + + + 190 32 16.8 emarginata (Peru) + + + + 0 0 cordata + neocordata + 0 0 186 .. Characters are as follows: 1, presence of mesonotal pattern; 2, presence of subcarinal hairs; 3, dark body color; 4, yellow body color; 5, presence of sensilla of first sternite; 6, presence of sensilla of seventh sternite of male; 7, presence of vesti~ial plates of first sternite, female (R=reduced but not wholly absent); 8, as 7, male; 9, total area of sixth tergite of fema e, in arbitrary units; 10, area of dull, velvety tergital marks; 11 , per cent of tergite occupied by dull tergital marks; 12, area of sixth tergite occupied by yellow coloration, female; 13, per cent of sixth tergite with yellow coloration. 1 Since there are two du11 areas on the sixth tergite, the figure may be doubled to obtain the total area occupied by such marks. TABLE 3 Variation in sixth tergite of females of D. saltans and prosaltans from various geographic strains. The morphological characters are the same as those with corresponding numbers in Table 2 9 10 11 12 13 salt ans strain 1 123 51 41.4 28 22.7 strain 2 177 65 36.7 46 25.9 strain 3 150 64 42.6 28 18.6 strain 4 157 57 36.3 25 15.9 strain 5 142 52 36.6 30 21.1 strain 6 162 61 37.6 33 20.3 strain 7 160 66 41.2 26 16.2 Average 153 59.4 38.9 30.8 20.1 prosaltans strain 127 39 30.7 30 23.6 strain 2 166 58 34.9 30 18.0 strain 4 149 49 32.8 33 22.1 strain 5 161 50 31.0 32 19.8 strain 6 162 50 30.8 32 19.7 strain 7 162 56 34.5 34 20.9 Average 155 50.3 32.5 31.8 20.7 only 4% of the females and 11 % of the males showed the vestiges), since 16% of the hybrid females and 20% of the hybrid males showed the plates. Even in this case, however, the genes for absence seemed to have more effect than those for presence. As was first pointed out by Sturtevant (1942), a most unique characteristic of the saltans group is the presence of velvety-opaque areas on the sixth tergite of the females of many species. Similar modifications have not been reported from flies of any other species group. The distribution of their presence in the saltans group is given in Table 2, columns 10 and 11. The shapes of these areas are often useful in identification, and the common types are illustrated, all to the same scale, in Figure 13. In cordata and neocordata, however, the modification takes the form of an internal flattened pouch which appears, after clearing and when viewed under high magnification, to be filled with minute hairs. 0 r genero "Drosophila." Rev. Brasil. Biol. 12: 79-96. Patterson, J. T., and R. P. Wagner. 1943. Geographical distribution of species of the genus Drosophila in the United States and Mexico. Univ. Texas Pub. 4313: 217-281. ----.,and W . S. Stone. 1951. Evolution in the genus Drosophila. New York, The Mac­millan Co., 610 pp. Pavan, C. 1950. Especies Brasileiras de Drosophila. II. Biol: Geral (Univ. Sao Paulo) 8: 1-37. ----. 1952. Thesis, Dept. Biologia, Univ. Sao Paulo, Brasil. ----, and L. E. Magalhaes. 1950. In Pavan, 1950. Magmhaes: The Saltans Group of Drosophila PLATE IV. Male genitalia and female spermathecae of the new species. Figs. 1-3, spermatheca, hypandrium and penis of D. lusaltan.s. Figs. 4--6, spermatheca, hypandrium and penis of D. nigrosaltan.s. Figs. 7-8, spermatheca and penis of D. milleri. Figs. 9-11, hypandrium, sperma­theca and penis of D. septentriosaltan.s. The Universi.ty of Texas Publication Salles, H. 1947. Sohre a genitalia dos drosofilidios (Diptera): I. Drosophila melanogaster e D. simulans. Sum. Brasil. Biol. 1: 311-383. Spassky, B. 1957. Morphological differences between sibling species of Drosophila. Univ. Texas Pub. 5721: 48-61. . Sturtevant, A. H. 1921. The North American species of Drosophila. Carneg. Inst. Wash. Pub. 301: 1-150. ----. 1942. The classification of the genus Drosophila, with descriptions of nine .new species. Univ. Texas Pub. 4213: 5-51. Wheeler, M. R. 1957. Taxonomic and distributional studies of Nearctic and Neotropical Dro­sophilidae. Univ. Texas Pub. 5721: 79-114. ----. 1960. Sternite modification in males of the Drosophilidae (Diptera). Ann. Ent. Soc. Amer. 53: 133-137. VIII. The Alagitans-Bocainensis Complex of the Willistoni Group of Drosophila 1 MARSHALL R. WHEELER AND LUIZ EDMUNDO DE MAGALHAES The alagitans species group of Drosophila was established by Patterson and Mainland (1944) for Drosophila alagitans and D. capnoptera, and was located in the subgenus Drosophila. Further study showed that the male genitalia of these species were quite similar to members of the willistoni group (Hsu 1949), and the group was therefore transferred to the subgenus Sophophora (Wheeler 1949). Carson ( 1954) found that there were three distinct entities involved under the name Drosophila bocainensis in Brazil; he identified one of them as bocainensis, described the other two under the names parabocainensis and bocainoides, and designated the complex as the bocainensis subgroup of the willistoni species group. In our collections of Neotropical Drosophilidae we have frequently taken specimens obviously related to bocainensis or to alagitans, and it is now clear that together they form a natural group of species which could with equal logic be termed a species group or one or two subgroups of the willistoni group. Since relatively few of these species have been successfully cultured in the laboratory, we are deferring a decision on the matter of "group" versus "subgroup(s)" until we have more evidence; meanwhile the expression "alagitans-bocainensis com­plex" describes this cluster of species adequately. With a few exceptions, the species are remarkably similar in their external morphology, but the male genitalia offer numerous useful characteristics. The dissections of the genitalia were done by the junior author (LEM) who also pre­pared camera lucida drawings of the various parts. The final drawings were made by Mrs. Linda Kuich, to whom we wish to express our sincere gratitude. All of the figures are drawn to the same scale; the true sizes are indicated by the 0.1 mm line which has been drawn on several of the figures. The known distribution of the species of this complex is shown in Figure 1 ; several of the points in South America are taken from the distribution records of Carson (1954) and Salzano (1955, 1956). Both morphologically and geo­graphically the species fall into two groups-the southern bocainensis-Iike forms and the more northern alagitans-like forms. D. bocainensis, however, reaches Colombia and thus overlaps the distribution of the "northern" members, while bocainoides, clearly a member of the northern group, occurs only in southern Brazil, as far as is known. Although we are not familiar with the species, kertes­zina Duda, kerteszina var. boliviensis Duda,2 and subinfumata Duda quite pos­sibly belong to this complex also. The known localities for these three species have also been added to the map. Rockefeller Foundation Fellow; permanent address: Departamento de Biologia Geral, Uni­versidade de Sao Paulo, Brazil. 2 We attempted to borrow the type of var. boliviensis from the Dresden Museum, but were informed by Dr. Hertel that he was unable to locate the specimen among Duda's type material. The types of kerteszina and subinfumata are believed to be in the Hungarian Museum, Budapest. KEY A alagitans Ba bocainens1s Bd bocainoides c capnoptera K kerteszina .. Kb var. boliviensis M megalagitans N neoologi tan s Pa parobocainensis Pd Ps Pl P2 s I 2 0 parobocainoides pseudobocoinensis " , type I " , type 2 sub infumoto species I species 2 species undetermined DISTRIBUTION OF SPECIES of the alagitans-bocainensis complex ....... (,)\ Ol ~ ~ ~ ~­ ~­..Q. ~ ~ ~ 0 ;;? l:!:: ·~· Ba ~ s· ;:s 0 1£ Pa ---p~ ~c: ~~~;------------- ---­ •-0£ Wheeler and Magalhaes: Alagitans-Bocainensis Complex LECTOTYPE SELECTIONS Since we had among our material an estimated twelve species belonging to the alagitans-bocainensis complex, it has been necessary to establish beyond any doubt the true identity of the described species: alagitans, capnoptera, bocainen­sis, parabocainensis, and bocainoides. Patterson ( 1943) stated that the type of alagitans, from Laguna Patzcuaro, Michoacan, Mexico, had been sent to the American Museum of Natural History in New York City. It cannot be found there at the present time and is presumed lost. In The University of Texas col­lection, however, are specimens from Valle de Huajumbaro and Uruapan, Mich., Mexico, collected by Gordon Mainland within a few days of the time he collected the type material, and which he clearly indicated belonged to the species ala­gitans. We have therefore used these specimens, which agree perfectly with the description, as representatives of the species, and believe that a neotype is not required. The holotype of capnoptera, from Perote, Veracruz, Mexico, as well as numerous other specimens from many localities, is in The University of Texas collection. The problem of the type specimens of bocainensis, parabocainensis and bocain­oides is rather complex, and has been explained in considerable detail by Carson (op. cit.). He has recently requested that the senior author (MRW) select and/or designate the types for these species in such a way that they will be valid under the International Code of Nomenclature. In naming the species parabocainensis and bocainoides Carson (op. cit.) set up a series of ten syntypes for each species; until recently they have been retained in the collection of Dr. Stalker at Washington University in St. Louis, Missouri. In order to clarify the identity of these species, and the new ones in our labora­tory, all of these syntypes were sent to us and several of them have been used to prepare the genitalia! figures. A lectotype has now been selected for each species, as follows: Drosophila bocainoides Carson 1954: 150. Lectotype selection. The specimen selected to be the lectotype is a male in good condition (one damaged antenna), and now bears the following four labels: "D. bocainoides, Vila Atlantica, ~,VII '51, H. Carson"; "S. P. Brazil"; "bocainoides"; (red label) "Lectotype, D. bocainoides, selected by M. R. Wheeler, 1960". This specimen, as well as four of the remaining paralectotypes, is being deposited in the American Museum of Natural History, New York City, in accord with Dr. Carson's wishes. The remaining five paralectotypes are in The University of Texas collection. These nine remaining paralectotypes are now so marked with a blue label bearing the word "Paralectotype". Some additional specimens, siblings of the original syntype series, are in the collection of the Museu Paulista, Sao Paulo, Brazil. Drosophila parabocainensis Carson 1954: 49. Lectotype selection. The speci­men selected as the lectotype is a male in good condition bearing the following four labels: "D. parabocainensis, ~, VII '51, H. Carson, Feliz, Rio Grande do Sul"; "Brazil"; "parabocainensis"; (red label) "Lectotype, D. parabocainensis, selected by M. R. Wheeler, 1960". This specimen, and four of the remaining paralectotypes, is being deposited in the American Museum of Natural History, New York City. The remaining five paralectotypes are in The University of The University of Texas Publication Texas collection. These nine remaining paralectotypes are now so marked with a blue label bearing the word "Paralectotype". Some additional specimens, siblings of the original syntype series, are in the collection of the Museu Paulista, Sao Paulo, Brazil. The type specimen(s) of D. bocainensis Pavan and da Cunha 1947 has appar­ently been lost. Carson (op. cit., p. 148 ff.) has described his attempts to locate this material. When he discovered that among the wild-caught specimens of this "species" there existed, in fact, three nearly identical species, two of which could be separated only on the basis of cytological differences, he attempted to find the types of bocainensis in order to apply the name to the correct member of the complex. He searched the collection in the Museu Paulista, Sao Paulo and the private collection of Dr. Pavan at the University of Sao Paulo, but was unable to locate any extant specimens of this species. Furthermore, the original living culture which had been used to describe the species and set up the types was no longer in existence. After examining the evidence he concluded that the species which was found abundantly in coastal Sao Paulo state was fairly certainly the true bocainensis, and supported his conclusion as follows: "The type material of Drosopldla bocainensis Pavan and da Cunha 1947 was collected in February 1944 at a remote fazenda called Campos da Bocaina in the extreme northeastern corner of the State of Sao Paulo, Brazil. This fazenda lies in a tongue of the state which projects eastward into the State of Rio de Janeiro and is approximately equidistant ( 30 km.) from Queluz, on the railroad Central do Brazil, and the town of Banana!. "As the laboratory strain from which this species was described is no longer in existence, it is not possible to assign the type material with certainty to one or the other of the species now recognized cytologically, nor has it been possible to make further collections from the type locality. Campos da Bocaina, however, lies in the coastal rainforest of the Serra do Mar, 270 km. (145 miles) north and east of Vila Atlantica, an area south of Santos from which extensive collections of Drosophila have been made. Both locations are closely similar ecologically and lie in the same mountain range. At Vila Atlantica, only one of the two sibling species has been found; this same sibling has likewise been found abundantly at two other locations in eastern Sao Paulo from which bocainensis-like flies have been analyzed cytologically. These locations are the Horto Florestal in the Serra Cantareira, near the city of Sao Paulo, and Mogi das Cruzes. "For the above reasons, it seems logical to utilize the original name Drosophila bocainensis Pavan and da Cunha to designate the particular member of the sibling species pair found abundantly in coastal Sao Paulo." He stated further that his choice ". . . conforms to the original description of this species by Pavan and da Cunha, 1947, pp. 18-20...." To represent the species bocainensis sensu strictu, he prepared pinned speci­mens (five males and five females) "... from a laboratory strain of flies derived from the offspring of a single wild female collected at Vila Atlantica in July 1951".Not being familiar with the requirements for the selection of neotypes, no single specimen was so designated at that time. This series of ten specimens was recently sent to us by Dr. Carson who requested that the present authors take whatever action seemed desirable to stabilize the name bocainensis. Wheeler and Magalhaes: Alagitans-Bocainensis Complex Drosophila bocainensis was thoroughly described by Pavan and da Cunha ( 194 7: 18)' but it is so similar in its general morphology to rxzrabocainensis Carson that their certain identification has been based only on the pattern of chromosomal inversions characteristic of each of them (Carson, op. cit.; Salzano, op. cit.). Mrs. Marta Breuer, of the University of Sao Paulo, compared a number of males of the two species and noted a small difference in the genital arch; although the difference was statistically significant there was sufficient overlap that identification on this character alone was not always reliable. We wish to thank Mrs. Breuer for permitting us to use some of her unpublished figures illustrating this character. We have now compared the male genitalia of the two species, using some of the syntypes of parabocainensis and some of the male siblings of the pinned series of bocainensis prepared by Carson. There are some consistent differences, as described below, but they are small differences and are not easily seen without well-prepared material. At the present time we feel that the selection of a neotype for bocainensis would serve no useful purpose. We are clearly dealing with a complex of closely related species in which Carson ( 1954), acting as a "first reviser", clearly re­stricted the name to a certain definable species. A neotype selection would do nothing that Carson's restriction has not already done; furthermore, a neotype, chosen from among the ten specimens from Vila Atlantica (about 145 miles distant from the original type locality), could lead to still greater confusion if it should be found that the species at Campos de Bocaina (Serra de Bocaina) is different than that at Vila Atlantica. MALE GENITALIA IN THE WILLISTON! GROUP The male genital arch of several species of the willistoni group was described by Hsu (1949), and Malogolowkin (1952, 1958), Nater (1953), and Spassky (1957) described all or part of the genitalia of various species of this group. Since the terminology employed by these various authors is frequently quite different, we feel that it is desirable to describe and illustrate here the nomenclature of the genitalia! pieces which we are using in the alagitans-bocainensis complex. An interpretation of the origin and homology of the genital parts among the various genera and subgenera of the family is also highly desirable, but we do not feel that we understand these relationships adequately to attempt such a treatment. Figure 2 illustrates the individual genital pieces of males of the alagitans­bocainensis complex as well as their appearance with different degrees of dis­section. The genital arch (A-1) is strongly narrowed dorsally and broadened ventrally; there are three margins: the anterior margin (A-2), in contact with the apical margin of the sixth tergite, the posterior margin (A-3) to which are attached the anal plates (A-8), and, on the ventral side between these two mar­gins, the undermargin (A-4) of Hsu (=bordo inferior of Malogolowkin) . The angle formed by the anterior and undermargins is termed the heel (A-5) by Hsu, the angulo anterior-inferior by Malogolowkin. Hsu described the angle formed by the posterior and undermargins as the "toe"; this is not well marked in the willistoni group (A-7) and is easily confused with a "posterior process" (A-6) which is often more highly developed in the Sophophora. The claspers (A-9; forceps) of each side are connected by the bridge (A-14; The University of Texas Publication 26 26 G Frc. Z. Diagrammatic representation of the complete male genital (copulatory) structure of an average member of the complex. A, genital arch, clasper, and bridge; B, hypandrium, clasper, and bridge; C, apodeme of the penis and pincers; D, penis; E = C + D; F = A + B; G = B + C +D. See text for key to numbers. Wheeler and Magalhaes: Alagitans-Bocainensis Complex ponte; decasternum of Okada); in this group the clasper is a small plate bearing a row of primary teeth (B-13) along the internal border (B-11 ) , and is elongated below into a hook (B-12). The external border is concave (B-10). From each clasper there is a prolongation of the bridge (B-15) which articulates with the hypandrium (B-16). Secondary clasper teeth are not present in this species com­ plex. The base of the hypandrium (novasternum plus ventral fragma of Okada) is its anterior margin (B-17); the lateral margins (B-18) are rolled or curved around. The distal (posterior) margin shows many landmarks; the median notch (B-19) may be shallow or deep; to each side is a paramedian expansion (B-20) which usually bear the paramedian spines (B-21), the latter sometimes being absent, sometimes doubled. Near each outer corner is an internal prolonga­tion (B-22) which articulates with the prolongation of the bridge. From the internal wall there also arise two hypandrial appendages (B-23) of varying size and shape. These do not appear to us to be homologous with the anterior gonapophyses (anterior parameres of Okada) of other species. Characteristically, the anterior gonapophyses possess minute sensilla, although Okada (1955) has shown that about 5 % of the species of Sophophora examined by him lack these sensilla. In some of our macerated preparations we have seen small clusters of sensilla, apparently located on small mound-like masses of soft tissue, sometimes near the caudal border of the hypandrium, and sometimes loosely attached (pre­sumably torn) along the apodeme of the penis. We believe that these masses may represent the extremely reduced anterior gonapophyses. The apodeme of the penis (C-24) is fused with the base of the pincers (C-25; pincas) ; the latter are sometimes straight but are more often bent or twisted apically. They may show an apical bifurcation (C-26); they may possess a lateral branch (C-27) and a secondary branch (C-29), the latter often protruding at a more extreme angle than the former. Near the base may be a basal expansion (C-28). The penis (D) is readily detachable from the apodeme and the pincers; from the base, on each side, there arises a ventral prolongation (D-30) the apices of which rest on the posterior margin of the hypandrium. There may also be lateral prolongations, the wings (D-31). On the median ventral line there is a prominence (D-32) and there is another on the back, dorsally (D-33). In Figure 2, E, is shown the typical position of the penis and the pincer­apodeme complex; F shows the relationship between the hypandrium, genital arch, clasper and bridge; and G shows the appearance of the copulatory apparatus after the genital arch and its associated parts have been removed. The species of the alagitans-bocainensis complex are mainly distinguished from the other members of the willistoni group by the features of the pincers, showing an apical bifurcation and/or lateral branches and/or secondary branches. (Com­pare, for example, the comparable figures for fumipennis, nebulosa, capricorni and sucinea in Malogolowkin, 1952: 91). The southern group of bocainensis-like forms all have the secondary branch of the pincer but lack apical bifurcations and lateral branches; the northern group of alagitans-like forms all have the lateral branches of the pincer and may also show the apical bifurcation and the secondary branches. The rather crude drawing of the male of kerteszina (Duda The University of Texas Publication 1927: 216) shows clearly the presence of the secondary branch of the pincer, thus leaving no doubt as to its position in this complex. Drosophila changuinolae, the last species described in this report, lacks these features of the pincer (Figure 8) and cannot be included in the complex. It is clearly a member of the willistoni group, however, the male genitalia appearing more similar to those of mangabeirai than to those of other species of the group (see Malogolowkin 1958: 444). SPECIES DESCRIPTIONS AND DISTRIBUTIONS The species of the alagitans-bocainensis complex can usually be recognized by the following combination of characters: wings usually darkened, lightly to intensely, the crossveins often strongly clouded; two subequal oral bristles; cheeks quite narrow; eyes with rather bare appearance; body color mostly tan, often with some pleural darkening, at least in males; basal scutellars divergent; only two distinct sternopleural bristles, the posterior one stout and long, the anterior one half as long or shorter; abdomen with dark bands, usually darkest apically on males, less heavily darkened on females. The two other species of the willistoni group that are most likely to be con­fused with the members of this complex are nebulosa and fumipennis. The simplest recognition character for these two concerns the sternopleural bristles; there are clearly three, the two anterior ones weak and approximately subequal, the posterior one long and strong. Drosophila alagitans Patterson and Mainland 1943. Univ. Texas Pub. 4313: 194. Recognition features: Upper half of pleura brown in both sexes, the lower half contrastingly pale; wings only slightly dusky, the posterior crossvein dark but only slightly clouded. Male genitalia as in Figure 3, A-C. Type locality: Laguna Patzcuaro, Mich., MEXICO. Also known from Valle de Huajumbaro, and Uruapan, Mich., Mexico. Drosophila neoalagitans, new species Mesonotum dark tan; acrostichal hairs in six rather regular rows. Pleura of male brownish, darkest on sternopleura; female pleura less brownish, the sterno­pleura paler than mesopleura. Male abdomen mostly black, the basal 3-4 tergites with yellowish bases; female abdomen paler, the tergites mostly tannish yellow with apical brown bands. Wings rather uniformly grayish, the posterior cross­vein a bit more heavily darkened. Costal index 2.4-2.6; 4th vein index 1.8-2.0. Male genitalia as shown in Figure 3, D-F. Body length 2.8-3.0 mm. Holotype male, allotype, and eleven paratypes (H410.7), Kenscoff, HAITI, June-July 1959, W. B. Heed and H. L. Carson. Additional specimens from Petionville, Haiti, and Hardware Gap, JAMAICA. Drosophila capnoptera Patterson and Mainland 1944. Univ. Texas Pub. 4445: 47. Recognition features: Pleura of both sexes brown, somewhat mottled above, on Wheeler and Magalhaes: Alagitans-Bocainensis Complex Fie. 3. Apodeme, pincers, hypandrium and penis of: A-C, alagitans; D-F, neoalagitans; G-I, m egalagitans. M any specimens of neoalagitans show only a single paramedian spine on each side of the hypandrium, rather than two, as shown. males becoming intense on lower sternopleura where the color contrasts strongly with the pale fore coxae; wings heavily clouded over both crossveins and over anterior half of wing; large species. Male genitalia as shown in Figure 4, A-C. The University of Texas Publication Type locality: Perote, V.C., MEXICO. Also known from Jacala, Hid., Tezuitlan and Huauchinango, Pueb., and San Cristobal, Chiap., Mexico; Volcan Santa Ana and Volcan Boqueron, EL SALVADOR; Monte Vyuca near Zamorano, HON­ B F1G. 4. Apodeme, pincers, hypandrium and penis of: A-C, capnoptera; D-F, bocainoides; G-1, parabocainoides. Wheeler and Magalhaes: Alagitans-Bocainensis Complex DURAS; Santa Maria de Ostuma, NICARAGUA; Volcan lrazu and Heredia, COSTA RICA; Boquete, PANAMA. Drosophila megalagitans, new species Holotype ~ : Mesonotum tan, scutellum a little darker; upper half of pleura brown, rather mottled, and the sternopleura with a brown area on upper posterior part surrounding the bristle bases, in front of this all pale. Wings dark, darkest along costal margin; crossveins clouded but the clouding diffuses into the general wing color. Costal index 3.0; 4th vein index 1.4. Male genitalia as shown in Figure 3, G-1. Body length greater than 3.0 mm. Female paratypes: Two females, taken at the same time and place as the holo­type, are not certainly conspecific. The wings are less heavily darkened, and the pleura is much paler. Holotype male, Cerra de Par Rico, east of Bucaramanga, COLOMBIA, Sept. 1956, H. L. Carson, M. Wasserman and H. Hoenigsberg. Male genitalia of holo­type mounted on slide No. 278, The University of Texas collection. Two paratype females, with the same collection data as the male. The collection was made at about 12,000 ft. in the Cordillera Oriental de Andes. Drosophila hocainensis Pavan and da Cunha 1947. Bol. Fae. Fil. Cien. Letr. Univ. S. Paulo 86: 18. Recognition features: Pale tan, both sexes with poorly defined dull brownish abdominal bands; pleura pale; wings slightly dusky, both crossveins with narrow clouds. Not readily separable from parabocainensis. Mrs. Marta Breuer (personal communication) has observed that on the ventral portion of the genital arch there is a posterior, rather knob-like process that is visible from external view, and which is useful in distinguishing these two sibling species. The "typical" position ofthis process is shown in Figure 5; in bocainensis it usually forms an angle of ±15° with the lower margin, while in parabocainen­sis the upper margin of the process is nearly parallel with the lower margin of the genital arch. However, there is 'considerable variation in this character in both species, even between sibling specimens, and the degree of overlap is such that certain identification cannot be based upon this feature alone. The male copulatory organs of the two species are also extremely similar (Figure 6), but a few fairly distinctive features are discernible in good prepara­tions. In bocainensis the outer corners of the posterior margin of the hypandrium B A t---:c:::-=-=~-L------------_ -___ __ __---_____,,. ________ Fm. 5. Ventral portion of genital arch of (A) bocainensis and (B) parabocainensis, showing the average difference in position and direction of the posterior process of the undermargin. The University of Texas Publication (near 22, Figure 2) are not so regularly curved toward the internal prolongation, the paramedian spines are usually larger, the apex and body of the penis are of slightly different shape, and the dorsal prominence shows some degree of a serrated edge. Type locality: Campos da Bocaina, Est. Sao Paulo, BRAZIL. Also known from several localities in the states of Sao Paulo, Minas Gerais, Parana, and Rio Grande do Sul, Brazil; Medellin and Rionegro, COLOMBIA; Merida, VENE­ZUELA; and from El Destino, ARGENTINA. D . ~ ({ _ t~ \ ! ' E FIG. 6. Apodeme, pincers, hypandrium and penis of bocainensis (A, C, E, G, I, K) and para­bocainensis (B, D, F, H, J, L). G, H-dorsal view; I, I-ventral view; K, L-lateral view. Wheeler and Magalhaes: Alagitans-Bocainensis Complex Drosophila parabocainensis Carson 1954. Evolution 8: 149. 1952. Science 116: 518; nomen nudum. Recognition features: Identical to bocainensis in gross morphology. The male genitalia are shown in Figure 6, and are discussed under bocainensis, above. Type locality: Feliz, Est. Rio Grande do Sul, BRAZIL. Also reported from Ponta Grossa, Emboaba, and Muitos Capoes, R.G.S., Pirassununga and Serra Cantareira, Est. Sao Paulo, and Montes Claros, Est. Minas Gerais, Brazil. Drosophila bocainoides Carson 1954. Evolution 8: 150. 1952. Science 116: 518; nomen nudum. Recognition features: Pleura of male brown, the fore coxae sometimes dark­ened; otherwise quite similar in appearance to bocainensis. Female pleura pale, rarely the mesopleura a trifle darkened. Male genitalia as shown in Figure 4, D-F. Type locality: Vila Atlantica, Est. Sao Paulo, BRAZIL. Also known from Serra Cantareira and Mogi das Cruzes, S.P., and Angra dos Reis, Rio de Janeiro, Brazil. Drosophila parahocainoides, new species Male. Tan; dark abdominal bands heavy; pleura mostly tan, but with a dis­tinctive brown area on pteropleura just anterior to base of haltere. Wings moderately dark on anterior half, gradually becoming paler behind; posterior crossvein with distinct cloud. Costal index 2.5-2.6; 4th vein index 1.5-1.6. Body length about 3.0 mm. Male genitalia are shown in Figure 4, G-1. Female un­known. This is possibly the same species as subinfumata Duda, described from Costa Rica. Holotype male and two paratype males, Boquete, Chiriqui Pr., PANAMA, June-July 1959, W. B. Heed and H. L. Carson. Six paratype males, same locality, August 1958, W. B. Heed and M. Wasserman. Drosophila pseudobocainensis, new species Tan, the scutellum a little darker; pleura mostly pale, the mesopleura a trifle grayish. Basal abdominal segments with prominent yellow basal bands, more apical tergites increasingly darker. Wings only slightly dusky, the posterior crossvein with a narrow cloud. Coastal index about 2.8; 4th vein index about 1.7. Body length 2.5-2.8 mm. Male genitalia as shown in Figure 7, A, B, D, F, I. This is possibly the same species as kerteszina Duda, described from Costa Rica. In studying preparations of male genitalia from specimens from all parts of the range, three types were found. We are considering the form found in Central America to be the "typical" one. The differences between the types are at least as great as those between bocainensis and parabocainensis, but since there is at this time no information on possible genetic isolation, hybridization, either natural or forced, or on other characteristics of the natural populations, we do The University of Texas Publication FIG. 7. Apodeme, pincers, hypandrium and penis of pseudobocain.ensis. Typical form (Boquete, Panama): A, B, D, F, I ; variant from Popayan, Colombia: G, J ; variant from Coroico, Bolivia: C, E, H, K. L, M-lower margin of genital arch of typical form (L) and of variant from Popayan (M). not feel that it is advisable to designate these variants as distinct species. The male genitalia of type 1, characteristic of northern South America, are shown in Figure 7, G, J, M; type 2, from Bolivia, is shown in Figure 7, C, E, H, K. Holotype male, allotype and ten paratypes, Boquete, Chiriqui Pr., PANAMA, June-July 1959, W. B. Heed and H. L. Carson. Also known from Cerro la Cam­pana, Panama; Barro Colorado Is., CANAL ZONE; Volcan Santa Ana and Laguna Alegria, EL SALVADOR; and San Jose, COSTA RICA. Type 1: Popayan, Colombia; 65 klm. southeast of Bogota, Colombia; Maracay, Wheeler and Magalhaes: Alagi.tans-Bocainensis Complex Venezuela. Type 2: Coroico, Bolivia. The latter is possibly the same as kerteszina var. boliviensis Duda. Unnamed species 1 A single male, from near Melgar, about 30 klm. west of Girardot, Colombia, collected Nov. 1955 by W. B. Heed, is clearly different from the other known members of the complex. This male was teneral, showing little color. The genitalia are illustrated in Figure 8, A-C. Unnamed species 2 A single male, from Maracay, Venezuela, collected Nov. 1956 by M. Wasser­man, has rather unique genitalia (Figure 9). The specimen was found among rather numerous individuals of pseudobocainensis, type 1, among which it was not recognized as being different until the male genitalia were examined. Drosophila changuinolae, new species Mesonotum dull, dark tan, becoming distinctly more brownish on scutellum. Pleura browner than disc of mesonotum, especially on mesopleura, the sterno­pleura often a bit paler. Acrostichal hairs irregularly 6-rowed. Front dull tan, the orbits dull grayish yellow, the ocellar area blackish. Arista with 6 dorsal and 3 ventral branches basal to the terminal fork. Head bristles stout and black. Anterior reclinate orbital about~ length posterior; face tan; carina rather large; 2nd oral bristle not quite as strong as first. Two anterior sternopleurals thin, subequal, the posterior one long and stout. FIG. 8. Apodeme, pincer, hypandrium and penis of unnamed species from Melgar, Colombia: A-C. D-H, D. changuinolae. The University of Texas Publication Legs all tan. Abdominal tergites with dull brown apical bands, those of more posterior segments becoming larger and blacker. Wings darkened, more so on anterior half; posterior crossvein with a narrow cloud. Costal index about 1.5; 4th vein index about 2.2. Body length about 2.0 mm. The male genitalia and female spermatheca are shown in Figure 8, D-H. Holotype male, allotype, and eleven paratypes, Changuinola, PANAMA, June-July 1959, W. B. Heed and H. L. Carson. Also known from Leticia, Ama­zonas, COLOMBIA. Relationships. Belongs to the willistoni species group, within which it shows affinities in both external and genitalia! characteristics with D. mangabeirai Malogolowkin. It is clearly not a member of the alagitans-bocainensis complex. EcoLOGY Very little is known about the ecology of the species discussed here. Heed (1957) has made the most extensive ecological study of Central American Dro­sophilidae (principally in El Salvador), a study which included two species of this complex: capnoptera and pseudobocainensis (=alagitans-Iike of Heed) . He reported on the composition of several micropopulations, defined as that group of flies collected over a definite, localized feeding site. Such sites were sometimes quite small, such as a single fallen fruit, or a single flower or fungus. On Volcan Santa Ana at an elevation of about 5000 feet, capnoptera and pseudobocainensis were elements of four of the seven micropopulations studied. The following table, extracted from his Table 9, shows the feeding sites involved and the number of individuals encountered: On acorns On flowers On figs On drupes Total no. flies 283 100 189 219 Per cent capnoptera 30% 3% 30% Per cent pseudobocainensis 28% 20% (Acorns: Quercus skinneri; flower: unidentified; figs: Ficus sp.; drupes: Citha­rexylum donnelrsmithii). The preponderance of capnoptera and pseudobocainensis over acorns is quite interesting in view of the fact that pseudobocainensis, type 2, from near Popayan, Colombia, was quite common in an oak grove; several hundred specimens were taken in less than an hour simply by sweeping with a net beneath the trees. Heed (op. cit.) felt that acorns, as such, were not actually used for egg deposition and larval development. However, Chymomyza amoena has been reared from acorns in eastern and central United States, so there is no reason to doubt that larvae of other species might not be similarly adapted to this food source. LITERATURE CITED Carson, H. L. 1954. lnterfertile sibling species in the willistoni group of Drosophila. Evolution 8: 148-165. Duda, 0. 1927. Die siidamerikanischen Drosophiliden (Dipteren) unter Beriicksichtigung auch der anderen neotropischen sowie der nearktischen Arten. Arch. f. Naturg. 91 A 11-12: 1-228. Heed, W. B. 1957. Ecological and distributional notes on the Drosophilidae (Diptera) of El Salvador. Univ. Texas Pub. 5721: 62-78. Wheeler and Magalhaes: Alagitans--Bocainensis Complex FIG. 9. Male copulatory apparatus of unnamed species from Maracay, Venezuela. Hsu, T. C. 1949. The external genital apparatus of male Drosophilidae in relation to sytematics. Univ. Texas Pub. 4920: 80-142. Malogolowkin, C. 1952. Sohre a genitalia dos Drosophilidae (Diptera) : III. Grupo willistoni do genero Drosophila. Rev. Brasil. Biol. 12: 79-96. -----. 1958. Sohre a genitalia dos Drosofilideos. V. A genitalia masculina em "D. man-­gabeirai" (Diptera, Drosophilidae) . Rev. Brasil. Biol. 18: 443--445. Nater, H. 1953. Vergleichend-morphologische Untersuchung des iiusseren Geschlechtsapparates innerhalb der Gattung Drosophila. Zool. Jb. (Syst.) 81 (5~): 437-624. Okada, T. 1955. Comparative morphology of the Drosophilid flies. II. Phallic organs of the subgenus Drosophila. Kontyu 23: 97-104. Patterson, J. T . 1943. The Drosophilidae of the Southwest. Univ. Texas Pub. 4313: 7-216. -----,and G. B. Mainland. 1944. The Drosophilidae of Mexico. Univ. Texas Pub. 4445: 9-101. Pavan, C., and A. B. da Cunha. 1947. Especies brasileiras de Drosophila. Bol. Fae. Fil. Cien. Letr. Univ. Sao Paulo 86: 3--46. Salzano, F. M. 1955. 0 problema das especies cripticas estudos no sub-grupo bocainensis (Dro­sophila) . Boletim Inst. Ciencias Nat. 4 (Porto Alegre, Brasil): f--88. -----. 1956. Chromosomal polymorphism and sexual isolation in sibling species of the bocainensis subgroup of Drosophila. Evolution 10: 288-297. Spassky, B. 1957. Morphological differences between sibling species of Drosophila. Univ. Texas Pub. 5721 : 48-61. Wheeler, M. R. 1949. Taxonomic studies on the Drosophilidae. Univ. Texas Pub. 4920: 157-195. IX. Genetic Characteristics of Island Populations WILLIAM B. HEED1 INTRODUCTION On September 17, 1835, Charles Darwin first set foot on Chatham Island of the Galapagos Archipelago. The diversity of obviously related birds which Dar­win encountered laid much of the groundwork for his thoughts later expressed in The Origin of Species. Eighty-one years ago Alfred Russel Wallace enumerated the advantages in studying island life for understanding the distribution of plants and animals ( 1880) . Every one of the advantages that he listed contained the one common denominator of simplicity. Ecologists have turned to the arctic, the desert and to oceanic islands to study interacting forces in a simplified environ­ment in order to better comprehend more complex environments (for instance Fosberg, 1951). And now geneticists, interested in population structures and histories, are capturing butterflies in the Isles of Scilly in order to count the num­ber of spots on the underside of the hind wing (E. B. Ford, 1960). Oceanic islands serve the student of organic diversity in the same way a well­equipped laboratory is necessary to the student of DNA. The relative simplicity of interacting factors and the contrast of conditions under which island popula­tions are maintained in comparison to continental ones enables the investigator to ferret out details often obscured on the mainland. The results of natural selec­tion (and sampling errors) are more open to inspection in island populations and this is especially true of biparental organisms that can be cross-bred under labora­tory conditions. The present report is a continuation of the genetic studies initiated by the author (1957) and continued in conjunction with N. B. Krishnamurthy (1959) on the dunni subgroup of the cardini species group of Drosophila inhabiting, as far as known, all the islands of the Lesser Antilles, Puerto Rico and Jamaica (Table 1) . Of the six island populations tested previously, five could be identified by their phenotype alone. The populations are: Puerto Rico (PR), St. Thomas (ST), Guadeloupe (GU), Barbados (BA), St. Vincent (SV) and Grenada (GR). Only SV and GR are phenotypically identical. All possible crosses among the stocks (30) have been made; a fertile second generation has been obtained in only five cases. Several instances of unequal sex ratios were obtained showing a Hal­dane effect in one series and a maternal effect in another series. The major chromosome differences between some of the populations consist of shifts in heterochromatin. The dot (IV) chromosomes of PR and ST have added hetero­chromatin making them short metaphase rods. This chromosome in the Jamaica population is a long rod due to the accumulation of heterochromatin. The differ­ences in the island populations were attributable to the lack of genetic continuity between the islands and the independent adaptation (genetic change) in each island form. 1 Department of Zoology, University of Arizona, Tucson. TABLE 1 Species of the cardini group and their distributions Neotropical mainland West Indies 1. Western hemisphere tropics D. cardini Sturtevant 2. Mexico to Brazil D. cardinoides Dobzhansky and Pavan 3. Mexico to Trinidad D. parthenogenetica Stalker D. neomorpha Heed and Wheeler 4. South America D. polymorpha* Dobzhansky and Pavan D. n. neocardini Streisinger D. n. mourensis da Cunha D. n. itambacuriensis da Cunha 5. Andes of Bolivia and Peru D. procardinoides Frydenberg 1. Florida, Cuba, Jamaica, Hispaniola D. acutilabella Stalker 2. Jamaica D. belladunni Heed and Krishnamurthy 3. Puerto Rico--St. Thomas D. d. dunni Townsend and Wheeler D. d. thomasensis n. subsp. 4. St. Kitts-Guadeloupe D. arawakana n. sp. D. a. arawakana, nom. subsp. D. a. kittensis n . subsp. 5. Martinique D. caribiana n. sp. 6. St. Lucia D. antillea n. sp. 7. Barbados D. nigrodunni Heed and Wheeler 8. St. Vincent-Grenada D. s. similis W illiston D. s. grenadensis n. subsp. • Also present on Grenada. The present paper attempts to give equal weight to the similarities, as well as the differences, between each island population. In so doing, it is believed that different types of selection pressures have been identified. The stocks recently tested are: Jamaica (JA), St. Kitts (SK), Martinique (MA) and St. Lucia (SL) . On the basis of the results of these tests and of the previous ones, each island population has been given a specific or a subspecific name. Three new species and three new subspecies are described. The cardini subgroup is established to contain all members not included in the dunni subgroup. The male genitalia of all members of the cardini group are illustrated and compared. The body and wing lengths observed in each island population are recorded and compared. Salivary chromosome analysis is extended. Finally, in the discussion, the be­havior of the island representatives of the dunni subgroup is compared with that of the continental representatives of the cardini subgroup. The differences are in accord with present day concepts of the different history, the isolation and the different ecologies that are bestowed upon island populations in contrast to mainland populations. Reference will be made repeatedly to the "previous" or "first" publication which refers to Heed and Krishnamurthy ( 1959) . Although each island popu­lation is now recognized as a taxon, they will be designated as before with the initials of the islands from which they were established. TAXONOMY AND MORPHOLOGY Table 1 lists the known forms in the cardini group according to their distri­ Heed: Island Populations of Drosophila bution. Fifteen species are represented, four of which are divided into two races each. The dotted line in Table 1 separates the two subgroups: the cardini sub­group here established as consisting of the seven species on the mainland and D. acutilabella, and the dunni subgroup which consists of seven species distributed on 10 islands in the Lesser Antilles, Puerto Rico and Jamaica. The chief pecu­liarity of the dunni subgroup is a genetic one; the seven species hybridize among themselves more easily than with any member in the cardini subgroup. Three new species and three new subspecies in the dunni subgroup are described below in a brief outline form, made possible because the characteristic features such as size, color pattern, and genitalia have been analyzed in conjunction with other members of the group and are presented separately. Figure 1 illustrates the similarities and differences in the male apodema in the cardini group as drawn from freshly prepared slides. Variations within a species from different areas are minute. For instance, when the apodema of D. cardini from Florida was compared to that from Chile, no consistent differences could be found. By contrast, the differences between species (where they exist) are con­stant and reliable. The nigrodunni complex includes all the middle island forms in the Lesser Antilles: St. Kitts, Guadeloupe, Martinique, St. Lucia and Barbados. It is impossible to distinguish any of the five island forms by this one character. Another surprising fact is the discovery of two triads of species on the mainland. Drosophila neocardini, polymorpha and neomorpha are closely related to one another. Drosophila cardinoides, procardinoides and parthenogenetica have more characters in common with one another than with other members of the group. Drosophila cardini unfortunately cannot be related by this method and stands alone with a unique apodema. The dunni subgroup as a whole is quite diverse and is united in Figure 1 because of the genetic affinities within the group and the distributional pattern. Drosophila acutilabella is not easily placed. According to the morphology of the apodema, it could have affinities with the neocardini­polymorpha-neomorpha triad as well as with the dunni subgroup. Some very obvious size differences have long been noted in the dunni subgroup strains. Figure 2 represents the results of measuring the body length (base of antenna to tip of abdomen, solid lines) and wing length (dotted lines) of 25 females and 25 males from 10 island strains reared for one complete generation at 19.5° C. in not too crowded conditions since the larvae were supplied yeasted kleenex on which to pupate. There were no controls in the sense that two strains from the same island were not tested. One might conceive that the differences between strains are those that have incidentally become fixed (homozygous) in the laboratory. This might be true to some extent but the correlations that do exist lend confidence to the data. For instance, the strains from Puerto Rico and St. Thomas belong to the same species and they differ from all other species by having very large females and small males. The St. Vincent and Grenada strains belong to the same species and they differ from all other species by having very large females and the largest males. The correlation breaks down with the third two-island species but even in this case there appears to be an explanation. The Guadeloupe and St. Kitts strains are interfertile and are regarded as belonging to the same species, although St. Kitts females have wings much shorter than the body. There is only one other strain in which the wings are shorter than the The Universi.ty of Texas Publication bel/odunni dunni nigrodunni complex similis JA PR,ST SK, GU, MA, SL, BA SV,GR ~ 1' I II I I I I I : I I I I I {, ,--~~~~~~~~~----J/'..~~~~~~~~~~~---., t l r ocufilobel/o neocordini polymorpho neomorpho Fw. 1. Tip of apodemas (penes) of males of the cardini group species in profile. A: transparent sheath; B: sperm duct; C: apodema. Heed: Island Populations of Drosophila 2 .80 +. +. SV ST FEMALES 2.10 ... 0 I -1-· _j__'t 11I:--GU I : i : I JA BA 2 .60 I t' :' T: +·L. SK I I I E ~' -t-: E2.50 t -I +~ MA :I: 2.40 ~ -' ' - I­ ·-r- C> t ' ' ' I z : : w ...J 2.30 .--1--: -+-' '· +I ' -1-' ­ ' ! i2 .2 0 -+­ i +r· 2 .10 MALES -t+ fi­ 2.00 ~ - DECREASING BODY SIZE OF FEMALES Fw. 2. Sample mean and its 95% confidence interval for body length (solid) and wing length (bars) of ten island strains reared at 19.5° C. n= 25. body; this is the Barbados strain. The males of St. Kitts and Barbados strains also tend to have shorter wings than body, while in all the other strains males tend to have wings longer than the body, except St. Vincent and Grenada which have exceptionally large males anyway. The situation may conceivably be explained as follows. St. Kitts and Barbados are very open islands and planted mostly in sugar cane. Shorter wings could be of value if Drosophila fly in open windy country, hopping from one orchard or wood patch to the next. The other islands have comparatively large continuous areas of woodland or fincas. St. Thomas has quite broken country with scanty vegetation but the wings of its inhabitants are not exceptional. Drosophila similis Williston. D. similis Williston 1896. Trans. Ent. Soc. London: 415. D. similis, Sturtevant 1921. Cam. Inst. Pub. 301: 79. Originally described by Williston (1896) from St. Vincent. Sturtevant (1921) recorded it from Jamaica and Barbados, and stated that it was similar to D. cardini. Metz (1916) reported the metaphase chromosomes as five pairs of rods and a pair of dots (Cuban strain). Stalker (1953) could find no living or pinned material of similis and suggested that his newly described D. acutilabella, which certainly is a sibling species of cardini, could be a synonym of similis if it were not for the fact that acutilabella has two pairs of V's instead of four rods in the metaphase. The University of Texas Publication I have recently compared two cotypes (#20353, AMNH, male; #2756.1, Cornell University, female) of D. similis with the dunni subgroup form from St. Vincent and Grenada and have concluded that they are the same species. A third cotype (#2756.2, Cornell University, female), however, is not a cardini group form but belongs in the tripunctata group. · I have selected cotype #20353, belonging to the American Museum of Natural History, as the Lectotype of Drosophila similis. Drosophila similis does not exist outside of the islands of St. Vincent and Gre­nada (unless it is present on the Grenadines and possibly Tobago). It does not have five pairs of rods and a pair of dots (see below) . The only member of the cardini group known to have this metaphase arrangement is D. cardini. Additional taxonomic points of value for D. similis are as follows: Abdominal pattern: As pictured in H. &K. ( 1959). Apodema: Figure 1, this publication. Size: Figure 2, this publication. Chromosomes: Two pairs of V's, a pair of rods and a pair of dots. Figure 2, H. &K. (1959). Drosophila similis is hereby separated into two subspecies. The form from St. Vincent is D. similis similis and the form from Grenada is D. similis grenadensis, new subspecies. The separation is based on genetic differences in crosses to other island forms. Grenadensis gives a more extreme sex-ratio with Martinique females, Barbados females and Guadeloupe males. The type material for the new subspecies is deposited in The University of Texas collection. Drosophila arawakana, new species This represents SK (St. Kitts) and GU (Guadeloupe) of this and previous publications. Hybrids between the two island forms are perfectly fertile. Abdominal patterns: AsillustratedinH.&K. (1959). Apodemas: As illustrated in Figure 1, in the nigrodunni complex. Sizes: As shown in relation to other members of the dunni subgroup, Figure 2. Chromosomes: Two pairs of V's, a pair of rods and a pair of dots, as illustrated in H. & K. (1959). Drosophila arawakana is divided into two subspecies: D. arawakana arawakana from Guadeloupe, and D. arawakana kittensis from St. Kitts. D. a. kittensis has significantly shorter wings than body whereas arawakana does not, nor does kittensis give as severe deviation from the normal sex-ratio in crosses to other island forms as does arawakana. Type material deposited in The University of Texas collection. Drosophila antillea, new species Represents SL (St. Lucia) of this and earlier publications. Abdominalpattern: Illustrated in H. &K. (1959). Apodema: Figure 1, this publication, in the nigrodunni complex. Size: Figure 2, this publication. Chromosomes: Two pairs of V's, a pair of rods and a pair of dots, as illustrated in H. & K. (1959). Heed: Island Populations of Drosophila Distribution and types: Restricted to the island of St. Lucia, B.W.I. Type material deposited in The University of Texas collection. Drosophila caribiana, new species Denoted as MA (Martinique) in this and earlier publications. Abdominal pattern: Illustrated in H. & K. (1959). Apodema: Figure 1, this publication, in the nigrodunni complex. Size: Figure 2, this publication; the smallest member of the dunni subgroup. Chromosomes: Two pairs of V's, a pair of rods and a pair of dots, as illustrated inH. &K. (1959). Distribution and types: Restricted to the island of Martinique, F.W.I. Type material deposited in The University of Texas collection. Drosophila dunni Townsend and Wheeler. D. dunni Townsend and Wheeler 1955. Jour. Agric. Univ. Puerto Rico 39: 60. D. dunni is hereby separated into two subspecies: D. dunni dunni, typical subspecies, from Puerto Rico; and D. dunni thomasensis, new subspecies, from St. Thomas. The distinguishing characters are that thomasensis has a darker grayish abdomen than the nominate race and its metaphase differs in that it has a J­shaped X chromosome instead of a rod-shaped X as in dunni, and one of the autosomes is J-shaped instead of V-shaped as in dunni. Type material of the sub­species deposited in The University of Texas collection. HYBRIDIZATION TESTS AND CYTOLOGY St. Kitts. Fertile with Guadeloupe and gives a normal sex-ratio in the progeny (Table 2). The sex-ratios among the hybrids from BA X SK and SV X SK are abnormal, but not as sharply so as those among the hybrids from BAX GU and SV X GU. Hybrids of SK with PR, ST and JA are completely sterile, but gene exchange is possible through backcrosses with BA and SV. Martinique. The most genetically isolated species in the dunni subgroup (Table 3). All hybrid males and females are sterile among themselves and are sterile when backcrossed to the parents (Figure 3). Dissections, where made, showed the presence of aborted ovaries and testes in the majority of cases. Mar­tinique males show extremely abnormal sex-ratios with other island females located south of Martinique in that the majority of the hybrid male zygotes are lethal. The Y chromosome from MA is lethal or near lethal in combination with southern island X's. Martinique females show a maternal cross-lethal effect in that many of the hybrid females do not survive in tests with SV and GR males. The cross-lethal gene or gene complex in the X chromosome of Grenada acts as a dominant, and that of St. Vincent as a semi-dominant, in Martinique cytoplasm. In Grenada cytoplasm with hybrid genotype, the females are viable. St. Lucia. Quite unambiguously isolated from PR and ST to the north and SV and GR to the south (Table 4). In most of these tests no hybrids were produced although dissections showed that the majority of the females in each cross were The University of Texas Publication TABLE 2 Hybridization Tests with St. Kitts P er cent No. F X M Total F1 males Remarks 1. 14SK X 11 PR 0 Females fertilized 2. 11PR X 14 SK 33 33.3 Males sterile 3. 10 PR/ SK X 35 SK 0 4. 7 SK X 7 ST 0 5. 13 SK X 12GU 53 43.4 Fertile 6. 24SK X 20GU 142 43.0 Fertile 7. 22GU X 28 SK many Fertile 8. 16BA X 6 SK 477 28.3 Males sterile; 95% males deformed wings 9. 27 BA/ SK X 26 BA 767 34.9 Fertile 10. 17 BA/ SK X 20 SK 421 40.0 Males sterile 11. 19 BA/ SK/ SK X 26 SK 170 44.7 Fertile 12. 14 SV X 15 SK 323 33.1 Males sterile; 44% males abnormal abdomen and deformed wings 13. 10SV X 7 SK 150 26.0 Few males fertile 14. F1 of No. 13 8 50.0 Males sterile 15. 14SV/ SK X 9 sv 78 43.5 Males sterile 16. 6 SV/ SK/ SV X 14 SV 33 30.3 Males sterile 17. 20 SV / SK/ SV / SV X 14 SV 54 40.0 Fertile 18. 21 SV/ SK X 24 SK 360 31.4 Males sterile 19. 30 SV / SK/ SK X 26 SK 40 57.5 Males sterile 20. 15 SK X 12 SV 49 47.0 Males sterile 21. 6 SK/ SV X 18 SK 24 50.0 Males sterile; 75% males deformed wings 22. 6 SK/ SV X 10 SV 11 27.3 Males sterile 23. 12SK X 10 GR 31 64.5 Males sterile; 35% males and females abnormal abdomen 24. 9 SK/ GR X 20 SK 4 0 25. 15 SK X 16 JA 0 Females fertilized 26. 8 JA X 10 SK 14 0 27. 14 JA/ SK X 21 SK 1 0 Most females aborted ovaries inseminated but the sperm were dead. In the series of crosses listed in Table 4 the hybrid males were sterile but the females were not tested for fertility in backcrosses. Comparison of Table 4 with Table 7 shows that while sperm from SL will not live in SV females, they can survive and function in SV / GU hybrid females. Table 7 also establishes that hybrid females with Y2 SL genes and 14 GU and 14 SV genes are at least partially fertile even to BA males and this is indicated in Figure 3 as GU X SL. Jamaica. A total of 38 tests were made with JA crossed to other members of the dunni subgroup in large and small mass matings, Table 5. JA males will mate with all other members but do not produce hybrids. JA males will, however, produce progeny with JA/ GR hybrids, but by the third backcross to JA males no offspring at all are produced (Nos. 24, 25 and 26). The situation is all the more unusual since the original cross involved JA egg cytoplasm. Theoretically Heed: Island Populations of Drosophila TABLE 3 Hybridization tests with Martinique (20 males and 20 females each) Per cent o. FXM Total F1 Males Remarks 1. PR X MA 242 40.1 2. ST X MA 380 41.8 1 % females and males abnormal abdomen. 3. GU X MA 540 31.7 5% females and males abnormal abdomen. 4. MA X MA 208 53.4 5. SL X MA 66 9.1 50% males rough eyes; ovaries aborted. 6. BA X MA 178 2.8 4% females abnormal abdomen. 7. sv X MA 331 0.3 9% females abnormal abdomen; ovarie aborted in all. 8. SV X MA 474 0.6 9. GR X MA 289 0.4 3% females abnormal abdomen; ovaries aborted in all. 10. GR X MA 271 1.1 11. *JA X MA 0 12. MA X PR 62 32.3 50% males and females abnormal abdomen. 13. MA X ST 67 34.3 All males rough eyes; 50% both sexes abnormal abdomen. 14. MA X GU 514 48.4 15. MA X MA 208 53.4 16. MA X SL 736 52.3 Ovaries aborted. 17. MA X SL 127 44.1 Testes aborted. 18. MA X BA 662 54.2 Ovaries and testes aborted. 19. MA X BA 563 53.3 20. MA X SV 98 62.2 All males rough eyes; testes aborted; 20% both sexes abnormal abdomen. 21. MA X SV 141 70.2 Same a above. 22. MA X GR 12 91.7 All males rough eyes; 30% males abnormal abdomen. 23. MA X GR 113 100.0 All male rough eyes; 3% abnormal abdomen, testes aborted. 24. MA X JA 0 • Two mass matings. by the third backcross the hybrid female should contain about 17% GR genes. Apparently this dosage is more critical than higher dosages of the foreign genes. This interesting cross is being repeated. J A females crossed to males from all other islands, except Grenada and Bar­bados, produce sterile progenies. JA females were not tested to SV males. Fertile progeny of both sexes are produced in the first backcross to GR males and in the second backcross to BA males. Unequal sex-ratios are evident when GU and SK are used as males. This was not totally unexpected since the Y chromosomes from these islands are known to be lethal or near lethal with other island X chro­mosomes. However, in Table 5, test No. 8, 18 of the total of 19 males emerged in the last three days of a 15-day emergence period as though nullo-Y sperm were produced in quantities by one of the GU males. It is also possible that the GU population has several types of Y chromosomes, some of which are not lethal in the hybrid males. To check this, 49 pair matings of JA females X GU males were prepared (Table 6). Out of the 21 pairs that were fertile, only one pair The University of Texas Publication xp R ST sK Gu MA sL BA sv GR JA p R s T s K Gu MA s L B A sv GR J A FIG. 3. Fertility relations in the dunni subgroup. Shaded squares: both sexes of hybrids fertile. Open squares: not tested. PF: hybrid females partially fertile in backcrosses. S: both sexes of hybrids sterile. N: hybrid females not tested for fertility in backcrosses. 0: no hybrids (or very few) produced. TABLE 4 Hybridization tests with St. Lucia (20 males and 20 females each) Per cent Per cent No. F X M Total F1 males No. F xM Total F1 males 1. PR X SL 0 13. tSL X PR 2 0 2. tST X SL 0 14. tSL X ST 6 0 3. GU X SL 361 42.4 15. SLX GU 0 4. *GU X SL 123 39.0 16. *SL X GU 62 19.4 5. MAX SL 127 44.1 17. SL X MA 1 0 6. SL X SL 298 47.3 18. SL X SL 298 47.3 7. *SL X SL 168 44.0 19. *SL X SL 168 44.0 8. BA X SL 145 36.6 20. SL X BA 163 54.0 9. *BA X SL 204 48.5 21. *SL X BA 520 50.2 10. :j:SV X SL 0 22. tSL X SV 4 25.0 11. tGR X SL 0 23. *SL X SV 58 17.2 12. JA X SL 82 46.4 24. SL X GR 0 25. SL X JA not made * Incompl ete mass mating. t Two mass matings. :!: Three mass ma tings. Heed: Island Populations of Drosophila TABLE 5 Hybridization tests with Jamaica (D. belladunni) Per cent No. Total F1 males Fertility Phenotype F XM 1. 2. 24JA X 26 PR 20 JA X 20PR 2 6 50.0 83.0 Sterile Sterile Equals JA Almost JA 3. 19PR X 6JA/ PR 0 Testes aborted 4. 2 JA/ PR X 16 PR 0 Ovaries aborted 5. 20 JA X 20 ST 2 50.0 Sterile Males JA; females almost JA 6. 7. 8. 8 JA X 10SK 14JA/ SK X 21 SK 19JA X 23 GU 14 1 105 0 I {Most females 0 s aborted ovaries 18.1 Sterile Females variable; males almost JA 9. 14JA/ GU X 14JA 0 Females sterile 10. 16 JA/ GU X 30 GU 0 Females sterile 11. 20 JA/ GU X 30 SK 0 Females sterile . ....... 12. 20 JA X 20 SV/ GU* 24 0 13. 22 JA/ SV/ GU X 23 GU 0 Females sterile 14. 25 JA X 18 MA 0 15. 20JA X 20MA 3 pupae 16. 22 JA X 21 SL 82 46.3 Males sterile Intermediate 17. 20JA X 20BA 126 47.6 Males sterile Intermediate 18. 14JA X 25 BA 133 54.1 Males sterile Intermediate 19. 27 JA/ BA X 30 BA 171 30.0 Males sterile BA to intermediate 20. 11 JA/ BA/ BA X 25 BA 104 4·1.1 Fertile BA to dilute BA 21. F1 of No. 20 many Fertile All equal BA 22. 19 JA X 19 GR 36 41.1 Males sterile Males JA; females almost JA 23. 10GR X 8 JA/ GR 0 24. 10 JA/ GR X 9 JA 79 36.7 M ales sterile All equal JA 25. 20 JA/ GR/ JA X 18 JA 21 33.3 M ales sterile 26. 14 JA/ GR/ JA/ JA X 16 JA 0 27. 11JA/ GR X 17GR 48 43.7 Fertile Variable to almost JAandGR 28. F1 of No. 27 Z5 68.0 Fertile Males vary; females almost GR 29. F2 of No. 27 196 38.7 Fertile All equal GR 30. 20 JA X 20 JA 432 46.8 Fertile 31. Females of all islands X JA males 0 All females inseminated • SV / GU is F12 hyb1·id . TABLE 6 D. belladunni (JA) retests with GU Aborted Eggs prcscnt F XM o. pairs Fertile Femal es Males ovaries but sterile JA (356.3f ) X GU 22 11 109 0 83 10 JA (356.3d) X GU 15 7 136 0 23 JA (356.3d) X GU 12 3 101 Total 49 21 346 The University of Texas Publication produced a male. The total sex ratio was 346 females to one male or 0.28%males. This is believed to be a truer ratio than that seen in the mass mating and places Jamaica as the extreme "southern" island in the sex-ratio dine with GU males crossed to southern island females (Table 13). Four-way island crosses. The list of crosses in Table 7 illustrates the theoretical possibility that genes from one island could be transmitted to the populations of other islands in the right combinations although the backcrosses were only carried for four generations. The fourth generation females in the majority of cases laid eggs and were no doubt at least partially fertile. It is also true that certain combi­nations which ought to work do not. For instance in test No. 6, it is surprising that MA males produced no progeny with the triple hybrid females, because MA males produce offspring with SV, GU and BA separately. It is also surpris­ing that SL males in test No. 7 produced very few progeny with the triple hybrid since it worked with BA and GU alone but not with SV (even after three attempts in mass matings, Table 4), but did work with SV/GU (test No. 9). Thus, the general combining ability among the island forms is fair to poor and predict­ability is not always possible. Tests with the cardini subgroup. All members of the dunni subgroup, except SK, were tested both ways in mass matings to D. acutilabella from Jamaica (stock no. H355.3 iso female). The only island forms that produced hybrids were SV and GR. GR X acutilabella gave 18 females and 12 males; acutilabella X GR gave one male; SV X acutilabella gave four females and five males. All the hy­brids were sterile. Another stock of acutilabella from the Everglades, Florida (No. H415.7c) produced a few hybrids with GU and SV (the only strains tested) and the hybrids were sterile. Drosophila cardini (No. H415.9 cardini) from the Everglades, Florida was tested with BA, PR, GU, SL and JA in small mass matings. No hybrids were produced except GU X cardini produced two pupae. Many of the females were inseminated in each cross but the sperm were usually dead. In the test BA X cardini, four of the eight BA females dissected had a typical insemination reac­tion, the first case encountered in the dunni subgroup. TABLE 7 Results of continued backcrosses to other islands No. FXM TotalF, Per cent males Fertility 1. 19SV X Z3 GU 342, Z3.4 Males sterile Z. ZO SV / GU X 41 BA 601 Z7.1 Males sterile 3. 31 SV / GU/ BA X 45 PR 10 60.0 Males sterile 4. 48 SV/GU/ BA X 43 ST 5 40.0 5. 55 SV/ GU/BA X 54 GU Z14 13.1 Males sterile 6. 14SV/GU/ BA X ZOMA 0 7. 30 SV /GU/BA X 30 SL z 50.0 8. 40 SV /GU/ BA X 40 SV ZZ8 4Z.5 Males sterile 9. ZOSV/GU X 31 SL Z6 Z3.1 Males sterile 10. ZO SV / GU/SL X 50 BA 89 40.4 Males sterile 11. 30 SV/ GU X 43 PR 48 Z5.0 Males sterile 1Z. Z6SV/GU/PR X 50BA Z3 17.4 Males sterile Heed: Island Populations of Drosophila TABLE 8 Mass matings and reciprocals of BA, SV and GR to other members of the cardini group D. cardini Florida X BA and GR Chile X BA and GR D. acutilabella Miami, Fla. X BA and SV* Everglades, Fla. X BA• and SV* D. neomorpha Nicaragua X BA and SV Panama X BA D. parthenogenetica Mexico X BA and.GR Colombia X BA and GR D. polymorpha Trinidad X BA and SV Brazil X BA and SV D. procardinoides Bolivia X BA and SV D. neocardini Colombia X BA and SV Brazil X SV • No hybrids produced except where marked. Hybrids with SV and acutilabella sterile Hybrids with BA and awtila­ bella fertile when backcrossed to BA males. · Table 8 shows the remainder of the crosses with members of the cardini sub-. group. In this set of crosses BA and SV or GR only were chosen for testing. The only tests that produced hybrids were again with acutilabella. An average of four females was dissected for each cross that did not go. Fertilization had oc­curred in most cases but the sperm were usually dead. The stock of acutilabella (No. H415.9 acuti) from the Everglades, Florida is an exceptionally fertile one to many cardini group species and this stock threw fertile hybrid females with BA: acutilabella females X BA males gave 27 females and 16 males and the hybrid females were fertile to BA males. Cytology. There are several corrections and additions to be made to the cyto­logical picture in the dunni subgroup as presented in the first publication. Figure 8 in H. & K. (1959) is not correct in the sense that there are no inversion differ­ences between PR, ST and SV, GR. Also the inversion difference between SV, GR and the central islands (SK, GU, MA, SL and BA) is not 2LB but rather 3RB. In other words 2LB does not exist and the only main inversion difference in the dunni subgroup is 3RB which is fixed in all the central islands. These errors were found when JA was being tested to all other island strains. There are no inversion differences between JA and PR, ST or SV, GR. When JA is tested with the central islands, 3RB shows up in heterozygous condition in every case. As a result of this PR, ST and SV, GR were tested with one another and found to be homozygous. Also the inversion difference between D. acutilabella, from Jamaica, and St. Vincent was previously reported as a small mid-inversion in the X chromosome. The strains of acutilabella from Jamaica, Cuba, St. Petersburg and Floral City, Florida, are identical in the gene arrangement in all chromosomes. The strain of acutilabella used recently is from the Everglades (H415.9 acuti). This strain shows overlapping inversions with very little pairing in the basal half of the X chromosome in hybrids with SV. This one fact illustrates that acutilabella cannot be related by a simple interpretation to the dunni subgroup. The new strain has not yet been tested with other acutilabella strains. There is one inversion that has recently been found heterozygous in St. Kitts. The University of Texas Publication It is a short basal inversion (about 1/7 total length of arm) in the left arm of chromosome II. The inversion attracted attention for study because the homo­zygous standard arrangement (that common to all the islands) was picked up only in very low frequencies at first. A more thorough analysis showed that the SK homozygote and heterozygote were almost three times more frequent in the SK stock than the standard homozygote: SK/SK 75, SK/ST 74, ST/ST 28 larvae. In order to test whether ST homozygous is not as viable as the other arrange­ments, two types of matings were prepared. In the first test the ST gene arrange­ment from Guadeloupe was tested with the SK arrangement from St. Kitts in the hybrids of the mating SK X GU and the reciprocal in one small mass mating each. From seven sample vials of the F4 generation, the total count of the three arrangements fits the expectation on the Hardy-Weinberg equilibrium basi~: SK/SK SK/ST ST/ST* Obs. 41 87 55 Exp. 45.75 91.5 45.75 P=.10 • From Guadeloupe. Of the seven vials, three gave results fitting the Hardy-Weinberg expectation, three gave deviations at the 1 to 5 % level of significance, and one showed extreme heterosis. In the second test the two gene arrangements in St. Kitts, previously made homozygous, were tested in small mass matings of SK/SK X ST/ST and the reciprocal. The data below· represent the total larvae dissected of the F 3 genera­tion from four sample vials: SK/SK SK/ST ST/ST* Obs. 46 77 20 Exp. 35.75 71.5 35.75 P< .01 • From St. Kitts. Of the four vials, two were very unbalanced against ST/ ST and two were balanced. The data indicate that the standard arrangement in St. Kitts is indeed less viable in competition with the inversion (SK). It is tempting to speculate that the value, and thus high incidence, of the new inversion (for it is apparently restricted to St. Kitts) is to shield a semi-vital gene complex. However, the strains have not been independently tested for fertility, fecundity and viability. It is interesting to compare the frequency of inversions in the dunni subgroup and in the cardini subgroup. Of the seven species in the former, two species are heterozygous for one inversion each. Da Cunha et al. (1953) found 6 inversions in polymorpha from three localities and 3 inversions in cardinoides from three localities. A recent survey in this laboratory showed the following interstrain inversions from two localities each of the following species: polymorpha 3, neomorpha 3, parthenogenetica 2 and cardini 1. The cardini subgroup has a higher frequency of inversions than the dunni subgroup even though the number of inversions is low. More information on inversions is reported in the discussion. Heed: Island Populations of Drosophila PHENOTYPES In many of the tests with the dunni subgroup the abdominal patterns of the hybrid females were scored as being intermediate or closer to one or the other parent. The data are too extensive and complex to analyze but a few generaliza-.. tions may be made. It should be recalled that all island stocks are quite uniform in phenotype within themselves. There are in general three classes of hybrid phenotypes. Some of the hybrids are "intermediate and uniform"; hybrids in other crosses are "variable", ranging from one parental type (or almost) to the other parental type (or almost) ; in still other crosses the hybrids show dominance in varying degrees of one parental type. Table 9 shows that when Barbados is one of the parents the hybrids are usually of the first type and when Guadeloupe is one of the parents the hybrids are of the second class. However single individuals of Guadeloupe can apparently be fixed homozygous for color genes since in one case with Grenada all hybrids (48 females) were close to GU and in another pair mating all hybrids (12 females) with Grenada were intermediate and uniform. Also, BA females throw variable hybrids, in mass matings, with ST and SK, and BA males are variable with SL females. Other hybrids that are variable in mass matings are: GR X ST, SL x MA, MA x SL, SL x SV and SV x ST. In general, the populations ST, SK, GU, MA and SL produce variable hybrid offspring when crossed to each other or to the populations of other islands. The island populations that generally give intermediate hybrids with other islands are: PR, BA, SV and GR. There is no relation of the above two classes to the third category: partial dominance. A "variable" island may be partially dominant to another "variable" island or to an "intermediate" island. However an "intermediate" island is partially dominant only to "variable" islands. Also a light phenotype may be dominant to a darker phenotype or vice-versa, and one cross may show dominance but the reciprocal may fall in one of the other two categories. One sequence of partial dominance is interesting: GR>ST>GU>SL>BA> MA in one or both of the TABLE 9 Phenotypes from pair matings No. FXM o. pairs Phenotypes of females 1. STX BA 6 Mostly intermediate and uniform 2. GU x BA 11 Mostly intermediate and uniform 3. BA X SV 4 Mostly intermediate and uniform 4. BA X GR 1 Mostly intermediate and uniform 5. SVx BA 6 Mostly intermed iate and uniform 6. GR X BA 6 Mostly intermediate and uniform 7. GU X SV 4 Variable 8. GU x GR 3 Variable and GR present in low frequency 9. BA X GU 10 Variable but closer to BA 10. SV X GU 9 Variable, SV and GU in low frequency 11. GR x GU 7 Variable and GR present in low frequency All 48 fema les close to GU 12. GR xGU ·1 All 12 females intermediate and uniform 13. GR xGU The University of Texas Publication reciprocal crosses. St. Vincent females are partially dominant in the cross to SK males. There are a few examples in which the hybrids were either lighter or darker than either parental type. For instance GU X .ST produced two females that were darker than either parent. The hybrid phenotypes of Jamaica (strain H356.3d) with other island males illustrate all three classes (Table 5). JA shows almost complete dominance to PR and ST in the north and GR in the south. Hybrids with GU are variable but inter­mediate with SL and BA. Table 10 shows the scoring of phenotypes from the 21 TABLE 10 Variability in F1 hybrid color patterns from pair matings of JA X GU Close to GU (Int.) Intermed ia te (Int.) Close to JA JA With JA (356.3d) Pair 1 Pair 2 Pair 3 7 pairs 8 8 24 17 1Z 10 22 31 24 6 11 23 10 4 7 2 4 0 0 14 0 0 0 Total With JA (356.3f) 11 pairs 57 0 75 2 64 0 23 3 18 63 33 pair matings made up to test for non-disjunction as mentioned previously. Two strains established from single females of JA were used in crossing to GU males (from a multiple stock). The difference between the two strains is striking. Strain "d" shows high variability and strain "f" shows 33% complete dominance in the hybrid females. This is good evidence that the abdominal patterns are under rather strict genetic control. The tests that gave an F2 are recorded in Table 11. There appears to be strong selection for the male parental phenotype in three of the four tests. There is good indication that only one or very few F1 males were fertile in each test. The data from many of the early backcrosses is given in Table 12. Each test is from a mass mating and many of the total numbers of offspring are well below the original parental crosses indicating that selection is strongly operating on the recombinations, and in favor of the phenotype to which the hybrids are crossed. T ABLE 11 Frequency of extreme phenotypes in F 2 No. FX;\l F phenotype females F phenotypes 1 F2 2 1. GU X BA Mostly intermediate 34 50% BA 50% BA to intermediate 2. BA x GU Variable 9 All intermediate 3. BA X SV Mostly intermediate 217 20% SV 5 % close to BA 6% lighter than either 4. SV X GU Variable 222 45% GU 4% SV Heed: Island Populations of Drosophila TABLE 12 Frequency of extreme phenotypes from backcrosses No. fem alesNo. Fx l scored Phenotype o[ females 1. ST/BA X ST 86 30.3% ST 2. ST/ V X ST 90 11.1% ST 3.3% close to SV 3. SV/ST X ST 217 25.3% T 1.3% close to SV 6.0% lighter than either (whitish) 4. GU/ BA X GU 220 21.8% GU 5. GU/ SV X GU 186 39.0% GU 6. GR/ GU X GU 223 39.1 % GU 7. ST/BA X BA 70 22.8% BA 8. GU/ BA X BA 178 21.9% BA 3.4% close to GU 9. BA/ SV X BA 276 37.8% BA 1.1 % darker than either parent 1.1 % lighter than either parent 10. BA/ GR X BA 183 8.9% BA 1.1 % clo e to GR 1 female lighter than either parent 11. GR/ BA X BA 175 25.7% BA 3.4% lighter than either parent 12. V/ST X SV 86 16.3% SV but darker yellow ground color 13. GU/ V x v 111 20.0% sv 9.0% GU 14. BA/ V x v 203 18.2% sv 2.9% close to BA 15. GU/GR x GR 95 31.5% GR 8.4% close to GU 16. GR/GU X GR 182 22.0% GR 15.3% close to GU 1 female lighter than either parent 17. GR/ BA X GR 212 10.0% GR 1.3% very close to BA 3.3% lighter than either parent SuMMARY OF IsoLATING MECHANISMS Island populations that can be tested in the laboratory are unique in that they show genetic differences produced through time and isolation unaffected by selection for or against isolating mechanisms after two closely related populations meet. They show that isolating mechanisms do evolve as a by-product of adapta­tion. Sexual isolation is rather completely absent in the dunni subgroup. This may indicate that behavior phenomena are under a more general genetic control than other mechanisms and may be the last mechanism to evolve in allopatric popu­lations through time. No sexual preference tests have yet been attempted. Gametic mortality is only fairly common. Jamaica males produce no hybrids with all other islands although insemination takes place regularly. St. Lucia males and females show a high frequency of dead sperm with PR and ST to the north and SV and GR to the south. The University of Texas Publication All cases of zygotic mortality in the subgroup are more or less in direct relation to the geography of the islands. This phenomenon is probably not completely fortuitous and illustrates increasing genetic divergence through space (but see Discussion). The best example is the sex-ratio dine exhibited when the Y chromo­some of Guadeloupe interacts with the X chromosome from islands to the south of GU (except Martinique), producing fewer males at each step. Since this was discussed in the previous paper only new information will be added here. Jamaica must be placed at the extreme end of the dine in the sense that the X-JA, Y-GU combination is almost completely lethal. The stage at which most of the males die, at least in the case of SV X GU, is the critical period of emergence from the pupal case and also the pupal stage. The majority of males remain half-way emerged until they die. The remainder die before the pupa case is opened. The Y chromosome from St. Kitts as far as it has been tested behaves similarly to the GUY but does not give such an extreme cross-lethal effect. The Y chromo­some from Martinique has a very extreme lethal effect in combination with the X's of SL, BA, SV and GR resulting mostly in female progeny. Comparisons are summarized in Table 13. T ABLE 13 Per cent males from tests showing cross-lethal effects in relation to geography Males Females SK GU MA SL BA SV GR JA 28.3 26.0; 33.1 * 0 19.4 20.0 8.3 0.9 0.3 9.1 2.8 0.3 ; 0.6* 0.4;1.1* • Two mass matings. Martinique and Barbados females show a maternal effect with the X's of SV and GR in that the sex ratio favors the males and more so with GR than SV. The gene or gene complex in the GR X behaves as a dominant lethal in MA cytoplasm resulting in death of all the female progeny. The phenomenon here compares favorably with the maternal effect in D. montana females to texana male reported by Patterson and Griffen ( 1944) in the virilis group which was a better analysis since the reciprocal hybrids were fertile. The reciprocal crosses (SV and GR X BA males) give a relatively normal sex ratio. The two islands of Barbados and Martinique may be compared in their maternal effect: BA x SV: 63.0% males MA x SV: 62.2 to 70.2% males (2 tests) BA x GR: 72.7% males MAxGR: 91.7to 100.0% males (2tests) Male sterility is by far the most common isolating mechanism in the subgroup. All hybrid males from interspecific crosses are sterile, although there have been several interspecies tests which gave a few second generation hybrids but they in turn were sterile. The one interesting case that did give at least one fertile male in which later progeny were fertile was the SV X GU test as reported in Heed: Island Populations of Drosophila the first paper. The cross has been repeated many times since then and again one fertile male was produced. The results will be presented elsewhere. Complete female sterility is found only in hybrids involving Martinique, how­ever not all the possible combinations were tested. All other islands have at least one combination that produces partially fertile females with males of another interspecific island (Figure 3). There is no well-defined correlation of hybrid female fertility with the geography of the islands. It should be emphasized that prediction is not possible for the combining abilities among the different islands and perhaps this is a special type of isolating mechanism. For instance Martinique males will produce hybrids with GU, BA and SV females separately but not in the "triple" combination females (Table 7). In other words any combination of genes does not work and each island has its own balanced system. ·The genotype as a balanced system has been emphasized by Stone, in Patterson and Stone ( 1952). The isolating mechanisms illustrated above are no different than those already reported for Drosophila but they are proof that such systems may arise in iso­lation as a consequence of general adaptation. (See Discussion.) THE SPECIES CoNCEPT IN THE DuNNI SUBGROUP There are six types of analysis reported on the dunni subgroup in this and the previous publication in order to distinguish similarities and differences between the island populations. The first three are genetic: fertility relations, metaphase configuration, and inversions. The last three concern the phenotype: abdominal color pattern, body size and male genitalia. The six characteristics are not super­imposed upon one another in random fashion but tend to show relationships, and the relations are generally in accord with the geographic position of the islands. Figure 4 shows three main divisions. The first and sharpest division separates Puerto Rico and St. Thomas from all other islands in all characteristics except for the standard gene arrangement which is common also to SV, GR and JA. Actually the metaphase configuration of these two islands are different, which Fert ile Met a phase In ve r sio n Ge nital ia S iz e Pot t er n PR S T ~ ~ --- --- ~ - ~ ~ --­-- SK GU D MA SL BA SV GR D D D D D D D C=:J D D D [] [==:J JA D D Fm. 4. Comparative characteristics in the dunni ubgroup. Equal width of bars indicates equality within each class. The University of Texas Publication is not indicated in the figure. The fertility relations beyond a fertile F 1 are too diverse to include and will be discussed later. The hybrid inviability effects have already been discussed and are shown to be in accord with the spatial relations of the islands. The second division separates the central islands from St. Vincent and Grenada in all characters but the metaphase type of 2 V's, a rod and ·a dot. The third division separates Jamaica in all characters except the gene arrangement and size. It is thus not difficult to establish three good species on the five most differ­entiated islands: D. dunni in Puerto Rico and St. Thomas; D. similis in St. Vin­cent and Grenada; D. belladunni in Jamaica. The five central islands offer more of a challenge. To a taxonomist working with only morphological characters, they would seem to represent a complex of subspecies that show a dine in the amount of pigment, since there are no good differences even in the male genitalia (Figure 1 ) . There is not one instance in Drosophila taxonomy that is familiar to the author where closely related species cannot be differentiated by genitalia. A good case in point is the taxonomically difficult tripunctata group in Middle and South America which consists of 45 described species and many more undescribed forms. The shape and coloring of the body are so similar in many instances that identification is possible only by dissecting out the genitalia. As far as has been tested, there is a direct correlation between the ability of two forms to hybridize and produce fertile progeny with the similarity in male genitalia and vice-versa. Patterson ( 195 7) mated 11 population stocks of the tripunctata group in ' 110 crosses. Two of the stocks had identical male genitalia and they were the only stocks that produced fertile hybrids in reciprocal crosses. It is well known that sibling species in the willistoni group, obscura group and melanogaster group can all be differentiated by genitalic structure. Biologically, however, the complex consists of one island that is genetically isolated from all the others, Martinique, and two islands that are completely inter­f ertile, St. Kitts and Guadeloupe. The central islands at once represent good biological species but also good morphological subspecies (Mayr, 1948). The subspecific category implies inter-fertility, allopatry and, usually, ecological similarity. Although many of the hybrid females are semifertile, continued out­crossing to other island males produces fewer and fewer offspring and in some cases the systems become so unbalanced that no offspring survive. A trinomial designation would clearly indicate that each form lives on a separate island. The ecological requirements are probably quite similar on each island and this is reviewed in the discussion. . The species category, of course, implies rather strict reproductive isolation, which is typical of most of the island crosses. It is more important to emphasize the hybridization ability of those forms that can be tested in the laboratory than to emphasize their morphological and ecological characteristics. The similarity in phenotype, except for abdominal pattern, and apparent similarity in ecology among the populations of the five islands is regarded not as a result of present day gene exchange, but as a result of the absence of any selection pressure to make the change necessary. Accordingly, the complex of five islands is divided Heed: Island Populations of Drosophila into four species: D. arawakana, St. Kitts and Guadeloupe; D. caribiana, Mar­tinique; D. antillea, St. Lucia; and D. nigrodunni, Barbados. On the other hand, the fact that six of the seven species in all of the 10 islands are potentially ·Capable of exchanging genes through partially fertile hybrid females with at least one other species, and in the case of arawakana and similis with four other species, has little taxonomic value and only serves to show close­ness of relationship which can easily be deduced from examination of the pheno­types in the first place. Also the recent evidence that a certain strain from the Everglades, Florida, of D. acutilabella, which is not included in the dunni sub­group, produces partially fertile hybrid females with Barbados, reduces the value of this type of fertility for taxonomic purposes. D. acutilabella is the pivotal species in the cardini group since Stalker ( 1953) has shown that it will hybridize with more cardini group species than any other form. If acutilabella were in­cluded in the dunni subgroup, then this category, if based completely on back­cross fertility relationships, would vanish. However, the dunni subgroup is a useful division genetically and morphologically and the line is drawn to exclude acutilabella. If subspecies are to be considered, it would be more legitimate to establish them within the three species that are distributed on six islands. Every one of the six islands has its own characteristics in its behavior in crosses with other island populations. St. Thomas differs from Puerto Rico in having a J-shaped X chromosome without satellites instead of a rod with satellites. St. Kitts differs from Guadeloupe by a heterozygous short inversion in the basal left arm of chromosome II. There is no assurance, however, that the inversion does not also exist in the island of Guadeloupe. The St. Kitts Y chromosome does not give as severe a sex-ratio with other island X's as does the Guadeloupe Y. The St. Vincent X chromosome differs from the Grenada X by being semi-lethal instead of lethal in Martinique cytoplasm. Also, the populations of four of the six islands can be differentiated by visible morphological traits. St. Thomas has a grayish abdominal ground color and Puerto Rico is more yellowish. St. Kitts females have much shorter wings than body while Guadeloupe females have only slightly shorter wings than body. St. Vincent and Grenada are similar in appearance. To name the inhabitants of these two islands as separate subspecies is an obvious act of "splitting" to any taxonomist, but it has value in showing that they are in some way different from one another. Accordingly, Puerto Rico becomes D. d. dunni and St. Thomas is D. d. thomas­ensis. Guadeloupe becomes D. a. arawakana and St. Kitts becomes D. a. kittensis. St. Vincent is D. s. similis and Grenada is D. s. grenadensis. Most assuredly the conclusions reached in the discussion of what constitutes a subgroup, a species and a subspecies in the island populations will not be accepted by all readers, but the author has attempted to be consistent. Nature will not be pigeon-holed and therein lies the species problem. DISCUSSION Perhaps the best approach to a description of the genetic characteristics of the dunni subgroup is to compare them with closely related forms on the South The University of Texas Publication American continent and in Middle America. The one comparison most signifi­cant in the cardini group is that the large land masses have evolved seven species while at least eight species are confined to the West Indian islands, seven in the dunni subgroup plus D. acutilabella. D. cardini ranges from Chile to Mexico, the Greater Antilles, and some of the islands of the Lesser Antilles, and thus is considered to be a continental form. Obviously the total land mass of the islands is only a very small fraction of the mainland indicating the effects of isolation and initially small populations to promote differentiation. There is a little information on the ecology of the cardini group, especially in respect to "animals of different kinds". In the islands, the dunni forms are the subdominant Drosophila, being outranked only by D. willistoni. Table 2 in the first publication shows a surprising amount of consistency in the frequencies of the dunni forms taken in the month of January in baited forest samples from three of the four islands: Puerto Rico 22.2 % ( 351 ) ; St. Kitts 25 .2 % ( 515) ; St. Lucia 23.8% (579); Barbados 5.3% (322). The numbers in parenthesis are the total numbers of Drosophila collected in each sample. Also, the fact that no two dunni forms are found on the same island attests to their ecological similarities. Two population samples from the continent, both taken in December, will suffice to illustrate that the other cardini group species are ecologically different from the island species and from one another. One baited sample from Trinidad, which is effectively a part of the South American continent, taken in a Mora forest at Sangre Grande yielded 1,136 individuals of which D. polymorpha, the most common of the five species in the group in Trinidad, made up 1.7%. It was the eighth species in abundance out of 29. The other sample taken in the Arima Valley of Trinidad yielded 1,273 individuals of which polymorpha made up 1.2%. It was the ninth most abundant species out of 15. The dunni forms have found their niche in the islands. If interspecific compe­tition is operating to produce more cooperative genotypes, it is less severe on the islands and is typical of island ecology in general. It is also quite possible that willistoni has replaced the dunni forms as the dominant species by later invasions on the islands. ' The taxonomic difficulties encountered in the island types have been discussed earlier. There are no taxonomic difficulties with the cardini group on the main­land, at least in morphology of the male genitalia, Figure 1, even though almost all members look superficially very similar in size, shape and general coloring. The unusual species relationships of D. cardini, acutilabella and belladunni were pointed out in the previous paper. In Florida, cardini and acutilabella are very difficult to separate; in fact, there is no reliable criterion by which the females may always be distinguished. In Jamaica, acutilabella and belladunni are morphologically very similar and the females cannot be distinguished with perfect accuracy. However, D. cardini can easily be identified from belladunni. It is the opinion of the author that both sets of species may be considered as sibling species. Apparently different genotypes are responding in a similar manner to the same environment. If it is true, then the evidence is good that the particular abdominal patterns are adaptive. Comparison of frequencies of inversions between island and mainland forms, as far as they have been tested, show that the former are more depai~perate of Heed: Island Populations of Drosophila inversions, as one would expect in island populations. A much better comparison of this type has been made in the West Indies for D. willistoni by Dobzhansky ( 195 7). Smaller islands usually have fewer inversions. The isolating mechanisms isolating the mainland forms, as far as they have been tested, are more rigid and efficient than those in the dunni subgroup. Zygotic mortality may prove to be the most common type. Sexual isolation is also preva­lent (Stalker, 1953; Streisinger, 1946). Streisinger tested interspecifically five strains of cardinoides, two strains of polymorpha and one strain of neocardini. The species were sexually isolated from one another. Stalker found cardinoides and neocardini females sexually isolated from each other and from cardini, parthenogenetica, and acutilabella males. D. cardinoides females were insemi­nated by polymorpha males but no hybrids were produced. D. acutilabella females from Florida produced various numbers of hybrids with cardinoides, cardini, parthenogenetica and polymorpha. The only fertile hybrids were those with cardinoides males. The only species with which acutilabella is sympatric in these tests is cardini and, thus, it is not surprising to find them not as sexually isolated. The other combinations usually produced dead eggs or larvae. Table 8 shows the tests of Barbados, St. Vincent and Grenada with other members of the cardini group. No hybrids were produced in the majority of cases, although insemination usually took place in all crosses. Mr. David Futch has recently hybridized many newly obtained stocks in the cardini group (Futch, this Bulletin) . He has confirmed that acutilabella will readily hybridize w1th the majority of the other species. However, if only the tests between the six species restricted to the mainland are considered, then five tests out of 30 produced hybrids. The species on the mainland, whether sympatric or not, are genetically more isolated than those which inhabit separate islands. Four of the five tests that were fertile included the cardinoides-procardinoides­parthenogenetica triad. In island populations that have no closely related forms within each island (Jamaica and Grenada are exceptions), there is no selective pressure within these species to reinforce the evolution of isolating mechanisms. Hence, those that are found take on special meaning, that is, are either by-products of selection pressure of a different kind (in this case of the physical environment) or show genetic divergence by distance along a migration route, since several examples of gametic mortality and all examples of zygotic mortality (cross-lethal genes) are correlated with the spatial relations of the islands. The initial genetic diver­gence in the X chromosome causing cross-lethality and semi-lethality in the male· hybrids, once initiated, might tend to accumulate with increasing distance by the addition of new alleles or modifiers as differently balanced systems were created on each newly invaded island. If the migration of the dunni subgroup­started from Grenada and went north through the islands to Guadeloupe and St. Kitts and Barbados was colonized from St. Lucia, then the sex ratio dines are in agre~ment with the migration route. Puerto Rico and St. Thomas fit only the dine with Martinique females, but this appears to be fortuitous since the males from the two stocks usually throw a low frequency of hybrid males with most other island females. Puerto Rico and St. Thomas are now believed to have been colonized from either Grenada or St. Vincent because they do not own inversion The University of Texas Publication 3R. Also, the fact that St. Lucia females show complete gametic mortality with Puerto Rico and St. Thomas on the one hand, and Grenada and St. Vincent on the other, fits the picture. In this sense, then, increasing genetic divergence with distance could account for the results. If the migration routes were more disjunct in the middle and southern islands, then the dines in sex ratio must be attribut­able to the selection of the physical environment. Mainland (1942) demonstrated a general east-west dine in sexual isolation and fertility between the several subspecies of D. macrospina and of D. sub­funebris in continental United States. He attributes these differences to genetic differentiation by distance. Patterson and Stone (1952) discuss this and other cases in more detail. A more direct indication that selection is acting in a general step-wise fashion is the increase in dark pigment in the southern islands. Rensch (1960) has stated that geographical variation can at times be a consequence of successive loss of alleles in the course of migrations and expansions of small peripheral populations without selection operating at all. He gives as an example the reduction from polymorphism to monomorphism in color pattern in the terrestrial snail, Papuina, in New Britain, as the probable course of migration is traced from the west to the eastern tip of the island, where the snail does not occur. That such events occur is undoubtedly true, but it is also probably true that marginal populations are under more extreme selection pressures and cannot afford major gene polymor­phism (Carson, 1959). Ifthe dunni forms spread by island-hopping, as they probably did, then at each step variability was lost. That variability has been regained and rebalanced as each island population expanded is evidenced by the isolating mechanisms and the variability in phenotypes in some of the inter-island hybrids. Ernst Mayr has laid the theoretical foundation for explaining why marginal island populations are not only different from one another and from large central island or main­land populations but different in a new way (1954). Such populations are not just variations on a single theme but are qualitatively different even when the environment is ostensibly similar among the small islands. Mayr emphasizes the drastic change in the genetic environment that must take place when a few indi­viduals or one fertile female is removed from an area of active gene flow to an area where inbreeding suddenly becomes the only means for survival. There results a sudden change in the selective value of many genes simultaneously, a veritable "genetic revolution". This may be true, but if a single female, pre­viously fertilized by one male, is transplanted, there still remains a maximum of four alleles at each locus to give variability by crossing over and recombination. The establishment of the dunni forms as a separate subgroup is indication that its members are distinct from the remainder of the cardini group. The long para­median markings that extend the length of each tergite on the abdomen is a qualitatively new character. D. belladunni is an exception. It owns an abdominal phenotype found only in the Jamaican D. acutilabella. Mayr is, of course, correct in stating that a genetic revolution need not take place every time a few indi­viduals are isolated. Among the factors which will allow it to occur at the most rapid rate, however, Mayr lists two which can be measured in the cardini group. The first is a "shift into a vacant ecological niche somewhat different from the Heed: Island Populations of Drosophila parental one". The previous discussion on ecology has shown this to be true. The second condition is "if the parental population was particularly variable and subject to much gene flow." The second condition could also have been true in this case. The cardini group shows more externally visible phenotypic variability than any other species group in the genus. Species like D. polymorpha, neomorpha and parthenogenetica are polymorphic for the abdominal pattern which is con­trolled by one or a few sets of alleles. However, at least in the case of polymorpha, the marginal populations in northern South America, Trinidad and Grenada are monomorphic light. It has recently been demonstrated by Mr. Patrick Blake of this laboratory that the light alleles in Trinidad, B.W.I., and Santa Marta, Colom­bia, are dominant to the dark allele in at least two strains from Brazil. In most areas of Brazil, three phenotypes exist, controlled by a single pair of alleles with no dominance ( da Cunha, 1949) . A full report on the study will be reported later. D. cardini and acutilabella possess light and dark phenotypes. Stalker (1953) reports that in acutilabella the pattern is apparently controlled in a simple Men­delian fashion, the heterozygote being very close to the light form. Tests with acutilabella in this laboratory do not confirm his findings and it was not possible to establish any rules for inheritance. Sturtevant ( 1921) had the same trouble with tests on D. cardini. His conclusions were that the conditions under which the larvae develop determined the phenotype. The problem remains unsolved. Inheritance of the color pattern in two races of D. neocardini has been reported by da Cunha ( 1955). The situation here resembles more closely the condition in the dunni subgroup. The two races are morphologically distinct and uniform (not polymorphic), but the first generation hybrids are variable from one parental type to the other. However, the lighter parental type is favored in much higher frequency than the darker form (23.5% vs. 0.8% females, n=498). The data may be directly compared to the variable hybrids obtained from the test JAX GU in Table 10. The light phenotypes are 20% and the dark phenotypes (=JA) are 0.4% of the total F1 (238). Da Cunha's conclusion is that the two races are homozygous for the gene or genes controlling pigmentation but are heterozygous for factors which act as modifiers in the hybrid genotypes. The same conclusions apply to several of the dunni tests. Thus, there is now living in South America a species that is variable in the same way the dunni species are variable. A closer scrutiny of neocardini is thus warranted. Figure 1 demonstrates that the three most similar species to the island forms according to male genitalia are polymorpha, neomorpha and neocardini. Exami­nation of the salivary chromosome arm 3R indicates that each of the three species is homozygous for the standard arrangement as found in the two southern islands, the two northern islands and Jamaica. Hybrids of acutilabella with St. Vincent and Barbados show that no inversion in 3R is present in the former case but that the 3R inversion is present with Barbados. Acutilabella then also has the standard gene sequence. Also, hybrids of this species with polymorpha show no inversion in 3R. Therefore inversion 3RB probably originated on the central islands. This is a good indication that the central islands were not the first ones to be invaded either from the north or the south and is important since the central islands from Guadeloupe to St. Lucia are apparently geologically older than the islands on The University of Texas Publication either end of the chain in the Lesser Antilles (Beard, 1949). The degree of band­ing differences on 3R between the three species, as compared to a photograph of the 3R of St. Vincent, showed that neomorpha was most distinct (many bands could not be matched), polymorpha was next in distinctness (several to many bands could not be matched), and that neocardini was strikingly similar to St. Vincent (almost band for band) . The geographic origins of the three strains are: neomorpha, Honduras; polymorpha, Trinidad; and neocardini, Brazil. The 3R arm has not yet been compared with the other species in South America. Futch (this Bulletin) has contributed more evidence on the relationship of the n£omorpha-polymorpha-neocardini triad of species to one another and to the dunni subgroup. He has found that of all the possible crosses, polymorpha and momorpha will produce sterile hybrids and that polymorpha and neocardini will produce at least partially fertile female hybrids with acutilabella. This con­firms the order of relationship written above. Futch has tested all members of the cardini subgroup with belladunni, dunni and arawakana. Aside from acutilabella which produced hybrids with arawakana and belladunni, neocardini is the only other species in the cardini subgroup that hybridized with the dunni subgroup. In two tests neocardini males produced eight sterile females and two sterile males with belladunni females. This confirms our interest in neocardini. It is possible that the island complex could have arisen from neocardini or a species with the same characteristics. The northern limit of neocardini may be the Amazon basin. It is not found in Colombia except at Leticia, on the Amazon, and has not been taken in Trinidad. The island founders could have been carried by floating vegetation from the Amazon, or possibly the Orinoco, to the islands. James Bond, a well-known West Indian zoogeographer, in personal communica­tion, described the very strong current he encountered, while swimming off the island of Tobago, that came from eastern South America and headed straight for Grenada. Beard ( 1949) has found the vegetational affinities of the Lesser Antilles to be with both Puerto Rico and South America. In relation to the latter he says that much beach drift on the islands today is derived from the rivers of eastern South America. The few recent and fossil mammals present in the Lesser Antilles according to Simpson ( 1956) have their relations with the present Trinidad and eastern Venezuela fauna, not the Greater Antilles. In a sense then, this fulfills the last of Mayr's condition, "that the parental population be subject to much , gene flow", in that marginal populations may not be involved. Invasion of the Lesser Antilles from Puerto Rico is, of course, also possible, as discussed in the previous paper, but does not now seem as likely in view of the similarity in banding pattern in the 3R chromosome between D. similis (St. Vincent) and D. neocardini (Brazil) . T1ie interpretation of a "genetic revolution" must take into account the amount of time involved for later differentiation by selection to occur, as Mayr well knew. The factors acting together, however, should produce qualitative differences. Recent experimental studies by Dobzhansky and Pavlovsky (1957) and Dobz­ hansky ( 1960) indicate that genetic drift and natural selection may operate simultaneously, at least if the original material has more than the usual store of genetic variability. The authors have shown that replicate experimental popula­ tions initiated by a small number of founders (20) are more variable after 20 Heed: Island Populations of Drosophila generations, for the frequencies of two different gene arrangements from differ­ent localities, then replicate populations started with 4000 flies, even though in both sets of experiments the cages are saturated after the first generation. This shows how each new founder group, arriving from a large population and ex­panding in numbers, can be genetically different after many generations even in the same environment. In different words, selection has a greater variety of genotypes from which to choose, in relation to initial population size, in the small founder cages. In this respect genetic drift and natural selection can operate simultaneously. The authors have speeded up the evolutionary process by endowing the founders with more genetic variability than probably ordinarily exists in nature. The original hybrids were derived from Texas and California. Shifts in inversion frequencies do not constitute a genetic revolution, but they indicate how each new group of founders can be genetically different. There is some indication that genetic drift has operated in the dunni subgroup for the dine in abdominal pattern in the islands is not a perfect dine. The Martinique population has almost as much dark pigment as Barbados but the island is located north of St. Lucia, whose population is only slightly darker than the population from Guadeloupe. Genetic drift must be more important in permanently small populations. Good observational data by Lowe ( 1955) on the vertebrate faunas of three islands in the Gulf of California of different sizes showed that the smallest island contained the highest frequencies of endemic species and subspecies, although the popu­lation sizes were not measured. That only one major genetic revolution took place on the islands is evidenced by the basic similarity in phenotypes of most of the species, Jamaica being an exception. The similarities between the island forms and neocardini have been pointed out. In this sense the island species were preadapted for the type of differ­entiation under discussion. D. polymorpha inhabits Grenada, but it is still the same species. Naturalists have observed for years that well-spread contiguous continental populations show the effects of selection only as geographical races while island populations show more drastic differences. It follows that gene flow, migration, is a very strong deterrent to selection. However, the observation that environ­mental selection may actually be different on islands was pointed out by Mayr (op. cit.) because the environment is different, especially the biotic environment. Islands have long been known to harbor unbalanced floras and faunas. The rarity of "animals of different kinds" (Andrewartha and Birch, 1958) is most evident in the Lesser Antilles. The largest number of species in the family Drosophilidae collected on any of the islands was 34 on the island of Guadeloupe in six collecting days. It is not uncommon to collect four to five times this number of species on the mainland in the same amount of time. The consequences of these conditions, lack of migration and the absence of closely related species, are that island populations show the effects of selection by the physical environment more openly than do continental populations. The dine in phenotype in the dunni subgroup within approximately 700 miles illu­strates the point. Also, the shorter wings on the flies from St. Kitts and Barbados, the two most "exposed" islands, illustrates the point. E. B. Ford and his co­ The University of Texas Publication workers ( 1960, for review and references) are demonstrating that selection has a tight control on the number of spots on the underside of the hind wing of the Satyrine butterfly, Maniola jurtina. Also, the observation that on five small ecologically different islands in the Isles of Scilly the modal number of spots differs from one island to the next while on the three large islands, ecologically similar, the number of spots is the same, illustrates that selection is now operating. The spots are apparently under control of multiple factors, which makes it more difficult for genetic drift to be important. The multitude of isolating mechanisms that have arisen in the dunni subgroup indicates that most of the island populations are genetically closed systems. That they are even more genetically closed than isolating mechanisms show is revealed by the fact that in no case has it been possible to obtain a phenotypically variable stock after several generations in any of the backcrosses even when the hybrid males are fertile after the first backcross. The great majority of hybrid combi­nations apparently do not work. The most extensive data concerns hybrids with St. Vincent and Guadeloupe. In several generations of mass breeding after the initial backcross to either St. Vincent or Guadeloupe, the total population cannot be distinguished from the parent to which it was backcrossed. Also, in two cases out of 10 attempts in the initial cross of SV females to GU males at least one hybrid male was fertile in each case. In both cases the hybrid stocks were carried many generations and were phenotypically like GU. When the hybrids were out­crossed to both parents, they were fertile to GU and sterile to SV. Also, the inversion 3R in both instances was homozygous for the GU type. This study will be reported in more detail at a later date. Carson (1959) has used the term homoselection for small isolated populations. The phenomenon in the dunni subgroup may be termed interspecific homo­selection, which undoubtedly is not a special property of island populations, but it is a constant property. This is in marked contrast to recent hybridization experiments with a few continental species of Drosophila. For instance, Mettler (1957) was able to establish a hybrid colony between D. arizonensis and mojav­ensis that had a combination of each species' inversions. Not all combinations worked but a new balance was maintained in some of them. Bruneau ( 1955) found the same phenomenon in several species of the virilis group. Carson ( 1959) reports that two out of four hybrid stocks in D. bocainensis and parabocainensis were interchromosomally balanced even after two and one-half years in the laboratory. A study of heterosis is supposedly not legitimate between species that show some sterility as some of the above cases do, but there is a little data for two island populations that are completely fertile. Table 14 shows, insofar as total number of offspring is a measure of heterosis, that none exists between St. Vincent and Grenada, either in mass matings or pair matings. The data were extracted from the first publication and the crosses were not made as a heterosis experiment. Each stock, originally composed of many males and females, had undergone about six generations in the laboratory before the tests were made. The island populations then may be at a different level of fitness than many continental populations. They are not so flexible among themselves; they are under stronger environmental selection pressure since among other things there is little or no migration from distances greater than within each island to disrupt Heed: Island Populations of Drosophila T ABLE 14 Total number of offspring of St. Vincent and Grenada and their hybrids Two mass matings each of Pair matings show ing20 fcrnales and 20 males average per pair sv x v 2702 SV x SV 105.9 GR x GR 24-83 GR x GR 197.2 SV x GR 25 18 V x GR 92.3 GR X SV 2754 GR X SV 160.6 selection. The amount of time in isolation is of course important, but it is also important in the above examples from mainland populations. That migration can disrupt selection has been abundantly illustrated by Brncic (1954), Wallace (1955) and Vetukhiv (1957, and earlier) for large continental populations such as D. melanogaster and D. pseudoobscura. The interpopulation hybrid heterosis, demonstrated by these authors, is usually lost in the next generation because of crossingover and recombination in the original hybrids. The influx of new genotypes by migration into a locally coadapted gene pool undoubtedly produces more variability but at the same time breaks up the integrated systems in all probability in proportion to the rate of migration. Wasserman (1960) has excellent cytological data that the 12 species in the mulleri complex of the repleta group initiated their divergence as a mosaic of small semi-isolated desert populations. The unusual sharing of six inversions between the 12 species has led Wasserman to propose an important function of inversion differences between populations, i.e., to reduce the effects of ill-adapted migrants by suppressing recombination. The lack of heterosis between St. Vincent and Grenada is more similar to the report of Stone et al. (1954) for small isolated populations of D. novamexicana and D. hydei in the southwest. Stone measured egg production and egg hatch in interpopulation hybrids and found that egg hatch was improved, but there was no heterosis for total egg production in comparison to the controls. Also later generations did not fall below the parentals in egg hatch. Stone concluded that the egg-laying pattern is at its optimum in these small inbred populations, thus making it difficult to improve by outcrossing. The present author has collected novamexicana as well as the island flies and knows first hand that the latter have far larger populations. In fact, the largest collection of novamexicana amounted to only 58 individuals in several days' trapping. One, however, cannot measure the degree of homozygosity in each population so readily. St. Vincent and Gre­nada are two of the island types that give intermediate and uniform hybrid color patterns to most other islands. In this sense they are more homozygous than Guadeloupe for instance. Even so, they are no doubt more heterozygous than the extremely small, shrunken, isolated populations of novamexicana. Different kinds of genes and balanced systems are effective under different conditions. The absence of heterosis for egg production in both types of populations, so different in history and structure, may have basically the same underlying cause. They are under more strict control of natural selection since both types of populations are isolated populations. There is one direct measure of the results of natural selection in island popula­ The University of Texas Publication tions of various sizes. Stone et al. (1957) and Stone and Wilson (1958) measured the fertility, fecundity and viability of D. anana.ssae from four of the Marshall Islands and Ponape by inbreeding, crossbreeding and doublecrossing the progeny from three successive years of collecting (1955 to 1957) after thermonuclear tests on one of the island atolls (Bikini in 1954 and other lesser tests in 1956). Although the measurements and times of collecting were not as favorable as desired, there are some very dramatic recoveries in viability (egg development) within one year. Rongelap atoll, which received fallout, changed from 45 per cent (1955) to 83.7 per cent (1956) egg development. Bikini atoll, where the tests were made, changed from an average of 46.1 to 63.4 per cent egg develop­ment. The control islands did not change this much. Interestingly enough, egg production between the three years did not change in any population but inter­island hybrids did show a little improvement in two out of the three years. It appears that the island populations in the cardini group have been drastically rebalanced only once, at the initial invasion from the mainland. The best evi­dence for this is the presence of a species on more than one island. For instance, St. Thomas was probably colonized from Puerto Rico (because its metaphase is more different than Puerto Rico's in relation to other dunni forms) but St. Thomas is completely fertile with Puerto Rico. The interfertility also between St. Kitts and Guadeloupe, St. Vincent and Grenada illustrates that an invasion from one island to an uninhabited island does not necessarily imply a drastic genetic rebalance on the part of the invader. Therefore, the populations that are sterile between themselves must have had more time to become different, barring differences in island ecologies. The fact that no two species in the dunni subgroup exist on the same island may be explained in several ways. Either migration between islands is taking place and interspecific homoselection occurs for the phenotype of the island being invaded in each case or the ecological properties of each island population are so similar that competition would not allow the invaders to get a foothold. Perhaps both phenomena are at work, but in the case of Martinique, ecological barriers seem to explain the situation, since this island is sterile to all other island species and thus interspecific homoselection could not occur. The situation is in contrast to the secondary island invasions believed to have occurred in the evolution of the drepaniids in the Hawaiian Archipelago (Amadon, 1950) and the geospized finches in the Galapagos Islands (Lack, 1947). Both authors agree that most of the species initiated their divergence on separate islands even though at present there are 15 species of honeycreepers on Hawaii itself and 10 species of Darwin's finches on James, Albermarle and Indefatigable. It is also true that insects in general are more sedentary than birds. Zimmer­ man (1948), in his analysis of the insects of the Hawaiian Islands, has recorded that most of the species endemic to the archipelago are restricted to only one island or small part of one island. Of the 222 endemic species of c;arabid beetles, only five occur on more than one island. Eighty per cent of the anobiid beetles ( 140 species) are confined to one island. Zimmerman believes the island having the largest number of species of a particular group is the one in which the original colonization occurred. The restricted microenvironn;ients of many of the insects of these islands accounts for the secondary isolation necessary for divergence. Heed: Island Populations of Drosophila The Drosophila under study in the Antilles have a much greater ecological lati­tude for they can be found in domestic as well as undisturbed sites in many areas of each island. There is a little evidence that the presence of closely related species restricts genetic variability. Da Cunha et al. (1959 and earlier) have found that the frequencies of inversions in D. willistoni are less in areas where several or all of the three other sibling species are present, especially in high proportions, than in ecologically similar areas, where they are absent or in low proportions. Ives ( 1954) suggested that the reason for the drop in frequencies of homozygous lethal second chromosomes from 65.3% (337 chromosomes) to 51.1 % (131 chromo­somes) over a period of nine years in D. melanogaster in F1orida may have been caused by the observed rise in frequency of the sibling species, D. simulans, from 15% (300 males) to 86% (990 flies). The dunni forms have no sibling species with which to compete in any of the islands except Jamaica allowing in part a greater genetic, and thus ecological, latitude for each island type. This in turn supports the thesis that each form is restricted to its own island because of eco­logical barriers. Incidentally, the drosophilid fauna in the Hawaiian Islands is extraordinarily rich. Zimmerman (1958) has estimated perhaps 300 species may eventually be described by Hardy and others, and rightly challenges: "Where else has such a drosophilid fauna developed?" In August, 1958, in Boquete, Chiriqui Province of Panama, Dr. Marvin Wasserman and the author collected 230 species in eight collecting days in an area of about two square miles, an area smaller than any of the main islands in Hawaii. It is believed that the study of island populations of Drosophila has disclosed the relative importance of two major types of selection pressure, among many, that lead to differentiation. The first type is selection by the physical environ­ment, that is, temperature, humidity, solar radiation, etc., the type of selection that produces differences in relation to the geography of the landscape inhabited by a similar group of organisms. Possible examples of such selection have been discussed previously, but the dine in phenotype is the best example. The dine in sex-ratio may also be an example. The examples do not appear to be fortuitous because of their regularity. The second type of selection pressure is that caused by the presence of closely related species. This type is absent on most of the islands and imparts to the populations certain characteristics usually otherwise present in species on. the mainland. The first is the complete lack of sexual isolation be­tween island populations. The second is the similarity in male genitalia among four genetic species in the five central islands. The third is the apparent similarity in ecology in most of the island populations. If the three latter characteristics are present solely because of recent habitation of the islands by the dunni subgroup, then it not only strengthens the thesis that physical environmental selection is operating but that it is working at a rapid pace. Anderson ( 1960) has been able to isolate other selective forces operating on island populations of salamanders in the San Francisco Bay region. The lack of predators and competing species has allowed these forms to build up dense popu­lations which resulted in reproductive inhibition and this in turn permits new modes of selection to operate by regulating population size. The University of Texas Publication Parenthetically, the role of peripheral populations in originating really new characters (evolutionary novelties) has recently been emphasized by Carson and Mayr. Oceanic islands harbor numerous examples of a radical change in function and sometimes structure. The genus Drosophila is not without representatior. since an entirely new ecological niche has recently been discovered (verifying an old museum label) for a species inhabiting Mona Island, west of Puertd Rico. Drosophila carcinophila is an ordinary-looking fly which apparently lives its entire life cycle on a cave-inhabiting land crab. The unusual story has been reported by Wheeler, 1960. SUMMARY Ten islands in the West Indies harbor seven closely related species and three of the species are newly described. The complex of species is a natural one and is known as the dunni subgroup of the cardini species group of Drosophila. Three two-island groups contain members which can be distinguished usually morpho­logically and always genetically in tests to other island strains; They are described as subspecies. The similarities and differences between each island population are described and discussed on a genetic and morphologic basis. Comparisons are made with the cardini group species in Middle and South America. The original differentia­tion of the dunni subgroup can be explained by the thesis presented by Ernst Mayr that the change in the genetic environment in founder individuals and the subsequent different kind of selection pressure on island populations promotes novel differences. Some of the present-day differences between the island strains are believed to result from selection pressure of the physical environment. Many of the similarities are believed to result from the absence of closely related species. The island populations are characterized as large, fairly heterogeneou,s but closed genetic systems. The majority of interspecies crosses produce sterile males. The one intersubspecies cross tested gave no heterosis for total number of progeny. The results are compared to several contrasting types of continental and other island Drosophila populations. ACKNOWLEDGMENTS This paper is presented in the memory of Professor John Thomas Patterson. Just 10 years ago in June, Dr. Patterson sent his last fledgling graduate student on his maiden collecting trip for Drosophila. Then, as now, the wealth of material contained a high potential for understanding organic diversity, reward enough for anyone, as he well knew. The author expresses his appreciation to Dr. Th. Dobzhansky for reading the manuscript and for making valuable suggestions. However, he is not to be held responsible for the opinions expressed. I also wish to thank Mrs. Marion Weiden, Mrs. Jean Russell and Mr. Don Harrington for laboratory assistance. This work is supported by the National Science Foundation. LITERATURE CITED Amadon, D. 1950. The Hawaiian Honeycreepers (Aves, Drepaniidae). Bull. Am. Mus. Nat. Hist. 95 Article 4. N.Y. Heed: Island Populations of Drosophila Anderson, P. K. 1960. Ecology and evolution in island populations of salamanders in the San Francisco Bay region. Ecolog. Monog. 30 No. 4: 359-385. Andrewartha, H. G., and L. C. Birch. 1954. The distribution and abundance of animals. Univ. of Chicago Press. Chicago. Beard, J. S. 1949. The natural vegetation of the windward and leeward islands. Oxford at the Clarendon Press. Brncic, D. 1954. Heterosis and the integration of the genotype in geographic populations of Drosophila pseudoobscura. Genetics 39: 77-88. Bruneau, H. L. 1955. Ph.D. dissertation. Univ. of Texas. Carson, H. L. 1959: Genetic conditions which promote or retard the formation of species. Cold Spring Harbor Symp. Quant. Biol. 24: 87-104. Da Cunha, A. B. 1949. Genetic analysis of the polymorphism of color pattern in Drosophila polrmorvha. Evolution 3: 239-251. -----.. 1955. Sohre duas racas de Drosophila neocardini Streisinger. Rev. Brasil. Biol. 15: 117-125. Da Cunha, A. B., D. Brncic, and F. M. Salzano. 1953. A comparative study of chromosomal polymorphism in certain South American species of Drosophila. Heredity 7: 193-202. Da Cunha, A. B., T. Dobzhansky, 0. Pavlovsky, and B. Spassky. 1959. Genetics of Natural Populations XXVIII. Supplementary data on the chromosomal polymorphism in Drosophila willistoni in its relation to the environment. Evolution 13: 309-404. Dobzhansky, T. 1960. Evolution and environment. In Evolution after Darwin, Vol. 1. Ed:S. Tax. Univ. Chicago Press. Chicago. -----. 1957. Genetics of Natural Populations XXVI. Chromosomal variability in island and continental populations of Drosophila willistoni from Central America and the West Indies. Evolution 11: 280-293. -----, and 0. Pavlovsky. 1957. An experimental study of interaction between genetic drift and natural selection. Evolution 11: 311-319. Ford, E. B. 1960. Evolution in progress. In Evolution after Darwin. Vol. 1. Ed. S. Tax. Univ. Chicago Press. Chicago. Fosberg, F. R. 1951. Ecological research on coral atolls. Atoll Research Bull. 1: 6-8. Futch, D. G. Hybridization studies within the cardini species group of the genus Drosophila. This bulletin. Heed, W. B. 1957. A preliminary note on the cardini group of Drosophila in the Lesser Antilles. Univ. Texas Pub. 5721: 123-124. -----, and N. B. Krishnamurthy. 1959. Genetic studies on the cardini group of Drosophila in the West Indies. Biological Contributions. Univ. Texas Pub. 5914: 155-179. Ives, P. T. 1954. Genetic changGs in American populations of Drosophila melanogaster. P.N.A.S. 40: 87-92. Lack, D. 1947. Darwin's Finches. Cambridge Univ. Press. London. Lowe, C. 1955. An evolutionary study of island faunas in the Gulf of California, Mexico, with a method for comparative analysis. Evolution 9: 339-344. Mainland, G. B. 1942. Genetic relationships in the Drosophila funebris group. Univ. Texas Pub. 4228: 74-112. Mayr, E. 1948. The bearing of the new systematics on genetical problems. The nature of species. Adv. in Genetics 2: 205-237. -----. 1954. Change of genetic environment and evolution. In Evolution as a Process. Ed. J. Huxley, A. C. Hardy, and E. B. Ford. London. Allen and Unwin Ltd. Mettler, L. E. 1957. Studies on experimental populations of Drosophila arizonensis and Dro­sophila mojavensis. Univ. Texas Pub. 5721: 157-181. Metz, C. W. 1916. Chromosome studies on the Diptera. III. Additional types of chromosome groups in the Drosophilidae. Amer. Nat. 50: 587-599. The University of Texas Publication Patterson, J. T. 1957. A study of interspecific hybridization between members of the tripunctata group of Drosophila. Univ. Texas Pub. 5721: 7-14. -----, and A. B. Griffen. 1944. A genetic mechanism underlying species isolation. Univ. Texas Pub. 4445: 212-223. -----, and W . S. Stone. 1952. Evolution in the Genus Drosophila. The Macmillan Co. New York. Rensch, B. 1960. Evolution above the Species Level. Columbia Univ. Press. New York. Simpson, G. G. 1956. Zoogeography of West Indian land mammals. Am. Mus. Novitates No. 1759: 1-28. Stalker, H. D. 1953. Taxonomy and hybridization in the cardini group of Drosophila. Ann. Ent. Soc. Amer. 46: 343-358. Stone, W. S., M. L. Alexander and F. E. Clayton. 1954. Heterosis studies with species of Dro­sophila living in small populations. Univ. Texas Pub. 5422: 272-307. Stone, W. S., F. D. Wilson, J. T. Neuenschwander, and T. G. Gregg. 1957. Factors affecting viability in Drosophila ananassae populations from the Marshall Islands. Section V in Genetic studies of irradiated natural populations of Drosophila. Univ. Texas Pub. 5721: 291-316. Stone, W. S., and F. D. Wilson. 1958. Genetic studies of irradiated natural populations of Drosophila II. 1957 tests. P.N.A.S. 44: 565-575. Streisinger, G. 1946. The cardini species group of the genus Drosophila. Jour. N. Y. Ent. Soc. 54: 105-113. Sturtevant, A. H. 1921. The North American species of Drosophila. Carneg. Inst. Wash. Pub. 301. 150pp. Townsend, J. I., and M. R.. Wheeler. 1955. Notes on Puerto Rican Drosophilidae, including descriptions of two new species of Drosophila. Jour. Agric. Univ. Puerto Rico 39: 57-64. Vetukhiv, M. 1957. Longevity of hybrids between geographic populations of Drosophila pseudoobscura. Evolution 11 : 348-360. Wallace, A. R. 1902. Island Life. Macmillan and Co. N. Y. 3rd edition and revised. Wallace, B. 1955. Interpopulation hybrids in Drosophila melanogaster. Evolution 9: 302-317. Wasserman, M. 1960. Cytological and phylogenetic relationships in the repleta group of the genus Drosophila. P.N.A.S. 46: 842-859. Wheeler, M. R. 1960. A new genus and two new species of Neotropical flies (Diptera; Dro­sophilidae). Ent. News 71: 207-213. Williston, S. W. 1896. On the Diptera of St. Vincent (West Indies). Trans. Ent. Soc. London 1896 (3): 253-446. Zimmerman, E. C. 1948. Insects of Hawaii. Volume 1: Introduction. Univ. Hawaii Press. Honolulu. -----. 1958. 300 species of drosoph~la in Hawaii? A challenge to geneticists and evolu­tionists. Evolution 12: 557-558. X. The Problem of Phylogeny In the Genus Drosophila1 LYNN H. THROCKMORTON2 INTRODUCTION Some time ago the writer began investigations directed toward determining the usefulness of biochemical characteristics in Drosophila taxonomy. Before this problem could be resolved satisfactorily it was necessary to determine to what extent biochemical and morphological characteristics followed the same patterns of behavior during evolution, and to what extent these patterns could be inter­preted to produce a model for phylogenetic analysis consistent both with morpho­logical and biochemical evidence and with our present concepts of the dynamics of adaptive change. In determining patterns of morphological evolution the cen­tral problem is, of necessity, the problem of phylogeny. One cannot follow the behavior of characteristics, either morphological or biochemical, during evolution until phylogenetic relationships have been shown with reasonable accuracy. Several phylogenies for the species involved are available, but they show con­siderable inconsistency, both as to method of derivation and as to result. It has, therefore, been necessary to re-evaluate phylogenetic relationships within the genus and to develop a conceptual model for phylogenetic analysis. The present paper deals with the distribution and phylogenetic significance of certain morpho­logical features. Biochemical aspects will be covered separately (see Throck­morton, and Throckmorton and Magalhaes, This Bulletin) . In order to evaluate the behavior of morphological characteristics during evo­lution it is necessary to have available a phylogeny based on characteristics other than morphological ones. Several phylogenies based on cytological evidence have been prepared for a number of species groups in the genus. The most extensive ones are those for species of the virilis group (Stone, et al., 1960) and for species of the repleta group (Wasserman, 1960). It is thus possible to relate distributions of distinct morphological characteristics to the evolutionary sequences which produced them. When this is done, inferences can be drawn regarding the be­havior of the genetic systems producing the various phenotypes during the spe­ciation events initiating the phyletic lines under consideration. The results from these analyses indicate useful methods for phylogenetic study, and these methods can be applied to investigate distributions of morphological features throughout the genus. The morphological characteristics cho.~en for this study are, for the most part, those used widely in Drosophila taxonomy. Several, however, are described in detail for the first time here. The features investigated are as follows: 1) morpho­logy of paragonia and relationship between paragonia and vasa deferentia, 2) morphology of ejaculatory bulb and of ejaculatory apodeme, 3) morphology of the testes, 4) characteristics of the first and sixth abdominal sternites in the male, 1 Supported in part by grants NSF-G 4999, and RG-6492 (NIH); additional laboratory support came from the Rockefeller Foundation. 2 Present address: Department of Zoology, The University of Chicago. The University of Texas Publication 5) characteristics of the spermathecae, 6) morphology of the ventral receptacle, 7) characteristics of the Malpighian tubules, 8) arrangement of branches of the anterior pupal spiracle, and 9) characteristics of the egg filaments. Several addi­tional features have been noted also, and these will be commented on briefly but not considered in detail. The characteristics of these structures provide a large amount of anatomical detail, adequate both for correlations with known phylogenies and for determin­ing the main outlines of evolution within the genus. The phylogeny arrived at through their use is broadly consistent with those previously produced by other means (Sturtevant, 1942; Hsu, 1949; Patterson and Stone, 1952; Malogolowkin, 1953; Okada, 1956 and 1958), and, in addition, the conceptual framework upon which it is based allows data from species in closely related genera to be inte­grated with data from species in the genus Drosophila. When this is done, rela­tionships between these genera and certain phyletic lines within the genus are indicated, thus opening the possibility for extending this type of phylogenetic analysis to problems of relationships between genera and between groups at higher taxonomic levels. One major conclusion indicated by the analysis of this material is that the majority of groups are derived from populations "heterozygous" for the genetic determiners of various alternate forms of a given characteristic. Thus, for any single morphological feature there has been a considerable amount of "parallel" evolution. A consequence of this is that two or more phenotypes, expressing sub­stantially homologous genotypes, may arise separately in phyletic lines, them­selves derived from a single ancestral population. This appears to complicate, but really adds precision to, the phylogenetic analysis, and the problems involved will be discussed in detail later. This consideration should, however, be kept in mind during the presentation of the data. · MATERIALS The materials used in this investigation have come from several sources. The great majority of species and strains were from those maintained as stocks at the University of Texas Laboratories. Some of the stocks have been provided by other workers, and their sources will be acknowledged elsewhere in this publica­ tion. In addition to utilization of laboratory strains, it has been advisable t() include other species which cannot be reared in the laboratory. For the most part these have been collected by the author, either in the vicinity of Austin, Texas, or in the vicinity of Riverside, California. The species and strains used, together with their classification, University of Texas collection numbers (where applicable) and collection localities are listed in the Appendix. The classification followed is that of Wheeler ( 1949b), Patterson and Stone (1952), Okada (1956) and others. Two changes involving subgeneric reference have been made to conform with the results of the present study. D. nannoptera, previously assigned to the subgenus Soplwphora (Wheeler, 1949b), has been transferred to the subgenus Drosophila. Species of the bromeliae group, previously assigned to the subgenus Sophoplwra (Patterson and Stone, 1952) are likewise transferred to the subgenus Drosophila. Species of both of Throckmorton: Phylogeny in Drosophila these groups are related most closely to species of the virilis-repleta section of this subgenus. Only two minor changes have been made. D. guttifera is here included as a member of the quinaria group, and D. aureata has been removed from the repleta group. METHODS Individuals to be used for dissection have come from laboratory cultures or from local collections. When laboratory cultures were used flies to be dissected were taken directly from the stock, and no effort was made to obtain flies of uniform age. Age may have an effect upon the appearance of some structures, and these effects will be indicated during the discussion of the individual char­acteristics. When flies were from local collections, age was also unknown. Indi­viduals from only one strain of a given species were used, and the size of the sample varied, depending on the characteristic being investigated. Therefore, few conclusions regarding intraspecific variability of a trait can be drawn from the present data. From ten to twelve individuals were used to determine the internal characteristics of the male. Approximately five individuals were used when the internal characteristics of the female were recorded. If details were observed in cleared material, as for the ejaculatory bulb in the male and spermatheca in the female, general characteristics were noted from uncleared material and only one specimen was cleared ana used to prepare the figures. Additional specimens were cleared when unusual features were noted. Samples for other characteristics varied between five and ten individuals. The methods used for demonstrating internal structures have been quite simple. The fly was dissected in Drosophila Ringer's solution (Ephrussi and Beadle, 1936) . The relationships in situ of the various organs were noted, and then the structures of particular interest were separated and figures made of their major features. The low powers of a binocular dissection microscope were generally adequate, both for dissection and for observation of structure. Most of the organs were observed without further treatment. To determine the character­istics of the ejaculatory bulb and of the ejaculatory apodeme it was necessary to clear the material. This was done by transferring the bulb, generally with external genitalia and ejaculatory duct attached, to a drop of phenol on a slide. In a short time, often less than one minute, the cellular envelope and the contents of the bulb were dissolved away leaving the transparent, thinly-chitinized lining with the ejaculatory apodeme attached to it. During this treatment the shape of the bulb was retained, except for the shape of very long and fine caecae ( ejacu­latory sac diverticula of Rosenblad, 1941 ), when present. In the cleared condition it was a simple matter to rotate the organ to any position desired and to determine its structure in some detail, even when using low magnifications. Only rarely was it necessary to check the bulb at higher magnifications by the use of the com­pound microscope. The figures of the ejaculatory bulb show the organ after clearing in phenol. It was also necessary to clear the spermathecae, and this was done by methods similar to those used for the ejaculatory bulb. The female reproductive tract was dissected free from other internal organs. The ovaries, and generally the egg The University of Texas Publication guides, were removed, and the major portion of the vagina with its attached spermathecae, parovaria and ventral receptacle was transferred to a drop of phenol on a slide. Clearing time varied greatly from species to species, although it generally was considerably less than five minutes. The spermathecae, paro­varia and ventral receptacle were observed, first with the binocular dissecting microscope and then with the low power of the compound microscope. Figures for the spermathecae were made from the cleared material as seen through the compound microscope. Figures of ventral receptacles involve observations from both cleared and uncleared material. Where pertinent, the source of certain details, whether from cleared or uncleared material, will be indicated during the description of the individual characteristics. The morphology of the Malpighian tubules was checked from both males and females. About five individuals of each sex were dissected. The general features of the Malpighian tubules were noted by using the magnifications of the dissect­ing microscope. The posterior Malpighian tubules were then transferred to a drop of Drosophila Ringer's solution on a slide, and they were observed with the compound microscope to determine whether the tips were apposed or fused. Free tips could be identified readily without use of the compound microscope. Larvae and pupae were taken directly from stock cultures. Only third instar larvae were used. For observation of pupal spiracles the pupae can be of any age, so long as the anterior spiracle is completely everted. Pupae taken from the medium rather than from the walls of the culture vial were preferable for ob­serving spiracle characteristics. The anterior spiracle must always be checked to be sure that the branches are completely everted. This is not hard to do so long as the necessity for the check is recognized. Some previously published figures (e.g., busckii in Patterson and Stone, 1952) appear to be from specimens in which the branches of the anterior spiracle had failed to evert, or had everted only partially. In some species a large proportion of the pupae have spiracles which. fail to evert completely. Figures were made only from pupae with the branches completely everted, regardless of the proportion of everted and uneverted spira­ cles in the total sample. Eggs were taken directly from fresh cultures. At the time of observation free-hand sketches were made from the material, and note was taken of any unusual features. Within each figure relative pro­ portions were, as nearly as possible, as observed. No effort was made to draw the figures to any set scale. Thus, strict size comparisons between individual species cannot be made from these figures. General comparisons can be made between species in the genus Drosophila, and in certain cases (the ejaculatory bulb and the spermathecae) the relative sizes compare very closely with those observed. In some cases it has been necessary to exaggerate the size of certain structures, and these cases will be noted individually. All of the figures show semi-diagrammatic representations of the structures involved. The organs, par­ ticularly the testes and paragonia, are often twisted around each other in such a way as to make a photographic representation of their relationships quite uninformative . . THE CHARACTERISTICS Direction of evolution is an immediate problem in any phylogenetic study. Throckmorton: Phylogeny in Drosophila Most of the characteristics selected for use in this investigation show directional change, and the more primitive form of expression of each trait can generally he inferred, either on theoretical grounds or by analysis of the distribution of the various forms of the trait among related groups. In this section the various characteristics will be described individually, and the data for each characteristic will be presented. Here the major emphasis will be on the description of the different forms of the trait and on the features which indicate its evolutionary status (primitive vs. derived) . The emphasis here will also be on broad general similarities which tend to unite major groups. This overemphasizes uniformity, but discussion of types of variation will make up a major part of a later section, and the significance of the variants will be emphasized at that time. The data from each trait will also be summarized in the form of a phylogeny. These initial phylogenies should be interpreted only as working models, although they may also have a certain usefulness as pictorial keys. Integration of all the data for all characteristics will be deferred until the next section, and a general phylogeny will be developed there. Data from other genera and families will be included, where applicable, to support conclusions regarding the evolutionary status of a characteristic. Internal Reproductive Structures of the Male-The terminology to be used follows, for the most part, that of Patterson ( 1943) or of Okada ( 1956). Figure 1.1 is labeled to indicate the major features. The male reproductive system of melanogaster has been described elsewhere (Miller, 1950), and since other species, from other families of Diptera as well as from the genus Drosophila, differ from melanogaster only in detail, it is not necessary to discuss the system more fully here. In the case of material from other families there may be some doubt regarding homology of certain structures. For the most part the structures involved are of little significance for the present purpose and are included in the figures primarily to give an indication of the range and type of variation en­·countered outside of the family Drosophilidae. Figures 1 and 2 show internal structures for males from other families of Diptera and from the family Drosoph­ilidae, respectively. With reference to terminology, one major departure from usual usage has been made. The common practice has been to refer to the "inner" and "outer" coils of the testes in species where the testes have a spiral form. Stern (1941a, 1941b) has described the growth of the testes in several .species of Drosophila, and it is clear from his description that the testis proper makes up only the outer coil. The so-called inner coil of the testis is a part of the vas deferens caused to coil mechanically by the coiling of the testis following its attachment to the vas d~­ferens during development. It therefore seems advisable to drop the term inner coil, and to use the term vas deferens to apply to the duct leading from the base of the paragonia to the testis (i.e., to the beginning of the outer coil) . This duct may be differentiated in several ways, and its characteristics will be commented on later. The term testis will be used only to refer to the outer coil. Morphology of paragonia and relationship between paragonia and vasa de­ferentia-Aside from the extensive work of Hori (1960) describing the male internal organs of Calyptrate Muscoid flies, little information is available re­garding the types and distributions of the internal reproductive organs of the The University of Texas Publication -o.e.d. .2 & .\ -e.b. -p.e.d. .6 12 Frc. 1. Internal male reproductive structures from representatives of various families of (Aulacigastridae) ; .11 ) Periscelis annulata (Periscelidae) ; .12) Diastata vagans (Diastatidae) . obscurella (Ephydridae) ; .4) Phy tobia sp. (Agromyzidae); .5) Sepsidimorpha secunda (Sepsi­dae); .6) M umetopia occipitalis (Anthomyzidae) ; . 7) T haumatomyia glabra; .8) Oscinella coxendix (Chloropidae); .9) Prochyliza xanthostoma (Piophilidae) ; .10) Aulacigaster leucopeza (Aulacigastridae); .11 ) Periscelis annulata (Periscelidae); .12) Diastata vagans (Diastatidae) . Symbols in Figure 1.1 indicate: t-testis; v-vas deferens; p-paragonium; a.e.d.-anterior ejaculatory duct; e.b.-ejaculatory bulb; p.e.d.-posterior ejaculatory duct. male. Although the sample from other Acalypterate families is not large, the general types encountered are consistent enough, and compare closely enough with types seen by Hori (op. cit.) among the Muscoids, that they may be used as Throckmorton: Phylogeny in Drosophila a point of departure in discussing the direction of evolution for the different char­acteristics. As a general rule the paragonia are rather thin, but they may be quite variable in length (Fig. 1). Within the family Drosophilidae, and particularly within the genus Drosophila, the paragonia become unusually robust and exhibit highly specific orientation and morphology (Figures 2-14). The vas deferens in other families is almost always a short, rather heavy duct which may or may not be pigmented. Aside from the Drosophilidae and the Diastatidae (Figure 1.12), Frc. z. Internal male reproductive structures from representatives of the family Drosophilidae . .1 ) Zapriothrica dispar; .Z) Gitona americana; .3) Rhinoleucophenga obesa; .4) Dro ophila victoria; .5) D . pseudoobscura; .6) D. tolteca; .7) D. yakuba; .8) D. equinozialis; .9) D. busckii; .10) D. aldrichi; .11 ) D. bi/urea; .JZ) D. subbadia. The University of Texas Publication both closely related, a basal fusion of the vasa has been seen only in the Ephy­dridae. In at least one of the Ephydrids figured (Figure 1.3) it is probable that the apparent vas deferens is not homologous with that seen in other families. In this instance the testes appear to have established a new connection with the ejaculatory duct, and the old vasa remain as a Y-shaped remnant (still showing basal fusion, however) . Among the Chloropids at least two genera (an example of only one, Figure 1.7, shown) show no recognizable vasa and the testes attach directly to the bases of the paragonia. Other genera of Chloropids (e.g., Figure 1.8) show the more usual configuration. These variants, however, do not seri­ously detract from the impression that short, unfused vasa are the general con­dition. This is the type seen in such species as victoria (Figure 3.1), and, since the other types seen in the genus Drosophila are highly distinctive (e.g., Figures 2.8, .12) the assumption that this is a primitive type for the group seems most reasonable. The basal fusion seen in busckii (Figure 2.9) apparently repre­sents a modification of this type, but since similar fusions of the vasa are evident in other, more primitive genera (Figures 2.2-.3), it cannot be considered as being derived within the genus. Aside from the development of specific relationships between the vasa and the paragonia, which will be discussed shortly, the major change of the vasa with evolution has involved lengthening and regional differentiation. The differentia­tion generally takes the form of an expansion of the region of the vas just prox­imal to the testis proper, and it is seen in almost all species in the genus. In forms with spiral testes it is this region which is coiled mechanically by the asymmetri­cal growth of the testis. In no cases does the number of coils in this region exceed the number of coils in the testis, and it is generally considerably less. This region is pigmented if the testes are pigmented, apparently due to the migration of pig­ment cells from the testes (Stern and Hadorn, 1940). This differentiated region has been named the inner coil of the testis by some, and the seminal vesicle by others (Miller, 1950; Okada, 1956). In species where the vasa remain short, differentiation involves the whole structure (e.g., Figure 13. 7). In others, the proximal portion appears as a thin, hyalin duct while the distal portion, generally one-fourth to one-third the total length, is expanded and has an opaque, granular appearance. Presumably, it is this portion which represents the original vas, with the proximal, hyalin duct representing the modification added during evolution. This last assumption can, however, only be verified by more detailed, compara­tive histological and developmental studies. Until such studies are made, applica­tion of a particular name such as seminal vesicle, to the differentiated region seems premature. If the term seminal vesicle were generally used, it would have to be applied in some cases to the entire duct between the bases of the paragonia and the testis. In others it would apply only to a part of this duct. It seems best to def er naming this region until its functional and evolutionary significance throughout the family is known. Within the genus Drosophila the most extreme development of this region is seen in species from the immigrans group and in species belonging to the quin­ aria section of the subgenus Drosophila (see Patterson and Stone, 1952, p. 81, for general phylogeny). Many other species in the subgenus Drosophila, par­ ticularly within the repleta group, also have this region strongly developed. Using Throckmorton: Phylogeny in Drosophila the methods of simple dissection, no sharp distinctions based on this feature can be made within the genus, although it seems probable that more detailed investi­gations, and particularly a histological study, would show significant patterns of distribution. One apparent exception to this general statement involves the type of vas development seen in some members of the hydei subgroup of the repleta group. The most extreme example is seen in bi/urea (Figure 2.11) where there is no distal swelling of the vas, and instead the vas in this region is thrown into fine, regular coils. One other species in the genus, nannoptera, shows this same type of vas development. D. neohydei and eohydei show this to a limited degree, and hydei does not show it at all, although its vasa show little differentiation and are almost uniform in diameter throughout. D. castanea strongly resembles hydei in this respect. D. nigrohydei has the usual, moderately differentiated vasa. The most conspicuous and most useful details of structure for the vasa and the paragonia are shown in Figures 3 to 12 for the genus Drosophila and in Figure 13 for species from related genera. Although changes in the major components, i.e., the vasa and the paragonia, have, in the evolutionary sense, developed inde­pendently of each other, it is most convenient to treat them both at the same time. As has been said earlier, the type of vas development seen in such species as victoria (Figure 3.1) represents a primitive form, and the abbreviated phy­logeny shown in Figure 14 is based on this assumption. The members of the subgenus Pholadoris (Figures 3.1-.7) show either the primitive, short, pigmented form of the vas or a simple modification of this type. The paragonia are rather variable, but almost all are relatively long and slender. · They are usually folded at least twice, and most often are folded three or more times. In this case it is not possible to specify a .distinct primitive type. Paragonia in other families are too variable, as are those from other genera in the family. The great majority of types within the genus Drosophila are themselves distinc­tive. Those seen among the species of the subgenus Pholadoris . resemble types seen in other families and in more primitive members of the family. The as­sumption most consistent with all available data is that the types seen in victoria and pattersoni are near the primitive, although types seen in some species of the virilis group (Figures 4.25-.28) and in Phloridosa species (Figures 6.10-.12) may be equally primitive if not more so. When both the vasa and the paragonia are considered, the subgenus Pholadoris is placed at the base of the phylogeny of the genus (Figure 14) . Several types are seen in the subgenus Sophophora. In none of these is there an association between the vasa and paragonia such as is seen in the other major phylogenetic branch of the genus. The first type is that of populi (Figure 14.15) , showing the short, pigmented vasa. It has paragonia which are, at present, distinctive in the genus, unless those of busckii (Figure 3.11) represent a variant of this form. The paragonia are quite robust, more so than any others in this subgenus, and are bent so that their ends point in opposite directions. It is of interest to note that Okada (1956, Figure 34, p. 62) shows a figure of Chymomyza nigrimana in which the paragonia appear to be of this general type. In species of the obscura group (Figures 3.13-.22) the vasa are all short and pigmented, and the paragonia have a characteristic twist which usually results in their ends being directed anteriorly. This same type is seen in species from The University of Texas Publication FrG. 3. Morphology of paragonia and relationship between paragonia and vasa deferentia in Drosophila species. Subgenus: PHOLADORIS coracina group Subgenus: SoPHOPHORA victoria group .1 victoria .2 pattersoni .5 cancellata .6 lativittata . 7 novopaca Subgenus: HrRTODROSOPHILA .12 populi obscura group obscura subgroup .13 miranda bryani group .8 duncani .14 pseudoobscura .3 bryani .9 pictiventris .15 persimilis .10 thoracis .16 anibigua latifasciaeformis group Subgenus: DoRSILOPHA .4 latifasciaeformis .11 busckii Throckmorton: Phylogeny in Drosophila the saltans and willistoni groups (Figures 4.1-.24; Figures 14.10-.11). In some members of the obscura, saltans and willistoni groups the paragonia are reduced in size. In the sturtevanti subgroup of the saltans group the paragonia may ap­pear almost as vestiges. Most species of the melanogaster group lack the twisted paragonia seen else­where in the subgenus, although both ananassae and bipectinata approach this condition, or are intermediate between this and the type seen in populi. rfhe paragonia are generally large and show a second fold distal to the major arch (Figures 14.12-.13; 3.23-.33; bipectinata has not been included in the figures and is substantially as in ananassae). All species of the melanogaster group have the testes located anterio-poste~iorly rather than laterally (Figure 2. 7). In cases where the vasa are not fused to form a common duct basally there is generally a crossing of the vasa. For melanogaster this cross is clearly shown by Miller (1950, Figure 38A). Stern (1941b) describes this cross as being due to asym­metric growth of the testes. The left testis unites with the left vas, its continued growth causes the vas to be swung to the right, and the testis comes to lie in the anterior part of the abdomen. The opposite is true for the right testis. Even in melanogaster, however, the cross is not always seen since the two vasa occa­sionally appear to arise dorso-ventrally rather than laterally, or, in other species, they may be fused basally. The anterio-posterior orientation of the testes is therefore more diagnostic for the group than are the crossed vasa. Among species in the saltans and willistoni gr9ups the testes are located later­ally, but their free ends either cross or lie parallel to each other and point in opposite directions (e.g., Figure ~.8). Basally the vasa cross at least once and often twice. This cross is shown for nebulosa by Patterson ( 1943) . In these spe­cies the mass of testes and paragonia is tightly packed and strongly bound to­gether with tracheae. This makes the demonstration of the details of the system very difficult, and dissections were directed only toward determining the presence of crossed vasa. It should be possible to determine the cause and significance of the double cross of the vasa, but no attempt was made to do that at this time. Superficially, it would appear that the species of the saltans and willistoni groups exhibit the same general phenomenon seen in species of the melanogaster group. They would differ mainly in having . "completed" the cross, so that the larval right and left testes are seen as left and right testes respectively in the adult. A more detailed investigation will be required to show whether or not this is actually the case. In the saltans, willistoni and melanogaster groups the vasa may or may not be fused basally to form a common duct. In the saltans and willistoni groups the affinis subgroup .17 affinis .18 algonquin .19 narragansett .20 tolteca .21 athabasca .22 azteca melanogasler group melanogaster subgroup .23 simulans .24 melanogaster .25 yakuba takahashii subgroup .26 takahashii ananassae subgroup .27 ananassae montium subgroup .28 rufa .29 nikananu .30 serrata .31 auraria .32 seguyi .33 kikkawai The University of Texas Publication ~@,f}) ~~ @p &ffJ ~~~~t ~~·~~~· ~ ~~0f? ~ Frc. 4. Morphology of paragonia and relationship between paragonia and vasa deferentia in Drosophila species. Subgenus: SoPHOPHORA willistoni group .1 equinoxialis .2 paulistorum .3 tropicalis .4 willistoni .5 fumipennis .6 nebulosa .7 sucinea .8 capricorni .9 changuinolae .10 pseudobocainensis sahans group parasaltans subgroup .11 subsaltans .12 parasaltans cordata subgroup .13 neocordata elliptica subgroup .14 neoelliptica .15 emarginata sturtevanti subgroup .16 sturtevanti .17 milleri saltans subgroup .18 1 usaltans .19. nigrosaltans .20 pseudosaltans · .21 austrosaltans .22 prosaltans .23 septentriosaltans .24 saltans Subgenus: DROSOPHILA virilis group .25 virilis .26 americana .27 novamexicana .28 littoralis .29 ezoana .30 montana .31 flavomontana .32 lacicola .33 borealis Throckmorton: Phylogeny in Drosophila M~~ ~~~ .4 .5 MM~~ ·~~ ~Jmb~ ~ , Fie. 5. Morphology of paragonia and relationship between paragonia and vasa deferentia in Drosophila species. Subgenus: DROSOPHILA robusta group .11 species D .6 lacertosa melanica group· .7 colorata immigrans group .1 micromelanica .8 sordidula .12 hypocausta .2 melanica .9 robusta .13 spinofemora .3 paramelanica .14 immigrans .4 euronotus annulimana group Subgenus: HIRTODROSOPHILA .5 nigromelanica .10 gibberosa .15 histrioides The University of Texas Publication ffl .11 .12 M .15 FIG. 6. Morphology of paragonia and relationship between paragonia and vasa deferentia in Drosophila species. Subgenus: DROSOPHILA .6 nannoptera .8 aracea .13 peruviana funebris group bromeliae group .14 species H .1 macrospina .7 species I .15 species G .2 subfunebris .16 tumiditarsus .3 funebris polychaeta group .9 polychaeta Subgenus: PHLORIDOSA carbonaria group .10 species 0 .5 carbonaria miscellaneous species .11 species Q nannoptera group .4 carsoni .12 species P Throckmorton: Phylogeny in Drosophila ~~~ ~~~ ~ ~~~ ~~~ ~~~ ~~~~ FIG. 7. Morphology of paragonia and relationship between paragonia and vasa deferentia in Drosophila species. repleta group .6 fascioloides .13 mulleri .7 moju .14 longicomis fasciola subgroup .8 mojuoides .15 arizonensis .1 fulvalineata mulleri subgroup .16 mojavensis .2 fasciola .9 tira .17 martensis .3 coroica .10 pachuca. .18 stalkeri .4 pictura .11 nigricruria· .19 hamatofila .5 pictilis .12 aldrichi .20 eremophila The University of Texas Publication M~~ m~~ MJ~ ffl " ~ ~~~ ~~M© FIG. 8. Morphology of paragonia and relationship between paragonia and vasa deferentia in Drosophila species. repleta group .7 anceps .14 limensis .9 peninsularis .15 repleta mulleri subgroup mercatorum subgroup .16 canapalpa .1 buzzatii .10 paranaensis .17 melanopalpa .2 pegasa .11 mercatorum not assigned to subgroup .3 meridionalis melanopalpa subgroup .8 serenensis .4 promeridiana .12 fulvimacula ungrouped species .5 meridiana .13 fulvimaculoides .6 aureata ~~ ~ .2 ¥ ~ .4 .5 ~ ~ ~ ~ Fie. 9. Morphology of paragonia and relationship between paragonia and vasa defei-entia in Drosophila species. replete group .5 hydei mesophregmetica group canelinee group .1 0 gaucha hrdei subgroup .6 canalinea .11 pavani .1 bifurca .7 paracanalinea ungrouped species near replete group .2 nigrohydei dreyfusi group .12 castanea .3 eohydei .8 camargoi .13 species F .4 neohydei .9 briegeri fused section is short when present (e.g., Figures 4.3-.11) . In the melanogaster group the common duct may reach considerable length. It is longest in ananassae and rufa (Figures 3.27-.28). A similar fusion has already been men­tioned for busckii and for two Plwladoris species. Since these species have the The University of Texas Publication FIG. 10. Morphology of paragonia and relationship between paragonia and vasa deferentia in Drosophila species. ungrouped species near cardini group testacea group .14 falleni .7 putrida .15 phalerata .1 species K .8 testacea .16 species J .2 species L .17 occidentalis rubrifrons group macroptera group .18 tenebrosa .3 parachrogaster .9 submacroptera .19 subquinaria .4uninubes .10 macroptera .20 transversa ungrouped species quinaria group .21 palustris .5 sticta .11 innubila .22 subpalustris pallidipennis group .12 quinaria .23 guttifera .6 pallidipennis .13 rellima Throckmorton: Phylogeny in Drosophi.la \M) D¥D UfU M lM1 00 ~ ~ Fw. 11. Morphology of paragonia and relationship between paragonia and vasa deferentia in Drosophila species. tripunctata group .6 unipunctata guarani group .7 trapeza .13 guaramunu .1 mediodiffusa .8 bandeirantorum .14 guaraja .2 albicans .9 paramediostriata .15 griseolineata .3 albirostris .10 mediostriata .16 subbadia .4 tripunctata .11 mediopictoides .17 guarani .5 mediopunctata .12 crocina primitive, short vasa, the fusion here may only be analogous to that in Sopho­phoran species. As will be seen shortly, many species in the subgenus Drosophila also have the vasa fused basally. The fusion of the vasa seen in species of the The University of Texas Publication ®) .I fAro .6 FIG. 12. Morphology of paragonia and relationship between paragonia and vasa deferentia in Drosophila species. cardini group .6 neocardini calloptera group .1 dunni . 7 parthenogenetica .12 ornatipennis .Z belladunni .8 acutilabella .13 calloptera .3 nigrodunni .9 procardinoides .14 schildi .4 polymorpha .10 · cardinoides .5 neomorpha .11 cardini melanogaster group, however, may be of a type distinct from that seen in species of the subgenus Drosophila. In at least some of the melanogaster group species the two vasa only tightly adhere to each other. In such cases they can be sepa­rated and then they are seen to be' truly fused only in a very short basal section. In some species of the melanogaster group there is an extreme expansion of Throckmorton: Phylogeny in Drosophila the anterior end of the ejaculatory duct. This is most prominent in takahashii (Figure 3.26), moderate in members of the melanogaster subgroup, and slight or absent in members of the montium and ananassae subgroups. This feature is also seen in some members of the willistoni group (Figures 4.5-.6; 4.8-.10). It is, at best, only slightly developed in members of the saltans group. This feature is present in some members of the subgenus Pholadoris (Figures 3.1-.2, .7) and in a few scattered species in the subgenus Drosophila. In these cases, however, its expression is not as extreme as in some species of the melanogaster group. In the subgenus Drosophila the number of types is too great to allow inclusion of all of them in the space available for a pictorial phylogeny. Figure 14 shows only some of the major types. Others will be pointed out below. In this subgenus, none of the vasa are of the primitive type, and evolution has been toward the acquisition of a specific relationship between the vasa and the paragonia. The final configuration arrived at has been determined, in part, by the presence or absence of basal fusion of the vasa. Two major phyletic lines are indicated within the subgenus. One line leading to such forms as mediostriata (Figure 14.6) has involved increasing length of the fused section of the vasa and an in­creasingly regular association between the. vasa and the paragonia. There has also been a shortening and thickening of each paragonium to form a high arch with slight or no indication of a second fold. In the other section (Figures 14.2-.3) there has been almost no basal fusion of the vasa, and there has been an increasingly regular association between the vasa and paragonia. In this sec­tion shortening of paragonia is not pronounced. Indeed, sine a primitive type cannot be surely specified, it is not improbable that an increase in length may have been involved. Some forms from this section also show a reduction in the ~~~fa~~ ~­.~TflJ ~(jf) * Fw. 13. Morphology of paragonia and relationship between paragonia and vasa deferentia in species from other genera of Drosophilids ..1) Chymomyza amoena; .Z) C. aldrichi; .3) C .. procnemis; .4) Mycodrosophila dimidiata; .5) Scaptomyza adusta; .6) S. pallida; .7) S. hsui; . .8) Zaprionus ghesquierei (3/ 1Z individuals); .9) Z. ghesquierei (9/1Z individuals); .10) z_ vittiger. to to U U 7 j \J I I tripunctata ~ ~ J1 """ "~p \ m•I:;~;'"' // " (1) -____, I I '""''" \\ I I Qi) II '~ c::: ;:i ~­ (1) ~ ~ -· ~ )( '-/ \\ grJ]oup~ .IO .. virilis 0 ~ group "i:I >:: ~ ...... §' .9 ,.,., viri lis-replete branch . I I . ' c;· ®~ ;:i 1mm1grans group 5 -----:?ubgenu s Subgenus Drosophila Pholadori s '""''""' ,,,,,,,.,,,, ~ ~ @ Fw. 14. Pictorial phylogeny of paragonia and vasa deferentia. . 1) cancellata; ,2,) lacicola; .3) repleta; .4) species O; .5) guarani; .6) mediostriata; .7) quinaria; ,8) subfunebris; .9) immigrans; .10) neocordata; .11 ) equinoxialis; .1Z) simulans; .13) ananassae; .14) algonquin; .15) populi, Throckmorton: Phylogeny in Drosophila diameter of the paragonia, followed apparently by a reduction in length also, so that only vestiges remain. This is most conspicuous in the members of the hydei subgroup of the repleta group where the paragonia are very small (Figures 9.1-.5). In these figures the size of the paragonia is somewhat exaggerated. An apparent reduction is also seen in some members of the virilis group (Figures 4.25-.33) and in members of some other subgroups of the repleta group (Figures 8.10, .16). These latter cases, however, probably represent retention of an an­cestral, less robust form of paragonia rather than an actual reduction. The association between the vasa and the paragonia is one of the most charac­teristic features of species in the subgenus Drosophila. Although no sharp distinc­tion can be made, species from both sections of this subgenus fall into two major categories. In one the vasa are in close association with the paragonia but not actually·adhering to them. In the other the vasa adhere so closely to the walls of the paragonia that they can be separated from them only with difficulty, if at all. Between these two extremes are many species in which the vasa adhere to the paragonia, but so loosely that careless dissection will disrupt the associa­tion. These categories can be further subdivided into forms in which the vasa are rather irregularly associated with the paragonia, generally crossing their ventral surface, and forms in which the association is very regular and the vas adheres to, and follows; the major curvature of the arch of the paragonium. This associa­tion generally ends at the point where the differentiation of the vas begins, al­though in species where the vas is unusually long a considerable portion of its distal end may be free. Generally, forms having the most irregular association between the vasa and the paragonia also have a more loose attachment between the two. Forms having a very regular association between the two may occupy either extreme and have the vas strongly adhering or completely free (but still following the major curvature of the paragonia very closely). Presumably, the forms in which the vasa are free but associated with the paragonia approach the primitive type for the subgenus. In the figures a common line serving both to delineate the outer surface of the paragonium and the inner surface of the vas indicates that the vas adheres to the paragonium in that region (e.g., 14.2-.3, .5-.6). Where separate lines are used the association is loose and generally the vas does not adhere (Figures 14.4, .7, .9). The great majority of species belonging to the virilis-repleta section (Figures 4.25-.33; 5.1-.11; 7; 8; 9; as well as others in Figure 6 which can be picked out by inspection) show a regular association between the vasa and paragonia and a relatively tight adherence between the two. An irregular association is seen only in some species of the repleta group, and even here the association is not highly irregular. The examples ofthis type are seen in Figure 7 (.12, .14-.16, .18) and in Figure 8.6. In most of these cases the irregularity lies in the vas crossing a surface of the paragonium rather than following the major curvature. In some,. however, the association is very brief, as in longicornis (Figure 7.14). Generally speaking, none of the species of the virilis-repleta section appear to be primitive with respect to the characteristics just discussed, although the repleta species· just mentioned appear to have retained some features of more primitive forms. As was indicated earlier, in members of the virilis-repleta section the paragonia The University of Texas Publication generally are large and robust, and exceptions to this have already been noted. Since most of the individuals involved are themselves rather large, it is difficult to say, without careful measurements, whether there actually has been a relative increase in size of the paragonia during the evolution of this branch. The general impression is that the paragonia are distinctly more robust than in presumed primitive forms, but this will need to be verified. It is pertinent in this respect to note that there is some difference with age in the characteristics of the para­gonia. In newly emerged and very young adult males of some species, the para­gonia are not fully developed. That is, they appear to have their normal length and characteristic folding, but they are not completely filled out and may have a somewhat shriveled appearance. Paragonia in very old males may also have · this appearance. This condition generally changes within a day or two of emer­gence, and, during dissection, an effort was made to use only fully mature males. However, the possibility remains that an entire sample may have been of rela­tively young, or of very old, males. This seems somewhat improbable, but it may explain why species such as paranaensis (Figure 8.10) differ in this respect from their close relatives. It is highly improbable that age differences are responsible for the small paragonia seen in virilis group species, and age definitely is not responsible for the previously mentioned characteristics of hydei subgroup species. In other respects the morphology of the paragonia in species from the virilis­repleta section is relatively constant. Generally the paragonia are folded at least three times, and often more. Some species of the virilis group are not of this type (Figures 4.25-.28) and resemble Phloridosa species (Figures 6.10-.12) to a limited extent. A few species from this section have the paragonia only twice folded, or folded only once, and these can be picked out by inspection of Figures 5 through 9. The next cluster of species to be considered is, in many ways, intermediate between the two major branches. Species of the immigrans group (Figures 5.2­ .14), for example, show the basal fusion of the vasa generally associated with species of the quinaria section, but the paragonia are twice folded, and they may show a reduced (or incipient) third fold. In two of the species, immigrans and spinof emora, the vasa closely follow the major curvatures of the paragonia but do not adhere to them. As was said earlier, this presumably approaches the primitive condition for the subgenus Drosophila. In the third species, hypocausta, the vasa adhere to the paragonia and follow their curvature. Except for thei presence of a second fold of the paragonia, this configuration is very similar to that seen in species from the quinaria branch. In the funebris group (Figures 6.1-.3) a somewhat similar pattern is seen. The paragonia vary from three folds in funebris to one fold in subfunebris. The association between the vasa and the paragonia is loose and irregular in two species but regular and of the quinaria type in the third. The pattern seen in carbonaria (Figure 6.5) is also generally of the type seen in subfunebris, although the vasa more nearly follow the major curvature of the paragonia. Among the species which most distinctly belong to the quinaria section there is one group, and another cluster of species from various groups, which closely resemble the types seen in the funebris and immigrans groups. Species from the Throckmorton: Phylogeny in Drosophila calloptera group closely resemble immigrans group species (Figure 12.12-.14), except that they lack the second fold of the paragonia. The vasa uniformly follow the major curvatures of the paragonia, but they do not adhere to them. In only one of the three species available, schildi, are the vasa fused basally, and this ' only for a short distance. The other group of species (Figures 10.1-.3, .6, .9; 1 Li) may or may not have the vasa fused basally, and the vasa are generally short and cross the surfaces of the paragonia rather than follow the major curvatures. A comparison between Figure 10.1 and Figure 6.2 will indicate the general similari­ ties between these two types. Other species belonging to this section (Figures 10, 11 and 12) are very similar in general appearance of the vasa and paragonia. A few species (Figures 10.5; 11.2, .5, .8), mostly belonging to the tripunctata group, show the virilis-repleta pattern with the vasa unfused basally and closely adhering to the paragonia. The paragonia, however, are typical of those seen in other members of the quinaria section. The remaining species vary chiefly in the closeness of the association between the vasa and paragonia. In some (e.g., quinaria, Figure 14. 7) the associ­ ation is very loose. In others (all species of the cardini and guarani groups, most species of the tripunctata group, and about half of the species in the quinaria group), the association is strong and the vasa adhere closely to the paragonia. The three species available from the subgenus Phloridosa are all very similar and differ among themselves primarily in that the association between the vasa and paragonia is loose in one but stronger in the other two (Figures 6.10_::.12). The paragonia are only slightly arched and have only one fold. In general aspect they resemble types seen in virilis group species (Figures 4.25-.28) more closely than any others, and thus they are placed in the phylogeny as shown in Figure 14.4. Only four species were available from the subgenus Hirtodrosophila, and these are quite variable (Figures 3.8-.10 and 5.5). Even from such a small sample this subgenus appears to be very heterogeneous. Thus far, pictiventris is the only species seen to have spirally coiled paragonia. D. dlfncani has the primitive vasa and has paragonia which probably are not too different from the primitive type. D. thoracis and histrioides are somewhat similar to species of the virilis-repleta section, although neither resembles them closely. Perhaps histrioides is closer to the immigrans group (and was placed on the same figure with these species for comparison), but such a relationship is highly improbable for pictiventris and duncani. The relationships of the Hirtodrosophila to other members of the genus will be discussed later. They have not been included in the initial phylogenies. Paragonia and vasa of species from other genera are shown in Figure 13. In all Chymomyza species available for dissection (Figures 13.1-.3) the vasa are completely free of association with the paragonia and generally resemble Sopho­phoran types. The single available species of Mycodrosophila (Figure 13.4) is similar in general respects to the immigrans type, although it lacks the basal fusion of the vasa. The paragonia are strongly arched, as are those of most species from the quinaria section of the subgenus Drosophila, but they have a long, strongly attenuated section distal to the major arch. Species from the genus Scaptomyza are of two general types. The paragonia and vasa of S. ailusta and S. pallida (Figures 13.5-.6) are similar to the type seen in the funebris group, The University of Texas Publication although the vasa are much shorter. S. hsui belongs to a type of its own, although others similar to it are figured by Okada (1956). Here the vasa are greatly expanded, and the figure (13.7) includes the testes. Species from the genus Zaprionus resemble species from the subgenus Drosophila but have characteris­tics of both major branches. The paragonia are very robust and are folded at least three times. In Z ghesquierei the individuals were variable, showing two major types with only slight intergradation between them. Of twelve males dissected, three were as in Figure 13.8 and nine were as in Figure 13.9. Associa­tion between the vasa and paragonia was almost absent in the first type since the vasa are short and almost immediately expand to form the strongly differ-· entiated region typical of so many Drosophila species. In the second type the vasa are somewhat longer and associate closely with the ventral surface of the paragonia, somewhat after the fashion of some species of the repleta group. Ex­cept for the extra folds of the paragonia, Z. vittiger resembles species from the quinaria section. The ejaculatory bulb and the ejaculatory apodeme-There is considerable variation in the characteristics of the ejaculatory duct among the various Aca­lypterate families. In the Piophilidae, Chloropidae and Anthomyzidae the duct is undifferentiated and shows neither ejaculatory bulb nor ejaculatory apodeme (Figures 1.6-.9). In the family Ephydridae (Figures 1.2-.3) there may be one or several enlargements of the ejaculatory duct, but these bear very little re­semblance to those seen in other families. In one case, Scatella stagnalis, there is a dark-pigmented, chitinized structure within an enlargement. In the other, Discocerina obscurella, the enlargement of this region is very slight and there is a very small chitinized structure resembling a limpet shell in shape and at­tached to the outside of the chitinized lining of the duct. In position at least, this resembles the condition seen in other families. This "ejaculatory bulb" has been shown, relatively much enlarged, in Figure 22.16. Figures 22.11-.15 show ejacu­latory bulbs and ejaculatory apodemes from several other families of Diptera. Although varying in detail, the ejaculatory bulbs and ejaculatory apodemes seen in these families (Diastatidae, Periscelidae, Aulacigastridae, Sepsidae, Sphaeroceratidae) appear to be similar in their basic characteristics. The ejacu­latory apodeme seen in one member of the family Agromyzidae is rather com­plex and has not been figured in detail. In its general features, however, it appears to resemble the types just mentioned. Many types of variation are seen within the family Drosophilidae. In Gitona bivisualis and Rhinoleucophenga obesa the ejaculatory duct is undifferentiated, with neither ejaculatory bulb nor ejaculatory apodeme. At the other extreme are seen such species as Zaprionus vittiger (Figure 22.8) where the ejaculatory bulb is large and itself further differentiated, with paired and bifurcated posterior caecae. Okada (1958) has made an extensive survey of the characteristics of ejacu­latory bulbs within the family Drosophilidae, and he places them in seven types, as follows: Type 0: no bulb Type A: bulb, no caecae Type B: bulb, a single posterior caecum Throckmorton: Phylogeny in Drosophila Type C: bulb, paired posterior caecae Type D: bulb, paired and bifurcated posterior caecae Type E: bulb, paired anterior and posterior caecae Type F: bulb, three anterior, and paired posterior caecae. All of these types, except type B and type F, were present among the species available for this study. Okada (op. cit.) concludes that type A or type 0 repre­sents the primitive type in the family, and he recognizes the possibility that type 0 may be a degenerate form of type A. Considering the types of bulbs seen among the Acalypteratae, it seems more probable that type 0 is primitive among the acalypterates and that type A and type 0 represent primitives for the family Drosophilidae. Type A would be primitive for the genus Drosophila. The type classification of Okada is apparently based on uncleared material. A brief inspection of Figures 15-19 will show that there is some overlapping of types, and there are many more types apparent when cleared material is used. In addition there are some useful differences in the morphology of the caecae which complicate any simple system of types. While reference to types is useful for generalization, it is not adequate for the present purpose, and Okada's types will be referred to only occasionally in the following discussion. Even a moder­ately complete division into types, based on the data presented in Figures 15-22, would more than quadruple Okada's number of seven. Such a system would probably be intelligible only to its author, and the multiplication of types would serve no useful purpose in the present discussion. Nater (1950) made a survey of the ejaculatory apodeme (Sammenpumpen­sklerit) in Drosophila and related species, and he also summarizes its major fea­tures in a later paper (Nater, 1953). All of his types were seen in the present study, although some differences of interpretation are apparent. He does not indicate a primitive type, but since the spade type (e.g., Figure 15.6) is found in species from all the major subgenera, it will be considered as near the primi­tive. If the absence of an ejaculatory bulb is considered primitive among the Aca­lypteratae, then the most primitive type in the genus Drosophila should be that showing the least differentiation of the bulb relative to the ejaculatory duct, and the apodeme should be only slightly developed. As indicated by Okada (1958), such types might be degenerate, and this must be recognized as a possibility. On the whole, however, interpretation of these types as primitive seems more reason­able, and the phylogeny (Figures 20 and 21) will be based on this assumption. Ejaculatory apodemes are arranged in a phylogeny in Figure 24. When the characteristics of the ejaculatory bulb are considered, the type seen in coracina group species (Figures 15.3-.5) of the subgenus Pholadoris (Figures 15.1-.7) appears to be nearest the primitive. (Note: in situ the apodeme is di­rected ventrally. In the figures, the bulb has been inverted so that ventral is toward the top. Anterior is toward the right.) In these species the bulb is very small and only slightly more enlarged than the ejaculatory duct. In Figures 15.3-.5 the size of the bulb is exaggerated. In relative proportion, the bulb of victoria (Figure 15.1) is at least six times as large as that of cancellata (Figure 15.3). The ejaculatory apodeme in these species is small, almost unpigmented, and weakly chitinized. The ejaculatory bulb and apodeme in latifasciaeformis The University of Texas Publication -/-----­ @ 33 34 Throckmorton: Phylogeny in Drosophila species from this section of the subgenus show various modifications of this type of bulb, and some representatives of this type are shown in Figure 23, together with figures of two "normal" bulbs·for comparison. Figures 23.1-.2 show lateral and ventral views of the ejaculatory bulb for castanea. Figures 23.3-.4 show ventral views of the bulbs of lacertosa and sordidula respectively. Only a part of the long caecae of lacertosa has been shown in the diagram, and the anterior ejaculatory duct and ejaculatory apodeme have been omitted. Figures 23.7-.10 show lateral and ventral views of species having the more normal type of ejacu­latory bulb. Species belonging to the annulimana group (Figures 16.19-.21) have two major types of ejaculatory bulbs. D. gibberosa has posterior but no anterior cae­cae. The others have both anterior and posterior caecae. Except for the length of these caecae, these latter types resemble that seen in colorata (compare Figures 16.16 and .20). Ejaculatory apodemes in this group are also of two types. Two species have the spade type. The third has a modified type similar to that seen in some Sophophorans (compare Figures 15.45 and 16.21). The ejaculatory bulb in species of the repleta group varies from very small and simple to very large. The simplest types are seen in the fasciola subgroup (Figures 17.1-.8). In most of these species the bulb is very small and has neither lateral lobes rior caecae (e.g., Figure 17.7). In others (e.g., Figure 17.3) the bulb is somewhat larger and there are short, blunt, posterior caecae (lateral lobes tilted posteriorly?). Both in size and in extent of differentiation, ejaculatory bulbs in this subgroup are among the simplest in the genus. The bulb itself is only a slight enlargement of the ejaculatory duct. The ejaculatory apodeme is of the spade type. Species of the mulleri subgroup (Figures 17.9-.25) show almost all of the types of variation seen in the repleta group as a whole. The ejaculatory bulb in mulleri is virtually identical with those seen in some fasciola subgroup FIG. 19. Ejaculatory bulbs and ejaculatory apodemes of Drosophila species. tripunctata group .1 mediodiffusa .2 albicans .3 metzii . 4 albirostris .5 tripunctata .6 mediopunctata .7 unipunctata .8 trapeza .9 bandierantorum .10 paramediostriata .11 mediostriata .12 mediopictoides .13 crocina guarani group .14 guaramunu .15 guaraja .16 griseolineata .17 subbadia .18 . guarani cardini group .19 dunni .20 belladunni .21 nigrodunni .22 polymorpha .23 neomorpha .24 neocardini .25 parthenogenetica .26 acutilabella .27 procardinoides .28 cardinoides .29 cardini calloptera group .30 ornatipennis .31 calloptera .32 schildi .33 detail of ejaculatory apodeme, generalized cardini type .34 detail of ejaculatory apodeme, generalized tripunctata type The University of Texas Publication Subgenus Sophophoro Throckmorton: Phylogeny in Drosophila quinorio .9 2 ~ / / group 3 ~corocino v1ctorio quinorio branch ~'~~~u p\g\ou\p '""'~\\ Subgenus Drosophi lo Subgenus Pholodoris ,@ \ FIG. 2L Pictorial phylogeny of ejaculatory bulbs (Drosophila and others) , .1) bryani; .2) cancellata; .3) victoria; ,4) virilis; .5) micromelanica; .6) paramelanica; ,7) mojavensis; .8) species P; .9) innubila; , 10) parthenogenetica; J 1) crocina, The University of Texas Publication jk ~ .10~ II~ ;;J ~,.J; .15~ FIG. ZZ. Ejaculatory bulbs and ejaculatory apodemes from other Drosophilid genera, and from other families of Acalpyteratae..1) Chymomyza amoena; .Z) C. aldrichi; .3) C. proc­nemis; .4) Scaptomyza adusta; .5) S. pallida; .6) S. hsui; .7) Zaprionus ghesquierei; .8) Z. vittiger; .9) Mycodrosophila dimidiata; .10) Zapriothrica dispar; .11) Diastata vagans (Dia­statidae); .1Z) Periscelis annulata (Periscelidae) ; .13) Aulacigaster leucopeza (Aulacigastridae) ; .14) Sepsidimorpha secunda (Sepsidae) ; .15) Leptocera sp. (Sphaeoceratidae); .16) Discocerina obscurella (Ephydridae) . species. The bulb is very small and simple. Most species in this subgroup (e.g., Figure 17.12) have moderately large bulbs with short posterior caecae. Three species (Figures 17.15-.16, .27) have a bulb derived from a type seen in species from the melanopalpa and mercatorum subgroups. The ejaculatory apodemes are all of the spade type. According to Wasserman (1960) species of the mulleri and fasciola subgroups are related cytologically as shown in Figure 25. Correla­tion of cytological and morphological details will be deferred until a later section, but it is helpful to refer to the cytological phylogeny at this time. As can be seen from Figure 25, there are three major phyletic lines within the group. All of these derive from the repleta standard gene sequence. There are several species, chiefly members of the mulleri subgroup, which are derived independently from the standard. The next cluster of species to be considered inakes up the second major branch · Throckmorton: Phylogeny in Drosophila within the repleta group. The species involved belong to the melanopalpa and mercatorum subgroups. In addition, peninsularis, generally considered a mem­ber of the mulleri subgroup, belongs cytologically to this branch. Ejaculatory bulbs of these species are of three types. The first, seen in mercatorum (Figure 17.30), is moderate in size and has short posterior caecae. This closely resembles types seen in the mulleri and fasciola subgroups. The second, seen in [N!ninsularis (Figure 17.29) and several other species, is rather elongate and has short posterior caecae which curve inward and generally touch at the midline. The third type (Figure 17.31 and others), almost certainly a modification of the type just men­tioned, is the same in all details except that the caecae fuse at their tips and form a continuous tube around the posterior ejaculatory duct. It is this type which was seen in some of the members of the mulleri subgroup mentioned previously (e.g., Figures 17 .15-.16) . Two types of apodemes are seen in members of this branch. Most are of the usual spade type. In some (e.g., Figure 17.36) the plate is much flattened. .2 FIG. 23. Lateral and ventral view of certain ejaculatory bulbs. Anterior ejaculatory duct and ejaculatory apodeme have been omitted from ventral views . .1 ) castanea (lat.); .2) castanea (vent.); .3) lacertosa (vent.); .4) sordidula (vent.); .5) peruviana (vent.); .6) camargoi (vent. ); .7) gaucha (lat.); .8) gaucha (vent.); .9) pavani (lat.); .10) pavani (vent.). The University of Texas Publication The third major branch of the group consists of hydei subgroup species (Figures 17.38-.42) . Here the ejaculatory bulbs are very small, and they are much alike in general morphology. Posterior caecae are small and generally out­curved. All have the spade type of apodeme. The ejaculatory bulb of castanea is shown in Figure 17.49. This species is derived, cytologically, from the repleta standard. Species Fis, on the basis of internal morphology, very close to castanea and it has an ejaculatory bulb identical with that figured for castanea. These two species have the posterior part of the ejaculatory bulb expanded, and they resemble some species of the robusta group in this respect. Aside from the repleta group species and castanea, two other groups, canalinea and dreyfusi, are derived cytologically from the repleta standard. The cytological evidence indicates the relationship shown in Figure 25. The ejaculatory bulb of paracanalinea (Figure 17.44) is small and has two short, blunt posterior caecae. The bulb of canalinea (Figure 17.43) is larger and has long, slender posterior caecae. Both species have the spade type of apodeme, although that of paracana­linea is rather small and weakly chitinized. The ejculatory bulbs of species from the dreyfusi group (Figures 17.45-.46) have moderately long posterior caecae. There are also very small anterior caecae resembling those seen in robusta (Figure 16.18). In these two species, the anterior caecae are in close contact with the anterior ejaculatory duct and appear to clasp it basally. In ventral view (Figure 23.6) the bulb is seen to be of the type mentioned earlier for castanea and many robusta group species. The general similarities between the ejacula­tory bulbs of dreyfusi and robusta group species can be seen by comparing Figures 23.3, .4 and .6. Ejaculatory apodemes are of the spade type. Species of the mesophragmatica group are also cytologically related to the repleta complex, but not as closely as are the species just covered (Wasserman, pers. comm.) . They have the normal type of bulb with two moderately long posterior caecae. The ejaculatory apodemes are different in the two species. That in pavani (Figure 17.48) is of the usual spade type. That of gaucha (Figure 17.47) has the plate shaped more like a spoon, and in this respect resembles somewhat the types which will be seen shortly in species from the quinaria section. The plate, however, is much shorter than that of the quinaria type. On the basis of external morphology the next group of species is heterogeneous. Some features, particularly of the ejaculatory apodeme, suggest a relationship between them. The simplest type is that seen in an undescribed form (species H, Figure 18.7). Here the bulb is moderately large but lacks both lateral lobes and posterior caecae. The ejaculatory apodeme is minute, and its size is exag­ gerated three to four times in the figure. The size of the bulb is as it should be in proportion to other bulbs in the genus. Ifthe ejaculatory apodeme were drawn to the same scale it would appear as a fine line in the figure. The apodeme plate is almost non-existent. The base of the handle simply flares slightly and curves around the base of the ejaculatory duct. D. peruviana is, on the basis of external morphology, related to this species, but it has a quite different ejaculatory bulb and apodeme. This is figured in lateral view in Figure 18.6 and in ventralview in Figure 23.5. Another undescribed form (species G, Figure 18.5) has an ejaculatory apodeme very similar to that seen in species H. The apodeme is minute, and the plate is only slightly more developed than in the previously Throckmorton: Phylogeny in Drosophila mentioned species. The ejaculatory bulb is moderately large with stout posterior caecae. A species of the bromeliae group (Figure 18.3) has an ejaculatory bulb and apodeme very similar to that seen in species G. The major difference between the two is in the length of the posterior caecae. The last species having this type of apodeme is nannoptera (Figure 18.2). The ejaculatory bulb has short posterior caecae. The apodeme is minute and almost identical with that seen in species H. The ejaculatory bulb of carbonaria (Figure 18.1) shows few features which clearly associate it with any group in the genus. The bulb. is small with lateral lobes and short anterior and posterior caecae. The ejaculatory apodeme is of the spade type and, on the whole, the ejaculatory bulb most nearly resembles a type seen in the melanica group (compare Figures 16.14 and 18.1). Species of the immigrans group (Figures 16.22-.24) have large ejaculatory bulbs. That of spinofemora is undifferentiated. That of hypocausta has slightly developed lateral lobes and folded posterior caecae of a type to be seen among species from the quinaria section of the subgenus. D. immigrans has both anterior and posterior caecae, although all are very short. In this respect the bulb some­what resembles that of robusta (Figure 16.18), although the position of the anterior ejaculatory duct and the ejaculatory apodeme is quite different in the two species. The ejaculatory apodeme is very large in two species of the immi­grans group. That of spinofemora is, relative to the others in the group, rather small, but this appearance is due mainly to the short handle. The plate is of the spade type, but it is about twice as long as those previously seen of this type. In length it approaches types seen in the quinaria section. The apodeme of immi­grans is large and heavy. The plate is long and its anterior angles are rounded, a feature which is also characteristic of many species of the quinaria section. The apodeme of hypocausta is unique. The handle is large, but very thin and flat. It is continuous with a thin keel which extends down the mid-ventral line of the plate. The plate is also thin and flat. In species of the funebris group the ejaculatory bulb is rather uniform (Figures 16.26-.28). In two species the bulb is almost spherical. Lateral lobes are incon­spicuous and the posterior caecae are folded. In the third species the bulb is elongate and somewhat expanded posteriorly. Ejaculatory apodemes are of the spoon type and have short handles. In general aspect they most closely resemble types seen in species from the quinaria and calloptera groups. Ejaculatory bulbs of species in the quinaria group fall into four types. In innubila (Figure 18.23) the bulb is moderate in size and has two short posterior caecae. In falleni and phalerata (Figures 18.26-.27) the bulb is elongate and has short, folded posterior caecae. In quinaria, rellima, tenebrosa and occidentalis (Figures 18.24-.25, .29-.30) the bulb is elongate and the posterior caecae are moderately long. The caecae in these four species are almost of uniform diameter throughout. In the remainder of the species in the group (Figures 18.28, .31-.35) the caecae are longer and have two distinct sections. The proximal section has the same diameter as that of caecae seen elsewhere in the group. Somewhat before the midpoint of the caecum there is a constriction, and the distal section has a diameter roughly half that of the proximal section. In most individuals the con­striction is sharp and the two sections are distinct. In a few the diameter tapers gradually, but over a distance of less than one-tenth the total length of the cae­ The University of Texas Publication cum, to the smaller size. Although the apodemes varyin detail, they are all of the spoon type. Species of the calloptera group (Figures 19.30-.32) have very elongate ejacu­latory bulbs. That of schildi is flexed near its midpoint and has moderately long, and precisely folded, posterior caecae. The other two species have elongate bulbs, but they are not flexed. The posterior caecae are long and much. folded, but they do not show the regular long folds seen in schildi. Ejaculatory apodemes are all of the spoon type. Species of the testacea group (Figures 18.19-.20) have small ejaculatory bulbs with short posterior caecae. In testacea the posterior caecae are somewhat dis­placed anteriorly, a condition seen often and more conspicuously in many of the species remaining to be covered. Ejaculatory apodemes are of the spoon type. Species of the rubrifrons group (Figures 18.15-.16) have ejaculatory bulbs of moderate size with short posterior caecae. The bulb of one (Figure 18.16) is . slightly flexed and the caecae are displaced somewhat anteriorly. Apodemes are of the spoon type. Ejaculatory bulbs from species in the macroptera group (Figures 18.21-.22) have long posterior caecae. Both in respect to characteristics of the bulb and of the apodeme, they resemble some ~pecies of the tripunctata group. Most species of the tripunctata group (Figures 19.1-.13) have moderate sized ejaculatory bulbs which are elongate to very elongate. Caecae may arise almost at the posterior end of the bulb (e.g., Figure 19.2), but they are generally displaced anteriorly (e.g., Figure 19.3) and appear to arise from the apex of a lateral fold. This appearance is due, in part, to a flexure of the bulb similar to that seen in schildi (Figure 19.32). The posterior caecae are generally short (and folded when long enough). In crocina the caecae are very long. Ejaculatory apodemes are rather variable. Some (e.g., Figure 19.4) have an elongate spade type. Most (e.g., Figure 19.5) are of the spoon type. Species of the cardini group (Figures 19.19-.29) have ejaculatory bulbs of moderate size. The posterior caecae are rather long and generally are folded. Often they are displaced anteriorly, sometimes conspicuously so (Figure 19.24). In this last case, neocardini, the bulb is not flexed, and the caecae arise at the apices of distinct lateral folds. In this case also, the caecae cross over each other and are often intertwined. This crossing of the caecae is seen in some individuals from all the species in the group but it cannot be considered characteristic of any species (with neocardini a possible exception) . The ejaculatory apodemes of species in the cardini group .are quite distinctive, although they show obvious resemblances to the spoon type seen in other species from this branch. of the genus. In most cases the plate is extremely long and slender, generally equaling the length of the handle. The plate is of the usual width across the anterior angles, but it immediately constricts sharply and becomes very narrow and pointed. Figures 19.33 and 19.34 show details of the apodeme for generalized cardini and tripunctata types. Only polymorpha has an apodeme which approaches the tri­punctata type. Ejaculatory bulbs of guarani group species are generally large. Posterior cae-· cae are rather short and generally are folded. Anterior displacement of the cae­cae is not pronounced, and lateral folds are not Conspicuous. The ejaculatory apodemes are large and of the spoon type. Ejaculatory bulbs and apodemes of I Throckmorton: Phylogeny in Drosophila miscellaneous species from this branch fall into one or another of the types al­ready discussed. D. pallidipennis (Figure 18.18) is similar to trapeza (Figure 19.8); sticta (Figure 18.17) to mediodiffusa (Figure 19.1), etc. In species from the subgenus Phloridosa (Figures 18.9-.11) the ejaculatory bulb is small with lateral lobes and folded posterior caecae. In one species the caecae are relatively long. The apodeme is of the spoon type similar to those seen in species from the quinaria section or in polychaeta. Species from the subgenus Hirtodrosophila are quite variable (Figures 15.8-.10; 16.25). The ejaculatory bulbs in duncani and thoracis have short, blunt posterior caecae. The posterior caecae in pictiventris (Figure · 15.9) are peculiar to that species. In histrioides (Figure 16.25) the ejaculatory bulb is small with blunt posterior caecae. The bulb itself resembles types seen in the repleta group (e.g., Figure 17.20), but the ejaculatory apodeme is quite different. The plate is rather broad, flat and long. In general shape it resembles types seen in species of the quinaria branch, but it is too flat to appear very similar to them. The ejaculatory apodemes of duncani, pictiventris and thoracis are slight modifications of the spade type. The ejacu­latory bulb of busckii is moderately large and has lobose posterior caecae (Figure 15.11). The apodeme is of the standard spade type. The ejaculatory bulbs of species in the genus Chymomyza (Figure 22.1-.3) most nearly resemble those of species in the obscura group (compare Figures 22.1 and 15.17 or 22.2 and 15.13). The apodemes are of two types. The first (Figures 22.1-.2) resembles the obscura type but is much more delicate. The apodeme of C. procnemis (Figure 22.3) approaches the standard spade type. The lateral borders of the plate are weakly chitinized and not pigmented. Nater (1953) gives a figure of the apodeme of C. procnemis which differs from this. His figure shows an apodeme of the type seen in C. amoena and C. aldrichi. The ejaculatory bulbs of Scaptomyza species vary (Figures 22.4-.6). That of S. adusta is moderate sized. In the individuals dissected the caecae were un­branched. However, Rosenblad (1941)' describes this species as having branched posterior caecae (as figured for S. hsui, except that the caecae are longer), and Patterson ( 1943) gives a figure for this species in which one caecum is branched and the other not. Apparently this species is variable in this respect. The caecae of S. hsui are short and branch once. Those of S. pallida are longer and branch twice. The apodemes are of the spoon type. Ejaculatory bulbs of Zaprionus species are moderately large and almost spherical. One (Figure 22.7) has long, unbranched posterior caecae. The other (Figure 22.8) has long caecae, each of which branches four times. The apodemes are modifications of the spoon type (compare Figures 22.8 and 18.21). The ejaculatory bulb of Mycodrosophila dimidiata (Figure 22.9) is distinctive, although one might force a resemblance between it arid that of pictiventris (Figure 15.9) . The ejaculatory apodeme is of the spoon type. Morphology of the testes-In the majority of the Acalypteratae the testes are elliptical or nearly so (Figure 1). Spiral testes have been seen only in the Drosoph­ilidae, and in this family there is great variability in testis form (Figure 2). Investigations by Stern (1940, 1941a, b) indicate the mechanisms involved in the formation of the spiral testes in several species from both the subgenus Drosophila and the subgenus Sophophora. During early developmental stages The University of Texas Publication Subgenus Drosoph i la Subgenus ~ophophora Fw. 24. Pictorial phylogeny for ejaculatory apodemes . . 1) victoria; .2) bryani; .3) cancel­lata; .4) virilis; .5) melanica; .6) peninsularis; .7) melanopalpa; .8) canalinea; .9) nannoptera; .10) species O; .11) palustris; .12) tripunctata; .13) cardini; .14) calloptera; .15) guarani; .16) funebris; .17) immigrans; .18) neocordata; .19) neoelliptica; .20) subsaltans; .21) parasaltans; .22) pseudobocainensis; .23) willistoni; .24) melanogaster; .25) serrata; .26) ananassae; .27) pseudoobscura; .28) populi. the testes are completely free, and it is only during later ontogeny (pupal) that they become attached to ducts which arise from the genital imaginal disc. Stern (1941a) finds that the form of the testes is determined by the thin inner mem­brane of the testis capsule, and growth occurs by terminal addition to this mem­brane. Differential elongation is explained as being due to the release of the growth promoting substance from the vas in different amounts to opposite sides of the testis. The membrane elongates most near the point of attachment and curvature of the testis is thus away from this point (see Figures 2.9, 13.7 and 26.2). This is important, since it provides the criterion by which true spiral coiling is differentiated from pseudo-coils or curvatures produced primarily by crowding of an elongate testis in the body cavity (e.g., Figures 2.2-.3). There can be little doubt that the elliptical form of the testis is primitive and that the spiral form is derived within the family Drosophilidae. To some extent the number of testis coils is an index of the evolutionary status of a species, in that the more primitive forms tend to have a lower number of coils. Within the family testis form varies from true elliptical (e.g., Figures 2.1, .4-.5), to elongate elliptical (e.g., Figures 2.2-.3) , to true spiral (e.g., Figures 2.6-.13). Other forms are figured by Okada (1956), but in many cases these cannot be evaluated as to whether they belong to the true spiral form or to some other type. This can only be decided by use of the criterion mentioned earlier, i.e., direction of testis curvature relative to its point of attachment to the vas. In some cases the presence of counter-coiling of the vas deferens (inner coil of testis) can be taken Throckmorton: Phylogeny in Drosophila as indicative of true spiral coiling in the testes, since the counter-coil of the vas results mechanically from asymmetric growth of the testis. However, not all species having true spiral testes exhibit counter-coiling of the vasa (e.g., Figures 2.6, .10 and 13.7). In these cases, curvature of the testes relative to the point of attachment is the only available criterion. Since the need for this type of dis­tinction has not been recognized previously, earlier descriptions and figures of testes are of limited value. In most cases it seems probable that those described as spiral exhibit true spiral coiling, but some doubt must remain until these spe­cies have been re-examined with appropriate criteria in mind. Within the genus Drosophila there is little difficulty on this point. All species examined fall into two categories, true elliptical and true spiral. The second category can be arbi­trarily subdivided according to the number of coils in the spiral. Since only a few (10-12) individuals have been examined for any one species, and since there is some individual variation in number of coils (generally dif­ferences are in fractions of a coil over or under a modal value), the data for all the species will not be tabulated. Table 1 summarizes the data on testis coils by groups. Most of the miscellaneous species have been omitted. As can be seen by inspection of Table 1, the number of testis coils ranges from zero (elliptical testis) to approximately twenty. Probably several distinct factors influence the number of coils in the adult testis. Two obvious factors will be the amount of growth stimulation substance produced by the vas and the duration of the period during which this substance is provided to the testis (and/ or the duration of the period during which the testis is capable of responding to this substance). The amount of growth substance appears to be the major factor in determining the number of testis coils in some affinis subgroup species, and differences between the num­ber of coils in melanogaster (three) and in virilis (six) are due to difference in rate of testis growth relative to the pupal period (Stern 1941b). At the present time it is not possible to determine which of these factors are operating, singly or in combination, to produce the various numbers of coils and other variations in testis form seen among Drosophila species. Thus, the major distinction which can safely be made is between elliptical and spiral testes, and, as was said pre­viously, the higher number of coils generally indicates more derived forms. Inferences regarding types of spiral testes must be tentative, but conclusions drawn from certain closely related species will probably be valid. Figure 26 gives a much abbreviated phylogeny of testes forms. Representatives chosen as ex­amples of forms from the two major branches of the subgenus Drosophila indicate the more extreme types. Intergradations between these two types do exist, but generally the two types are recognizably distinct. Species in three of the four groups in the subgenus Phola,doris have elliptical testes similar in type to that shown in Figure 26.1. The testes of bryani (Figure 26.2) have about one-half coil. Of the Hirtodrosophila, one, duncani, has ellipti­cal testes. The others vary from approximately one, to approximately three coils. D. busckii (Dorsilopha) has about two coils (Figure 2.9). In the subgenus Sophophora, populi has elliptical testes. Species of the obscura group fall into two types. Those of the obscura subgroup have elliptical testes (e.g., Figure 26.12) and those of the affinis subgroup have spiral testes varying from one-half coil (affinis) to three coils ( azteca) . Details will not be presented for most of the The University of Texas Publication TABLE 1 Testes coils of Drosophila and other species. The denominator indicates the number of species examined in each group. The numerator indicates the number of species in the group which have the indicated number of coils in the testes. Zero (no coils) indicates elliptical testes. Number of coil s 0 0--1 1-3 3-6 6-9 9-1 2 12--15 15-20 Pholadoris victoria gp. coracina gp. latifasciaeformis gp. bryani gp. H irtodrosophila Dorsilopha Sophophora populi obscura gp. melanogaster gp. willistoni gp. saltans gp. Drosophila virilis-repleta section virilis gp. melanica gp. robusta gp. annulimana gp. canalinea gp. dreyfusi gp. mesophragmatica gp. repleta gp. mulleri sbgp. fasciola sbgp. mercatorum sbgp. melanopalpa sbgp. hydei sbgp. polychaeta gp. nannoptera gp. bromeliae gp. 2/ 2 3/ 3 1/ 1 1/ 4 1/ 1 4/10 1/ 1 1/ 10 3/4 1/ 1 5/ 10 5/12 10/ 10 1/9 1/ 6 12/19 1/ 2 1/ 1 6/ 12 5/ 9 4/ 6 3/ 4 1/2 2/ 2 2/2 7/ 19 3/8 1/2 3/6 1/ 1 1/ 12 10/ 14 2/ 9 1/ 6 1/ 4 4/ 4 1/2 5/ 8 3/ 6 4/ 14 1/9 1/5 2/5 2/ 5 1/ 1 N umher o f coil s quinaria section 0 0-1 1-3 3-6 6-9 9-1 2 12--15 15-20 immigrans gp. funebris gp. testacea gp. calloptera gp. quinaria gp. guarani gp. cardini gp. tripunctata gp. Phloridosa Scaptomyza Zaprionus M ycodrosophila Chymomyza 1/ 3 2/3 1/2 1/3 3/3 3/ 3 2/ 2 3/3 2/13 1/5 6/ 11 9/13 6/ 13 1/5 5/11 4/13 1/2 1/3 5/13 3/5 1/ 1 1/3 3/3 Throckmorton: Phylogeny in Drosophila remainder of the groups in the genus or they will be given later with the per­tinent discussion. Morphological differences, aside from differences in coil number, are of several types. In the majority of groups in the genus, and particularly where the number of testes coils is low (below six), the testes are thick and heavy (e.g., Figures 26.3-.4, .8-.11). Where the number of coils is relatively high, the coils are thin (e.g., Figure 2.11), although the thickness may be somewhat variable throughout the length of the testis. In the majority of species from the quinaria section of the subgenus Drosophila the testes coils are relatively thin, regardless of the number of coils. It is also in these species that the distal region of the vas is strongly dif­ferentiated, and in many species from this section the counter-coils of the vasa are larger and more bulky than the coils of the testes (e.g., Figures 26.5-.6). Taken together, the thin testes and the pronounced counter coils of the vasa produce a pattern which is almost constant for the more derived species belonging to the quinaria section. As was mentioned previously, species from other sections of the genus may have strongly differentiated vasa, but almost invariably these are associated with thick, heavy testes (e.g., Figures 26.3-.4) . It should be noted that the counter-coils of the vasa are often pigmented. In Figures 26.3 and 26.5-.6 the counter-coils have been left unshaded to provide a contrast between them and the testes. Data for testes coils seen in other genera are included in Table 1. The testes of Chymomyza species are relatively heavy. In general they resemble the spiral types seen in Sophophoran species. Okada (1956) figures one species (C. cauda­tula) having elliptical testes. Scaptomyza species also have thick, heavy testes, as do species of the genus Zaprionus. Mycodrosophila dimidiata has around ten thin coils in the testes, and these somewhat resemble the testes seen in species of the quinaria branch. The vasa, however, are only slightly differentiated, and in this respect the pattern does not follow that mentioned for species of the quinaria section. Internal Reproductive Structures of the Female-Characteristics of the sper­ conopalpo (8) melonopolpo (8) mojuoides (6) moju (6) (3) aldrichi foscioloides (8) limensis (5) (3) mulleri (9) pictilis fasciolo (5) (6) p1c turo coroico (5) m.mercotorum (3) '"''";"'"•181 p •eoo•l>l hydei (15) replete (8) n1grohydei (10) neohydei (20) eohydei (15) f. fu lvimocula (6) {fulvimoculoides bifurco (20)(5) comorgoi(6) / / ..;,,.,;(6) r conolineo (5) dreyfusi 1porocanohneo. group v{8) canalinea group to FrG. 26. Pictorial phylogeny of testes forms. . 1) lativittata; .2) bryani; .3) nigricruria; .4) fulvalineata; .5) rnediostriata; .6) subfunebris; .7) subsaltans; .8) sucinea; .9) rnelanogaster; .10) ananassae; .11) algonquin; .12) pseudoobscura; .13) populi. Throckmorton: Phylogeny in Drosophila mathecae, ventral receptacle and parovaria have been widely used by Drosophila taxonomists and they require little introduction. Only the characteristics of the spermathecae will be considered in detail. Those of the ventral'receptacle will be summarized briefly. Some additional, general features can be noted before turn­ing to detailed descriptions. Both the spermathecae and the parovaria arise from a small, chitinized plate in the antero-dorsal vaginal wall. In all species investigated there are four stalks originating from this plate, the two anterior being the spermathecal ducts, the two posterior the parovarial ducts. There is considerable variation in the shape of this plate and in the relative positions of the stalks which arise from it. Character­istics of this plate are rather difficult to determine and they were not routinely recorded. The few notes available suggest that its characteristics will prove use­ful, particularly for classification at the higher taxonomic levels. Characteristics of the parovaria and their ducts have been recorded, but they will not be reported in detail. In most species the length of the parovarium plus its duct is equal to or less than the length of the spermathecal duct (to the base of the spermatheca), and the parovaria are markedly smaller than the sperma­thecae. In a few species, mainly some more derived forms from both the virilis­repleta and quinaria sections of the subgenus Drosophila, the length of the paro­varial duct equals that of the spermathecal duct. In some species, again generally more derived forms from the subgenus Drosophila, the parovaria and sperma­thecae are almost equal in size. There is no apparent correlation between length of duct and size of parovarium. Among Sophc_>phorans, the montium subgroup species of the melanogaster group have spermathecae and parovaria which are almost equal in size and whose ducts are almost equal in length. The length of the parovarial duct cannot be determined accurately unless the material is cleared. In many cases a duct may be of usual length but appear shorter due to coiling which is not apparent before clearing. Or the duct may have a surprisingly long section folded and imbedded within the muscular wall of the vagina. In some species a thinly chitinized and crumpled lining of a parovarium may re­main after clearing. In these cases, the cleared parovaria strikingly resemble the weakly chitinized and non-telescoped spermathecae seen in some species (e.g., montium subgroup species of the melanogaster group, etc.-see Figures 27.29­.33) . Pigmentation of spermathecae varies within the genus. In the more primitive forms (Pholadoris, virilis group, obscura group, etc.) the spermathecae are black. More derived forms vary from brown, to golden, to pale yellow in color. Some species of the tripunctata group have spermathecae which are yellow basally but dark apically (see Frota-Pessoa, 1954). Weakly chitinized spermathecae are also weakly pigmented. At the present time judgment of color would necessarily be subjective, and variation is gradual in some groups. Also, changes seen during clearing suggest pigmentation differences not evident by simple inspection. Two forms having identical color may vary greatly in clearing time and in final ap­pearance. This may reflect heavier chitinization, but it seems probable that other factors (qualitative differences in pigment, differences in dispersion, etc.) con­tribute. For these reasons it seems advisable to defer treatment of spermathecal pigmentation until more reliable criteria are available for its evaluation. The University of Texas Publication ,err ,f Hf .~5cr .lychaeta (Figure 35.9) would better be placed in this type than in the type mentioned earlier. The type seen in Figure 36.7 is found in several species of the mulleri (Figure 36.7) and melanopalpa (Figure 36.12) subgroups of the repleta group. It is also seen in one species of the canalinea group (Figure 37.5) and the mesophragmatica type (Figure 37.2) may repre­sent a variation of this. The type shown in Figure 36.8 is generally typical of species from the hydei subgroup of the repleta group, although some of these species (e.g., bifurca) have four or five larger coils basally. D. aracea (Figure .2 .5 .6 .7 .9 FIG. 38. Ventral receptacle types ..1) mediodiffusa; .2) albicans; .3) unipunctata; .4) tri­punctata; .5) guaramunu; .6) dunni; .7) cardini; .8) ornatipennis; .9) schildi. The University of Texas Publication 35.8) and canalinea (Figure 37.4) also have a ventral receptacle of this type, and that of fulvalineata (fasciola subgroup of the repleta group) somewhat re­sembles it. As can be seen from this brief survey, ventral receptacle types are quite varied, and in several cases the types are restricted to closely related species. At the present time sharp distinctions between some types cannot be made, and the above assignment to types must be considered tentative. The present data do serve to show that evolution of ventral receptacles has followed the same general pattern as was seen for characteristics covered earlier. The primitive type for the subgenus Drosophila was probably similar to that shown in Figure 35.7 or 36.5 (i.e., short, not appressed to the vagina, either weakly or strongly attenuated distally, and it may or may not have been folded). The type shown in Figure 35.4 (histrioides, Hirtodrosophila) suggests derivation of a coiled type from a folded type. Also many of the coiled types show a reversal of coiling at approxi­mately every fifth coil, suggesting that coils have been superimposed on a folded structure, with the points of reversal representing the original folds. It was not possible to check all ventral receptacles for this feature, but all of those which were checked showed the reversal. In the figures (e.g., Figure 34.14) the reversal has been shown only for cases which have been specifically checked for it. If it is not shown in the figure, reversal may or may not be present, but it probably is. Regardless of the origin of coiling, subsequent evolution in both major branches of the subgenus Drosophila has involved increasing length and addition of spe­cific coiling relationships. There is no general correlation between length of the ventral receptacle and number of coils in the testes. Ventral receptacles of the Hirtodrosophila (Figures 34.4; 35.4) and of busckii (Figure 34.5) are of the obscura type (folded and appressed to the vagina). That of histrioides (Figure 35.4), previously mentioned, is distinctive, but it is still basically a folded type appressed to the vagina. Okada (1956) figures several other species having the same type of ventral receptable. Figure 39 shows ventral receptacles from other genera. Members of more primitive genera (Figures 39.1-.3) have short and folded ventral receptacles Frc. 39. Ventral receptacles from other genera: .1) G. bivisualis; .2) R. obesa; .3) Zapri­othrica dispar; .4) C. aldrichi ; .5) S. adusta; .6) Z. ghesquierei; .7) M. dimidiata. Throckmorton: Phylogeny in Drosophila which are appressed to the surface of the vagina. Those from the genus Chymo­myza vary in length, but all are strongly attenuated distally. They are rather irregularly folded, there is no indication of coiling in those seen (Okada, 1956, shows both coiled and folded types), and they tend to be appressed to the surface of the vagina. The ventral receptacles from Scaptomyza species resemble the . obscura type in being folded and generally appressed to the vagina. Ventral re­ceptacles of Zaprionus species resemble those of Scaptomyza species. They are folded and appressed to the vagina in Z. ghesquierei, folded and rather loosely associated or free of the vagina in Z. vittiger. M. dimidiata (Figure 39. 7) has a ventral receptacle which is free of the vagina, coiled basally but folded distally. It somewhat resembles the type seen in histrioides, except that the latter type is appressed to the surface of the vagina. Malpighian Tubules-The various types of Malpighian tubules are shown phy­logenetically in Figure 41. Table 2 summarizes data relating to the condition (free, apposed, fused) of the tips of the posterior tubules. Only the major groups of species have been included in the table. In related Acalypterate families, Malpighian tubules have the posterior tips free (Sturtevant, 1942), and this is also the condition in the more primitive genera in the family Drosophilidae. Okada ( 1955), after a survey of Malpighian tubules within the family, concluded that the type seen in most Sophophoran species (posterior tips free) represents the primitive. Where tips are free the posterior tubules pass from their common stalk back toward the posterior end of the abdomen, then turn forward and their tips lie lateral and adjacent to the gut a short distance posterior to the origin of the stalk. 12~ II me~~~~osler I~ // obscuro z~ Sub9enus Sophophoro /@ FIG. 40. Pictorial phylogeny of ventral receptacles..1) victoria; .2) nigromelanica; .3) aldrichi; .4) longicornis; .5 canalinea; .6) mediodiffusa; .7) mediopunctata; .8) subpalustris; .9) immigrans; .10) equinoxialis; .11) austrosaltans; .12) serrata; .13) affinis; .14) populi. all other species / ~ i~ 1/4 0==1~ 2/4 c:~=>{rc::::~,:~~~~' ~~ ~~ \ ~~0.71~ ~~ ~ ~ ('>) ~ i:;' ('>) ~ ~· ~ "'-3 ~ -~ ~ "3"' K c;· ;::i --... ~F==C r;---,'===l ~ virilis-repletJ 217 ~) ~ Subgenus~Pholadoris, Subgenus I~~""' Drosophila Fw. 41. Pictorial phylogeny of Malpighian tuhules. For explanntion see text. Throckmorton: Phylogeny in Drosophila 281 TABLE 2 Characteristics of posterior Malpighian tubules Free Apposed Fused Pholadoris victoria group 2/ 2 coracina group 3/3 latifasciaeformis group 1/ 1 bryani group 1/ 1 Hirtodrosophila 4/ 4 Dorsilopha 1/ 1 Sophophora populi 1/ 1 obscura group 10/ 10 melanogaster group 12/ 12 willistoni group 10/ 10 saltans group 14/14 Drosophila virilis-repleta section virilis group 9/ 9 melanica group 6/ 6 robusta group 3/4* annulimana group 4/ 4 canalinea group 1/2 1/ 2 clreyfusi group 2/2 mesophragmatica group 2/ 2 repleta group mulleri subgroup 11 / 19 8/ 19 fasciola subgroup 5/ 8 3/ 8 mercatorum subgroup 2/2 melanopalpa subgroup 6/6 hyclei subgroup 1/ 5 4/5 polychaeta group 1/ 1 nannoptera group 1/ 1 bromeliae group 1/ 1 Free Apposed F used quinaria section immigrans group 3/3 funebris group 3/3 testacea group 2/2 calloptera group 2/ 3 1/3 quinaria group 13/13 guarani group 3/5 2/5 carclini group 11 / 11 tripunctata group 3/13 10/ 13 Phloridosa 2/3 1/3 Scaptomyza 2/3 1/3 Zaprionus 2/2 M ycodrosophila 1/ 1 Chymomyza 3/3 Zapriothrica 1/ 1 Gitona 2/2 * Posterior MalpiglUan tubule in coloratu unbranched. The University of Texas Publication In these forms the tips of the posterior tubules are lightly attached to the gut by tracheae. In these forms also, both the anterior and posterior tubules are differ­entiated, with the proximal portion of the tube being yellow in color, the distal portion being somewhat irregular in shape and white in color. During the course of evolution a change in orientation appears to have occurred, with the tips of the posterior tubules curving around the gut and meeting in the midline. The tips retain their close association with the gut and are lightly held together by tra­cheae. This is, presumably, the most primitive type within the category of ap­posed tips. In most species, however, the walls of the tips have fused, but the inner lining has not broken down and the lumen is not continuous. In these cases also, the tips of the tubules are often much less closely associated with the gut and may be quite independent of it. In most species having the posterior tips of the Malpighian tubules apposed, the individual tubules are differentiated as were those where the tips are free (i.e., a proximal yellow, and thin, white, distal section). In some, however, the posterior tubules lack the distal white section and are of uniform color and diameter all the way to the tips. In species having the posterior tips fused, the inner walls have broken down and the lumen is continu­ous between the right and left tubules. In many of these forms, association between the gut and the tips of the tubules is absent, but in some the association with the gut has been retained. Here, as in the apposed category, the posterior tubules are generally differentiated, but in some the distal white section is absent and the ring formed by the fusion of the tips is of uniform appearance and di­ameter throughout. For brevity, only the three major categories (free, apposed, fused) will be used in the following discussion. As was noted earlier, Figure 41 shows the major types of Malpighian tubules seen in the various groups. A fraction is also included with most figures. The numerator of this fraction indicates the proportion of the group having that type of Malpighian tubule. The denominator indicates the total number of species in­volved. For example, the data from seventy-seven species from the virilis-repleta section of the subgenus Drosophila have been summarized in the figure. Seventy­three species ("all other species," upper left) have the usual type of Malpighian tubules. Twenty-six of these have the tips apposed and forty-seven have the tips fused. Where needed, further explanation will be given below. In most respects species from the subgenus Pholadoris show primitive char­acteristics. Their Malpighian tubules, however, are of derived types. In species from the victoria and coracina groups the posterior tips are apposed, in victoria very loosely so and almost free. In bryani and latifa.sciaeformis the stalks of both the anterior and posterior tubules are relatively very long (Figure 41). In lati­fasciaeformis the posterior tips are apposed, in bryani they are fused. Only one of these types is shown in the figure, and the fraction (2/7) in this case indicates the proportion of Pholadoris species having Malpighian tubules with long stalks. With the exception of populi, all Sophophoran species have Malpighian tubules with the posterior tips free. In populi they are · apposed. In some species of the melanogaster group the stalks of both the anterior and posterior tubules are longer than usual, being about twice as long as in other species in the subgenus. This is particularly noticeable in anana.ssae and takaha.shii. Species in the mon­tium subgroup vary from auraria with short, to seguyi with longer (about 1 ¥2X) Throckmorton: Phylogeny in Drosophila stalks. Members of the melanogaster subgroup and bipectinata have the usual short stalks. Only the usual type is shown in Figure 41. In the subgenus Drosophila species from both sections have the same general types of Malpighian tubules. In the majority of species the posterior tips are fused (see Table 2). Many species from both sections have the apposed type. None have tips which are completely free, although some species from the tri­punctata group approach this type. In these species (e.g., crocina) the tips curve toward each other around the gut and are very loosely tied together with tra­cheae. In this respect they differ from the typical Sophophoran type and so were classified as apposed. Frota-Pessoa (1954) describes other species from the tri­punctata group as having free tips, and presumably these are all of this same general type. Only two major groups of species in the subgenus Drosophila have Malpighi­an tubules which depart from the usual type. These species are in the robusta group from the virilis-repleta section and in the cardini group of the quinaria section. Some species of other groups from the quinaria section vary in the direc­tion of cardini types, as do some of the undescribed forms (e.g., species K, see Figure 41). Species in the robusta group have three types of Malpighian tubules. Those of lacertosa are of the usual type with posterior tips apposed. Those of robusta and sordidula resemble each other in that both the anterior and posterior stalks are very long. In sordidula the stalks are about three-fourths the total length of the tube, in robusta they are about one-half the length of the tube. The posterior tips are apposed in both species. This type of Malpighian tubule is the same as that seen in latifasciaeformis, and it is seen also in species of the cardini group. In colorata the stalk of the anterior tubule is about four-fifths the total length and the posterior tubule is long and unbranched. The types of Malpighian tubules shown in Figure 41 for species of the cardini group represent the two extremes of types seen within the group. Both the an­terior and posterior stalks are very long and the posterior tips are apposed. The Malpighian tubules of species K have stalks markedly longer than the usual type, but they are not as long as those seen in species from the cardini group. In this species the posterior tips are fused. Several other species (griseolineata and sub­badia of the guarani group; uninubes of the rubrifrons group) have Malpighian tubules of this type, and the posterior tips are fused in some, apposed in others. Another cluster of species have Malphighian tubules intermediate between this last type and the usualtype with short stalks. These are: putrida (testacea group), calloptera and ornati pennis (calloptera group), guaramunu, guara;a, guarani ,(guarani group), submacroptera (macroptera group), parachrogaster (rubri­frons group), and albicans, metzii, mediopictoides (tripunctata group) . The in­termediate types between that of the cardini group and normal were omitted when the fraction of apposed vs. fused was calculated for "other species" in Fig­ure 41. The only groups from the quinaria section in which all species have the usual short stalks are the funebris, immigrans, and quinaria groups. However, only in species K, in two species of the guarani group, and in one species of the rubrifrons group does the type approach that of the cardini group. Within the genus Drosophila the type of Malpighian tubules seen in members The University of Texas Publication of the subgenus Phloridosa is distinctive. All of these species have very short anterior tubules. In species P and species Q the anterior tubules are reduced to small lobes at the tip of the stalk. In species 0 they are somewhat longer (see Figure 41). The posterior tips in species P are fused. In the other two they are apposed. In species Q the posterior stalk is somewhat longer than usual and it expands greatly toward its point of union with the posterior tubules. Zapriothrica dispar has Malpighian tubules identical with those described for species 0, except that the posterior tips are fused. Malpighian tubules from species in other genera (see Table 2) are of the usual types. Anterior and posterior stalks are short in all species examined. Other Features of Internal Anatomy-Wheeler (1947) described a structure, apparently attached to the base of the ovipositer, in females of species belonging to the willistoni group. During dissection of all species this feature was checked, and it was found in species of the willistoni group but nowhere else. It varies from species to species in this group, generally being a rather short, corrugated and indistinctly bilobed pouch. In fumipennis it has a long, narrow stalk and a long, fusiform bulb. Itis very conspicuous when fully distended and, as indicated by Wheeler (op. cit.), it appears to arise from the base of the ovipositer. When the vagina is punctured this pouch collapses also, suggesting a connection be­tween them. A similar structure is present in Chymomyza coxata (Wheeler, 1952). . Another interesting structure was noted in peruviana. In this species the rectum appears only as a slight enlargement of the posterior digestive tract. It lacks the usual four rectal papillae. Instead, there are two large lateral pouches arising on narrow stalks from the posterior end of the rectum, and each of these has two typical rectal papillae. The position of these papillae varies somewhat, but generally they are apical in each pouch. These pouches are present in both males and females. Anterior Spiracles of the Pupa-For the most part characteristics of the pupa have been little used in Drosophila classification, although horn indices have been useful as species characteristics. The horn index will not be covered at the present time. In species of the subgenus Pholadoris the anterior horn is very short (almost non-existent) relative to the length of the puparium. It is also short in all Sophophorans. In species of the subgenus Drosophila the anterior horn varies from short to very long. The characteristics of the branches of the anterior spiracle are occasionally noted in species descriptions, but no extensive systematic study has been made of them. They provide one of the best indications of subgeneric relationships available, and it is unfortunate that they have not been reported in greater detail. Representatives of spiracle types for the genus Drosophila are shown in Figures 42 and 43. They are summarized phylogenetically in Figure 44. Types from other genera are shown in Figure 45. Figures are not drawn to scale, so no size comparisons can be made. The type shown for pattersoni (Figure 42.1) is pre­sumably near the primitive for the genus. The branches are short, generally recurved, and arranged in a simple whorl around the spiracle opening. The ring of branches does not make a complete circle. It is interrupted and there is a dis­tinct gap (basal gap) directed dorso-laterally. The branches adjacent to th,e gap Throckmorton: Phylogeny in Drosophila ~~ .3 .4 7~ ~ 9~ .6 .10 /f II~ 12~ I~ 14~ 15~ 16~ .17 FIG. 42. Characteristics of the anterior pupal spiracle . .1 pattersoni . 9 tolteca .17 ezoana .2 novopaca .10 melanogaster .18 melanica .3 bryani .11 takahashii .19 robusta .4 duncani .12 serrata .20 gibberosa .5 thoracis .13 willistoni .21 mojuoides .6 busckii .14 fumipennis .22 longicornis .7 pseudoobscura .15 parasaltans .23 aureata .8 athabasca .16 neoelliptica .24 canalinea .25 camargoi (basals) are short and the following ones increase in size regularly, with thei ones opposite the gap (antibasals) being longest. For descriptive purposes the branches can be divided into three categories: basals, laterals, and antibasals. In some species, particularly from the subgenus Drosopmla, there are branches within the ring and these will be referred to as centrals. For-discussing certain Tlze University of Texas Publication .6 I~'~JI~ ,., " .II )~~"''' ,, .24 FIG. 43. Characteristics of the anterior pupal spiracle . . 1 gaucha .9 subfunebris .17 parachrogaster .2 nannoptera .10 macrospina .18 pallidi pennis .3 carbonaria .11 hypocausta .19 testacea .4 polychaeta .12 innubila .20 macroptera .5 species I (bromeliae gp.) .13 species J .21 neocardini .6 species G .14 tenebrosa .22 griseolineata . 7 peruviana .15 pal ustris .23 tripunctata .8 funebris .16 species K .24 ornatipennis types it is convenient to designate the lateral branch adjacent to the basal as the sub-basal. The sub-basals often differ in position from other branches in the ring, being displaced inward and standing erect rather than being recurved. A spiracle with sub-basals of this type is shown in Figure 42.4 . . The ring is always sym­metrical (in fully everted spiracles), and the two halves are mirror images of each other. All figures except 43.22-.24 show dorsal views of the right spiracle. Throckmorton: Phylogeny in Drosophila FIG. 44. Pictorial phylogeny of anterior pupal spiracles. For further explanation see text. Frc. 45. Anterior pupal spiracles from members of other genera. .1 C. amoena . 4 S. adusta .7 Z. ghesquierei .2 C. aldrichi .5 S. pallida .8 M. dimidiata .3 C. procnemis .6 Z. vittiger All species of the subgenus Pholadoris have the same basic type of spiracle (Figures 42.1-.3). Branches are arranged in a simple whorl and they increase slightly in size from basals to antibasals. The chief variation within the group is in number of branches. Species of the victoria group have eleven, those in other groups have five to eight. The spiracle of victoria differs from the others The University of Texas Publication in having sub-basals erect and slightly displaced inward. Branches are uniform in color and are of the same color as the puparium. Species from the obscura, saltans and willistoni groups of the subgenus Sopho­phora have the same general type of spiracle as was seen in Pholadoris species. Branch number varies from seven to eleven in species of the obscura group, and all members of this group have the sub-basals erect and displaced inward (Figures 42.7-.9). In some members of the group the sub-basals are rather strong and almost completely displaced inward so that they appear to be within the ring (Figure 42.9) . In most of these species the branches form a flat whorl, al­though in some they are more nearly erect (this varies with age and dryness of pupa) . In narragansett the branches are shorter and strongly recurved. In most cases the length of the branches increases regularly and markedly from basals to antibasals and the color of the branches is brown with tips slightly darkened. The number of branches varies from nine to eleven in members of the willistoni group. The type shown in Figure 42.13 represents willistoni, equinoxialis, tropi­calis and paulistorum. That shown in Figure 42.14 represents sucinea, capricomi, nebulosa and fumipennis. Pupae of changuinolae and pseudobocainensis were not seen. In most species of this group the branches are uniform brown, although in some the tips are dark. The number of branches in species from the saltans subgroup varies from eleven (parasaltans) to seven (neoelliptica and emargi­nata), and all have erect sub-basals (Figures 42.15-.16). Color of branches is variable. Spiracles from species in the melanogaster group are difficult to interpret. In most species the branches, particularly the antibasals, are much longer and more erect than in the previous type. The number of branches in the outer whorl varies from seven (yakuba) to eleven (ananassae). Members of the montium and ananassae subgroups have from two to four distinct long central branches (Figure 4:2.12). It seems probable that at least two of these represent sub-basals displaced inward. In melanogaster there appear to be two branches within the ring but their position and appearance suggests that they are enlarged sub-basals. In takahashii (Figure 42.11) and yakuba the true basals appear to be almost miss­ing (only vestiges seen, and these were not shown in the figure) and the enlarged sub-basals now occupy the position of the basals in the ring. This interpretation is, however, quite tentative. The branches in montium (except serrata) and takahashii subgroups are black with light tips. Those of the melanogaster sub­group are a uniform brown. Spiracles from species in the subgenus Drosophila are of several types (Figures 42.17-.25; 43.1-.24), and all of them are distinct from those types seen in mem­bers of the subgenus Pholadoris and in Sophophoran species. The type shown in Figure 42.17 is restricted to members of the virilis group. Branches are heavy, nearly erect, and recurved only at their tips. There is a very marked and regular increase in length from basals to antibasals. In most species the median edge of the spiracle (from which the antibasals arise) is extended as a lip. This is most extreme in ezoana (figured) . There are thirteen branches in the outer ring and a cluster of about four centrals which appear to originate at the bases of the antibasals. There are also some rudimentary central branches. Branches are brown with dark tips. Throckmorton: Phylogeny in Drosophila Although the number and length of branches varies, most of the remaining species from the virilis-repleta section have spiracles of the same general type (the robusta type, Figure 42.19). That of aureata (Figure 42.23, and see Wheeler, 1957) is an aberrant type seen only in this one species. The type shown in Figure 42.20 is restricted to members of the annulimana group, to eremophila (mulleri subgroup of repleta group) and to melanura (melanica group). A very similar type (Figure 42.21) is seen in several members of the fasciola subgroup of the repleta group. Among these last species there is an intergradation of types, with fulvalineata having spiracles of the robusta type and others being intermediate between the robusta type and the annulimana type. The remainder of the species from this section (Figures 42.24-.25; 43.1-.7) are all of the robusta type. In all of these the branches are heavy and increase regularly in length from basals to antibasals. There are always at least two centrals present and the basal gap is always distinct. Generally the color of the branches is brown with dark tips but in castanea the branches are almost black. There are several distinct spiracle types from species in the quinaria section (Figures 43.8-.24). Those of species in the funebris group vary (Figures 43.8­.10). In one, funebris, the spiracle is of the robusta type. In subfunebris the basals are very small and two of the centrals have moved out so that they are almost on the ring. In macrospina the basals are absent and there are four long branches in their position. Judging from the situation in subfunebris, it would appear that the basals have been lost and the centrals have moved out to the ring, but this is only speculative. For convenience, these long branches in the basal position will be referred to as pseudobasals. The present terminology should be considered only as descriptive. Homologies of various branches can probably be determined eventually, but they must remain uncertain for the present. The type seen in macrospina is seen in the great majority of species from the quinaria section. This type can be defined as follows: basal gap very small or absent, pseudobasals present and longer than laterals, generally equalling antibasals in length, centrals always present. All species of the immigrans group have spiracles of this type, and it will be referred to as the immigrans type (Figure 43.11). Those of spino­femora have two pseudobasals, those of hypocausta (figured) and immigrans have four. Laterals are very short and the number of centrals varies from three to six. Spiracles from some species in the quinaria group resemble the annu­limana type (compare 42.20 and 43.12). Although the number of branches varies, innubila (figured), phalerata, falleni and transversa are all of this type. Species J (Figure 43.13) and occidentalis differ from this type primarily in having a short and erect pair of pseudobasals. The type seen in guttifera and tenebrosa is the same as the latter type except that the branches are longer (Figure 43.14). The spiracles of palustris (Figure 43.15) and subpalustris are completely of the immigrans type. Several species from the tripunctata, testacea, macroptera and rubrifrons groups have spiracles resembling those seen in tenebrosa (a pair of short pseudobasals), and these can easily be confused with the robusta type if they are not checked closely. In these species the branches are longer and more slender than those of tenebrosa. Most species of the tripunctata group and all species of the guarani and calloptera groups (schildi not seen) have spiracles of the immigrans type. Three representatives are shown in face view (looking The University of Texas Publication directly toward the pseudobasals) in Figures 43.22-.24. The spiracle branches of guaraia are very long and delicate. They seldom stand erect, so this type ap­pears to be distinct from the others. Relative length and position of branches, however, is as seen in the immigrans type. The species of the cardini group (Figure 43.21) and species K (Figure 43.16) and L have unusual types of spiracles. In some of these species the number of branches is reduced (neocardini has the lowest number, belladunni the highest). There is a pair of very long and recurved pseudobasals, then a pair of up-curved small laterals. The remainder of the laterals are short and recurved, while the antibasals are long and tend to be recumbent rather than erect. Types having the higher branch numbers tend to resemble the immigrans type more closely, but most of them are distinct from it. Color of branches is variable from species to species, but most of the branches have black bases (and/ or the edge of the spiracle itself is black) and the tips are pale. This same color pattern is seen in many spe­cies from other groups in the quinaria section. Species from the subgenus Hirtodrosophila have very variable spiracles. Those of duncani are of the Pholadoris type with erect sub-basals (Figure 42.4). Those of pictiventris and thoracis (Figure 42.5) are of the immigrans type. Those of histrioides resemble the robusta type, except that the branches tend to be recum­bent rather than erect. The spiracles of busckii resemble the immigrans type. Spiracle types seen among species from other genera are shown in Figure 45. Those of Chymomyza species vary as shown in Figures 45.1-.3. That of C. amoena (Figure 45.1 ) is basically as in obscura group species, except that the rim of the spiracle is extended slightly as an antibasal lip. This extension is ex­treme in C. procnemis. Spiracles from Scaptomyza species (Figures 45.4-.5) more nearly resemble Chymomyza types than any other. The two species from the genus Zaprionus have two types of spiracles (Figures 45.6-.7). One is of the robusta type, the other of the immigrans type. The spiracles of M. dimidiata were difficult to draw. Basically there is an outer ring of eleven, increasing regularly from basals to antibasals, and an inner ring (all almost equal) of seven. Some of the centrals tend to crowd the basals and may give the appearance of pseudo­basals. In summary (Figure 44) types of anterior spiracles provide very useful char­acteristics, both for separating subgenera and for separating some groups within subgenera. When investigated in greater detail the arrangement of branches may also provide useful specific characteristics. The figures give only a limited indica­tion of the distinctive features of each type. The types seen in species of the melanogaster group were particularly difficult to draw. However, once seen, melanogaster types cannot be mistaken for other types, at least among the species included in the present discussion. Among species from the subgenus Drosophila, spiracle types rather sharply distinguish species of the quinaria section from those of the virilis-repleta section. Other characteristics of the pupae-Aside from the characteristics of the an­terior spiracle, presently useful characteristics of the pupae are limited. Morpho­logical characteristics of the posterior spiracles vary both within and between species and they follow no distinct pattern. Color of posterior spiracles is generally that of the puparium. In several species from the quinaria section the posterior Throckmorton: Phylogeny in Drosophila spiracles are dark, at least basally, and in some they are shining black. If no dis­tinction is made between these two types, the distribution is as follows: immi­grans, 2/3; guarani, 3/5; tripunctata, 9/ 13 (the fraction indicates the proportion of the group having dark posterior spiracles). A feature of possible future use may be found in the characteristics of the pair of anal plates which lie lateral to the anal pore. These vary in shape and dis­ tinctness from species to species. The anal pore also varies in shape and in some species it appears double. These features have not been investigated in any detail. Color and position of posterior spiracles appears to be the same in both larvae and pupae, so characteristics of the larvae will not be considered separately. Egg Filaments-Information regarding egg filaments is summarized phylo­genetically in Figure 46. The phylogeny is the same as that used for spermathe­cae, etc. Types seen in Scaptomyza and Chymomyza species are also included in this figure, but they are not placed in the phylogeny. Egg filaments in species from the subgenus Phola.doris are thin, irregular in position, and variable in number. In the few eggs checked from these species counts ranged from five to nine. Wheeler (1949a) tabulates filament numbers from three species (victoria, nitens, latifasciaeformis) with a range from two to eleven. An egg with six fila­ments is shown in Figure 46 since this is near the modal number for the group as a whole. All Sophophoran species have eggs with two filaments. Those of populi are short and heavy, and they continue as narrow ridges to the posterior end of the egg. In the melanogaster group all takahashii and montium subgroup species (except nikananu) have egg filaments which are thick at the base, then irregu­larly tapering. The egg filaments of nikananu are of the more usual type for the subgenus and are seen also in species from the obscura, saltans and willistoni groups. In this type the filaments are heavy at the base, then very much flattened and expanded apically (see Figure 46). Egg filament number among species from the subgenus Drosophila ranges from one (e.g., bandeirantorum, tripunctata group) to four. The great majority of species have four-filament eggs. Aside from variations in length of filament, four-filament eggs can be roughly divided into two classes. In the first, the pos­terior filaments are thick and flattened basally, tapering gradually to a fine point apically. The anterior filaments are thin and fine. This type is seen in two species of the immigrans group, species of the calloptera group ( schildi not seen) , in species of the annulimana group and in castanea and species F. In the second type all four filaments are thin, although they may expand very slightly at the base. This type is shown in Figure 46 for species of the virilis group, funebris group,. etc. It is present in approximately 80 per cent of the species in the subgenus. Species from all of the groups not included in Figure 46 have egg filaments of this type. D. aracea (not figured), a species not assigned to any group, has four­filament eggs in which both the anterior and posterior filaments are heavy and flattened basally, tapering gradually to fine points. Two-filament eggs are present in all species of the melanica group, bromeliae· group and nannoptera group. They are present also in hypocausta of the immi­grans group and have been reported (Frota-Pessoa, 1954) in some members of the tripuncata group. D. aureata (not figured) also has eggs with two filaments. Egg filaments among species in the melanica group are thick and taper only to <.O o~ Scaptomyza er to ~ ~ ~ / C::! i:!. ~­ .Q.. ~ ~~\'\~ ;f·W V'\\ Jl:~~u U }f ~ ~ ~ g: ~ §" 5· ;:s populi * o~ ~ OmmOg""'0~ ~ ''"' Subgenus . a irilis-repleta ""-J quinaria section section ~"'""'"""'~ enus ~ ~ Subgenus Drosophila / oladoris @ Chymomyza FIG. 46. Pictorial phylogeny of egg filaments. For further explanation see text. Throckmorton: Phylogeny in Drosophila slightly. Those of the bromeliae group are also thick and somewhat more distinctly tapering. In nannoptera they are flattened apically as in some Sophophorans, and in hypocausta they are very long, and thin throughout. In aureata the filaments are short, heavy and blunt. Species of the quinaria group have three-filament eggs, the two anterior ones are thin, the single posterior one is short and heavy. Most species in the tripunc­ tata group have the usual four-filament type, but some species have one (figured), two, or three egg filaments. Members of the subgenus Hirtodrosophila have four­ filament eggs, with those of thoracis being rather peculiar. In this species the posterior filaments are short, blunt lobes, and the anterior ones are very short and extremely thin. In spite of their thinness they are quite rigid and perfectly straight. The eggs of the other three species are of the usual four-filament type, as are those of busckii (Dorsilopha). Eggs of Phloridosa species were not seen, but those of fioricola are without filaments (Sturtevant, 1942). Eggs from species in the genus Chymomyza are distinct. The general type is shown in the lower right corner of Figure 46. There are approximately eight short, in-curved filaments decreasing in size regularly from posterior to anterior. Types seen in Scaptomyza species are shown in Figure 46 also (upper center). That of S. pallida (left) has two short heavy posteriol'. filaments. That of S. adusta (right) has four short, heavy and bent filaments. The eggs of Zaprionus species have four long filaments. They are all heavier than in the usual four-filament type. Eggs of M. dimidiata have four filaments. In all four the basal one-third is heavy, the remainder thin. Characteristics of the Abdominal Sternites of the Male-Wheeler (1960) ·describes and discusses sternite modifications in male Drosophilidae. Number of sternites varies from six in more primitive forms to four in the more derived species. Reduction in number involves changes in, or eventual loss of, the first and sixth sternites. Detailed demonstration of sternite characteristics requires clearing of the ventral abdominal wall in phenol and examination at the higher magnifications of the compound microscope. Time did not permit a study of this sort, but it was possible to check those features which could be seen in the living fly using the dissecting microscope. At these magnifications well-chitinized and pigmented plates can be detected and, using the characteristics of the second sternite (see Wheeler, op. cit.) as a guide, the specific sternites represented could be determined. Results of this limited survey are of interest, and they are useful so long as the limitations of the observational methods are recognized. Most species of the subgenus Pholadoris have visible remnants of the first and/or the sixth sternite. In species of the victoria group the first sternite is present as two plates (see Wheeler, op. cit., for figure of victoria) and the sixth sternite is present as a plate comparable in size and appearance to the o+her abdominal sternites. In species of the coracina group no first sternite was visible. (In the following descriptions, "not visible" should not be interpreted as imply­ ing absence. The sternite in question may be present but not pigmented and of any of several forms, or it may actually be absent.) In novopaca the sixth is present much as in victoria. In cancellata and lativittata the sixth is present as a thin, chitinized bar. In bryani and latifasciaeformis the first is not seen. In The University of Texas Publication bryani the sixth is a narrow bar, concave posteriorly so that it has a shallow U­shape. A sixth was not seen in latifasciaeformis. All Sophophoran species have the sixth sternite present as a well-chitinized, but generally not bristled plate. In most species this plate is somewhat polished in appearance. Shape of the plate varies. It is largest in populi, very narrow and almost interrupted in the midline in several species of the melanogaster group. In populi the first sternite is present as two bristled plates much as in victoria. In most of the remaining Sophophorans the first sternite is not seen. Two small dark plates in this region are seen in neoelliptica, emarginata and milleri (saltans group), although there is apparently individual and interstrain variation of the first sternite in some of these species (Malgalhaes, this Bulletin). With two exceptions, first and sixth sternites are not seen in species from the subgenus Drosophila. In montana (virilis group) there may be a thin bar in the position of the first sternite, and there are faint indications of this in littoralis and ezoana. In the last two species there is considerable individual variation. If the presence of the first sternite as a bar is verified, this fact would be of consider­able interest. The available evidence, while fragmentary, suggests that in most evolutionary lines in the genus the first step toward loss of the first sternite has involved its separation into two pieces. The presence of a complete, but narrow, bar in some species of the virilis group would suggest that loss of the first sternite may have evolved independently in this group. More detailed investigation of this characteristic may prove fruitful. In the remainder of the species belonging to the virilis-repleta section there was no visible evidence of either the first or the sixth sternites. In some species there were a few scattered bristles in the region of the sixth (in position suggesting that the sixth had split in the midline, then moved laterally to fuse with the seventh tergite) but observations were too limited to warrant further description. The second exception involves the presence of the sixth sternite in testacea. In this case the sternite is rather polished and equals the fifth sternite in width. There is a conspicuous bristle at each lateral margin. Among species of the quinaria section no other structures were seen which could be interpreted un­ equivocally as a sixth sternite, although many species of the cardini group ap­ pear to have the anterior wall of the genital chamber weakly chitinized, suggest­ ing the possible presence of a remnant of the sixth sternite in this region. The first and sixth sternites are absent in Phloridosa species, in busckii (Dor­ silopha), and in most species from the subgenus Hirtodrosophila. In duncani the sixth is present as a narrow bar. The sternites in members of other genera are variable. The first and sixth were not visible in R. obesa. (In this species, something is present in the position of the sixth sternite, but its characteristics could not be determined by methods used here.) In G. americana the first is a very large plate, wider than the other sternites and concave at its anterior margin. The sixth is present, but turned on edge to lie along the anterior wall of the genital chamber. In Zapriothrica dispar the first is a broad plate, concave at its posterior margin, and the sixth is present. as a narrow, rectangular and weakly bristled plate. In species from the genus Chymomyza both the first and the sixth sternites vary. In C. aldrichi the first and the sixth are both present as paired plates. In C. amoena the first is not visible Throckmorton: Phylogeny in Drosophila and the sixth is present as a large U-shaped plate (arms directed posteriorly), and the arms of the U are weakly bristled. In C. procnemis the first was not seen, the sixth is present as paired plates (see Wheeler, op. cit., for figure). No first or sixth stemites were seen in Scaptomyza, Zaprionus or Mycodrosophila species. GENERAL APPROACH FOR THE ANALYSIS OF PHYLOGENETIC RELATIONSHIPS It has often been stated that the characteristics of a population reflect its evo­lutionary history, and this postulate should be applicable to the characteristics of groups of species as well as to those of single populations. Most often, this statement is applied to the genetic structure of a population, but morphological characteristics are a reflection of the genotype, and if this truisln is to have more than an abstract significance it should be possible to infer the history of different genotypes from the distribution of the morphological characteristics which they determine. The basic problem of phylogeny is that of detecting genetic continuity, and it is therefore necessary that phylogenetic analysis be made in terms of genotypes and their history. If the characteristics of populations actually do reflect the history of the populations, such an analysis should be possible. The following discussion is directed toward developing a method of phylogenetic analysis based on this assumption. When phylogenetic analysis is to be approached in genetic terms it is first necessary that correlations be made between morphological characteristics and the genotypes which determine them. The difficulties inherent in making such correlations are well recognized, and some of the problems involved have been discussed by Dobzhansky (1959), Mayr (1959), Simpson (1961) and others. For the most part these discussions emphasize the fact that morphological identity and genetic identity need not be precisely related, and emphasis of this point is certainly justified. Intentionally or not, however, such discussions recognize only the negative aspects of the problem, and they appear to assert that inferences regarding the history of a genotype cannot be drawn from consideration of morphological characteristics. Emerson (1961) has pointed out a logical de­ficiency of such a position, and a completely negative approach does not appear to be justified. Several considerations suggest that a certain amount of correlation between genotype and phenotype can reasonably be assumed. First among them is the fact that any operational, and to a large extent any theoretical, approach to phylogenetic analysis must be concerned primarily with the most probable course of events in an evolving system. It is certainly necessary to recognize that some things, e.g., mimetic mutations, can happen. It by no means follows that one must assume such events to happen routinely, or even with such frequency as to seriously affect the validity of conclusions regarding the phylogeny of a group as a whole. As stated by Emerson (1961), "the replicative capacity of the genetic system is even more apparent and precise than its capacity to change," and it is this stability which provides a reasonable basis from which to infer the most probable ways in which genotypes may come into being and change during the course of evolution. The stability of the genetic system, and its limitations with respect to its The University of Texas Publication capacity to change rapidly (Haldane, 1957), suggest that a considerable portion of the genetic material in related species will have been derived substantially unchanged from their ancestral populations. Still, genetic differences below the level for morphological detection must be recognized as possible, and it is there­fore necessary to make a distinction between gene homology and genotypic homology as these terms are to be used in making taxonomic comparisons. For gene homology the basic unit of reference is a specific allele at a specific locus. For genotypic homology the basic unit of reference is the portion of the gene pool which directly influences the expression of a given phenotype. When making comparisons between species, and when inferences are drawn regarding geno­typic homologies underlying their morphological similarities, the genetic com­ponents compared cannot be less than the total controlling the expression of the characteristic throughout the range of the species and throughout their seasonal fluctuations. A given morphological feature of the individuals within a species will often, if not always, be under the control of a large number of genes, and the characteristics of any two individuals may not rest on identical genotypes. It would therefore be quite unrealistic to adopt a criterion for genetic homology which could not be applied uniformly to comparisons between individuals of one species. On this basis, for example, complete gene identity would not be a suitable criterion. Simpson (1961, p. 78) has defined morphological homology as "resemblance due to inheritance from a common ancestry," and this definition applies equally well to genotypic homology. In the case of genotypic homology we deal with the inheritance from a common ancestry of the genetic elements which comprise the complex genotypes underlying the morphological expression of a given charac­teristic seen in several species. A large number of loci will be involved and complete identity of all alleles at all loci is not required, both because apparently identical phenotypes within a species may not rest on identical genotypes and because complete identity of structure between species may not be involved. In this sense then, the taxonomist deals with the degrees of genotypic homology rather than with the absolutes of gene homology, and the assumption that more nearly identical phenotypes reflect more nearly identical genotypes appears to be justified. Genotypic homology neither requires nor precludes complete gene identity. It does require that a large fraction of the genes controlling the expres­sion of a given characteristic be derived from a common ancestral gene pool. A special aspect of the problem of genotypic homology relates to the possibility that a developmental pattern, and therefore a particular phenotype, may be adaptive and so be perpetuated, but the genotype controlling this pattern may be sharply changed during evolution (Dobzhansky, 1959). Since the simplest and perhaps most effective way to retain a developmental system is to retain the genotype which originally produced it, this type of change may not be an inevit­able consequence of evolution. Operationally, however, the reality of this possi­bility does not influence phylogenetic analyses. If a specific developmental pat­tern is to be perpetuated in two separately evolving lines, it must have been present in their common ancestral population. When morphological homologies are identified and genotypic homology inferred, the proper inference is not that . the present genotypes are identical, but that the ancestral genotypes at the time Throckmorton: Phylogeny in Drosophila of the initial divergence were identical. Genetic continuity is determined by con­ditions at the time of separation of two populations, not by the terminal geno­types of these populations. Whether the terminal phenotypes result from reten­tion of identical genotypes or from "gradually evolving functional analogy" (Emerson, 1961), the phylogenetic relationships suggested by these phenotypes will remain the same. Finally, if gene complexes within the species gene pool are taken as the basic unit of reference for making taxonomically relevant identifications of genotypic homology, the possibility for the origin of identical or closely similar phenotypes through parallel mutation becomes rather small. For this to happen, substantially similar mutations would have to occur at a large number of loci and in such a genetic context that selection could mold them into an integrated gene complex which would be expressed as a phenotype so similar as to be confused with that produced by another system of independent evolutionary origin. If only one gene is involved this chain of events is conceivable. When gene complexes are considered, it becomes highly improbable. The possibility is certainly not great enough to invalidate the assumption that morphological identity indicates sub­stantial genotypic homology, or to require us to discard concepts of genotypic homology as theoretical or operational tools. This should not be taken as imply­ing that the possibility for independent origin of "homologous" phenotypes is something to be ignored or lightly dismissed. The problem still exists, but it is not a problem which should be of first concern when preparing a working model for phylogenetic analysis. In passing it may be well to emphasize one feature of the chain of events just outlined which will be of significance later. This is the fact that, at the time of its mutational origin, whether this involves one mutational event or several, a necessary condition is that a new gene be adaptive. And this requires not only a congenial genetic background but also a congenial environment. Therefore, if a particular gene or genotype is to be established in a population, it must not only occur in a gene pool in which the appropriate genetic background is available (i.e., a background in which it may be expressed as a specific or mimetic pheno­type), but this gene pool must also be present in an environment such that the new allele be adaptive. The establishment, which is not to say the fixation, of a gene in a population, i.e., its primary integration into the gene pool, is the critical evolutionary step which makes that gene or genotype available for future change. It is this integration, rather than the origin of the gene by mutation, which has such a low probability of occurrence in independently evolving gene pools. The same mutational events probably occur many times, but only those which occur in the "right" context are of phylogenetic significance. The probability that the "right" context has been produced independently several times in several phyletic lines is almost surely small. The possible fates of genotypes in evolving systems-As stated previously, the basic problem of phylogeny is that of detecting genetic continuity. Before this problem can be approached it is necessary to determine the ways in which geno­types may traverse a succession of gene pools and the ways in which morpho­logical characteristics may be distributed to descendent populations. For sim­plicity, the genotype for only a single trait will be followed. Also for simplicity, The University of Texas Publication it is convenient to treat the genetic system which determines the trait as involv­ing only a pair of alleles at a single locus. Thus, the following discussion is a great over-simplification, phrased in the language of "bean-bag genetics" (Mayr, 1959), but it illustrates in general terms the possible sequences of genetic events which may have occurred during the speciation processes which established the various phyletic lines within the genus. As a starting point a hypothetical population, genetically uniform (AA) with respect to the determiners for a given characteristic, may be assumed. In the course of time, through mutation and selection, this uniformity is lost and new genetic complexes arise (AA') which have a direct effect on the trait in question. It is not strictly necessary that the phenotypic expression at this time (AA') be specified, but it seems probable that it would be shifted somewhat from the earlier form and that this altered phenotype, whether defined in morphological or physiological terms, would be maintained in the population by one or several of the balance mechanisms currently postulated as being of importance in the perpetuation of genetic heterogeneity within a population. Mechanisms adequate for preserving heterogeneity have been proposed and discussed by many workers (e.g., Dobzhansky, 1951, 1955; Dempster, 1955). For the present discussion, mechanisms, so long as they exist, are not the important thing. What is important is that, during the course of evolution, the transition from a population showing one form of expression of a trait to another showing a derived form of the trait must involve temporally intermediate populations heterozygous for the geno­types determining the alternative forms of expression. If such a heterozygous population can come into being and be maintained for an evolutionarily signifi­cant period of time, there is a distinct problem concerning the fates of the possible genotypes (AA, AA', A'A') which will be produced within this gene pool. Theo­retically, there are three possibilities. First, the old genotype may be completely replaced by the new (AA to AA' to A'A'). Secong, an equilibrium level may be attained and perpetuated so that both genes (genotypes) persist in the population (AA to AA' to AA', etc.) . Third, the direction of selection may be reversed, and the old genotype will again predominate (AA to AA' to AA). There is, in addi­tion, the problem of the hypothetical, heterozygous population if, for one reason or another, it becomes subdivided and its subdivisions diverge genetically to such an extent that they would be recognized as distinct species. Again, the possib1e fates in these subpopulations are the three just listed, and we cannot say what would happen in any instance. There is no a priori basis for asserting that only one or another of these events occurs during evolution. The first alternative (AA to AA' to A'A') is often the only one recognized by practicing taxonomists or phylogeneticists. A discussion by Hennig (1956) pro­ vides an example of an approach which recognizes only this alternative. Here he says, "Nothing is more obvious than the conclusion that two species which share the derived form of a characteristic have acquired this from a common stem species. This would, accordingly, indicate that they are more closely related to each other than to species which do not show this form of the characteristic" (writer's translation and italics) . The general sequence of mutation and specia­tion events which would be required for the second statement (italicized) to hold true is shown in Figure 47 (left). Minor changes could be made in this sequence. Throckmorton: Phylogeny in Drosophila species species species species species species y A B w x z A 1 A' ti. A' A' A' AA AA' A' ti. \I \I \I ti. ti. Ats.. AA' ~/ AA' species c A+A' AA AA \I AA AA i i FIG. 47. Two alternative evolutionary sequences. The one on the left illustrates evolution according to the "classic" hypothesis. The one on the right illustrates a possible sequence based on the "balance" hypothesis. See text for further explanation. For example, the mutations of A to A' could have occurred during the first specia­tion process, but only in the subpopulation shown on the left. Or, if they occurred in both subpopulations, which seems more probable, those in the subpopulation at the right would have experienced strong negative selection while those in the subpopulation at the left would have been strongly selected for. Such slight shifts, however, would not alter the overall requirements or the general pattern. In this sequence species A and B share the derived form of the trait and are phylo­genetically more closely related to each other than to species C, which shows the primitive form of the trait. Such a series of events has undoubtedly occurred during the formation of many groups of species. However, other sequences must be recognized as possible, and one such is shown in Figure 47 (right) . In this sequence, the A' allele is established, but not fixed, in the population prior to the first speciation event, and heterozygosity at this locus is maintained during and after speciation. Thus, the next speciation event in each new phyletic line in­volves a population heterozygous for the locus in question, and the alleles (geno­types) may segregate by selection in various ways to descendent populations. Among the possibilities shown, species W and Z share the derived form of the The University of Texas Publication trait. Species X shows the primitive form and species Y might show some inter­mediate form of expression. If genetic and speciation events followed some pattern such as this, at least in some instances, the assumption that two species showing a derived form of a characteristic are most closely related would not hold true. Species W and Z share the derived form of the trait, but species W is phylogenetically most closely related to species X, which shows the primitive form of the trait. . The two models, Figure 47 left and right, are related somewhat in the same manner as are the "classic" and "balance" hypotheses of population structure outlined by Dobzhansky (1955). That on the left follows the "classic'' system, with its populations being primarily homozygous and its genetic diversity "either neutral, or transient or morbid" (Dobzhansky, 1955). In the example, the genetic diversity is transient and interposed between speciation events. The sequence on the right incorporates elements of the "balance" hypothesis, at least insofar as it emphasizes the role of heterozygous populations in evolution. To some extent, events shown in the sequence on the right may involve homoselection and hetero­selection as defined by Carson ( 195 9) . With reference to problems of phylogeny, if speciation may occur in this fashion (Figure 47, right; and see Carson, 1959), the inevitable result will be the perpetuation, for a considerable period of time, of a heterozygous population potentially capable of giving rise to descendent populations homozygous or heterc ozygous for alternative genotypes (species Y in Figure 47 could continue to evolve and speciate in this fashion, etc., etc.). Thus, within a phyletic line where evolution has occurred on some variation of this pattern, alternative forms of a trait may be distributed somewhat at random. Possession by two species of a particular characteristic will indicate only that they are derived from some com­mon heterozygous population, which may be quite distant from either of them. They may belong to separate phyletic lines, each taking its origin from the heter­ozygous population in which the potentiality for segregation first appeared. A "sequence" of morphological types will not necessarily indicate an evolutionary sequence in the sense that species showing the primitive form appeared first, those intermediate appeared second, and those showing the derived form appeared last. Thus, before one can determine what significance is to be attributed to a morpho­logical sequence of types, one must determine whether the evolution of the group has followed a pattern similar to that shown on the left in Figure 47 or similar to that shown on the right. These two patterns themselves only represent ex­tremes, particularly if more complex genetic systems are considered and if several characteristics are considered simultaneously. A complete spectrum of alterna­tives lies in between, and the actual evolutionary pattern of a group should· fit somewhere in the spectrum of possibilities. Although the alternatives just noted appear to be possible, this alone does not make all of them probable. Events may be possible without their making any significant contribution to evolution. It is therefore necessary to seek evidence which might give some indication that evolution would be expected to follow, and at times has followed, a pattern similar to that outlined to the right in Figure 47. At a general level, the fact that heterozygosity contributes to the evolutionary potential of a population is widely recognized, and there is more than ample Throckmorton: Phylogeny in Drosophila evidence that heterozygosity exists in Drosophila populations in nature (e.g., Dobzhansky, 1951; da Cunha, et al., 1959). When developing a phylogeny from living species, only those ancestral populations which gave rise to descendent species are detected. These are the populations which, at a given time level, had the greatest evolutionary potential, and, as a rough approximation, their evolu­tionary potential can be related directly to the number of species derived from them. It seems reasonable to assume that the detectible populations depended, at least in part, upon heterozygosity for their demonstrated capacity to evolve. Further, it seems reasonable to assume that those populations which continued to evolve most rapidly, and which produced the most descendent species, were able to do so because they fell heir to a part, or all, of the heterozygosity of their ancestral population and did not have to accumulate a complete new store of variability from which to mold adaptive genotypes. Conversely, a population which lost a large share of its genetic diversity, perhaps through homoselection, would have little, or at least less, capacity to proliferate rapidly. Before such a population could proceed to further subdivide and speciate, it would need to build up a new store of variability from which to produce a diversity of adaptive genotypes. Thus, in rapidly evolving populations heterozygosity may have been perpetuated, and, if so, it will have provided some of the genetic variability from which descendent adaptive genotypes have been fashioned. The assumption that all, or even most, speciation events entail a restriction or elimination of genetic variability, so that future evolution must depend on and reflect only variability injected into the gene pool by new mutation, seems unnecessarily rigid and not completely consistent with available information. Some speciation events almost surely do entail a substantial loss in heterozygosity, but it is not anticipated that populations passing through such evolutionary bottlenecks will have contributed as significantly to living populations as have those whose genetic variability has been less restricted. And there is no reason to assume that genetic determiners for characteristics of taxonomic value have been immune to retention or per­petuation in the heterozygous state, or that stem populations for families, genera, species groups, etc., have been conveniently homozygous. While these speculations may have some merit of plausibility, it is still neces­sary to obtain more nearly objective evidence that such evolutionary sequences have occurred. For this we turn to an evaluation of the distribution of morpho­logical characteristics among species for which a defined phylogeny is available. The most convincing and extensive evidence to this end comes from species in the repleta complex of the subgenus Drosophila. Extensive investigations, pri­marily by Wasserman (Wasserman, 1960, and This ·Bulletin), have provided a cytological phylogeny for these species (see Figure 25), and this can be cor­related with the morphological characteristics already described. Such correla­tions have been made for the characteristics treated in this study, and they all show the same general pattern. Thus, details for only two of these, the charac­teristics of the spermathecae and of the ejaculatory bulb, will be summarized. The complete data for spermathecae from species of the repleta complex are given in Figure 29. Figure 48 shows a portion of this data as it appears when com­bined with phylogenetic information from Figure 25. It should be emphasized that the cytological data show the various branches and single species to be de­ TheUniversity of Texas Publication FIG. 48. Evolution of spermathecae in the repleta complex. Based on the cytological phylog­ eny given in Figure 25. .1 castanea .8 pegasa .15 nigrohydei .2 meridionalis .9 mercatorum .16 bifurca .3 mulleri .10 limensis .17 briegeri .4 pachuca .11 canapalpa .18 paracanalinea .5 fulvalineata .1Z fulvimacula .19 canalinea .6 moju .13 peninsularis .7 buzzatii .14 hydei rived from the repleta standard gene arrangement. Cytological data give no information regarding the order or position of these branches relative to each other_ The arrangement of branches in Figures 25 and 48 is that most compatible with morphological characteristics of the different species. In Figure 48 the Throckmorton: Phylogeny in Drosophila position of castanea and of species of the canalinea and dreyfusi groups has been shifted downward. This was done for convenience in making the figure and no particular significance should be attributed to it at the present time. As has been noted previously, no primitive spermathecal type can be desig­nated, although that shown in Figure 48.18 seems a reasonable choice. Since the major problem is that of the role of heterozygous populations during evolution, designation of primitive types is not critical. The behavior of primitive types can be followed more readily with characteristics of the ejaculatory bulb, and this aspect of the problem will be considered later. Species of the repleta complex present several spermathecal types, and, as can be seen from inspection of Figures 27-30, the majority of these types are restricted to this section of the genus. The distribution of types among these species is, however, not completely regular. If only major branches are con­-sidered, the type shown in Figure 48.3 is seen in members of two branches (Figures 48.3 and .13) and in several species derived independently from the standard (e.g., Figure 48.7) . The majority of species having this type, however, belong to the mulleri complex shown to the upper left in Figure 25. The type shown in Figure 48.6 is seen in the majority of species from the fasciola subgroup and in both species from the mercatorum subgroup. Again, each subgroup belongs to a different phyletic line. The type shown in Figure 48.14 is seen in three mem­bers of the hydei subgroup, and a very similar type is seen in species of the dreyfusi group (Figure 48.17). Other types (Figures 48.4 and .16; 48.2, .5 and .15; 48.18 and 29.27; etc.) show similar distributions, but their resemblances are not as sharp as those within the types first cited. It is obvious that a basic tenet of the classical approach to phylogeny, that species sharing a given trait are phylogenetically more closely related to each other than to species lacking the trait, does not hold when applied to relationships among species of the repleta complex. Almost exactly the same type of distribution is seen with the characteristics of the ejaculatory bulb (Figure 49). We have good reason to assume that the simple ejaculatory bulbs (e.g., Figure 49.4) are the more primitive. Ejaculatory bulbs shown in Figure 49.4 and 49.8 are virtually identical, and they are pre­sumably near the primitive, not only for the repleta complex but also for the genus as a whole. The type shown in Figure 49.5 is derived, and it is widely distributed among the various branches of this complex (Figures 49.2, .5-.7, .9­.10, .20). The type shown in Figure 49.3 is found in only two of the branches (Figures 49.3, .11, .13-.14), and that shown in Figure 49.12 has thus far been identified only from members of one branch (Figures 49.12, .15-.16). Here also then, we have a pattern similar to that shown to the right in Figure 47, and a comparable pattern is seen for the distribution of number of testis coils among these same species (see numbers in parentheses with the species names in Figure 25). There is, therefore, more than ample indication that independent phyletic lines have been initiated by populations heterozygous for genotypes determining several forms of expression of a given trait. Further, there is also evidence (e.g., Figures 49.3-.5 and 49.7-.8) that primitive genotypes may be "fixed" in species descended from populations presumably heterozygous for primitive and derived genotypes. Thus, we do see a type of pattern among these species which is sub­ The University of Texas Publication .1 castanea .8 moju .1 5 fulvimacula .2 meridionalis .9 pegasa .16 peninsularis .3 arizonensis .1 0 mercatorum .17 hydei .4 mulleri .11 paranaensis .18 bifurca .5 aldrichi .12 limensis .19 camargoi .6 tira .13 canapalpa .20 paracanalinea . 7 mojuoides .14 fulvimaculoides .21 canalinea stantially that predicted if heterozygous populations play a major role in evolu~ tion and if genotypes may traverse gene pools in substantially the way outlined earlier. In short, if we take this data at face value, without elaborate rationaliza~. tion and without attempts to explain it away, we see a pattern of evolution consistent with the best documented characteristics of evolving systems and of the behavior of the genetic material itself. Before turning to comparable evidence from other sections of the genus, the implication of morphological evolution in the repleta complex can be evaluated in somewhat greater detail. It would appear that the population ancestral to these species was moderately heterozygous at the loci whose activity is detected morphologically in characteristics of the spermathecae, the ejaculatory bulbs and the testes. Its complete store of variability was not passed on equally to all subpopulations deriving from it, or, more correctly, as its subpopulations di­verged, re-integration of their gene pools resulted in somewhat different gene complexes being maintained in each. Thus, while some morphological types are common to two or more lines, the total pattern of variability in any one line is slightly to markedly different from that seen in other lines. These different patterns may have arisen in two different ways. From a given array of genetic material, several to many adaptive combinations may be possible, and we may . Throckmorton: Phylogeny in Drosophila infer that most of these different combinations will be reflected in slightly differ­ent phenotypes. Thus we see an array of types (Figures 48.2, .5, .8, etc.) which are substantially the same in general features but which differ from each other in detail. Such slight variations reflect either the different ways in which sub­stantially similar genotypes may be integrated with each other and with other elements of the. gene pool, or they reflect the contribution from new mutations which have occurred subsequent to the separation of the phyletic lines, or they reflect both of these factors. The contribution of mutation to the diversity of phenotypes can be inferred chiefly through noting the number of phenotypes distinctive to any one line. That shown in Figure 48.11 may be one such. This method of assessing the contribution of mutation is, of course, only an approxi­mation, since the "distinctive" genotype may actually have been potential in the ancestral population but eliminated from all but one of the descendent populations. Or it may still be "potential" in genotypes of some of the existing species from several of the phyletic lines. We have no way of predicting the morphology of individuals of ancestral populations, and thus we cannot detect species retaining in toto the ancestral genotype. In spite of the limitations of the estimate, it seems probable that new mutations have contributed very little to the morphology of the species in this complex. By far the greater proportion of phenotypes are of such wide distribution that their genotypes must have been derived from genes already established in the ancestral population common to these species. Thus, distinctive phenotypes may owe their origin to new mutation or to an integration of a different array of genes from an ancestral population. The major features of types recurrent in several phyletic lines take their origin from the gene pool of ancestral populations, and hence they evidence genotypic homology. In the light of this situation, the implications of genotypic homology must be reassessed. Genotypic homology does not require the direct derivation of a geno­ type from a population expressing that genotype. To return to the earlier, simple model, the genotype, A'A', is homologous with other A' A' genotypes regardless of whether it was derived from a population having the genotype A'A', or from one having the genotype AA'. In the model shown in Figure 47, species Wand Z have homologous genotypes, but neither of them was derived from a population expressing this genotype. Genotypic homology may be determined by common ancestry. from an original population in which the genotype, A'A', was potential, as well as from a population in which this genotype was fixed. Thus, a given genotype may be produced independently in separate phyletic lines, but the production of this genotype can only take place from a gene pool in which the necessary elements were already established. Genotypic homology requires genetic continuity, but it does not require genotypic continuity. Parenthetically, it may be noted that recognition of this fact removes so-called parallel evolution from the ranks of the inexplicable to those of the expected. Ifwe do not accept independent integration of genotypes in descendent popula­ tions as an adequate explanation of data such as is shown in Figures 48 and 49, we are left with a remarkable system of convergent evolution. The series of types seen on the central branch of Figure 49 may serve as one of many possible ex­ amples from the repleta complex. The type of ejaculatory bulb shown in Figure The University of Texas Publication 49 .11, .13 and .14 must have arisen independently at least three times on this branch alone and probably at least three times in species from the mulleri sub­group; once in the population from which the sibling species arizonensis and mo;avensis arose, once to produce anceps, and probably once also to produce serenensis (see Figure 17). This type is not seen elsewhere in the genus, and its appearance within this closely related cluster of species strongly suggests that its occurrence here results from genetic continuity between these species. The distribution of this type is such, however, that its appearance cannot be attributed to genotypic continuity. In the figure, the type shown in 49.13 and .14 and the type shown in 49.12 and .15 cannot both owe their origin to genotypic continuity. One or the other must have originated independently, and from the standpoint of the present discussion it is immaterial which one did. In all probability, both owe their appearance to independent integration of their genotypes from a com­mon heterozygous gene pool. The simplest alternate to this explanation would assume that the type shown in Figure 49.13 differs genetically from that seen in close relatives by a very low number of genes. Thus, the conversion of the type shown in Figures 49.12 and .15 to that shown in Figures 49.13 and .14 might involve a relatively simple genetic change and hence may have arisen independ­ently by mutation. This interpretation would not be too difficult to accept if we had to consider only the sequence from Figure 49.12 to 49.15. At their time of divergence the ancestral gene pools would have been very similar. Specific muta­tions would probably occur with about equal frequency in each of these diverg­ing gene pools and might be integrated independently in each. This possibility, however, becomes increasingly unlikely when we consider the distribution shown in Figure 49 .3 to .5 and 49 .10 and .11. The basic question here is whether origin by independent mutation or by independent integration of the genotype is more probable. If we assume common origin from a heterozygous gene pool, all that is required is a reintegration of the gene pool from elements already tested', and probably at frequency levels such that the changes involved could be quite rapid. If we must start each time with mutation as the origin of the genotype, the ap­propriate genes must first arise, then they must be brought to a frequency level such that selection may act, then they must be integrated with the gene pool, and finally they must reach a level of integration such that they are uniformly ex­pressed as characteristic of the species in question. While such a sequence is con­ceivable, it hardly seems to represent the most probable explanation. The time element alone, and the expense to the population (Haldane, 195 7) , would seem to preclude this. Only the last step is required if separation of the phyletic lines involved populations already heterozygous for the requisite genotypes. The origin of the genes and their primary integration into the gene pool need only have happened once. The expense of completely replacing one gene with another would be spread through a long period of evolutionary time and would not be borne by a single population during the formation of a single species. Once the primary integration into the gene pool is completed, the fates of potential genotypes are determined by adaptive situations to which the populations and their subdivisions are subjected. The number and distribution of types seen in the repleta complex suggests that, from the original heterozygous gene pool, only a limited number of adaptive combinations of genes controlling a given characteristic was possible. Throckmorton: Phylogeny in Drosophila These combinations have recurred, and one or another of these has been incor­porated into the gene pool of each descendent species. Contributions from muta­tion subsequent to separation into independent gene pools cannot be excluded, but they seem to have played a minor role in determining the phenotypes ex­pressed. Mutations which have been incorporated into the gene pool of a species since that gene pool became an independent evolutionary entity will, in most cases, have their major effects on the phenotypes to be produced in the future. Rapid evolutionary change (phenotypic change) is possible from such mutations, but evidence from the repleta complex suggests that it is not common. Genes and genotypes controlling separate characteristics, e.g., those of the spermathecae and those of the ejaculatory bulbs, have been relatively independ­ent of each other. The origin of the phyletic line leading to the mulleri complex species apparently involved a population in which the genotypes for character­istics of the ejaculatory bulbs (Figure 17) were relatively heterozygous while that for spermathecal type was substantially fixed. The line leading to the melanopalpa subgroup retained heterozygosity for characteristics of both the spermathecae and the ejaculatory bulbs, etc. (The term, fixed, as used here and elsewhere in this communication, must be interpreted rather loosely. At present there is sufficient uncertainty regarding the genetic structure of populations, e.g., Crow, 1961, to require a somewhat imprecise form of reference. The term, stabil­ized, might be a suitable alternative.) Thus, the general pattern of evolution dur­ing the transition from one morphological type to another becomes evident. It need not, and probably does not, involve a stepwise alteration of the phenotype. Rather, it involves the production of an array of types. When two or more sys­tems are in transition in the same gene pool a great number of genotypic combinations are potential, and many of these are actually realized in descendent species. If evolution within the genus has been of this general type, it will be necessary to modify methods of phylogenetic analysis to accommodate for the irregular distributions of characteristics which result from such a system. Evi­dence that this has been the general pattern of evolution throughout the history of the genus comes from several sources and from all of the species groups avail­able for study (except those consisting of only a single species). Only two ex­amples will be treated in any detail. First, we can see that evolution of this type is not peculiar to species of the repleta complex by analyzing the distribution of characteristics among species of the virilis group. A cytological phylogeny for these species is also available (Stone, et al., 1960) and is shown, correlated with certain of the morphological characteristics, in Figure 50. The species of this group do not vary morphologi­cally among themselves to the extent that species of the repleta complex do Still, the same pattern of distribution is seen here. Two general types of para­gonia are present in the group (e.g., that of virilis and that of ezoana) . The population designated as Primitive I was probably heterozygous for genotypes determining these two types. The derivation of the population designated as Primitive II from this population may have involved homoselection toward one of these genotypes, but the population designated as Primitive III apparently retained heterozygosity for both genotypes. The population from which ezoana and littoralis were derived likewise retained heterogygosity, but that from which The University of Texas Publication loci cola bore al is M T-3 T-O M CffJ ezoono M // Ii ttorolis Frc. 50. Evolution of paragonia, spermathecae and testes of species in the virilis group. Cor­related with cytological phylogeny of Stone, et al. (1960). Throckmorton: Phylogeny in Drosophila the other four species were derived became homozygous for the second genotype. A comparable pattern of distribution is seen for characteristics of the spermathe­ cal duct and for number of testes coils (T-6, T-8, etc. in Figure 50). Other char­ acteristics of these species, e.g., those of the ejaculatory bulb (Figure 16), those of the anterior pupal spiracle (Figure 42.17), etc., are distinctive to this group al1.d almost completely constant within it. Apparently their genotypes originated and were fixed prior to the speciation events which produced the group as we now see it. These then, are the "good" taxonomic characteristics for the group. The others, those which are variable within the group and which do not define phylogenetic separations, are "bad" characteristics, and they fall in this category because the species showing them originated from populations in which genotype fixation had not yet occurred. Extensive cytological phylogenies are not available for other groups of species, so further evidence for this type of evolution must be obtained in a different way. The evidence required can be deduced from consideration of the type of char­acter distribution expected if evolution has followed the "classic" pattern (Fig­ure 47, left). In such a case, genotype fixation is assumed to occur during each speciation process, those which established populations initiating phyletic line­ages as well as those producing individual living species. Thus, a complete cor­relation of all characteristics is required. Any deviation from this pattern is assumed to result from mutation following speciation, and recurrence of types in separate phyletic lines should be rare. It should be possible to arrange morpho­logical types in a stepwise sequence, and the sequences for one characteristic should correspond to those for any other-not necessarily in the sense that each morphological characteristic changed at the same rate and time as any other, but in the sense that primitive characteristics will generally be associated, derived characteristics will always be associated, and intermediate associations will show a recognizable sequence from the primitive association to the derived. Admit­tedly, these specifications are extreme, but they indicate the general rquirements of such a system. One example will serve to show the type of character distribu­tion actually encountered in all species groups in the genus. Characteristics from eight of the thirteen species investigated from the quinaria group are soown in Figure 51. Space does not permit showing all thirteen in one figure, but these suffice to indicate the general pattern. The morphological sequence from left to right is arbitrary and follows the evolutionary change in the ejaculatory bulb (line 3 of Figure 51) . Those types shown ~t the right in line 3 are distinctive, and their genotype presumably arose during the evolution of this group. If this sequence is held constant (changing the order within one of the bulb types will not help matters), then characteristics of the spermathecae, paragonia, vasa deferentia, anterior pupal spiracles and testes follow no regular pattern. There is a recurrence of general types, but it is impossible to arrange these species in any order (branching or linear) such that a regular sequence for all types is seen. The situation is further complicated by the fact that many of the types seen here are also seen in species outside of this group. The general type of anterior pupal spiracle seen in tenebrosa and guttifera, for example, occurs in species from the testacea group and in many species from the tripunctata, rubrifrons and macroptera groups. It must have arisen at least twice in the quinaria group, or w - @ ® @ 1 rm uw 11 rt ~fiW flfJ ~~~WU ITTN f ~ b~~~ ~~~~f ~ .f/l ~ ' • T-12 T-12 T-5 T-12 T-8 T-8 T-12 T-8 in nu bi I a pha lerata rell ima tenebrosa tran sversa guttifera pal ustris subpa lustris FIG. 51. The arrays of types produced by evolution from a heterozygous stem population. For further explanation see text. Throckmorton: Phylogeny in Drosophila else the type of ejaculatory bulb seen in guttifera arose twice. It must also have arisen many other times in other species of the quinaria section, but not else­where. Either the recurrent types arose by mutation, or these characteristics represent an array of types derived through evolution from a common heterozy­gous gene pool, and, on the whole, the latter process is more probable. Evidence that this type of evolution has occurred throughout the history of the genus will become more evident as the phylogeny is developed. To anticipate this, however, data included in the pictorial phylogenies (Figures 14, 20, 21, 24, 26, 31, 32, 40, 41, 44, 46) show that the stem populations from which major lines developed were themselves heterozygous for genotypes which later became fixed and characteristic of groups of species. Thus, analysis of phylogenetic relationships within the genus will require methods which recognize the taxonomic implica­tion of heterozygous populations in evolution. Method for analyzing phylogenetic relationships-Standard correlation meth­ods (Sturtevant, 1942; Malogolowkin, 1953) are adequate for deriving static classification, but they are not completely suitable for analyzing the dynamic pattern of change which attends the evolutionary development of large groups of species. They establish with reasonable certainty what groups exist, but they do not tell how these groups came to show their particular combinations of characteristics. As indicated by Michener and Sokal ( 195 7), the systematist must first determine relationships in a nonhistorical sense, and correlation methods are suitable to do this. The systematist may then decide on the most probable lines of descent. Some of the factors which must be considered when such de­cisions are made have already been discussed. The following method of analysis is one which takes into account the possibility that evolution may have proceeded from heterozygous populations. It is, however, quite suitable for analysis of phylogenetic relationships in groups whose evolution has followed the classic pattern, so it is not strictly necessary that the manner of evolution be known before the method is applied. The results of the analyses themselves will indicate the type of evolution involved. Three major correlations may be made when a phylogenetic analysis is begun. The first of these is the total phenotypic comparison of the type described by Michener and Sokal (op. cit.). This is the correlation most useful to the sys­tematist in determining a basis for classification, and its chief purpose is to sub­divide large complexes of species into groups of closely related forms. Such cor­relations may provide some indication of phylogeny, but their chief usefulness is in breaking a group down into subunits which may be conveniently analyzed. The second type of correlation is one involving only primitive characteristics, and the third is one considering only derived characteristics. These last two correla­tions often need not be made in a formal sense. Inspection of the distributions of such characteristics among previously established groups will usually be suffi­cient. The purpose of these correlations is to establish major dichotomies and to indicate evolutionary trends in the group as a whole. Often, but not always, they will be a routine part of a taxonomist's evaluation of species relationships. Since these phases of analysis are generally standard practice, and since descriptions of general methods applicable to such study are readily available (Sturtevant, 1942; Michener and Sokal, 1957), they will not be considered here. The University of Texas Publication For the purpose of phylogenetic analysis, the genus Drosophila provides ex­ceptional material. Classification within the genus has been established on a firm and objective basis by Sturtevant (1942), and this classification has been con­firmed and extended by many workers since that time (Wheeler, 1949b; Patter­son and Stone, 1952; Okada, 1956; etc.). Of primary importance for this analysis is the fact that the objective reality of the taxonomist's estimation of relationship has been substantiated by genetic and cytological investigation. Such studies have been made for many of the major species groups within the genus, and they confirm the geneological propinquity of species placed together in species groups (see Patterson and Stone, 1952, for summary to that time). It is therefore pos­sible to restrict analysis of phylogeny to relationships between species groups rather than to relationships between individual species, and this greatly simplifies the problem. From genetic and cytological evidence, we may infer that species within a species group have originated from a common ancestral population and that this population was evolutionarily distinct from other such populations. The absolute validity of this inference is not required since the methods to be de­veloped will indicate occasional erroneous classifications. Its validity is more than adequate for the present purpose. Once groups of species have. been established, further analysis proceeds by examination of the distributions of single characteristics. When living species provide the material investigated, phylogeny is determined by inference, and specifically by inference regarding the history of distinct genotypes. It is not to be expected that genotypes for several characteristics will have identical histories, and therefore they cannot be considered en masse. Herein lies one of the disad­vantages of character correlations as indications of phylogeny. Of necessity, correlations must deal with many characteristics (genotypes) simultaneously, and hence they cannot give a detailed picture of the history of any one of them. They do give a picture of total genotypic change, but sequence of change must be determined subjectively if such methods are used. As already indicated, the method for analysis of phylogenetic relationships within the genus must be one which does not involve the assumption that gene fixation has occurred at any one level. When and where character fixation first appeared may be inferred after the major outline of the phylogeny is determined, but not before. The method of analysis is eminently simple and may represent only a formalization of intuitive methods utilized by many systematists. Its steps are as follows: 1) Determination of general direction ( s) of change for all characteristics. This_ need only be done tentatively since the major purpose of this step is to provide orientation for further analysis. This part of the present analysis was included with the earlier descriptions and need not be recapitulated here. 2) Analysis begins with choosing one characteristic, preferably one which allows subdivisions into the largest groups or which appears to separate distinct phyletic lineages. Characteristics of the paragonia, for example, are of this type and were used as the starting point for phylogenetic analysis of the genus. Char­acteristics of the spermathecae, on the other hand, are much too .detailed to be used in initial stages. They could be used. In fact, any characteristic could be used, but the analysis is simplified if it is begun with the more general character­ Throckmorton: Phylogeny in Drosophila istics. The distribution of the different forms of a trait among the species in each species group indicates the genotypes which were potential in the stem population for that group. From the inferred genotypes of these stem populations, one then infers the genotypes of the populations from which they were derived, etc. When two or more stem populations appear to be derived from populations having the same general genotypic composition, lines coalesce into single populations. Es­sentially, we are tracing genealogical streams back through time to their point of confluence (Hennig, 1956). Examples will be given later to illustrate the steps and interpretations involved. A basic assumption to be made at all levels is that the population antecedent to the one in question was heterozygous for alternate genotypes, even though the population from which one is proceeding (backward) may appear to have been homozygous. The antecedent population may actually have been homozygous, but assumption of heterozygosity is least limiting. It is the most objective assumption which can be made at the start of the analysis. As analysis of other characteristics proceeds, accumulating evidence will often indicate general levels at which fixation of a given genotype has occurred. 3) Analysis of remaining characteristics is carried out in the same manner, each time working backward from the stem populations of species groups. Char­acteristics are analyzed one at a time, in sequence from the more general to the more specific. The results of each analysis are superimposed on the cumulative pattern produced from the previous ones. Addition of information from each new characteristic may suggest shifting the point of origin of the stem populations for various species groups, but the magnitude of these shifts will be limited within the pattern already adduced. 4) It will often be convenient to omit difficult groups from the first analysis. Once the major sequences of populations have been determined, genotypic char­acteristics of the stem populations of the difficult groups can be inferred. These can then be compared with the genotypes already established within the major pattern to determine their probable origins. As information from additional spe­cies becomes available it can be treated in the same fashion, as can information from species in other genera, etc. 5) The last step in the analysis is the determination of the histories of the geno­types of the individual characteristics. Once the major sequences of intermediate populations have been inferred from all characteristics, individual genotypes may be followed from the stem population to the terminal groups of the estab­lished phyletic lineages. This constitutes the final evaluation of the phylogeny. Complete details of this analysis for the genus Drosophila are too voluminous to be outlined in the space available. A representative analysis, utilizing only a limited number of characteristics, will be given. Abbreviated analysis of relationships within the subgenus Drosophila-The salient features of the method can be seen most easily when a limited number of species groups from the subgenus Drosophila is used. The first two steps in the analysis of this subgenus are shown in Figure 52. For brevity, three general types of paragonia may be identified, and the distribution of these types is shown in the inset at the upper left of the figure. The two major types are both robust forms, one having a single, high arch; the other having multiple folds. Species from the tripunctata, cardini, guarani, quinaria and calloptera groups all have JA~ ~Ii}) replete (A canalinea drefusi m~~~pah~ogmatica ~ n1co r obusta a~nuli!l'ano funebris 1mm1grons ~ replete canal inea dreyfusi mesophragmatica melonico robusto annuli mo no ~ M vi ri Ii s w (A(f] tripunctato cardini guaran i quinario cal laptera ...... ~ ~ t:! ~­ I::: ..,~ ~­ .Q.. qu inur1u ~ ;;? ~ M ~ K ~ call optera §" funebris ~ immigra ns ~ l ~ FIG. 5Z. Schematic representation of the first two steps in the phylogenetic analysis of the subgenus Drosophila. For further explanation see text. Throckmorton: Phylogeny in Drosophila paragonia of the first type (Figures 10-12), their stem populations are assumed to be derived from a common gene pool (not necessarily directly), and this stem population is assumed to have arisen from an earlier, heterozygous population. A similar situation exists for members of the repleta, canalinea, dreyfusi, me­sophragmatica, melanica, robusta and annulimana groups, except that the para­gonia here are of the second type (Figures 5-9). Characteristics of the stem population for the immigrans group (Figure 5) are somewhat less certain, but, on the whole, these paragonia are also of the second type and the immigrans group is therefore included with these forms. The stem population for the fune­bris group was heterozygous for both types (Figure 6). This stem population is assumed to have been derived from a population likewise heterozygous for these types, and this establishes the heterozygous population from which the other two major groups of species were derived. Essentially then, three major heterozygous populations are "identified," each heterozygous for approximately the same genotypes, and these coalesce as shown in the inset of Figure 52. Members of the virilis group have two types of para­ gonia (Figure 4). One type, not quite so robust as in other species, has multiple folds and suggests origin from the same gene pool as the repleta group and its neighbors. There is, however, another type which is seen nowhere else in this subgenus, unless the extremely shrunken types seen in species of the hydei sub­ group of the repleta group are of this pattern (Figure 9) . However this may be, the position of the virilis group is uncertain. For the time being, it is derived directly from the population ancestral to the whole subgenus. Although this by no means exhausts the information available from the paragonia, this part of the analysis will be considered as substantially complete, and the tentative phy­ logeny arrived at is that shown in the inset of Figure 52. It should be obvious, but perhaps it should be mentioned, that the "populations" inferred are not populations of the same type as are seen in living species. These hypothetical populations had an extensive time dimension and most probably consisted of complexes of species. Since we are inferring broad, temporal sequences rather than detailed speciation events, it would not be legitimate to refer to these popula­ tions as species or any comparable equivalent thereof. They represent clusters of gene pools having the same general properties relative to the genotypes under consideration and having an indefinite, but perhaps considerable, time dimension. The next step in this analysis utilizes characteristics of the vasa deferentia. Again, any of the remaining characteristics could have been used, and the final result would be the same. The characteristics of the vasa are examined, as before, to determine the distribution of genotypes in the stem populations of species groups. Characteristics of these stem populations and the ways in which they appear to coalesce will confirm the previous disposition of the groups or will indicate alternate groupings. They may also suggest subdivisions of the groups and indicate the sequence in which various groups were derived. With one exception, characteristics of the vasa confirm the unity of the cluster of species groups (repleta, etc.) to the upper left in Figure 52. The exception is the immigrans group. From the distribution of types within this group, it may be inferred that the genotypes of its stem population were not comparable with those of other species groups in this section. Further, the genotypes of this stem The University of Texas Publication population appear to have included more primitive alternatives, and so the group is tentatively derived from the stem population for the subgenus (lower right, Figure 52). Characteristics of the vasa of virilis group. species follow the same pattern as for species in the repleta cluster. Its position in the phylogeny is left unchanged (lower left, Figure 52) . Characteristics of the vasa in species of the quinaria section are variable (Figure 52, right; Figures 10-12) , showing different degrees and types of association with the paragonia and different degrees of basal fusion. Species of the tripunctata group show almost the complete array of types, lacking only that type seen in species of the calloptera group and in two out of three species in the immigrans group. Stem populations of the quinaria, cardini and guarani groups were each heterozygous for two of the major types. The distinctive characteristics seen in species of the calloptera group suggest its de­rivation from a gene pool having somewhat different characteristics than that from which the other four species groups arose. Since its stem population appears to have had some genotypes in common with that of the immigrans group, they are both shown as derived from the same ancestral population, i.e., the stem population for the subgenus. The remaining groups in this section ( quinaria, cardini, guarani and tripunctata) had stem populations of approximately the same character and so these unite as shown in the figure. In species of the funebris group, characteristics of the vasa are substantially as seen in species of the quinaria section. They are not sufficient to allow the stem population of the group to be definitely related to that of the quinaria section, however, and the group retains its original position. Thus, we now have six tentative branches from the ancestral population, four of which are individual species groups. The two remaining branches are made up of several species groups, with the individual groups in each of these branches being phylogeneti­cally equivalent, i.e., no subbranches are indicated at the present time. Grouping of the cardini and guarani species groups was done to save space in Figure 52 and has no significance. From this point onward, complete details of the analysis need not be given. Most of the more useful information has been included in the pictorial phy­ logenies seen earlier. The major objective now is to determine to what extent clusters of species may be separated to indicate sub-branches within major line­ ages. Only those characteristics which establish cleavages will be considered. Characteristics of the ejaculatory bulb are useful here (Figures 16-21). Those of species in the virilis group are quite distinctive and do not closely resemble any of the types seen elsewhere in the subgenus. This suggests a less close rela­ tionship with other groups, i.e., a longer period of independent evolution, so this group remains as derived independently from the ancestral population. Species of the melanica and robusta groups show an array of types, with most of the ejaculatory bulbs having four caecae. Both groups have one species show­ ing the most extreme type, that figured for lacertosa (Figure 16.15) of the robusta group being .also present in melanura (not figured) of the melanica group. These two groups are assumed to have a common ancestral population in which the genotype for this extreme condition was potential. Species of the annulimana group (see Figure 16) also have ejaculatory bulbs of types .seen in both the melanica and robusta groups, so we can tentatively place these three groups on Throckmorton: Phylogeny in Drosophila a common sub-branch (see Figure 53 for final disposition of groups). This sub­branch is, itself, derived from the stem population for the major branch, which takes its origin from the ancestral population for the subgenus. By inference, the stem population for this branch is assumed to have been heterozygous for the genotypes common to the melanica, robusta, and annulimana groups. It probably was not heterozygous for those determining the extreme types seen in species from the melanica and robusta groups, but it may have been. The remainder of the species groups (repleta, canalinea, dreyfusi, mesophragmatica) would be derived as the other sub-branch of this major stem. Some species of the robusta group and species of the dreyfusi group have the posterior part of the ejaculatory bulb expanded laterally, and this suggests that species of the dreyfusi group may be more closely related to the robusta complex than are the other remaining species groups. Subgroups within the repleta group can be handled just as if they belonged to species groups and their stem populations identified accordingly. Species of the mulleri, fasciola, mercatorum and melanopalpa subgroups have over-lapping arrays of bulb types and so would be assumed to stem from a popu­lation distinct from that for species in the hydei subgroup. Species of the cana­linea group have one type of bulb in common with repleta species, so are assumed to be derived from the same ancestral gene pool and hence are derived at the same level (see Figure 53), etc. Characteristics of the spermathecae also provide information concerning the subdivisions of this branch. The array of types seen in the melanica group some­what resembles that seen in species of the virilis group (Figure 28), but since the stem population for this branch is derived from the same stem population as was that of the virilis group, this does not alter the position of the melanica group within the branch. Genotypes for these characteristics are presumed to have been derived from the stem population for the subgenus. Retained geno­types might potentially have been present in all the stem populations of this branch, so their distribution cannot determine the position of the group in ques­tion. Spermathecae of species in the robusta and annulimana groups have many characteristics in common (Figure 28), and also some characteristics in common with types seen in members of the immigrans group (compare Figures 28.19 and 28.24). This, together with information from the melanica group, indicates that we must consider the possibility that these three groups (robusta, melanica, and annulimana) belong to an independent line derived directly from the ancestral population for the genus, rather than from the stem population in common with the repleta group and its relatives. Without going into details, however, informa­tion from remaining characteristics tends to confirm the disposition shown in Figure 53. Spermathecae from the repleta, canalinea, dreyfusi and mesophrag­matica groups (Figure 29) form a distinctive array of types which suggests that they have a stem population in common and that this population was distinct from that of the robusta complex. Distribution of spermathecal types among sub­groups of the repleta group tends to confirm the relationships indicated by the characteristics of the ejaculatory bulbs. They indicate also that the dreyfusi group originated from much the same population as did the hydei subgroup and that the canalinea group originated from much the same population as did the mulleri subgroup. Hence the final disposition shown in Figure 53. The University of Texas Publication mercatorum subgroup spe cie s Kand L group tri punctata group macroptera r:~:~;,ons /I pallld1penn1s species quinaria G and H group aureata I// te stacea group % annulimana group robusto ~""' group~ me Ionic~ group viri lis group melanogaster \ \ ' ,,~JI"'°'"''" ",lg~;~~'" populi "'--\ bryoni~ lotifosciae form is corocin~ group victorio~~ group~ ~ F1G. 53. The phylogeny of the genus Drosophila showing the relationships of other genera to the various phyletic lines within the genus. The inset (lower left) shows an alternate phylog­eny which may be derived if subgeneric boundaries are ignored and the species groups are taken as the basic phylogenetic units. For further explanation see text. Throckmorton: Phylogeny in Drosophila While the analysis is by no means complete, this brief description has covered the major points of the methods involved. Similar methods are applied to species groups on the quinaria branch, and to groups from other genera and subgenera. This finally results in the delineation of a series of populations having intergrad­ing or overlapping genotypic characteristics. Since the direction of evolution of each characteristic has been considered during the analysis, these populations are defined in a distinct time sequence. The genotypic change from one popula­tion to the next has been gradual, and some latitude exists as to the time of derivation of any particular group. However, change of all characteristics has not been synchronous, and·the total genotypes at the different time levels are relatively unique and allow placing the various groups with some precision. The precision of such an analysis will increase as the number of characteristics con­sidered increases, but it will always be limited by uncertainties regarding the length of time during which certain genotypes have been potential in the stem populations. In this manner, the extent to which stem populations have retained heterozygosity sets the final limit on the accuracy of any phylogeny, whether it be developed from living species, from fossil forms, or from chromosome struc­tures. Phylogenetic relationships-Figure 53 shows the phylogenetic relationships of genera, subgenera, species groups and species included in this study. Interpreta­tion of this phylogeny must be made within the context of the methods utilized in obtaining it, and its major limitations and uncertainties will be pointed out in the following discussion. Points of origin of the various groups are essentially probability values based on consideration of the history and distribution of all characteristics and on an evaluation of the evolutionary characteristics of the alternate populations from which they may have been derived. The phylogeny is based completely on characteristics described earlier. It is, of course, not inde­pe:r;tdent of other characteristics, since definition of species groups depends on total phenotypic comparisons among species. Figure 53 shows only the general sequence of origin of the different groups. Vertical separations and other dimen­sions were adjusted to avoid crowding in the figure, so they do not reflect an objective judgment of elapsed time. In deriving the relationships shown at the base of the phylogeny, two alterna­tive assumptions could be made. The relationships included in the total phylogeny result if it is assumed that both the subgenus Pholadoris and the subgenus Sophophora are monophyletic. If subgeneric boundaries are ignored and the species groups are taken as the phylogenetic units, the relationships shown in the insert (lower left, Figure 53) may be derived. The first alternative places both of these subgenera on side-branches. The second places both on the main se­quence of evolution in the genus. Recalling the evolutionary models discussed earlier, there is no completely objective way to decide between these two alter­natives. Consideration of other characteristics, for example, chromosome homol­ogies or features of male genitalia, might provide information which would indicate the more probable relationship, but extension of the analysis to include such features is beyond the scope of the present discussion. The alternatives shown in the inset will not be included in the following discussions. There can be little doubt that the stem population for the subgenus Pholadoris The University of Texas Publication originated at an earlier time than did those for the other subgenera. Almost all of the distinctive features of these species are primitive, and for most of those which are not, other evidence indicates either their independent origin within the group or their origin prior to the establishment of the stem population for the genus. The peculiar type of ejaculatory apodeme (Figure 15) seen in species of the victoria group falls in the first category. Characteristics of the Malpighian tubules (posterior tips apposed or fused rather than free) fall in the latter cate­gory. Zapriothrica dispar, which is surely more primitive than the complex associated with the genus Drosophila, has the posterior tips of the Malpighian tubules fused, indicating that genotypes for this fusion were potential in the stem population for the genus. The subgenus Pholadoris is the only group which in­cludes species having an unreduced sixth sternite in the male, and most of these species have the primitive elliptical testes, the primitive vasa, etc. The only population which was potentially capable of producing these genotypic com­binations was the stem for the genus. The subgenus Sophophora was derived from a population which had under­gone a certain amount of evolutionary modification and in which genotypes for several of the more derived types were potential. These genotypes may have been potential in earlier populations, but the evidence available at present sug­gests their origin during the time immediately following the separation of the Pholadoris stem population. As more extensive information from more primitive genera becomes available, it may be necessary to modify this assump­tion. This, incidentally, emphasizes the impracticality of attempting to determine the detailed phylogeny of a single genus without considering the characteristics of species in related genera. The major uncertainties of the present phylogeny stem almost directly from the fact that the available material from·more prim· itive genera was not adequate to allow definition of the evolutionary events which preceded the origin of the stem population for the genus. Be that as it may, the available evidence allows tentative conclusions to be reached, and it would appear that genotypes for derived characteristics of the vasa, the testes, the ejaculatory bulb and apodeme, etc., either arose or reached greater significance as adaptive components of the gene pool during the period intervening between the time of origin of the stem population for the subgenus Pholadoris and the time of origin of the stem population for the subgenus Sophophora. The stem population for the genus Chymomyza arose from a population having the same general char­acteristics as that from which the Sophophoran stem originated. There is a dis­tinct possibility that the genus Chymomyza was derived from the basic Sopho­phoran stem rather than from the major stem population for the genus. These two groups have in common characteristics of the ejaculatory bulb, the ejacula­tory apodeme, the vasa, the ventral receptacle, the abdominal sternites, and the anterior pupal spiracle (see earlier descriptions) . Since the genotypes for the distinctive features may have been present in the earlier population, Figure 53 shows Chymomyza to be derived at that level. This is the most objective assess­ment of its relationships. It probably was derived from the Sophophoran stem, but the evidence is not conclusive. The same general conditions also apply to the relationships of populi. This species was detived either from very near the base of the Sophophoran stem or directly from the major stem population of the genus. Throckmorton: Phylogeny in Drosophila Phylogenetic relationships within the subgenus Sophophora remain substan­tially as indicated earlier by Patterson and Stone (1952), except that no groups are visualized as derived from presently existing groups. The stem population for the subgenus became separated into _three major subpopulations. Their se­quence of origin is somewhat uncertain. The most obvious interpretation would be that the stem for the obscura group separated first. Then, following a period of evolution, the melanogaster stem separated from the saltans-willistoni stem. The alternative one would suggest that the stem population separated into two subpopulations, a saltans-willistoni stem and an obscura-melanogaster stem. These possibilities are equally likely, so the obscura, melanogaster and saltans­willistoni stems must be shown as derived at the same level. Evidence from the characteristics used here does not permit finer discrimination. As a convention in making Figure 53, groups were shown as derived at the same level when further details as to the precise sequence of origin could not be adduced from the data. Such a depiction is intended to be noncommittal. It does not indicate that the groups in question arose simultaneously, although they may have. The evolutionary branch leading to the subgenus Drosophila was derived at approximately the same level as was the Sophophoran stem. Changes leading to the establishment of genotypes for still more derived conditions of the vasa, the ejaculatory bulbs, the ejaculatory apodemes, the ventral receptacles, the sperma­thecae and the anterior pupal spiracles were all in progress in this stem popula­tion: Since these changes were not synchronized, an overlapping pattern is produced which allows the origin of the various groups to be placed in an approxi­ -mate time sequence. For example, the genus Scaptomyza arose at a time when genotypes for the primitive (Sophophoran) type of anterior pupal spiracle were still available in the gene pool. It arose at approximatedy the time when geno­types leading to an association between the vasa and paragonia were becoming available but before those for the coiled ventral receptacle became conspicuous. It arose at a time when the genotypes for the four-filament egg were becoming prominent, but before they had been stabilized to produce the typical four­filament egg of Drosophila species. It also arose from a population in which the genotypes for the spoon type of ejaculatory apodeme were potential. When all characteristics are considered, the genus Scaptomyza becomes one of the earliest branches to be derived from the stem population leading to the subgenus Drosophila. The subgenus Dorsilopha was also derived at the same general time, perhaps somewhat earlier, but probably somewhat later than was the genus Scaptomyza. There is a considerable amount of character overlap between species of the genus Scaptomyza and those of the genus Zaprionus. Both have substantially the the same type of ejaculatory apodeme, the same type of ventral receptacle, and one species of Zaprionus has the same type of vasa as is seen in Scaptomyza species. Also, both Scaptomyza and Zaprionus species show the peculiar branched posterior caecae on the ejacultory bulb. All of these characteristics indicate that these two groups have originated from a common stem. Evidence from other characteristics indicates that these two genera originated in the sequence shown in Figure 53. Species of the genus Zaprionus have anterior pupal spiracles of the types seen in species from the subgenus Drosophila, one species has paragonia The University of Texas Publication and vasa of types seen in species from this same subgenus, egg filaments are ' basically of the Drosophila type, etc. The major alternative to the arrangement shown in Figure 53 would assume both Scaptomyza and Zaprionus to be derived from a single stem which was itself derived from the base of the branch leading to the subgenus Drosophila. This stem population would have been heterozygous for almost all of the genotypes seen in later species. During subsequent evolu­tion of this stem population it would have become subdivided into the stem popu­lations for the two genera. The origin of the Zaprionus stem would have involved homoselection toward almost all of the more derived genotypes and that of the Scaptomyza stem would have involved homoselection toward all of the primitive genotypes. This seems unlikely. In most of its characteristics, the genus Zaprionus shows a marked resemblance to species of the immigrans and funebris groups, and it is assumed to have arisen at a time either slightly antedating the origin of the immigrans group or even somewhat later. As far as the characteristics under consideration here are concerned, the genus Zaprionus is just a moderately distinctive species group in the subgenus Drosophila. Most of the relationships within the subgenus Drosophila have been indicated during the earlier discussion. The immigrans group is derived almost directly from the stem population for the subgenus and the same is true for the virilis group. Within the virilis group the general combination of characteristics is that of the virilis-repleta section, and it is shown in Figure 53 as derived from the very base of this branch. Its most probable alternate position would be at an earlier level, derived directly from the stem population for the subgenus and in­dependent of all other groups. Evaluation of the relationships between the other groups from the virilis-repleta stem indicates three major clusters of species, the robusta complex, the repleta complex and the bromeliae complex. The latter cluster of species (nannoptera,' bromeliae, etc.) is, for the present, a very ill-defined group. There is a relatively large number of undescribed or ungrouped species which are generally of the peruviana type, and its seems probable that as these species become more fully investigated and understood a third major phyletic branch, most 'closely related · to the repleta complex, will be recognized. Even at the present time there would be some justification for more sharply separating the robusta complex from the other two, rather than deriving all three at the same level. As indicated pre­viously, this method of depiction is intended to be noncommittal, and an ob­jective evaluation of relationships does not allow finer subdivision at the present time. If, however, taxonomic studies of these species develop as anticipated, the virilis-repleta section will eventually be separated into two distinct phyletic lines. One will include the virilis group and the robusta complex. The other will include the repleta and bromeliae complexes. Within the repleta complex itself, general relationships are as shown in the figure. Since this complex has already been discussed in some detail, and since a cytological phylogeny is available for many of these species, little more needs to be added. D. aureata, formerly placed in the repleta group (Wheeler, 1957), · most probably arose much earlier. A conservative estimate of its position places jt as shown in Figure 53. Alternate positions would range from this level down­ ward about to the level indicated for the origin of polychaeta. The mulleri sub­ Throckmorton: Phylogeny in Drosophila group of the repleta group is almost certainly not an evolutionary unit in the same sense that the hydei, fasciola, mercatorum and melanopalpa subgroups are. Many, if not most, of the species shown derived independently from the repleta standard gene sequence (see Figure 25) are probably of independent origin from the stem population for the repleta group. They might be shown as a scattering of "twigs" arising between the base of the hydei stem and the base of the mulleri stem as it is shown in Figure 53. The position of fX>lychaeta is somewhat equivocal. It shows some relationships with both sections of the subgenus, but the great majority of its characteristics are those of species in the virilis-repleta section. In characteristics of its ejacula­tory bulb and apodeme it most nearly resembles Phloridosa species, and the sub­genus Phloridosa itself is closest to the virilis-repleta section. It seems probable that both polychaeta and Phloridosa were derived from the early stem population for the virilis-repleta section. The position of the funebris group remains doubtful. Since the majority of its characteristics are those of species in the quinaria section, it has been shown de­rived from the early stem population for this section. However, all of the geno­types for the characteristics seen in species of the funebris group were potential in the gene pools from which the stem populations for the two major sections took their origin. It is possible, for example, that the funebris group arose from an early population which was evolving toward the integration of virilis-repleta genotypes. It could simply have separated from this major population before the distinctive combinations of genotypes had been established. This is somewhat less probable than the alternative indicated in Figure 53. Within the quinaria section itself, the calloptera group has a somewhat isolated position. It is unquestionably a member of this section, but its stem population appears to have separated from the main stem at an early time. The remainder of groups and species in this section form a rather closely knit cluster, with the quinaria group being the most distinct and derived earlier than the others. The terminal groups in this section appear to have been produced, as it were, by explosive subdivision of a large and complexly heterozygous stem population. Examination of all characteristics shows that the population at about the level of the quinaria group had retained a large amount of ancestral genetic variabil­ity, and it was acquiring and retaining a great deal more. The guarani, cardini and testacea groups are distinct and are each derived from stem populations which included a moderate "sample" of the genetic variability in the major stem populations. The remainder of these groups are much less well defined. It seems probable that species now placed in the tripunctata group have been derived from the major stem population at a sequence of time levels, and that the macroptera and rubrifrons groups are only moderately distinctive clusters of species derived in the same way. Thus, some of the species of the tripunctata group were prob­ably derived at about the time of origin of the quinaria group, others somewhat later at about the time of origin of the guarani group, etc. This group or cluster­of species probably provides the major exception to the earlier assumption that species in a group are derived from a single, independent ancestral population. Species of the mulleri subgroup of the repleta group are another exception, and it seems probable that the stem population for the subgenus Drosophila evolved The University of Texas Publication in much this same fashion. The semblance of order which is now seen at the base of the subgenus reflects the pruning effects of time, with most of the minor twigs removed and only the derivatives of the major populations remaining. Some twigs still exist, however. Both carsoni and carbonaria were probably de­rived from the stem population for the subgenus, and both have peculiar com­binations of characteristics which can best be accounted for if they arose at an early time. D. carsoni most probably arose from a very early population of the virilis-repleta section, and carbonaria from an early population of the quinaria section. D. tumiditarsus likewise arose from an early stem population of the subgenus. It probably is a member of another major phyletic line of the subgenus, however, rather than an isolated "twig." From inspection of descrip­tions and figures by Okada (1956), it is apparent that several groups exist which probably represent an oriental equivalent of the primarily newworld, virilis­repleta stem. The branch to which tumiditarsus belongs is possibly the earliest branch from this stem, but it may have arisen independently from the stem population of the subgenus. There is probably also an oriental equivalent of the quinaria stem which originated at about the level indicated for aracea in Figure 53. Most of the oriental Hirtodrosophila, as well as several other oriental groups which Okada (1956) refers to the quinaria section, probably belong to this branch. D. aracea itself is most closely related to the immigrans group and is probably a member of still another independent lineage deriving from the early stem population for the subgenus. Since only a single female was available for dissection, characteristics of pinicola (collected at Andreas Canyon, California) have not been discussed earlier. The characteristics of this female, however, are distinctive enough to allow it to be placed with reasonable certainty at about the level indicated in Figure 53. Thus far, this is the only species distinctly related to this stem which retains the first sternite as paired and pigmented plates. Its ventral receptacle is a mixture of the coiled and folded types, having about four large coils basally and a couple of flat loops distally. It is completely free and not appressed to the surface of the vagina. Except for the number of coils and folds, it is identical with that type figured for histrioides (Figure 35.4) and M. dimidiata (Figure 39.7). Its spermathecae are of the obscura or Pholadoris types, dark-pigmented and elliptical in outline with a slight apical indentation. The spermathecal duct is flexed just proximal to the spermatheca. The spermathecal envelope is thin and uniformly distributed. It is interesting to note that Sturtevant (1942) found the minimum number of differences between the species included in his study to be between pinicola and Scaptomyza terminalis, an observation which is completely consistent with the positions indicated in Figure 53. D. pinicola is one of the most primitive members of the subgenus Dr.osophila, but it can no longer be considered as near the primitive for the genus as a whole. The position of the Hirtodrosophila must remain uncertain. As presently clas­sified, they are almost surely a polyphyletic group. If subgeneric boundaries are ignored and the individual species placed as their characteristics would seem to indicate, duncani would be derived somewhat earlier than Dorsilopha, pictiven­tris at about the level of Dorsilopha, thoracis near Mycodrosophila, and histrio­ides at the level of the immigrans group. It would appear that the subgenus Throckmorton: Phylogeny in Drosophila Hirtodrosophila is badly in need of revision. Until it has been re-evaluated, these species cannot be placed in the phylogeny of the genus. Genotypic histories-Once phylogenetic relationships have been indicated, final evaluation of the phylogeny is made through tracing the history of genotypic change for the individual characteristics. These analyses have the formal purpose of detecting inconsistencies and providing additional insight into the type of rela­tionship which exists between the different groups. Since the genotypic histories for all characteristics follow substantially the same pattern, only one will be covered in any detail. Sufficient data have been provided in earlier sections so that the reader may develop the others if this seems desirable. As described earlier, phylogenetic analysis proceeds backwards from the geno­types of the species in a group. From these genotypes, the genotypic composition of the stem population of the group is inferred, and so on until the stem popula­tion for the genus is reached. The summation of the analyses for all individual characteristics defines an evolutionary series of populations, the major genotypes potential within their gene pools, and the major changes which took place in these gene pools as subpopulations diverged from each other. Once this outline is com­plete and the major pathways of evolution established, the genotypic histories for individual characteristics may be traced, starting with the probable genotypes potential-within the stem population for the genus. One of the more simple se­quences of change, that involving the ejaculatory apodeme of Sophophoran spe­cies, has been diagrammed in Figure 54. If this figure is read from top to bottom, the phlyogenetic analysis based on the characteristics of the ejaculatory apodeme is followed. If it is read from bottom to top, the history of genotypic change is traced. The sequences leading to the terminal species of the subgenus Drosophila have not been included, although a much abbreviated summary for some dis­tinctive types has been shown to the left-center in the figure. Since the method of phylogenetic analysis has already been described, this aspect of the figure will not be discussed. It should, perhaps, be emphasized, that the sequence of populations shown in Figure 54 is that determined by the total analysis and not just by the analysis of the ejaculatory apodeme. In Figure 54, genotypes are shown figuratively by the phenotypes which they determine. At each level, an array of genetic variability existed, and an array of phenotypes is shown. It is not intended that these be considered the only geno­types potential in these populations. They are simply the genotypes which may be inferred to have been present then. When genotypes are shown as being perpetuated unchanged (solid arrows) , it should be inferred that later genotypes were homologous in the sense defined earlier, but they were not necessarily completely identical with similar genotypes shown elsewhere. In most cases it seems probable that by far the major fraction of the genes comprising a given genotype were derived unchanged from ancestral gene pools. When new geno­types are indicated as originating (dashed arrows), most of these will have arisen by simple modification of a pre-existing genotype, not by a complete replacement of genes at all loci which influence the characteristic. A single possible exception to this will be noted later. Only the more general of the major genotypes are included in the figure. The complete array of variability, reflecting both the dif­ferent ways in which these genotypes may be integrated with other elements of The University of Texas Publication ~!!./I! /J../.; ,,004.. ~\ t / \ I ' Jjj dd J,/ oo' Riverview, Florida Palmira, Colombia Saint Petersburg, Florida Mexico, D.F., Mexico Patagonia, Ariwna St. Lucia, B.W.I. Coroico, Bolivia Lago Pichichuela, El Salvador Port of Spain, Trinidad Barro Colorado, Canal Zone Las Cruces Trail, Canal Zone Arima Valley, Trinidad Angra dos Reis, Brazil Minatitlan, Veracruz, Mexico Las Cruces Trail, Canal Zone Turrialba, Costa Rica Lima, Peru Barbados, B.W.I. Gomez Farias, Tams., Mexico Ramsey Canyon, Arizona Antigua Road, Guatemala Patagonia X Klondike, Ariz. Bucaramanga, Colombia Carpentaro, Venezuela Cave Creek, Arizona Chile Throckmorton: Phylogeny in Drosophila Collection Species number Locality ungrouped species near repleta group Castanea Patterson and Mainland species F peruviana Duda species G species H aureata Wheeler bromeliae group species I nannoptera group nannoptera Wheeler polychaeta group polychaeta Patterson and Wheeler quinaria section immigrans group hypocausta Osten-Sacken spinofemora Patterson and Wheeler immigrans Sturtevant funebris group macrospina Stalker and Spencer subfunebris Stalker and Spencer funebris Fabricius quinaria group innubila Spencer quinaria Loew rellima Wheeler falleni Wheeler phalerata Meigen occidentalis Spencer species J tenebrosa Spencer subquinaria Spencer transversa Fallen subpalustris Spencer palustris Spencer guttifera Walker testacea group testacea von Roser putrida Sturtevant calloptera group ornatipennis Williston calloptera Schiner schildi Malloch guarani group guaramunu subgroup guaramunu Dobzhansky and Pavan griseolineata Duda guaraja King guarani subgroup guarani Dobzhansky and Pavan subbadia Patterson and Mainland H360.24 H407.140 H181.41 2507.20 H194.8 H180.42 H435.21 H381.6 H435.44 2502.5 2372.16 2321.9 1784.12 2181.3 2082.1 2076.8B 1753.7 2068.9 1062.6 1915.1 2175.3 9'29.8 2072.6 2516.1 2549.1 1877.9 1757.13 2086.3 3005.4 2539.1 2378.2 H19'2.11 2543.2 2271.1 2211.3 H407.20 2211.2 2262.24 Boquete, Panama Boquete, Panama Barro Colorado, Canal Zone Angra dos Reis, Brazil Villavicencio, Colombia San Jose, Costa Rica Leticia, Colombia Acatlan, Puebla, Mexico Leticia, Colombia Ponape, E. Caroline Is. Queensland, Australia Cross Anchor, South Carolina Durango, Mexico Willow Creek, California Minneapolis, Minnesota Silver City, New Mexico Lake Shetek, Minnesota Oakdale, Nebraska Gt. Smoky Mt. N. P., Tenn. Beirut, Lebanon Mt. San Jacinto, California Cave Creek, Arizona Eagle Nest, New Mexico Aspen, B. C., Canada England Georgetown, South Carolina Bemidji, Minnesota Austin, Texas Fairbanks, Alaska Austin, Texas St. Vicente, Cuba Rio Negro, Colombia Barro Colorado, Canal Zone Brazil Ponta Grossa, Brazil Boquete, Panama Feliz, Brazil Huatusco, Veracruz, Mexico The University of Texas Publication Collection Species number Locality cardini group dunni Townsend and Wheeler belladunni Heed and Krishnamurthy nigrodunni Heed and Wheeler polymorpha Dobzhansky and Pavan neomorpha Heed and Wheeler neocardini Streisinger parthenogenetica Stalker acutilabella Stalker procardinoides Frydenberg cardinoides Dobzhansky and Pavan cardini Sturtevant ungrouped species near cardini group species K speciesL rubrifrons group parachrogaster Patterson and Mainland uninubes Patterson and Mainland species M species N macroptera group submacroptera Patterson and Mainland macroptera Patterson and Wheeler pallidipennis group pallidipennis Dobzhansky and Pavan tripunctata group mediodifjusa Heed and Wheeler albicans Frota-Pessoa metzii Sturtevant albirostris Sturtevant unipunctata Patterson and Mainland mediopunctata Dobzhansky and Pavan tripunctata Loew trapeza Heed and Wheeler bandeirantorum Dobzhansky and Pavan paramediostriata Townsend and Wheeler mediostriata Duda mediopictoides Heed and Wheeler crocina Patterson and Mainland ungrouped species near tripunctata group sticta Wheeler species not placed in section aracea Heed and Wheeler carsoni Wheeler carbonaria Patterson and Wheeler tumiditarsus Tan, Hsu and Sheng Subgenus: PHLORIDOSA species 0 species P species Q Genus: Chymomyza amoena (Loew) 2532.4 aldrichi Sturtevant 2199.9 procnemis (Williston) 1782.7 Rio Piedras, Puerto Rico Hardware Gap, Jamaica Barbados, B.W.I. Montes Claros, Brazil Lancetilla, Honduras Angra dos Reis, Brazil Atlixco, Puebla, Mexico St. Vicente, Cuba Coroico, Bolivia Lago Pichichuela, El Salvador Tezuitlan, Puebla, Mexico Boquete, Panama Boquete, Panama Zacatecas, Mexico Boquete, Panama Guatemala Boquete, Panama Antigua Road, Guatemala Rocky Mt. Nat'l Park, Colo. Bucaramanga, Colombia Ocho Rios, Jamaica Merida, Venezuela Changuinola, Panama Barro Colorado, Canal Zone Montero, Bolivia Ponta Grossa, Brazil Austin, Texas San Salvador, El Salvador Coroico, Bolivia Mayaguez, Puerto Rico Vila Atlantica, Brazil Boquete, Panama Rio Piedras, Puerto Rico Lancetilla, Honduras Santa Tecla, El Salvador Bridgewater, Vermont Austin, Texas Hangchow, China El Colegio, Colombia El Colegio, Colombia El Colegio, Colombia Austin, Texas Morgan, Utah Durango, Mexico 2327.1 H356.3d H247.1 2374.8 H51.10 H339.1 1802.17 2378.3 H346.8 H27.9 2263.6 H407.23 H407.139 1787.3 H407.126 2510.6 H360.21 2522.9 2580.1b H191.48 H352.4 H209.17 H404.15 2538.4 H3#.33 2211.4 2539.2 H29.16 H346.42 H254.12 H341.13 H407.32 H131.2 H51.15 H46.28 2551.4 2540.2 1736.6 H4#.2B H4#.2C H4#.2D Throckmorton: Phylogeny in Drosophila Collection Species number Locality Genus: Scaptomrza adusta (Loew) 2368.1 pallida (Zetterstedt) 2532.3 ?hsui Hackman Genus: Zaprionus ghesquierei Collart 2371.3 vittiger Coquillett 1974.4 Genus: Mrcodrosophila dimidiata (Loew) 254-0.1 Genus: Zapriothrica dispar (Schiner) H444.2A Genus: Gitona americana Patterson bivisualis Patterson Genus: Rhinoleucophenga obesa (Loew) Family: DIASTATIDAE Diastata vagans Loew Family: SPHAEROCERATIDAE Leptocera sp. Family: EPHYDRIDAE Scatella stagnalis (Fallen) Discocerina obscurella (Fallen) Family: PERISCELIDAE Periscelis annulata (Fallen) Family: SEPSIDAE Sepsis violacea Meigen Sepsidimorpha secunda Melander and Spuler Family: CHLOROPIDAE Oscinella carbonaria (Loew) Oscinella coxendix (Fitch) Thaumatomyia glabra (Meigen) Family: ANTHOMYZIDAE Mumetopia occipitalis Melander Family: AGROMYZIDAE Phrtobia sp. Family: AULACIGASTRIDAE Aulacigaster leucopeza (Meigen) Family: PIOPHILIDAE Piophila casei (L.) Prochrliza xanthostoma Walker Austin, Texas Austin, Texas San Jacinto Mts., California Africa Africa Austin, Texas El Colegio, Colombia Austin, Texas Austin, Texas Austin, Texas Fairbanks, Alaska Riverside, California San Jacinto Mts., California Riverside, California Austin, Texas Austin, Texas Austin, Texas Austin, Texas Austin, Texas Austin, Texas Austin, Texas Austin, Texas Austin, Texas Austin, Texas Austin, Texas XI. Effects of X-Ray Irradiation in Drosophila virilis at Different Stages of Spermatogenesis1 2 FRANCESE. CLAYTON The genetic effects of irradiation on meiotic cells in different stages of sperma­togenesis have been reported by a number of investigators. These tests have shown a variability in the. sensitivity of spermatogenic cells during different stages in spermatogenesis. The present series of tests was designed to correlate genetic effects, as measured by dominant lethals and translocations, with the histological changes in testes following irradiation at different pupal and adult periods in Drosophila virilis. Since no complete study on meiosis in D. virilis has been reported, investigators using virilis in radiation tests have had to assume the similarity of development with spermatogenesis in D. melanogaster, taking into account the longer life cycle in virilis. Cooper (1950) described the process of spermatogenesis and differentia­tion of the testes in melanogaster, and Metz (1926) reported the course of meiosis in virilis and several other Drosophila species. Metz found no essential differ­ences in the chromosomal behavior in the species investigated; his account, how­ever, did not include the time at which various meiotic stages first occur. Clayton (1957) reported the differentiation of virilis testes during pupal development and in young adults. In order to present a more thorough analysis of the effects of irradiation on the meiotic cells, the normal differentiation of the testes in virilis adults is described in the present study. This analysis is a continuation of the earlier investigation in which the sequence of changes in the pupal testes was described. MATERIALS AND METHODS Observations of spermatogenic cells have been made from several different types of preparations; various aspects of spermatogenesis are revealed by different techniques, and each method has advantages for analysis of particular aspects of cell structures. The differentiation of the testes and the development of the meiotic cells were analyzed from sectioned material collected at stages during pupal differentiation and at regular intervals during adult development. The ages of the males were determined from the time of egg deposition and are accurate to within one hour. Material collected at regular intervals during pupal and adult periods was fixed in a modified Carnoy's solution, sectioned at ten microns, and stained with Heidenhain's hematoxylin. The cells in the testes were scored according to the stage in spermatogenesis; sperm bundles and mature spermatozoa were not counted. 1 This investigation was performed under contract AT-(4-0-1)-1974 with the United States Atomic Energy Commission. 2 Department of Zoology, University of Arkansas, Fayetteville, Arkansas. The University of Texas Publication Details of the structure of meiotic cells and spermatids were obtained through observations of living cells using phase-contrast microscopy. Testes were dissected from pupae and adults and placed in a drop of saline on a slide. A cover glass was placed over the material and slight pressure applied; this pressure caused the wall of the testis to rupture and living cells were suspended in the saline. The prepa­ration was then sealed with a 1: 1 mixture of paraffin and petroleum jelly. The testes prepared in this way could be examined and individual cells studied, both through the testis wall and as free cells from ruptured portions of the gonad. Cells were examined under the phase-contrast microscope using the dark-M series of objectives with 10, 20, 43, and 97X magnification. In order to determine the age at which spermatozoa become functional in adult males, several tests were carried out on living flies. Males were dissected in saline and the testes and accessory structures were examined. The organs were checked for rhythmic contractions in the ducts, the presence of sperm in the testes and ducts, the degree of coiling of spermatozoa, and the motility of the sperm. Mature females, which had been placed with the males, were dissected and the female reproductive organs examined. The females were checked for the presence and motility of sperm in the spermathecae and seminal receptacles. When eggs had been deposited by the females, those cultures were retained and checked for hatch­ing. Ten males and ten females were dissected at regular intervals during the first seven days of adult development. Observations of the chromosome configurations in meiotic cells were made by staining the testes in aceto-orcein for fifteen minutes. The gonads were then placed on slides in fifty percent acetic acid and squashed, using the technique utilized for salivary gland preparations. X-ray tests were carried out at a temperature of 4 to 5°C. using a Westing­ house 150kv Industrial X-ray unit. The machine was operated at 150kv and 5ma, with the exception of test 1 where it was operated at 90kv and 5 ma. Using a 0.5 mm aluminum filter and at a distance of 19.5 cm from the center of the X-ray tube, the dosage was approximately 200 r per minute at 150kv and 100 r per minute at 90kv. Calibration for dosage was made at each test using a Victoreen Condenser r-meter. The Drosophila virilis stock used in these tests was from Texmelucan, Mexico (UT 1801.1) and was maintained on the standard banana-agar-malt-yeast food at a constant temperature of 21°C. Translocations were detected by using a marker stock of D. virilis containing five mutants on the major autosomes: broken (b, 188.0, chromosome 2), tiny bristle (tb, 104.0, chromosome 3), gapped (gp, 118.5, chromosome 3), cardinal (cd, 32.2, chromosome 4), and peach (pe, 203.0, chromosome 5) . NORMAL SPERMATOGENESIS IN PUPAE AND ADULTS Cytological Analysis The spermatogonia in the testis are small cells forming a compact mass in the apex of the gonad. They undergo active mitotic divisions to provide a continuous supply of primary spermatocytes. The predefinitive spermatogonia do not occur in cysts but appear in layers in the rounded apex of the testis. The spermato­ Clayton: X-ray Efjects on Spermatogenesis gonial cell has a well-defined nucleus containing the nucleolus and other, more diffuse, material (Figure 1). In the prometaphase stage the duplicated chromo­somes are easily distinguished; the spermatogonial metaphase is a typical cell, showing the somatic pairing of the ten rods and two dots characteristic of this species. The anaphase is regular and proceeds to a typical telophase. In the nucleus of a very young primary spermatocyte, the nucleolus is distinct, and well-defined dark bodies are frequently associated with the nucleolus; in others diffuse material is present in the nucleoplasm. The cytoplasm of these cells contains some granular material, which is usually concentrated around the nu­clear membrane (Figure 2). As growth of the early spermatocyte progresses, sev­eral distinct bodies appear in the nucleus (Figure 3) . Previous investigations (Huettner, 1930; Cooper, 1950), and the present study, have not clearly indi­cated whether these bodies are chromatin in nature. Cooper, however, reported that these structures were not stained by aceto-orcein and therefore were prob­ably not nucleoprotein in nature. The nucleus continues increasing in size during growth until it is almost three times the size of the early spermatocyte nucleus. In these later stages the nu­cleolus is a spherical structure with the sex chromatin frequently in close associ­ation (Figure 4). The cytoplasm becomes more granular in appearance, and mitochondria begin to appear as threadlike structures scattered throughout the cytoplasm. The mitochondrial threads later become oriented around the nucleus, at first filling the cytoplasm and later becoming more compact around the nucleus (Figure 5). As growth of the primary spermatocyte continues, the association of the sex chromatin and the nucleolus disappears, the nucleolus fragments~ and the nucleolar material is scattered throughout the nucleus. As spindle formation pro­gresses, some of the mitochondria appear to be oriented with. individual astral rays while others surround the nucleus (Figure 6) . In the prometaphase stage the mitochondria are aligned parallel to the long axis of the spindle and are not as distinctly associated with the asters (Figure 7). The .bivalents become distinct during the period of spindle formation and are located usually in the peripheral area of the nucleus, moving to a more central position during prometaphase (Figure 8). Cooper (1950) has discussed the descriptions, by a number of investigators, on the appearance of the metaphase in the primary spermatocyte; observations on living cells during the present study indicate a regular alignment of the bivalents on the spindle (Figure 9). Anaphase begins irregularly as the chromosomes move .toward the poles. It is during this stage that the thin line separating the mito­chondrial sheath and spindle may be seen most distinctly (Figure 10). The nuclei forming during the late telophase appear almost empty as the chromatin masses seem to aggregate near the nuclear membrane (Figure 11). The mito­chondrial sheath persists through the first meiotic division and is divided pas­sively into the daughter cells during cytokinesis (Figure 12). In aceto-orcein preparations the bivalents of the primary spermatocytes are more clearly dis­tinguished than in either sectioned material or living cells. The formation of a regular metaphase plate was observed, as well as the V-shaped configuration of the rod-chromatids, held together at the centromere region, during the first anaphase. The University of Texas Publication PLATE I. Phase-contrast photographs of spermatogenesis. All photographs were taken at an original magnification of 970X, using the dark-M series of objectives. All prints were enlarged to the same degree. Fig. 1. Apex of testis showing spermatogonia, photographed through wall of testis. Fig. 2. Early primary spermatocyte. Fig. 3. Primary spermatocyte, midgrowth. Fig. 4. Primary spermatocyte, late growth; mass associated with nucleolus probably sex chromatin. Fig. 5. Primary spermatocyte, late growth; threadlike mitochrondria oriented around nucleus. Fig. 6. Spindle formation in primary spermatocyte. Fig. 7. Prophase 1, primary spermatocyte. Fig. 8. Prometaphase, primary spermatocyte. Fig. 9. Metaphase 1; spindle elongated with one bivalent in focm in center and another bivalent out of focus on right. Clayton: X-ray Effects on Spermatogenesis There appears to be a short interphase stage prior to the second meiotic division. The cytoplasmic elements, mitochondria, and probably Golgi, remain oriented about the clear nucleus. The second metaphase and anaphase stages are regular, with a normal metaphase plate and synchronous separation of chromosomes in early anaphase. The mitochondrial elements appear to be coarser during this PLATE II. Fig. 10. Anaphase 1, primary spermatocyte. Fig. 11. Late telophase, pnmary spermatocyte. Fig. 12.. Interphase. Figs. 13, 14. Second meiotic division, secondary spermatocyte. Fig. 15. Telophase of secondary spermatocyte. Fig. 16. Young spermatid, nuclear growth period; double nature and conc!)ntric layers of nebenkern visible. Fig. 17. Young spermatid, nuclear growth; centriole (C) between nucleus and nebenkern. Fig. 18. Spermatid, nuclear growth; plasmosome evident in nucleus, vesicle of acrosome may be distinguished. The University of Texas Publica:tion second division, lacking the fine threadlike appearance of these elements during the first meiotic division (Figures 13, 14). In the spermatid the chromosomes are distinct for a very short time. As the cells begin to differentiate, the chromatin appears as aggregates of rather diffuse material near the nuclear membrane; it later appears as a condensed mass near one side of the nucleus in the elongating spermatid. The early spermatids are round cells containing small nuclei, which are surrounded by the cytoplasmic elements still retaining their orientation of the second division spindle (Figure 15). As nuclear growth occurs in the differentiating spermatid the nebenkern develops and its double nature and the concentric layers which comprise this structure become apparent (Figure 16). The nucleus enlarges rapidly and the centriole moves to a position between the nucleus and the nebenkern (Figure 17). The acroblast becomes differentiated and the axial filament may be seen between the two portions of the nebenkern, growing out from the centriole (Figure 18). As elongation of the spermatid occurs, the double nebenkern loses its rounded appearance and becomes progressively longer, appearing eventually as a struc­ture composed of paired, twisted ribbons. During the progressive elongation of the spermatid, the large nucleus, with its distinct plasmosome, becomes condensed, forming the small head of the mature sperm and the cytoplasm, present during early elongation, is shed. Further differentiation of the sperm involves greater condensation of the head region and coiling of the tail portion. Details of the cyto­logical changes in the spermatids have not been included in the present report; reference is made to the description of spermiogenesis by Momma (this publica· tion). HISTOLOGICAL ANALYSIS The histological investigation of normal spermatogenesis reveals the types of cells present in the testes during pupal differentiation and adult development through the first ten days. This information offers a method of more clearly defining the critical period for irradiation of particular types of immature germ cells. As described in the previous investigation (Clayton, 195 7), larvae and early pupae in D. virilis males contain only spermatogonia and primary spermatocytes in the early growth stages. Meiotic divisions were first observed at about 230 hours, approximately two days after pupation. The spermatids were first observed in pupae about ten hours later. As the testis continued differentiating, the maxi­mum number of immature cells in the gonad was observed in pupae at 300 hours. The normal progress of spermatogenesis during adult development is sum­marized in Table 1. The meiotic cells present in the testes of young adults during the first 24 hours after emergence vary only slightly in mean frequencies, but, as may be seen in Figure 19, cell counts of spermatogonia, spermatocytes, and spermatids during development beyond the first day reveal an apparent two-day cycle in meiosis. The means for total number of cells and for primary spermato­cytes are significantly higher on the second, fourth, and sixth days as compared with the alternate days. The means for meiotic cells in pupae of 320 hours are included since adults emerge shortly after 320 hours; therefore, values for this period represent the differentiation of meiotic cells approximately two days Clayton: X-ray Effects on Spermatogenesis earlier than in the 48-hour adults. These values indicate a significant difference in cell counts obtained on the first, third, fifth, and seventh days, when compared with those from the 320-hour pupae, second, fourth, and sixth-day adults. The means of primary spermatocytes on the "odd-numbered" days show no signifi­cant change, while these same cells in 320-hour pupae, and on the second, fourth, sixth, and eighth days indicate steadily diminishing means until the value on the eighth day is at the same level as found on the alternate "low" days. The peaks of the total number of cells are a reflection of the changes in frequencies of pri­mary spermatocytes since spermatogonia and spermatids fail to show these con­sistent changes. D. virilis males produce functional spermatozoa at six days of age, and, although coiled sperm are present in the ducts earlier, very few fertile eggs may be obtained from matings with males prior to the sixth day. It may be noted that, following this sixth day, the numbers of meiotic cells no longer show .a two-day cycle but a steady decrease in frequencies. This indicates a possible correlation between the age of the fully mature male and the meiotic cycle. The relative frequencies of spermatogonia, spermatocytes, and spermatids are plotted in Figure 20, shown as percentage of total cells present in each of the first nine days of adult development. The percentages of the primary spermatocytes and spermatids which are present on succeeding days also give some indication of a two-day cycle; in general, an increase in spermatocytes on one day is followed by an increase in spermatids on the following day. Fluctuations in the relative frequencies of spermatogonia are small when compared with spermatocytes and spermatids, but the frequencies during the first eight days correspond more closely to those of the primary spermatocytes than to those of the spermatids. TABLE 1 Mean number of meiotic cells in testes of adult males Age in days Stage of meiosis 2 3 4 5 6 8 9 13 Spermatogonia 184 297 21 1 276 284 318 283 232 217 145 Primary sperma tocytes 1,099 1,444 1,092 1,379 1,027 1,273 1,125 1,026 848 782 Prophase, growth 1,084 1,420 1,079 1,368 1,009 1,255 1,113 1,009 838 772 Meta phase 6 2 4 2 5 6 3 3 1 6 Anaphase 2 1 5 2 4 6 2 2 2 Telophase 7 21 4· 7 9 12 3 15 7 2 Interphase 11 11 8 1 6 Secondary spermatocytes 31 21 15 13 5 9 3 2 15 6 Prophase 11 2 Meta phase 6 16 6 6 3 9 1 2 2 Anaphase 3 2 5 3 2 2 7 2 Telophase 11 3 2 4 2 6 1 Spermiogenesis 609 564 436 501 377 383 358 302 321 212 Early spermatid 55 34 13 15 10 16 16 6 7 3 Nuclear growth 148 103 125 11 5 107 86 118 76 109 47 Linear growth 146 226 170 230 149 190 131 143 111 106 A xial filament 260 201 128 141 111 91 93 77 94 56 TOTAL 1,934 2,337 1,762 2,170 1,694 1,983 1,770 1,562 1,401 1,151 2800 2600 2400 2200 2000 1800 1600 1400 1200 1000 800 600 400 200 The University of Texa.s Publication % \ Fig. 19 70 60 50 40 0 100 90 • BO 70 60 50 40 30 20 20 -;1-·-. ~.........__ . ....-......._~Spermotids '--·=-:-.~~.·--•-"'•yl ···11-.... .... 10 ···.........-······.. ··::·Spermotogonia ....1 1-.~ 320 3 4 5 6 7 B 9 13 Doys Hou~ Emergence Age of Adults Fig. 21 100 ,,...--, % / Pupae-\ \ \ 90 \ \ \ 80 A\ \ \ // \, _,..:...Pupae \ \ \ 70 + AdullS \ \ \· \ .\ 60 \1 I ' ' \ -Matdre Adults \ 50 ' ' \ \ -Larvae I 500 r I\ I 40 I I I __...\ \ \I ". \, \...._LarYoe " 250 r 30 A c D Motin9 Pe riods \ \ '-·--........... Spermot1ds /'·'>­ ......_ _-......., . ·'--J \ .......,.-·~permotogonia ..··········· .·· 2 3 4 5 6 7 8 9 13 Days Age of Adults Fig. 22 .// \ I -, _,,..... Pupae + \ \ Adults \ \ \ ', \ ' Adu lts '\ \ \ Pupoe I I Mature Adults I I I I I I , I , I,// Immature Sperm Meiotic Early Spermatogonio ISpermotids ISpermotocytes I Frc. 19. Mean numbers of meiotic cells per male in Drosophila virilis adults and 320-hour pupae. Frc. ZO. Relative frequencies of meiotic cells in adult males of Drosophila virilis. Frc. Zl. Dominant lethal rates from irradiation at different stages of development, as single or fractionated doses. Pupae: single dosage of 1089r; young adults: single dosage 'of 1089r; mature adults: single dosage of 1089r; larvae: single treatments of 500r or Z50r; pupae+ adults: 1000r administered as 500r to young pupae and 500r to young adults. Periods A through H represent eight successive two-day mating periods. Clayton: X -ray Effects on Spermatogenesis The results of this study and the previous investigation present the histological changes in the testes of D. virilis from late larval stages through the first nine days of adult development. The cell frequencies determined for normal spermato­genesis during these different stages make it possible to correlate the normal con­dition with histological and genetic effects from X-ray tests on males of different ages. PRODUCTION OF FUNCTIONAL SPERMATOZOA The age at which adult D. virilis males produce functional spermatozoa is of importance in determining the time for mating in irradiation tests where specific types of meiotic cells are to be checked. This information, obtained from dissec­tion of males and females, is summarized in Table 2. In flies just after emergence, rhythmical contractions of the ducts occur, but no sperm were present in these regions. The peristaltic contractions extended anteriorly into the testis and some cells were forced posteriorly by these move­ments. The spermatozoa were long and still in bundles and no coiling had oc­curred. During the first twenty-four hours, the location and activity of the sperm showed little change. In one male at six hours, some sperm bundles were observed in the upper portion of the anterior ejaculatory duct, but these cells were not motile and were not convoluted, as is characteristic of the mature functional TABLE 2 Location and activity of spermatozoa in adults of Drosophila virilis Males Females Con-Ant. Post. Egg Age tractions Coiling Ej. D. Ej. D. Motility Sper. S.R. :vlotility Fertility 0-5 minutes 10 0 0 0 0 0 0 0 6 hours 9 0 1 0 0 0 0 0 12 hours 10 0 2 0 0 0 0 0 18 hours 10 0 0 0 0 0 0 0 24 hours 10 0 1 0 0 0 0 0 36 hours 10 0 2 0 0 0 0 0 48 hour 10 0 0 0 0 0 0 60 hours 10 0 7 1 0 0 0 0 3 days 9 0 8 0 0 0 0 0 4 day 10 2 10 2 0 0 0 0 5 days 10 10 10 4 0 0 0 0 6 days 10 10 10 10 10 10 10 10 + 7 days 9 10 9 9 9 10 10 10 + The numbers given in the table indicate the positive j·esults based upon ten dissections of each sex. Males Females Contractions. Rhythm ic movement of the male Sper. Prcse'1ce of sperm in sperma thecae. reproductive ducts. S.R. Presence of spenn in scm in()l receptacles. Coi ling. Tightly convoluted spenn tails. Motility. Active movement of sperm in either Ant. Ej. D. Presence of spcnn in the anterior spcrma lhccae or seminal receptacles. ejacul atory duct. Post. Ej. D. Presence or sperm in the posterior ejaculatory duct. l\lotility. Active movement in any region. Fie. 22. Dominant lethal rates by stage in spermatogenesis at time of irradiation. Pupae: single treatment of 1089r; adults: 4-6 hours, single treatment of 1089r; mature adults: single treatment of 1089r; pupae+ adults: 1000r administered as 500r to young pupae and 500r to voung adults. The University of Texas Publication spermatozoa. The first occurrence of sperm in the posterior ejaculatory duct was observed in one male at sixty hours, but nine of the ten specimens examined at this time lacked sperm in this posterior duct; no expansion of the duct had oc­curred. At the end of four days, all ten males had sperm in the anterior duct, but only two of these had sperm in the posterior duct. The lumen of the posterior ejaculatory duct at this time was filled with fluid and expanded. Two of the males contained spermatozoa that possessed the highly coiled tails. After five days, all males had coiled spermatozoa and these cells were massed in the anterior ejaculatory duct, but only four had sperm in the posterior ducts. None of the females examined at this time had been inseminated. These results agree with those from the control test on virilis reported by Stone, Haas, Alex­ander, and Clayton ( 1954), in which only two percent of the males produced any offspring during the first five days after emergence. At the end of the sixth· day the females had been inseminated and the sperm were highly motile. One male showed no peristaltic contractions and no sperm in the ducts; however, those . spermatozoa in the posterior portion of the testes were highly coiled. Not included in the table are observations made on males after thirty days. The sperm supply was less abundant in these older males and the lumen of the testis was sometimes filled with cellular debris. Some spermatogonia and sperma­tocytes were still present in the apex of the testes, although they were scarce. Females placed with these males contained motile sperm, although not as abun­dantly as in the younger stages. . EFFECTS OF X-RAY IRRADIATION Mature adults (tests I and V) were irradiated to check the damage to mature spermatozoa. Because of an error in the original calibration, flies in test I re­ceived a total dosage of only 100 r. Mature adults were irradiated in a second test, with the dosage of 1,089 r, and results from this experiment are presented as test V. During normal development the adult males, shortly after emergence, do not have functional spermatozoa in the testes although sperm ·bundles are present; the predominant types of cells are spermatids in various stages of dif­ferentiation. The immature adults in test II were four-to-six hours old at the time of irradiation; therefore, this test would expose large numbers of developing spermatids and nonfunctional spermatozoa. In normal development, the period of 30-to-40 hours after pupation is the time during which cells are dividing rapidly and spermatocytes are numerous. Very young spermatids are being produced but elongation and differentiation of the spermatids has not begun. Males of this age, test IV, were tested to determine effects on cells prior to spermiogenesis. In larvae, prepupae, and early pupae, only spermatogonia and very young primary spermatocytes were found in the normal testes. Three experiments, using different X-ray dosages, were carried out on larvae to determine the irradiation effects on premeiotic spermatocytes and spermatogonia (tests III-a, -b, -c). In this series, lower dosages of 250 r (test III-a) and 500 r (testUI-b) were used in addition to 1,089 r (test III-c). In test III-c, most larvae died shortly after treatment; however, a few adults were obtained but all were abnormal in structure and none survived long enough for Clayton: X-ray Effects on Spermatogenesis mating. Normal adults were obtained from both tests where lower X-ray dosages were used. A control test was carried out to determine the percentage of dominant lethals and translocations in the stock to be used for the irradiation tests. No transloca­tions were recovered and the dominant lethal rate was 7.9 percent. Dominant Lethals Six test series were analyzed for the frequencies of dominant lethals. Lethal rates were determined by percentage of pupae which developed from the total number of eggs laid. Results of these tests are presented in Table 3 and in Figure 21. Males which were irradiated as mature adults, tests I and V, were mated within one hour after irradiation to virgin females. The pairs were trans£ erred to fresh food daily and egg counts were made every 24 hours. After the second day, males were removed. and the eggs deposited by the females were counted for three additional days. The two-day mating period eliminated those sperma­tozoa which had been in early spermatogenic stages at the time of irradiation. Two X-ray tests were carried out using mature males. The dominant lethal rate from exposure to 100 r was very low, within about one percent of the control rate. In test V, mature adults were given 1,089 r and the dominant lethal rate was 45.6 percent. In both of these experiments mature spermatozoa were tested. Males treated as immature adults, test II, were placed with mature virgin females within one hour after irradiation and left for five days. After this pre­liminary period, during which no fertile eggs were deposited, the males were remated to mature virgin females (period A) and daily egg counts were made. After a two-day mating period, the males were remated (period B) and egg counts made. This procedure was continued for a total of eight consecutive mating periods (A through H) . For each period the male remained with the female for two days, during which time egg counts were made; then, after removal of the males, daily counts were made for three additional days, thus giving a total of five days for the egg counts for each period. The most mature cells present in adult males four-to-six hours after emergence are sperm bundles; these cells are morphologically quite similar to motile spermatozoa but are not functional at this age. The dominant lethal rate for the most mature cells is only slightly higher than that for mature spermatozoa, but the rate increases rapidly during periods G and D, reaching a peak at period E. These periods correspond to early and developing spermatids. Earlier stages in spermatogenesis are represented by periods F, G, and H, which probably correspond to dividing spermatocytes, young premeiotic spermatocytes, and spermatogonia. In test IV, pupae of 30-to-40 hours were treated with 1,089 r. The testes at this time contain early spermatids as the latest stage of differentiation. The successive mating periods (A through H ) therefore indicate the dominant lethals produced in: very young spermatids, followed by dividing spermatocytes, young growing spermatocytes, and spermatogonia; If the curve showing the dominant lethals from irradiation of pupae is shifted on the graph to the corresponding stages of spermatogenesis in young adults, the close correlation of the dominant lethal rates for similar stages in meiosis may be seen. (Figure 22). The University of Texas Publication TARLE 3 Dominant lethal rates from X-ray irradiation at different developmental stages in Drosophila virilis males Per cent Per cent lethals Percenta7e males Per cent Average Test Dosage ir/~~~:i~n Eggs Pupae Adults (pupae/ eggs) (adults pupae) producln g fem ales no. ergs offspring in seminated dai y Control 14 days 1,674 1,555 1,524 7.9 98.0 100.0 100.0 16 (Adults) I 100r 14 days 3,2 12 2,923 2,858 9.0 97.8 100.0 100.0 14 (Adults) II 1089r 4-6 hours (Adults) A 1,548 782 769 49.6 98.3 89.7 95.0 13 B 1,290 554 540 57 .1 99.3 96.3 100.0 12 c 791 122 121 84.6 99.2 81.5 100.0 7 D 55 7 39 37 93.0 94.9 50.0 100.0 5 E 309 12 12 96.1 100.0 26.9 94.7 3 F 707 30 24 95.8 80.0 57.7 100.0 7 G 793 128 121 83.9 94.5 88.5 100.0 8 H 1,760 796 766 54.8 96.2 96.2 100.0 17 III-a 250r Mature Larvae A 918 634 618 30.9 97.5 100.0 100.0 18 B 659 445 429 32.5 96.4 93.3 100.0 11 c 687 536 523 22.0 97.6 100.0 100.0 13 III-b 500r Mature Larva e A 1,363 647 621 52.5 96.1 94.1 100.0 20 B 470 302 295 35.7 98 .0 93.3 100.0 8 c 755 497 480 34.2 96.8 100.0 100.0 14 IV 1089r 30-40 hrs. ( Pupae) A 58 5 5 91.4 100.0 6.5 20.0 1 B 94 5 3 94.7 60.0 9.7 80.0 1 c 298 15 14 95.0 93.3 33 .3 90.0 3 D 594 116 109 80.5 94.0 57.7 61.7 6 E 944 364 343 61.4 94.2 69.0 83.3 8 F 449 199 184 55 .7 92.5 50.0 65.0 4 G 236 164 159 30.5 97.0 39.0 75 .0 2 H 904 569 562 37.l 98.8 60.0 95.0 9 v 1089r 6 days 3,120 1,697 1,652 45.6 97.5 96.2 100.0 15 (Adults) VI 1000r 230-240 hrs. pupae (5 00r)+4-6 hrs. adults (500r) A 202 56 54 72.3 96.4 66.7 50.0 3 B 947 333 3 12 64.8 93.7 91.7 100.0 13 c 976 315 299 67.7 94.9 100.0 100.0 15 D 841 218 212 74.1 97.2 100.0 100.0 12 E 536 175 170 67.3 97.1 100.0 100.0 8 F 546 258 243 52.4 94.1 100.0 100.0 9 Clayton: X-ray Effects on Spermatogenesis In each of the tests described above, the females were dissected following the five-day period of egg laying. The spermathecae and seminal receptacles were examined to determine whether motile sperm were still present at this time. The percentages obtained for each of the tests are included in Table 3. The only strik­ing variation from the control was found in test IV, where young pupae were irra­diated. In period A, which is the first mating period and therefore a test of the most mature cells present at the time of irradiation, only twenty percent of the females were inseminated. During this period the number of eggs deposited was very small, averaging less than one per day per female. The percentage given for the dominant lethal rate in this period, therefore, must be considered in rela­tion to the small number of functional spermatozoa available for fertilization of the eggs. The percentage of insemination increased in period B and later periods, approaching the normal condition only during the last mating period (H). Irra­diation during early pupal development, when the cells are spermatogonia, spermatocytes, and young spermatids, apparently destroys or prevents normal development of spermatozoa so that there is a scarcity of functional cells for fertilization. The percentage of adults which emerged from pupae is given for each test. Except for two periods (II-F and IV-B) where the numbers are very small, the percentage of emergence is above 90 percent, which would indicate that most of the dominant lethality occurred prior to pupation. At the time of counting the adults a check was made on the sex ratio. These data have not been included 'in the table as no significant variation was obtained from any of the tests. A test (VI) was carried out in which pupae were treated with 500 r and then given a second 500 r dosage as young adults. The dominant lethal rates from this test were compared with the dominant lethals obtained from single treatments of 1,089 rat each of the comparable developmental periods (tests II and IV). The dominant lethals from successive two-day mating periods, A through H, of these three tests are shown graphically in Figure 21. The dominant lethals from a single mating of six-day males exposed to 1,089 r have been included for com­parison of the effects of irradiation on mature spermatozoa with the more imma­ture cells. When young adults received 1,089 r of X-ray irradiation, the rate of dominant lethals increased rapidly from period A through E and then decreased sharply in periods G and H. Irradiated pupae showed the peak of dominant lethals during the first three mating periods, A through C, followed by a sharp decrease in periods D through H. When flies were irradiated in two different stages, 500 r as young pupae and 500 r as young adults, the dominant lethals for each of the early periods were intermediate between those obtained from the two tests of 1,089 r treatments. The curve lacks the distinct peak found in both of the single treatments. The correlation between the dominant lethals produced and the stage of spermatogenesis of the cells may account for the differences in the curves. The correlation is more easily shown as presented in Figure 22, where the rates are given according to types of cells irradiated. Irradiation of young adults results in exposure of all stages in spermatogenesis and spermiogenesis but does not include motile, functional spermatozoa. The sperm bundles present at this time are non­motile and are not functional until the males are about six-days old. Since the The University of Texa.s Publication lethals from the first mating period (A) result from irradiation to the most mature cells present in the testes at the time of exposure, the rate for each stage has been plotted with period A indicating the dominant lethal rate from immature sperma­tozoa. The successive mating periods present the rates obtained from younger spermatids and meiotic cells, and, in the final periods, the spermatogonia. If the results from irradiation of pupae are plotted according to spermatogenic cells rather than mating periods, the close similarity of the two curves may be seen. In the pupae, at the time of irradiation, the most mature cells in the testes were spermatocytes undergoing meiosis. The first mating period, therefore, would correspond with period D of the test with adult males. The flies receiving two dosages of 500 r were irradiated as pupae containing only spermatogonia and spermatocytes and as very young adults containing all stages except mature spermatozoa. The dominant lethals from the first period of this test would indi­cate the irradiation effects from 500 r as primary spermatocytes and a second 500 r as elongated spermatids or non-functional spermatozoa. The lethal rate from this period is intermediate between the rates from period A of the adult test and period A of the pupal test. If the latter percentages are averaged, the expected dominant lethal rate, intermediate between the two, would be 70.5 percent. The actual value obtained from test VI was 72.8 percent. In the suc­ceeding stages the overlapping of the two dosages results in a leveling of the curve found in the 1,089 r tests. The peak of the curve in this test probably represents the dominant lethals produced from treatments of spermatogonia and primary spermatocytes undergoing meiosis. The final period (F) would give the dominant lethal rate from two successive treatments of spermatogonia; thus, the dominant lethal rate for this period approaches those from the single 1,089 r treatments of pupae (period F) and adults (period H). There is necessarily some overlapping of the stages as to the types of cells in the successive periods, but the variation in sensitivity of the different stages of spermatogenesis can be seen. Spermato­gonia are least sensitive, spermatozoa are slightly more sensitive, and young spermatids and spermatocytes during meiosis have the highest susceptibility. On the basis of these tests, the sensitivity of spermatogenic cells to X-ray irradiation, as detected by dominant lethal rates, agrees with the results obtained by other investigators; greatest susceptibility is found among early and develop­ing spermatids, and many of the spermatogonia and premeiotic spermatocytes are killed after irradiation, with few functional spermatozoa produced. Trans locations In conjunction with the dominant lethal tests, a series of translocation tests were undertaken in which D. virilis males from each of the radiation tests were crossed with mature virgin females of the stock in which recessive mutants marked the four major autosomes. The treated males were remated to virgin marker fe­males at two-day intervals, resulting in groups which corresponded to periods A through H in the dominant lethal tests. The results of these experiments have been reported by Chang (unpublished) and are summarized in Table 4 and Figure 23. The highest percentage of translocations from irradiated immature adults was recovered at period E, the period of highest damage in the dominant Clayton: X-ray Effects on Spermatogenesis lethal tests. In the test with pupae, very few F 2 off spring were obtained from periods A through F and no translocations were recovered. The number of viable sperm increased during the last two periods ( G and H) and several translocations were recovered. No translocations were recovered from larvae after 500 r treat­ment, although one was recovered from larvae after 250 r. The rate of transloca­tions from mature adults after 1,089 r was about ten percent. Two "unexpected" translocations were recovered, one from the 250 r test on larvae and the other from the 100 r test on adults. Alexander and Stone ( 1955) reported that, on the basis of their results, translocations would not be expected below a minimum X-ray dosage of 295 r. The recovery of the highest number of translocations at period E indicates a high sensitivity of spermatids and meiotic cells. The preme1otic cells lack this sensitivity evidenced by the later stages. The test on pupae resulted in the recovery of only two translocations, one in period G and one in period H. These results also indicate the resistance of premeiotic cells to chromosome breakage by X-ray irradiation. The irradiation of mature spermatozoa, test V, resulted in 9.4 percent TABLE 4 Translocation rates at different stages of spermatogenesis T est Dosage Age of male at irradiation N umber of sperm tes ted N umber of Lranslocations Percentage transloca tions I 100r Adult, 14 days 207 0.5 II 1089r Adult, 6 hours A 78 6 9.4 B 231 32 13.9 c 169 36 21.3 D 31 13 41.9 E 13 8 61.5 F 6 0 0.0 G 12 0 0.0 H 206 0 0.0 III-a 250r Mature larvae A 99 1 10 B 79 0 0.0 III-b 500r Mature larvae A 193 0 0.0 B 27 0 0.0 c 159 0 0.0 IV 1089r Pupae, 220-230 hours A 0 0 0.0 B 2 0 0.0 c 6 0 0.0 D 8 0 0.0 E 122 0 0.0 F 48 0 0.0 G 54 1.9 H 119 0.8 v 1089r Adult, 6 days 234 22 9.4 The University of Texas Publication 't. 't. 70 60 50 40 30 20 10 % 130 120 110 70 60 160 Fig. 23 150 140 130 120 --Adults 110 'Mature Ad ults Mating Periods I I Fig. 24 70 I /\ 280 /// / \ \ / \ // \ 't. .,,,,.._ Spermatocytes \ \ \ \ \ I I 110 \ I I I I I I I I I I Control 100 90 80 _/~permotOQOnio \ ---· 60'---'-~~~~~~~~~~~~~~ 300 24 46 Hours Pupae Adu Us Fig. 26 f, I I 180 190 200 210 Hours 6 9 10 Days F1G. 23. Reciprocal translocation rates from irradiation tests at different developmental stages. Periods A through H represent eight successive two-day mating periods. Pupae: 30-40 hour pupae, 1089r; adults: 4-6 hours, 1089r; mature adults: 6 days, 1089r; larvae: just prior to pupation, 250r. FIG. 24. Meiotic cells from irradiated larvae (test III-c) plotted as percentage of control cells present in testes at the same periods. FIG. 25. Meiotic cells from irradiated pupae (test IV) plotted as percentage of control cells present during the same pupal and adult stages. FIG. 26. Meiotic cells from irradiated adults (test V) plotted as percentage of control cells present during the same periods. Clayton: X-ray Effects on Spermatogenesis translocations, identical with the percentage recovered in period A following irradiation of young adults. These tests indicate that the spermatozoa are less sensitive than the spermatids and meiotic cells but more sensitive than those cells in premeiotic stages. Histological Study of the Testes Following Irradiation Immediately following irradiation in tests III-c, IV, V, and VI, some of the treated males were separated for the histological study. These males were fixed and imbedded at intervals following X-ray exposure, using ten males at each period. The imbedded material was sectioned and stained and complete counts of the immature spermatogenic cells in both testes were made; results are summar­ized in Table 5. The counts from control groups of comparable stages are based on an earlier study (Clayton, 1957) and on additional counts from control adult males given in Table 1. T ABLE 5 Comparison of cell frequencies in testes of irradiated and control D. virilis males Spermatogonia Primary !'pennatocytes Spermatids Total Test Test/control % Test/control % Test/control % Test/control % III-c Late larvae 1089r Post-irradiation: 2 hrs. 382/ 585 65 .3 685/778 88.0 1160/ 1270 91.3 16 hrs. 344/492 70.0 1323/1054 125.5 1667 / 1546 107.8 21 hrs. 335/345 97. 1 1346/141 8 94.2 1681 / 1763 95 .3 26 hrs. 264/387 68.2 1418/1955 78.1 1713/2342 73.1 IV Pupae, 250 hrs. 1089r Post-irradiation: 23 hrs. 139/ 173 80.3 1195/1798 66.5 568/559 101.6 1916/2605 73.6 42 hrs. 258/247 108.5 1316/1409 93.3 565/ 795 72.3 2167/2491 87.6 69 hrs. 273/310 88.1 1223/1609 76.0 469/ 609 77.0 1982/2574 77.0 93 hrs. 145 985 454 1603 6 hrs., ad ults 226/ 182 124.2 11 29/1101 102.5 708/613 11 5.5 2030/ 1933 105.0 24 hrs., adults 296/184 160.8 1080/ 1099 98.2 548/ 609 90.0 1956/1934 100.1 48 hrs., adults 308/297 104.0 1106/1444 76.6 429/564 76.0 1854/2338 78.0 v Adults, 6 days 1089r Post-irradiati on: 2 hrs. 238/318 74.8 861/ 1273 67.7 265/383 69.2 1365/1983 68.8 24 hrs. 189/ 283 66.8 940/ 1125 83.6 280/359 78.0 1401/1771 79.1 48 hrs. 162/ 232 69.8 744/1028 72.+ 207/ 299 69.3 111 6/1565 71 .3 3 dnys 153/21 7 70.5 961/848 11 3.3 242/372 65 .1 1363/1454 93.7 4· days 168/ 259 64.9 840/ 11 62 72.3 197/288 68.4 1212/1718 70.5 VI Pupae, 225-230 hrs., 500r; + Adults, 4~6 hrs., 500r Post-irradiation: Pupae, 5 hrs. 269 / 349 77.1 1900/1889 100.6 2168/2265 95.3 15 hrs. 309/359 86.1 1959/1893 103.5 0/134 2225/ 2445 90.0 25 hrs. 23 7 / 284 83.4 2148/1743 123.2 0/522 2392/ 2591 92.3 45 hrs. 235/ 25 7 90.7 2051/1516 135.3 436/722 60.0 2762/ 255 1 108.3 The University of Texas Publication Using the numbers in the controls as 100 percent for each period, the counts from irradiated larvae (test 111-c) are plotted as percentages of the control cells present at the periods indicated (Figure 24). Inthis test the survival of larvae and pupae following 1,089 r exposure in the larval stage was so low that no pupae could be collected beyond the 210-hour period. During pupation the percentages of spermatogonia and spermatocytes in the controls showed considerable variation for a particular hour and it is probable that the only significant portion of the curve in Figure 24 is the reduction in both spermatogonia and primary spermato­cytes at 210 hours. The total number of cells present in the irradiated specimens did not vary considerably from the control until that period and the relative frequencies of spermatocytes and spermatogonia at 180 and 190 hours may be indicative of the same variability found in the control. It. is within this period that the spermatocytes increase rapidly and the relative frequency of the sperma­togonia shows a corresponding decrease. During these periods, therefore, the frequencies of the two types of /cells are dependent upon the specific time in development during which the increase in spermatocytes began. The values for irradiated pupae (test IV) have been plotted in Figure 25. When pupae were irradiated at about 250 hours, emergence of the adults was delayed about two days. For this reason there are no control counts for comparison during the last two days of pupal differentiation. It is during this period that the total number of cells in the irradiated testes reached the lowest level, prior to a rapid increase in the young adults. This would indicate that irradiation at the 1,000 r level resulted in cell deaths among spermatogonia, spermatocytes, and young spermatids. This reduced the number of these cells during pupal differentiation, with recovery during early adult development. Comparisons of the mean num­bers of spermatogonia and spermatids from the test with those of the late pupal differentiation of the controls reveals a correlation between the development of the germ cells of these two periods. The numbers in the young adults from the test are very similar to the numbers in the control just before emergence of the adults. The primary spermatocytes during the first 24 hours are at the normal level and the total number of cells present in the testes is at the normal level. During the second day of adult development, the cells in testes of irradiated flies failed to increase in numbers but maintained the same level or decreased slightly. In the controls the second day is characterized by a significant increase in meiotic cells. Mature adults were treated with 1,089 rat six days (test V) and sections were examined from flies at six, seven, eight, nine, and ten days. The mean numbers of spermatogenic cells plotted as percentage of control cells are presented in Figure 26. The mean numbers of spermatogonia and spermatids were lower than those of the control but comparison of the means indicates that the changes in numbers during the five-day period correspond to the variations in the control. The fre­quencies of primary spermatocytes from the test series are quite different from the controls during the same periods. The two-day cycle found in young adults from the control series is indicated in the test series following irradiation. The numbers of spermatocytes in the control drop during the sixth through ninth days and increase on the tenth day. All types of cells were reduced below the Clayton: X-ray Effects on Spermatogenesis level of the control except that primary spermatocytes increased above the con­trol level on the ninth day. Irradiation with 500 r in the pupal stage and an additional 500 r in the young adult stage (test VI) gives some indication of damage to spermatogenic cells irradiated at two different stages in the meiotic cycle. Although results are not complete, several differences during pupal development may be seen from stages already studied. Irradiation in the early pupal period before active meiotic cells are present results in damage to spermatogonia with the resulting decrease in the number of spermatocytes present. The recovery of spermatogonia from radiation effect is indicated by the relative percentage of these cells reaching the control level about two and one-half days after irradiation. The primary spermatocytes are at about the same level as the controls through the first day but the percent­age subsequently increases above the control value. This may be correlated with the day in spermatid formation. The meiotic divisions in the control series re­sulted in the presence of young spermatids at 240 hours, with a corresponding decrease in the frequency of primary spermatocytes. In the test series, spermatid formation was delayed and none were observed prior to 270 hours. This delay in the progress of spermatogenesis may be correlated with the corr~sponding delay in eclosion of adults. Normal males emerged from the pupae at about 320 hours, whereas irradiated males emerged at about 360 hours. DISCUSSION Dominant Lethals A number of investigators have determined the degree of dominant lethality resulting from radiation at different dosage levels and at different stages in spermatogenesis. Since dominant lethals are detected by determining the percent­age of adults or pupae which develop from eggs, any radiation effect which pre­vents this development would be scored as a dominant lethal. The failure of spermatozoa from X-rayed males to fertilize eggs has been reported by several workers. Kaplan (1958) examined sections of eggs and reported that unfertilized eggs contributed significantly to dominant lethals following irradiation of 500 r, 1,500 r, and 2,500 r. The sterility increased with dosage and was highest during the period corresponding to irradiation of spermatocytes and spermatogonia. Stone, Haas, Alexander, and Clayton (1954) examined eggs by the smear tech­nique devised by Patterson, Stone, and Griffen (1942); they attributed a large proportion of dominant lethals during sensitive spermatogenic stages to failure of irradiated males to produce functional sperm. During radiation tests on various stages of meiosis and spermiogenesis, the number of functional spermatozoa available to fertilize the eggs varies with the stage irradiated. Stone (1956) reported 92 percent fertility in Drosophila virilis during periods corresponding to irradiation of spermatozoa and spermatids, with a reduction to 36 percent fertility during spermatogonial periods. Alexander and Stone ( 1955) reported late spermatogonia so susceptible to radiation that many cells were destroyed and few sperm produced. In an earlier study, Moore (1932) reported the highest percentage of sterility from irradiation of D. melanogaster The University of Texas Publication larvae, 24-36 hours of age, obtaining fifty percent sterility as compared with 16.7 percent from irradiated adult males. In the present series of tests the absence of functional spermatozoa or the in­ability of the sperm to fertilize eggs was significant only in the test in which 30­40 hour pupae were X-rayed with 1,089 r (test IV). This is indicated by low percentages of insemination and also by the small percentage of males producing any offspring. The sensitive cells were spermatids and spermatocytes; recovery during later mating periods would indicate a greater resistance among early spermatogonia. Irradiation of young adults (test 11) resulted in a number of males showing temporary sterility; during the most sensitive period only 26 per­cent of the males produced any offspring; however, among the eggs examined there was a high percentage of insemination. The dominant lethals detected from irradiation at different stages of spermato­genesis indicate varying sensitivity at different periods. It has been reported by Moore (1934), Serebrovskoya and Shapiro (1935), Timofeeff-Ressovsky and Zimmer (1947), Muller (1950), and others, that the lethal frequency is signifi­cantly higher in spermatozoa than in spermatogonia. Experiments designed to test other stages in the meiotic cycle have shown that spermatids have the highest sensitivity, and that there is decreased susceptibility among spermatozoa, and greatest resistance in spermatogonia (Oster, 1958b; Fahmy and Fahmy, 1958; Belgovsky, 1958; Muller, Herskowitz, Abrahamson, Oster, 1954; Liining, 1952a; Alexander, Bergendahl, and Brittain, 1959). Khishin (1955) analyzed muta­genic effects by measuring frequencies of sex-linked lethals from irradiated eggs, larvae, pupae, and young adults. The mutation rate was low through the larval period although he found a trend toward increasing sensitivity with larval age. In young prepupae the rate was three times higher than in the larvae; there was a gradual and significant increase to the mid-pupal period followed by decreasing rates to eclosion. His results also indicate highest susceptibility among spermatids and meiotic cells and greatest resistance among premeiotic spermatocytes and spermatogonia. The present tests indicate similar variations in sensitivity of the meiotic stages. The highest percentages of dominant lethals were obtai~ed from mating periods corresponding to spermatids and meiotic cells following irradiation of pupae and young adults. The dominant lethals decreased rapidly during spermatogonial stages. Auerbach ( 1954) stated that failure of eggs to hatch could not be equated with dominant lethals; there is considerable evidence that a large number of the so­called dominant lethals are the results of inviable rearrangments. At lower dosage levels, however, dominant lethal frequencies are essentially linear (De_merec and Fano, 1944; Demerec, Kaufmann, Sutton, and Hinton, 1940; Catcheside and Lea, 1945) and only at higher levels do the frequencies increase in a manner similar to rates from rearrangements. Demerec and Kaufmann (1941) described tests at different dosage levels from 5,000 r to 10,000 r; some fertilized eggs failed to develop and death was attributed to loss or duplication of chromosome regions. In studies of sex-linked lethals, Oliver ( 1932) reported 5.3 percent associated with chromosome aberrations following 385 r, 7.3 percent after 1,540 r, and 24.6 per­ Clayton: X-ray Effects on Spermatogenesis I cent after 3,180 r. Demerec (1937) found lethals coinciding with one of the breakage points in 24 of 26 sex-linked lethals examined. Pontecorvo (1942) examined the problem of dominant lethals and proposed that lethal effects proportional to dosage may be explained on the basis that single breaks which do not undergo restitution or participate in rearrangements under­go fusion of sister chromatid broken ends, starting a breakage-fUsion-bridge cycle leading to genetic unbalance and death. Bonnier and Liining (1950) reported an mcreasing effect of X-rays due to increasing sensitivity of spermatozoa stored within the testis and proposed that this sensitivity may be the result of a reduction in the rate of restitution in immature sperm. Alhough no cytological analyses were made in the present study, abnormal mitotic figures were observed in some late-pupating larvae from the dominant lethal tests. This may be an indication that chromosomal aberrations contribute to the dominant lethality. In some cases these larvae were still active in the food while adults were emerging from the pupae; a few pupated but adults did not emerge. Counts for dominant lethal rates would include these larvae in addition to the eggs which failed to develop due to absence of sperm or inviable chromo­somal aberrations. The high percentage of lethals from lower dosages exposing larval stages indi­cates that dominant lethality cannot be accounted for entirely by two-break aberrations since the translocation tests indicate a very low percentage of recover­able translocations from exposure of spermatogonia and early primary spermato­cytes. If dominant lethals, to a considerable extent, are the result of chromo­somal aberrations, the frequencies obtained from fractionation, as well as fre­quencies from different stages of meiosis, should in general agree with those from studies of chromosome aberrations. It has been found, however, that translocation rates do not correspond to dominant lethals during the same periods. This has been reported by Catsch and Radu (1943), in a study of 11-111 translocations in Drosophila melanogaster. In D. virilis, Stone, Haas, Alexander, and Clayton ( 1954) reported that peaks of translocation frequencies during successive mating periods did not correspond to peaks in the frequencies of dominant lethals. In the present tests, the differences are seen most clearly in the results from test 'IV in which pupae were irradiated. Dominant lethal percentages were high throughout the mating periods, highest in the early periods which correspond to young spermatids. The pupae checked for translocations were about 10-20 hours younger at the time of irradiation so that no spermatids were exposed; only two translocations were recovered. In the test (II) on young adults, the peak of trans­locations occurred at period E; no translocations were recovered from spermato­gonial periods (F, G, H), whereas dominant lethals remained high at periods F and G, dropping rapidly at period H. Translocations A number of investigators have found that translocation frequencies vary with the stage in the meiotic cycle exposed to irradiation. Sparrow ( 1951), working with Trillium, reported break frequencies fifty times greater at the most sensitive stage in the meiotic cycle than at the least sensitive. He found that the condensed The University of Texas Publication stage of the chromosomes has a high potential breakability and that an inverse correlation exists between the length of the chromonemata and breakability. Sensitivity increases after pachytene; diplotene, diakinesis, and first metaphase are periods of high sensitivity, and susceptibility decreases steadily through first anaphase and second division, reaching a low at postmeiotic interphase. Schacht ( 1958) reported that 85 percent of the translocations recovered in Drosophila melanogaster were from cells irradiated during meiosis and early spermiogenesis. Oster (1958b), Auerbach (1953), Liining (1952b), and Alexander, Bergendahl, and Brittain (1959) have all found highest susceptibility among spermatids and greatest resistance among spermatogonia and premeiotic spermatocytes. Liining recovered translocations from irradiated spermatids in a frequency several times as great as recovered from fully mature spermatozoa. Alexander et al. recovered no translocations from premeiotic cells and found the rate lower during meiotic stages than during spermiogenesis. Glass (1955, 1956) and Auerbach (1954) also found that spermatogonia were least sensitive and that during meiosis sensi­tivity increases, reaching its peak during spermiogenesis. Oster recovered no translocations following radiation of larvae ( spermatogonia), with the highest frequency recovered from radiation of pupae ( spermatids). The present tests indi­cate the same varying sensitivity at different stages. In those tests (II and IV) in which successive two-day mating periods were checked, the translocation fre­quencies were highest at periods corresponding to spermiogenesis and lowest among spermatogonia and premeiotic spermatocytes. Translocations recovered from period E of test II, which corresponds to the spermatid stage, were six times as frequent as those recovered from irradiation of mature adults. The types of translocations and the chromosomes involved from all tests are summarized in Table 6. In all tests, a total of 121 translocations was recovered. TABLE 6 Types of translocations and chromosomes involved Test Dosage Age at irradiation Type of translocation T, Ta T :! +:? y Chrnmosomes in volved 2 3 4 I 100r Adult, 14 days II 1089r Adult, 4-6 hours A 6 1 3 2 3 3 B 32 6 14 12 13 19 c 31 3 5 19 15 12 24 D 13 6 6 6 7 E 7 2 3 6 6 III-a 250r Late larvae A IV 1089r Pupae, 220-230 hours G H v 1089r Adult, 6 days 22 4 10 13 6 11 Total 115 4 18 55 54 48 71 Clayton: X-ray Effects on Spermatogenesis Of these, 115 (95.0%) were translocations involving two chromosomes (T2), four (3.3%) were complex translocations involving three chromosomes (Ta), and one involved two separate translocations (T2+2) in one nucleus. The more complex rearrangements were recovered from periods C and E of test II. The total dosage in test II was 1,089 r at the rate of 100 r per minute and periods C and E represent those spermatogenic cells during sensitive stages, early spermiogenesis and meiotic cells. Alexander, Bergendahl, and Brittain (1959) also reported an increase in multiple breaks at period E following irradiation of D. virilis with 500, 1,000, and 2,000 rat the-rate of 202 r per minute. The results of these tests correspond closely with those reported by Haas, Dudgeon, Clayton, and Stone (1954) from irradiation of D. virilis with 2,000 r. Of a total of 2,417 translocations from all tests, 94.3 percent were of the T 2 type, 4.5 percent involved three chromosomes, and 1.2 percent involved two different translocations within one nucleus. In their tests, four translocations involving four chromosomes were recovered and one involving two different translocations, one with two chromosomes and the other with three chromosomes (T2+a) , was found. None of these latter types was recovered in the present tests. The summary of the frequency of involvement of the chromosomes in translo­cations is given in Table 6. The samples are small as compared with those of Baker (1949) and Haas et al (1954), and the results of a larger sample would probably be more consistent with their results. Baker reported the Y chromosome involved in 9.6 percent of the translocations, or about 0.4 as frequent as the autosomes. Haas et al found the Y involved in 12.7 percent of the translocations, or 0.6 as frequent as the autosomes. They found that chromosome 2, which is slightly longer than the other autosomes tested (chromosomes 3, 4, and 5), was involved more frequently, 23.0 percent as compared with about 21.0 percent for the other autosomes. Results from the present tests were not this consistent with chromo­some length. Chromosome 2, the longest of the autosomes, was involved more frequently than chromosome 4, but chromosome 3 was involved in translocations with a frequency equal to that of chromosome 2, and chromosome 5 was involved in translocations more frequently than any of the other autosomes. Transloca­tions involving the Y chromosome were 7.3 percent of the total, or about 0.4 as frequent as the autosomes. This low frequency may have resulted from failure to detect translocations in viable aneuploids or from sampling error, since the total number of recovered translocations was small. Results of a number of other investigators also indicate that break distribution following radiation is proportional to chromosome length at the time of radiation (Bauer, Demerec, Kaufmann, 1938; Bauer, 1939). Further evidence for this pro­portionality has been presented by Kaufmann and Demerec (1937), who studied salivary chromosomes following irradiation, and by Bauer (1939), who computed frequencies on the basis of length of chromosomes in mitotic cells. Evidence for the susceptibility of certain stages may be seen in test II-C, with the largest number of translocations in the period corresponding to spermiogene­sis. The frequency recovered from irradiation of young adults, when immature spermatozoa were irradiated, was higher than from irradiation of young pupae, containing only spermatogonia and growing primary spermatocytes. If reunion The Unive~sity of Texas Publication of chromosomes occurs prior to fertilization, more recovery may be expected from irradiation of pupal stages than from treatment of adults. Muller ( 1954) interpreted the lack of the square of the dose for translocations at high dosages as the result of complex rearrangement involving more than two breaks per nucleus. On the basis of his earlier work (1940), he reported that dosages from 375 r to 1,500 r result in frequencies which do correspond to the square of the dose but higher dosage levels approach a 3/2 ratio, with a leveling off of the frequencies. This was confirmed by the work of Catsch, Kanellis, and Radu (1943) and by Catsch (1948), when a comparison was made between the 1,000-2,000 r level and the 2,000-4,000 r range. The complex rearrangements in the present series of tests tend to confirm these findings. The only translocations involving more than two breaks were recovered when susceptible meiotic stages were exposed to 1,089 r. Radiation Damage to Spermatogenic Cells A limited number of studies have been carried out on the histological aspects of radiation damage to gametogenic cells. Oakberg (1955) analyzed the sensi­tivity and degeneration of germ cells in the mouse following irradiation at differ­ent stages of maturation. Bryan and Gowen ( 1958) studied mouse testes by quan­titative histological techniques and found that, following 320 r, temporary inhi­bition of spermatogonial mitoses occurred, but that after 2,560 r necrosis of cells became an important factor; a permanent depletion of spermatogonia occurred and the surviving cells were incapable of sustained regenerative efforts. W elshons and Russell ( 195 7) studied the histological condition of Drosophila melanogaster testes following 4,000 r irradiation. They reported that large pri­m<:1ry spermatocytes were not killed but that cells in the gonial region were greatly reduced in number; the gonial region was repopulated by the fourth day following irradiation; spermiogenesis stages were normal until the fourth day, decreased on the fifth day, and were absent from the sixth to the ninth days. After a series of tests using levels or irradiation as high as 9,687 r they found secondary spermatogonia and young primary spermatocytes sensitive to killing effects, resulting in a period of sterility at those mating periods corresponding t? irradiation of germ cells at these sensitive stages. In the present study, the meiotic cells were counted at intervals following irra­diation of larvae, pupae, and adults with approximately 1,000 r. At this lower level of radiation, no period was found in which spermiogenesis was absent fol­lowing radiation. There was a reduction in all types of spermatogenic cells at intervals following radiation, in the sequence that would be expected: spermato­gonia, spermatocytes, and spermatids. Destruction of spermatogonia, followed by recovery, affects the number of primary spermatocytes and spermatids present in the testes during subsequent periods. The results of the histological analyses on spermatogonia, spermatocytes, and spermatids are summarized in Figures 27, 28, and 29. No comparison can be made between late pupal stages in the tests a::id cell coun~s from the control since one effect of the radiation was to delay eclosion by approximately 48 hours. In Figure 27, the spermatogonia present after radiation are compared "'.iththe Clayton: X-ray Effects on Spermatogenesis controls at the same periods. It may be seen, from tests III and IV, that the sig­nificant decrease in spermatogonia occurs about 30 hours after irradiation; in test IV, the spermatogonial cells had reached the control level by emergence of the adults and the recovery may be seen during the first two days of adult de­velopment. In test V, the spermatogonia were at a reduced level throughout the period of the study. 600 Fig. 27 - ~ 0 I Test I:Sl' Test Il 100 180 200 220 240 260 280 300 32 6 24 48 6 9 10 Hours Days ' II ' Fig.28 I"' 2000 / /Test lll J Control Test :nr ' . ·­ 1000 ! ,//\\ //',',,, \ I 800 \I vTest Jl 100 200 220 240 260 200 300 ~o G 24 48 G s e s 10 Hours Days FrG. 27. Mean numbers of spermatogonial cells from X-ray tests as compared with the control series. Test III: larvae, 1089r; test IV: pupae, 1089r; test V: young adults, 1089r; test VI: 1000r administered as 500r to pupae and 500r to young adults. Fm. 28. Mean numbers of primary spermatocytes from X-ray tests as compared with the control series. Test Ill: larvae, 1089r; test IV: pupae; 1089r; test V: young adults, 1089r; test VI: 1OOOr administered as 500r to pupae and 500r to young adults. . The University of Texas Publica.tion 800 700 600 500 400 300 200 \ \ Fig29 A i\ \ ~'""'' I I I I I I I I I I I I 1 Test :Ill I I I I .___.._.___..___._ _.______.__ ____!._ 220 240 260 280 300 320 6 24 48 6 7 8 9 10 Hours Days Fie. Z9. Mean numbers of spermatids from X-ray tests as compared with the control series. Test III: larvae, 1089r; test IV: pupae, 1089r; test V: young adults, 1089r; test VI: 1000r administered as 500r to pupae and 500r to young adults. The primary spermatocytes from irradiated larvae, pupae, and adults are plotted in Figure 28. In test III, the spermatocytes did not increase rapidly in number as in the control at 210 hours; this reflects the destruction of spermato­gonia and probably primary spermatocytes from radiation of the larvae just before pupation. The results of test IV indicate a decrease in spermatocytes, which also may be correlated with the reduction in spermatogonia; although spermatocytes reached the control level during the first 24 hours in the adult, the significant increase found in the control during the second day was not found in the test. The results of test V indicate a reduction in spermatocytes and the two­day cycle seen in the control series at earlier stages during adult development.' In Figure 29, the effect of radiation on spermatids is presented. Development of early spermatids was delayed by about 30 hours when pupae were irradiated at the critical period when meiosis is first seen in the controls. This delay in the formation of spermatids may be one of the factors related to the cell counts during the additional two days of pupal differentiation in the test series. If differentiation in general is inhibited by irradiation, the two-day period may be an indication of the extent of radiation damage in the metabolic process of differentiation and the length of time required for recovery from the effects of the radiation. The spermatids in test IV were at a normal level in early periods following irradiation but were significantly decreased in late pupal stages. The numbers of cells under­going spermiogenesis during early adult periods resemble the figures obtained in Clayton: X-ray Effects on Spermatogenesis the controls during late pupal differentiation. Recovery to the control level is not indicated in any of the adult periods of either test IV or test V. On the basis of the results of these tests, it is evident that in future experiments cell counts should be extended beyond the periods counted in the present series. It may be that, with a continuation of cell counts for several additional days, the recovery to the control level could be seen. In correlating the cell counts with the genetic tests, the sensitivity of sperma­tids may be seen. The spermatids are present in the highest relative frequency during the first 24 hours of adult development (31.5% at 12 hours); irradiation of males during this period resulted in the largest number of translocations and in the highest percentage of dominant lethals. When pupae were irradiated be­fore meiotic stages were present, translocations were few; larvae also were re­sistant to radiation as measured by translocations. On the basis of all the present series of tests, the sensitivities of the various types of cells to radiation agree with the results found by previous investigators. Cells are most sensitive to radiation when undergoing spermiogenesis; spermatozoa and meiotic cells are more sensi­tive than spermatogonia, but less susceptible than spermatids. Spermatogonia are the most resistant to genetic damage, as measured by translocations and dominant lethals, but are inhibited in mitosis or destroyed by radiation, as evi­denced by the reduction in numbers when cell counts were made at periods folowing X-ray exposure. SUMMARY Normal spermatogenesis in Drosophila virilis pupae and adults was studied by phase-contrast microscopy, aceto-orcein smears, and by histological analysis of sections from adult males. Cytological aspects of meiosis are described and the number of spermatogenic cells in testes of adults through the first nine days of development are presented. Cell counts from adults indicated a two-day cycle in meiosis until the males are sexually mature. There appear to be steadily diminish­ing peaks in the mean numbers of primary spermatocytes in pupae just before emergence, and in adults of two, four, and six days. Examination of dissected male and female reproductive systems in adults revealed that spermatozoa of D. virilis are not motile and functional until the sixth day. Females were not inseminated and no fertile eggs were recovered prior to the sixth day. . Dominant lethals and translocations from X-ray exposure of larvae, pupae, and adults were determined for successive two-day mating periods. Results indi­cated that spermatids are highly susceptible to radiation damage; meiotic stages and mature spermatozoa are also sensitive periods, with greatest resistance to chromosome breakage among spermatogonia. Histological analysis of testes collected at intervals following irradiation of larvae, pupae, and adults revealed decreases in all types of meiotic cells followed by recovery to the control level in some tests. Radiation by 1,089 r resulted in €rmatocytes At metaphase a few bundles of mitochondria are arranged in rhombic shape, putting up two opposite apexes at the poles of the spindle, while the others are at the equatorial region of the cell body facing each other (Figs. 35,'51). Golgi ele­ments are accumulated at the poles of the spindle as a well-defi,ned group of granules. From anaphase to telophase "the mitochondrial bundles stretch directly between the poles along the spindle (Figs. 36, 37). Prior to the beginning of the formation of a cleavage furrow, the cell body elongates along the_spindle axis (Fig. 38). When the cleavage furrow appears in the cell body, the middle part of each of the mitochondrial bundles decreases in diameter (Figs. 39, 52). During these stages Golgi elements have always been observed as a granular shape sur­rounding each pole. Then the mitocho_ndrial bundles are divided into two ap­proximately equal halves, and each ofi he bundles shows a transformation into two masses holding the daughter nucleiis between them (Figs. 40, 52). Soon the mitochondrial bodies condense into round masses (Figs. 41, 54) . Afterwards, they fuse into a single body coming into contact at the base of the nucleus (Fig. 42). Eventtially, the body becomes spherical in shape and is called the "nebenkern" in the sperinatid (Fig. 55). The nebenkern is very voluminous, having a diameter of ,' ·' 'more thantwice that of the nucleus. The duration of each mitotic phase in the spermatocytes was measured in the successive series of a division process followed through the same cells beginning with prophase or metaphase and ending with telophase. The data are presented in Table 1. The time required for the completion of the cleavage furrow, from the beginning of the formation of the furrow to complete division of the cell body, was about 25 minutes in tlie primary spermatocyte, and 20 minutes in the sec­ondary spermatocyte. The division cycle in, ;the . secondary spennatocytes was ·1completed in approximately 170 minutes. The length of time for the completion of the nebenkern in the new spermatid was about 50 minutes. D1scuss10N Most of the descriptions concerning the segregation of the mitochondria to the daughter cells were made by the earlier cytologists. There are remarkable differ­ences among species in regard to the precision and orderliness of the distribution TABLE 1 Time relations in the division of the spermat6cytes of D. virilis, at Z0° C. Metaphase Anaphase Telophase 1st meiotic division "· 55 min. 30min. 45min. Znd meiotic division 35min. 2,5 min. 30min. The University of Texas Publication of the mitochondria. Wilson ( 1953) has given the opinion that in general the chondriosomes, in which mitochondria are included in general cytology, are passively divided into two nearly equal groups as a mechanical result of the cell­constriction. In a scorpion, Centrurus, however, all the chondriosome-material Momma: Behavior of Mitochondria in the primary spermatocytes aggregates into a single ring-shaped body, and is cut by the division accurately into two half-rings. The activity of the chondrio­some-material has also been investigated in the spermatocytes of the scorpion, Butkus, described in detail by Sato ( 1940) . According to him the chondriosome­ring is composed of a definite number of coils of the "chondrionema," and they are accurately halved into exact parts containing an equal number of turns of the chondrionemal spiral. In Drosophili:z, as is seen in most insects, the mitochondria in the spermatocytes at early telophase have the form of numerous rods or threads between the two poles, parallel with the elongated spindle as shown in D. melanogaster (Cooper 1950), D. pseudoobscura (Dobzhansky 1934) and D. virilis (Metz 1926). Naka­hara (1952), studying living grasshopper spermatocytes, suggested that such an elongation of the mitochondrial elements is due to end-to-end connection of the two or more elements in a linear series. Tahmisian et al (1956) with the aid of an electron microscope have observed that mitochondria in a grasshopper, Melano­plus, are divided rather than distributed at random to a daughter cell. Such phe­nomena may occur with some of the mitochondria in D. virilis. However, certain appearances at mid-telophase in spermatocytes of D. lacertosa (Momma, unpub­lished) indicate that the mitochondrial threads may be formed by fusion or close union of several smaller elements, even though they are conspicuously shortened by fixatives. Transverse sections of these threads at this time show their shapes EXPLANATION OF FIGURES The microphotographs were taken with Mifilmca (Leitz) at a magnification of approximately X94-0 (Figs. 1-28, 35-42), X830 (Figs. 29-34) and X1740 (Figs. 43-55). Figures from 1 to 42 are from living material and from 43 to 55 are from sections. . FIGs. 1-34. Primary spermatocytes. Figs. 1, 2. Late prophase showing the development of asters. Figs. 3, 4. Diakinesis showing the formation of the spindle body, mitochondrial bundles and their orientation, and diminution of nucleoli, Fig. 5. Metaphase. Fig. 6. Latest prophase showing nucleolus. Figs. 7-9. Early metaphase showing diminution of nucleolus and orientation of mitochondrial bundles. Figs. 10, 11. Metaphase showing chromosomes. Fig. tz. Late metaphase showing chromosomes arranging at equatorial plate. Fig. 13. Telophase showing tension of mitochondrial bundles and formation of cleavage furrow. Figs. 14--16. Telophase showing the formation of daughter cells. Figures from 17 to 28 show successive changes of a single cell from diakinesis of primary spermatocyte to the formation of secondary spermatocyte; material was mounted in .mineral oil. Figs. 17, 18. Diakinesis showing five large and one small dot-like bivalents (Fig. 17), and the formation of spindle body (Fig. 18). Nucleolus is still observed as a white granule. Fig. 19. Early metaphase. Figs. 20-22. Metaphase showing the stretching of mitochondri11l bundles. Figs. 23-27. Telophase showing the formation of cleavage furrow. Fig. 28. Newly formi!d secondary spermatocytes. Figures from .29 to 34 show successive changes of a single giant cell including tri-spindles during the telophase, and also showing the formation of three binucleate secondary spermatocytes. FIGs. 35-42. Secondary spermatocytes. Fig. 35. Metaphase. Fig. 36. Anaphase. Figs. 37-42. Telophase showing tension of mitochondrial bundles (Fig. 37), cell elongation, segregation of mitochondrial elements (Figs. 38-40), formation of cleavage furrow, and of nebenkern (Figs. 39-42). . FIGS. 43-49. Primary spermatocytes. Fig. 43. Metaphase. Figs. 44, 45. Anaphase. Figs. 46-49. Telophase showing segregation of mitochondrial elements. FIGS. 50-54. Secondary spermatocytes. Fig. 50. Newly formed daughter cells. Fig. 51. Meta­phase. Figs. 52--54. Telophase showing segregation of mitochondrial ·elements. FIG. 55. Early spermatid showing well-stained large spherical nebenkern. The University of Texas Publication as cylinders or splitting ones. This evidence makes firm the interpretation about the fusion of smaller mitochondria, and may signify the division of the threads at an earlier stage as described by Wilson (1953). Indeed, most mitochondrial com­ponents at late telophase seem to actually separate into two essentially equivalent groups without regard to the process of the cleavage furrow. This active behavior of the mitochondria was also observed in living material, especially secondary spermatocytes, of this species. From the foregoing it is probable that some mito­chondria are passively and mechanically divided as a result of the cell constric­tion, but on the whole, at least in the case of the second meiotic division, the pres­ent evidence may point to the conclusion that the division of the mitochondrial elements is active. The difficulty in determining the duration of mitosis lies in the accurate de­tection of the onset of the process. In general the limits are decided by the study of nuclear cycles. Nuclei and spindles observed in Drosophila spermatocytes are so small in size, however, that the study of the nuclear cycles is very difficult. But it is possible to say from the present study that the time required for the com~ pletion of one mitotic cycle in the spermatocytes of D. virilis is extremely long as compared with that in the early Drosophila egg reported by Rabinowitz (1941)~ SUMMARY Behavior of cell-components, especially mitochondria, during meiosis of Dro­sophila virilis have been investigated by the use of living material. The sperma­tocytes are very large in size. There exist in the cytoplasm visibly clear mito­chondrial elements of a characteristic shape together with other cytoplasmic in­clusions. ._., The general aspect of the mitochondria during the meiotic division seems to indicate that most of them actually separate into two essentially equivalent groups without regard to the process of the cleavage furrow. The time required for division of the spermatocytes was observed. The data are shown .in Table 1. · LiTERATURE CITED Clayton, F. E. 1957. Absolute and relative frequencies of spermatogenic stages at different pupal periods in Drosophila virilis. J. Morph. 101: 457-476. . .. · Cooper, K. W. 1950; Normal sperwatogenesis in Drosophila. Biology of Drosophila edited by Demerec. John Wiley and Sons1 NewYork, pp. 1~61. ., Dobzhansky, Th. 1934. Studies 'o~ hybrid sterility. I. Spermatogenesis iri pJlre and hybrid Drosbphila pseudoobscura: Z. Zellforsch. u. Mikroskop. Anat. 21: 169-223. · Guyenot, E., and A. Naville. 1929. Les, ~hr~~olomes et la .reduction chromatique chez Vr~­sophila melanogaster. (Cinese somatiques, spermatogenese, ovogenese). ·cellule 39:' 2fi~2. Heuttner, A. F. 1930. The spermatogeriesis of Drosophila melanogaster. Z. Zellforsch.J u. Mikroskop. Anat. 11 : 615--637. ,., Metz, C. W. 1926. Observations on spermatogenesis in Drosophila. Z. Zellforsch. u. MikrC>Skop. Anat. 4: 1-28. Nakahara, H . 1952. Beha~ior of the mitochondria in cell division, with evidence concerning the kinetic function. Cytol0gia 17: i68-178. , · Rabinowitz, M . 1941. Studies on the cytology and early embryology. of the egg of Drdsophila melanogaster. four. Morph. 69: 1-:49. Momma: Behavior of Mitochondria Sato, I. 1940. Studies on the cytoplasmic phenomena in the spermatogenesis of the oriental scorpion, Buthus martensii, with special reference to the structure of the chondriosome ring and the dictyokinesis. Jour. Sci. Hiroshima Univ. Ser. B, Div. 1-8: 1-116. Tahmisian, T. N., E. L. Powers, and R. L. Devine. 1956. Light and electron microscope studies of morphological changes of mitochondria during spermatogenesis in the grasshopper. J. Biophysic. and Biochem. Cytol. 2 (Suppl): 325-330. Wilson, E. V. 1953. The cell in development and heredity. 3rd ed. Macmillan Co., New York. XIII. Further Observations on the Relation Between Gas Pressure and the X-Ray Damage in Drosophila melanogaster TSUENG·H~ING CHANG1 INTRODUCTION An investigation of the oxygen effect seemed open to a different type of ap­proach after Ebert and Howard (1957) and Ebert, Homsey, and Howard (1958) reported that nitrogen, hydrogen and inert gases, under high pressure, were able to counteract the oxygen effect in the broad bean (Vicia faba) root tip and in tumor cells. Later it was shown (Chang, Wilson, and Stone, 1959) that nitrogen, argon, and methane similarly counteract the oxygen effect in Drosophila. When nine atmospheres of either argon, nitrogen or methane were added to one atmosphere of oxygen, X-ray damage in D. melanogaster, as measured by the frequencies of sex-linked recessive lethals, was about the same as that caused by the same amount of radiation in 10 atmospheres of pure argon or nitrogen. Sim­ilar results were also observed in D. virilis using dominant lethals as a measure of genetic damage. Some of the experiments reported in the previous paper (Chang et al., 1959) have been repeated and some additional experiments have been performed. The results which are to be reported here include, in part, those reported in the previ­ous paper. MATERIAL AND METHODS D. melanogaster (Oregon-R) males were used in this study. The stock was carried by mass inbreeding in one-half pint milk bottles using cornmeal or ba­nana food, at a temperature of 20-21 °C. Males were collected over several hours and were separated from females immediately after the end of the collecting period. The age of these males at the time of irradiation was approximately 24 hours, and in all but a few experiments males differed in age by no more than five hours at irradiation. · Unless noted otherwise, the dose of X-rays administered to males was uni­formly 1,000 r. Irradiation of males was carried out at 20-21°C. with a Westing­house quadrocondex X-ray machine operated at 200 kvp (in previous experiments 250 kvp) , 15 ma, and with a filter of Yz mm Cu +1 mm Al. When males were X-rayed in various gas mixtures (with or without pressuri­zation) they were held in the mixtures before and after irradiation for 15 min­utes. Less than two minutes was taken in bringing up the pressure to the desired level or in releasing the pressure after the completion of post-irradiation treat­ment in a gas mixture. Immediately after X-ray treatment, .the males were mated individually to three 2-4-day-old virgin Muller-5 females for detection of sex-linked recessive lethals. At two-day intervals thereafter, males were remated to another set of virgins for six times. This gave seven (A-G) mating periods. Females from each 1 Department of Zoology, The University of Texas, Austin, and Department of Biology, The University of Texas, M. D. Anderson Hospital and Tumor Institute, Houston. The University of Texas Publication period were allowed to lay eggs for four days. For details of the Muller-5 tech­nique see Spencer and Stern (1948). The F1 females were mated, again individ­ually, to their brothers. In the F2 cultures if there were two or more wild-type males it was scored as a non-lethal X-chromosome. Those with one or no wild­type male were reported here as a lethal X-chromosome. Cultures having a lethal-X were checked through F3 • For dominant lethals the procedures were the same except that treated males were mated to three 2-4-day-old Oregon-R virgin females. Eggs laid by the fe­males were counted daily for three days. The percentage of dominant lethals was determined by the number showing eclosion vs. the number of eggs laid. Since none of the gases used here react with each other, gas mixtures were ob­tained simply by regulating the partial pressures of the gases. For example, 10% of oxygen in argon at the total pressure of 10 atmospheres was obtained by intro­ducing one atmosphere of oxygen and nine atmospheres of argon into the pressure chamber, after the air originally present in the chamber was flushed out with argon. The mixtures for series 11 and 26 were obtained in a similar way except that the pressure was lowered to one atmosphere at the time of pre-irradiation treatment. Each period of mating represents a certain stage of cell development in sperma­togenesis at the time of X-irradiation. As has been pointed out (Chang et al., 1959) each mating period in this investigation has been correlated to the stages of spermatogenesis as follows: A, sperm; B, sperm bundle; C and D, late and eai-ly ;;permatids respectively; E, meiosis; F, spermatocytes; G, spermatogonia. RESULTS The results of the experiments are summarized in the tables. The 95% confi­dence intervals for dominant lethals were calculated with the approximate and conventional method of ±1.96 SE. However, that for recessive lethals was calcu­lated in accordance with Stevens' recommendation (1942). Recessive lethals a) The effect of increased air pressure. One atmosphere of air (series 5 and 16) and 1,000 r of X-rays produced the highest number of recessive lethals in period D, the next highest in C and E. The frequencies in A and B were about the same. A similar pattern of genetic damage variation was seen when males were X­rayed in 10 atmospheres of air. Two experiments each were done with both one and 10 atmospheres of air. In each case the second experiment was repeated more than one year after the first one. The repeated ones were in good agreement with the earlier ones; as a matter of fact, all of the repeated experiments reported in this paper show recessive lethal frequencies similar to those of comparable experi­ments. This fact illustrates how reproducible this type of experiment is. To em­phasize this point, some data from the previous publication (Chang et al., 1959) are included here. All comparable experiments were pooled together in the tables. This is justified for the reason given above. The data also show that irradiation in 10 atmospheres of air produced con­siderably more recessive lethals than in one atmosphere (of air). This is defi­ Chang: Gas Pressure and Mutations nitely so in period D (early spermatids) which is the most radiosensitive stage of spermatogenesis. The definite increase of lethals caused by irradiating in 10 at­mospheres of air becomes more apparent when all the figures (period A through G) are added together. (Admittedly this is not a good way, but it does show the obvious effect of increased air pressure.) When flies were irradiated in.10 atmos­pheres of air, 5.5% recessive lethals was produced as compared to 2.8% in the series which was irradiated in one atmosphere of air (see series 2 and 8 and 5 and 16, Table 1). Approximately twice as many lethals were produced in period D in the former as in the latter. b) The effect of increased oxygen pressure. It was shown that in D. virilis as many dominant lethals were produced by X-rays in one atmosphere of oxygen as in 10 atmospheres of oxygen. An increase in oxygen pressure does not increase the X-ray damage. The same is true of sex-linked recessive lethals (compare series 7 and 10 and 17 in Table 2). It was previously assumed (Chang, 1960) that X-rays would produce more lethals in 10 atmospheres of oxygen than in one atmosphere (of oxygen). This proved to be untenable. · c) Various combinations of argon and oxygen. X-raying in eight atmospheres of argon plus two atmospheres of oxygen showed essentially the same results as those obtained in 10 atmospheres of air (Table 1) . This is understandable since the partial pressure of oxygen in both treatments was equal (i. e., two atmos­pheres). However, in every mating period the frequencies of lethals obtained in the former were smaller than in the latter, although not significantly different from the values obtained for the "10-atm.-air" series. When the irradiation of flies was carried out in a mixture of nine atmospheres of argon plus one atmosphere of oxygen, the frequencies of sex-linked recessive lethals was greatly reduced. In fact, their values were close to those for the pure gas series (i.e., 10 atmospheres of argon or of nitrogen). The radiation damage in this gas mixture was also less than that observed in the series treated in 90% nitrogen and 10% oxygen at one atmosphere of total pressure (series 11, Table 1). There was ten times as much oxygen in the former mixture as in the latter~ although the ratios of the two gases were the same, namely 9: 1. d) The effect of increasing inert gas pressure. X-ray damage in one atmosphere of argon or nitrogen was not different from that induced in 10 atmospheres of either argori or nitrogen (Table 2). Therefore, it can be concluded that a pressure per se of either argon or nitrogen as high as 10 atmospheres has no effect on the induction of mutations. Dominant lethals a) Treatment of males in gases under pressure only. When males were treated in one atmosphere of oxygen or in 10 atmospheres of argon without radiation, no· effect on the percentage of hatch was observed (series 19 and 20, Table 3). In both series, the percentage of hatch in every mating period was the same as that in the control series (series 18). Since the treatment in 10 atmospheres of argon does not have any effect it is very likely that a similar treatment in one atmos­phere of argon would not affect the hatchability of fertilized eggs. No data for a similar treatment of flies in 10 atmospheres of oxygen were available. However TABLE 1 No. of recessive lethals Sex-linked recessive lethals induced by 1,000 r of X-rays in various gas mixtures. a, ; b, per cent of lethals; No. of X-chromosomes checked w c, 95% confidence limits. CXJ CXJ STAGES OF SPERMATOGENESIS (See Text) A B c D E F G Series Gas mixlure b b b b b c b b 1 atm 4 0.5 2 0.1 13 2.7 8 1.4 9 2.0 1 .03 1 .02 5• , air 237 1.7 4.3 . 273 I 0.7 2.6 262 5.0 8.3 247 3.2 6.3 211 4.3 7.9 911.1 6.0 145 0.7 3.9 1 atm 9 0.9 6 0.6 9 1.2 24 4.6 19 3.4 8 0.9 2 0.1 16 5 & 16 2* 8 air 1 atm air 10atm air 10 atm air 446 13 683 9 221 11 419 2.0 1.9 4.1 2.6 3.8 1.0 3.2 1.9 7.6 1.3 4.6 403 8 676 9 264 18 445 1.5 1.2 3.4 4.0 3.2 0.5 2.3 1.6 6.4 2.4 6.3 361 22 623 14 221 19 327 2.5 4.7 2.2 3.5 5.3 3.5 6.3 10.4 3.5 5.8 8.9 343 7.0 10.2 32 3.8 590 5.4 5.9 32 11.3 200 16.0 21 .9 90 7.6 961 9.4' 11.4 346 5.5 8.4 28 3.4 557 5.0 7.2 13 5.4 130 10.0 16.5 25 7.2 229 10.9 15.8 376 2.1 4.2 9 0.9 467 1.9 3.7 6 0.9 242 2.5 5.3 10 179 5.6 10.0 331 3 476 1 235 13 875 0.6 0.6 0.4 1.5 2.2 0.1 1.8 .01 2.4 0.8 2.5 '""-3 ~ (':) ~ ;::s;:: · 3-· ~ ~ 2 & 8 10 atm air 8 atmA + 20 640 8 3.1 1.9 4.8 1.2 27 709 8 3.8 2.5 5.5 1.1 33 548 19 6.0 4.2 8.4 4.0 122 1161 25 8.8 10.5 12.4 7.8 38 7.6 359 10.6 14.2 10 4.1 16 421 0 3.8 2.2 6.1 0.0 14 1110 4 1.3 0.7 2.1 0.5 ~ !--! Cl 3* 13 3 & 13 2atm02 8 atmA + 2atm02 8atmA + 2atm 0 2 298 12 501 20 799 2. 7 2.4 2.5. 5.2 1.2 4.1 1.5 3.8 303 4 477 12 780 2.6 0.8 1.5 5. 1 0.2 2. 1 0.8 2.7 288 14 279 33 567 6.6 5.0 5.8 10.1 2.8 8.3 4.0 8.1 214 11.7 16.8 72 7.1 802 9.0 11 .2 97 7.8 1016 9.5 11 .5 120 8.3 14.8 10 2.8 175 5. 7 10.3 20 4.2 295 6.8 10.3 110.0 33.5 2 0.1 273 0.7 2.6 2 0.1 284 0.7 2.5 2 14 4 383 8 597 1.9 1.0 1.3 4.8 0.3 2.7 0.6 2.6 ~ <::l'"' ....... ;:::; · >:) g. ;::s 1 • 9atmA + 1 atm02 2 214 0.9 0.1 3.4 151 0.7 .02 3.7 4 167 2.4 0.7 G.O 8 207 3.9 1.7 7.5 5 144 3.5 1.1 7.9 3 178 1.7 0.3 4.8 3 165 1.8 0.4 5.2 9atmA + 15 0.9 25 1.6 21 2.5 6 2.1 5 1.1 3 0.7 4 0.4 12 1atm02 905 1.7 2.5 1010 2.5 3.6 517 4.1 G.1 107 5.6 11.8 154 3.2 7.4 88 3.4· 9.7 249 1.6 4.1 1 & 12 9 atmA+ 1atm02 17 1119 1.5 0.9 2.4 26 1161 2.2 1.5 3.3 25 684 3.7 2.4 5.4 14 314 4.5 2.5 7.4 10 298 3.4 1.6 6.1 6 266 2.3 0.8 4.8 7 414 1.7 0.7 3.5 1 atm 11 (90% N 2 + 10% 0 0 ) •From Chang el ; l .. 1959. 9 426 2.1 1.0 4.0 9 581 1.5 0.7 2.9 12 292 4.1 2. 1 7.1 16 206 7.8 4.5 11.8 21 228 9.2 5.8 13.7 8 298 2.7 1.2 5.2 6 555 1.1 0.5 2.3 TABLE 2 No. of recessive lethals Sex-linked recessive lethals induced by 1,000 r of X-rays in pure gases. a, ; b, per cent of lethals; c, 95% confidence limits. No. of X-chromosomes checked STAGES OF SPEHMATOGENESIS (See Text) A B c D E F G Series Gas b b b b b b b 7' 10 atm 0 .. 10 atm 10 211 9 4.2 2.0 7.9 1.7 15 250 15 6.0 3.4 9.7 3.9 23 266 10 5.6 8.7 12.7 3.2 24 7.7 204 11.8 17.0 16 6.4 7 76 17 3.8 9.2 17.9 7.2 2 52. 14 3.9 0.5 13.4 5.2 5 186 2 2.7 0.1 6.1 0.1 C":l s­ 10 0 ., 250 3.6 6.7 213 7.0 11.4 150 6.7 11.9 147 10.9 17.1 141 12.1 18.6 150 9.3 15.1 247 0.8 2.9 ~ 7& 10 10atm 0 ., 19 491 3.9 2.4 5.9 30 453 6.5 2.4 9.3 33 416 7.9 5.5 10.9 40 8.3 351 11.4 15.2 24 217 11.1 7.3 16.0 16 202 4.6 7.9 12.5 7 433 1.6 0.5 3.3 ~ "'Cl 17 1 atm 0 , 16 391 4.1 2.4 6.6 11 382 2.9 1.4 5.2 21 358 5.9 3.7 8.8 42 9.5 327 12 8 17.1 30 334 9.0 6.2 12.6 4 365 1.1 0.3 3.2 3 358 0.8 0.2 2.5 ~ ~ 4·* 10 atm A 2 161 1.2 0.2 4.4 149 0.7 .02 3.7 5 204 2.5 0.8 5.6 3 171 1.8 0.4 5.0 3 161 1.9 0.4 54 82 1.2 .03 6.6 5 142 3.5 1.2 8.0 ~ § \:I.. 9 4& 9 10atm A 10atm A 15 714 17 875 2.1 1.9 1.2 3.4 1.1 3. 1 13 700 14 849 1.9 1.6 1.0 3.2 0.9 2.8 8 422 13 626 1.9 2.1 0.8 3.7 1.1 3.5 9 388 12 559 2.3 2.1 1.1 4.4 1.1 3.7 7 351 10 512 2.0 2.0 0.8 4.1 0.9 3.6 4 361 5 443 1.1 1.1 0.3 2.8 0.4 2.6 4 361 9 503 1.1 1.8 0.3 2.8 0.8 3.4 ~ !:: ..... ~ g· "' 6* 10 atm N 2 6 256 2.3 0.9 5.0 6 299 2.0 0.7 4.4 5 261 1.9 0.6 4.4 10 288 3.5 1.7 6.3 6 244 2.5 0.9 5.3 5 226 2.2 0.7 5.1 217 0.5 .01 2.6 14 1 atm A 5 284 1.8 0.6 4.1 6 283 2.1 0.8 4.6 10 280 3.6 1.7 6.5 5 318 1.6 0.5 3.6 9 294 3.1 1.4 5.7 6 325 1.8 0.7 4.0 0 257 0.0 0.0 1.4 15 1 atm N z •From Chang e l al., 1959. 5 249 2.0 0.7 4.6 5 302 1.7 0.5 3.8 12 297 4.0 2.1 7.0 7 293 2.4 1.0 4.9 9 288 3.1 1.4 5.8 6 291 2.1 0.8 4.4 3 274 1.1 0.2 3.2 w 00 <.O ().) ID 0 TABLE 3 Dominant lethals caused by 1,000 r of X-rays in various gas mixtures. a, No. of hatch . No. of eggs ; b, per cent of hatch±2. SE STAGES OF SPERMATOGENESIS (See T ext) A B c D E F G Series Gas mixture a b a b a b a b a b a b a b 1 atm air• 605 515 11 56 1661 2.731 353 42.7 ~ ~ 18 (con trol) 673 89.9± 2..3 549 93.8±2..0 1306 88.5±1.7 1790 92..8± 1.2. 2.960 92..3± 1.0 379 93.1 ± 2..6 484 88.2.± 2..9 (I) 994 2257 579 750 374 82. 116 t:!;::s..... 22 1 atm air 1379 72.1±2.4 2.960 76.3 ± 1.5 1060 54.6± 3.0 2091 35 .9± 2.0 1198 31.2±2.6 240 34 2. ± 5.9 145 80.0±6.5 ~ 2276 688 393 35 4 205 307 1613 ..... 21 10 atm air 3924 58.0± 1.5 1161 59.3 ± 2.8 943 41.7± 3.1 1356 26.1±2.3 i164 17.6± 2.2 2019 15.2±1.6 1935 83.4±1.7 ~ 2997 617 439 212 227 150 2.414 -a 23 15 atm air 4867 61.6± 1.4 1061 58.2.±2.9 1332 32.9± 2.5 1060 20.0± 2.4 1390 16.3± 1.9 132.8 11.3± 1.7 3392. 71.2± 1.1 ~ ~ 911 1596 621 879 1575 61 3 11 38 24 10 atm A 11 38 80.1 ±2..3 2069 77.1±1.8 886 70.1±3.1 1200 73.3 ± 2.5 2510 62.8±0.6 1051 58.3 ± 3.0 1301 87.5± 1.8 ~ ~ 3041 1647 1235 3543 882 1485 512 ......(:;• 19 10 atm A• 3252 93.5±0.8 1738 94.8± 1.0 1314 94.0±1.3 3779 93.8±0.8 958 92.1±1.7 1562 95.1±1.1 555 92..3 ±2.2 !:).....5· 32.45 1650 3280 302.5 1450 1269 965 ;:s 20 1atm02 • 3446 94.2.±0.8 1767 93.4·± 1.2 3557 92.2.±0.8 322.2 93.6±0.8 152.2 95 .3 ± 1.0 1334 95. 1±1 .2 1051 91.8± 1.7 9atmA + 1347 91 5 943 816 1001 2202 1530 25 1atm02 2300 58.6 1358 67.4 2091 45.1 2.3 11 35.3 1951 51.3 2910 75.7 1857 82.4 1 atm (90 % A+ 1012. 1022 - 777 1083 2809 359 26 10 % 0 2) 1710 59.2 1448 70.6 2317 33 .6 3270 33. 1 3433 81.8 445 80.7 • No X·inadiation. Chang: Gas Pressure and Mutations a batch of males was treated in 10 atmospheres of oxygen for 35 minutes; no male survived the treatment. b) The effect of increased air pressure. The data show that an increase in air pressure increases the frequencies of dominant lethals caused by X-rays (Table 3). Except for period A the results show decreased hatch with increase in air pressure, although the differences in hatch for periods B and E between treat­ments in 10 and 15 atmospheres of air are not significant. These results corrobo­rate those obtained for recessive lethals. c) Counteracting effect of argon against oxygen. The percentages of hatch for series 25 and 26 are rather low, due to the fact that when the hatch was checked there were too many second generation larvae in the vials. This was responsible for the low percentages of hatch in these two series. Since the figures were biased to some degree the standard error was not calculated. However, a clear pattern is still recognizable, namely that the extent of dominant lethal damage in series 25 was at the same level as in series 26. The same result was obtained for recessive lethals. D1scuss10N The data presented here, although not very extensive, show several points. First, an increase in air pressure increases the X-ray damage, whether the dam­age be measured as sex-linked recessive lethals or as dominant lethals. On the contrary, an increase in pressure of oxygen does not increase the frequencies of X-ray induced recessive lethals since X-irradiation in one and in 10 atmospheres of oxygen gave the same degree of damage. Similar experiments were not done for dominant lethals but this was shown to be so in D. virilis (Chang et al., 1959). The explanation for the above facts is not clear. It is possible that, since cells are adapted to functioning in air (20% oxygen), flies may be most sensitive to change in air (thus sensitive to change in oxygen partial pressure). Or, there is a point beyond which increase in oxygen partial pressure does not affect its effect. Secondly, nine atmospheres of argon can counteract the effect of one atmos­phere of oxygen present in the gas mixture. Ebert et al. (1958) proposed that this is a phenome,non of deplacement of oxygen from the lipid phase by argon, making the biological structure less susceptible to radiation damage. Argon and nitrogen have greater fat/ water partition coefficients. This explanation appears plausible since methane, also with a high fat/water partition coefficient, possesses the same capacity as argon and nitrogen and other inert gases. However, Hung, Gray, and Boag (1960) reported that an argon/ oxygen pressure ratio of over 500 was not enough to restore the electron spin resonance signal (which is suppressed by oxygen) from activated charcoal. Thus in the system they used, argon was not able to deplace oxygen from the sites. These authors concluded that "if compe­tition is indeed the mechanism underlying the radiobiological action of the inert gases, then the oxygen must be more loosely bound on the sites in the cell than it is on charcoal." Thirdly, as has been shown repeatedly (cf. Stone, 1955), cells in different stages of spermatogenesis respond differently to X-rays. The data show that, for recessive lethals, early spermatids (period D) are the most sensitive; late sperma­tids and meiotic cells, periods C and E respectively, are the next most sensitive. The Universityof Texas Publication For dominant lethals, spermatocytes (period E) are most heavily damaged. How­ever, it is important to point out that when the males were t~eated in the absence of oxygen the differential response to X-rays is very much less obvious~ This is so both for recessive and dominant lethals. As a matter of fact the recessive lethals produced by X-raying in 10 atmospheres of argon were about the same in each mating period. This observation indicates that the higher X-ray sensitivity of spermatids normally observed is due mostly, if not eni:irely, to the presence of oxygen. In this connection, it is interesting to note that Oster (1957) reported greater reduction of translocation frequencies from "in-air" to "in-nitrogen" treatment for spermatids than for mature sperm in D. melanogaster. On the other hand, the increase in translocations from "in-air" to "in-oxygen" treatments was greater for sperm than for spermatids. Oster suggested that the oxygen tension is higher in spermatids than in mature sperm, and that this fact is at least one of the reasons for higher radiosensitivity of the former. Gray, Chase, Deschner, Hunt, and Scott ( 1958) estimated that there is ten times as much oxygen in spermatids as there is in sperm. Fourthly, one of the reasons for including data from Chang et al. (1959) is to show that these experiments can be repeated with a high degree of reproduci­bility. It has been pointed out that all the repeated experiments were in good agreement with the ones done previously. Finally, one can ask: Do not all these gases (or gas mixtures) cause lethals by themselves (without X-rays)? This can be answered by examining the data pre­sented above. They show that an equal amount of recessive lethals is caused by radiating in one and in 10 atmospheres of oxygen. Hence it can be concluded that treatment of flies in a pure oxygen atmosphere of up to 10 atmospheres of pres­sure without radiation, for the duration of 40 minutes (from pre-irradiation treat­ment to the end of post-irradiation treatment), does not induce recessive lethals. For dominant lethals we know, at least, that one atmosphere of oxygen does not affect the hatchability. In a series of tests (Chang, unpublished), spermatids (the most sensitive stage) were treated in an atmosphere consisting of nine atmos­pheres of argon + one atmosphere of oxygen for 40 minutes. Seven hundred and eighteen X-chromosomes were tested, but no recessive lethal was recovered. This is reasonable assurance that treatment in the above gas mixture itself -does-not increase mutation rate. Since treatment of gametes in 10 atmospheres of oxygen without X-ray for 40 minutes does not increase the rate of X-ray induced muta­tions, the writer is inclined to the view that none of the gases and gas mixtures used in these experiments has mutagenic action by itself. SUMMARY 1. Increase in air pressure increases the X-ray damage measured as sex-linked and dominant lethals. On the contrary, an increase in pressure of pure oxygen does not increase X-ray damage. The same is true for argon and nitrogen. 2. Whereas eight atmospheres of argon were not able to counteract the effect of two atmospheres of oxygen, nine atmospheres of argon definitely reduces the effect of one atmosphere of oxygen, almost to the anoxic level. 3. The typical differential response of spermat()genic cells to X-rays was ob­ Chang: Gas Pressure and Mutations . served: early spermatids are most sensitive to recessive lethal induction, sperma­tocytes are most sensitive to dominant lethal induction. This differential response is much less obvious when the flies were treated in anoxia. Hence the high radio­sensitivity of spermatids (and spermatocytes) is attributable, partially at least, to the presence of oxygen. 4. Treatments of flies, without radiation, in the gases and mixtures used do not have an effect. ACKNOWLEDGMENTS I wish to thank Professor Wilson S. Stone for his suggestion of undertaking this problem, and his guidance throughout the course of the investigation. This work was supported by grants from The National Institutes of Health (RG-6492) and the Atomic Energy Commission (AT-40-1-1323) . A portion of this work was performed after the writer was appointed as a post­doctoral research fellow at M. D. Anderson Hospital, Houston, paid from a train­ing grant, U. S. Public Health Service CRT-5047. LITERATURE CITED Chang, Tsueng-Hsing. 1960. Reduction of the sex-linked recessive lethal frequencies in Drcr sophila melatWgaster by a·rgon, nitrogen, and methane under pressure. Dissertation, The University of Texas. ----, F. D. Wilson, and W. S. Stone. 1959. Genetic radiation damage reversal by nitro­gen, methane, and argon. Proc. Natl. Acad. Sci. 45: 1397-14-04. Ebert, M., S. Homsey, and A. Howard. 1958. Effect of inert gases on oxygen dependent radio­sensitivity. Nature, 181: 613-616. ----, and A. Howard. 1957. Effect of nitrogen and hydrogen gas under pressure on the radiosensitivity of the broad bean root. Rad. Res. 7: 331-341. Gray, L. H., H. B. Chase, E. E. Deschner, J. W. Hunt, and 0. C. A. Scott. 1958. The influence of oxygen and peroxides on the response of mammalian cells and tissues. Proc. 2nd U.N. lntematl. Conf. Peaceful Use of Atomic Energy 22: 413-419. Hunt, J. W., L. H. Gray, and J. W . Boag. 1960. The effect of paramagnetic gases on the electron spin resonance signal from activated charcoal. Rad. Res. 12: 319-324. Oster, I. I. 1957. Suggested mechanism underlying the differential radiosensitivity of cells having condensed chromosomes. (Abstract) Rec. Genet. Soc. America, p. 387. Genetics 42: 387. Spencer, W. P., and C. Stem. 1948. Experiments to test the validity of the linear r-dose/ muta­ tion frequency relation in Drosophila at low dosage. Genetics 33: 43-74. Stevens, W. L. 1942. Accuracy of mutation rates. Jour. Genet. 43 : 301-307. Stone, W . S. 1955. Indirect effects of radiation on genetic material. Brookhaven Symp. Biol., No. 8 (Mutation): 171-190. XIV. The Flavopilosa Species Group of Drosophila 1 2 MARSHALL R. WHEELER, HARUO TAKADA, AND DANKO BRNCIC We are establishing this new species group of Drosophiw, subgenus Drosophila, for Drosophila fiavopilosa Frey and thirteen new species (including six unnamed) from the Neotropical region. With few exceptions the species are entirely or mostly all dull yellow; they are of small to medium size, have a rather high costal index, a single strong oral bristle, an arista formula of 3/2 (the number of dorsal and ventral branches, excluding the terminal fork, expressed as a fraction), and six acrostichal rows. Females have unusually strongly spined ovipositors, and most of them have apical caps on the spermathecae. The male genitalia are of characteristic structure (see figures) : lower portion of genital arch usually with two long bristles; "toe" strongly bent forward, usually elongate and narrow, not covering clasper; anal plate oblong and fused with genital arch; primary clasper broad, its under margin basally convex. Penis slender and long, curved ventrad and with a pair of apical lobes; hypandrium simple; anterior gonapophyses usu­ally lacking or fused with hypandrium; posterior gonapophyses apparently ab­sent; phallosomal index more than 4.0 (ratio of penis length and length of its apodeme; see Okada, 1953, Zool. Mag. [Japan] 62: 278-293). D. fiavopilosa is known to oviposit in flowers of the solanaceous plant, Cestrum parqui, in Chile; a similar habit for the other species of the group is suspected but is as yet not proved. D. neochracea Wheeler (1959, Univ. Texas Publ. ·5914: 183), a replacement name for ochracea Duda 1927, not Grimshaw 1901, is superficially much like members of this group, but it has a low costal index, convergent basal scutellars, and eight acrostichal rows. Males have recurved hairs along the inner side of the fore tarsi, and the ovipositor of the females is thin and pale, with normal den­tition. HISTORICAL PERSPECTIVE Dr. 0. Duda described Drosophila dentata in 1927 (Arch. f. Naturg. 91 A12 [1925]: 201-202) from eleven females from La Paz, Bolivia, one female from Los Andes, Chile, and one female from Cuzco, Peru. Through the courtesy of Dr. Rolf Hertel of the Dresden Museum, we were able to borrow one of the "cotype" females from La Paz. In 1924, however, Dr. Duda had described a different fly under this same name (Arch. f. Naturg. 90 A3: 204; 242) by including, in his new subgenus Hirtodrosophila, the nominal combination, Drosophila longecrinita var. dentata n. var., and later (1926, Suppl. Ent. 14: 69) elevated this dentata to the rank of species. Since the name dentata Duda 1927 was a clear homonym of dentata Duda 1924, Wheeler (1959, op. cit.: 183) proposed the new name, ten­data, to replace the rejected homonym. While reading the description of Drosophila fiavopilosa Frey (1918. Finska 1 Visiting investigator, Genetics Foundation, The University of Texas, 1960-1961; perma­nent address: Zoological Institute, Hokkaido University, Sapporo, Japan. 2 lnstituto de Biologia "Juan Noe", Sanitago, Chile; work performed under A.E.C. Contract AT(30-1) 2465. The University of Texas Publication Vetenskaps-Societetens Fordhandlingar 60 A14: 14) from Valparaiso, Chile, it was decided that fiavopilosa might possibly belong to the "dentata" complex. With the kind assistance of Dr. Walter Hackman we were able to borrow the type female of fiavopilosa from the Zoological Museum in Helsmki. Although the specimen was in rather poor condition, the ovipositor with its characteristic teeth was clearly visible and this, together with the other usable characters, showed clearly that it represented the same species as dentata Duda 1927, and hence, tendata Wheeler 1959. Malloch (1934. Dipt. Patagonia and S. Chile, Pt. VI, Fasc. 5: 444) reported additional specimens of dentata from Chile and Argentina. But he was clearly misled by the remarkably spined ovipositor of the female, writing "I strongly suspect that Duda had two species mixed here and that the supposed female was a male belonging to the next species [dudai Malloch]. ...... I have thus selected the form he described as his dentata, and the one he figured as a new species." For this new species, dudai, from Angol, Chile, he mentioned only a holotype male and an allotype. We found, however, in the U.S. National Museum collec­tion, the holotype and two paratype females (dated 1927 and 1928; one labelled "on decayed squash"). The abdomen of the male has been lost. Through the as­sistance of Dr. Willis Wirth we borrowed the two paratype females for study; both appeared to belong to funebris Fabricius. Dissection and study of the ovi­positor and spermathecae of one of the paratypes confirmed this synonymy. The male genitalia of the dudai holotype no longer being available, a positive identifi­cation is not possible, but the rather rough drawing of the visible external geni­talia of this holotype (Malloch, op. cit., p. 441, Fig. 75, d) can quite readily be interpreted as being that of funebris. Thus we have no hesitation in assigning dudai as a junior synonym of funebris. We have also re-examined two Chilean specimens determined by Malloch as dentata; both are flavopilosa. One is a female, labelled "Casa Pangue, Llanqui­hue, Chile; Dec. 1926; R. and E. Shannon," and the other is a male labelled "Angol, Chile, 17 Dec. 1926." The absence of further records of fiavopilosa in the more recent, modern studies on South American Drosophila is probably explained by the fact that it is a flower-breeder. Heed, Carson and Carson (1960. Drosophila Information Service 34: 84) have shown very forcefully the advantages of flower-collecting. In his monograph on Chilean species of Drosophila, Brncic (1957. Las especies Chilenas de Drosophilidae; Monograf. Biol. de la Univ. de Chile) failed to include "den­tata," but it was later discovered to be present by the thousands all over central Chile, living in the flowers of Cestrum parqui of the Solanaceae. It is the only species of the group whieh has been studied while alive. Meanwhile· additional specimens of fiavopilosa and of some of the new species were located in the U.S. National Museum collection and in South American material from the collection of the California Academy of Sciences, San Francisco, sent to us by Dr. Paul Arnaud. In addition collectors from the University of Texas laboratory, working under a National Science Foundation grant (NSF G-4999, to M. R. Wheeler) have found several new species at anumber of N eotropical localities. In the present report we are describing seven of these new species and are sum­marizing the known distributions. None of the species has as yet been satisfac­torily raised in the laboratory on the usual media, but details of the internal anatomy and cytology of Chilean fiavopilosa have been secured by one of us (D. B.) since this species is readily available in the gardens of the University in Santiago. Types of the new species are located in the following collections: U.S. National Museum, Washington, D. C. (USNM), California Academy of Sciences, San Francisco ( CAS), and Drosophila Type and Reference Collection of the Genetics Foundation, University of Texas (UT). Careful study of the male and female genitalia has required the use of microscope slide preparations; many of these were made permanent and are numbered to correspond to the specimen from which they came. Such slide preparations are indicated in the text as "prep. 245," etc.; the slides are in the University of Texas collection. THE SPECIES OF THE FLAVOPILOSA GROUP The species of this group, and their known distributions (the first named lo­ cality is that of the holotype), are as follows: fiavopilosa Frey-Chile; Bolivia; Peru; Argentina; Uruguay. incompta n. sp.-Panama; Dominica, B. W. I.; Mexico; Colombia. · nesiota n. sp.-Haiti. acroria n. sp.-Colombia. lauta n. sp.-Haiti. crossoptera n. sp.-Panama; Colombia. gentica n. sp.-El Salvador. Jamaica? gilvan. sp.-St. Lucia, B. W. I.; Panama; Colombia. unnamed species 1-El Salvador. unnamed species 2-El Salvador. unnamed species 3-Colombia. unnamed species 4-Ecuador. unnamed species 5-Haiti. unnamed species 6-Colombia. Although several of these species, at least in one sex, are recognizable on ex­ ternal characters alone, this is not generally true. Careful study of the male ·genitalia seems to be the most satisfactory way to separate the similar species since very little variation within a species has been seen. Female ovipositor struc­ture shows some variability, and will not always separate females with certainty. We cannot be sure that the differences seen in the spermathecal shape are entirely real; the method of preparation (hydrating pinned specimens in steam, boiling the abdomen in 10% sodium hydroxide, clearing in phenol, examination in oil of creosote) often results in rupture of the inner sclerotized capsule, and this could easily result in alterations of overall shape. In general, however, we believe that the figures represent fairly typical and repeatable appearances of these structures. Drosophila ftavopilosa Frey 191.8. Mitteilungen iiber siidamerikanische Dipteren. Finska Vetenskaps-Socie­tetens Fordhandlingar. Bd. LX, Afd. A, No. 14: 14. Type, from Valparaiso, The University of Texas Publication Chile, in Zoological Museum, Helsinki, Finland. =dentata Duda 1927. Arch. f. Naturg. 91 A12 [1925]: 201; not dentata Duda 19·24. Arch. f. Naturg. 90 A3:204; 242. =tendata Wheeler 1959. Univ. Texas Puhl. 5914: 183. =dentata Duda, Malloch 1934. Dipt. Patag. S. Chile, VI (5): 441; 444. Frey's description, not being generally available, is here quoted in its entirety: "Eine einfarbig gelbe Art, die durch die intensiv gelbe Farbe der Beborstung und Behaarung des ganzen Korpers wohl ziemlich leicht wieder zuerkennen sein wird. " Species DHO XAN SP CA AHP BIO BI UHC XIC willistoni g1·oup pseudo bocainensis ++++ ++ + + + +++ + ++ +++ +++ ++ nebulosa ++++ ++ ++ + +++ ++ + +++ ++ + + + ++ sucinea ++++ +++ ++ + ++ + +++ ++ +++ ++ capricorni ++++ ++ +++ ++ ++ + ++ + +++ +++ +++ changuinolae + +++ ++ ++ + +++ +++ ++ +++ ++ neoalagitans + +++ ++ +++ + ++ +++ +++ +++ ++++ fumipennis tropicalis ++++ ++++ +++ ++ +++ +++ ++ ++ ++ +++ + ++ +++ +++ +++ +++ +++ ++++ +++ ~ ~ ti) willistoni equinoxialis paulistorum + + ++ ++++ ++++ +++ +++ +++ ++ +++ + ++ + ++ + + ++ + + + + +++ +++ + ++ ++ +++ + + +++ + + + +++ ++++ +++ +++ ~ ;:l-·<::::: ti) Subgenus: Drosophila ~ -· ...... virilis-repleta section ~ virilis group ~ virilis ++++ + ++ + + + + + ·······­ ++ ----­--­ ~ americana ++++ ++ ++ + ++ ++ ····-·-­ + + --······ ~ novamexicana ++++ ++ ++ + +++ + + ++ ····---· ~ littoral is ezoana montana ++++ ++++ ++++ ++ ++ ++ ++ + + + + + + + ++ + ++ ++ + ++ + ++ ---­--·····­---····· +++ + + ++ ----···· -------­ 'tj l::: '3'" ......-· (") flavomontana lacicola ++++ ++++ ++ + + ++ ++ + + +++ ++ + + + ++ ---····· ++ ++ + ······-­---····· I:) ......5· ~ borealis + + ++ ++ ++ + ++ ++ ++ robusta group lacertosa + + ++ + ++ + ++ ++ ······-· +++ colorata ++ ++ + ++ + ++ ++ ····---­ + + + robusta ++ ++ + ++ + + + + + +++ m elanica group micromelanica ++++ ++ + + + ++ + ++ + + + -------· +++ melanica +++ + + +++ ++ +++ ++ + ---····· + + + paramelanica + +++ + +++ ++ + ++ +++ ····· +++ euronotus + +++ + + + + ++ ++ ++ + nigromelanica ++ ++ + +++ ++ ++ ++ ·····--­ + + + annulimana g roup gibberosa canalinea group canalinea paracanalinea dreyfusi group camargoi briegeri mesophragmatica group gaucha pavani repleta group mulleri subgroup aldrichi anceps an zonens1s buzzatii hamatofila longicornis martensis m eridiana m eridionalis mojavensis mulleri nigricruria pegasa peninsularis promeridiana stalkeri fasciola subgroup fulvalineata fasciola pictilis pictura fascioloides moju mojuoides +++ + ++++ ++++ +++ + ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ + + ++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ +++ + + + + + + + + + ++ + ++ + + + + ++ + ++ ++ ++ ++ ++ + + + + + + + + ++ +++ +++ ++ .i.++ ++ ++ + +++ ++ ++ + + + + ++ ++ ++ + ++ ++ ++ + + + ++ ++ ++ ++ ++ + + + + + + + + ++ + + + + + + + + + + + + + + + + + + + + +++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ + + + + ++ ++ ++ + ++ ++ ++ + ++ ++ ++ ++ ++ ++ ++ +++ ++ ++ +++ +++ ++ ++ + ++ ++ ++ + + + + ++ ++ ++ ++ ++ ++ ++ + ++ ++ ++ ++ ++ ++ ++ -······· ........ ··-····· ........ ...... ....... ........ ........ ........ .... ........ ······-· ·-····· ....... ........ ....... ···•··· ........ +++ +++ +++ +++ ++ + +++ +++ +++ +++ +++ ++ ++ ++ +++ ++ +++ +++ +++ +++ +++ +++ +++ ++ ++ ++ +++ +++ +++ +++ +++ ++ + ······ ··-···· -·-· ······ ....... .... ...... + ..... ·····-·­ + ..... + + + ++ ++ ++ ++ ++ "--3 a ~ (') ;>;-< ~ 0 "'( ..... 0 ;:::! .. b::l ...... 0 (') ~ (I) ~ ~· ..... ~ ~ ;::s ~ ~ 0 ;::s 0 ~ ~ ~ "I T A BLE Z-Continued Results from chromatograms of the heads of Drosophila species and of species from related genera ~ CX> Species DHO XAN SP CA AHP BIO BI UHC XIC mercatorum subgroup mercatorum ++++ ++ ++ + ++ ++ ........ +++ paranaensis ++++ ++ ++ + ++ ++ +++ melanopalpa subgroup fulvimacula ++++ ++ ++ + ++ ++ -------­ +++ fulvimaculoides ++++ ++ ++ -T­ ++ ++ .... +++ limensis repleta canapalpa melanopalpa hydei subgroup ++++ ++++ ++++ ++++ ++ ++ ++ ++ ++ ++ ++ ++ + + + + ++ ++ ++ ++ ++ ++ ++ ++ ----·--­........ ········ + + +++ ++ +++ + + + "'-3;:r. II:) C'.:! ;:l-. <::! nigrohydei bifurca ++++ ++++ + + ++ ++ + + ++ ++ ++ ++ ........ ----···· +++ +++ ------­ ~ "'..... ..... eohydei ++++ + ++ + ++ ++ +++ ~ 0- neohydei ++++ + ++ + ++ ++ ........ +++ hydei ungrouped species near the repleta group castanea ++++ ++++ + ++ ++ ++ + + ++ ++ ++ ++ ........ ........ +++ +++ ----···· ~ 1-i &l "tl I:: peruviana aureata ++++ ++++ ++ + +++ ++ ++ + ++ ++ ++ ++ ........ ----­ +++ ++ ++ + ~ .......... (") ~ bromeliae group bromeliae ++++ ++ ++ + + + + ··-··--· ++ + .....-· 0 ;:l nannoptera group nannoptera ++++ ++ +++ ++ +++ +++ .... +++ ++ q uinaria section immigrans group hypocausta ++++ ++ + ++ ++ +++ +++ ---· +++ ++++ spinofemora ++++ + +++ ++ ++ +++ .... +++ +++ immigrans ++++ + +++ ++ +++ +++ --······ ++ +++ funebris group m acrospina ++++ ++ +++ ++ ++ +++ --······ +++ + subfunebris ++++ ++ +++ ++ + ++ +++ ·······­ +++ + funebris ++++ + ++ + ++ ++ -------­ ++ + quinaria group innubila quinaria rellima falleni phalerata occiden talis species J subquinaria tenebrosa subpalustris palustris guttifera testacea group putrida calloplera group ornatipennis calloptera guarani group guaramunu subgroup guaramunu griseolineata guaraja guarani subgroup guarani subbadia cardini group dunni belladunni nigrodunni polymorpha neomorpha neocardini parthenogenetica ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++++ ++ ++ ++ ++ +++ ++ ++ ++ ++ ++ ++ ++ ++ ++ + ++ ++ +++ ++ ++ +++ + ++ ++ +++ + +++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ +++ ++ ++ ++ +++ ++ ++ ++ ++ +++ +++ +++ ++ ++ + + + + + + + + + + + + + + + + + ++ + + + + ++ ++ ++ + + +++ ++ + ++ ++ +++ +++ +++ ++ +++ ++ +++ +++ +++ ++ +++ ++ +++ ++ ++ ++ +++ +++ ++ ++ +++ ++ +++ +++ ++ ++ + ++ ++ ++ ++ ++ + ++ +++ +++ +++ ++ ++ +++ ++ ++ ++ +++ +++ ++ ++ +++ +++ ++ ---·· -------· ........ ---·--·· +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ +++ ++ +++ +++ +++ +++ ++ ++ ++ +++ +++ +++ +++ ++ + ++ +++ ++ ++++ ++++ +++ +++ ++++ +++ ++ ++ ++++ ++++ ++ +++ ++++ ++++ ++ ~ a (') ;:i ~ 0 "'1 c ~ b::i o· g.. ~ c;;· ..... ~ >:i ;:::i !=l... J...3 ~ 0 ;:::i 0 ;:i ~ .j::.. to co TABLE 2.----Continued Results from chromatograms of the heads of Drosophila species and of species from related genera ~ w 0 Species DHO XAN SP CA AlJP BIO Bl UHC xrc acutilabella ++++ + ++ + +++ ++ ........ +++ +++ procardinoides ++++ +++ ++ + +++ +++ ........ +++ ++++ cardinoides ++++ +++ ++ + +++ ++ ........ +++ ++++ cardini ++++ ++ ++ + ++ ++ ····---­ +++ +++ l'ttb1·ifrons g1·oup ~ pa rnch rogaster ++++ + ++ + ++ ++ ...... +++ +++ ~ uninubes ++++ + ++ + ++ ++ ........ +++ +++ ~ 1nacroptera group su bmacroptera ++++ +++ ++ + +++ +++ ........ +++ +++ ;:s.... ~ pallidipennis gl'Onp pallicli pennis ++++ + ++ + +++ +++ ........ +++ + + + ~ .... ~ tl'ipunctata gl'o np al bicans ++++ ++ +++ + +++ +++ ........ +++ ++ .a metzii ++++ + ++ + +++ +++ ····-··· +++ +++ ~ unipunctata ++++ +++ +++ ++ ++ +++ ........ +++ ++++ ~ mecliopuncta ta ++++ + ++ + +++ ++ ........ +++ +++ tripunctata trapeza bancleirantorum mediostriata mecliopictoicles ++++ ++++ ++++ ++++ ++++ +++ + ++ + +++ ++ ++ ++ ++ ++ + + + + + ++ +++ ++ ++ +++ ++ +++ +++ ++ ++ ........ ····---· ........ ........ ...... +++ +++ +++ +++ +++ +++ ++++ +++ ++++ +++ ~ g:f:;· ~ ....c;·;:s crocina ++++ + ++ + ++ ++ ····· +++ +++ ungrouped species near the tripunclala group sticta ++++ + ++ + ++ ++ .... . .. +++ ++++ species not placed in section aracea ++++ + ++ + ++ ++ ++ +++ ca rbonaria ++++ + ++ + ++ ++ +++ tumiclitarsus ++++ + +++ ++ ++ ++ ++ ++ Genus: Chymomyza aldrich i ++++ ++ +++ ++ ++ ++ + +++ ++++ procnemis ++++ ++ +++ ++ ++ ++ +++ ++++ Genus: Scaptomyza ad us ta ++++ ++ +++ ++ +++ +++ +++ ++++ pallid a ++++ ++ +++ ++ +++ +++ +++ ++++ Genus: Zaprionus ghesquierei ++++ ++ ++ + ++ ++ ++ +++ +++ vittige1· Genus: Mycodrosophila dimidiata ++++ ++++ ++ ++ ++ ++ + + +++ ++ ++ ++ ++ +++ ++ +++ -------­ "-3;:::.. a (') ?!­ ;:i c.., ..... c ~ to s· (');:::.. (I) ;:i <::;· ...... ~ ~ ;::s ~ ~ 1-i c ;::s c ;:i ~ ..p. (.)J ....... ~ TABLE 3 (.).) to Results from the chromatograms of the bodies of Drosophila and related species Species DHO ISO XAN SP CA HFL AilP BIO UHC YI OLher Genus: Drosophila Subgenus: Pholadoris victoria group pattersoni ······-· + + --······ +++ ······ +++ victoria ........ + + ........ ........ +++ ········ ·--·--­ +++ coracina group ~ lativittata ++ + ++ + ----···· -------­ + + +++ G ~ (I) cancellata ++++ ++ + ........ + ........ + +++ ........ G ~ novopaca ++++ ++ + ........ -------­ +++ ++ + +++ . ..... ····-­ ;::!-· latifasciaeformis group latifasciaeformis ++ ........ +++ ........ + +++ ·······­ ··-·---­ ~ (I) ""I "'-· Subgenus: Hirtodrosophila duncani + ++ + ++ + ++ +++ + ++ ----···· ........ ..... ~ 0- pictiventris +++ + -------­ ........ ++++ + + +++ +++ ---····· histrioides +++ + +++ ++ +++ ++ + +++ ----···· GZ ~ Subgenus: Dorsilopha ~ busckii Subgenus: Sophophora ........ +++ + +++ ++ ++ + + +++ ---·--­ GZ "ti.:: ~- obscura group -· ~ obscura subgroup ~ .....-· persimilis pseudoobscura ++++ ++++ ++ ++ + + + + ........ ·-·-·-· ++ ++ ++ ++ ........ +++ +++ ····--­ G G 0 ;::! miranda ++++ ++ + ++ + ++ ++ ........ + + + -----·· G ambigua ++ + + + + + ++ + ++ ++ ··-··-·· ++ + ....... G affinis subgroup affinis +++ ++ + + + + ++ ++ ++ ··-· G3 &V tolteca +++ +++ + ++ + +++ ++ ++ ······· G3 &V algonquin +++ +++ + ++ + +++ ++ ·····--· ++ -·-····· G3 &V azteca +++ ++ + + + + + +++ ++ -------­ ++ ········ G3 &V narragansett +++ +++ + + + + + +++ ++ ........ ++ --···-­ G3 &V athabasca +++ +++ + +++ + +++ ++ --······ ++ ·····--­ G3 &V melanogaster group melanogaster subgroup melanogaster simulans yakuba takahashii subgroup takahashii ananassae subgroup ananassae montium subgroup rufa serrata kikkawai aurana seguyi nikananu saltans group cordata subgroup neocordata elliptica subgroup emarginata neoelliptica sturtevanti subgroup sturtevanti milleri parasaltans subgroup subsaltans parasaltans saltans subgroup lusaltans a ustrosaltans pseudosaltans nigrosaltans saltans ........ ........ ........ ········ ...... ········ ······­ ----· --·-···· ...... ...... ·-···· ........ +++ +++ +++ +++ ++++ ++++ ++++ ++++ ++++ ++++ ++ +++ ++ + ++ +++ ++ ++ +++ +++ +++ +++ +++ + + + + + + + + + + + + + + + + + + + + + + + ++ +++ +++ + +++ ++++ +++ +++ +++ +++ + +++ +++ +++ +++ ++ ·······­ ........ +++ +++ +++ +++ ++ + ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ ++ + ++ ++ ++ ++ + ++ + + + ++ ++ ++ ++ ++ ++ + ++++ +++ +++ ++ +++ ++ ++ +++ +++ +++ +++ +++ + . ....... + ...... + ++ ++ ++ ++ ++ + ++ +++ + + + + + + + + + ++ + + + + + + + + ++ ++ ++ ++ ++ + + ++ ++ ++ ++ ++ ++ +++ +++ +++ +++ +++ +++ +++ +++ +++ ++ +++ ++ +++ ++ +++ ++ ++ ++ ++ ++ ++ +++ ······· ........ --·-·· --······ ...... --······ -----· ....... ........ ........ . ....... ··-----­ G2 G2 G2 G2 G2 G2 G2 G2 G2 G2 ...... G2 G2 G2 G2 G2 G2 G2 G2 G2 G2 ~ ~ g­ 0 ..., 0 ~ ~ s· ~ ~ ;3 f,;; • .... ~ ~ ~ ~ 0 ;:s 0 ;3 ~ -f>. w w -£>. TABLE 3-Continued w -£>. Results from the chromatograms of the bodies of Drosophila and related species Species DRO ISO XAN SP CA HFL AHP BIO UHC YI Other prosaltans +++ + ++ + +++ + +++ ----···· GZ septentriosaltans +++ + -­-----­ ........ +++ ········ + +++ willistoni group pseudobocainensis ..... ++ + ++ + +++ ...... + ++ . ...... GZ nebulosa sucinea ..... +++ +++ + + ++ + + +++ + + + +++ ++ ---· GZ -------­ 1...3 ~ capricorni changuinolae neoalagitans fumipennis tropicalis + ······· ....... ++ ++ ++ ++ ++ + + .... + + + ++ + + ······ + ........ ++ +++ +++ ++ +++ ········ + + + + + ++ +++ ++ ++ ++ ·····-­···--· ----·· GZ ·····-­----·· (I) ~ ;::i..... <::::: (I) "I "'..... ..... willistoni equinoxialis paulistorum Subgenus: Drosophila virilis-repleta section ······ · +++ +++ ++ + + ·-·­ + ++ + + ----··-· + ++ + ·---­ + ++ ++ +++ +++ +++ ··-­··-···-· ...... G2 ~ -Q. i "ti virilis group I:: ~ virilis americana novamexicana ++++ ++++ ++++ ++ +++ +++ + + + ++ +++ +++ + ++ ++ +++ +++ +++ ++ ++ ++ + + + +++ +++ +++ + ++ ++ GZ GZ GZ .......... (') l::i .....c; · littoralis ++++ +++ + +++ ++ +++ ++ + +++ ++ GZ ;::i ezoana ++++ +++ + +++ ++ +++ ++ + +++ ++ GZ montana ++++ +++ + +++ ++ +++ ++ + +++ ++ GZ fl avomontana ++++ +++ + +++ ++ +++ ++ + +++ ++ GZ lacicola ++++ +++ + +++ ++ +++ ++ + +++ ++ GZ borealis ++++ +++ + +++ ++ +++ ++ + +++ ++ GZ robusta group lacertosa ++++ ++ + ++++ +++ ++ ++ + +++ GZ colorata ++++ ++ + +++ ++ +++ ++ + ++ + GZ robusta ++++ + .... + ·····--­ + ·····--· + +++ m e lanica group micromelanica melanica paramelanica euronotus nigromelanica annulimana group gibberosa canalinea group canalinea paracanalinea dreyfusi group camargoi briegeri 1nesophrag1natica group gaucha pavani reple ta group mulleri subgroup aldrichi anceps arizonensis buzzatii hamatofil a longicornis martensis meridiana m eridionalis mojavensis mulleri nigricruria pegasa peninsularis promeridiana stalkeri +++ +++ +++ ++ ++++ ········ + + + + +++ ........ + ++++ + --······ + + + ........ ........ ........ ·····--­ ++ ++ ++ +++ +++ + +++ +++ ++ ++ +++ +++ ++++ +++ ++++ ++ +++ +++ +++ ++ +++ ++++ +++ ++ +++ +++ +++ +++ + + + + + + ++ ++ ········ ........ ........ ........ ·-·----­ ........ ········ ........ ........ ........ ........ ---····· ........ ···-···­ ---····· ··-····· ........ ++ ++ ++ ++ ++++ +++ ++++ ++++ + + ++++ +++ ++++ ++++ ++++ ++++ +++ ++++ ++++ ++++ +++ ++++ ++++ ++++ ++++ ++++ +++ +++ + + + + ++ ++ ++ ++ ........ +++ ++ ++ ++ ++ +++ + ++ ++ + ++ +++ ++ +++ ++ ++ + + +++ + + ++ +++ ++ +++ +++ + + +++ +++ +++ +++ ++++ ++++ +++ +++ +++ +++ ++ +++ +++ ++ +++ +++ +++ +++ ++ + + ++ +++ ++ ++ ++ -------· ----· ++ ++ +++ ++ +++ +++ ++ +++ +++ ++ ++ +++ +++ +++ ++ +++ ++ ++ + + + + + ++ ++ ++ + + ++ ++ +++ + +++ +++ + +++ ++ + + +++ ++ + + ++ + + ++ ++ ++ +++ +++ + +++ +++ ++ ++ +++ ++ +++ +++ ++++ ++ + +++ +++ +++ +++ + ++ +++ +++ +++ +++ ++ +++ +++ +++ ++++ ++++ ++ ++ ++ ++ ---····· + ........ ++ + + ++ + ---····­ ........ + + + -----··· + G2 G2 G2 ........ ........ G2 G2 G2 G2 G2 G2 G2 G2 G2 G2 G2 G2 G2 G2 G2 G2 G2 G2 "-3 ~ ~ () 3 ~ c "'; ..... c ~ tl:::I ..... c () ~ (1:) 3 c;;· ..... ~ ~ s. "-3 ~ c ;::i c 3 ~ -!'­ w (Jl T ABL E 3-Continued w """' 0) Results from the chromatograms of the bodies of Drosophila and related species Species DHO ISO XAN SP CA HFL AIIP BIO UHC YI OLher fasciola subgroup fulvalineata + +++ ++ + + ++ ++ +++ + + GZ fasciola ++ +++ ++ + +++ + ++ + ++ ++ GZ pictilis ++ ++ ++ + +++ ++ ++ + + + GZ pictura ++ ++ + + + +++ + ++ +++ ++ GZ fascioloides + +++ -······ ++++ + + +++ +++ ++ +++ ++ GZ "":3 m o ju + ++++ ++ + ++ + +++ ++ GZ ~ (':) mojuoides mercatorum subgroup + ++ ······· + ········ ++ ······ + ++ + GZ t:1;::s..... mercatorum paranaens1s melanopalpa subgroup ...... +++ +++ + + + +++ + + ++ +++ ++ ++ + + ++ + +++ + + GZ GZ <::: (':) "'! "'..... ..... ~ fulvimacula fulvimaculoides ········ +++ ++ + + + +++ + + ++ + + ++ + + +++ ++ + + GZ GZ .a lim ensis +++ + ........ ++ + +++ ++ GZ ~ repleta ++++ ·····-­ ++ + +++ ++ + +++ ++ · GZ ~ canapalpa m elanopalpa .... ++++ ++++ ······ ++ + ++ + + +++ +++ ++ ++ + + +++ +++ ++ + GZ GZ "ti I:: l::l"'- hydei subgroup nigrohycl ei bifurca eohydei + ········ ·····-· ++++ ++ +++ ........ ..... ........ + + ++ + + + + + ++ + + ++++ + ++ +++ ++ ++ + + + +++ ++ ++ + + +++ + GZ GZ GZ ..... (") i::i .....c;· ;::s neohycl ei ........ + + + + ++ ++ +++ ++ + +++ ++ GZ hydei +++ + ........ + + + ++++ ++ + + ++ + ++ GZ ungrouped species near the repleta group Cas tanea ..... ++ . ..... ++ + + ++ + ++ + GZ peruviana ---····· ++ ··-----­ + + + + + ++ + + + GZ a ureata +++ + +++ + ++++ ++ + +++ ++ GZ bromeliae group brom eliae + + ++ + +++ + +++ + + + ++ nannoptera group nannoptera ++++ +++ ++ +++ ++ +++ ++ ++ +++ + GZ quinaria section immigrans group hypocausta +++ + + +++ ++ ++ ++ ++ ++ + GZ spinofemora ++ ++ + + ---·---­ + + + ++ + immigrans + ++ + ++ + ++ -----­ + ++ f unebris group macrospina ++ ++ + ++ + ++ + ++ + ++ + GZ ~ subfunebris ++ ++ + +++ ++ ++ ++ + ++ + GZ d funebrjs + ++ + + ··-····· ++ ........ ++ ++ ---····· ~ ~ ~ quinaria group innubila quinaria + + +++ ++ +++ + ++ ........ + ++ ·····-· ·--··-· + + +++ +++ + + GZ ----·· 0 ""1 C' ~ rellima falleni + + +++ +++ + ....... + ++ + ++ ++ + + +++ +++ + + ........ GZ ttl5· phalerata occidentalis + ++++ +++ +++ + +++ ++++ ++ +++ +++ +++ ++ + + ++ +++ + GZ GZ ~ ;:i-.. (1:) ~ species J subquinarja ++ ++ +++ +++ ........ +++ +++ ++ ++ +++ ++ ........ + + +++ +++ + + GZ GZ ~· ..... ~ tenebrosa +++ +++ + +++ ++ +++ + +++ GZ i:::i subpalustris palustris + + ++ +++ ····-· ++ ++ + + ++ ++ + ........ ........ + + +++ +++ + + GZ GZ s.. ~ guttifera + +++ + +++ ++ +++ ........ ++ +++ + GZ ~ testacea group 0 ~ putrida + +++ -------­ -------­ -------­ + + ++ ++ ~ calloptera group ~ omatjpennis ········ + + ······-­ ···----­ ++++ + +++ + ++ calloptera ........ ······-­ + ······-· -------­ ++++ -------· + ++ +++ guarani gro up guaramunu subgroup guaramunu ++ + + ........ +++ + +++ ++ griseolineata guaraja ........ ........ ++ + + + + ++ ........ + ++++ +++ -------­ + + +++ +++ + + ····-·­ -i;:.. (.)J 'I TABLE 3-Continued ~ w Results from the chromatograms of the bodies of Drosophila and related species Species DRO ISO XAN SP CA Hl' L AHP BIO UHC YI Other guarani subgroup guarani ·----··· ++ + ++ + +++ ........ + +++ + subbadia ++ + ++ + +++ + ++ + cardini grou p dunni -------­ ++ + + ·-······ +++ ++ KYL belladunni ++ + ········ ++ + ++++ + KYL 1-..3 nigrodunni ++ + ........ ++ + +++ + KYL ::l­~ polymorpha + +++ ++ + + +++ + ++++ ++ KYL ~ neomorpha + ++ ++ + ++ + +++ ++ KYL ;:s-· neocardini + + + ++ + +++ + +++ ++ KYL ~ ~ parthenogenetica ++ + + ···---·· + ········ + ++++ +++ KYL "'j"'-· acutilabella + ++ ++ + ........ + + ++ +++ KYL ...... ~ procardinoides cardinoides cardini ........ ---····· ++ ++ ++ + + ++ + + + ----·· .... ........ ++++ ++ ++++ ·------­ + + + +++ +++ +++ ++ ++ + +++ KYL KYL KYL .Q_ 1-..3 ~ rubrifrons group ~ parachrogaster ------·· +++ -------­ + ........ ++ + +++ +++ "'ti uninubes 1nacroptera group ........ ++ ........ ---····· -----­ +++ ········ + +++ +++ ···· ···· i:::: <::!"'--· ~ submacroptera pallidipennis group pallidipennis ----·--­ ++ ++ ····­·····-·· + + -------­........ +++ ++ ······­+ + + ++++ +++ ++ ' ······­ $:) ......c; ;:s tripunctata group albicans + + + -------­ ........ + + ++ + metzii + ..... ........ ++ + ++ + unipunctata + + ++ + ++ + +++ ·------­ + +++ -r mediopunctata + ++ + + .... .. + + +++ tripunctata + ++ + + ·------­ + ........ + +++ + trapeza ++ ........ ··-····· ++ + +++ + bandeirantorum + + -------­ -------­ + -·-····· + +++ + mediostriata + +++ + + -------­ +++ ········ + +++ + m ediopictoides crocina ungrouped species near the tripunctata group sticta species not placed in section aracea carbonaria tumiditarsus Genus: Chy momyza aldrichi procnemis Genus: Scaptomyza adusta pallida Genus: Zaprionus ghesquierei vittiger Genus: M ycodrosophila dimidiata + + ...... ........ ++++ ······-· ........ ··----·· .,._... ····---­ -----· ++ +++ + +++ ++ +++ +++ +++ ++++ +++ +++ +++ ++ + + + ----···· ........ ........ ---·· + -----· ·-----·­ + ····-·-­ + ++++ ++ ++++ ···-·-·­ ........ +++ +++ ++++ ++++ + ---····· -------· ········ +++ + +++ ........ ++ ++ +++ +++ + + ++ + ++++ +++ ++++ +++ ++++ +++ ++ ++ ++ ++ ····--­ ........ +++ ++ +++ ······ + + +++ +++ + + + + +++ ++ + + ++ + + +++ +++ +++ ++ +++ +++ +++ +++ +++ ++ ++ + +++ + + + ++ ........ v ? ------­- -·----·· ······-­ -----· ++ G2 ++ G2 G2 ........ G2 ........ "'-3 ;:s-­ Cl (') ;:>r­ ;:i 0 "j c ~ ~ ..... 0 (') ~ ;:i &:;· ..... ~ ~ ~ ~ 0 ;:s 0 ;:i ~ ...... (.).) <.O T ABLE 4 + + Results from the electrophoresis of the testes of Drosophila species Species DHO ISO SP BL YG BIO AHP GH Genus: Drosophila Subgenus: Pholadoris victoria group pattersoni ++ ···--· ........ ........ . ....... . ....... ++ v:ictoria ++ ----···· ···-···· ---····· ·-----·· ++ ~ coracina group ~ lativittata + + + ++ -------­ + + cancellata novopaca + + ++ ++++ ++ ++ -------­ ........ ----··-­ . ....... ····-··· + + + + ........ ~ ;::s...... ~ ~ latifasciaeformis group latifasciaeformis ++ ........ -·-----­ ······-· + -----·-­ "'t "'::::.· Subgenus: Hirtodrosophila duncani + ++ ++ -------­ ++ ········ . ... ~ .Q.. pictiventris ··-····· +++ -------­ ........ ++ ++ ++ -----­ ~ Subgenus: Dorsilopha busckii Subgenus: Sophophora +++ +++ + ++ +++ ++ ........ ~ "ti i::: obscura group obscura subgroup persimilis pseudoobscura ++++ ++++ ++ ++ + + ·--····­-------­ ...... ........ ++ ++ ~ .....c:;· I::)..... s· ;::s miranda ++++ ++ ++ ........ -------­ ........ ++ ambigua ++++ ++ ++ -------­ ........ ··-·-··· ++ affinis subgroup affinis +++ ++ ++ ···-·--­ ··-····· ++ ++ tolteca +++ +++ ++ ------·· ........ + +++ +++ algonquin +++ +++ ++ ----···· ········ + +++ +++ azteca +++ +++ +++ ··--···· + +++ +++ narragansett +++ +++ +++ -------­ ······· + +++ +++ athabasca +++ +++ +++ ........ . ....... + +++ +++ 1nelanogaster group melanogaster subgroup melanogaster + +++ +++ ++ .... +++ + simulans + +++ +++ ++ + +++ + yakuba + +++ +++ ++ +++ + takahashii subgroup takahashii + ++ ++ + ++ + ananassae subgroup ananassae + +++ +++ ++ ++ +++ ++ ...... . "'-3 monlium subgroup rufa + +++ +++ ++ ...... +++ + ~ .., -0 () serrata + +++ +++ ++ + +++ + .... .... ~ ~ kikkawai aurana + + +++ +++ +++ +++ ++ ++ + +++ +++ + + 0.., ..... 0 seguyi + +++ +++ ++ +++ + ·· ··­- ~ nikananu + +++ ++ + + ++ + b::1 saltans g1·oup s· () cordata subgroup neocordata ++ +++ + + ++ + .... ... ~ 11) ~ elliptica subgroup !:;• ..... emarginata + +++ ++ + ++ +++ -------· ....... . ~ neoelliptica + +++ +++ + + +++ ...... • ······ ~ slurtevanti subgroup ~ ~ sturtevanti -1­ +++ +++ + ++ + milleri ++ ++ ++ +++ + ~ parasaltans subgroup ~ subsaltans ........ ++ + + ++ -------­ ........ 0 parasaltans ........ ++ ++ ++ + ........ ~ ~ saltans subgroup lusaltans + +++ +++ + ········ +++ austrosaltans + +++ ++ + ········ +++ + pseudosaltans + +++ +++ + +++ + nigrosaltans t +++ ++ ++ +++ + saltans +++ ++ + + ++ + prosaltans ........ +++ ++ + + ++ + ········ ~ septentriosal tans ······· +++ . .... .... ········ ++ ·-·· ~ ........ T ABLE +-Continued "*'" "*'" Results from the electrophoresis of the testes of Drosophila species to Species DRO ISO SP BL YG BIO Al-IP GI\ willistoni group pseudobocainensis +++ +++ +++ ++ ++ + nebulosa ........ +++ ++ ++ +++ + sucinea ·------­ ++ + ········ ++ + capricorni ······· ++ + .... ++ + changuinolae ++ +++ ++ + + +++ + neoalagitans ·------­ ++ + + ---­ ++ + ~ fumipennis ·--­ ++ + + ---­ ++ + ~ (I) tropicalis willistoni ........ +++ +++ ++ + ++ + -----­ +++ +++ + + ······ C'.1 ;::l..... equinoxialis paulistorum ....... +++ +++ + + + + ········ ·---·--· +++ ++ + + ...... ~ (I)...., "'..... Subgenus: Drosophila ..... ~ virilis-repleta section .Q. virilis group virilis ++++ ++ ++ + ---····· + ++ ····-··· ~ americana ++++ +++ +++ + --······ + ++ &l novamexicana ++++ +++ +++ + ······ + ++ "ti littoralis ++++ +++ +++ + ······-· + ++ -----· !:::: ezoana ++++ +++ +++ + + ++ ........ "'""......... ~ montana fl avomontana ++++ ++++ +++ +++ +++ +++ + + ........ + + ++ ++ ...... l::i .....5· lacicola ++++ +++ +++ + ---····· + ++ ;::l borealis ++++ +++ +++ + ---····· + ++ rohusta group lacertosa ++++ +++ ++++ ++ --······ ++ + colorata ++++ +++ +++ ++ -----­ + +++ robusta +++ ++ + + + rnelanica group micromelanica +++ ++ ++ ········ ········ ++ ++ melanica +++ ++ ++ -·-----­ ++ ++ paramelanica +++ ++ ++ -------­ + + ++ euronotus ++ +++ ++ ---· .. ..... ++ + ++ nigromelanica ++++ +++ ++++ + ........ ++ +++ annuli1nana group gibberosa ······· ++ +++ + + +++ +++ + + canalinea group canalinea + ++ + ++++ + + ++++ ++ ++ paracanalinea + + + + ++++ + + ++++ +++ + dreyfusi group camargoi + ++ + + ++ + ........ ~ briegeri m esophragmatica group + ++ + ++ -······· ++ + ........ cs (") gaucha + + + + ++++ ++ + ++++ ++ -----­ ~ ~ pavani +++ + + ++ + + + +++ + -·-····· 0 "'j reple ta group mulleri subgroup c ;::s•.. aldrichi ++++ +++ + + + + + + +++ ++ ........ tl:l-· anceps .... +++ ++++ ++ ++ +++ +++ ·---··· 0 (") arizonensis + +++ ++++ ++ ++ +++ +++ -------­ ~ ~ buzzatii ++++ +++ ++++ ++ ······-· ++ + ·-······ ~ hamatofila +++ + + + ++ ++ +++ + !;; • ..... longicornis + +++ ++++ ++ + +++ +++ +++ ········ ~ martensis m eridiana + +++ ++ ++ + + ++++ ++ ++ + + +++ +++ +++ +++ -------­···----­ ~ ;::s ~ meridionalis + ++ + + ++ ++ + +++ +++ ........ ~ mojavensis mulleri + + +++ +++ + + ++ ++++ + + ++ ++ +++ +++ ++++ ++ +++ ........ ........ ~ 0 ;::s nigricruria pegasa -------­ +++ + ++ ++++ ++++ +++ + + + +++ + +++ ++ +++ + --··-··· ·-······ 0 ~ ~ peninsularis ·---···· +++ ++++ +++ + +++ +++ promeridiana + ++ ++++ ++ + +++ +++ stalkeri +++ +++ +++ + +++ +++ fasciola subgroup fulvalineata ........ +++ ++ +++ -------­ +++ + fasciola ++ +++ ++ + ----­ +++ + pictilis pictura ++ ++ ++ + +++ + + ++ +++ +++ + + +++ +++ ++ + ........ -------­ t w T ABL E 4-Continued Results from the electrophoresis of the testes of Drosophila species t Species DRO ISO SP BL YG BIO AHP GR fascioloides + +++ +++ + + + + +++ ++ moju + +++ + +++ --·· +++ + m ojuoides + +++ + + ........ ++ + mercatorum subgroup me)·catorum +++ +++ ++ +++ +++ +++ paranaensis ........ +++ +++ +++ + +++ + melanopalpa subgroup fulvimacula ·-······ +++ +++ +++ ++ ++ ++ ---· "":'l ;:i-. ~ fulvimaculoides ----·-·· +++ +++ +++ ++ ++ ++ -------­ C::! limensis ........ ++ + ++ + ++ + --­----­ ;:i..... repleta canapalpa ----·--­........ ++ ++ ++ ++ ++ ++ + + ++ ++ + + ····---­----­ ~ ~ "'j "'..... melanopalpa ··-····· ++ ++ ++ + ++ + ........ ~ hydei subgroup .a nigrohydei + +++ ++++ +++ + + + -------­ ~ bifurca ........ +++ ++ +++ + ++ + -------­ 1-l eohydei -------· +++ ++ +++ + ++ + -------­ t; neohydei hydei ungrouped species near ...... +++ +++ +++ +++ +++ +++ ++ + +++ ++ ++ + ········ ·····-·­ 'ti;:: ~ .......r:; · ~ the repleta group castanea peruviana ········ ........ +++ + +++ ++ ++ +++ + ........ ++++ ++ + + + ........ ....5· ;:i aureata ....... ++ +++ ++ + ++ + nannoptera group nannoptera + +++ +++ ++++ ++ ++ ++ + + quinaria section immigrans group hypocausta + ++ + ++++ + ++ ++++ ++ spinofemora ++ ++ + + + ++ + immigrans + ++ +++ ++ + +++ + funehris group macrospina ++ ++ ++ + +++ + + + subfunebris ++ ++ +++ + ++ ++ + funebris + ++ ++ + + + + quinaria group innubila + +++ +++ + ++ ++ + quinaria + ++ + ·····--­ ++ ++ + rellima + ++ ····---­ ++ ++ falleni + ++ + + ++ ++ + phalerata + +++ +++ + ++ ++ + occiden talis +++ +++ +++ + ++ ++ + · ······· "'-3 species J ++ +++ +++ + ++ ++ + -------­ ~ subquinaria tenebrosa ++ +++ ++ ++ +++ ++ + + +++ ++ ++ ++ + + -------­-----··· (') ;>\"" ~ subpalustris + ++ ++ + +++ ++ + ------­- c..., palustris + ++ +++ + ++ ++ + -------­ 0 guttifera + +++ +++ + --······ ++ ·····--­ -------­ ~ testacea group putrida + ++ + -------­ ++ + -----­·· b::i g· calloptera group ~ ornatipennis calloptera ........ ···-··-· . ....... ···--··­ ........ ........ -------· . ....... -------· ++ ++ ........ -------· ---····· -------· ~ t;;·.... guarani group ~ guaramunu subgroup ~ guaramunu + ++ ---····· -·--·· ++ + ~ griseolineata + ++ ........ ··-­ ++ ······-­ 1--3 guaraja guarani subgroup + ++ ----···· ++ ++ + -------­ ~ c ;:s guarani subbadia + + ++ ++ ++ ++ ........ ........ -------· ++ ++ ++ .+ + --·--··· ······ c ~ ~ cardini. group dunni .. .. ++ -----­ ·······­ + belladunni ·····--­ ++ ...... ........ ..... + nigrodunni ........ ++ ........ ...... + polymorpha + ++ + ···--··· -------­ ++ + neomorpha + ++ + -------­ --··· ++ + neocardini parthenogenetica + + ++ ++ + + ········ ........ ··--···· ++ ++ + + -------­ t ()\ t Ol TABLE 4--Continued Results from the electrophoresis of the testes of Drosophila species Species Dl\O ISO SP BL YG BIO AHP GI\ acu tila bella + ++ + ------­- -----· ++ + procardinoides + ++ + -------­ ---­ ++ + cardinoides + ++ + -­-----­ --······ ++ + cardiui. rubrifrons group + + + + ........ . ....... ++ + ........ "'-3 ~ parachrogaster + ++ + + + ++ + ++ + ........ uninubes 111acroptera group submacroptera pallidipennis group pallidipennis tripunctata group + ++ + +++ ++ ++ --­····· + ------·· ........ ···-···· ·­······ ++ + ++ ++ ++ -t + + ·····-­·-····-· ---­--­- ~ ;;:s.... ~ (I) ;;i.... ~ .Q.. albicans ++ + + ---­ ........ +++ ........ ........ ~ m etzii + + ++ + -----­ ++ + + ·····-·· H unipunctata m ediopunctata tripunctata trapeza bandeirantorum m ediostria ta ++ ++ ++ ++ + + + + ++ + + ++ + ++ ++ + + + + + + + + -------­........ ........ ----·-­· + + ........ + ·-·· ····· ..... +++ ++ + ++ ++ + ++ ········ ········ ........ ----···· -------· ...... . ....... -------­........ ~ 'ti $::: <::!"'....c:; · ~ ....c;· ;;:s mediopictoides + ++ + + + +++ crocina + ++ + ........ ........ +++ ungrouped species near the lripunctata group sticta + ++ + -------­ ........ + ++ + species not p laced in section carbonaria ++++ + + ++ ........ . ....... ········ ++ ? TABLE 5 Results from electrophoresis of the Malpighian tubules of selected Drosophila species Species RFL ISO BL YI BG GR XIC URC Species RFL ISO BL YI BG GR XIC UHC Subgenus: Pholadoris victoria group pattersoni ++++ + ........ + ++ ----·--­ + ~ coracina group ~ novopaca +++ + -------­ ........ + +++ + ++ Cl (') Subgenus: Hirtodrosophila ;ii.. duncani ++ + ....... ------­- ++ + ++ ++ ~ pictiventris +++ + ++ +++ +++ + +++ +++ 0 "'! c Subgenus: Dorsilopha ~ busckii + + ++ +++ ++ + + + b::l Subgenus: Sophophora .... 0 ohscura grou p ~ obscura subgroup pseudoobscura +++ + ........ ········ ++ +++ ++ + ~ ·e;;. affinis subgroup ..... algonquin +++ + ........ . ....... ++ ++ ++ + ~ · l::i 1ne lanogaster group ;:s ananassae subgroup ~ ananassae ++ + ----··-· ........ ++ + ++ ++ ~ montium subgroup 0 rufa ++ + + +++ + +++ ++ ++ ;:s 0 seguyi +++ + + +++ + +++ +++ + ~ saltans g roup ~ elliptica subgroup neoelliptica ++ + -------­ -------­ + ++ + +++ parasaltans subgroup parasaltans +++ + ··-----­ ········ + ++ + +++ willistoni g roup willistoni ++ + ........ . ....... + + ---­ ++ neoalagitans +++ + ·····--· -------­ +++ +++ + ++ t 'I ~ TABLE 5-Continued """00 Results from electrophoresis of the Malpighian tubules of selected Drosophila species Subgenus: Drosophila virilis-repleta section virilis group virilis + + ···-·--· + ++ ........ ---·· +++ montana 1·obusta group + + ++ +++ ++ + ...... ++ 1--3 ~ lacertosa +++ ++ + + +++ color a ta robusta ++++ ++ + + ++ ····--·· ++++ + +++ + +++ + ++ ····--·· ++++ +++ ~ ;::s-. ~ ~ melanica group euronotus + + .... .... + + +++ "1 "'-· ~ nigromelanica annulimana group +++ + -----­ + +++ +++ ...... +++ .Q_ gibberosa canalinea group ++ + + ++++ + +++ ······· ++++ ~ "~ canalinea mesophragmatica grou p pavani repleta group +++ + + + + ···----­ ++++ + ++ + ++ +++ ······· ........ +++ ++ "1::1 ~ <::3­.......c:; · ~ mulleri subgroup aldrichi +++ + ........ + +++ ++ + ++++ .....-· 0 ;:i buzzatii ++++ + ........ ++ +++ ++ +++ longicornis ++ + + +++ + ++ + +++ martensis +++ + + ++ ++ +++ ++ +++ meridiana ++ + + + ++ +++ +++ ........ +++ mojavensis + ++ + + + ++ ········ ++++ mulleri + ++ ·­······ ---­ +++ +++ ........ +++ fasciola subgroup fulvalineata + + + +++ + ++ ++ ··-----­ ++ + moju ++ + ++ +++ +++ ++++ + ++ mercatorum subgroup mercatorum melanopalpa subgroup rep I eta hydei subgroup bifurca hydei ungrouped species near the repleta group castanea peruviana quinaria section immigrans group hypocausta immigrans funebri~ group macrospina quinaria group occidentalis subpalustris calloptera group calloptera guarani group griseolineata cardini group belladunni rubrifrons group uninubes tripunctata group mediopunctata species not placed in section carbonaria +++ +++ +++ ++ ++ + +++ + ++ ++++ + ++ ++ +++ + + ++ +++ + + + + + ++ + + + + + ........ + + + + + + + ++ + ++ + + + ·--· + + ++ ........ + +++ ++ + ++++ +++ ++++ +++ +++ ···----­ ++++ + + ++++ ++ +++ + ++ ··--··-· ++ ++ +++ ++ +++ + +++ +++ + ++ + + ++ + + ++ +++ +++ ++ + +++ + +++ +++ +++ ++ ++ + ++ + + ++ + + --····-­ ........ ........ ........ + + ...... -····· ........ ........ ---··-·· ------·­ ········ ++ ++ +++ +++ ++ +++ ++ ++++ +++ +++ +++ +++ ++++ +++ +++ +++ +++ ++ ~ d ~ ~ 0 ""j ... 0 ~ tl::I g· ~ ~ ;:i c;:;· ..... ~ i::i ~ ~ ;:i 0 ;:i ~ :t \0 The University of Texas Publication TABLE 6 Results from the chromatograms of late third instar larvae of Drosophila species and of species from related genera Species ISO XAN RFL GI URC KYL YI Genus: Drosophila Subgenus: Pholadoris victoria group pattersoni + + +++ ++ +++ + victoria + + ++ ++ +++ ++ coracina group lativittata + + +++ +++ +++ +++ cancellata + + +++ +++ +++ +++ novopaca + + +++ +++ ++ ++ latifasciaeformis group latifasciaeformis + + ++ +++ ++ ++ Subgenus: Hirtodrosophila duncani + +++ + + + + pictiventris + + +++ + + ++ + histrioides + + +++ ++ + ++ + Subgenus: Dorsilopha busckii + + + + + + Subgenus: Sophophora ohscura group obscura subgroup persimilis + +++ +++ +++ ++ pseudoobscura + +++ ++ +++ ++ miranda + ++ ++ ++ ++ ambigua + +++ +++ ++ ++ affinis subgroup affinis + + ++ +++ +++ ++ tolteca + +++ +++ +++ +++ algonquin + ++ ++ + ++ azteca + + + ++ + ++ narragansett + + ++ + +++ ++ athabasca + + ++ ++ + m elanogaster group melanogaster subgroup m elanogaster + + + ++ + simulans + + ++ +++ ++ + yakuba + + ++ +++ ++ ++ takahashii subgroup takahashii + + ++ +++ ++ ++ ananassae subgroup ananassae ++ + +++ +++ +++ +++ montium subgroup rufo ++ + +++ ++ +++ ++ + serrata + + ++ ++ +++ ++ ++ kikkawai + + ++ ++ ++ + ++ auraria ++ + ++ ++ +++ +++ +++ seguyi ++ + +++ +++ +++ +++ +++ nikananu + + +++ +++ +++ +++ +++ saltans group cordata subgroup neocordata + + ++ ++ + + Throckmorton: Biochemistry and Taxonomy TABLE 6--Continued Results from the chromatograms of late third instar larvae of Drosophila species and of species from related genera Species ISO XAN RFL GI URC KYL YI elliptica subgroup emarginata ++ + ++ ++ ++ + neoelliptica + + ++ + ++ + sturtevanti subgroup sturtevanti + + ++ ++ +++ + milleri + + ++ ++ +++ + parasaltans subgroup subsaltans + + ++ ++ +++ + parasaltans + + ++ + ++ ++ saltans subgroup lusaltans ++ + ++ +++ +++ ++ austrosaltans ++ + ++ ++ +++ ++ pseudosaltans ++ + ++ ++ +++ ++ nigrosaltans ++ + ++ + +++ +++ saltans + + ++ ++ +++ + prosaltans + + ++ ++ +++ ++ septentriosaltans + + ++ ++ +++ ++ willistoni group pseudobocainensis ++ + ++ ++ +++ ++ nebulosa ++ + +++ +++ +++ +++ suc1nea + + + +++ ++ capricorni + + + ++ + + neoalagitans ++ + +++ ++ +++ +++ fumipennis + + +++ +++ +++ + tropicalis + + ++ +++ +++ ++ willistoni + + + +++ ++ equinoxialis + + + ++ ++ paulistorum + + + ++ +++ Subgenus: Drosophila virilis-repleta section virilis group virilis + + ++ + + ++ americana + + + ++ + ++ novamexicana + + ++ ++ + ++ littoralis + + ++ ++ + ++ ezoana + + + + + ++ montana + + ++ ++ + ++ fl avomontana + + ++ ++ + ++ lacicola + + + ++ + ++ borealis + + + ++ + +++ rohusta group lacertosa + ++ ++ + +++ +++ colorata + + ++ + + robusta + + ++ + + m elanica group micromelanica + + ++ ++ ++ ++ melanica + + ++ ++ ++ ++ paramelanica + + ++ ++ ++ + ++ euronotus + + +++ +++ ++ + nigromelanica + + ++ + +++ The University of Texas Publication TABLE 6-Continued Results from the chromatograms of late third instar larvae of Drosophila species and of species from related genera Species ISO XAN RFL GI URC KYL YI annulirnana group gibberosa + + + +++ +++ + canalinea group canalinea + + ++ + ++ paracanalinea + + +++ +++ + + ++ dreyfusi group camargoi + + + ++ + ++ briegeri + + + +++ ++ + + mesophragrnatica group gaucha + + +++ +++ + ++ pavani + + ++ ++ + + repleta group mulleri subgroup aldrichi + ++ ++ ++ ++ anceps + + ++ ++ + + -r arizonensis + ++ + ++ + buzzatii + + ++ + + hamatofila + + ++ + + longicornis + + +++ ++ + + martensis + + ++ + ++ meridiana + + + ++ +++ + + meridionalis + + ++ ++ ++ + mojavensis + + ++ +++ + mulleri + ++ ++ + + nigricruria + + ++ ++ + ++ + pegasa + ++ ++ + peninsularis + + ++ + + promeridiana + +++ ++ ++ ++ stalkeri + + ++ + + fasciola subgroup fulvalineata + + + + + fasciola + + ++ ++ + ++ pictilis + + ++ +++ ++ ++ pictura + + ++ +++ ++ ++ fascioloides + + ++ ++ ++ ++ + moju + ++ ++ + ++ mojuoides + + ++ +++ ++ ++ ++ mercatorum subgroup mercatorum + + + + + + paranaensis + + + + + + melanopalpa subgroup fulvimacula + ++ ++ ++ + fulvimaculoides + + ++ + + limensis + + ++ + + repleta + + ++ + + ++ canapalpa + + ++ + + ++ melanopalpa + + ++ + ++ ++ hydei subgroup nigrohydei + + ++ ++ + bifurca + + ++ + + ++ TABLE 6-Continued Results from the chromatograms of late third instar larvae of Drosophila species and of species from related genera Species ISO XAN RFL GI URC KYL YI eohydei + ++ + + ++ neohydei + ++ + + ++ hydei + + ++ + + +++ ungrouped species near the repleta group castanea + ++ ++ + +++ peruviana + + ++ ++ ++ ++ ++ bromeliae group bromeliae + + +++ + + + ++ +++ nannoptera group nannoptera ++ + +++ + +++ +++ + quinaria section immigrans group hypocausta + + ++ + ++ spinofemora + + ++ + + + ++ immigrans + + ++ ++ + ++ funebris group macrospina + + +++ + ++ ++ ++ subfunebris + + +++ + + ++ ++ funebris + + +++ + ++ ++ ++ quinaria group innubila + + ++ ++ + + + quinaria + + ++ ++ + ++ ++ falleni + + + + + ++ phalerata + + ++ + + ++ occidentalis + + ++ + + ++ species J + + + ++ + + ++ subquinaria + + + ++ + ++ ++ tenebrosa + + ++ ++ + + + subpalustris + + ++ + + ++ ++ palustris + + ++ + + ++ guttifera ++ + ++ + ++ ++ ++ testacea group putrida ++ + ++ +++ +++ + calloptera group ornatipennis + + +++ ++ +++ + + ++ calloptera + + ++ + ++ + guarani group guaramunu subgroup guaramunu + + +++ +++ + + + ++ griseolineata + + +++ +++ + ++ ++ guaraja + + +++ +++ + +++ +++ guarani subgroup guarani + + ++ ++ ++ ++ + + subbadia + + +++ ++ + ++ + cardini group dunni + + + ++ + ++ +++ belladunni + + + ++ + ++ +++ nigrodunni + + + ++ + ++ +++ polymorpha + ++ ++ + + +++ The University of Texas Publication TABLE 6-Continued Results from the chromatograms of late third instar larvae of Drosophila species and of species from related genera Species ISO XAN RFL GI URC KYL YI neomorpha + + ++ ++ + +++ +++ neocardini + + ++ +++ + ++ +++ parthenogenetica + + ++ ++ + ++ +++ acutilabella + ++ ++ + ++ + procardinoides + + ++ ++ + ++ ++ cardinoides + + ++ +++ + +++ + cardini + + ++ ++ + ++ ++ rubrifrons group parachrogaster + ++ ++ + ++ + uninubes ++ + +++ ++ ++ ++ +++ macroplera group submacroptera + + ++ ++ + +++ + pallidipennis group pallidipennis + + + + + + + ++ tripunclata group al bi cans + + + + + +++ ++ metzii + + ++ ++ + +++ +++ unipunctata ++ + ++ +++ ++ ++ ++ m ediopunctata + + ++ ++ + + + ++ tripunctata + + ++ ++ + ++ +++ trapeza + + ++ ++ + ++ + + bandeirantorum + + +++ ++ + + ++ m ediostriata + + +++ ++ + ++ + + mediopictoides + + ++ ++ + +++ + croc1na + + +++ ++ + ++ ++ ungrouped species near the lripunctata group sticta + + ++ + + +++ +++ species not placed in section aracea + + ++ ++ + + ++ tumiditarsus + + ++ + ++ + ++ Genus: Chymomyza aldrichi ++ + + + + + procnemis + + ++ + + Genus: Scaplomyza adusta + + + ++ + + + pallida + + ++ +++ + ++ ++ Genus: Zaprionus ghesquierei + + + + ++ vittiger + + + + + Genus: Mycodrosophila dimidiata + + +++ + +++ +++ were present, for example, in Sophophoran species. The general indication from the available data is that drosopterin content of the eyes varies in parallel with that of the other pteridines. Aside from overall quantitative levels, interspecific variation in pteridine ac­cumulations in the eyes is slight among species in the genus Drosophila and among those species from related genera which are included in Table 2. The Throckmorton: Biochemistry and Taxonomy same pteridines are accumulated in approximately the same relative amounts in all species. Insofar as the present methods may indicate, variations of pteridine accumulation in the eyes of Drosophila species are not of a type which would make them useful as taxonomic characteristics within the genus. The greater uniformity of accumulation among closely related species (e.g., species of the obscura group, species of the virilis group, etc.) probably reflects the fact that more nearly comparable samples were obtained from these species. Body weight is not a completely adequate indication of eye size, with the body weight increas­ ing somewhat more rapidly than eye area if one compares individuals from a small species with those from a much larger species. Eye sizes among closely related species were more nearly the same, the samples obtained from them were more nearly comparable, and the results were more uniform. Thus, it seems probable that more critical methods for determining sample sizes of eyes, perhaps using the area of the eye or the number of ommatidia, would allow the demon­ stration of significant differences in quantitative levels but more nearly uniform relative accumulations of the different pteridines. It will be of interest to explore this aspect of pteridine accumulation through more detailed studies. Only limited samples were available from species in other genera in the family Drosophilidae. Results from those most clos.ely related to the genus Drosophila are given in Table 2, and their patterns of pteridine accumulation are substan­ tially as are seen from species in the genus Drosophila. As indicated elsewhere (see Throckmorton, this Bulletin, and Figure 4A), these genera appear to have arisen from the basic Drosophila stem populations, and their near identity in pteridine accumulations is not unexpected. In addition to these, samples have been obtained from species in several other genera. The major features of these are shown in the top li.ne of Figure 3. Species from two genera, Gitona and Leuco­ phenga, differ sharpl)i from other genera, and from each other, in their relative accumulations of the major pteridines in the eyes. Species in the genus Amiota show approximately the same relative accumulations, but they accumulate all pteridines in the eyes at much higher levels than are seen from species in the genus Drosophila. Species from the remainder of the genera sampled show pat­ terns of pteridine accumulation which are directly comparable to those from the genus Drosophila. Although this sample from the other genera is too limited to allow drawing conclusions regarding the characteristic patterns for these genera, the results do indicate variations of a type which may be of considerable useful­ ness for classification at the high taxonomic levels. They also suggest that the metabolic.systems involved in the production and accumulation of pteridines in the eyes are considerably more amenable to change than might be indicated by the data from the genus Drosophila. Very few samples were available from species outside of the family Droso­philidae. Among the Acalypterate Diptera, representatives from only fifteen other families (twenty-four genera) have been examined. Except for the absence of drosopterins, many species from these families showed relative accumulations . of known pteridines which were quite comparable to those seen from the family Drosophilidae (see Figure 3). Drosopterins have been seen from only one non­Drosophilid species among the Acalypteratae. This species was in the family Diastatidae, and since the Diastatidae and Drosophilidae are closely related, such The University of Texas Publication Amiot a Drosophila, Chymomyzo, Gitono Leucophengo (2 species) Scoptomyzo, Zoprionus, (I species) (3 species) Mycodrosophilo, Poro ­mycodrosophilo, ZygothricoDrosophilidae Microdroso phi lo . ~11111F:Ja .& Periscelidoe Agromyzidoe Otitidoe Sepsidoe (I genus) (2genero) (3 genera) (2 genera) Ephydridoe ChloropidoeAnthomyz idoe Ephydridoe (I genus) (3 genera) (I genus) (I genus) Aca lypteratae .JIIh I a ~ la ORO ISO XAN SP AHP BIO Conopidoe Bombylii doe Bombyliidoe (I genus) (I genus) (I genus) Legend Non-aca lypterate FIG. 3. Representative accumulations of the known pteridines in the eyes of species from several families of Diptera. Abbreviations are listed in Table 1. a resemblance is not surprising. One instance, however, of drosopterin accumu­lation in a non-acalypterate species was noted. This was in a species from the genus Phthiria in the family Bombyliidae. Another species from this family (genus Sparnopolius) did not accumulate the drosopterins, so here there would appear to be an intra-family difference in this respect. A great variation in accumulation of pteridines was seen among the non-drosophilid genera, but there was a general tendency for genera within the same family to exhibit very similar patterns of accumulation. There was also a rather general clustering of families into larger groups, and more extensive . studies of the accumulations of the fluo­rescing compounds in the eyes of dipteran species should provide unusually fruit­ful material for taxonomic investigation. Throckmorton: Biochemistry and Taxonomy Aside from the known and presumed pteridines, two other compounds are of interest. These are xanthurenic acid (XIC) and the unknown compound, BL Indications are that this latter compound is a kynurenine derivative (Hatfield, unpublished), and since xanthurenic acid is also a product of tryptophane metab­olism (e.g., see Wagner and Mitchell, 1955), it is possible that both of these compounds are involved in the metabolism of the brown eye pigments. Within the genus Drosophila the B1 compound has been identified with certainty only from species belonging to the saltans-willistoni branch of the subgenus Sopho­phora. It may also be present at very low levels in ananassae (melanogaster group). If it is present in other species its accumulations are so small as to be undetectable by the present methods. Accumulations of this compound, there­fore, are good indications of taxonomic relationship, at least among the species included in this investigation. Outside of the genus Drosophila the B 1 compound is seen at trace levels in some species from the three genera, Chymomyza, Micro­drosophila and Leucophenga, and at low levels in species from the genus Zapri­onus. As may be seen from Table 2, xanthurenic acid accumulates in the eyes of most species from the genus. Exceptions are some species in the subgenus Phola­doris and many species from the virilis-repleta section of the subgenus Drosophila. Morphological evidence (see Throckmorton, this Bulletin) suggests the possibility that the virilis-repleta section might be subdivided into two branches, with species from the virilis, melanica, robusta and annulimana groups belonging to one branch, and species from the remainder of the groups in this section belonging to the other. On this basis, species of the virilis branch of this section fail to accumulate xanthurenic acid and those from the other branch are variable in this respect. Species from the quinaria section of the subgenus Drosophila uni­formly accumulate this compound, although some, e.g., species from the funebris group, accumulate it only at low levels. Species from other genera in the family Drosophilidae uniformly accumulate xanthurenic acid, although some, e.g., Gitona, appear to accumulate it only at trace levels. Accumulations of xanthu­renic acid are also seen in the eyes of species from most, but not all, of the non­Drosophilid families iri.vestigated. Results from chromatograms of bodies and electrophoresis of testes-Accumu­ lations of fluorescing compounds and uric acid in the decapitated bodies are shown in Table 3. Here, marked differences in the accumulations of known pteridines are seen among the different species. Since visual inspection indicates that pteridine pigments accumulate in the testis sheath, it seemed probable that testis-accumulations made the major contribution to the patterns of fluorescing compounds seen from chromatograms of the bodies. The results of the prelimi­nary survey (Hubby and Throckmorton, 1960) suggested that the primary change of pteridine metabolism with evolution in the genus had involved an organ-specific divergence, with major changes occurring in the pteridine metab­olism of pteridines and only minor changes occurring in the pteridine metabolism of the eyes. It was therefore desirable that data be obtained directly from the testes themselves, and this proved to be feasible through the use of paper electro­phoresis. Paper chromatograms of testis material gave uniformly poor results. The results from electrophoresis of the testes are given in Table 4. Most of the The University of Texas Publication important pteridines (SP, ISO, BIO, AHP, DRO) could be estimated reliably from both types of analysis. Since the electrophoresis samples included 3 X as much testis material as did those for the chromatograms of bodies, it was possible to detect the presence of certain compounds with much greater sensitivity by this method. Thus, drosopterins were found to be present in the testes of many more species than had been indicated by the results of chromatograms. One compound, BL, could be identified with certainty as different from any seen on chromatograms. Since it was often present in rather large amounts, lack of de­tection on the chromatograms was not solely due to the larger samples used for electrophoresis. It may have been masked by more strongly fluorescing com­pounds on the chromatograms, but time did not permit investigating its chro.­matographic behavior in any detail. Sufficient checks were made to be certain that it caused no difficulty in determining the amounts of biopterin and 2camino­4-hydroxypteridine on the chromatograms, and, tentatively, it seems to have been masked there by isoxanthopterin and sepiapteridine. The spot designated as YG may have been a mixture of several compounds, among them the xanthop­terin-like compounds. The spot designated as GR was generally present at such low levels as to have been undetectable on chromatograms. It was also a com­pound spot, occasionally separating into two spots of identical fluorescent ap­pearance in some species. As a general rule, the testes appear to contain the major accumulations of pteridines from the bodies, and most of those not accounted for by the testes can be attributed to accumulations in the Malpighian tubules (see later). Perhaps the major exception to this is seen in the accumulations of 2-amino-4-hydroxypteri­dine: The accumulations of this compound in the testes and Malpighian tubules often make up only a minor fraction of that seen from the decapitated bodies. This can be seen, for example, by comparison of accumulations of AHP seen from chromatography and electrophoresis of novopaca (coracina group of the subgenus Pholadoris, Tables 3 and 4). Presumably the fat bodies and hemolymph contain quantities of this compound. Minor accumulations of other compounds outside of the testes and Malpighian tubules would not be detectable by the present methods. Quantitative estimates of testes accumulations of pteridines encounter some­what the same difficulties as do estimates of accumulations in the eyes. Body weight need not be a completely accurate indication of testis size. In addition, shape of the testes differs among these species, with a few having elliptical testes and the majority having spirally coiled testes. There is also a considerable inter­specific difference in the length of the coiled testes. Since the pteridine pigments at least appear to accumulate only in the testis sheath, sample sizes might be determined more realistically if they were based on surface area of the testes. Such determinations were impossible for a survey of this type. Visual observa­tions of testis pigmentation (intensity of the red or yellow color of the testis sheath) indicate that quantitative estimates of the responsible pigments (DRO and SP) from chromatography or electrophoresis samples correlate very closel,y with the observed color of the testis. Thus, red testes show the highest amounts of drosopterins following electrophoresis, lemon-yellow testes have the highest accumulations of sepiapteridine and little or no accumulation of drosopterins, Throckmorton: Biochemistry and Taxonomy etc. Since greater significance is to be placed on changes in relative accumula­tions of the different compounds than on general quantitative levels, samples based on body weight appear to be adequate for the present purposes. Here, as for the eyes, closely related species have nearly comparable testis sizes, so in most cases the significant differences can be attributed to changes in pteridine metabolism rather than to errors introduced by methods of determining sample size. Phylogenetic relationships suggested by analysis of the evolutionary change of morphological features will be used as the framework for the organization of the descriptions of testes accumulations of fluorescing compounds. On this basis, species belonging to the subgenus Pholadoris are among the most primitive, i.e., separated earliest from the major evolutionary line, in the genus (see Figure 4A). All of these species included in this survey have elliptical testes, although they differ considerably in color of the testis sheath. Species in the victoria group of the subgenus Pholadoris have testes which are black-or brown-pigmented. Oc­casionally, especially in immature individuals, testis color may approach an orange-brown, particularly when viewed through the body wall. In these cases, visual inspection cannot reliably determine whether or not drosopterins are present in the testes. Species from several other Drosophilid genera (e.g., Gitona, Rhinoleucophenga) have testes of this indeterminate color, and electrophoresis shows that they do not accumulate drosopterins in the testes. Species from most other families of Diptera have testes which are heavily pigmented and either black or brown in color. Exceptions to this have thus far been seen only in species from the family Chloropidae. Here, some species have testes which are a bright lemon-yellow and which accumulate large quantities of sepiapteridine. Other species from this family have the more typical, dark-pigmented testes. Tenta­tively, dark-pigmented testes can be considered as the general condition among the Acalypteratae, and the presence of forms having black testes among the species from the subgenus Pholadoris tends to confirm the primitive position of these species within the genus Drosophila. Few pteridines are accumulated, either in the bodies or specifically in the testes of species from the victoria group. Only isoxanthopterin among the known pteri­dines is present in the testes (Table 4) . Both isoxanthopterin and the xanthop­terin-like compounds are present at trace levels from the decapitated bodies (Table 3). The green compound (GR) may be one of the xanthopterin-like com­pounds, but this seems doubtful since these compounds are seen from most species in the genus and the green compound only from a more limited group. Most of the riboflavin-like material visible on chromatograms of bodies is contributed by the Malpighian tubules, but a small amount of it is present in the testes. In species such as victoria_ which do not accumulate sepiapteridine in the testes, the pres­ence of the RFL compounds could be detected in testes samples. However, this material was usually obscured by sepiapteridine on the electrophoresis samples and so was not included in the tabulation of the data. Species from the coracina group of the subgenus Pholadoris have red testes and accumulate large quantities of drosopterins. They also accumulate traces of 2-amino-4-hydroxypteridine and bi<;>pterin in addition to isoxanthopterin. D. latifasciaeformis has yellow testes, but it accumulates only isoxanthopterin and traces of biopterin. The yellow color The University of Texas Publication might be attributable in part to the presence of the RFL compounds, but another yellow-fluorescing compound appears to be present (remaining at the point of application of the sample) and this probably contributes most to the color of the testes in this species. With respect to the known pteridines, species from the subgenus Sophophora exhibit three general patterns of accumulations in the testes: with drosopterins at high levels, drosopterins at trace levels and sepiapteridine at high levels, and low levels of all pteridines. In addition, chromatograms of bodies show several com­pounds which appear to have a restricted distribution among these species. Ac­cumulations of the known pteridines are relatively constant among species from the obscura group, although species from the obscura subgroup accumulate most of them in lower amounts ( drosopterins excepted) than do species from the affinis subgroup. Within the affmis subgroup, the increase in accumulations roughly fol. lows the increase in number of testis coils and so may be a reflection of sample size. The general change in relative accumulations of these compounds suggests that a decrease in drosopterin accumulation has been correlated with an increase in the accumulation of the other pteridines, but the present methods do not permit such discriminations at this level of difference. With respect to the electrophoresis samples from species of the obscura group and from other species accumulating drosopterins in the testes, it may be noted that a dark, non-fluorescing substance often was seen obscuring the first of the four drosopterins which were visible. This is shown as a stippled area in Figure 2. This susbtance was not always present, but it was visible in near maximum amounts in samples of the testes from species of the obscura group. It was present only at low levels in samples from species belonging to the coracina group of the subgenus Pholadoris. It may be the black pigment observed in the testes of species from the victoria group, but if so its migration in the electric field must depend on the presence of some carrier compound such as, for example, a drosopterin. Elec­trophoresis samples from species in the victoria group do not show the presence of this compound. The dark pigments in these cases remained as a dense, black spot at the point of application of the sample. On chromatograms of bodies, several interesting differences are seen between the accumulations in species of the obscura subgroup and those of the affinis sub­group. Species of the obscura subgroup accumulate a compound designated as G, and this is apparently the same compound that is accumulated by lativittata and cancellata of the coracina group. On the other hand, species in the affinis subgroup accumulate two distinctive compounds, G3 and V. The G3 compound is seen only in these species, and Vis seen elsewhere only in carbonaria. The compound desig­nated as G2 may be present in species from both the obscura and affinis subgroups and obscured by the drosopterins. However, G2 could be detected in species of the virilis group having comparable accumulations of drosopterins, so these species of the obscura group would appear to accumulate G2 at low levels, if at all. Most species from the melanogaster, saltans and willistoni groups accumulate drosopterins in the testes only at very low levels, and for many of these species no drosopterin accumulations are detectable. In most cases, levels of drosopterin ac­cumulation were so low that they could not be detected chromatographically, and only traces of them were visible on the electrophoresis samples. All species of the Throckmorton: Biochemistry and Taxonomy melanogaster group showed these traces and the relative accumulations of the other compounds were much the same in all species. Quantitatively, the species in this group might be arranged in series, with rufa showing the highest levels of accumulation and takahashii and nikananu showing the lowest. Here, and throughout most of the genus (affinis subgroup species a possible exception), amounts of compound A and G2 roughly parallel those of sepiapteridine. Neither of these two compounds accumulates to the extent that sepiapteridine does, how­ever. Amounts of 2-amino-4-hydroxypteridine and isoxanthopterin tend to be higher in species of the montium subgroup when samples from the bodies are considered, but this relationship does not hold for the testes themselves. A greater uniformity is seen among these species when testes samples only are considered and this probably indicates that testes metabolism of pteridines is relatively similar among these species. lnterspecific differences in pteridine metabolism elsewhere in the bodies may exist, however. Although the range in quantitative levels is greater, variation among species in the saltans group is similar to that seen among species in the melanogaster group. In species of the parasaltans subgroup and in septentriosaltans, pteridine accumu­lations have been reduced to very low levels, and the testes have a white, or nearly transparent, appearance. Only about half of these species show traces of drosop­terins in the testes, and these are among the species which accumulate the other pteridines at higher levels. Species in the willistoni group show approximately the same range of distribution of testes accumulations as is seen in species of the saltans group, although a greater proportion of these species accumulate the compounds at the lower levels~ (D. neoala,P.tans, H410.7 from Kenscoff, Haiti, is included among these species. It was not listed among the species in the appen­dix to the earlier paper.) Pteridine accumulations in species from the subgenus Drosophila are quite varied. In many cases, patterns of accumulation similar to those seen in Sopho­phoran species are evident. Others are very different. This subgenus may be di­vided roughly into two major phyletic branches, one of them constituted of spe­cies assigned to the virilis-repleta section, the other of species assigned to the quinaria section. Within the virilis-repleta section, the virilis group is derived earliest and occupies a somewhat isolated position. Its species show very uniform accumulations of pteridines. Those of virilis itself may be somewhat reduced relative to accumulations seen in other species of the group, but the difference is slight. Aside from somewhat higher quantitative levels, which may be attribu­table to sample size, accumulations of pteridines in these species are roughly com­parable to those seen in species of the obscura group. The testes are orange in color and both the drosopterins and sepiapteridine accumulate in large amounts. The relative accumulations of 2-amino-4-hydroxpyteridine and biopterin are similar in these two groups, and just the reverse of those seen in species of the saltans, willistoni and melanogaster groups. In these latter groups, biopterin ac­cumulates in the testes at much higher levels than does 2-amino-4-hydroxypteri­dine, and the drosopterins are present only at low levels, if at all. In species from the virilis and obscura groups, drosopterins are accumulated in the testes at very high levels, and 2-amino-4-hydroxypteridine accumulates in greater quantities relative to biopterin. Thus, biopterin levels appear to increase and AHP levels The University of Texas Publication appear to decrease when drosopterin accumulations in the testes are reduced. This general relationship does not appear to hold true for all groups in the genus, however. All species in the virilis group accumulate an unknown compound, YI, in the bodies. The major site of accumulation of this compound appears to be in the Malpighian tubules, but it has been seen occasionally from electrophoresis samples of testes. Ifdrosopterins are present in the testes, however, this compound could not be detected reliably in the electrophoresis samples and so is not indi­cated with that data. All species in the robusta group accumulate drosopterins in the testes. If data from chromatograms of the bodies are considered, lacertosa and colorata would appear to be very similar to species in the virilis group. However, when testes samples are considered, differences are apparent. Both accumulate the unknown compound, BL, in greater quantities, and the relative accumulations of BIO and AHP differ in the two species. Accumulations of known pteridines are low in robusta, although the drosopterins are present at relatively high levels. Generally, pteridine accumulations in robusta are quite similar to those seen, for example, in cancellata of the subgenus Pholadoris. The difference, trace amounts of SP in the robusta sample, may be due to the fact that robusta has proportionally larger testes than cancellata, and thus sample sizes were not quite comparable. Species of the melanica group also accumulate drosopterins in the testes, but there is considerable intragroup variation in this respect. The accumulations of these compounds are highest in nigromelanica and lowest in euronotus. Relative accumulations of the different fluorescing compounds vary considerably from species to species when chromatographic data is considered, but a relative con­stancy is seen from the testes samples themselves. The quantitative differences in accumulations of the drosopterins correlate closely with visual observations of the color of the testes, so they appear to indicate real differences in accumulation rather than differences in sample size. The unknown compound(s), GR, seen in species of the victoria group and in species of the affinis subgroup may be present in some of these species. D. gibberosa, of the annulimana group, does not accumulate detectable amounts of drosopterins in the testes, and its general pattern of accumulation is that seen in many of the remaining species in this section of the subgenus Dro­sophila. The chief difference between these accumulations and those seen earlier lies in the relatively high accumulations of the compounds designated as BL and YG. Here, as elsewhere, such differences might reflect greater amounts of testis material in the sample from these species. However, great increases in the amount of material in the samples from selected species, e.g., lusaltans, which appear to accumulate low amounts of these compounds in the standard sample, do not pro­duce marked increases in the apparent accumulations of these compounds. Thus, sample size does not appear to be wholly responsible for differences in accumula­tion of the types seen here. The problem still requires further investigation, but it may be assumed that these represent real differences. Among the remainder of species in this section, variation of accumulations in the bodies and testes are roughly similar to those seen in species of the saltans, willistoni and melanogaster groups. One of the major differences, that involving the compounds BL and YG, has already been mentioned. The other difference Throckmorton: Biochemistry and Taxonomy is that the relative accumulations of biopterin and 2-amino-4-hydroxypteridine in the testes do not show the same constancy here as they did, for example, in the Sophophoran species. As mentioned earlier, most Sophophorans accumulate higher amounts of biopterin than of 2-amino-4-hydroxypteridine in the testes. Many of the species in this section accumulate both of these compounds at high levels and in relatively similar amounts. Some, however, as for example the species of the fasciola subgroup of the repleta group, appear to follow the general Sophophoran pattern. Species in the repleta group show substantially the com­plete range of variation seen in this section. D. aldrichi and buzzatii both accumu­late drosopterins at relatively high levels, others accumulate them at intermedi­ate or low levels, and many show no detectable accumulations of these com­pounds. The general level of accumulation of other fluorescing compounds is so high in most of these species that correlated changes in relative accumulations are not detectable. It is true that sepiapteridine accumulates in almost excessive amounts in many of the species which show reduced accumulations of drosop­terins, for example, but similar high accumulations of sepiapteridine are also seen in those species which do accumulate drosopterins, and the inference that a reduction in accumulation of drosopterins was accompanied by an increase in accumulation of sepiapteridine and/or other pteridines cannot be drawn from the present data. A reduction in sample size might allow such quantitative dis­crimination, however. Here, as for the Sophophorans, certain groups or species show strongly reduced accumulations of pteridines in the testes. Species of the dreyfusi group, for example, accumulate the pteridines in the testes at quite low levels. Many species of the fasciola subgroup of the repleta group (e.g., moiuo­ides) do also. The pattern of reduction here, however, does not seem to be quite the same as that seen in Sophophoran species. There, traces of drosopterins were observed when other pteridines were accumulated at relatively high levels, but they were absent when accumulations of other pteridines dropped to low levels. Among the species of the repleta section, many species accumulate other pteri­dines at very high levels but do not show any traces of the drosopterins. In others (e.g., species of the dreyfusi group, species of the fasciola subgroup, etc.) accumu­lations of sepiapteridine and the other pteridines have dropped to low levels, but traces of drosopterins are still very evident. Other instances (e.g., species of the canalinea group, or limensis of the repleta group) appear to follow the Sopho­ phoran pattern. Species from the quinaria section likewise show a broad range in testis accumu­ lation of pteridines, and there is often a considerable amount of variation within any one group. Within this section the immigrans group was probably derived earliest, either from the very base of the quinaria branch or from an earlier stem population for the subgenus as a whole. One species, hyf)Ocausta, accumulates drosopterins in large quantities. The testes of this species are a dirty brown in color, and the high accumulation of drosopterins was not suggested by visual inspection. In general appearance, these testes somewhat resembled in color the testes of species from the victoria group since they seemed to have large quantities of a brown pigment present. Following electrophoresis of the testes sample, the obscuring brown substance covering the first of the drosopterins (See Figure 2) was present in large amounts. Sepiapteridine and biopterin also accumulate in The University of Texas Publication these testes at high levels, but isoxanthopterin is present only in trace amounts. As a general rule, in other species, accumulations of isoxanthopterin only de­crease when the amounts of other pteridines reach their lower levels, and this species appears to be an unusual exception. Other species in this group also ac­cumulate drosopterins, and their general patterns of accumulation of the other pteridines are similar to those seen elsewhere. Species in the funebris group ac­cumulate pteridines in the testes in much the same patterns as do those of the im­migrans group. Here, traces of drosopterins are still seen when amounts of the other pteridines have dropped to the lower levels. In both the immigrans and funebris groups, and generally throughout this section, the unknown blue com­pound (BL) is present only at low levels, even when the other compounds are present in large amounts. The unknown, YG, may or may not be present at high levels. Species from the quinaria group show almost the complete range of variation seen among species from this section. Some species, e.g., occidentalis and tene­brosa, accumulate drosopterins and other pteridines at relatively high levels. Others, e.g., rellima, accumulate the pteridines at low levels, but they still show traces of the drosopterins. In this respect, the pattern somewhat resembles that seen in some species of the virilis-repleta section (e.g., species of the dreyfusi group) although species of the quinaria group accumulate the YG rather than the BL compound. Species of the calloptera group are distinctive in that they ap­pear to have eliminated the accumulations of all pteridines except biopterin from the testes. Drosopterins may have been present, but if so their level of accumula­tion was such as to make their identification uncertain. The remainder of species from this section have much the same patterns of pteridine accumulation. The major · differences are at a general quantitative level. All show strongly reduced accumulations of pteridines in the testes. Generally, species in this section appear to accumulate relatively greater quantities of biopterin than of 2-amino-4-hy­droxypteridine in the testes. Aside from the species in the calloptera group, the most extreme reduction in pteridine accumulations are seen in some species (dunni, etc.) of the cardini group. Here only isoxanthopt~rin and traces of biop­terin remain. Chromatograms of species in the cardini group are distinctive in that they show accumulations of the kynurenine-like (KYL) compound in the bodies. There is some quantitative variation in the level of accumulation, with cardini, procardinoides and parthenogenetica showing only trace amounts and neomorpha, polymorpha and acutilabella accumulating great quantities. The other species in the group accumulate this compound in intermediate amounts. The species not placed in a section follow one or another of the patterns seen from other species in the subgenus, although the pattern of carbonaria is basically more similar to those seen from species in the obscura group of the subgenus Sophophora. This resemblance is further enhanced by the presence of the un­known violet compound (V) on the chromatograms of this species. Morphological evidence suggests that carbonaria was derived from an early stem population for the subgenus Drosophila, but its biochemical distinctness within the subgenus, and its strong similarity to Sophophoran species is still rather surprising. Patterns from species in other subgenera, e.g., Dorsilopha and Hirtodrosophila, show nothing of a very distinctive nature. Their patterns of pteridine accumula­tion are generally of types seen among species from the subgenus Drosophila, but such similarities are not marked. Species from two other subgenera, Phloridosa and Siphlodora, have also been examined chromatographically. Material was not available for testes samples from these species. The single species from the sub­genus Phloridosa (collected by Dr. A. Faberge in Chiapas, Mexico) had a chro~ matographic pattern identical with that tabulated for lacertosa of the robusta group. That of fiexa (H410.3, from Kenscoff, Haiti) was identical with that tabu­lated for immigrans. Testes samples have not been taken from species in other genera. Chromatographically (see Table 3) species from the four genera, Chy­momyza, Scaptomyza, Zaprionus and Mycodrosophila are generally of Dro­sophila types. Aside from the distribution of the different pteridines, the distribution of the Y1 compound is of interest. This compound is seen on the chromatograms of most species from the subgenus Drosophila, but many groups are variable in this re­spect. It has not been seen from chromatograms of species in the subgenus Phola­doris or in the subgenus Sophophora. This latter observation is in some doubt, however, since the Y1 compound is seen from samples of Malpighian tubules from several species in the melanogaster group. Since this compound could be detected more reliably following electrophoresis of the Malpighian tubules, its distribution will be discussed with that material. Results from the electrophoresis of Malpighian tubules-Only selected species have been used for investigating the accumulations of fluorescing compounds in Malpighian tubules. The results of this analysis are listed in Table 5. Satisfactory chromatograms may be obtained from samples of the Malpighian tubules, and a limited number of chromatographic samples were run in order to confirm the identification of certain of the compounds seen from the Malpighian tubules. Ac­cumulations of the riboflavin-like compounds (RFL) vary considerably from species to species and show no particular group correlations. Isoxanthopterin i~ almost uniformly present in trace amounts. The BL compound is variable and it is generally present when Y1 is present in large amounts. The amounts of the BG and GR compounds are also variable and follow no apparent pattern. Xanthu­renic acid (XIC) is present in all samples from species in the subgenus Sopho­phora. It is also _present in the Hirtodrosophila and Dorsilopha. It is variable in species from the subgenus Pholadoris and in species from the virilis-repleta sec­tion of the subgenus Drosophila. Except for being present in species of the immi­grans group, it is absent in all samples from species belonging to the quinaria section of the subgenus Drosophila. The compound, Y1, appears to be absent from species in the subgenus Pholadoris and from most Sophophorans. Species from the montium subgroup are an exception. Most, but not all, of the species in the sub­genus Drosophila accumulate Y1 in the Malpighian tubules. Insofar as may be judged by this limited sample, the accumulation of Y1 in adult Malpighian tu­bules follows much the same distribution as is seen for the same compound in samples from larvae. It seems probable that a more complete sampling of the Malpighian tubules from Drosophila species would provide useful information for taxonomic purposes. Results from chromatograms of larvae-The larvae of most species provide ex­ cellent material for chromatographic purposes. In most cases the number of com­ The University of Texas Publication pounds seen is small, the spots are compact, and the characteristics seen are very constant on repetition. Both propanol-ammonia and butanol-acetic chromato:. grams have been prepared from larvae. Most of the more useful informationwas present on the butanol-acetic chromatograms, and only this data has been re­corded in Table 6. Information from the propanol-ammonia chromatograms is of interest, however. As was mentioned, the number of fluorescing compounds seen from the larvae is generally small. However, the propanol-ammonia chromato­grams of the larvae of many Sophophoran species show an array of pteridines similar to that seen from the testes of adult males. In particular, many of these larvae show accumulations of sepiapteridine, 2-amino-4-hydroxypteridine and biopterin, as well as increased accumulations of isoxanthopterin. Those Sopho­phoran species in Table 6 which accumulate the higher amounts of isoxanthop­terin (++) are the ones which also show accumulations of the other pteridines. Several species in the saltans group are most conspicuous in this respect. These species also have larval fat bodies which are visibly pigmented yellow, and it seems probable that the fat bodies are the site of this unusual accumulation of pteridines in the larvae. There is no evidence that the fat bodies of these species retain the capacity to accumulate pteridines in the adult stage, although they may do so to a limited extent. For most compounds from larvae, no distinct pattern of distribution is seen. There is a general tendency for species from the subgenus Drosophila to accumu­late uric acid at low levels. Most Hirtodrosophila and Dorsilopha also accu­mulate uric acid at low levels. Most Sophophoran species and species from the subgenus Pholadoris accumulate uric acid at relatively high levels. The distribu­tion of the kynurenine-like compound (KYL) and Y1 is generally in accord with taxonomic grouping. Thus, KYL accumulates in most, but not all, Sophophoran species, in species from the subgenus Pholadoris, from the subgenus Hirtodro­sophila and from the subgenus Dorsilopha. It is variably present in species from the virilis-repleta section of the subgenus Drosophila, and present in all except the immigrans group of the quinaria section. Itis variable among species from the genus Chymomyza, present in species from the genus Scaptomyza, absent in species from the genus Zaprionus, and present in Mycodrosophila. D1scuss10N Most of the problems encountered when dealing with biochemical features are similar to those met with during the interpretation of morphological ones. Per­haps the chief problem with both types of material is that of identifying the characteristic. Morphological features may be distinctive enough so that they can be identified readily through examination of gross anatomy. Often, however, particularly between distantly related species, sufficient differences exist so that more detailed studies into the ontogenetic and phylogenetic development of a structure are needed to obtain information required to substantiate the inference of homology. Superficially, investigations of biochemical characteristics, especi­aJly those involving the presence or absence of a specific compound, may not ap­pear to meet the problem of identity at quite the same level as do investigations of morphological characteristics. Such accurate methods of chemical analysis exist that the identity of many compounds need not long remain in doubt, and it is this Throckmorton: Biochemistry and Taxonomy feature (precise identification) of biochemical characteristics which makes them seem attractive as indications of relationship. However, the appearance of a given compound may depend both on the existence of mechanisms for its biosynthesis and on the existence of mechanisms for its accumulation. Since several alternate pathways of biosynthesis may be possible, and since mechanisms of accumulation may be varied, the potentiality for convergent evolution is at least as great for biochemical characteristics as for morphological ones. Thus, for biochemical characteristics, the identity upon which taxonomic inference depends is the iden­tity or homology of the metabolic systems responsible for the existence and ac­cumulation of certain compounds or groups of compounds. The chemical identity of a specific compound may be only a minor part of the more directly relevant problem of homology at the developmental and genotypic level. Biochemical and morphological characteristics can only be given equal weight in taxonomic analy­sis when both reflect substantial homology of the respective metabolic and/or developmental systems involved. The validation of morphological characteristics as indications of homologous developmental systems, and hence of homologous genotypes, has been accom­plished through detailed studies of comparative anatomy dating back to the time of Cuvier. A tremendous body of information has been derived from such studies, and the results of these studies have had sufficiently general applicability as to make unnecessary (but certainly not undesirable) exhaustive studies of similar types for each new species or new characteristic investigated. No comparable body of information relating to the dependability of biochemical characteristics as indications of homologous metabolic systems exists at the present time. The problem of the validation of biochemical characteristics may be ap­proached in two ways. The first of these assumes that systems governing bio­chemical characteristics will have had approximately the same evolutionary sta­bility as those determining morphological characteristics. The distributions of biochemical characteristics among related species can be determined and com­pared with those of morphological characteristics of known taxonomic value. If both morphological and biochemical characteristics follow the same general pat­terns of distribution, it may be assumed that the evolutionary stabilities of both types of systems are roughly comparable and both types of characteristics have approximately equal taxonomic values. Thus, if biochemical characteristics show the same evolutionary behavior as morphological ones, they are assumed to be valid because morphological characteristics are (generally) valid. This approach is obviously an expedient which can only give a tentative indication of the va­lidity of the biochemical characteristics dealt with. If these characteristics are to be evaluated at other than this empirical level a different approach is necessary­one which can give information for metabolic systems comparable to that ob­tained for morphological systems from studies of comparative anatomy, compara­tive embryology, etc. Such an approach will be one which involves the critical dissection of metabolic systems using methods of comparative biochemistry. Potentially, investigations of this type may carry the validation of biochemical homology to somewhere near the genotypic level (e.g., see Anfinson, 1959) and it is in this respect that comparative biochemistry may be of greatest value for the study of taxonomy and evolution. The University of Texa.s Publication Obviously, only the empirical method can be used to evaluate the results of a survey of the present type. In addition, however, the results themselves may indi­cate what possibilities exist for carrying out more detailed biochemical studies and whether or not such studies may provide information relative to more gen­eral problems of evolution. The steps in the evaluation may be phrased as three questions: 1) Do biochemical characteristics follow the same general patterns of evolu­ tion as morphological characteristics? ' 2) Does the potentiality exist for relating biochemical changes to the genetic changes which have occurred during evolution? 3) Can biochemical changes be related to an evolutionary model of suitable complexity so that information from them may be of value in studying mechanisms of evolution? The answer to the first question will constitute the empirical evaluation of the validity of the biochemical characteristic under consideration. The answers to the second and third questions can only be tentative, but they may give an indi­cation of whether or not more detailed investigations might provide worthwhile information. The taxonomic and phylogenetic distribution.s of individual compounds-In Tables 2 through 6 the accumulations of the different compounds are shown by taxonomic grouping and, generally, in phylogenetic sequence. As was indicated in the earlier descriptions, several compounds show distributions which reflect the taxonomic position of the species concerned. Some of these compounds appear to have a restricted distribution and are confined to species in a particular group or subgroup. For example, the kynurenine-like compound accumulates only in the bodies of species in the cardini group, and the compound designated as G3 ap­pears only in species of the affinis subgroup of the obscura group, etc. (Table 3). These particular compounds show a type of distribution seen for many morpho­logical characteristics. It may be assumed that the genotypes responsible for their appearance arose and became fixed or stabilized in the stem populations for the individual groups or subgroups and are thus seen in all the species derived from these stem populations. At an empirical level, then, they appear to be quite com­parable in value to morphological characteristics and might be included in a taxonomic listing of the characteristics which distinguish these groups from their neighbors. Other compounds are often restricted to a particular group or subgroup, but are also seen in some species from elsewhere in the genus. The compound G, for ex­ample, is seen in species from the obscura subgroup of the obscura group ( Sopho­phora) and in two species from the coracina group (Pholadoris). The compound designated as V is seen in all species from the affinis subgroup of the obscura group, but it is also seen in carbonaria of the subgenus Drosophila, etc. It is this latter type of distribution, consistently present (or absent) in some groups of species but variable in others, which is by far the most common for the com­pounds observed here. Itis seen for the compounds B1, XIC, KYL and Y1, for ex­ample, and also for the known pteridines. The distributions of species accumulat­ing B1 and XIC in the eyes of adults, and of those accumulating KYL and Y1 in the larvae, are shown related to the phylogeny of the genus in Figure 4. The fEmoecunu,olo tu D? 0 C. FIG. 4. A. General phylogeny of the genus (after Throckmorton, this Bulletin) and the distribution of species accumulating the compound Bl in the eyes. B. Distribution of species accumulating xanthurenic acid in the eyes. C. Distribution of species accumulating KYL in the larvae. D. Distribution of species accumulating Y1 in the larvae. For further explanation, see text. phylogeny followed here is that given in Figure 53 of Throckmorton, this Bulle­tin. The basic skeleton of the phylogeny is identical in Figures 4A through 4D, and group names are given only in Figure 4A. In these figures, if the particu­lar compound is accumulated by all species in the group, the entire branch for the group is darkened. If some species in a group accumulate the compound and others do not, the tip of the branch is shown forked, with one fork darkened and the other not. At the observational level, the constitution of the stem populations The University of Texas Publication which gave rise to the different groups is unknown, and these populations have been shaded lightly in the figure. As may be seen from Figure 4, and from Tables 2 and 6, the distributions of these four compounds tend to be uniform in certain phylogenetic branches of the genus but variable in others. Thus, for the accumu­lation of xanthurenic acid in the eyes (Figure 4B), species in the subgenus Sophophora and those from the quinaria section of the subgenus Drosophila all accumulate the compound, whereas species in the subgenus Pholadoris and in the virilis-repleta section of the subgenus Drosophila are variable. Only the species in the quinaria section uniformly accumulate KYL in the larvae (Figure 4C), and the same is true for those accumulating Y1 (Figure 4D). When distributions such as these are encountered, one of the first questions which comes to mind is whether or not the phylogeny is valid. This problem may be approached in several ways, the most obvious being to attempt to adjust phylogenetic relationships so that all species showing a given characteristic (the accumulation of XIC in the eyes, for example) are derived, either directly or in sequence, from a common ancestral population. Even using only the four charac­teristics from Figure 4 (or even any two), however, it is impossible to derive a pattern of phylogenetic relationships which will bring the distributions of all characteristics into accord. The problem becomes still more difficult if morpho­logical characteristics are considered simultaneously with the biochemical om~s. No matter what system of relationship is used, some characteristics continue to vary within certain groups. However, there may always be some doubt concern­ing the objective accuracy of a phylogeny based on morphology, and it is always possible to hope that if we only knew how to interpret the characteristics properly, all would be well. While hopeful optimism of this type cannot be com­pletely erased, it may be dimmed somewhat if the characteristics under consider­ation are related to a more objective phylogeny. This cannot be done for the entire genus, but a cytological phylogeny is available for species in the repleta group of the subgenus Drosophila. A cytological phylogeny is at least the most objective type available, and if objectivity is going to improve matters, this should be apparent in Figure 5. Figure 5 is constructed in the same manner as was Figure 4. The skeleton of the phylogeny follows Wasserman (1960, and personal communication) and the subgroup names are given only in one of the three figures shown. Since the compound, B 1, is not accumulated by any of these species (see Figure 4A), the distributions of XIC, KYL and Y1 only are shown in Figure 5. Here, as elsewhere, the distributions of these .compounds remain persistently variable in certain of the subgroups. Such patterns of distribution, however, are not peculiar to biochemical charac­teristics. Fully comparable distributions of morphological features are also seen (e.g., see Figures 14, 20, 21 , 24, 31, 32, 40, 41, 44, 46, 48 and 49 of Throckmorton, this Bulletin). Again at the empirical level, we may con.elude that biochemical and morphological features show comparable evolutionary behavior and are thus of comparable taxonomic usefulness. Their resemblances run the entire gamut from constancy to caprice, and there is little to choose between them in either respect. Taxonomic practice, of course, is adjusted to accommodate for such behavior. Taxonomic groups are generally based on complexes of characters, and the methods of interpreting and utilizing variability have been discussed, Throckmorton: Biochemistry and Taxonomy melanopalpa subQroup KYL XIC YI FIG. 5. General phylogeny of the repleta group (after Wasserman, 1960), and the distribu­tion of species accumulating XIC, KYL, and YI. For further explanation, see text. for· example, by Simpson (1961) and need not be considered here. So long as biochemical characteristics are treated as characters added to an already ex­tensive catalogue of morphological ones, they may be used as indications of taxo­nomic grouping. If attempts are made to give them over-riding significance, they will meet with the same difficulties as would be encountered if a morphological character were given undue respect. Biochemical characteristics as evidence of homologous metabolic systems­ Although biochemical characteristics show constancy within certain groups, the most conspicuous feature of their distribution is variability. The variability, how­ever, is not completely random, and there is a distinct tendency for variable groups to be related in clusters. That is, most of them appear to be related through rather immediate, common ancestral populations. The question as to the sig­nificance of such a distribution must be raised. Do the independent appearances of characteristics in such groups represent convergent evolution, i.e., the repeated origin of characteristics through independent mutation, or do they reflect some other evolutionary mechanism? The clustering of the variable species groups suggests that the answer to this question lies in some characteristic of the ancestral · populations other than a general capacity for mutation. However, if convergent evolution is the answer, it is of interest to note just what it has produced. Con­sidering only the characteristics shown in Figure 4, genotypes leading to the accumulation of B1 (Figure 1A) must have arisen independently at least three times; those responsible for lack of accumulation of XIC (Figure 4B) have arisen The University of TeXmore probable than paracentric inversions> peri­centric inversions > fusions. Translocations should be about four times as fre­quent as paracentric inversions; with a break in one rod, the probability that a second break will occur in one of the other four rods or the dot is somewhat over four times as great as the probability that it will occur in the same rod. If one or more centric fusions are present, the probability of a translocation goes to be­tween three and four times that of a paracentric when a break is in an area of a fusion. The second arm of an acrocentric chromosome is quite short and hetero­chromatic. Therefore the probability of a paracentric inversion (both breaks in the long arm) is much greater than that of a pericentric inversion (one break in the long euchromatic arm and the second in the short heterochromatic arm on the other side of the centromere) . Following the same line of reasoning, a pericentric inversion is more probable than a fusion of two acrocentric rods since the latter event requires one break in the heterochromatin of the long arm very near the centromere of one chromosome and one break in the short arm of a second chro­mosome followed by suitable combination of parts, or both breaks in the short arms followed by attachments so the two centromeres are close enough together that they can disjoin as one unit. In the derived (advanced) species where a Stone: Natural Selection and Superspecies fusion between acrocentrics has been shown to have occurred ( 3 rods, 1 V, 1 dot, etc.), the probability of occurrence in the V-shaped chromosome is equal for para­centric and pericentric inversions. The frequency of the several types of rearrangements that have been detected in the genus Drosophila is very different from the relative probable frequencies on random breakage and rejoining to produce rearrangement. It is clearly sig­nificant that there have been only four good cases of heterozygous translocations detected in populations of Drosophila (ignoring the several cases in Drosophila melanogaster which has been studied so extensively in the laboratory): one in Drosophila ananassae by Dobzhansky and Dreyfus ( 1943), a second in the same species by Freire-Maia (see discussion of Stone, 1955), a third in Drosophila melanica by Ward (1952), and a fourth in Drosophila prosaltans by Dobzhansky and Pavan ( 1943). The latter is an interesting case of whole arm exchange. Dobzhansky and Tan (1936) suggested that several small translocations had occurred to produce the cytological differences between Drosophila miranda and Drosophila pseudoobscura. This is the first case reported of translocation differ­ences fixed between species. The second case is that of Drosophila anana.ssae, where Kaufmann (1937) and Kikkawa (1938) showed that the locus of the gene bobbed and the nucleolus organizer had been translocated from the X to the 4 (dot) chromosome but no corresponding transfer had occurred from the Y to the 4 chromosome. This system sometimes acts as a complex sex chromosome system. Hsiang (1949) showed that the dot chromosome of Drosophila tumiditarsus carries the nucleolus organizer and a mass of heterochromatin (probably trans­located from the sex chromosome with the nucleolus organizer). Thus, there are just three known cases of translocations being fixed in the evolution of species. In contrast paracentric inversions are common features of the evolution of species. At least 500 paracentric inversions (perhaps double that number) have been detected heterozygous in populations, demonstrated by hy­brid synapsis from species crosses or inferred from differences detectable by care­ful comparisons of the banding sequence in the polytene chromosomes of related species. This large number is certainly in marked contrast to the small number of translocations, although the probability of random breakage and rejoining favors translocations. The same sort of situation holds for the comparison of peri­centric inversions and fusions, see Figures 4 and 5. The majority of pericentric inversions we know about are the cases of changes from acrocentrics to V-or J­shaped chromosomes. Although a few more cases are known, the comparison given in Stone (1955) is adequate. There are 32 pericentric inversions listed, to 58 fusions (and 38 additions of heterochromatin which we cannot compare to the others on a probability basis). These are the minimum number of different cases of each kind but certainly the nearly 2: 1 majority of fusions is far from the probability predictable on random breakage and recombination of chromosome parts which would favor pericentric inversions by a large factor. The reasons for these departures from the relative frequencies of these several types of rearrangements that would be predicted from random breakage and joining is to be found in the relative selective disadvantage, in terms of differ­ential survival, of zygotes produced by heterozygotes for these several types of rearrangements in Drosophila. The majority of heterozygous translocations pro­ The University of Texas Publication duce about 50% aneuploid gametes and so produce only about half the normal number of progeny. Heterozygous pericentric inversions produce aneuploid gametes in proportion to the number of strands in each tetrad that have single or other odd-numbered exchanges within the limits of the inversion. Therefore the percent of aneuploid gametes from females (males produce normal gametes since crossingover is absent) varies from Oto 50%, 0% if no crossingover occurs be­cause of the size, position or other factors interfering with crossingover within the inversion, and increasing toward 50% as the limit if the inversion has always at least one exchange within the inverted section. Cases of pericentric inversions heterozygous in populations are known to be either those too small for appreci­able crossingover to occur inside the inversion or to be those protected from such crossingover by associated paracentric inversions, as Miller (1939) has shown for Drosophila algonquin. Stone (1949) showed that many of the females hetero­zygous for a fusion of the X and 4 chromosome in Drosophila americana ameri­cana/Drosophila americana texana heterozygotes gave a normal egg develop­ment, about 95%, as good as in the best controls. Therefore heterozygosity for a fusion per se does not lead to the production of aneuploid gametes. Sturtevant and Beadle (1936) showed that the proportion of aneuploid gametes produced by a: heterozygous paracentric inversion is determined by the frequency with which all four chromatids are involved in an odd number of exchanges within the limits of the heterozygous inversions. Therefore these two types of rearrange­ments seldom produce more aneuploid gametes than normal except that the prob­ability of survival of a very long paracentric inversion may be reduced compared to medium or short rearrangements. Perhaps, like some pericentrics, long para­centrics may survive best if protected from crossingover by other inversions. The details and references to papers by many others who have contributed to our understanding of the effects of chromosomal abnormalities are given in Patter­ son and Stone (1952). If these facts have been established earlier, why do we now wish to restate them? It is because they constitute a demonstration of the overriding importance of Darwin's (1859) Natural Selection in the evolution of this genus through millions of years, under tremendously diverse conditions. In terms of their rela­tive frequencies, the several types of chromosomal rearrangements which have been used in the cytological evolution of these species have survived in inverse proportion to their selective disadvantage (in terms of reduction in number of progeny due to the presence of the heterozygous rearrangements) rather than their probable frequency of occurrence. Fisher (1929; revised edition, 1958) demonstrated that in a large sexual population an individual mutation with no selective advantage had only about 1.5% chance of surviving 127 generations and no chance of surviving to the limit, whereas a mutation with 1%selective ad­vantage had about 2.7% chance of surviving 127 generations and just under 2% chance of surviving to the limit. Chromosomal abnormalities of the type we are discussing are unique events for, unlike gene mutations, the probability of occur­rence of any abnormality is very low and, in addition, the probability of re­currence involving precisely the same breakage points and reattachment is ex­ceedingly small. Therefore we must assume that the rearrangements that sur­vived had an initial selective advantage, perhaps because of position effect, or Stone: Natural Selection and Superspecies more probably because they were or became linked (within a limited number of generations) with mutations which had selective advantage. Wright ( 1931) also pointed out the major importance of natural selection in the evolution of populations, and he has discussed ( 1941) the special case involving the possibility of fixation of reciprocal translocations despite the reproductive disadvantage of the heterozygotes and concluded that it could occasionally occur in very small populations. He showed that if aneuploid types survived and were fertile, the probability of fixation of the translocation is appreciably greater. It is interesting to note in this connection that in two of the three cases in Drosophila where translocations were fixed, namely D. ananassae and D. tumiditarsus, the trans­locations involved chiefly heterochromatin and could survive and be fertile as aneuploids. In fact, the male of ananassae is hyperploid since it carries the locus of bobbed and the nucleolus organizer on the Y as well as having these in the pair of 4 chromosomes. There are not only no exceptions to the general ordering of expected utilization of these rearrangements in terms of their relative selective advantages, but also the least probable types to be expected to survive on the theory of natural selection, translocations, are represented in two out of three cases by the types with minimum selective disadvantage in Drosophila. Another illustration of the effectiveness of natural selection in discriminating against those rearrangements which reduce effective fertility in the heterozy­gotes is to be seen in the comparisons of the survival in natural populations of the different types of rearrangements. Four cases of heterozygous translocations are recorded; in no case could the translocation be recovered again in further samples. These represent, therefore, the detection of the occurrence of a new translocation, not the retention of a translocation heterozygote-an event which is common for paracentric inversions due to heterosis. In contrast to this both Miller (1939) and Carson and Stalker (1947) describe the presence of pericentric inversions in natural populations of Drosophila algoriquin and Drosophila robusta respectively. The latter authors report two inversions, one rare but the other common. The subspecies americana and texana differ by a fusion, and there is an overlap zone between them with free crossing extending from Texas to Vir­ginia. There is no indication that one or the other gene arrangement is being lost. In contrast to translocations, which are rare and fleeting, paracentric inver­sions heterozygous in some populations of a species are ubiquitous. Almost any Drosophila species sampled may have some or many. For example, Stone (1955) lists for Drosophila willistoni around 41 paracentric inversions ( da Cunha and Dobzhansky, 1954), for D. pseudoobscura around 22 (Dobzhansky, 1951), and for D. montana, 23 (Hsu, 1952; Moorhead, 1954). Additional inversions have been found in some of these species since the 1955 summary. Although willistoni, which is the most abundant Drosophila species known, has more inversions than any other, montana, which has now small populations scattered along water systems, also has a large number of inversions. There are several species which have no inversions recorded even though they have been sampled widely, for example, Drosophila repleta, mulleri and aldrichi. Others have had no inversions reported but have been insufficiently sampled. However, an inversion may occur occasionally even in species which are now characteristically free of them; Chino (1936) found an inversion in Drosophila virilis although many strains have been The University of Texa.s Publication examined since without finding other cases. Inversions are not characteristically fixed only in small populations. There are a large number of inversions present in the members of the successful willistoni group (Dobzhansky et al., 1950) and in the virilis group (Hsu, 1952; Moorhead, 1954; Stone, Guest and Wilson, 1960) which is characterized by fairly small populations. The largest number of rearrangements estimated is between Drosophila pseudoobscura, a large popu­lation, and D. miranda, a small one (Dobzhansky and Tan, 1936). The repleta group, which is characterized by small and scattered desert populations, usually has few inversion differences fixed between species; in fact the sibling species Drosophila mulleri, aldrichi, and wheeleri have no gross inversion differences (Wasserman, 1960). Obviously, the presence in abundance of heterozygotes or fixation of inversions is not necessarily a small population phenomenon. Paracentric inversions and phylogeny-Let us return to the unique role af­forded by paracentric inversions, making use also of other infrequent rearrange­ments, in determining the phylogeny of Dipterous forms such as Drosophila. Paracentric inversions are often found (heterozygous) within populations or as differences between some populations and between most species. Very often their frequency is not the same in the several chromosomes, presumably due to differ­ences in selective advantage in a particular set of circumstances in a species, for clustering does not necessarily involve the same chromosome in different species. They may occur as independent, included, or as overlapping inversions, and changes in sequence with overlapping inversions increases the power of this · method of analysis, see Figures 6, 7 and 8 (independent and included inversions may be able to recombine by segregation or crossingover). The polytene chromosome is a very large and highly organized structure, compared to the usual molecular dimensions even of large proteins. Bridges' (1938) revised polytene map of the X chromosome of Drosophila melanoga.ster has 1024 numbered lines and corresponds to the large uniform stretched chromo­some of 414 microns in length and from 10 to 20 microns in diameter for the most part. In comparison, the Drosophila repleta chromosomes in Figure 1 are much less stretched and are comparatively wider. The chromosomes of repleta are larger than those of melanogaster but both are small compared to those of Rhyn­chosciara angelae studied by Pavan and Breuer ( 1952, and other publications). Each species has a set of chromosomes with a unique pattern of component parts; in the most cytologically versatile there may be a number of different gene orders (banding sequences) for the same chromosome within a species, sometimes several for each chromosome. In more conservative species such as those of the repleta superspecies, there are a number of species which have only one known gene order in each chromosome; in fact, three species, isolated by cross sterility or hybrid sterility factors, have the same gene order. These differences in gene order or arrangement (inversions and fusions) allow us to establish a sequence of relationships between species. In favorable cases it allows us to determine the sequence of genes in the primitive ancestor(s) of components of species groups. It must be emphasized that the gene order is usually very different between even closely related superspecies. For example, Wassennan could not relate the gene orders between the virilis and repleta superspecies, although he could recognize the corresponding chromosomes. Stone: Natural Selection and Superspecies These highly differentiated chromosomes, which carry the genetic informa­tion, have been accumulating their uniquely organized complexity since the primitive origins of this order of insects. With some 5000 bands of different sizes, shapes and sequence arrangements, the di.stinctive parts (if of sufficient size) of a particular chromosome of a species can easily be recognized by a trained expert familiar with that species. Furthermore, the rules governing changes in order or arrangement are well known and involve two-break exchanges, with acentrics and dicentrics (unless the centromeres are placed very near together) and almost all aneuploid conditions being eliminated (very small duplications may survive). The two breaks occur at random or nearly so. Rearrangements are rare events; the probability of the same two-break rearrangements occurring more than once is quite small; the probability that one will occur twice in closely related species and that both are associated with alleles providing enough selective advantage to maintain them is vanishingly small. The probability of pattern equivalence due to any factor other than origin from the same line of descent is too small to consider. In other words, convergent evolution from distantly related lines of descent to produce the same definitive pattern of the polytene chromosomes of two species does not occur. Mutations not only occur and recur, but their effects may be simulated by mutants at other loci; the polytene chromosome is a unique, visibly organized and ordered aperiodic solid, therefore the power of the analysis using rearrangements is extraordinarily great. Sturtevant and Dobzhansky ( 1936) and especially Dobzhansky and Sturtevant (1938) discussed the role of paracentric inversions and their geometric proper­ties, illustrating from their studies of Drosophila pseudoobscura. They gave a very clear analysis of independent, included and overlapping inversions, the latter producing non-reversible novel sequences especially useful in establishing phylogenies, see Figure 6, 7 and 8, as well as the implications of complex re­arrangements with more than two breaks which have not been found persisting in natural populations (see references in Stone, 1955). In addition to the im­portant contributions made to evolutionary phylogeny from a study of Drosophila pseudoobscura as mentioned above (see Dobzhansky, 1951, for a summary and references to these and other related studies), the virilis group studied by Patter­son, Stone, Hsu and their colleagues (Patterson and Stone, 1952; Stone, 1955; Stone, Guest and Wilson, 1960; Hsu, 1952) has contributed in a major way to our understanding of cytogenetic evolution, as has the large repleta group dis­cussed by Wasserman ( 1960 and in this bulletin). In the latter articles especially, Wasserman provides diagrams showing the details of the relationships in this superspecies and to some closely related groups. He has examples showing the use of the basic repleta sequence, inverted according to some designated additions of inversions, and then indicates the breakage points of particular additional in­versions. This is supplemented by the details of the changes in sequence in given lines of descent. Before discussing the virilis group, it is best to present the limitations of the method. One limitation is that we must work with the forms we can collect and test. This means that intermediate stages may be missing, especially in old groups. Therefore complete analysis of the differences within single chromo­somes or all chromosomes may be impossible. For example, Drosophila pseudo­ The University of Texa.s Publication obscura and Drosophila miranda hybridize but there are so many differences (around 100 breaks) that there is no possibility of complete analysis (Dobzhan­sky and Tan, 1936). An even more extreme case is that of the hybrids obtained between Drosophila pseudoobscura, an American member of the obscura group, and Drosophila ambigua, a European member of that group. Buzzati-Traverso (1950) found that the chromosomes do not pair; the order of banding has been sufficiently changed by rearrangements that no effective comparison can be made. In fact some one chromosome in both the virilis and the repleta groups has, in several cases, been so changed as to be not amenable to determining the number and extent of rearrangements present. In these cases, however, the other chromo­somes gave a clear phylogeny. Inversion phylogenies are not unidirectional, although the sequence of changes from one to the next form is readily determined in favorable cases. Therefore we cannot determine ancestral forms from cytology alone with complete satisfaction. The species Drosophila virilis is placed close to the ancestral form because of its central position genetically, demonstrated by hybridization tests as well as its central position cytologically, for the americana complex evolved cytologically in one direction, while the littoralis and montana complexes evolved in still different directions from the primitive stems, Figures 9 and 10. In the repleta group, Wasserman places Xa Xb Xe 2a 2b 3b as closest cytologically to the center of the repleta group (see example, Figure 2), in part due to the fact that three closely related species groups appear to have originated from this stem. These latter are not primarily desert forms and therefore most probably split off before this series of adaptations, again indicating the primitive nature of this sequence. In this case (the same is true in the virilis group for Primitive I and D. virilis) the basic map is that drawn of Drosophila repleta by Wharton ( 1942), while the primitive stem is indicated as differing from the map by these six inversions. Evolution probably occurred from the primitive Xa Xb Xe 2a 2b 3b stem to repleta. We will use the virilis superspecies to illustrate the reality of the species group concept as an expression of evolutionary relationship through descent and to illustrate some of the limitations of the method. The cytology was worked out mainly by Patterson, Stone, Griffen, Spe:p.cer, Hughes and others, especially Hsu (1952; see Patterson and Stone, 1952, for references to earlier work), and byMoorhead (1954) and Guest (Stone, Guest and Wilson, 1960). Thephylogeny as we know it is given in Figures 9 and 10. The figures show the named species and subspecies (Drosophila americana americana and Drosophila americana texana). It illustrates four forms of Drosophila montana: basic montana, cyto­logically equivalent to Primitive IV, which gave rise to D. fiavomontana, D. lacicola and D. borealis and which occurs as a segregant in the regular montana population; the regular montana type, giant montana and Alaska-Canada mon­tana. These last three might have been designated as subspecies since they differ in size, distribution, and grouping of (shared) inversions. In addition to these named species, there are four primitive cytological types, although the last is unnecessary in that it may be cytologically, but not necessarily genetically, equivalent to the basic montana gene sequence which still persists. Three of the species, Drosophila virilis, D. ezoana, and D. littoralis exist in the Palaearctic (although virilis occurs as small populations usually associated with man in Stone: Natural Selection and Superspecies many parts of the world including the Nearctic). The other six species ( 9 forms) are Nearctic. We cannot assign the Primitives to only one region with certainty, although Primitive II gave rise to the americana, texana, novamexicana complex in the Nearctic, and Primitive IV (perhaps Primitive III to IV) gave the montana complex of species in this same region, while Primitive III gave rise to ezoana and to littoralis in the Palaearctic. This leads to one of the limitations of this method of phylogeny, namely, that we may be able to work out the directions and relationships but lack knowledge of certain intermediate steps when not enough forms survive (or are known) at present. Our difficulty in this case arises from the complexity of the inversion differences between the montana X chromosome and that of ezoana and in tum texana, the latter representing the gene sequence of Primitive IL If the ezoana and montana X chromosomes have X a b, these inversions would mean that Primitive III came from Primitive II and that Primitive I can be omitted, as an intermediate step to Primitive II, in the ancestry. We do not know if the ezoana X chromosome sequence is equivalent to Primitive II, and is incorporated in the basic montana X, although the latter is certainly ancestral to the other forms; the only possible exception is that Primitive IV may have had a simpler sequence which gave rise to the lacicola sequence, for the latter is too different from mon­tana for analysis of the X chromosome. Despite these deficiences in the record, the line and direction of change is clear, and follows the direction of the arrows, which are double headed where exchanges occur between the forms within a species. We make one assumption; Simpson (1944) presented convincing evi­dence that there are in evolution low rate lines, which often give rise to inter­mediate (normal) rate lines and high rate lines. We assume that in this phylog­eny we have a minor example of this situation. Drosophila virilis is cytologically and genetically primitive, the latter in the sense that it is genetically least isolated from other species. We assume that virilis gave rise to Primitive I (or II) which had a higher rate of evolution leading as indicated to the several diverse lines of related species. Although as an alternative, Primitive I might have given rise to virilis, we prefer the former alternative since virilis seems to be close to the an­cestry of the group. There remain a few comments about details of the phylogeny. Just as with basic montana, strains of texana occur that have none of the heterozygous in- Frns. 9, 10. These diagrams show the virilis group phylogeny. There has been no attempt to show which inversions are overlapping but the sequence relations in the phylogeny are clear (see discussion in text of the differential accumulation of paracentrics in different lines, the irreversibility of pericentrics and fusions, etc.). The figures illustrate the diploid chromosome configurations of the males of the species, subspecies and necessary primitive forms although other intermediate steps occurred in the evolution of the several species. The X and the several autosomes are numbered, so pericentrics and fusions are easily identified. To reduce the para­centric inversion designations to minimum repetition, we have used capital letters where the inversion occurred and small letters in the descendent forms with the inversion; if the inversion was heterozygous the letter is in italics. The chromosomes are therefore designated by numbers and the accumulated inversions through the line of descent are lettered a to z (if necessary; e.g., in giant montana chromosome 4 goes to Z) . There are so many inversion changes in the X of ezoana, littoralis, montana and lacicola that they could not be analyzed. The direction of descent is indicated by arrows, two-headed if exchange is occurring. The University of Texas Publication x y VIRIL/S FIG. 9. This shows the relations starting with virilis, and going through Primitive I and Primitive II to texana, americana and novamexicana, and through Primitive I and Primitive III to ezoana and littoralis. The fusion of X and 4 in americana is interesting since Y and 4 are free, forming a trivalent sex chromosome complex. versions, especially in regions distant from the long overlap zone withamericana. It is unnecessary to show this since these inversions are indicated as heterozy­gous and so can be absent in all independent combinations, and no basic texana was necessary to establish the phylogeny. In addition to the heterozygous in­ Stone: Natural Selection and Superspecies b e / ti/ d /i' w cJ'A:' ~@" ~~{" ~f T 6 -!;/' 5 5 5 kj X Cd '~fi~:~1;1~:~ii~ 2 Ii .N XMY XMY XMY ALASKAN-CANADIAN GIANT BOREALIS MONTANA MONTANA XMY FLA VOMONTANA FIG. 10. This shows the sequence from Primitive III to basic montana (Primitive IV) which gave rise to the three montana forms (or subspecies) and to borealis, lacicola and fla vomontana. Note the variation of ZF in descendants of Primitive III and 4H in those of basic montana. versions shared between different subspecies or forms of a species, there are two heterozygous inversions of particular interest, shared in different ways between spe(:ies. The inversion 2 F occurred heterozygous in Primitive III. It is present heterozygous in ezoana but absent in littoralis. However it is fixed in the other line of descent from Primitive III for basic montana and all its descendants have The University of Texas Publication 2f fixed homozygous. Another example is 4 H which was heterozygous in Primi­tive IV, basic montana. The 4 h inversion is heterozygous in all the montana forms, but homozygous in borealis and fiavomontana. It is noteworthy that it is still heterozygous in lacicola, although this is the species in which the X chromo­some has changed radically. These two persistent heterozygous inversions must necessarily have possessed extraordinary heterotic properties, but it is noteworthy that alternative descendent genotypes can utilize one homozygous or lose it instead. The several species illustrate fixation either in the original normal, in the inverted sequence, or retention of the sequence heterozygous, in these descend­ants of the heterozygous (heterotic) ancestor. We have shown one alternative for the origin of the americana, texana, nova­mexicana complex, choosing the one that represented a type of introgressive hybridization (Anderson, 1949) . It is quite clear that the combination of the novamexicana and basic texana inversion complexes which sum to give the mix­ture in americana may have occurred in the past when fewer isolating factors were present between novamexicana and texana and the distribution of the forms was more favorable, perhaps a more pluvial period, since members of the virilis superspecies are ordinarily found along streams and close to water. We do not . know if the original mixture was formed by a one-way migration from the nova­mexicana (ancestor) to the texana (ancestor) area or if two-way migration occurred but the present small limited species novamexicana lost residual texana gene sequences. Alternatively, the original form may have had a mixture of the component sequences now found in these three strains, which sorted out into the several types, plus those resulting from the continuing hybridization between texana and americana. The pericentric inversion of Primitive III and the three fusions found in texana, americana and littoralis provide added evidence of the direction of evolu­tion in the phylogeny. The primitive chromosome configuration in the genus is five pairs of rods (acrocentrics) plus one pair of dots (Figure 4); the polytene chromosome configuration is that of five long arms and a very short arm, due to the intimacy of somatic synapsis. The process of pericentric inversion from an acrocentric to a V-or J-shaped chromosome is relatively difficult and infrequent. The reverse process from a two-armed chromosome to a rod is-equally or more difficult; only two cases where this is the more probable explanation for the ori~in of a known chromosome configuration are known in the genus. Therefore the Primitive III pericentric in chromosome 2 is retained in all descendants of that form. Fusions are even more improbable than pericentric inversions. The re­verse process seems exceedingly difficult in Drosophila since there are no extra centromeres available for easy use except perhaps that of the Y. There are no known cases of opening of a fusion to rods in the genus although this is theo­retically possible. The three fusions are therefore clear determinates of line of descent relationships. In cases favorable for analysis, those with a sufficient array of forms to de­termine all or most of the gene sequence changes such as the virilis and repleta groups, the superspecies-species group-emerges as a real evolutionary taxo­nomic unit. This is a true lineage relation to a common ancestor or ancestors, with the members related to each other in a determined sequence. It consists of Stone: Natural Selection and Superspecies a group of species related cytologically, and genetically where testable, and differing from other superspecies which in turn form a separate cytological array of related species. The repleta group with desert adaptation may be used to illustrate the fact that a superspecies develops with the attainment of some useful unique genetic adaptation which allows the expansion and survival of a series of related but divergent species to develop. In the most favorable cases the main cytological relationships to related species groups can be determined, as Wasser­man has done with the three superspecies related to the primitive cytological stem of the repleta group. Here the flies are phenotypically different, for they are not the spotted species characteristic of the repleta superspecies, nor are they desert adapted forms as most but not all subgroups of the latter. As we find and describe more species, we may form new groups, add to old ones, or leave them undetermined for the time being. It has not been necessary to split off species. Even the very large repleta superspecies, with more than 60 species,. is most reasonably regarded as a unit, although subgroups have been formed (see Wasserman's discussion). The detailed divisions of the genus will be revised with further study and some errors in species groups will be found among those not analyzed cytologically. Even now the phylogeny of the superspecies of the genus is not agreed to completely between Patterson and Stone (1952), Hubby and Throckmorton ( 1960), and Throckmorton (this bulletin) . Simpson (1961) discussed the problems of parallelism and homology from the standpoint of propinquity of descent versus convergent evolution. These poly­tene chromosomes leave no doubt about propinquity of descent and ordering of relationships in favorable cases. Distantly related forms are obvious despite phenotypic similarity by the accumulated divergence in the p0lytene chromo­some pattern. When sufficient rearrangements have accumulated, no further discrimination is possible by this method, but such divergent forms are not as­sumed in error to be closely related. An example of such a case is that given by Wasserman in removing Drosophila peninsularis from the mulleri subgroup (see Wasserman, this bulletin). The development of desert adaptation by D. pseudoobscura is a case of convergent evolution, here related to ecological versa­tility. In this case the convergent evolution did not lead to phenotypic similarity; the gene sequences in the repleta and pseudoobscura superspecies are exceed­ingly different. One important question has too little evidence bearing on it in Drosophila. This is the time involved in the evolution of superspecies The number of polvtene chromosome rearrangements involved in the evolution of a different species is not the same in different groups, even if we knew how frequently inversions, etc., were added, which we do not. In favorable cases we can determine the extent of cytological change between two species and infer very roughly the amount of genetic change but we cannot estimate the number of allele replacements. We can nevertheless make some interesting estimates. Sturtevant (in press) and Wheeler (in press) were provided the opportunity to examine some snecimens of Diptera in amber, estimated to be some 20 to 30 million years old. The speci­mens examined by Wheeler were tentatively placed in Neotanvgastrella. a genus of the Drosophilidae close to Chymomyza. The very P-ood specimen studied by Sturtevant was described by him as a new species of Periscelis of the family The University of Texas Publication Periscelidae. It resembles P. wheeleri more than the other five species in this genus now living. The differences between it and P. wheeleri are of about the same order of magnitude as the differences between the six living species of this genus. The specimens studied by Wheeler are in poor condition. Nevertheless, they suggest a species similar to the seven known Neotropical Neotanygastrella, but with somewhat longer head bristles and arista. Wagner (1961) in discussing the organized chromosome system in relation to organized enzyme complexes in the cytoplasm at the Symposium on enzyme complexes, used the polytene chromosome as an example of a very complex ordered system which has existed in the same sequence of gross organization for a great period of time despite an accumulation of allele changes. He used the data from Wheeler and Sturtevant to estimate the age of the repleta superspecies at 50 to 100 million years. This great age is estimated not only from the fact that the phenotypic differences in the most diverse members of this group are as great as that found by Sturtevant and Wheeler to exist between the fossil amber forms and their modern representatives, but also because the repleta group is not too far removed from the pinicola group which Sturtevant ( 1942) regards as most nearly resembling the primitive ancestral forms for the genus. Zimmer­man (1948) in his introduction to the Insects of Hawaii estimated the Hawaiian Islands to be about 5 million years old (although other people have estimated a greater age) . These islands had a most extraordinary proliferation of species, with the formation of a great many endemic plant species followed by the great development in number of endemic animal species, among them more than 350 species of Drosophila (estimate, in 1961, of Prof. D. Elmo Hardy; personal communication). However, the environment was exceptional, with a tremendous variety of open ecological niches which the few invading species filled by formation of the multiple new endemic forms. It does not seem likely that the same diversity of opportunity was available to the repleta superspecies even after they made the basic desert adaptations that allowed them to invade and proliferate in the Southwest desert area of North America. Therefore we consider that the time of development of this superspecies was more than the 5 million years minimum, perhaps 5 to 20 times this period. However, such an estimate serves better to point out the antiquity of the greatly organized struc­ture, the polytene chromosomes, and the fact that gene change without change in gross gene order could persist for millions of years. When fossils of Drosophila in amber have been found, much more meaningful estimates may be made. Estimates of total rearrangements in Drosophila-The polytene chromosomes are the largest and most complex ordered living system with a determined and consistent inherited pattern that is changed in geometrically simple and limited ways through two-break rearrangements. No other cellular organelle in natural populations of living organisms has an equivalent ordered complexity in which changes can so easily be determined. Therefore chromosome evolution in the genus Drosophila and in other Diptera with polytene chromosomes is most readily and accurately measurable. In fact within the limits discussed, many important taxonomic evolutionary systems can be established with an exactitude impossible elsewhere. Furthermore, the changes which can be used are very frequent in the genus and the rules that govern the relative frequency of occurrence of the several Stone: Natural Selection and Superspecies types of changes can be established experimentally, especially with X-rays. Some estimates of the frequencies of these changes that occurred in the evolution of the genus have been made by Stone ( 1955) and Stone, Guest and Wilson ( 1960). To show the overwhelming effectiveness of natural selection we will introduce a few rough estimates. Wheeler (personal communication) estimates now that there are over 1500 species in the genus, perhaps as many as 3000. We will use an estimate of 2000 living species and not attempt to assess the number that lived in the ancestry nor the unsuccessful side evolutionary sequences. These species would increase the estimate greatly. We will use the data compiled by Stone (1955) that there are 592 variable (heterozygous in some or many individuals) paracentric inversions known in 42 species or an average of 14.1 per species. In the virilis group there are 1.2 paracentrics fixed in the evolution of the super­species to 1 heterozygous while there are 2: 1 in the repleta group according to Wasserman. We will use the figures 1.5 fixed: 1 variable inversion since the repleta group may be more freq_uently homozygous for sequences than the majority of the genus. We can use these figures to make some estimates of the variability in the genus and the variability that made these results possible, Tables 1 ,and 2. Using 2000 as the number of species and 14.1 as the average number of heterozygous paracentric inversions, our estimate of the number now present heterozygous in the genus is 28,200. At 1.5 fixed to 1 variable inversion, this gives us a figure of 42,300 fixed paracentrics. If a species has the primitive chromosome configuration of five pairs of rods and one pair of dots, there would be on the average somewhat over 4 times as many translocations as paracentric inversions among two-break rearrangements. The vast majority of analyzed species have this basic configuration or one involving that with a fusion so we can use a figure of 4 translocations to 1 paracentric. If the survival of transloca­tions and paracentric inversions was equally probable, we would expect 112,800 heterozygous to 169,200 homozygous translocations in the genus. If we just take the 42 species used in this calculation (the number of paracentrics was so large this seemed an adequate sample) , we would expect about 1/50 (42/ 2000) of these translocations in that sample. In reality a larger number of species have now been checked and there are just four reports of the fleeting occurrence of a heterozygous translocation which could not be recovered in subsequent samples. There are the three cases of fixed translocation but two of these are cases of heterochromatic regions that include the nucleolus organizer, and which can live aneuploid without serious adverse effects [the fixation of one of these trans­locations in evolution resulted in the male of Drosophila ananassae being hyper­ploid for the bobbed and nucleolus organizer region transferred from the X to the fourth (dot) chromosome since the Y also carries this region]. It is obvious that the actual frequencies are very different from the relative frequency of expectation, judged by paracentric inversions. In order to estimate the number of occurrences of these two types of two-break rearrangements to get some feeling for the number of trials of such changes that go into the evolution of such a group, we have to make some further (and very conservative) estimates. Fisher (1929, revised ed. 1958) calculated that only about 2 in 100 of the mutations that occurred would survive in sexual reproduc­tion when they had 1 percent selective advantage. An even smaller proportion The University of Texas Publication of those with less selective advantage would survive and those with no advantage (much less those with a disadvantage) would all be eliminated through time. We can estimate that at least fifty times the number occurred that survived from their unique origin, since these two-break rearrangements are unique events, with a very small probability of the same rearrangement occurring twice. Fur­thermore, a very optimistic estimate of the number of beneficial mutations to those with no selective advantage (including detrimental mutations) would be 1: 100. Therefore a very conservative estimate of the number of initial inversions that gave the 28,200 heterozygous paracentrics is 1,410,000 advantageous ones and 141,000,000 disadvantageous or neutral ones; for fixed paracentrics it is 2,115,000 advantageous plus 211,500,000 neutral or disadvantageous para­centrics. This gives a total of over 350 million paracentric inversions involved in the history of the genus, ignoring the fact that the number of inversion differ- TABLE 1 Number of trials The number that occurred with of occun·ences Type of abnormality Average number per species Estimated number in genus (2000 species) selective advantage (only 2% of those which have 1% advantage survive) (trials), if one in 100 had selective advantage and the others 0 or negative Paracentric inversions heterozygous; estimated from sample with 592 heterozygous in 42 species 14.1 28,200 1.41 x 106 1.41 X 108 Paracentrics homozygous (estimated as 1.5 times the number heterozygous per species) 21.15 42,300 2.115 x 106 2.115 X 10s TOTAL (estimate from sampling) 35 .25 70,500 3.525 x 106 3.525 X 10s Translocations Probability of heterozygous translocations, based on a probability of 4 times that of paracentrics (estimate based on the measured frequency of paracentric inversions) 56.4 112,800 5.64 x 106 5.64 X 108 Probability of translocations fixed in species, based on the same assumptions 84.6 169,200 8.46 x 106 8.46 X 108 TOTAL (estimate from sampling paracentrics) 141 282,000 1.41 x 107 1.41 x 109 All homozygous fusions in sample (58/215 species) 0.27 540 27,000 2.7 x 106 Translocations estimated from fusions, based on 1000 translocations: 1 fusion 270 540,000 2.7 x 107 2.7 x 109 Actual translocations: heterozygous ( 4/150 species) 0.03 60 (fleeting) homozygous (3/150 species) 0.02 40 Stone: Natural Selection and Superspecies TABLE 1 Paracentric inversions, translocations and fusions The estimate of the number in the genus and the number that occurred is based on an a priori probability of over 4 translocations to 1 paracentric inversion as discussed in the text. The actual translocations are quite different. In the data on the 42 species used to calculate para­centrics (Stone, 1955), there was 1 translocation heterozygous (in Drosophila ananassae) and 2 fixed (1 in ananassae and 1 between D. pseudoobscura and D. miranda). Natural selection had eliminated the others and actually the heterozygous translocation in ananassae could not be recovered in further samples. Perhaps half of the translocations fixed could be detected in any species where the metaphase is known, especially if the polytene chromosomes are also examined because unequal exchange of material would lead to large size discrepancies in the chromosomes. In perhaps 250 species now checked no additional fixed translocations have been found, other than the three cases discussed in the text. Freire-Maia (appended to Stone, 1955) mentions five pericentric inversions, one deletion, one translocation and two transpositions in addition to 24 paracentrics found in populations of ananassae in Brazil. This species has more recorded cases of these heterozygous abnormalities which are selectively disadvantageous than any other species. Unlike paracentric inversions, big classes of heterozygous and homozygous translocations (see calculated expectations) are absent. Therefore the estimate of the number of heterozygous and homozygous translocations that would be expected in a species or the genus, based on 4 translocations to 1 paracentric inversion, does not correspond to the true numbers, due to the fact that natural selection so effectively eliminated the heterozygous translocations. The number of trials is however a reasonable minimum estimate. The actual number of trans­locations in the genus of 2000 species are 60 fleeting, usually detected once and not again recoverable (4/150 species examined well enough to make an estimate) , and 40 fixed translo­cations, based on the three detected in these 150 species. The number of fusions, the type of translocation that has no selective disadvantage in the heterozygote, is estimated for the genus and for its history, based on 58 fusions in 215 species. Using an estimate of the relative frequency of ordinary translocations to fusions of 1000 to 1 (see text), the number of homozygous translocations and the number of trials of translocations is estimated. T ABLE 2 Mutations available for the evolution of the genus Estimate per large species based on general mutation rate of: M utation rate per locus per generation, based on measurements of spontaneous mutation rate in Drosophila Number of loci estimated at 10,000 (5,000-20,000) in Drosophila from mutation studies. Therefore new mutations per fl y per generation Average number of generations per year (5 to 40 is range, depending on species and temperature) Over 10% (200) of the species have populations ranging from 106-1012. The others have smaller numbers, with fluctuation. Therefore the range in number of mutations per large species per year is: Estimated age of species groups (but perhaps genus is older) is 5 X 106 to 10s. Therefore the mutations per large species (200, with 1800 small-1/ 10 to 1/ 100 as many) during this time 10-6 10-2 20 2 X105 to 2 x 1011 1012 to 2 x 1019 10-5 10-1 20 2 X 10G to 2 x 1012 101a to 2 x 1020 The University of Texa.s Publication ences between species groups is probably much greater than the number detect­able within one of the groups favorable for study, and keeping in mind all of the other minimum estimates we made in the calculations. The number of initial translocations is at least 4 times as great, yet we have found only 3 fixed and 4 heterozygous in the 150 plus species (over 50 in the repleta group alone) that have been examined for translocations, at least to some degree, see Table 1. Although the number of pericentric inversions (32) and fusions (58) that have been established in the genus is small compared to paracentric inversions, the evidence that natural selection determined the relative frequency of their retention in the genus is very clear. Figures 4 and 5 show that the relative prob­ability of pericentric inversion of a rod chromosome (one break anywhere in the long euchromatic arm and one break across the centromere in the very short heterochromatic arm) is at least 100 times as great as that for fusion of two rods (one break in the heterochromatin next the centromere of the major arm of one chromosome and the other break in the short heterochromatic arm of another rod chromosome) . The actual frequencies of those detected in the genus (on minimum estimates) are 32 pericentrics to 58 fusions, which again shows the determining role of natural selection in the addition of chromosome rearrange­ments to a species. A glance at Figures 4 and 5 shows that we are calling a particular class of trans­location centric fusions. The reason for this special designation is the frequency and role of fusions in evolution. White ( 1954) has discussed the role of centric fusions in evolution, beginning with Robertson's original findings in the grass­hopper. Fusions represent a special type of translocation, unusual in that they have little or no detrimental effect on disjunction when heterozygous and so produce normal gametes. The probability of translocation with a break in each of two major arms compared to that of a fusion, the type of translocation with one break in the heterochromatin of one acrocentric and the second break in the short heterochromatic second arm of another, is between 100: 1 and 10,000: 1. The lower figure is based on the fact that we examined well over 100 simple trans­locations in Drosophila virilis without detecting a fusion, and other investigators have also examined a number of translocations, and only one fusion has been reported. The higher estimate is based on the relative sizes of the parts of the chromosomes involved and the scarcity of translocations involving the short heterochromatic arms of the rod chromosomes. We shall use a probability of occurrence estimate of 1000 translocations to 1 fusion for our calculations. We will consider the estimated 58 fusions as homozygous, since the X-4 fusion is homozygous or hemizygous in Drosophila americana americana. A minimum number of 58 fusions in the 215 species checked up to 1951 was reported by Patterson and Stone (1952). This is 0.27 fusions per species or 540 in the genus (2000 species) . If there occur 1000 translocations of the more usual types to 1 fusion, the number of translocations in the genus would be 540,000 and the number of trials of translocations with selective advantage would be 2.7 X 107, whereas the number of trials on minimum estimates would be 2. 7 X 109 (see Table 1). This figure agrees well with the estimate made from inversions. We consider that there are most probably thousands of mutations for every Stone: Natural Selection and Superspecies rearrangement. Chromosomal rearrangements and mutation are the sources of variability available with which Darwin's natural selection shaped evolution. A better feeling for the frequencies of trials leading to the success of evolution through natural selection of these two basic sources of novelty is obtained from the figures summarized in Tables 1 and 2. SUMMARY 1. The existence and properties of the polytene chromosomes in the Diptera, as demonstrated adequately first by Painter ( 1934) for Drosophila melanogaster, are of such a nature that we can establish many important evolutionary taxo­nomic problems. These evolutionary problems have been analyzed recently by Simpson (1951) and we have discussed some of the definitive evidence that it is possible to obtain, at least in the families Drosophilidae, Phoridae, Sciaridae, Culicidae, Simuliidae and Chironomidae. The examples discussed involve the genus Drosophila but the polytene chromosomes seem to be an ordinal character, present in most of the main subdivisions of the Diptera as given by White ( 1949). 2. The chromosomes with the genes distributed along them are the carriers of most genetic information and control of activity from generation to generation. Their stability preserves continuity; their change (mutation) is the source of novel variation and together with recombination by crossingover and random segregation of non-homologous chromosomes, provides the great heritable vari­ability necessary whereby natural selection leads to evolution. One type of change, that leading to gross reorganization of chromosome material, not only provides novelty used in evolution but also provides us with simple non-contro­vertible evidence of the relationships in evolutionary sequences leading to taxo­nomic relationships. Furthermore it provides a remarkable proof of the primary effectiveness of Darwin's Natural Selection based on differential survival of gam­etes and zygotes in the process of evolution. 3. The chromosome may be regarded, following Schrodinger's description, as an aperiodic solid. The polytene chromosome is a very large differentiated chro­mosome. The complexity of pattern of light and dark bands of differing size and visible local structure, with around 5000 dark bands (local heavy concentrations of DNA) , gives an organization in which particular chromosomes, in fact each short region of each chromosome, can be recognized from cell to cell and from generation to generation, see Figure 1. Furthermore the geometric rules govern­ing two-break rearrangements, which are the major or only source of change in these organized patterns of each chromosome, have been worked out very com­pletely from spontaneous and radiation-induced rearrangements. A particular spontaneous rearrangement, such as that leading to the formation of a para­centric inversion, is so rare that the frequency has not been measured. The prob­ability that the same rearrangement will occur twice in evolution in a short time (ca. 1 million years) is very small, and the probability that the same rearrange­ment will occur twice in a limited time in a closely related series of species such as a species group is so small as to be approximately zero. 4. One important result of this organization, subject to orderly change, is that the superspecies, designated alternatively as the species group in Drosophila, The University of Texas Publication emerges as a real and important taxonomic unit. In favorable species groups (those with sufficient surviving species) we can determine a directly observable phylogeny in which the sequence of relationships can be followed in every direc­tion. The primitive stem gene sequences ·of a superspecies can be established as probable in favorable cases, although side branches and intermediate forms may be missing in known living material. This allows us to determine unambiguously relationships and propinquity through common descent versus parallel or con­vergent evolution within and between superspecies. These favorable cases in Drosophila where a definitive check of relationship by common descent versus convergence or parallelism show that careful conservative taxonomists using their usual tools of analysis will usually come to the correct conclusion about relationships. Wasserman needed to remove Drosophila peninsularis from the mulleri subgroup of the repleta superspecies to a monotypic subgroup of its own. The low frequency of errors in judgment, except for sibling species, is a remark­able tribute to the accuracy of the better taxonomists. 5. Nature has provided us with a record of a genus-wide experiment which gives direct and complete proof of the primary importance of Darwin's Natural Selection in evolution, an experiment carried out over millions of years under conditions as they occurred and changed all over the world. For example the data show that the genus, estimated at 2000 living species, has at least 28,200 heterozygous paracentric inversions within species and 42,300 fixed as differences in gene order between them. There are 592 heterozygous inversions in 42 species that have been examined in sufficient numbers to detect much of the inversion variation. There are four fleeting heterozygous translocations and 3 fixed trans­locations in perhaps 150 species where the latter may have been detected (some types of translocations would have been detected in all of the more than 300 species whose metaphase chromosomes have been analyzed). Translocations are about 4 times as probable as paracentric inversions, based on the relative fre­quency of the various chromosome metaphase configurations known in the genus. The comparative frequencies of paracentrics to translocations, especially as 2 of the 3 fixed translocations are the types where aneuploids survive and so have a much higher probability of survival in a population, leave no doubt that natural selection has differentially determined the survival of the type which has no selective disadvantage in the heterozygote in normal gamete formation, i.e. para­centrics. A similar and equally convincing argument can be made from the survival and fixation of fusions, a special type of translocation with no selective disad­vantage since segregation is normal in the heterozygote, as compared with other types of translocations which produce aneuploid gametes from non-disjunction in the heterozygotes. Although the probability of occurrence of an ordinary translocation is at least 1000 times as great as that of a fusion, the numbers fixed in the species sufficiently tested are 3 translocations in 150 species rtwo of them having greater than usual probability of survival because of aneuploid viability and fertility as pointed out by Wright (1941)] to 58 fusions in 215 species. The number of trials of the different kinds of rearrangements is so great that we can rule out chance as the controlling factor in adding appreciable numbers of any type of rearrangement to a species. Stone: Natural Selection arul Superspecies 53:3 6. Using two recent studies of fossil flies, by Wheeler on a member of the Droso­philidae and by Sturtevant on a member of a genus fairly closely related to the Drosophilidae, we have made some age estimates. These specimens were in amber, about 20 to 30 million years old. They resembled the related modern species closely enough to fall within the range found within modern super­species. Remembering Zimmerman's (1948) short time estimate of evolution in the Hawaiian Islands, we estimate that the species groups in Drosophila may be from 5 to 100 million years old. Either figure, at 20 or more generations per year, gives a great deal of time for evolution. When we estimate that some gene orders now in existence may go back that far, we can compare this with the conservative estimate that 350 million paracentric inversions and 4 times that number of translocations were tried for the few that were added to species. An­other and perhaps more important statement can be made about the number of mutations per generation. Perhaps one tenth or more of the species of Drosophila 1012 have large · populations ranging from 106 to individuals per generation. Spontaneous mutation rates usually range from 10-s to 10-6 mutations per gene per generation. One can see how this period of time and these rates of mutation (gene and chromosomal) have provided adequate novel and changing variability for the known evolution of the genus Drosophila to occur. ACKNOWLEDGMENTS I wish especially to thank Professor Marshall Wheeler, who was kind enough to discuss with me many facts and theories based on his extensive knowledge of the taxonomic and evolutionary problems in the Diptera, which were utilized in this analysis. Professor Robert Wagner was kind enough to let me read and refer to his unpublished discussion of the long persistence of highly organized biological systems. Dr. Burke Judd and Mrs. Nancy Parker were kind enough to prepare the slides and photographs of the polytene chromosomes of Drosophila, and the figures were prepared by Mrs. Pauline West. The work was supported (in part) by a PHS research grant (RG-6492) from the National Institutes of Health, Public Health Service, and especially by grants from the Rockefeller Foundation, which provided funds necessary for the many years' work on evolution in Drosophila that made this analysis possible. LITERATURE CITED Anderson, Edgar. 1949. lntrogressive Hybridization. New York: John Wiley and Sons, Inc. Bauer, H., M . Demerec, and B. P. Kaufmann. 1938. X-ray induced chromosomal alterations in Drosophila melanogaster. Genetics 23 : 610-630. Bridges, Calvin B. 1938. A revised map of the salivary gland X-chromosome. J. Hered. 29: 11-13. Buzzati-Traverso, A. 1950. lnterspecific hybrids in the "obscura group" of Drosophila. Society for the Study of Evolution, Program of 1950 meeting. Carson, H. L., and H . D. Stalker. 1947. Gene arrangements in natural populations of Drosophila robusta Sturtevant. Evolution 1: 113-133. Chino, M. 1936. A case of inversion in the fifth chromosome of Drosophila virilis. Jap. Joum. Genet. 12: 63-65. Clausen, Jens, David D. Keck, and William M. Hiesey. 1939. The concept of species based on experiment. Amer. J. Bot. 26: 103-106. da Cunha, A. Brito, and Th. Dobzhansky. 1954. A further study of chromosomal polymorphism in Drosophila willistoni in its relation to the environment. Evolution 8: 119-134. The University of Texas Publication Darwin, C. 1859. On the origin of species by means of natural selection or the preservation of favoured races in the struggle for life. London, Murray. Dobzhansky, Theodosius. 1951. Genetics and the Origin of Species. New York: Columbia University Press. Dobzhansky, T., H. Burla, and A. B. da Cunha. 1950. A comparative study of chromosomal polymorphism in sibling species of the willistoni group of Drosophila. Amer. Nat. 84: 229-246. Dobzhansky, Th., and A. Dreyfus. 1943. Chromosomal aberrations in Brazilian Drosophila ananassae. Proc. Natl. Acad. Sci. 29: 301-305. Dobzhansky, Th., and C. Pavan. 1943. Chromosome complements of some South-Brazilian species of Drosophila. Proc. Natl. Acad. Sci. 29: 368-375. Dobzhansky, Th., and A. H. Sturtevant. 1938. Inversions in the chromosomes of Drosophila pseudoobscura. Genetics 23: 28-64. Dobzhansky, Th., and C. C. Tan. 1936. Studies in hybrid sterility. III. A comparison of the gene arrangement in two species, Drosophila pseudoobscura and Drosophila miranda. Z. i. A. V. 72: 88-114. Fisher, R. A. 1929; revised ed., 1958. The Genetical Theory of Natural Selection. New York: Dover Publications, Inc. Hsiang, W. 1949. The distribution of heterochromatin in Drosophila tumiditarsus. Cytologia 15: 149-152. Hsu, T. C. 1952. Chromosomal variation and evolution in the virilis group of Drosophila. Univ. Texas Publ. No. 5204: 35-72. Hubby, Jack L., and Lynn H. Throckmorton. 1960. Evolution and pteridine metabolism in the genus Drosophila. Proc. Natl. Acad. Sci. 46: 65-78. Kaufmann, B. P. 1937. Morphology of the chromosomes of Drosophila ananassae. Cytologia, Fujii Jubilee Vol., pp. 1043-1055. Kikkawa, H. 1938. Studies on the genetics and cytology of Drosophila ananassae. Genetica 20: 458-516. Miller, D. D. 1939. Structure and variation of the chromosomes in Drosophila algonquin. Genetics 24: 699-708. Moorhead, Paul S. 1954. Chromosome variation in giant forms of Drosophila montana. Univ. Texas Publ. No. 5422: 106-129. Okada, Toyohi. 1956. Systemntic Study of Drosophilidae and Allied Families of Japan. Tokyo, Japan: Gihodo Co., Ltd. Painter, T. S. 1934. Salivary chromosomes and the attack on the gene. J. Hered. 25: 465-476. Patterson, J. T., and W. S. Stone. 1952. Evolution in the Genus Drosophila. New York: The Macmillan Co. Pavan, C., and M. E. Breuer. 1952. Polytene chromosomes. J. Hered. 43: 151-157. Schrodinger, Erwin. 1945. What ls Life? New York: The Macmillan Co. Simpson, George Gaylord. 1944. Tempo and Mode in Evolution. New York: Columbia Uni­ versity Press. -----. 1961. Principles of Animal Taxonomy. New York: Columbia University Press. Stalker, Harrison D. 1954. Banded polytene chromosomes in the ovarian nurse cells of adult Diptera. J. Hered. 45: 259-264. Stone, W. S. 1949. The survival of chromosomal variation in evolution. Univ. Texas. Publ. No. 4920: 18-21. Stone, Wilson S. 1955. Genetic and chromosomal variability in Drosophila. Cold Spring Harbor Symposia on Quant. Biol. 20: 256-270. Stone, Wilson S., William C. Guest, and Florence D. Wilson. 1960. The evolutionary implica­tions of the cytological polymorphism and phylogeny of the virilis group of Drosophila. Proc. Natl. Acad. Sci. 46: 350-361. Sturtevant, A. H. 1942. The classification of the genus Drosophila, with descriptions of nine new species. Univ. Texas Publ. No. 4213: 5-51. Sturtevant, A. H. A Fossil Periscelid (Diptera) from the amber of Chiapas, Mexico. J. Pal., in press. Storze: Natural Selection and Superspecies Sturtevant, A. H., and G. W . Beadle. 1936. The relations of inversions in the X chromosome of Drosophila melanogaster to crossing over and disjunction. Genetics 21: 554-604. Sturtevant, A. H., and Th. Dobzhansky. 1936. Inversions in the third chromosome of wild races of Drosophila pseudoobscura, and their use in the study of the history of the species. Proc. Natl. Acad. Sci. 22: 448-450. Throckmorton, L. H. The problem of phylogeny in the genus Drosophila. This Bul~etin. Wagner, R. P. 1961. Genetic organization and enzyme complexes. Symposium on Enzyme Complexes. Meeting of the American Chemical Societies, St. Louis, Mo. Ward, Calvin L. 1952. Chromosome variation in Drosophila melanica. Univ. Texas Puhl. No. 5204: 137-157. Wasserman, Marvin. 1960. Cytological and phylogenetic relationships in the repleta group of the genus Drosophila. Proc. Natl. Acad. Sci. 46: 842--859. Wasserman, M. Cytological studies of the repleta group of the genus Drosophila. This Bulletin. Wharton, Linda T. 1942. Analysis of the repleta group of Drosophila. Univ. Texas Puhl. No. 4228: 23-52. Wheeler, Marshall R. 1959. A nomenclatural study of the genus Drosophila. Univ. Texas Puhl. No. 5914: 181-205. Wheeler, M. R. A note on some fossil Drosophilidae (Diptera) from the amber of Chiapas, Mexico. J. Pal., in press. White, M. J. D. 1949. Cytological evidence on the phylogeny and classification of the Diptera. Evolution 3: 252-261. -----. 1954. Animal Cytology and Evolution. 2nd Edn. Cambridge: The University Press. Wright, Sewall. 1931. Evolution in Mendelian populations. Genetics 16: 97-159. -----.. 1941. On the probability of fixation of reciprocal translocations. Amer. Nat. 75: 513-522. -----. 1959. Physiological genetics, ecology of populations, and natural selection. in Perspectives in Biology and Medicine, Vol. III, No. 1: 107-151. Zimmerman, E. c.· 1948. Insects of Hawaii. I. Introduction. Univ. Hawaii Press, pp. 1-206. XVIII. Hybridization Tests Within the Cardini Species Group of the Genus Drosophila DAVID G. FUTCH INTRODUCTION At various times in the past, members of the cardini species group of Drosophila have come under the scrutiny of geneticists interested in some of the unique characteristics encountered in certain species of this group. In two of the species, D. parthenogenetica Stalker and D. polymorpha'Dobzhansky and Pavan, Stalker (1954) was able to observe, for the first time in the genus Drosophila, a mech­anism for the parthenogenetic production of fertile offspring. The polymorphism exhibited by several of the species with respect to banding pattern and the degree of darkness or lightness of the abdomen has afforded da Cunha (1949) and, cur­rently, Heed, a supply of intriguing material with which to work. In a series of two articles (Heed and Krishnamurthy, 1959, and Heed, this bulletin), Dr. William Heed has found the small island populations of species of the dunni sub. group in the West Indies to be an excellent source of material for studying speci­ation and evolution in such geographically isolated populations of relatively few individuals. In the course of his work with the dunni complex, Heed has completed hybrid­ization tests among all of the various species and subspecies of this subgroup thus far known. However, the other subgroup of the cardini species group, the cardini subgroup, has remained relatively untouched in this respect in recent years. No experimental hybridizations are known to have been attempted since 1953, when Stalker (1953) tested the six species available to him at that time. The number of species identified as members of the cardini species group has now grown to fifteen, of which eight, including the six tested by Stalker, have been placed in the cardini subgroup. The remaining seven are members of the dunni subgroup (Heed, this bulletin). The purpose of this report, therefore, is to relate the results of hybridization tests recently carried out by the author with the eight currently known species of the cardini suba:rouo. These tests included intrasubPToup crosses in all possihle combinations and also some intersubgroup crosses between members of tlie cardini subgroup and three members of the dunni subgroup. Some of these tests were designed to determine the repeatability of crosses made by earlier investi­p,-Ht0rs, esoecially when stocks of different geographic origins were involved, as well as to attempt hybridization between species not available to the earlier workers. Although Heed has summarized the taxonomy and distribution of the entire cardini species group in his current article, a more detailed resume of the species used in these particular experiments is included in this report. This seems de­sirable at this time since brief descriptions of hybrids obtained from these crosses are given and since a comparison of apparent morphological relationships within The University of Texas Publication and between the two subgroups and those implied by genetic compatibility is also discussed. TAXONOMY AND DISTRIBUTION , In distribution the members of the cardini species group are confined almost entirely to the Neotropical region, although a few species have been found to extend into areas in the lower part of the Nearctic. Seven of the eight members of the cardini subgroup are continental species, being found principally on the mainland of Central and South America; the island of Trinidad is considered as part of South America. The other member of the cardini subgroup and the seven species of the dunni subgroup are known almost exclusively from the islands of the West Indies, the exception being the occurrence of the former in Florida. Aside from parthenogenesis and polymorphism, two distinctive characteristics found within the cardini species group are skipping larvae and larval cannibal­ism. Third instar larvae of all fifteen species have the ability to skip, an ability which enables them to attempt to migrate away from the food site just prior to pupation. A low level of cannibalism has been seen by Heed and Krishnamurthy (1959) in all stocks of the dunni subgroup which they examined. Apparently occasional cannibalism is not confined to the island species, for a survey of stocks used in these experiments revealed a few pupae in almost every stock in the process of being victimized and eaten by small, active larvae. It appears obvious that these two phenomena of behavior are probably related. However, since they have neither been studied in nature nor to any extent in the laboratory, the prob­lem of how this increased mobility and cannibalism operate in the adaptation of these species to their environment is open to speculation. THE SPECIES* 1. Drosophila cardini Sturtevant Distribution: Collections from Florida, Central and South America, all of the Greater Antilles, and the four southern islands of the Lesser Antilles (Martinique, St. Lucia, St. Vincent, and Grenada). Reported from Texas, New Mexico, and Arizona (Patterson, 1943). Metaphase Chromosomes: Five pairs of rods and one pair of dots; the only member of the species group having thi~ somatic metaphase figure. Other Remarks: Both dark and intermediate forms are found throughout the range. One stock of light flies which originated in Chile was received in the laboratory, but it has since "reverted" to a darker phenotype. 2. Drosophila acutilabella Stalker Distribution: Collections from Cuba, Jamaica, Hispaniola, and Florida. Metaphase Chromosomes: Two pairs of V's, one pair of rods (sex chromo­somes), and one pair of dots. Other Remarks: This is a plastic species which resembles D. belladunni on Jamaica and D. cardini in Florida. Males are easily recognized by the pres­ence of a prominent antero-ventral process on the labellum. This is the only *Nos. 1--8 belong to the cardini subgroup; 9-11 belong to the dunni subgroup. Futch: Cardini Group of Drosophila species in the cardini species group having such a structure unless D. ros­trata from Costa Rica is eventually found to belong to the group. 3. Drosophila cardinoides Dobzhansky and Pavan Distribution: Collections from various localities in Central and South Ameri­ca from Mexico to Brazil. Metaphase Chromosomes: Two pairs of V's, one pair of rods (sex chromo­somes) and one pair of dots. Other Remarks: This species has limited variability in phenotype from in­termediate to dark. 4. Drosophila procardinoides Frydenberg Distribution: Collections from Bolivia and Peru; has been found only at high altitudes in the Andes. Metaphase Chromosomes: Two pairs of V's, one pair of rods (sex chromo­somes), and one pair of dots. Other Remarks: This is the darkest species of the cardini species group. The dark abdominal pattern is extensive and black, as are the pleurae. The , mesonotum is dark brown. Except for being larger, it resembles D. cardi­noides ingeneral conformation. 5. Drosophila parthenogenetica Stalker Distribution: Collections from various localities from Mexico to Trinidad. Metaphase Chromosomes: Two pairs of V's, one pair of rods (sex chromo­somes), and one pair of dots. Other Remarks: This is another species which closely resembles D. cardi­noides. It is also one of the species displaying considerable polymorphism. Females of a stock from Atlixco, Mexico, are capable of diploid partheno­genesis (Stalker, 1954). 6. Drosophila polymorpha Dobzhansky and Pavan Distribution: Collections from various localities from Panama to Brazil; also from Trinidad and Grenada. Metaphase Chromosomes: Two pairs of V's, one pair of rods (sex chromo­somes) , and one pair of dots. Other Remarks: This species shows a marked degree of polymorphism. In Brazil three phenotypes, light, intermediate, and dark occur which are de­termined by a single pair of alleles neither of which is dominant ( da Cunha, 1949) . However, the light form, which is monomorphic in the northern limits of the range, is dominant when crossed to dark forms from Brazil. 7. Drosophila neomorpha Heed and Wheeler Distribution: Collections from various localities from Mexico to Trinidad. Metaphase Chromosomes: Two pairs of V's, one pair of rods (sex chromo­somes), and one pair of dots. Other Remarks: This species is also polymorphic and is very similar to D. polymorpha in external morphology. Males can be distinguished by differ­ences in the arrangement of secondary teeth on the forceps of the genitalia. The arrangement in D. polymorpha is circular. This is not the case in D. neomorpha. The University of Texas Publication 8. Drosophila neocardini Streisinger Distribution: Collections from various localities in Brazil and from the Amazon basin in Colombia. Metaphase Chromosomes: Two pairs of V's, one pair of rods (sex chromo­somes), and one pair of dots. Other Remarks: Two subspecies, D. n. mourensis and D. n. itambacuriensis, occur in Brazil ( da Cunha, 1955). Each subspecies is distinct and mono­morphic in abdominal morphology. 9. Drosophila belladunni Heed and Krishnamurthy Distribution: Jamaica. Metaphase Chromosomes: Two pairs of V's, one pair of rods (sex chromo­somes), and one pair of dots. Other Remarks: This species is found in greatest numbers in the highlands while its sibling species, D. acutilabella, is found in greater numbers in the lowlands (Heed and Krishnamurthy, 1959). 10. Drosophila dunni Townsend and Wheeler Distribution: Puerto Rico and St. Thomas Island, Lesser Antilles. Metaphase Chromosomes: Two pairs of V's, one pair of rods (sex chromo­somes) , and one pair of dots. Other Remarks: Each of the two island populations represents a subspecies, D. d. dunni on Puerto Rico and D. d. thomasensis on St. Thomas (Heed, this bulletin) . 11. Drosophila arawakana Heed Distribution: St. Kitts Island and Guadeloupe Island, both of the Lesser An­tilles. Metaphase Chromosomes: Two pairs of V's, one pair of rods (sex chromo­somes), and one pair of dots. Other Remarks: Each of these two island populations also represents a sub­species, D. a. arawakana on Guadeloupe and D. a. kittensis on St. Kitts (Heed, this bulletin). One further note concerning taxonomic relationships within the cardini sub­group comes from Heed's current report. On the basis of morphological similar­ities seen in the appearance of male apodema freshly mounted from live speci­mens, Dr. Heed has been able to envision possible affinities in the form of two triads among the mainland species: D. neocardini, D. polymorpha, and D. neomorpha forming one, and D. cardinoides, D. procardinoides, and D. parthe­nogenetica forming the other. Furthermore, D. acutilabella possesses affinities with both the neocardini-polymorpha-neomorpha triad and the dunni subgroup. In this respect, D. cardini stands alone, being distinct from all other species. HYBRIDIZATION TESTS Results of Earlier Studies Previous mention has been made of Stalker's (1953) efforts in crossing mem­bers of the cardini species group. He tested the six species: D. cardini (Florida), D. acutilabella (Florida), D. cardinoides (Guatemala and Brazil), D. neocardini (Brazil), D. polymorpha (Brazil) , and D. parthenogenetica (Atlixco, Mexico). Futch: Cardini Group of Drosophila lnterspecific matings, and their reciprocals, in all possible combinations were made with the exception of those which would have involved females of D. parthenogenetica, since this species had earlier been found capable of partheno­genesis. In all matings involving D. acutilabella females, with one exception, viable hybrids were produced. The one exception occurred when D. neocardini males were involved, although insemination was frequent. Only two other fertile crosses were recorded from among these matings: D. cardini females crossed successfully to D. polymorpha males and also to D. acutilabella males. A further note concerning these matings is that in no case could insemination be detected in D. neocardini females mated with males of the five other species. Prior to this study, Streisinger (1946) had attempted crosses involving D. cardinoides, D. neocardini, and D. polymorpha. His results were all negative except that some insemination did occur in crosses between D. cardinoides and D. polymorpha. D. acutilabella has been found to cross to some members of the dunni subgroup (Heed and Krishnamurthy, 1959, and Heed, this ' bulletin). Reciprocal crosses between D. arawakana arawakana, D. similis similis, and D. nigrodunni and various stocks of D. acutilabella have produced small numbers of offspring. Stocks and Methods Table 1 lists all of the stocks used in this investigation. The symbols were de­signed to indicate specific names and, in instances in which two or more stocks of a species were used, also geographic origins of the stocks. Since two stocks of D. cardinoides from Trinidad were tested, the year of collection is also included. Mass matings were made of from 20-30 males and 20-30 females, all aged for TABLE 1 List of the stocks used in this investigation, their geographic origin, University of Texas stock number, and the stock symbols used in the tests Subgroup Species Locality Stock number Symbol cardini cardini Mayaguez, Puerto Rico H260.26 CA cardini acutilabella St. Vicente, Cuba 2380.2 AC-C cardini acutilabella* Riverview, Florida 2303.11 AC-F cardini acutilabella* Hermitage Reservoir, Jamaica H1 37.5 AC-J cardini polymorpha El Recuerdo, Colombia H1 86.49 PO cardini neomorpha Cerro la Campana, Panama H1 83.4 M cardini neocardini Sao Paulo, Brazil H340.7 NC ca rdini cardinoid es Port of Spain, Trinidad H234.1 CS-T57 cardini cardinoides Port of Spain, T rinidad H332.22 CS-T58 cardini cardinoides Puebla, Mexico 2253.2 CS-M cardini procardinoides Coroico, Bolivia H346.8 PR cardini parthenogenetica Monteria, Colombia H313.7 PA-C cardini parthenogenetica Atlixco, M exico 1802.17 PA-M dunni d. dunni Rio Piedras, Puerto Rico H1 30.5 DU dunni a. arawakana Guadaloupe, F .W.I. H 252.7 AR dunni belladunni Hardware Gap, Jamaica H356.3 BE • Tltese stocks used only in crosses with D. cardinoid.P.s. The University of Texas Publication 5-6 days prior to mating. When possible, two sets of mass matings were made for each possible cross, with the exception of homogamic (control) crosses which were made only once for each stock. The mated flies were kept in standard size culture vials of banana-agar-malt­yeast medium, supplemented by a small amount of dry Fleischmann's yeast, which was sprinkled onto the food just prior to transfer of the flies into the vial. Transfer to fresh food was made every four days. Evacuated vials were ex­amined for the presence of larvae, pupae, and dead embryos at least once every two days. Dead embryos could be detected by the mottled brown coloration found in eggs in which partial development, death, and decomposition had oc­curred. All of these tests were followed for a forty-day period or until all of the flies were dead. Females from infertile matings were not checked for insemi­nation. Before hybridization experiments were attempted utilizing females from either stock of D. parthenogenetica, both stocks were examined for parthenogenesis. The results obtained by testing females of the PA-M stock were in accord with Stalker's ( 1954) original tests with flies of this same stock. Thus females of this stock could not be used. However, all of the tests with PA-C females were nega­tive even with regard to incomplete parthenogenetic development. This was fortunate, for it permitted the use of females of this species in these studies. The first generation hybrids of all successful crosses were tested for fertility. Since no F2 progeny were ever produced from inbreeding offspring of interspeci­fic crosses, backcrosses were attempted with both parental species whenever pos­sible. In several instances, testing for hybrid fertility was halted prematurely because of a heavy mite infestation which occurred before all of the tests were concluded. RESULTS Table 3 lists all but four of the matings made. It also indicates the fertility observed in each case and, if adults were produced, the number of male and female progeny. The four exceptions noted above involved two stocks of D. acutilabella which were used only in four specific instances. They are included in Table 2 which analyzes each cross that produced viable adults. With respect to interspecific crossability, F1 adults were produced by 18 of the 110 interspecifi.c crosses possible between the eleven species utilized in this study. Five other interspecific crosses produced inviable hybrids (embryos and larvae) in enough numbers to be considered significant. The several instances in which a few dead embryos and/ or one or two inviable larvae were observed were noted with the reservation that parthenogenesis at such abbreviated levels is known to occur in a number of the species of the cardini group (Stalker, 1953, and Heed and Wheeler, 1957). From the information given in Table 2 it can be seen that each of the eleven species tested in this study was crossed successfully with at least .one of the other ten species. As was expected, D. acutilabella was involved in more fertile inter­ specific crosses (six) than any other species. D. arawakana and D. belladunni each crossed with four other species, while both D. cardinoides and D. neccardini each crossed with three other species. Futch: Cardini Group of Drosophila Of the six fertile crosses which involved D. acutilabella, four represented suc­cessful efforts to repeat work of Stalker ( 1953) and Krishnamurthy ( 1959). The other two, crosses of D. acutilabella with D. neocardini and D. belladunni, had been tried but had not previously been found to produce adult offspring. Another difference between these results and those reported earlier was noted in crosses between D. acutilabella females and D. parthenogenetica males. Whereas Stalker was able to report the production. of adult hybrids from a cross between these two species, the only indication of fertility observed in the present tests was the production of inviable larvae by two of the four crosses attempted. Both of the crosses producing larvae involved the same stock of D. parthenogenetica (Atlixco) that Stalker used. However, Stalker's D. acutilabella came from Florida while the stock selected for these studies was from Cuba. It seems probable that the dis­crepancies noted here were due to genetic differences in the stocks of D. acutila­bella. Further evidence for the operation of such genetic differences in the interspe­cific crossing ability of different stocks of the same species was indicated by a series of tests involving D. cardinoides. These tests were especially interesting since two of the three stocks of this species used originated from flies collected from the same locality, i.e., Port of Spain, Trinidad. The first of this series of crosses was between D. acutilabella females and D. cardinoides males. AC-C~ X CS-T57 ~ produced adult progeny every time it was tried. However, AC-C~ x CS-T58~ always resulted in inviable hybrids. Only one larva from two attempts to complete this cross managed to survive to pupate; it developed no further than this. When CS-T58 males were crossed to fem ales of two other stocks of D. acutilabella, AC-F and AC-J, adult progeny were pro­duced. A difference was also found when females of the two Trinidad stocks of D. cardinoides were mated with males of the two stocks of D. parthenogenetica., These crosses were fertile in every case except one. No fertility at all was oh· served in two mass matings between CS-T57 females and PA-M males. Since females were not dissected it is not known if they were inseminated. Intraspecific crosses between CS-T57 and CS-T58 went well and indicated nothing out of the ordinary about the two stocks. Likewise, salivary chromosomes of the intraspecific hybrids were paired normally indicating no cytological dif­ferences. The crosses between the three members of the cardinoides-procardinoides­parthenogenetica triad proposed by Heed were another interesting feature of this study. Each species crossed successfully with the other two species in at least one direction, including reciprocal crosses between D. cardinoides and D. procardinoides. In all cases adults of both sexes were produced, and the hybrid females were fertile in every case in which backcrosses were attempted. In addi­tion, hybrid females of the cross between CS-M females and PA-M males were crossed successfully to D. procardinoides males. Apparently this complex as visualized by Heed is valid on the basis of both morphological and genetic affinities. Such convenient support was not lent the other triad proposed by Heed D. neomorpha females crossed to D. polymorpha males, but no significant fertility The University of Texas Publication TABLE 2 Summary of results from all fertile crosses and the fertility of the F1 offspring No. Total Fertility of F1 mass no. Percent l' 6 matings offspring males FXFBackcross 11 CA XCA 1170 51.0 fertile (control) CA x PA-M 1 0 F1 'i' sterile (no eggs) AC-C X AC-C 1 864 46.4 fertile (control) AC-C X CA 3 7 100 F1 6 6 :sterile with AC-C & CA 'i' 'i' AC-C X PO 3 35 0 F1 'i' 'i' :fertile with AC-C & PO 6 6 AC-C X NC 3 167 29.9 sterile F1 'i' 'i' :fertile with AC-C & NC 6 6 F1 6 6 :sterile with AC-C & NC 'i' 'i' AC-C X CS-T57 3 117 0.9 sterile F1 'i' 'i' :fertile with AC-C & CS-T57 6 6 F1 'i' 'i' :fertile with CS-T58 6 6 F1 6 not tested AC-C XCS-M 1 9 0 (not tested) AC-C X AR 2 145 0 F1 'i' 'i' :fertile with AC-C & AR 6 6 AC-F X CS-T58 5 0 (not tested) AC-F X CS-M 24 0 (not tested) AC-J X CS-T58 7 0 (not tested) AC-J X CS-M 1 0 (not tested) PO X PO 592 47.0 fertile (control) NM XNM 1 696 50.3 fertile (control) NM x PO 2 30 70 sterile F1 'i' 'i' :fertile with NM & PO 6 6 F1 6 6 :sterile with NM & PO 6 6 NC X NC 477 51.6 fertile (control) CS-T57 X CS-T57 1 699 49.2 fertile (control) CS-T57 X CS-T58 1 871 46.2 fertile (not tested) CS-T57 X CS-M 677 44.9 fertile (not tested) CS-T57 X PR 144 40.9 sterile (not tested) CS-T57 X PA-C 86 44.2 sterile (not tested) CS-T58 X CS-T58 612 48.0 fertile (control) CS-T58 X CS-T57 718 45.1 fertile (not tested) CS-T58 X CS-M 723 44.3 fertile (not tested) CS-T58 X PR 2 167 43.1 sterile F1 'i' 'i' :fertile with CS-T58 & PR 6 6 F1 6 6 :sterile with CS-T58 & PR 'i' 'i' CS-T58 X PA-C 2 103 47.6 sterile F1 'i' 'i' :fertile with CS-T58 & P A-C 6 6 F1 6 6 :sterile with CS-T58 & P A-C 'i' 'i' CS-T58 X PA-M 118 39.0 ;terile F1 'i' 'i' :fertile with CS-T58&PA-M 6 6 F1 6 6 :sterile with CS-T58 & PA-M 'i' 'i' CS-M X CS-M 790 48.6 fertile (control) CS-M X CS-T57 829 51 .9 fertile (not tested) CS-M X CS-T58 494 42.1 fertile (not tested) CS-M X PR 198 47.0 sterile F1 6 6 :fertile with CS-M & PR 6 6 F1 6 6 :sterile with CS-M & PR 'i' 'i' CS-M X PA-C 2 110 48.2 sterile F1 'i' 'i' :fertile with CS-M & P A-C 6 6 F1 6 6 :sterile with CS-M & PS-C 'i' 'i' CS-M X PA-M 1 137 42.3 sterile (not tested) (CS-M X PA-M) X PR) 1 31 25.8 sterile (not tested) PR X PR 1 640 47.7 fertile (control) PR X CS-T57 156 55.8 sterile (not tested) PR X CS-T58 2 111 63.1 sterile F1 'i' 'i' :fertile with PR & CS-T58 6 6 F1 6 6 :sterile with PR & CS-T58 'i' 'i' PR X CS-M 36 44.4 sterile (not tested) Futch: Cardini Group of Drosophila TABLE 2--Continued Summary of results from all fertile crosses and the fertility of the F1 offspring o. Total Fertility of F1 mass no. Percent O' matings offspring males F1 XF1 Back cross PR X PA-C 2 8 75 sterile F1 'i' 'i' :fertile with PA-C 'i' 'i' (only test made) F1 6' 6' :sterile with P A-C 6' 6' (only test made) PR x PA-M 5 40 sterile (not tested) PA-C X PA-C 1 912 49.3 fertile (control) PA-C X PA-M 2 1269 50.4 fertile (not tested) PA-M x PA-M 907 48.5 fertile (control) DU X DU 893 48.4 fertile (control) DU X AR 59 44.1 sterile F1 'i' 'i' :sterile with DU & AR 6' 6' F1 6' 6' :sterile with DU & AR 'i' 'i' AR X AR 1282 47.5 fertile (control) AR X NC 1 100 sterile AR X DU 8 25 sterile (not tested) BE X BE 671 45.2 fertile (control) BE X AC-C 2 10 0 F1 'i' 'i' :slightly fertile with BE 6' 6' F1 'i' 'i' :sterile with AC-C 6' 6' BE X NC 2 10 20 sterile (not tested) BE x DU 2 10 40 sterile (not tested) BE XAR 2 45 0 (not tested) was observed in attempts to cross D. neocardini to either of these two species. The possibility of a relationship between D. neocardini and the dunni sub­group was established by the two fertile crosses between this species and two of the three members of the dunni subgroup used in these tests. Thus, D. neocardini is the second member of the cardini subgroup, D. acutilabella being the other, which has sufficient genetic compatibility with the dunni subgroup to be cross fertile with members of that complex of the cardini species group. DESCRIPTION OF HYBRIDS AC-C'i' XCS-T57 6'. Hybrid 'i' 'i' from a cross between intermediate D. acu­tilabella and dark D. cardinoides. Abdominal pattern: Intermediate in extensive­ness; darker and more distinct than AC-C; no medial anterior extensions of dark bands of tergites 2 and 3 as in AC-C; no lateral anterior extension of dark band of tergite 3 as in CS-T57. Palpi: Intermediate in shape; one or two fairly strong apical bristles on anterolateral edge, similar to AC-C but less prominent. AC-C 'i' X PO 6' . Hybrid 'i' 'i' from a cross between intermediate D. acutilabella and light D. polymorpha. Thorax and abdomen had yellow color of PO. Bristles of flexor surface of fore femur were similar to AC-C. Abdominal pattern: Fairly light, intermediate between light and intermediate parental forms; medial an­terior extensions of bands were very faint. AC-C 'i' x NC 6' . Hybrid 'i' 'i'. Abdominal pattern was much like that of D. neocardini. Hybrid 6' 6'. Abdominal pattern: Some variation through a range of The University of Texas Publication TABLE 3 Total number of hybrids and control offspring from one or two mass matings of 20-30 males and 20-30 females each. Female progeny are shown to the left of the colon and male progeny to the right. Results of the second mass mating, when made, are shown after the semicolon. The following abbreviations are used: F: nonfertile; P: pupae; L: larvae; L-f-: numerous larvae; DE: dead embryos; DE-f-: numerous dead embryos. PA-M females used only in homogamic mating, due to parthenogenesis Female Male Results Female Male Hesults CA CA 573:597 PR 45:39;49:34 AC-C L,DE-f-;L,DE+ PA-C 29:25;25:24 PO NF;NF PA-M 72:46 NM NF;NF DU NF;NF NC DE;L,DE-f- AR NF CS-T57 NF BE NF;NF CS-T58 NF;NF CS-M CS-M 406:384 CS-M NF CA NF PR DE;DE AC-C NF PA-C NF; F PO F PA-M 1:0,L-f-,DE-f- M F DU NF; F NC NF AR NF CS-T57 399:430 BE DE;NF CS-T58 286: 208 AC-C AC-C 463:401 PR 105:93 CA 0:3;0:2 PA-C 5:6;52:47 PO 22:0;9:0 PA-M 79: 58 NM NF;NF DU NF NC 60:35;25:20 AR NF CS-T57 22:0;30: 1 BE NF CS-T58 P,L,DE-f-;L,DE-f- PR PR 335:305 CS-M 9:0 CA NF;NF PR NF;NF AC-C F · ' F PA-C NF; F PO NF;NF PA-M L,DE-f-;L,DE M NF; F DU F;DE c NF;NF AR 123:0;22:0 CS-T57 69:87 BE DE;NF CS-T58 8: 11;33:59 PO PO 314:278 CS-M 20: 16 CA NF;NF PA-C L;2:6 AC-C NF;NF PA-M 3:2 NM NF;NF DU oneL;NF NC NF;NF AR F CS-T57 NF BE NF; F CS-T58 NF; F PA-C PA-C 462:450 CS-M F CA NF; F PR NF;NF AC-C NF; F PA-C NF;NF PO NF;DE PA-M F M NF;NF DU NF;NF c NF;NF AR NF CS-T57 NF BE NF;NF CS-T58 NF;NF NM NM 346;350 CS-M NF CA oneL;NF PR NF;NF AC-C NF;NF PA-M 273 :280;356:360 Futch: Cardini Group of Drosophila TABLE 3-Continued Total number of hybrids and control offspring from one or two mass matings of 20-30 males and 20-30 females each. Female progeny are shown to the left of the colon and male progeny to the right. Results of the second mass mating, when made, are shown after the semicolon. The following abbreviations are used: NF: nonfertile; P: pupae; L: larvae; L+ : numerous larvae; DE: dead embryos; DE+ : numerous dead embryos. P A-M females used only in homogamic mating, due to parthenogenesis Fema le Male Results Female Male Results PO 2:6;7: 15 DU NF;NF NC NF;L AR NF CS-T57 NF BE NF;DE CS-T58 NF;NF PA-M PA­M 467:440 CS-M NF DU DU 461:432 PR NF;NF CA NF;NF PA-C NF;NF AC-C L,DE;NF PA-M NF PO DE;DE DU NF;NF NM NF;NF AR NF NC NF;NF BE NF;NF CS-T57 NF NC NC 231:246 CS-T58 · NF;NF CA NF;NF CS-M NF AC-C NF;NF PR NF;NF PO NF;NF PA-C NF;NF NM oneL;NF PA-M NF CS-T57 NF AR 33:26 CS-T58 NF;NF BE DE+ ;DE+ CS-M NF AR AR 673: 609 PR NF;NF CA P,L+,DE+ PA-C NF;NF AC-C NF PA-M NF PO NF DU NF;NF NM NF AR NF NC 0: 1 BE NF;NF CS-T57 NF CS-T57 CS-T57 355:344 CS-T58 NF CA NF CS-M NF AC-C NF PR NF PO NF PA-C oneL NM NF PA-M NF NC NF DU 6:2 CS-T58 469:402 BE DE CS-M 374:304 BE BE 368:303 PR 85:59 CA NF;NF PA-C 48:38 AC-C 4:0;6:0 PA-M NF;NF PO DE;NF DU NF NM NF;NF AR NF NC 4:2;4:0 BE NF CS-T57 NF CS-T58 CS-T58 318: 294 CS-T58 NF;NF CA NF;NF CS-M NF AC-C NF;NF PR NF;NF PO NF;NF PA-C NF;NF NM NF;NF PA-M NF NC NF;NF DU 4:2;2:2 CS-T57 393:325 AR 26:0;19:0 CS-M 403:320 The University of Texas Publication intermediate phenotypes; however, most were more similar to AC-C than to NC. Labellum: Most had a much reduced acutilabella-like anteroventral process; a slight degree of variation between individuals was also seen in this characteristic. Palpi: Shape and bristles much like D. acutilabella. AC-C