No. 4720 May 22, 1947 STUDIES IN THE GENETICS OF DROSOPHILA V. ISOLA TING MECHANISMS Directed by J. T. PATTERSON Professor of Zoology The University of Texas PUBLISHED BY THE UNIVERSITY OF TEXAS AUSTIN Copies of this publication may be procured for $2.00 each from the University Publications, The University of Texas, Austin, Texas. The University of Texas Publication No. 4720: May 22, 1947 STUDIES IN THE GENETICS OF DROSOPHILA V. ISOLATING MECHANISMS Directed by J. T. PATTERSON Professor of Zoology The University of Texas PUBLISHED BY THE UNIVERSITY FOUR TIMES A MONTH AND ENTERED AS SECOND·CLASS MATTER AT THE POST OFFICE AT AUSTIN, TEXAS, UNDER THE ACT OF AUGUST ll.41 1!111~ The benefits of education and of useful knowledge, generally_ diffused through a community, are essential to the preservation of 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 CONTENTS I. Sexual Isolation Between Members of the Virilis Group of Species -------------------------------------------------------------------------------------------7 J. T. Patterson, Linda Wharton McDanald, and Wilson S. Stone. II. Sexual Isolation in the Mulleri Subgroup_______________ __ ____________________ 32 J. T. Patterson III. The Insemination Rea.ction and Its Bearing on the Problem of Speciation in the Mulleri Subgroup_____.___ _________ ____________ ____ 41 J. T. Patterson IV. The Insemination Reactio'n in Intraspecific Matings of Dro­ sophila --------------------------------------------------------------------------------------------78 Marshall R. Wheeler V. Two Strains of Drosophila Peninsularis With Incipient Repro­ductive Isolation ----------------------------------------------------------------------------116 J. T. Patterson and Marshall R. Wheeler VI. A Study of the Isolating Mechanisms Found in Drosophila Arizonensis and Drosophila Mojavensis ----------------------------------126 William K. Baker VII. Relationships Within the Quinaria Species Group of Dro­ sophila -· ------------------------------------------------------------------------------------------137 Jae~ W. Sears VIII. The Species :Relationships in the Virilis Group___________________________ _ 157 Wilson S. Stone and J. T. Patterson IX. Gene Replacement in the Virilis Group___________________________________________ 161 Wilson S. Stone · X. Interrelationship Between Eye Color and Facet Arrangementin Lozenge Alleles of Drosophila Melanogaster_____ ____ __ _________ 1'67 C. P. Oliver PREFACE The ten articles included in this bulletin constitute number V of the series of publications to be issued from our genetics laboratory. As indi­cated in the subtitle, the majority of these articles deal with types of isolating mechanisms which operate in bringing about evolutionary diver­gence, either by reducing or preventing the exchange of genes between the different forms. Such studies now in progress in this and other laboratories are certain to shed much light on the mode of evolution as it occurs in groups of living organisms. We wish to express our appreciation to the University Research Insti­tute, which, under the direction of Dr. A. P. Brogan as Dean of the Graduate School, has furnished the funds for meeting the expenses of publication. J. T. PATTERSON Austin, Texas, February 21, 194 7. I. SEXUAL ISOLATION BETWEEN MEMBERS OF THE VIRILIS GROUP OF SPECIES J. T. PATTERSON, LINDA WHARTON McDANALD, WILSON S. STONE INTRODUCTION Dobzhansky (1937) proposed the term "isolating mechanisms" as a generic name for the various agents which prevent interbreeding of groups of individuals. These mechanisms therefore play a very impor­tant role in the evolutionary process. As he points out, it is not necessary to assume that a single mechanism is involved in preventing the inter­breeding of a given· pair of species. As a matter of fact it is usually the total effect of several such mechanisms which finally brings about the more or less · complete isolation of two forms, by reducing or preventing the exchange of genes between them. It has been abundantly demon­strated that this is true in the genus Drosophila. Various mechanisms have been found to occur among members of this genus, but the one most frequently encountered is what has been termed sexual isolation. Sexual isolation is sometimes referred to as psychological isolation, since it involves behavioristic phenomena. In this type of mechanism complete copulation resulting in insemination does not occur frequently between individuals of different species, and sometimes it may not occur between individuals of different geographic strains of the same species. The failure to mate may be due to differences in odor, courtship, behavior, sexual recognition signs, or other possible types of stimulation. It is the purpose of the present article to present the results which the authors have obtained on tests between members of the virilis species group. Three different methods have been employed to determine the possible occurrence of sexual isolation in Drosophila. One of these involves direct observation. By watching cultures containing males and females of different species, it is possible to determine whether crossbreeding is taking place. This method, however, is too tedious and too tb:ne-con­suming where it is desired to obtain data sufficient for statistical pur­poses. This method was employed in many of the early studies on the . mating habits of Drosophila. A second method of testing for sexual isolation is by the use of pair matings. Under this method a number of cultures are made up with each containing a male and a female belonging to different forms (species, subspecies, or strains). In due time the cultures are checked and the number failing to produce offspring recorded. From the data thus obtained the degree of sexual isolation may be determined. This method has been used extensively in the Texas laboratory, and from time to time reports have been made on a series of tests carried out on forms belonging to several different species group (virilis group, Patterson, Stone, and Griffen, 1940, 1942, Patterson and Griff en, 1944; repleta The University of Texas Publication group, Wharton, 1942; mulleri subgroup, Crow, 1942; funebris group, Mainland, 1942; quinaria group, Sears, 1947 in this publication). The third method ·of testing for sexual isolation is what has been termed the multiple choice technique. In employing t}).is method, the usual procedure is to place two kinds of females in a container with males belonging to 'one of the two types of females, and, after a certain elapsed time, the females are dissected to determine whether or not spermatozoa are present in their seminal receptacles. If the length of exposure of the sexes is properly regulated, and provided that not all of the females have been fertilized, it is possible to determine whether the insemination of the two kinds of females has taken place selectively, or at random. If it has occurred selectively, and not at random, sexual isolation will be indicated (Dobzhansky and Mayr 1944). The following investigators have used this method in studying sexual preference among several different Drosophila forms: Dobzhansky and Koller 1938, pseudoobscura subgroup; Stalker 1941, 1942, virilis and americana; Dobzhansky and Mayr 1944, strains of willistoni; Dobzhansky and Streisinger 1944, strains of prosaltans; Dobzhansky 1944, strains of sturtevanti; Mayr and Dobzhansky 1945, pseudoobscura, persimilis, and strains of prosaltans; Levene and Dobzhansky 1945, pseudoobscura and persimilis; Mayr 1'946a, 1946b, pseudoobscura and persimilis; Wallace and Dobzhansky 1946, pseudoobscura, persimilis, and subobscura. As indicated in these references, this method has been used for the study of sexual preference both between different species and between geo­graphic strains of the same species. Sturtevant (1920, 1921) had previously obtained evidence that in mixed cultures of D. melanogaster and D. simulans the flies exhibited preference for mating with members of their own species, and Lancefield ( 1929) found the same was true for races of pseudoobscura. Further comments on the second and third methods of testing for sexual isolation seem desirable. There are several factors which must be considered in any attempt to evaluate the results obtained by the pair-mating method. Not all sterile cultures, upon which the magnitude of sexual isolation is based, are due to a lack of copulation, for at least some of the females from such cultures have been inseminated, as has been shown by the presence of sperm in their seminal receptacles. The failure of inseminated females to produce offspring may be due to any one of several different causes. In interspecific inseminations the sperm may be killed or inactivated in the reproductive tract of the alien female, either immediately or shortly thereafter (Patterson, Stone, and Griffen 1942). Even though in some cases the sperm may succeed in penetrating the egg, yet development does not take place. This represents an initial form of zygotic mortality. Zygotic inviability is very common in inter­specific crosses of Drosophila, due to genie unbalance (Patterson 1942), or to complementary gene systems (Crow 1942, Patterson and Griffen 1944). Studies in the Genetics of Drosophila The third method of multiple choice technique is a test designed for the detection of sexual preference, rather than one intended primarily for use in the discovery of sexual isolation. Sexual isolation if present will usually be revealed in the results obtained in the experiment. Sexual preference may reduce the exchange of genes between two forms, but will not necessarily prevent such exchanges, whereas sexual isolation as strictly defined will prevent these exchanges. In tests between two species with the use of the pair-mating technique, only heterogamic matings are possible, and if none of the females has been inseminated it is evident that complete sexual isolation exists. On the other . hand, in sexual preference tests both homogamic and heterogamic matings are possible. In view of these facts we decided to carry.out a series of tests in which the results from the pair-mating and multiple choice techniques could be compared. Members of the virilis group were selected for the experiments, because we already had accumulated a considerable amount of data on these forms from our pair-mating tests. We also hoped that it might be possible to isolate and study some of the factors underlying sexual isolation, since it seems clear that this phenomenon must be due to genetic factors rather than to extrinsic environmental influences. EXPERIMENTAL METHODS In this study we have used principally the multiple choice technique, and wherever possible have compared the results obtained with those derived from the pair-mating tests for the corresponding crosses. As indicated in the introduction, the multiple choice method has been used by several different workers for studying sexual preference between different species or different geographic strains of the same species. In the recently reported investigations of Dobzhansky and co-workers the practice was followed of placing ten virgin females of each of two species or strains in a food vial together with ten males of one of the forms. After the females had been exposed to the males for a certain period of time they were dissected and their seminal receptacles examined under the microscope for the presence of sperm. From the results thus obtained it is possible to determine whether sexual preference is present. In general we have followed this same procedure, with but one impor­tant exception. In making up the cultures we have used only five females of the two forms under test and a single male of one of the forms. We adopted this method after considerable experimentation with other meth­ods, not only because the results were more consistent, but also because they revealed the performance of a single male. We suspected that in mass cultures not all of the males might possess the genetic factor or factors responsible for sexual preference. Such individual differences, if they exist, would be masked in the results from tests in which several males had been used. It was not considered practical to make similar tests of females, although it could be done by the use of direct observa­ The University of Texas Publication tions. Even in that case the aggressive role of the male in courtship might conceal any preference on the part of the female. However, the results derived by our technique will show the effect of female preference to some degree, as well as male preference. If all of the females were equally responsive to the courtship of the male, only male preference would determine the results. Any rejection of the courtship of the male by . one type of female will reduce the number of matings of that type, . especially in the presence of several females of the other type. In our experiments we used flies which were mature at the beginning · of the test. The length of time for the flies to reach maturity varies in different species, and sometimes even in the two sexes of the same species. In the virilis group, for instance, the females on the average mature and are ready to mate about one day earlier than the males. This fact was taken into consideration in making up the cultures for tlie tests. Once the age of maturity is determined, it becomes a matter of impor­tance to ascertain the proper length of time to expose the sexes before dissecting the females. This period of time is especially important in intraspecific tests between geographic strains, and in those interspecific crosses which yield fairly high degrees of homogamic and heterogamic matings. J:t is somewhat less important in interspecific crosses in which heterogamic matings either do not take place or occur only rarely. In recording the data derived from the experiments, we used long mimeo­graphed sheets ruled with horizonal lines for listing the individual females, and vertical columns for showing the type of cross, age of the female, age of the male, length of exposure, presence of the sperm in the recep­tacles, and the motility ·and quantity of spermatozoa present in the receptacles. The age of the flies at the start of the experiment was given in hours, as was the length of time the males and females were exposed to each other. In attempting to determine whether or not the sperm were alive and motile, it was sometimes · necessary to tap the cover glass with the point ·of a pencil. To determine the exact quantity of sperm present is difficult, and any method adopted to estimate the amount will not be entirely satisfactory. We finally established three categories, under the terms "many," "moderate," or "few." INSEMINATION AND THE DISTRIBUTION AND MOTILITY OF THE SPERM IN THE SEMINAL RECEPTACLES OF WILD STRAINS OF DROSOPHILA FEMALES As a preliminary study to our proposed experiments on sexual isola~ tion; we desired to know more about the degree of insemination in wild strains of Drosophila. We also wanted to learn more about the distribu­tion and motility of the sperm in the seminal receptacles of such females. Accordingly, we collected several of the species which are common in the vicinity of Austin, especially those found at the nearby Aldrich farm. The only species in the list that was not collected in the country was D. repleta. This form rarely frequents wild habitats, and we there­ Studies in the Genetics of Drosophila fore collected it at one of the produce houses in Austin where it was known to live and breed. As soon as the collections were brought into the laboratory, the females were isolated and placed separately in food vials. No selection was practiced in isolating these females, other than to exclude specimens which had been injured by the collecting net. These cultures were held long enough to allow development to occur in case the females had been inseminated in the field. The length of time to hold the vials would obviously depend on the length of the life cycle of any given species. At the end of this period the cultures were examined and the number containing offspring was used as a basis for determining the per cent of fertile females. The females from the sterile cultures were then dissected and their seminal repeptacles examined for the presence_of sperm. The number containing sperm was added to -the • number fertile to give the per cent inseminated. This procedure obviates the necessity for large-scale dissections, and in addition to giving 'the per cent inseminated, also shows the per cent of fertile females. In Table 1 are listed the data tor thirteen wild-type species, or rather species collected in the wild. The first column, following the name of the species, gives the total number of females isolated and tested. The second column gives the per ce:nt of fertile females, that is, females which actually produced offspring.•The third column shows the number of dissected females from sterile cultures. The fourth column gives the number of these females which contained sperm, and the fifth gives the per cent inseminated. The last column gives the per cent -of each species in the population at t~e time the collections were made. TABLE 1 Per cent of inseminated females in wild strains of Drosophila per cent per cent Species females isolated per cent fertile females dissected females with sperm insem­inated in popu­lation D. busckii___________________ 51 96.0 2 1 98.0 1.8 D. putrida_____________ __ ___ 203 92.1 16 1 92.6 15.4 D. melanogaster_________ D. simulans____________ ____ 547 302 95.0 92.0 27 24 3 3 95.6 93.0 23.2 18.1 D. affinis____________________ 88 84.l 14 0 84.1 5.5 D. pseudoobscura______ 331 96.6 11 5 98.5 27.0 D. tripunctata____________ 42 92.8 3 0 92.8 0.89 D. macrospina___________ 105 83.8 19 3 86.6 4.8 D. repleta__ _ ______________ 152 73.0 41 32 94.0 18.7 D. hydei_______ __ ___________ 552 80.8 106 55 90.7 23.4 D. mulleri___ _____________ 201 50.2 100 40 70.1 12.7 D. longicornis____________ 130 30.8 88 48 67.7 11.6 D. m eridiana_, ________ ____ 63 85.7 9 5 93.6 3.2 In the first eight species listed in the table the percentage fertile is very close to the percentage inseminated, due to the fact that only a few of the females from sterile cultures possessed sperm. In two cases (affinis, tripunctata) all inseminated females produced offspring. In the next four species, all members of the repleta group, there is a wide divergence between the per cent inseminated and the per cent fertile. Just why so many of these females with sperm in their receptacles should fail to produce offspring is not clear, unless it be due to the laboratory . food. It should be stated that every chance was given these females to breed. In most cases they were held for three weeks and over, with several changes to fresh food, and still no off spring appeared in the cul­tures. The last species on the list is D. meridiana, also a member of the repleta group. The difference between the two percentages of this form is not as striking as for the other four members of the group, but it is still greater than that of any of the first eight species. Drosophila longicornis shows the greatest difference between the two percentages of any of the thirteen species. It also has the lowest fer­tility percentage. The fact that only about half of the females brought in from the field were found to be inseminated may be due to the long maturing period of this species. Another point of interest is the apparent lack of correlation between the per cent inseminated and the per cent in the population of this group of species. There seems to be very little correlation between these two percentages. For example, for D. busckii the percentages are 98 and 1.8, respectively, whereas for D. pseudoobscura they are 98.5 and 27; again, for D. meridiana they are 93 and 18.1. Although in some cases the per­centages are nearly equal (e.g., simulans and repleta), yet on the whole there is very little correlation between the density of population and the degree of insemination. In Table 2 are presented the f::i,cts concerning the distribution and motility of the sperm in inseminated females. The data are presented under two headings for each of the thirteen species, for fertile and sterile females. The data in the fertile series were obtained from dissec­tions of about fifty specimens, whenever that number was available. The only exception is D. pseudoobscura for which dissections were not made. The data for the sterile series are rather meager, since so many of the females from the sterile cultures did not contain sperm. In the fertile series sperm were present both in the spermathecae and the ventral receptacle in all except two species. The two exceptions were D. mulleri and D. longicornis, in which sperm were present in the ventral receptacle only. These two forms belong to the mulleri subgroup of species in which the spermathecae do not function as sperm receptacles, but have evolved into glandular structures. At the bottom of the table are given the totals for the fertile and sterile cultures, together with their grand totals. These figures present. some interesting facts with reference to the distribution and motility of the sperm. In the fertile series there are 281 cases of motile sperm as against 214 non-motile in the spermathecae, while in the sterile series there were 12 motile and 94 non-motile cases for the same organ. The numbers of motile and non-motile cases in the fertile series were 51'0 and 47, respectively, for the ventral. receptacle, or a ratio of about Studies in the Genetics of Drosophila eleven to one. In the sterile series there were 99 motile and 30 non­motile cases for the ventral receptacle, or a ratio of slightly more than three to one. TABLE 2 Distribution and motility of sperm in fertile and sterile females Spermathecae recep­tacle Species Type of culture motile non-motile Ventral motile non-motile Sperm absent Totals D. busckii__________________ __ ________ __ fertile 35 13 43 5 0 48 sterile 0 l 0 1 1 2 D. putrida_____ _______________________ ___ fertile 53 6 58 1 0 59 sterile 1 0 1 0 15 16 D. melanogaster____________________ fertile 29 43 65 7 0 72 sterile 1 2 3 0 24 27 D. sirnulans ··--------·--------­-------· fertile 38 37 64 11 0 75 sterile 1 2 2 1 21 24 D. a!Jin.is.....·--·---­·------·---------­ fertile 25 10 36 2 0 38 sterile 0 0 0 0 14 14 D. pseudoobscura......._____ __·-· fertile not dissected sterile 2 3 5 0 6 11 D. tripunctata__________ _____ _______ _ fertile 11 2 13 0 0 l3 sterile 0 0 0 0 3 3 D. macrospin.a ....................... fertile 54 2 49 6 0 57 sterile 2 1 3 0 16 19 D. repleta...·-----·-···-------·--------· fertile 8 44 37 5 0 52 sterile 1 31 7 6 9 41 D. hydei ·--­------------­·-------------­ fertile sterile 0 1 52 54 10 5 15 5 0 51 53 106 D. rnulleri ····--·­---········----·--·--· fertile 0 0 49 1 0 50 sterile 0 0 28 14 58 100 D. longicomis.............------·-­-­ fertile 0 0 36 4 0 40 sterile 0 0 40 3 45 48 D. rneridiana........·--------------·-· fertile 28 5 50 0 0 50 sterile 3 0 5 0 4 9 Total, fertile series_____ __________________________ 281 214 510 47 0 607 Total, sterile series ------------------------·-------12 94, 99 30 267 460 Grand total ----------------------------·--····------·-293 308 609 87 267 1,067 These figures establish the fact that in fertile females there is a larger proportion of motile sperm in the ventral receptacle than is the case for the spermathecae. In contract to this condition, in sterile females the proportion of motile to non-motile cases is reversed for the sperma­thecae (12 vs. 94), while for the ventral receptacle the proportion is still in favor of the motile cases, yet the ratio is much less than for fertile females. We shall delay any further discussion of the possible significance of these data until a later section where larger numbers will be available. The University of Texas Publication TESTS FOR SEXUAL PREFERENCE AMONG MEMBERS OF THE VIRILIS SPECIES GROUP There are six known members of the virilis species group, and five of these were available for the study of sexual preference. The five forms are D. virilis Sturtevant, D~ americana americana Spencer, D. americana texana* Patterson, Stone, and Griffen, D. montana Patterson and Wheeler, and D. lacicola Patterson. Two different series of tests were carried out, one on interspecific crosses between the five forms, and one on intra­specific crosses between a selected list of geographic strains of three of the forms. There are twenty possible reciprocal crosses between the five forms, using a single strain for each form (Table 4). For each cross we dissected at least one hundred females for each of the two types of matings. The numbep dissected f9r the two types were usually not equal, due to the fact that some few females die during the course of the experiment, or are otherwise lost. The terms homogamic and heterogamic matings are employed in the same sense as was used by Dobzhansky and co-workers. The former term includes all matings occurring between the male and the females of his own species, subspecies, or strains, while the latter term includes all matings between the male and the alien females. The strains used for the five different species were, virilis, Pasadena strain, americana, a strain from Anderson, Indiana, texana, strain 1128.10 from New Orleans, Louisiana, montana, strain 1318.8a from Hart Prairie, Arizona, and lacicola, strain 1360.2 from Fairbanks, Minnesota. The values of x2 were calculated by the use of the fourfold table, and those for the isolation index by the formula developed by Stalker (1942). It was found necessary to adjust the length of exposure and the age of the flies in making tests for sexual preference, otherwise the results will not be entirely consistent. Two different series were carried out for the purpose of demonstrating this fact. Reciprocal crosses of virilis and americana were made and the cultures divided into two groups, one of which was kept at a temperature of 72° F., the other at a tempera­ture of 62° F. The form1~r represents .the temperature of the air-con­ditioned laboratory. The flies were collected once each day and aged in the laboratory for ninety-six hours before making the crosses . . Five fem ales of each species were placed with one male of one of the forms in food vials. The females were dissected at 24-hour . intervals (using ten cultures for each interval), and their sperm receptacles examined for the presence of sperm. Fifty females were therefore dissected for each possible homogamic and heterogamic mating. The number of females inseminated is recorded in per cents (Table 3). The results from the seventy-two degree series are listed in the upper half of the table, and those of the sixty-two degree series in the lower_ half. In *We now regard americana and teiana as constituting a pair of subspecies (see Article VIII). Studies in the Genetics of Drosophila referring to these data, we shall use the term "interval" to cover each of the 24-hour exposure periods. Thus the period from f-24 hours will be referred.to as the 24-hour interval, that of 25-48 hours a's the 48-hour interval, and so on through the series. TABLE 3 Effects of temperature and length of exposure on the per cent inseminated in virilis/americana crosses (5V ~ ~ +5A ~ ~ ) X V 5 (5A ~ ~ +5V ~ ~) X A 5 Hours per cent of per cent of per cent of per cent of Temperature exposed homogamic heterogamic homogarnic heterogamic 24 0 0 0 0 48 8 0 0 0 72 10 0 6 0 72 degrees % 94 4 62 6 Fahrenheit 120 86 12 88 16 144 100 4 42 16 168 90 32 42 16 24 0 0 0 0 48 0 0 0 0 72 0 0 0 0 62 degrees % 0 0 0 0 Fahrenheit 120 0 0 0 0 144 6 0 0 0 168 52 0 0 0 192 56 0 0 0 At seventy-two degrees temperature, with virilis used as the male parent the homogamic matings began to appear during the 48-hour interval, whereas the heterogamic matings did not start until sometime during the 96-hour interval. For these three intervals, 56 of the 150 ·dissected virilis females' had be.en inseminated, or 37:3%, while only two of the 150 dissected americana females contained ,sperm, . for a percentil.ge ·of 1.3. For the 120-hour interval there were 86 % of homogamic matings and 12% of heterogamic. For the 144-hour ·interval the per cent· of homogamic matings rose to 100, while that of the heterogamic fell to 4. Finally, in the last or 168..:hour interval the two percentages 'were 90 and 32, respectively. In order to understand these variations in the inseminatio~ percent­ages, it is necessary to refer to our recorded observations made during the course of the dissections. It was observed for the 96-hour .interval that a large majority of the females contained many motile sperm in their seminal receptacles. In contrast · to this condition, most of the females of the 120-hour interval, especially the virilis ones, had very .few sperm in these receptacles, frequently making it necessary to examine minutely both the ventral receptacle and the spermathecae in order to determine whether a few sperm might not be present. For the 1'44-hour interval all of the virilis females were fertile; and a majority of them contained many motile sperm. Finally, for the 168-hour interval not as large a proportion of the virilis females had as many sperm as those for The University of Texas Publication the 144-hour interval, but most of the americana females did have many sperm in their receptacles. The inference from these observations is that the virilis males first mate with their own kind of female, and do not copulate with the alien females until most of the conspecific females have been inseminated. This is almost certainly true up to the beginning of the 144-hour inter­ val, but during the latter period they again mate with their own kind of females, probably after such females have used up most of the sperm. It is possible that the seven virilis females, which had no sperm at the 120-hour interval, had already exhausted their supply of sperm at the time the dissections were made. In the reciprocal cross, with americana used as the male parent, none of the females were inseminated during the first two intervals. This _is undoubtedly due to the fact that the flies were too immature, as they were for the first interval of the preceding series. During the 72-hour interval 6 % of homogamic matings occurred, and none of heterogamic. Sixty-two per cent of homogamic and 6%of heterogamic matings occurred up to end of the 96-hour interval, and 88 % and 16% , respectively, for the 120-hour interval. For the last two intervals these two per cents were identical, 42 % homogamic and 16% heterogamic. The same variations in per cents inseminated in this series of tests up to the end of the 120­hour interval were observed as for the corresponding intervals of the first series, and the same explanation may be offered for both. There is considerable difference in the per cent inseminated in the two series for the last two intervals, due in part at least to differences in the two species. Of the two forms, americana is slower in maturing,. is less fertile, and on the whole is somewhat less vigorous than virilis. It was recorded that the americana females of the homogamic matings for the last two intervals contained relatively few sperm in their seminal recep­tacles, indicating that remating had not taken place. The results from the tests on flies exposed to the sixty-two degrees temperature are listed in the lower half of Table 3. In the cross in which virilis was used as the male parent no inseminations was found until the 144-hour interval. During this period homogamic matings alone occurred, to the extent of 6 % . The same was true for the 168-and 192­hour intervals, which gave 52% and 56 % of homogamic matings, respec­tively. Not a single inseminated female was found throughout the en­tire series in which americana was used as the male parent. These results demonstrate that temperature does modify the degree of sexual isolation. The most obvious effect is the striking retardation of copulation. In the cross in which virilis was used as the male parent, homogamic matings began during the 48-hour interval at 72°, and not until the 144-hour interval at 62°. Likewise, the heterogamic matings began during the 96-hour interval at 72°, and not at all at 62°. Therefore, at the lower temperature virilis males prefer their own kind of female to the exclusion of the alien females. In the reciprocal cross, with ameri­ Studies in the Genetics of Drosophila cana used as the male parent, homogamic matings began during the 72-hour interv~l at 72°' and the heterogamic matings during the 96-hour interval. No matings of either kind occurred at 62° temperature. Since none of the americana females in either cross was inseminated at the lower temperature, it is possible that these females failed to respond to the courtship of either type of male. The above reported tests were carried out for the purpose of augment­ing certain results ·that had been obtained in a series of tests carried out at the beginning of this study. The tests were performed with the view of determining the optimum time of exposure, as well as to determine the best age at which to mate the flies. The results demonstrate that if consistent data are to be obtained, it is necessary to mate the virilis flies after they have reached maturity, and then expose the sexes to each other for a definite period of time. Any wide deviations in the lengths of these two periods may modify the results appreciably, especially in intraspecific crosses in which the per cents of homogamic and hetero­gamic matings are rather high. A single example will make this point clear. If two geographic strains of virilis under test show a significant difference in favor of homogamic matings after ·seventy-two hours of exposure, this difference may be reduced to a point at which it is no longer significant upon prolonging the length of exp0sure from twenty­four to forty-eight hours. This does not mean that sexual preference is absent, rather it signifies that the males prefer to mate with their own kind first before inseminating the alien females. · We finally adopted the plan of aging the flies from 120-144 hours, and then exposing them for seventy-two hours. This plan was followed, with but few minor excep­tions, throughout the experiments described below. The results obtained from the sexual preference tests in interspecific crosses are recorded in Table 4. In six of the crosses (1, 11, 13, f4, 17, 19) the isolation index is 1.00, which indicates that sexual isolation is com­plete. There are seven other crosses (3, 5, 7, 12, 15, 16, 18) in each of which the isolation index is <;lose to unity and sexual isolation is nearly complete. Even in the cross with the lowest isolation index (8, lacicola/ 2 virilis) the xvalue of 8.02 and the isolation index of 0.29 both indicate that the difference between homogamic and heterogamic matings is significant . x2 . In the americana/texana cross (10) the value and the isolation index show that the difference is significant. But the isolation index is negative (-0.28), which means that texana males prefer americana females to their own kind where choice of mate is possible. At the time this result was first obtained it seemed so unusual that the test was repeated. The percentages obtained in the second test were practically identical with those of the first test. We later obtained several cases of negative isolation indices involving geographic strains of the same species, and in some of these the differences were significant (Table 9). More­over, Dobzhansky et al. .have reported such indices in crosses between The University of Texas Publication geographic strains. In this connection it should be pointed out that texana and americana are now regarded as constituting a pair of sub­species. T ABLE 4 Number of females dissected (n) and per cent inseminated (% ) m various interspecific crosses of the virilis group Iso­Homogamic Heterogamic lation Females Males n % n % index 1. virilis, americana ___________________ virilis 102 92.1 103 0.0 175.3 1.00 2. 3. americana, virilis -------------------­virilis, texana____________________________ americana virilis 105 101 49.5 74.2 106 103 24.6 2.9 14.l 109.9 0.33 0.92 4. texana, virilis ____________________________ texana 101 58.4 102 26.4 21.2 0.37 5. virilis, montana ________________________ virilis 109 80.7 115 1.8 145.3 0.95 6. montana, virilis _______________________ montana 111 58.5 108 14.8 44.9 0.69 7. virilis, lacicola _________________________ virilis 102 71.5 102 0.98 109.9 0.97 8. lacicola, virilis __________________________ lacicola 102 39.2 108 21.3 8.02 0.29 9. americana, texana____________________ americana 115 60.8 115 31.3 20.2 0.32 10. texana, americana____________________ texana 111 30.6 112 55.3 13.9 -0.28 11. americana, montana________________ americana 120 50.0 102 0.0 70.6 1.00 12. montana, americana________________ montana 109 67.8 115 0.86 112.9 0.97 13. americana, lacicola _________________ americana 107 50.4 109 0.0 92.6 1.00 14. 15. lacicola, americana ---------------­texana, montana______________________ lacicola texana 100 107 75.0 83.1 100 104 0.0 0.96 120.0 145.7 1.00 0.97 16. montana, texana_____________________ montana 104 69.0 102 0.98 104.8 0.97 17. texana, lacicola ________________________ texana 101 60.4 102 0.0 88.1 1.00 18. lacicola, texana ----------------------· lacicola 100 50.0 100 2.0 61.6 0.92 19. 20. montana, lacicola ----------··-------­lacicola, montana ___________________ montana lacicola 100 111 . 52.0 73.8 105 105 0.0 6.6 69.6 100.6 1.00 0.83 The conclusion to be drawn from the results obtained in this series of sexual preference tests is that sexual isolation is a very important isolating mechanism in preventing the exchange of genes between the different members of the virilis group. For purposes of comparison we have brought together certain experi­mental results which are shown in Table 5. The figures represent per­centages and are listed under three main headings of heterogamic, homo­gamic, and paired mating3. The percentages under each heading are given in two columns, in one of which the male had a choice of mate and in the other no such choice was possible. In the heterogamic and homogamic series ten females were exposed to a single male for seventy­two hours and then dissected. In the column under the subhead "with choice" five females each of two forms were used, while under the column "without choice" all of the females belonged to the same form. In the pair mating ·series the percentages were based on the number of females producing progeny, and in the homogamic column the males and females used belonged to the same form, but in the heterogamic column they belonged to different forms. A comparison of the results listed in the six columns in the table allows us to determine the extent of variation in mating under the several different conditions. In comparing the same homogamic mating with and without choice. from columns III and IV, we find that usually the results are not statistically different. In fact, only crosses 1, 8, 15, and 20 show a statistically significant difference. Other conditions make it improbable that these differences are due to the choice or lack of choice of mates available. For example, in cross 1 the per cent (92.1) of virilis females mated with choice is significantly higher than the per cent (73.0) mated without choice. However, in crosses 3, 5, and 7 the per cents mated with choice (74.2, 80.7, 71.5) are not significantly higher than the per cent mated without choice (73.0). Probably some factor other than choice, or the lack of it, influenced the higher per cent in cross 1. Cross 15 appears to be a similar case with texana (compare with 4, 10, 17). Crosses 8 and 20, which are statistically different, prob­ably represent fluctuations of lacicola for causes other than choice or lack of choice (c.f., f4, 18). Per cents Type of matings Crosses L (V'i' +A'i' ) x V t 2. (A'i' +V'i') x A t 3. (V'i'+T'i') X V t 4. (T'i'+ Ve;>) X T t 5. (V 'i' + M 'i' ) X V t 6. (M'i'+V'i') x M t 7. (V'i'+L'i') X V t 8. (L'i' +V'i') x L t 9. (A'i'+T'i') X A t 10. (T'i'+A'i') X T t lL (A 'i' + M 'i') X A t 12. (M'i' +A'i') X M t 13. (A'i'+L'i') X A t 14. (L'i' + A'i') X L t 15. (T 'i' + M 'i' ) X T t 16. (M'i'+T'i') X M t 17. (T'i'+L'i') X T t 18. (L'i' + Tc;>) X L t 19. (M'i' +L'i') x M t 20. (L'i'+M'i') X L t TABLE 5 inseminated in homogamic and heterogamic matings in the virilis series of crosses Ten 'i' 'i' X one t One 'i' X one t Heterogamic matings Homogamic matings Pair matings with without with without homo­hetero­choice choice choice choice gamic gamic 0.0 3.8 92.l no 95.0 8-0 24.6 24.0 49.5 52.0 70.0 23.0 2.9 mo 74.2 no 95.0 14.0 26.4 37.0 58.4 54.0 87-0 36.0 ----··· LB 3.0 80.7 73.0 95.0 0.0 -----­14.8 37.0 58.5 65.6 87.0 54.0 0.98 LO 7L5 73.0 95.0 0.0 -------­2L3 16.5 39.2 55.0 75.0 26.0 -------­31.3 26.0 60.8 52.0 70.0 48.0 -------­30.6 44.0 55.3 54.0 87.0 25.0 ····---­0.0 0.0 50.0 52.0 70.0 2.0 -----­0.86 0.0 67.8 65.6 87.0 0.0 -------­0.0 0.0 50.4 52.0 70.0 0.0 -------­0.0 0.0 68.4 55.0 75.0 0.0 -------­0.96 1.0 83.1 54.0 87.0 0.0 -----­0.98 LO 69.0 65.6 87.0 2.0 ---­---­0.0 0.0 60.4 54.0 87.0 0.0 -------­2.0 0.0 50.0 55.0 75.0 12.0 -----­0.0 2.0 52.0 65.6 87.0 14.0 ----­6.6 29.0 73.8 55.0 75.0 12.0 Columns -----------------------------­ II III IV v VI In comparing like crosses in columns I and II only two crosses, 6 and . 20, show a statistically significant difference. These involve lacicola and montana which themselves show considerable fluctuation. It is probable, therefore, that no real differences exist between these two columns. When we compare columns IV and V, we find that in every case the per cent of females producing offspring (V) is higher than the per cent of females inseminated in homogamic matings (IV or III), and is usually_ The University of Texas Publication siguificantly higher. This undoubtedly represents a real difference, and must depend on the longer time of opportunity to mate. In fact, the iength of the period of exposure of females to males, usually 72 hours, was held to this short time in order to allow the expression of sexual preference, as longer exposures often resulted in so many females becom­ing inseminated that no choice could be detected. A .comparirnp between columns II (or I) and VI does not show any appreciable increase in the pair matings (VI). This point is of con­siderable interest when we compare it to the comparable results from homogamic matings. Apparently the factors preventing cross insemina­tion are not those of time, choice, or opportunity, such as influence homogamic matings. This makes such measurements of sexual prefer­ence or sexual isolation between species of even more evolutionary significance. When we compare columns I with Ill, II with IV, and V with VI, there is never any question of the significance of difference between hetero­gamic and homogamic matings. The only exception to statistically sig­nificant difference are those where it is expected, that is, between the subspecies texana and americana. At the time the sexual preference tests were being carried out a com­plete record was kept of the distribution and motility of the spermatoza in the seminal receptacles of all dissected females. These records are tabulated in Table 6, and are shown under two main headings, the paired spermathecae and the single ventral receptacle. Under each of these headings the numbers containing motile and non-motile sperm are given for both homogamic and heterogamic matings. Since none of the females in six of the heterogamic matings had been inseminated, the table lists but thirty-four of the forty possible combinations. There are· some variations in the distribution of the two types of sp~rm among the different crosses, but we are interested in the totals as 'given at the bottom of the table. For the paired spermathecae there were 817 fetnales with motile and 645 with non-motile sperm. This gives a percentage of 55.9 in favor of the motile type. For the si_ngle ventral receptacle there were 1,327 females with motile. and 151 with non-motile sperm, for a percentage of 89.8 in favor of the motile type. The ventral receptacle contains a much larger proportion of motile sperm than the spermathecae. These totals may be separated into their two natural groups of homo­gamic and heterogamic. The percentages of motile sperm in the homo­gamic series were 55.7 for the spermathecae and 91.9 for the ventral receptacle, and those in the heterogamic series were 57.1 and 76.2, re­spectively. The percentages of motile sperm in the spermathecae for the two sets of data (homogamic, heterogamic) are very similar (55.7, 57.1), but in the ventral receptacle the percentage for the homogamic is significantly higher (x2 = 46.8) than that of the heterogamic (91.9 vs. 76.2). . Studies in the Genetics of Drosophila 2l'. TABLE 6 Distribution and motility of sperm in the female receptacles in crosses between members of the virilis group Female Spermathecae Ventral receptacle Crosses dissected motile non-motile motile non-motile (V'i'+A'i') x V t .................. virilis 68 26 93 0 (A'i'+V'i') X A t ................ americana 29 14 46 7 virilis 12 13 20 5 (V'i'+T'i' ) X V t .................. virilis 34 41 71 3 texana 1 2 1 1 (T'i'+V'i') X T t ................... texana 23 29 53 6 virilis 9 15 21 5 (V'i'+M 'i') X V t .................. virilis 48 37 79 5 montana 1 0 2 0 (M 'i'+V'i') X M t ................. montana 33 32 59 4 virilis 5 9 7 7 (V'i'+L'i') X V t ................... virilis 30 38 67 4 lacicola 0 1 0 1 (L'i'+V'i' ) X L t ................... lacicola 9 25 32 2 virilis 18 4 16 6 (A'i'+ T'i') X A t ................. americana 58 14 69 3 texana 27 9 30 6 (T'i'+A'i') X T t ........ ---······­ texana 19 13 27 6 americana 26 26 47 16 (A'i' +M'i' ) X A t ................. americana 31 29 55 4 (M'i'+A'i') X M t ................ montana 60 14 64 4 americana 1 0 0 0 (A'i'+L'i' ) X A t .................. americana 21 20 40 3 (L'i'+A'i') X L t .................. lacicola 44 32 61 13 (T'i'+M'i') X T t .................. texana 47 40 75 14 montana 1 0 1 0 (M'i' + T'i' ) X M t ................ montana texana 42 0 30 1 67 1 3 0 (T'i'+L'i') X T t .................. texana 27 32 57 4 (L'i'+T'i') X L t ................... lacicola texana 22 1 29 0 46 1 2 1 (M'2+L'i') X M t ................. montana 21 31 46 6 (Lc;i+M'i' ) X L t .................. lacicola montana 43 6 38 1 66 7 10 0 Totals .................................................................... 817 645 1,327 151 55.9% 44.lo/o 89.8% 10.2% Homogamic .......................................................... 709 564 1,173 103 55.7% 44.3% 91.9% 8.lo/o Heterogamic -------------------------------------------------------· 108 81 154 48 57.1% 4·2.9% 76.2% 23.8% As we have already stated, the matter of determining the quantity of sperm present in the seminal receptacles of inseminated females is very difficult. To attempt to count the sperm in large scale experiments would have been an impossible task. We therefore established three categories of classification, many, moderate, and few. We placed in the first cate­gory all cases in which the receptacle was at least one-third full of sperm; in the second category all cases in which the receptacle was The University of Texas Publication approximately one-fourth full; and in the third category all cases in which the receptacle was less than a fourth full. In the third category many cases were observed in which the number of sperm might have been counted. At best such a method can give only a rough estimate of the quantity of sperm present in the seminal receptacles. ' The data on the quantity of sperm for the thirty-four combinations are tabulated in Table 7. We shall again consider the totals group first. TABLE 7 Distribution and quantity of sperm in the female receptacles in crosses between members of the virilis group Female Spermathecae Ventral receptacle Crosses dissected many moderate few many moderate few (V'i' + A'i') X V J ______ virilis 85 1 8 86 2 5 (A'i' +Ve,>) X A J ______ americana 33 3 7 37 4 12 virilis 9 2 12 10 2 11 (V'i' +T'i') XV J ______ virilis 62 8 5 57 10 7 texana 2 0 1 0 0 2 (T'i' + V'i') X T J ______ texana 38 7 7 46 5 s virilis 11 6 7 8 8 10 (V 'i' + M 'i' ) X V $ -----virilis 52 18 15 51 17 16 montana 1 0 0 1 0 1 (M 'i' +Ve,>) X MJ ____ montana 39 11 15 44 7 12 virilis 5 2 7 3 4 7 (V 'i' + L 'i') XV$------virilis 53 6 8 52 8 9 lacicola 1 0 0 0 .0 1 (1 'i' +Ve,>) X L J ______ lacicola 30 3 1 2S 2 2 virilis 18 0 4 6 5 11 (A 'i' + T 'i' ) X A$-----­americana 67 2 3 69 2 1 texana 31 2 ,3 33 2 1 (Tc,>+ A 'i' ) X T $ -----­texana 24 4 8 28 4 1 americana 41 3 8 51 6 6 (A 'i' + M 'i' ) X A $-----americana 42 12 6 15 2 0 (M 'i' +A 'i') X MJ ____ , montana 58 8 8 57 2 9 americana 0 1 0 0 0 0 (A'i'+ L'i') X AJ ______ americana 32 0 9 33 2 8 (L'i'+ A'i') X L J __ ___ lacicola · 71 3 2 62 3 9 (T'i'+ M'i') X T J _____ texana 77 3 7 82 3, 4 montana 1 0 0 1 0 0 (M 'i' + T 'i' ) X M $ ---­montana 49 8 15 52 7 11 texana 0 1 0 0 0 1 (T 'i' +L 'i' ) X T $ ------texana 53 3 3 52 3 6 (L 'i' + T'i') X LJ ______ lacicola 50 0 1 40 1 6 texana 1 0 0 1 0 1 (M 'i' + L 'i') X M J.-__ montana 46 3 3 44 4 .4 (L'i'+ Mc,>) X L J _____ lacicola 71 6 4 61 7 8 montana 4 2 1 4 3 0 Totals -------------------------------------------------------1,157 128 178 1,114 125 181 79.1% 8.7% 12.2% 78.5% 8.8% 12.7% Homogamic ---------------------------------------------1,022 109 135 996 95 129 80.7% 8.7% 10.6% 81.6% 7.8% 10.6% Heterogamic ---------------------;---------------------125 19 43 118 30 52 66.8% 10.2% 23.0% 59.0% 15.0% 26.0% Studies in the Genetics of Drosophila These show a very close similarity in the three categories of both the spermathecae and the ventral receptacle. The deviation is less than one per cent for the corresponding categories. The per cents are also very similar for the three categories of the homogamic series, and the devia­tion is likewise less than one per cent for the corresponding categories. The greatest difference is seen in the heterogamic series. This can be shown by using the number of specimens in the first or many category and the sum of the other two categories for the homogamic and hetero­gamic series. The totals series may be omitted, since it is a mixture of the other two. The calculations by the chi square method gives x2 values of 18.9 for the spermathecae and 49.0 for the ventral receptacles, both of which are significant. There are three possible explanations for the observed differences between the homogamic and heterogamic series. In the first place, it is possible that in heterogamic matings the males do not deliver as many sperm as in homogamic matings. In the second place, in forms which have the insemination reaction, fewer sperm succeed in entering the seminal receptacles in heterogamic crosses, due to disturbances in the female reproductive tract. Finally, in heterogamic matings the sperm may soon become inactivated and may eventually disappear altogether in the tract of the alien female. The second or third alternatives, or better still, a combination of the two, would seem to offer the best ex­planation. Repeated matings may also be effective. It was observed during the course of the dissections that the sperm were not always present in both receptacles: They were sometimes absent from the ventral receptacle or the spermathecae. These data are shown in Table 8 for the thirty-four combinations. The total number of inseminated females was 1,625, of which 1,516 had sperm in both recep­tacles, 62 in the ventral receptacle only, and 47 in the spermathecae only. In the homogamic series, there were 1,448 specimens, of which 1,357 had sperm in both receptacles, 52 in the ventral receptacle only, and 39 in the spermathecae only. In the heterogamic series these num­bers were 159, 10, and 8, respectively. As shown in the last column, there were 79 females which had sperm in only one of the spermathecae. There is one other point that should be mentioned, and this applies equally well to the data given in Tables 7 and 8. The quality of sperm present in the receptacles at the time of dissection must be influenced by the amount of sperm already used up in case the female had been laying eggs. However, this does not detract from the significance of the differ­ence between the amount of sperm present in the homogamic and hetero­gamic matings. In fact, females in homogamic matings laid more readily than those in heterogamic matings whenever any difference existed. The University of Texas Publication TABLE 8 Variation in distribution of sperm 111 female receptades Ventral Crosses Dissected Both receptacle Spermathe- One sperma­ female receptacles only cae only thecae only x v e; __________________ (V'i'+A'i') virilis 94 0 0 1 __________________ amencana 10 0 (A'i'+V'i') X A i!; 47 6 virilis 24 1 2 2 (V'i'+T'i') X V i!i------------------· virilis 75 1 0 1 texana 2 0 1 1 __________________ texana 7 0 (T'i'+V'i') X T i!; 52 1 virilis 23 3 1 1 x v e; _________________. (V<;>+M'i') virilis 80 3 3 3 montana 1 1 0 0 _________________ montana 0 2 4 (M'i'+V'i') xMe; 62 virilis 12 2 2 1 x v e; __________________. 2 (V'i'+L'i') virilis 65 5 2 lacicola 1 0 0 0 ___________________ 140 8 2 (L'i'+V'i') X L i!; lacicola 0 virilis 21 0 0 1 CA'i'+T'i') x A i!i ------------------americana 69 1 2 5 texana 36 1 1 2 (T'i'+A'i') X T i!; texana 50 12 1 4 americana 29 1 0 5 (A'i'+M'i') xAi!i -----------------americana 51 1 0 12 (M'i'+A'i') xM e; montana 68 0 6 6 amencana 0 0 1 0 (A'i'+L'i') X A i!; americana 39 3 2 5 W'i'+A'i') X L e; lacicola 75 3 3 0 (T'i'+M'i') xTi!; _________________ texana 87 2 0 6 montana 1 0 0 0 (M'i' +T'i' ) X M i!; montana 69 0 3 4 texana 1 0 0 0 (T'i'+L'i') X T i!; texana 59 2 0 3 (L'i'+T'i') X L i!; lacicola 47 1 3 1 texana 1 1 0 0 x M C; _________________ (M'i'+L'i') montana 52 0 0 0 (L'i'+M'i') X L e; lacicola 76 1 4 0 montana 7 0 0 0 Totals -------------------------------------------------------------------1,516 62 47 79 93.3% 3.8% 2.9% 4.9% Homogamic --------------------------------------------·-------------1,357 52 39 66 93.7% 3.6% 2.7% 4.6% Heterogamic --------------------------------------------------------159 10 8 13 89.8% 5.7% 4.5% 7.3% TESTS FOR SEXUAL PREFERENCE AMONG GEOGRAPHIC STRAINS OF THE VIRILIS GROUP For these tests we used four geographic strains each of virilis, texana, and montana. For virilis the strains were Pasadena, Florida from Brooks­ville, Mexico from Mexico City, and Shenking from China. For texana, New Orleans from Louisiana, Georgetown and Newton from Texas, and Studies in the Genetics of Drosophila Okefenokee from Georgia. For montana the strains were Yellowstone Park (1211.58) from Wyoming, Cottonwood Canyon (1218.8) and Puf­fer Lake (1220.2) from Utah, and Bonita Canyon (1324.8) from New Mexico. In recording the data in the table we have listed the strains of montana by their stock numbers, as shown here in parentheses. We selected one geographic strain for each species and tested it against the other three strains. The strains selected were Pasadena for virilis, New Orleans for texana, and Cottonwood Canyon for montana. The experi­mental procedure followed was the same as that for the interspecific tests. We crossed five females each of two strains to one male of one of the strains for each combination. The results obtained in this series of tests are shown in Table 9. TABLE 9 Number of females dissected (n) and per cent inseminated {%) in intraspecific crosses of the virilis group so-Homogamic Heterogamic lation Females Males n % n % x' Index 1. Pasadena, Florida____________________ Pasadena 105 80.9 106 53.6 17.2 0.202 2. 3. 4. 5. Florida, Pasadena____________________ Pasadena, Mexico____________________ Mexico, Pasadena____________________ Pasadena, Shen king ________________ Florida Pasadena Mexico Pasadena 104 100 100 116 53.8 30.0 61.0 90.5 100 100 100 116 54.0 81.0 18.0 43.9 0.001 52.7 38.7 57.1 -0.001 -0.459 0.544 0.339 6. Shenking, Pasadena ·--------------- Shen king 102 33.3 103 65.0 20.6 -0.322 7. N. Orleans, Georgetown________ N. Orleans 103 34.7 101 46.5 2.37 -0.145 8. Georgetown, N. Orleans________ Georgetown 130 40.0 149 32.1 1.52 0.108 9. N. Orleans, Newton N. Orleans 104 38.4 104 41.3 0.18 -(l.036 10. Newton, N. Or leans _______________ Newt0n 104 44.2 112 25.9 8.00 0.261 11. N. Orleans, Okefenokee________ N. Orleans 128 56.2 127 30.7 16.9 0.292 12. Okefenokee, N. Orleans________ Okefenokee 101 28.7 107 48.6 8.64 -0.257 13. 1218.8, 1211.58 __________________________ 1218.8 111 66.7 101 37.6 17.9 0.279 14. 15. 1211.58, 1218.8 _________________________ 1218.8, 1220.2 ____________________________ 1211.58 1218.8 98 104 44.9 63.4 98 106 61.2 33.9 5.24 18.29 -0.153 0.303 16. 1220.2, 1218.8 __________________________ - 1220.2 114 43.8 114 60.5 6.35 --0.160 17. 18. 1218.8, 1324.8 --------------------------­1324.8, 1218.8 ____________________________ 1218.8 1324.8 107 101 51.4 69.3 104 101 56.3 43.5 0.40 13.61 --0.045 0.207 In the virilis series the results show that the difference between homo­gamic and heterogamic matings is distinctly significant in five of the six different crosses (1, 3, 4, 5, 6). This is indicated by the x2 values of 17.2, 52.7, 38.7, and 20.6, respectively, as well as by the isolation in­dices. In the other cross (2) the difference is not significant. The isola­tion index is negative for three of the six crosses, including the one in which the difference is not significant. In two crosses (3, 6) the negative isolation indices show that the males prefer females of the other strains to their own kinds. The ;x2 values and the isolation indices in the texana series show that the difference between homogamic and heterogamic matings are signifi­cant for three (10, 11, 12) of the six crosses. It will be noted that three The University of Texas Publication of the crosses have a negative isolation index, but two of these (7, 9) are not significant, while the third is (12). In the latter cross, Okefenokee males preferred New Orleans females to their own kind. In the montana series the x2 value shows a significant difference for five.of the six crosses. In three of these (13, 15, 18), both the X2 value and the isolation index are rather high, while in the other two (14, 16) they are at or just above the borderline of significance. Here again, three of the isolation indices are negative. In one (17) the difference between homogamic and heterogamic matings is not significant, and in the other two (14, 16) it is low. Table lO gives certain information about the inheritance of the genes responsible for the sexual preference. The first eight experiments in­volved males and females which had been aged 120-144 hours and ex­posed for 72 hours to each other before the females were dissected. The ninth, where the mating was the same as for the eighth, the flies were aged 240-264 hours and then exposed 24 hours. If they had been exposed 72 hours, all of the females would have been inseminated, making it im­possible to detect any sexual preference that might have been present. Nevertheless, crosses 8 and 9 are essentially similar 1n showing preference. TABLE 10 Sexual preference tests using Pasadena and Mexico strains of virilis and their hybrids. Per cents inseminated based on dissections of 100 'i' 'i' . Isolation Male Per cents inseminated x2 Index P i!l ..------------------------­ M i!J ~ ----------------­ p 5 _____________________________ Mi!J ___________,________________ P i!J _____________ ______________ M i!J ___________________________ _ PMi!J _______:_________________ _ MP 5-------------------------· MP i!J ________ __________________ P'i' 'i' = 30.0 P 'i' 'i'= 18.0 P'i' 'i'= 26.0 P'i' 'i' = 29.0 M'i' 'i'=37.0 M'i' 'i'=49.0 P'i' 'i' = 58.0 P'i' 'i'= 55.0 P'i' 'i' =30.0 . M'i' 'i' = 81.0 M 'i' 'i' = 61.0 MP'i' 'i'= 80.0 MP'i' 'i' = 66.0 MP'i' 'i'= 64.0 MP'i' 'i'= 79.0 M'i' 'i'=88.0 M'i' 'i'=87.0 M'i' 'i'= 63.0 52.7 -0.459 38.7 0.544 58.5 -0.509 27.4 0.389 14.6 0.267 19.5 -0.234 22.8 0.205 24.9 0.225 21.9 0.355 If Pasadena and Mexico females are used, all four types of males, (Pasadena (P), Mexico (M), Fi males from Mexico females-X Pasadena males (MP), and Fi males from Pasadena females X Mexico males (PM)) inseminated more Mexico than Pasadena females. If MP females were used with Mexico or Pasadena females, both Mexico and Pasadena males . inseminated more of the heterozygous females. The Fi females-(PM females behave similarly, and PM males give results like those of ·MP males), therefore, may be said to display hybrid vigor, here expressed as greater per cent of fertility in a limited period of time. It is still impossible to say how much of a role, or how exclusive the activity of one or the other sex is in determining choice. We must assume that dominant genes exist that make MP females more frequently fer­tilized than either parent in competition. We cannot say whether males Studies in the Genetics of Drosophila prefer them, or whether they are more receptive than their parent types. No change in preference accompanied hybridization in the males (MP and PM). We can, however, say that gene differences influencing prefer­ence exist and can be demonstrated to be inherited in crosses. Certain of these experiments were repeated three times with over six months between the tests, and all gave qualitatively similar results, so it is certain that they represent an expression of a real difference. Table 11 is included to show the variation between individual males. It is obvious that placing all ten of the males with ten females of each kind in a vial and leaving them for an appreciable period of time, would have given results reflecting the activity of the most aggressive males, provided the females do not control the mating process. TABLE 11 Test showing mating variability, illustrated from one Pasadena male crossed to five Pasadena plus five Mexico females Five Pasadena females Five Mexico females Male Inseminated Not inseminated Inseminated Not inseminated 1 0 5 5 0 2 5 0 5 0 3 0 5 0 5 4 1 4 5 0 5 1 4 3 2 6 4 1 5 0 7 0 5 3 2 8 3 2 4 1 9 2 3 5 0 10 0 5 2 3 DISCUSSION Sexually reproducing animals as diverse as fish (Gordon 1947), mice (Blair and Howard 1944), and flies (see below) all exhibit a certain preference in choice of mates. This sexual isolation or sexual preference which may be either interspecific or intraspecific has been extensively studied in the genus Drosophila, as indicated by the list of species and species groups investigated: (1) D. melanogaster and D. similans by Sturtevant (1915, 1920) who also made checks on several other species. (2) D. pseudoobscura A and B (D. persimilis) by Lancefield (1929); Dobzhansky and Koller (1938); Levene and Dobzhansky (1945); Mayr and Dobzhansky (1945); Mayr (1946a, b); and Wallace and Dobzhansky (1946) who also tested D. subobscura. (3) The melanica group, by Griffen (1942). ( 4) The funebris group, by Mainland (1942). (5) D. prosaltans, by Dobzhansky and Stressinger (1944) ; Mayr and Dobzhansky (1'945). (6) D. willistoni by Dobzhansky and Mayr (1944). The University of Texas Publication (7) D. sturtevanti by Dobzhansky (1944). (8) D. pallidapennis by Patterson and Dobzhansky (1945). (9) The repleta group by Wharton (1942). (10) The mulleri group by Crow (1942), Baker (this bulletin) and Patterson (this bulletin). (11) The virilis group, by Spencer (1938) ; Stalker (1942) ; Patterson,_ Stone, and Griffen (1942) ; Patterson and Griffen (1944). This list, which is incomplete, indicates that a very representative sample of this genus has been investigated. Sexual preference resulting in failure to inseminate members of the other species seems universally present and complete sexual isolation exists in some cases even within these species groups. It is probably the rule between members of dif­ferent species groups. These numerous investigators have discussed the situation in the group they investigated, as well as the general problem. The new results of the present investigation of the virilis complex of species, using the presence or absence of sperm in the female as a measure of isolation, give a more complete picture of the situation in this diverse group. Before we discuss the virilis data, there are several results of general application. One factor all investigators of evolution must evaluate is the repro­ductive· efficiency of the forms investigated under natural conditions. Table l. shows that under such conditions, the females of the thirteen species .:of Drosophila investigated are exceedingly efficient in being fertilize'.d. In nine of these thirteen species, more than ninety per cent of the females had been fertilized; in two more about eighty-five per cent, and the last two about seventy per cent. A remarkably high pro­portion of the females of most species are fertilized, especially consider­ing that a number of the females trapped must be quite young. The dif­ference in fertilization cannot be explained by differences in population density (at least in so far as the numbers of flies in the traps are related to population density) for the four species which showed the lowest amounts of irisemination were from rather intermediate population densities. Except for some members of the repleta group (D. repletd, D. hydei, D. mulleri and D. longicornis) almost all fertilized females produced offspring. In these cases, we can infer that the majority of inseminations were by males of the same species. There is no completely satisfactory explanation for the -failure of certain females in the repleta group to produce offspring, although it is known that stocks of these forms are much more temperamental and difficult to test in the labora­wry than most of the other species listed. There seems to be no consistent relation between motility of sperm in the seminal receptacles and production of offspring except that a larger per cent of those failing to produce offspring had non-motile sperm (Table 2). Some of these cases may have been due to abnormality of the sperm or of the fluids in the receptacles. In others the sperm may Studies in the Genetics of Drosophila have been too old. In some of these cases the non-motile sperm may have been those of a different species, but there is no proof of this point. The differences in insemination of the same species at different tem­peratures (Table 3), ages, and length of exposure (see Table 10 and report of results) indicate that isolation is not a single constant thing with a fixed value which is dissociated from metabolic changes and environmental factors. Despite this lack of exact quantitative fixity, we can establish qualitative differences, and even get an approximation of the average quantitative degree of isolation that exists in a particular case. Table 4 shows that all species of the virilis group show a significant amount of preference for their own kind. The one exception in the table is between the subspecies americana and texana where (in 10) the isola­tion index is negative. Four matings show a positive isolation index of 1.0 indicating complete irnlation in these experiments, and in a fifth case the isolation is complete in both directions. Other experiments indi­cate that this isolation can sometimes be broken down in mass matings, or by using other strains. In the other tests varying amounts of isolation were encountered between species. The isolation index is calculated as suggested by Stalker (1942) ; the difference between the per cent of in­semination of the two types of females tested is divided by the sum of the per cent insemination. Table 5 summarizes data from several sources and gives the same general picture of relations. In this species group choice between their own and alien females or lack of this opportunity to choose mates does not seem to make very much difference in mate preference in homogamic matings, except that the pairs (column V) shows a higher per cent of insemination than the homogamic matings which had only a limited time (columns III and IV) as would be expected with the time factor. The percentage of heterogamic matings did not seem to differ very Eignificantly with varying conditions (choice, lack of choice, pair mat­ings). There seems to be -a rather fixed and limited amount of cross mating that will occur, despite the opportunities given. This indicates a qualitative difference in response to members of the opposite sex between species, regardless of opportunities for choice of mate. There does not seem to be much difference between heterogamic and homogamic matings as to djstribution and motility of sperm, although there seems to be more sperm present in homogamic mating. Perhaps this is due in part to remating -and various other reasons for reduction in amount of sperm in heterogamic matings, such as the insemination reaction discussed by Patterson and by Wheeler (this bulletin). Mayr (1946b) has shown that in matings between species, coitus interruptus occurs in a high percentage of attempted fertilizations, so that insemination is not complete and no sperm are found in the females. This did not occur between strain·s within a species. We do not know how The University of Texas Publication frequent this mechanism prevents insemination between species of th~ virilis complex. ' If we can generalize from Mayr's results, the number of insemina­tions represent the net reproduction activity in intraspecific crosses. In the virilis group tested here, intraspecific selection (Tables 9, 10) usually represents a greater efficiency in one type of female in being inseminated, whatever this efficiency depends on. Hybrid vigor of the female seems to be effective here (Table 1'0) and may even extend to hybrids between species (Mayr, 1946a). This superiority of one-genotype of female at attaining insemination seems to be the case in D. willistoni, D. nebulosa, and D. prosaltans, although D. sturtevanti and D. americana seem to have a real preference for their own kind. Therefore for most of the cases the intraspecific tests indicate a simple selection for sexual vigor, here represented by insemination, rather than any sort of isolation. BIBLIOGRAPHY Blair, W. Frank, and Walter E. Howard. 1944. Experimental evidence of sexual isolation between three forms of mice of the Cenospecies Peromyscus maniculatus. Cont. from Lab. of Vert. Biol. Univ. of Mich., No. 26:1-19. Crow, J. F. 1941. Studies in Drosophila speciation. I. The Drosophila mulleri group. Genetics, 26 :146. Crow, J. F. 1942. Cross fertility and isolating mechanisms in the Drosophila mulleri group. Univ. Tex. Puhl., 4228 :53-67. Dobzhansky, Th. 1937. Genetics and the origin of species. Columbia University Press, New York (see pp. 261-267 of 2d Edition, 1941). Dobzhansky, Th. 1944. Experiments on sexual isolation in Drosophila. III. Geo­graphic strains of Drosophila sturtevanti. Proc. Nat. Acad. Sci., 30 :335-339. Dobzhansky, Th., and P. Ch. Koller. 1938. An experimental study of sexual isolation in Drosophila. Biol. Zentralblatt, 58 :589-607. Dobzhansky, Th., and Ernst Mayr. 1944. Experiments on sexual isolation in Dro­. sophila. I. Geographic strains of Drosophila willistoni. Proc. Nat. Acad. Sci., 30 : 238-244. Dobzhansky, Th., and G. Streisinger. 1944. Experiments on sexual isolation in Dro-· sophila. II. Geographic strains of Drosophila prosaltans. Proc. Nat. Acad. Sci., 30 : 340-345. Gordon, Myron. 1947. Genetics of Platypoecilus maculatus. IV. The sex determining mechanism in two wild populations of the Mexican Platyfish. Genetics, 32:8-17. Griffen, A. B. 1941. Studies in Drosophila speciation. II. The Drosophila melanica group. Genetics, 26:154. Griffen, A. B. 1942. Relationships in the melanica group. Univ. Tex. Puhl., 4228: 68-73. Lancefield, D. E. 1929. A genetic study of two races or physiological species of Drosophila obscura. Z. anastamm. lehre, 52 :287-317. Levene, Howard, and Th. Dohzhansky. 1945. Experiments on sexual isolation in Drosophila. V. The effect of varying proportions of Drosophila pseudoobscura and Drosophila persimilis on the frequency of insemination in mixed populations. Proc. Nat. Acad. Sci., 32 :226-234. Mainland, G. B. 1941. Studies in Drosophila speciation. III. The Drosophila macro­spina group. Genetics, 26: 160-161. Mainland, G. B. 1942. Genetic relationships in the Drosophila funebris group. Univ. Tex. Puhl., 4228:74-112, Studies in the Genetics of Drosophila Mayr, Ernst. 1942. Systematics and the origin of species. Columbia University Press, New York. . . Mayr, .Ernst. 1946a. Experiments on sexual isolation in Drosophila. VI. Isolation between Drosophila pseudoobscura and Drosophila persimilis, and their hybrids. Proc. Nat. Acad. Sci., 32:57-59. Mayr, Ernst. 1946b. Experiments on sexual isolation in Drosophila. VII. The nature of the isolating mechanisms between Drosophila pseudoobscura and Drosophila persimilis. Proc. Nat. Acad. Sci., 32:128-137. Mayr, Ernst, and Th. Dobzhansky. 1945. Experiments on sexual isolation in Dro­sophila. IV. Modification of the degree of isolation between Drosophila pseudo­obscura and Drosophila persimitis and sexual preference in Drosophila prosaltans. Proc. Nat. Acad. Sci., 31:75-82. Patterson, J. T. 1942. Isolating mechanisms in the genus Drosophila. Biol. Symposia, 6:271-287. Patterson, J. T. 1944. A new member of the virilis group. Univ. Tex. Pub!., 4445: 102-103. Patterson, J. T. 1946. A new type of isolating mechanism in Drosophila. Proc. Nat. A cad. Sci., 32 : 202-208. Patterson, J. T., and Th. Dobzhansky. 1945. Incipient reproductive isolation between two subspecies of Drosophila pallidivennis. Genetics, .30 :429-438. Patterson, J. T., and A. B. Griffen. 1944. A genetic mechanism underlying species isolation. Univ. Tex. Pub!., 4445 :212-223. Patterson, J. T., W. S. Stone, and A. B. Griffen. 1941. Studies in Drosophila specia­tion. Extension of the D. virilis group. Genetics, 26: 164. Patterson, J. T., W. S. Stone, and A. B. Griffen. 1940. Evolution of the virilis group iri Drosophila. Univ. Tex. Publ., 4032 :218-250. Patterson, J. T., W. S. Stone, and A. B. Griffen. 1942. Genetic and cytological analysis of the virilis spec~es group. Univ. Tex. Pub!., 4228: 162-200. Spencer, W. P. 1938. Drosophila virilis americana, a new subspecies. Genetics, 23: 169-170. Stalker, H. D. 1941. Sexual isolation in the virilis complex of Drosophia. Genetics, 26:170. Stalker, H. D. 1942. Sexual isolation studies in the species complex Drosophila virilis. Genetics, 27:238-257. Sturtevant, A. H. 1915. Experiments on sex recognition and the problem of sexual selection in Drosophila. Jour. Animal Behavior, 5 :351-366. Sturtevant, A. H. 1920. Genetic studies on Drosophila simulans. I. Introduction. Hybrids with Drosophila melanogaster. Genetics, 5 :488-500. Sturtevant, A.H. 192la. Ibid. · II. Sex-linked group of genes. Genetics, 6:43-64. Sturtevant, A. H. 1921b. The North American species of Drosophila. Carnegie Inst. Wash., Publ. 301. Wallace, Bruce, and Th. Dobzhansky. 1946. Experiments on sexual isolation in Drosophila. VIII. Influence of light on the mating behavior of Drosophila sub­obscura, Drosophila persimilis and Drosophila pseudoobscura. Proc. Nat. Acad. Sci., 32 :226-234. Wharton, Linda T. 1941. Studies in Drosophila speciation. VI. The Drosophila repleta group. Genetics, 26:174; Wharton, Linda T. 1942. Analysis of the repleta group of Drosophila. Univ. Tex. Pub!., 4228 :23-52. II. SEXUAL ISOLATION IN THE MULLER! SUBGROUP J. T. PATIERSON INTRODUCTION The mulleri subgroup of species belongs to the l.arge repleta species group, of which more than fifty forms have been described. About six­teen species may be assigned to this subgroup, and eight of these have been selected for this study. These eight species, together with the strains used, are as follows : D~ aldrichi Patterson and Crow 1940; strain from Aldrich Farm, Aus­tin, Texas. D. arizonensis Patterson and Wheeler 1942; strain 928.5, Tucson, Ari­zona. D. buzzatii Patterson and Wheeler 1942; strain 190, Cordoba, Argen­tina. D. hamatofila Patterson and Wheeler 1942; strain from Aldrich Farm, Austin, Texas. D. mojavensis Patterson and Crow 1940; strain from Chocolate Moun­tains, California. D. mulleri Sturtevant 1921; strain 1434.4, Aldrich Farm, Austin, Texas. D. peninsularis Patterson and Wheeler 1942; strain 1148.7, Lake Mc­Kethan, Florida. D. ritae Patterson and Wheeler 1942; strain from Huachuca Mountains, Arizona. The members of this subgroup have certain characters in common, of which two may be mentioned. While the tergite pattern varies among the different species, yet there is a strong tendency for the abdominal bands to be broken up into spots, · either in one or both sexes. The second character, which is constant for the subgroup, relates to the structure of the spermathecae. In most members of the genus Drosophila these paired organs serve as sperm receptacles, but in the mulleri series they are small and have slender dalks, and only' rarely are the .spermatozoa successful in gaining entrance to them. These organs have become modi-· fied into gland-like structures, each surrounded by a thick envelope from which secretions are discharged into its cavity. In testing for sexual isolation in the muperi subgroup, the writer has followed closely the experimental procedure used in the preceding article for the virilis group. It is, therefore, unnecessary to give an extended account of the methods employed. The life cycles of the mulleri forms are somewhat shorter than those for the virilis species, consequently, the flies were aged for only four days after emergence before makirig the crosses. The females were dissected after ninety-six hours exposure • Studies in the Genetics of Drosophila to the males. In the sexual preference tests the same plan was followed as before; five females of each of the two species under test were placed in a vial with a single male of one of the species. The writer acknowledges his indebtedness for excellent technical as­sistance, both in this and the following article, to Mrs. Sarah B. Martin who made the dissections, and to Miss Johanna Blume! who bred and separated the flies used in the dissections. He is also indebted to Profes­sor A. H. Sturtevant of The California Institute of Technology for sup­plying a stock of D. ritae. FERTILITY TESTS AMONG THE EIGHT SPECIES OF THE MULLER! SUBGROUP Three papers have been published which give the degree of fertility among some of the members of this subgroup. The first of these included but three species (Patterson and Crow 1'940), while the second and third dealt with only five species (Crow 1942, Patterson 1942). At most this would involve only twenty different crosses, whereas, the eight species used in the present study give a total of fifty-six different combinations. Moreover, in the three published articles only a few of the exposed females were dissected, and the percentages of fertility were based mainly on the production of hybrids. In the present study we dissected at least one hundred females for each combination, and from these determined the exact number of females which had been inseminated, irrespective of whether or not hybrids had been produced. The number inseminated was obtained by the use of small mass matings of ten females and ten males each. The tests were carried out in this manner because mass matings often yield offspring in cr9sses in which pair matings fail to do so. The results obtained from the fifty-six different combinations are given in Table 1. The first column shows the nature of the cross, the second indicates whether or not hybrids were obtained, the third gives the number of females dissected, and the fourth and last shows the number of females inseminated. The different crosses have been num­ bered from 1 to 56, for ease of reference, and these numbers appear in parentheses in the text. In thirty of the crosses none of the females had been inseminated, and, consequently, hybrids were not obtained. The number of females inseminated in the other twenty-six crosses varied from one to ninety-three. Ten of these yielded some sort of hybrids (1, 3, 5, 7, 9, 15, 21, 27, 31, 32), and sixteen failed to do so. The failure of the females of the latter group to produce offspring was due, in a large measure, to the deleterious effects of the insemination reaction in heterogamic matings (Patterson 1946). This reaction will be fully con­ sidered in the next article. The University of Texas Publication A further analysis of the data displayed in Table 1 may be made by referring individually to each of the eight species. TABLE 1 Crosses between members of the mulleri subgroup Crosses Types of hybrids 2 2 dissected Inseminated 1. mulleri 2 X aldrichi CJ -----------------------------------------­ sterile 2 2 , CJ CJ 100 32 2. aldrichi 2 X mulleri CJ ---------------­---------------­--------­ none 100 18 3. mulleri 2 X arizonensis CJ ---------------------­-------------­ sterile CJ CJ 100 8 4. arizonensis 2 X mulleri CJ ----------------------------------­ none 100 0 5. mulleri 2 X buzzatii,(; -----­--------------­----------­-~-----­ abnormal flies 100 2 6. buzzatii 2 X mulleri-(; ------------------------­--------------­-­ none 101 0 7. mulleri 2 X hamato(ila CJ ----------­---­------------------­ sterile 2 2 , CJ CJ 100 10 8. hamatofila 2 X mulleri (; ---------------------------­----­--­ none 109 0 9. mulleri 2 X mo javensis CJ -------­---­----------------------­- fert. 2 2 , st. CJ CJ 100 21 10. mojavensis 2 X mul]eri (; ---------------------­-------------­ none 105 0 11. mulleri X 2 peninsularis CJ ------------------------------­ none 104 0 12. peninsularis 2 X mu'.lleri t -----­---­-----------­----­---­-13. mulleri 2 X ritae CJ ,. -------------­-------------­--­--­-----------­ none none 104 110 0 0 14. ri tae 2 X mulleri (; , ----­----------------­---------------------­~ 15. aldrichi 2 X arizonensis CJ -----------­---------------------­ none sterile 2 2 104 100 0 26 16. arizonensis 2 X aldr)chi CJ --------------------------------­ none 100 0 17. aldrichi 2 X buzzatii CJ -------------------------------------­ none 100 0 18. buzzatii 2 X aldrichr t ---------------------------------------­ none 105 0 19. aldrichi 2 X hamatofila CJ ----------------------------------­ none 100 11 20. hamatofila 2 X aldrii::hi CJ ----------------­----------------­ none 103 0 21. aldrkhi 2 X mojavensis (; 22. moiavensis ~ X aldrichi CJ -------­------------------­-----­-----------------------­---------­ sterile 2 'i? none 100 100 4ti 15 23. aldrichi 2 X peninsularis t ------------------------------­24. peninsularis 2 X aldrichi (; ----------­-----------------­--­25. aldrichi 2 X ritae (; ----------------­------------­--------­-­----­26. ritate 2 X aldrichi CJ ------------------------------------­------­ none none none none 102 105 102 106 0 0 0 0 27. arizonensis 2 X buzzatii (; ---­----------------------------­28. buzzatii 2 X arizonensis (; ---------------------------­-----­29. arizonensis 2 X ham!ltofila (; ----------------------­------­30. hamatofila 2 X arizopensis (; --­--------------------------­31. arizonensis 2 X mojavensis (; ---------------------------­32. mojavensis 2 X arizonensis (; _ _____ ___________________ _ 33. arizonensis 2 X peninsularis (; _________________________ _ larvae only none none none fert. 2 2 , st. CJ fert.22, (;$ none CJ 106 100 100 100 100 100 102 1 88 21 5 29 93 0 34. peninsularis 2 X arizonensis (; ------------------------­35. arizo nensis 2 X ritae(; ---------------------------------------­ none none 100 101 43 0 36. ritae 2 X arizonensis $ ---------------------------------------­ none 100 0 37. buzzatii 2 X hamatofila (; ---­--­----­-------------------­---­~~: ~~~~~rna~ -::.o~~~!~~ils~ -·:::::::-:::::::::::::::::::::::::: none none none 100 100 100 74 1 18 40. mojavensis 2 X buzzatii (; ---------------------------------­ none 104 0 41. buzzatii 2 X peninsularis (; 42. peninsularis 2 X buzzatii CJ ------------------------------­---­---------------------------­ none none 103 103 0 1 43. buzzatii 2 X ritae CJ -------------------------------­-------------­ none 100 0 44. ' ritae 2 X buzzatii (; ---------------------------­----------­-----­45. hamatofila 2 X mojavensis (; -­--­-----------------------­46. mojavensis 2 X hamatofila t ------­---­------­-----------­ none none none 100 100 100 0 19 33 47. hamatofila 2 X peninsularis (; ----------------------­----­48. peninsularis 2 X hamatofila (; ----­---­------------------­49. hamatofila 2 X ritae $ --------­----------------­----------­-- none none none 103 100 100 0 10 1 50. ritae 2 X hamatofila (; ------------------------------­_______ _ none 100 10 51. mojavensis 2 X peninsularis (; -------------------------­52. peninsularis 2 X mojavensis (; -----­-------------------­53. mojavensis 2 X ritae t --------­----- -----------------------­54. ritae 2 X mojavensis (; --------------­---------­----­---------­55. peninsularis 2 X ritae (; -----------­----­---­----------­---­56. ritae 2 X peninsularis $ ----­-----------------------------­--­ none none none none none none 103 100 110 102 103 100 0 0 0 0 0 0 Studies in the Genetics of Drosophila D. mulleri.-Females of this species were inseminated by the males of five of the other seven species. In four of the crosses (1, 3, 5, 7), such hybrids as were produced were sterile, and in the fifth cross (9) the F1 males were also sterile, but at least some of the hybrid females were fertile in backcross. In the reciprocal cross, the mulleri males failed to fertilize any of the females of six other species, but did inseminate 18% of the aldrichi females without producing offspring (2). Several years ago the writer found another member of the subgroup, D. longicornis, the males of which fertilized mulleri females, producing a few sterile hybrids. The results from this test have been reported by Wharton (1944). So far as the exchange of genes is concerned, the results from these tests show that D. mulleri cannot do so except with one other species. In the mulleri/mojavensis cross, Crow (1942) has reported that in pair mating tests, 5% of the females produce offspring, and that the F1 females are fertile in backcrosses to the parent types of males. D. aldrichi.-FemaJes of this species were inseminated by the males of four other forms, mulleri, hamatofila, arizonensis, and mojavensis. The first two crosses do not result in the production of offspring (2, 19) while the other two produce sterile females (15, 21). In the reciprocal crosses, aldrichi males inseminated mulleri females, producing sterile males and females ( 1) , and also mojavensis females without the produc­tion of offspring (22). This species cannot, therefore, exchange genes with any other member of the subgroup to which it has been tested. D. arizonensis.-The females of arizonensis were fertile to the males of two other species. The first successful cross was to buzzatii males and resulted in the production of larvae which died in mid-larval stages (27). The second cross to go was with mojavensis males, and yielded fertile females and sterile males (31). In the reciprocal crosses males of arizonensis inseminated the females of six other species (3, 15, 28, 30, 32, 34), but only three of these produced hybrids (3, 15, 32). The reciprocal crosses of arizonensis and mojavensis are two crosses in which an exchange of genes with other members of the subgroup is possible. The mojavensis/ ctrizonensis cross wa.s the most successful of the entire series, giving both fertile males and females. D. buzzatii.-The females of this form were inseminated by the males of three other species, but none of these crosses produced hybrids. In the buzzatii/arizonensis cross (28), 88% of the females had been insem­inated, and in the buzzatii/ hamatofila cross (37) 74%. In the third cross, buzzatii/mojavensis (39), 18% of the females had been inseminated. In the reciprocal crosses, the males of buzzatii had inseminated four types of females, mulleri, arizonensis, hamatofila, and peninsularis. In the mulleri/buzzatii cross most of the resulting zygotes die in larval stages, and only a few reach maturity and are abnormal and sterile (5). The a.rizonensis/ buzzatii cross gave a few lethal larvae (27). One female of hamatofila and one of peninsularis had been inseminated by buzzatii The University of Texas Publication males (38, 42). On the basis of these results this species cannot exchange genes with any other member of the subgroup. D. hamatofila.-The females of this species were inseminated by the males of arizonensis (30), buzzatii (38), mojavensis (45), and ritae (49), but hybrids were not produced. Only mulleri females were fertile to hamatofila males, but the hybrids were sterile (7). In addition to this cross, the females of the other six species had been inseminated to vary­ing degrees by hamatofila males (19, 29, 37, 46, 48, 50). This species is also not capable of gene transfer with other members of the subgroup. D. mojavensis.-The females of this species were inseminated by aldrichi and hamatofila males, without the production of hybrids (22, 46), and fertile to arizonensis males, giving fertile males andfe:µiales (32). The males were fertile to mulleri, arizonensis, and aldrichi females, pro­ducing fertile females and sterile males in the first two crosses and . sterile females in the third (9, 31, 21). These males also inseminated buzzatii and hamatofila females, but without producing offspring (39, 45) . There are, therefore, three possible chances of gene exchange with this form and other members of the subgroup (9, 31, 32). D. peninsularis.-The males of peninsularis failed to inseminate a single female of any of the other seven species, but males of arizonensis, buzzatii, and hamatofila did inseminate some of its females (34, 42, 48). However, no offspring was produced, and gene exchange is not possible. D. ritae.-Offspring were not produced in any of the fourteen possible crosses with this species, but ten females had been inseminated by hamatofila males (50), and one hamatofila female had been inseminated by a ritae male ( 49) . Gene exchange between this species and other members oi the subgroup is also not possible. While some sort of hybrid is produced in ten of the fifty-six crosses, yet.there are only three of these in which an exchange of genes is pos­sible. These are the fertile F 1 females from the mulleri/ mojavensis and arizonensis / mojavensis crosses, and the fertile F1 males and females ftom the mojavensis/ arizonensis cross. All other hybrids are sterile and some of them are abnormal, or else the zygotes die in larval or pupal stages. We have no information as to whether the flies which produce fertile hybrids ever mate in nature. So far as known, the distributional areas of mulleri and mojavensis are widely separated, but there is a possibility that those of mojavensis and arizonensis may overlap. The most extensive overlapping. of distributional areas are those of mulleri and aldrichi (Patterson and Wagner 1943), but the F1 flies from the 'only cross that goes, mulleri <;> X aldrichi ~ are completely sterile. The male hybrids from this cross have been found in nature and definitely identified, and since this is the only known instance in which it has been possible to identify Drosophila hybrids in nature, it seems worthwhile .to present the evidence upon which this statement is based. The hybrids from the mulleri/ aldrichi cross resemble mulleri in most of 'their characters, but the testis of the male hybrid is rudimentary and Studies in the Gen~tics of Drosophila shows little coiling, and, as seen through the abdominal wall, appears in the form of a question mark ( ?) . It is this character which enables one to identify these males in collections brought in from the field. Such males are identical with the hybrid males obtained in the laboratory from this cross. During July and August of 1940 and 1941 collections of Drosophila from the overlap zone of mulleri and aldrichi in Texas were examined fori the presence of these hybrid males. These data are summarized in Table 2. All of the collections were made in the country, and none in towns or cities. In recording the place at which a given collection was made the nearest town is listed, but if no towns were near, it is recorded as from a Roadside Park (R. P., in the table). The number of mulleri and aldrichi males are given for each collection, followed by the percentage of aldrichi males in the male population of both species. The number of hybrid males is shown in the last column. Hybrid females must also be present in some of these collections, but they could not be detected with certainty. TABLE 2 Per cent of hybrid males found in nature from the cross mulleri 'j> %inseminated hamatofila____ 11 4 4.1 7 18.4 arizonensis____ _ 9 3 11.1 6 34.8 mojavensis_____ 8 3 37.3 5 18.6 buzzatii___________ 7 3 51.6 4 0.5 aldrichi___________ 6 4 17.5 2 15.0 mulleri______ 6 5 9.6 1 12.0 peninsularis___ 3 3 12.0 0 0.0 ritae____________ 2 1 5.0 1 0.5 Drosophila hamatofila leads the list in the number of crosses in which it was involved. Its females were inseminated by the males of four other species, and its males in~eminated the females of seven different species. The highest per cent of females inseminated by other males was buzzatti females, with a percentage of 51.6 in three combinations. But the per cent of females inseminated by buzzatii males was one of the two lowest, with a percentage of only 0.5 in four combinations. The six combinations with arizonensis males gave the highest per cent (34.8) for inseminations of alien females. The results obtained in this series of experiments make it clear that · sexual isolation is a real barrier to gene exchange between the members of the mulleri subgroup. It represents one of the most important factors in the evolutionary divergence of this group of species. This conclusiOn is supported by the work of other investigators on several other different groups of Drosophila species. It is not, however, the only barrier to gene transfer. There are several other isolating mechanisms operating in the mulleri subgroup, including the effects of the insemination reaction. All of these will be considered in the next article. REFERENCES Crow, J. F. 1942. Cross fertility and isolating mechanisms in the Drosophila mulleri group. Univ. Tex. Puhl., 4228 :53-67. Patterson, J. T. 1942. Interspecific hybridization in the genus Drosophila. Univ. Tex. Puhl., 4228 :7-15. Patterson, J. T. 1946. A new type of isolating mechanism in Drosophila. Proc. Nat. Acad. Sci., 32 :202-208. Patterson, J. T., and J. F. Crow. 1940. Hybridization in the mulleri group of Dro­sophila. Univ. Tex. Puhl., 4032 :251-256. Patterson, J. T., and R. P. Wagner. 1943. Geographical distribution of species of the genus Drosophila in the United States and Mexico. Univ. Tex. Puhl., 4313 :217-281. Other references. to the literature on sexual isolation will be found at the end of Article I. III. THE INSEMINATION REACTION AND ITS BEARING ON THE PROBLEM OF SPECIATION IN THE MULLER! SUBGROUP J. T. PATTERSON INTRODUCTION In a recent article the writer (Patterson 1946) presented a brief ac­count of the changes which take place in the female reproductive tract in some species of Drosophila immediately following copulation. The changes occur mainly in the vagina, and the term "insemination reac­tion" was used to designate this phenomenon. In the present article it is proposed to give a much fuller account of this reaction and its effect on reproduction among members of the mulleri subgroup of species. The same Drosophila forms are included in this paper as were employed in the preceding article of this publication, namely -aldrichi, arizonensis, buzzatii, hamatofila, mojavensis, mulleri, peninsularis, and ritae. We shall also have occasion to refer to the data included in the preceding article, especially those listed in Table 1 of that paper. The insemination reaction occurs both in intraspecific and interspecific crosses. It usually follows immediately after coitus, but in some forms it may begin before copulation is completed. The reaction is revealed in the vagina through a rapid secretion of fluid into its cavity, resulting in an increase in size to three or four times normal. In females from homogamic matings the vagina returns to a normal condition, usually within a period of eight or nine hours, but in those from heterogamic matings it may remain swollen for several days. In the latter type of mating this reaction may greatly reduce the number of hybrids pro­duced, and in some crosses may eliminate them altogether, even though the female has been inseminated. The term "reaction mass" was used in the preliminary paper to desig­ nate the contents of the swollen vagina. It is especially appropriate as a descriptive term for heterogamic matings in which the contents of the vagina persist for a long time as a very definite and discrete struc­ ture. As will be shown later, the reaction mass in homogamic matings remains soft and is eventually expelled by the female. Consequently, it does not materially affect the production of progeny. Its long retention in the vagina by females of heterogamic matings, accompanied by a change in composition from a soft to a hard, and often, resistant object may prevent the fertilization of the eggs, resulting in a failure of hybrid production. In this and the succeeding article we are using the term vagina to designate that part of the female reproductive tract which extends pos­ teriorly from the base of the median oviduct to the ovipositor. However, considerable confusion exists in the literature concerning the terminology of the different parts of the female apparatus in Drosophila, especially The University of Texas Publication with reference to the use of the terms·uterus and vagina. Some authors refer to that part mentioned above as the vagina (Snodgrass 1935), others refer to it as the uterus (Pantel 1910), and still others would divide this region into two parts, a uterus proper and a vagina (Nonidez 1920, Wigglesworth 1939, Anderson 1945). Where this division has been suggested, the posterior part of the duct has been called the vagina, and the anterior part the uterus. Some authors have suggested that a uterus is present only in forms in which eggs are incubated (Townsend 1911). There is need for a study of the histology of the vagina of these forms, not only of the normal condition, But also of the changes which occur during the formation of the reaction mass. We have made sections of a limited number of stages, but the preparations were inadequate for a detailed study. They show many of the histological features described by Nonidez for this organ in virgin females of D. melanogaster. It is, therefore, necessary to defer giving an account of the histological changes which accompany the insemination reaction until a later date. In making preparations of whole mounts of the vagina for study of the insemination reaction, we have used the following procedure. The etherized female is placed on her side in a large drop of saline solution on a clean slide. The dissection is made by means of a pair of very fine needles. One needle is placed through the thorax and the other through the posterior tip of the abdomen, and by a gentle, but steady, pull the entire reproductive tract can be removed except the ovaries, which break off at the upper end of the median oviduct. After the removal of any . adhering materials, a cover slip is placed on the preparation, which is then ready for examination under the microscope. The preparation is first examined under the 16 mm. objective, but for a detailed study the 4 mm. objective is used. The use of a large drop of saline is recommended because otherwise the weight of the cover slip may cause the contents of the vagina to be expelled, either through the posterior orifice or the broken end of the oviduct. In the preparation of a large, turgid vagina it is often necessary to support the cover slip by bits of glass in order to prevent the expulsion of the reaction mass. For observations on the movement of the spermatozoa and the formation of the reaction mass, it is desirable to have the vagina somewhat flattened by the weight of the cover glass. If the above procedure is followed the reproductive tract will lie on its side, which is the best position for a detailed exam­ination. The photomicrographs with which this, and the next article by Mr. Wheeler, are illustrated were made by transmitted light, using a 16 mm. objective and a Leitz 8X Periplan ocular. The films were developed in a strong contrast developer and the prints were made on either No. 2 or No. 3 glossy paper. Neither the negatives nor the prints have been re­ touched, except in three cases (Figs. 5, 30, 45) in which air bubbles or a piece of tissue in the background, and external to the specimen, were blotted out on the· negative with opaque. All of the photographs were Studies in the Genetics of Drosophila made at the same magnification from freshly dissected, unstained speci­mens, and as reproduced in the plates are enlarged about seventy-eight times. The plates are numbered consecutively in the two articles, as are also the figures. The time given for the ages of the different specimens in the legends to these figures in Plates I-IV and XI-XIII in this article were accurately determined by direct observation and for each specimen it represents the elapsed time from the end of copulation to the beginning of the dissection. In Plates V-X, unless otherwise stated, the ages given represent the period during which the males and females were exposed to each other, and consequently indicate the approximate ages of the different specimens. HISTORY OF THE INSEMINATION REACTION IN HOMOGAMIC MATINGS The duration of the insemination reaction in homogamic matings is relatively short and does not exceed nine hours for any of the eight species studied. The method followed was to place one female with four or five males in a vial and watch until copulation had taken place. As soon as coitus was completed, the female was separated from the males and a series of such females was dissected from time to time throughout the duration of the reaction. A record of the length of copulation was kept for each female, and from these data the average time was derived. In some species the insemination reaction begins before the end of coitus, and in such forms it was necessary to make dissections before copulation had been completed. This procedure made it possible to determine the exact time at which the vagina begins to enlarge. The facts thus accumu­lated are rather extensive and too large to present in any great detail. They are, therefore, shown in condensed form in Table 1. TABLE 1 Duration of copulation and the insemination reaction in homogamic matings Duration of copulation Duration of insemination reaction Species range mean begins at ends at D. aldrichi ---------------­ 1'20"-5' 3''2:7" 1'0'"'' 9 hours D. arizonensis__ _ ____ ____ 30"-2'10" 1'34" 3'50" 9 hours D. buzzatii_____ __ __ _________ l '-3'10" 2'4" 6'0" 8 hours D. hamatofi/a____ _________ 5'-10" 7'3" 4'30'"'' 9 hours D. mojavensis____________ 1'20"-5' 3'27" 2'52" 9 hours D. mulleri__________________ 30"-3' 1'35" 1'30" 8 hours D. peninsularis_________ D. ritae___ _____ ___ _________ __ 1'20"-2'5" 6'-9'15" 1'58" 8'0" 6'0" 5'20'"'' 6 hours 8 hours There is considerable variation in the duration of copulation, both within the same species and among the different species. The shortest time of only thirty seconds was found in arizonensis and mulleri, which show about the same average time. The longest durations occurred in The University of Texas Publication hamatofila and ritae, and they also have the highest averages. The range and mean times are identical for aldrichi and mojavensis. In column four is given the time at which the insemination reaction begins in the dif­ferent species. The figures represent the elapsed time from the beginning of copulation to the start of the reaction. In three species (marked; *) the reaction began before coitus was completed. With the exception of . one species, the duration of the insemination reaction was either eight or nine hours. The exception was peninsularis, in which the reaction lasted for only six hours. By the end of these periods the vagina has lost its opacity and returned to its original size. Before giving the more interesting facts observed in the study of homogamic matings, reference to the accompanying text-figure should Studies in the Genetics of Drosophila be made. The figure illustrates the various parts of the female repro­ductive system of D. buzzatii, and is more or less typical of this system for all members of the mulleri subgroup. The paired oviducts ( d) arise from the bases of the ovaries ( o) and soon fuse to form thE! median, or azygos, oviduct (d'). At the lower end of this duct there arise from its dorsal side the paired spermathecae (s) and the paired parovaria (p), the former slightly anterior to the latter. At about the level of the stalks of the parovaria, but on the ventral side of the median oviduct, there arises the long coiled ventral receptacle (vr). Just posterior to the points of origin of the parovaria and the ventral receptacle the duct abruptly enlarges to form the vagina (v). In this, and many other species ..of Drosophila, the vagina extends antero-ventral to form what is known as the vaginal pouch ( vp). The wall of this region is much thinner than that of other parts of the vagina and is capable of great distention. It is here that the swelling caused by the insemination reaction chiefly takes place. The figures in Plates I and II illustrate the effects of the insemination reactiOn on the vaginae of the eight species used in this study. Their dark color is due to the opacity of the reaction mass within their cavities. Viewed through a binocular microscope, the specimens appear opaque white, but by transmitted light they photograph black. Figure 1 shows the reaction in D. aldrichi at one hour after copulation. The specimen had been intentionally flattened slightly by pressure on the cover glass in order to bring out more clearly the granular appearance of the reaction mass. This species is one in which the reaction begins before the end of copulation. In a series of dissections of four aldrichi females, all made before the termination of coitus, the following observa­ tions were record~d: (1) At thirty seconds after the start of coitus the vagina contained no sperm, and was clear and normal; (2) after one minute the vagina contained many sperm in its cavity and was beginning to enlarge; (3) after one minute and thirty seconds it had increased slightly in size, with the sperm surrounding a central core of reaction material; finally, after two minutes and fifteen seconds, the vagina was greatly swollen and some of the motile sperm had been expelled. One female, which was dissected five minutes after the end of copulation, had a very large, turgid and slightly opaque vagina. Figure 2 illustrates the condition of the vagina of D. arizonensis thirty minutes after copulation. In this species the swelling does not begin until after copulation has been completed, and the vagina never becomes as large as in aldrichi. It reaches its maximum size in about an hour and fifteen minutes, and soon thereafter begins to shrink and clear up, but does not return to a normal condition until about nine hours. Figure 3 shows the inseminated vagina of D. mulleri at the end of one hour. lt was much larger than is indicated in this picture, but "exploded" just as the photograph was being made and expelled a considerable .quan­ tity of its contents, which appear as a cloudy mass about the lower end The University of Texas Publication of the figure. Accidents of this kind occur frequently in handling such large turgid vaginae. Usually, however, any expulsion of the contents takes place through the vaginal orifice, rather than through a rupture of the wall. Another case of this kind is seen in Figure 2. PLATE I. Vaginae from homogamic matings of four species of the mulleri subgroup. Fig. 1, D. aldrichi at one hour. Fig. 2, D. arizonensis at thirty minutes. Fig. 3, D. mulleri at one hour. Fig. 4, D. peninsularis at forty minutes. Studies in the Genetics of Drosophila The inseminated vagina of D. peninsularis is illustrated in Figure 4. While the photograph was taken only forty minutes after the end of copulation, yet the vagina had reached approximately its maximum size. It never does become as large as in the other members of the mulleri subgroup, and returns to a normal condition in about six hours. Figure 5 (Plate II) illustrates the inseminated vagina of D. mojavensis at one and one-half hours. It reaches its maximum size one hour after copulation and remains large until some time during the fourth hour. From this time on it gradually clears and decreases in size, and by the end of nine hours it appears entirely normal. Figure 6 shows the enlarged vagina of D. buzzatii one hour after the end of copulation. Forty-seven females of this species in a timed series were dissected, beginning immediately after coitus and extending through the 135th hour. Examinations of the reproductive tracts of these females showed that all except three had been inseminated. After the sixth hour all sperm were found to be located in the ventral receptacle, and none in the rest of the tract. The vagina was cleared and normal by the end of the eighth hour. The females began laying eggs at about seventy-two hours after mating, and offspring developed in a large majority of the retained food vials. One of the most interesting features of the hbmogamic mating of D. hamatofila is the duration of copulation, which on the average lasts over seven minutes. It is one of three species of the mulleri subgroup in which the insemination reaction begins before the end of copulation. The vagina has reached its maximum size within thirty minutes after copulation, and the finger-shaped reaction mass has become extremely opaque (Fig. 7, Plate II). The longest time of copulation of any member of the subgroup is found in D. ritae. However, a more striking feature of the homogamic mating in this species is the peculiar type of reaction mass. It is also elongated, but as viewed from the ventral side, as shown in Figure 8, it has two constrictions which incompletely divide the mass into three parts of about equal size. The history of the insemination reaction in ritae deserves a somewhat fuller description, and the following account is a direct quotation from the notes made at the time the tests were carried out. "This species is especially favorable for a study of the insemination reaction, both because of the speed at which it occurs and the nature of the opacity of reaction mass. In ritae and some other forms this mass soon becomes intensely opaque and appears black by transmitted light. A closely graded series of stages was studied, based on the duration of copulation. The average duration of copulation is about eight minutes. In females dissected immediately following copulation the reaction mass is already well formed, with each end dark and the middle section grayish. In order to observe the beginning of the process, it was necessary to dissect females before coitus had been completed. It was found that sperm hild not been transferred up to about five minutes after the be­ The University of Texas Publication ginning of copulation. The first female found with sperm in her vagina had been in copula for five minutes and twenty seconds, and in this specimen it was possible to follow the progress of the formation of the reaction mass. When first examined under the microscope, the semen PLATE II. Vaginae from homogamic matings of four other species of the mulleri subgroup. Fig. 5, D. rnojavensis at one and one-half hours. Fig. 6, D. buzzatii at one hour. Fig. 7, D. harnatofila at thirty minutes. Fig. 8, D. ritae at one hour. Studies in the Genetics of Drosophila with its sperm was located at the posterior end of the vaginal cavity. Within a very short time the sperm began moving forward through the cavity, and by the end of three minutes they had reached the anterior end, with some of them passing up a short distance into the proximal end of tJ!e median oviduct. The formation of the reaction mass followed in the wake of the advancing semen, and by the end of five minutes the appearance of the mass was very similar to that found in females dissected shortly after copulation." In Plate III are displayed a series of stages showing the changes which take place in the vagina of females from homogamic matings of D. aldrichi. Figure 9 is from a virgin female of the same age (five days) as those which were inseminated, and from which Figures lO to 12 were made. The vagina illustrated in Figure 10 was dissected three and one-half hours after the end of copulation. It was very opaque and represents the maximum size reached by this organ during the history of the in­semination reaction. By the end of five hours (Fig. 11) much of the reaction material had been expelled, including any sperm that had failed to enter the ventral receptacle. The condition of the vagina at seven hours after insemination is revealed in Figure 12. The vagina is still slightly enlarged and all but a small amount of scattered material has disappeared. It would have taken about two hours longer, or nine hours in all, for this vagina to return to the condition seen in Figure 9. The history of the insemination reaction in homogamic matings is essentially the same in all snecies which have been carefully studied by us. There are, however. certain variations in the process, such as the time at which the reaction begins, the length of its duration, and the character of the reaction mass, but these are minor details. In all Roecies exi:imined the vagina eventually returns to a normal condition, that is. it simulate!'! the transnarency and size of this organ in virgin females. In the preliminary article (Patterson 1946) it was pointed out that the contents of the swollen vagina. consisting of excess sperm and reaction material, are exnelled in the form of whitish dronlets which are easily detected on thA food o-r sides of the container. If examined immediately after denosition under the microscope, such droplets ar~ Reen to contain motile snerm. This habit on the part of females has been observed several times in four different species. It may .therefore be assumed that this is the principal method of bringing about the return of the vagina to a normal condition. In most species which have the insemination reaction, the female usually does not remate for some time, and in heterogamic matings she may never remate. This subject will be discussed further in a later .sec­tion on the mating habits of inseminated females. · The University of Texas Publication PLATE III. Vaginae of virgin and three homogamic matings of D. aldrichi. Fig. 9, virgin. Fig. 10, at three and one-half hours. Fig. 11, at five hours. Fig. 12, at seven hours. Studies in the Genetics of Drosophila HISTORY OF THE INSEMINATION REACTION IN HETEROGAMIC MATINGS The early stages of the insemination reaction in heterogamic matings are very similar to those occurring in intraspecific crosses. The vagina undergoes about the same changes in both types of crosses during the first few hours after inseminaton. The reaction progresses somewhat more rapidly in the heterogamic type, and the vagina is frequently more densely opaque. It takes much longer for the sperm to move forward to the point of origin of the ventral receptacle, usually requiring about . forty-five minutes to reach this point as compared with five to ten min­utes in the homogamic type. Furthermore, the spermatozoa frequently form a tangled mass about the opening to the ventral receptacle. This has the effect of retarding or inhibiting their entrance into the lumen of the receptacle, often resulting in only a few sperm gaining access to this organ, and in some cases, none at all. The sperm which finally suc­ceed in entering the receptacle become located at the distal end of the long, coiled receptacle. · The excess of spermatozoa left in the vagina usually disappears by the end of seven hours, although small groups of non-motile sperm have been observed after this period in a few cases. Their fate has not been definitely determined. We have found no conclusive evidence that they are expelled as in intraspecific crosses. It is possible that these sperm have been digested, and the products either absorbed by the reproduc­tive tract, or else included in the reaction mass. It has been reported that such excess sperm are digested and absorbed in a few other insects. This is said to be the case for the beg-bug Cin'!-eX (Abraham 1934, Cragg 1'920, 1923), and also for the bug GraphoSOJYfe lineatum (Hanrlirsch 1900), as reported by Wigglesworth in his book, The Principles of Insect ' Physiology. The sperm that are located at the distal end of the ventral receptacle may remain alive for several days. Motile sperm have been observed in the receptacles of females dissected as late as 168 hours after copulation. In many interspecific crosses these sperm never fertilize the eggs, due to the deleterious effects of the insemination reaction and its resulting reaction mass. Even though this mass may finally disappear, the vagina is usually left too severely damaged to permit the passage and insemina­tion of the egg from the oviduct to the exterior. The eggs may disinte­grate in the cavity of the vagina, but if they succeed in passing this barrier and are laid, smear preparations show that they have not been inseminated. The consequence is that the female, although fertilized, cannot produc~ hybrid off spring. In the series of crosses be~ween members of the mulleri subgroup of species every degree of the severity of this reaction was encountered. At one end of this series, in the buzzatii/arizonensis cross, hybrids were never produced, although 88 % of the dissected females were found to contain sperm in their receptacles. At the opposite end of the series, in The University of Texas Publication the mojavensisI arizonensis cross, about 75 % of the females produced hybrids, while 93 % of the females had been inseminated. Here, not only is there a reduction of 18 % in the number of females producing offspring, but also the number of offspring per culture is reduced much below that found in homogamic matings. Between these two extremes, various de­grees of reduction of hybrid production were found among the different crosses. These will be described in the following pages in conjunction with the photomicrographs. The data for the twenty-six crosses in which one or niore females had been inseminated in heterogamic matings are summarized in Table 2. TABLE 2 Condition of the reaction mass, vagina, and motile sperm after the males and females had been exposed to each other for ninety-six hours Inseminated Reaction Vagina Motile Hybrids Crosses females mass normal sperm produced 1. aldrichi <;> X arizonensis J ---------------------------­ 36 28 8 24 yes 2. aldrichi <;> X hamatofila J ----------------------------­ 14 14 0 0 none 3. aldrichi <;> X mojavensis J ---------------------------­ 66 46 20 40 yes 4. aldrichi <;> X mulleri J --------------------------------­ 24 24 0 5 none 5_ mulleri <;> X aldrichi J ---------------------------------­ 36 20 16 18 yes 6. mulleri <;> X arizonensis J -­---------------------------­ 13 9 4 8 yes 7_ mulleri <;> X buzzatii J ----------------------------------­ 2 1 1 1 few 8. mulleri <;> X hamatofila J ----------------------------­ 13 12 1 2 yes 9. mulleri Q X mojavensis J -----------------------------­ 32 22 10 18 yes 10. arizonensis <;> X buzzatii J ___________ _,_ _______________ 1 0 1 1 larvae 11. arizonensis <;> X hamatofila J ----------------------­ 27 27 0 3 none 12. arizonensis <;> X mojavensis J ---------------------­ 39 16 23 19 yes 13. buzzatii <;> X arizonensis J ---------------------------­ 156 156 0 97 none 14. buzzatii <;> X hamatofila J ---------------------------­ 129 129 0 2 none 15. buzzatii <;> X mojavensis J -----------~ ­--------------­ 25 25 0 14 none 16. hamatofila <;> X arizonensis J -----------------------­ 7 7 0 2 none 17. hamatofila <;> X buzzatii J ---------------------------­ 1 1 0 1 none 18. hamatofila <;> X mojavensis J -----------------------­ 24 11 13 16 none 19. hamatofila <;> X ritae J -------------------------------­ 1 1 0 0 none 20. mojavensis <;> X aldrichi J ---------------------------­ 24 24 0 10 none 21. mojavensis <;> X ariz0 nensis J ---------------------­ 154 97 57 80 yes 22. mojavensis <;> X hamatofila J -----------------------­ 46 4°6 0 0 none 23. peninsularis <;> X arizonensis J -------------------­ 52 52 0 24 none 24. peninsularis <;> X buzzatii J -------------------------­ 1 1 0 0 none 25. peninsularis <;> X hamatofila J ··--------------------­ 19 19 0 0 none 26. ri tae <;> X hamatofila J ---------------------------------­ 10 10 0 2 none These figures were compiled from Tables 1 and 3 of the preceding article on sexual isolation. They show the condition of the reproductive tract ninety-six hours after the males and females had been placed together. The five columns, following the column giving the nature of the cross, indicate respectively the total number of inseminated females, the num­ber with the reaction mass still present in the vagina, the number in which the vagina had cleared, the number with motile sperm in the ventral receptacles, and whether hybrid offspring had been produced. An examination of these data shows that in all crosses in which hybrid offspring were not produced the vagina still contained a reaction mass. A single exception was the cross hamratofila S? X mojavensis J (18). Iri that Studies in the Genetics of Drosophila cross thirteen of the twenty-four dissected females had returned to an apparently normal condition, and a few motile sperm were still present at the tip of the ventral receptacle. This cross never result$ in the pro­duction of hybrids, not even of the inviable larvae. In contrast to this, in all ten crosses which yielded hybrids, including inviable larvae, a large proportion of their vaginae were already cleared and normal by the end of ninety-six hours. In five of the crosses which failed to produce hybrids, the vagina of each contained a reaction mass and all sperm had disappeared from the ventral receptacle (2, 19, 22, 24, 25). In the interspecific cross of D. buzzatii ~ X D. arizonensis c;, a timed series of sixty matings was obtained, and the dissections revealed that fifty-five of these females had been inseminated. This series gave a complete history of the insemination reaction for a period of ninety-six hours. For the first few hours after insemination the reaction is similar to that occurring in homogamic matings, but instead of clearing up in eight or nine hours, the vaginae retain the reaction mass for at least as long as 192 hours. Figures 13 to 16 in Plate IV show four stages in the history of the insemination reaction for this cross. The specimen illustrated in Fig­ure 13 was photographed twenty-seven hours after insemination. As the photograph clearly indicates, the reaction mass was very opaque and occupied the entire cavity of the vagina. The vagina remains essentially in this condition for the next twenty-five or thirty hours. At about fifty-six hours the reaction mass becomes distinctly pear-shaped, with its broad end occupying the vaginal pouch and then tapering down to a blunt point at the lower end of the vaginal cavity. By seventy-two hours, some of the reaction material begins to pass out through the vaginal orifice. This material is not suddenly expelled as in homogamic matings, but is best described as draining out gradually. Figure 14 illustrates the beginning of the drainage process. This continues up to ninety-six hours, when the reaction mass becomes a discrete object, s}ia:rply out­lined, and largely confined to the vaginal pouch (Fig. 15). It remains in this condition 'through the 1'20-hour stage (Fig. 16), and, in most cases, for many hours thereafter. A Jarge number of inseminated females have been dissected up to 192 hours after copulation, and in most of them the mass was still present. It may become smaller, either through further drainage or by disintegration, especially should the female begin laying eggs. · In three females, in which the reaction mass had decreased to about one-fourth the volume of the one illustrated in Figure 15, contractions of the dissected vagina in the saline medium had resulted in voiding the mass in two cases and a failure to do so in the third one. In the latter instance the reaction mass had become tightly wedged in the vaginal orifice. The possibility that a female may thus eventually eliminate the remnants of the reaction mass is supported by observations on a fourth case. The female in question had become moribund, and upon making The University of Texas Publication the. dissection it was found that the vaginal orifice had become occluded by the old reaction mass. A few similar c~ses were observed in females in other interspecific crosses during the course of our dissections. PrJATE IV. Four stages of the reaction mass of heterogamic matings of D. buzzatii'i' X D. arizonensis i!J . Fig. 13, at twenty-seven hours. Fig. 14, at seventy-two hours. Fig. 15, at ninety-six hours. Fig. 16, at one hundred and twenty hours. Studies in the Genetics of Drosophila It remains to comment on the fate of the sperm in the ventral recep­tacle, egg laying, and the possibility of hybrid production in the buzzatii/ arizonensis cross. As shown in Table 2 (13), ninety-seven of the 156 in­seminated females listed still contained motile sperm in their ventral receptacles at the end of ninety-six hours. A few have been found in this receptacle as late as the 168th hour, but eventually all such sperm disappear. In the meantime the females may lay a few abnormal eggs, but examination by the smear technique has failed to reveal a single spermatozoon in such eggs. Sometimes an ovulated egg will remain in the vagina and undergo disintegration. Three such eggs are illustrated in Plate X. In view of the facts just stated, it is not surprising that we have failed to obtain hybrid offspring from this cross, even though many tests have been carried out for this purpose. In addition to the series from the buzzatii/ arizonensis cross illustrated in Plate IV, photographs of the inseminated vaginae from twenty other interspecific crosses are shown in Plates V-IX. These are presented in the same order as these crosses are listed in Table 2, with the omission of five combinations (7, 10, 17, 19, 24) in which only one or two per cent of the dissected females were found to have been inseminated. Females of D. aldrichi were inseminated by the males of four other species of the subgroup (Table 2, crosses 1-4). Illustrations of these crosses appear in Plate V. The aldrichi/ arizonensis cross is shown at the seventy-two hour stage in Figure 17. The reaction mass is still large, but the drainage process is in progress, as is indicated by the broad band of reaction material extending down from the lower side of the mass to the posterior end of the vagina. Thirty-six inseminated females were obtained from this cross (1). Twenty-eight of these had some form of the reaction mass, and in only eight was the vagina entirely clear, although several others probably would have cleared in time. In twenty-four specimens the ventral receptacle contained motile sperm. In two of the females a disintegrating egg was present in the vagina. This cross yields sterile female hybrids. Fourteen inseminated females were obtained in the aldrichi/ hamatofila cross (2). A ninety-six hour stage is illustrated in Figure 18. The reaction mass was confined to the vaginal pouch in all fourteen cases, and in the one illustrated the last stages of drainage are indicated by the dark, irregular line extending down from the lower side of the mass. None of the vaginae were clear, and sperm were entirely absent from their ventral receptacles. This cross does not produce off spring. A seventy-two hour stage of the aldrichi/ mojavensis cross is shown in Figure 19. Sixty-six inseminated females were obtained from this cross (3), and the reaction mass was present in forty-six specimens. In twenty females the vagina was entirely clear, and forty had motile sperm in their ventral receptacles. In each of two females an abnormal egg was lodged in the vagina. This cross gives sterile females. In the aldrichi/ mulleri cross all twenty-four inseminated females had conspicuous reaction masses (Fig. 20), and consequently none of the The University of Texas Publication vaginae were normal. Only five of the females still had motile sperm in the ventral receptacle. There were five cases of disintegrating eggs in the vagina, and in one of these the debris of several eggs could be PLATE V. Reaction mass in vagina of D. aldrichi from heterogamic matings with ·four different types of males. Fig. 17, with D. arizonensis male at seventy-two hours. Fig. 18, with D. hamatofila male at ninety-six hours. Fig. 19, with D. rnajovensis male at; seventy-two hours. Fig. 20, with D. mulleri male at ninety-six hours. Studies in the Genetics of Drosophila recognized by the presence of the egg filaments. This cross does not produce offspring. · The females of D. mulleri were inseminated by the males of five other species (Table 2, crosses 5-9). The affected vaginae of four of these PLATE VI. Reaction mass in vagina of D. mulleri from heterogamic matings with four different types of males. Fig. 21, with D. aldrichi at seventy-nine hours. Fig. 22, with D. arizoni;nsis male at seventy-eight hours. Fig. 23, with D. hamatofila male at ninety-six hours. Fig. 24, with D. mojavensis male at eighty hours. The University of Texas Publication are shown in Plate VI. The vaginae of twenty of the thirty-s~x insem­inated females from the mulleri/ aldrichi cross still showed the reaction mass (Fig. 21), and in the remaining sixteen they had cleared and were practically normal. Motile sperm were present in eighteen specimens. In each of two females with large reaction masses a degenerating egg was present in the vagina. This cross produces sterile males and females. Only thirteen inseminated females were derived from the mulleri/ arizonensis cross (6). Nine of these had reaction masses, and in the remaining four the vagina was entirely clear. Eight specimens had motile sperm in the ventral receptacle. This cross produces sterile male hybrids only. Figure 22 illustrates a seventy-eight hour stage. The mulleri/ buzzatii cross (7) yielded but two inseminated females in the experimental tests. In one of these the reaction mass was present, but in the other the vagina was clear and its ventral receptacle contained motile sperm. This cross gives a few abnormal flies in mass matings. It · is not illustrated in the plates. The mulleri/hamatofila cross (8) gave thirteen inseminated females, and by the end of ninety-six hours, twelve of these still had reaction masses (Fig. 23). In one the vagina was cleared, and in two motile sperm were present in the ventral receptacle. The writer has obtained from this cross, by the use of large mass matings, a few sterile males ·and females. The mulleri/ mojavensis cross (9) gave thirty-two inseminated females, of which twenty-two showed a reaction mass, and ten had the vagina clear. Eighteen females had motile sperm in their ventral receptacles. In each of five females, which still had the reaction mass, a degenerating egg was present in the vagina. This cross gives fertile females and sterile . males. An affected vagina at eighty hours is illustrated in .Figure 24: The females of D. arizonensis were inseminated by the males of three , other species (Table 2, crosses 10-12). The arizonensis/ buzzatii cross gave but a single inseminated female in the experimental tests, and the vagina was clear by ninety-six hours and the ventral receptacle con:. tained motile sperm. This cross produces hybrid larvae which die durjng mid-larval stages. In order to obtain such larva,e, it was necessary to use large mass matings. This cross is not illustrated in the plates. The arizonensis/ hamatofila cross (l'l) yielded twenty-seven inseminated females, all of which had reaction masses, but only three with motile sperm in their ventral receptacles. Degenerating eggs were present in the vaginae of five females, and in one of these a second egg was present in the median oviduct. This cross never produces offspring. Figure 25 (Plate VII) illustrates the beginning of the drainage process at the seventy-two hour stage. From the arizonensis/ mojavensis cross thirty-nine inseminated females were obtained (12), of which sixteen showed the reaction mass and twenty-three had clear vaginae. Motile sperm were present in the ventral receptacle of nineteen females. The vagina clears up .rather rapidly in females of this cross, and the reaction mass is not clearly delimited Studies in the Genetics of Drosophila (Fig. 26) as it is in certain other combinations. Fertile females and sterile males result from this cross. PLATE VII. Reaction mass in vaginae of D. arizonensis and D. buzzatii, each mated to two different types of males. Fig. 25, D. arizonensis ~ X D. hama-tofilai!J at seventy­two hours. Fig. 26, D. arizonensis ~ X D. mojavensis i!J at seventy-two hours. Fig. 27, D. buzzatii~ X D. hamatofila i!J at sixty-two hours. Fig. 28, D. buzzatii~ X D.moja­vensis i!J at seventy-nine hours. The University of Texas Publication The females of D. buzzatii were inseminated by the males of three other species (Table 2, 13-15). The buzzatii/ arizonensis cross gave' 156 inseminated females (13), all of which had distinct reaction masses at the ninety-six hour stage. This cross has been considered above in con­nection with the illustrations in Plate IV. The buzzatii/ hamatofila cross (14) produced 129 inseminated females, all of which showed a reaction mass. In only two specimens, both recent inseminations, were live sperm present in the ventral receptacle. In ail other females the sperm had completely disappeared. The vagina of one female contained an abnormal egg. Figure 27 shows a typical reaction mass of this cross, which never results in the production of hybrids. Twenty-five females had been inseminated in the buzzatii/ mojavensis cross (15) , and all of these had reaction masses. Fourteen females still had live sperm in the ventral receptacle. A degenerating egg was present in the vagina of four females. Figure 28 illustrates a seventy-nine hour stage. This cross never results in the production of hybrids. The females of D. hamatofila were inseminated by the males of four different species (Table 2, crosses 16-19). In the hamatofila/ arizonensis cross only seven females were found to have been inseminated (16), and their vaginae all contained a reaction mass. Figu:re 29 illustrates a seventy-two hour stage. The mass is still very opaque, and the female tract had been severe1y damaged. Offspring were not produced in this cross .. The hamatofila/ buzzatii and hamatofila/ ritae crosses each gave a single inseminated female, with the reaction mass still present (17, 19). Neither cross yielded hybrids, even in mass matings. These two crosses are not illustrated in the plates. The hamatofila/ mojavensis cross, as pointed out above, is the only one in which most of the vaginae had become normal, and yet the cross had failed to produce hybrids (18). It should be pointed out that the vagin;:i returns to, normal rather early after insemination. For example, the one illustrated in Figure 30 was only forty-eight hours old at the time of dissection, and those dissected between fifty and sixty hours were mostly cleared, and motile sperm were still present in the ventral recep­tacle. In older specimens the sperm soon become inactivated or killed, as occurs in many interspecific matings. The females of D. mojavensis were inseminated by the males of three other species (Table 2, crosses 20-22). Twenty-nine inseminated females were produced in the mojavensis/ aldrichi cross (20), and a reaction mass was present in all. By seventy-two hours after insemination, the reaction mass begins to break up and is later eliminated (Fig. 31). No hybrids were obtained from this cross. The mojavensis/arizonensis cross (21) yielded 154 inseminated females ' ' of which ninety-seven contained reaction masses (Fig. 32). '}'he vaginae of fifty-seven were normal, and eighty had motile sperm in their ventral receptacles. This cross is by far the most productive of hybrids of any Studies in the Genetics of Drosophila combination in the mulleri subgroup, and the F1 males and females are both fertile. PLATE VIII. Reaction mass in vaginae of D. h.amatofila and D. mojavensis, each mated to two different types of males. Fig. 29, D. hamatofila 9 X D. arizonensis J at seventy-two hours. Fig. 30, D. hamatofila 9 X D. mojavc,nsis J at forty-eight hours. Fig. 31, D. mojavensis 9 X D. aldrichi J at seventy hours. Fig. 32, D. mojcivensis 9 X D. arizonensis J at seventy-eight hours The University of Texas Publication Figure 33, Plate IX, illustrates the reaction mass in an inseminated female from the mojavensis/ hamatofila cross. Forty-six inseminated Studies in the Genetics of Drosophila females were obtained and all possessed reaction masses (22). This inter­specific cross never produces hybrids. Females of D. peninsularis were inseminated by the males of three other species (Table 2, crosses 23-25). Fifty-two inseminated females were derived from the peninsularis/ arizonensis cross, and all had reaction masses (23). No hybrids were obtained. A typical specimen at seventy­two hours is shown in Figure 34. A few cases of abnormal eggs in the vaginae were observed, and in one of these there was a normal egg in the median oviduct in addition to the abnormal . one in the vagina. The peninsularis/ buzzatii cross (24) gave a single inseminated female, which had a reaction mass in the vagina and a few non-motile sperm in the ventral receptacle. No offspring resulted from mass matings of this cross, which is not illustrated in the plates. . The peninsularis/ hamatofila cross (25) yielded nineteen inseminated females. All of these showed old reaction masses (Fig. 35), and the sperm had completely disappeared from their ventral receptacles. No offspring were obtained from these matings. The hamatofila/ ritae cross (26) gave a single inseminated female, which had a reaction mass containing a degenerating· egg. Figure 3.6 was made from another specimen from mass matings. Sperm were absent from the ventral receptacles, and hybrids were not produced. Figures 37-39, Plate X, represent three vaginae, each with an egg which is undergoing disintegration as a result of the effects of the insemination reaction. The one illustrated in Figure 37 is from the mulleri/ arizonensis cross, from a specimen dissected 150 hours after insemination. The ,egg is embedded in the old reaction mass, and has already become abnormal with the cytoplasm condensed into a spherical mass near the center. Figure 38 illustrates a specimen from the hamatofila/ arizonensis cross from a female dissected seventy-two hours after copula­ tion. The abnormal egg is located in the vaginal pouch of a badly dam­ aged vagina. Figure 39 illustrates'a specimen from the mulleri/ mojavensis cross, dissected seventy-two hours after copulation. The egg is located in the vaginal pouch, and has turned 1'80 degrees on its long axis with the filament-bearing end directed posteriorly. The content of the egg has largely disappeared and only the partially collapsed shell remains. The scattered material of the reaction mass has been pushed down into the lower end of the vaginal cavity. Recently discharged eggs from the ovaries of these females, while still in the median oviduct, appear to be entirely normal. It is only after the egg has reached the vaginal cavity with its contained reaction material that the dissolution or "digestive" process begins. If the female is un­ able to lay the eggs, as usually happens in severe cases, the disintegra­ tion may continue .to the point where only the hexagonally sculptured chorion and egg filaments are left, and sometimes only the filaments. This made it possible to detect cases in which two or more eggs had been dissolved. The largest number found was in a vagina with the remains of twenty filaments, which would indicate that five eggs had been digested. The University of Texas Publication PLATE X. Figures 37 to 39 show abnormal eggs in the vaginae of three females. Fig. 37, D. niulleri 9. X D. arizoni;nsis i!) at 150 hours. Fig. 38, D. hamatofi,la 9. X D. arizonensis i!) at seventy-two hours. Fig. 39, D. mulleri 9. X D. niojavensis i!) at seventy-two hours. Fig. 40, vagina of D. niojavensis 9. mated to a sterile F1 male derived from the arizonensis/mojcrnonsis cross at three hours. Such males have sperm­free semen. Studies in the Genetics of Drosophila Some interesting facts concerning the nature of the reaction mass; not emphasized above, should be mentioned at this point. It has already been stated that in homogamic matings, the reaction mass remains soft and is soon eliminated, and consequently does not materially, if at a,ll,. affect the production of offspring. In the case of heterogamic matings,_ not only does the reaction mass persist for a much longer .period, of time, but it also changes in consistency, frequently becoming extremel¥ hard in texture. The degree of this change depends on the cross. . In general, crosses which produce viable hybrids have the softer type. of reaction mass and the vagina clears in a shorter time. This includ~s. crosses between aldrichi, arizonensis, mojavensis, and mull.eri (c.f., Figs. 17, 19, 21, 22, 24, 26, 31, 32). On the other hand, crosses involving the other four species usually produce the hard, discrete type of. reaction mass, which persists for a.. long time after copulation ( c.f., Figs. 15, 16,. 18,. 23, 25,. 27, .28, 29, .. 33-36), and hybrids usually do not r~sult .fro~. these crosses. By far the hardest type of reaction mass develops in crosses i:ti which D. hamatofila is used.as the male parent. The vagina of the female 'in SUCh 'crosses, irrespective of the species employed, frequently . }JeCO!'.fleS as hard as a grain of sand, and it is difficult to crush the organ by pres­sure on the cover glass. If examined under the high power of the inicro­scope, the reaction mass appears to be crystalline 'in character. This is clearly discernible in such Figures as · 23, 25, 33, and 35. This demon­strates that the .semen must play a definite role in determining the physical character of the reaction mass. ... RoLE OF THE SEMEN IN THE INDUCTION OF THE INSEMINATION REACTION One of the most important questions concerning the insemination re~. action is the possible role of the semen and its contained sperm in bring­ing about the formation of the reaction mass. It seemed clear from. the experimental results that the reaction might be induced either by· the living spermatozoa, or by the semen alone. In studying the early stages: of the formation of th.is mass, it had been observed that, in a few cases1 the reproductive tract of the female contained no detectable sperm, ·and' yet the typical insemination reaction had immediately followed cQ.pu­lation. A single observation will be sufficient to make this point clear. In .examining the early stages of the reaction ill homogamic, matillgs of D. aldrichi, one case was found in which the vagina had already rea'ched its maximum size, although the female was dissected before the termination ·of coitus and after the pair had been in copula for .two ·minutes and ten seconds. The vagina was large and turgid, and its cavity was filled with a clear fluid throughout which was .scattered a grayish. reaction material. The most careful examination, under optimum light­ ing . conditions, failed to reveal . a single spermatozoon in the d1stenqed vagina or the ventral receptacle. From this, and similar observationsr. The University of Texas Publication it may be assumed that the sperm are not the active agent in causing the reaction to take place. Fortunately, an opportunity was present for making a critical test of this assumption. In some of the interspecific crosses sterile F1 males are produced, and in two such crosses a sufficient number of hybrid males could be obtained for making the test. The two crosses are, D. arizonen­sis ~ X D. mojavensis ~ and D. mulleri ~ X D. mojavensis ~. In the first of these crosses the F1 males appear normal, and their testes are nearly as Jarge as those of fertile males, but spermatozoa are never formed in them. Sterile hybrid males from the mulleri/ mojavensis cross are also rather normal in appearance, except that their testes are small and abnor­mal. These males likewise fail to produce spermatozoa. Both types of males will sometimes mate in backcross to the parent and other kinds of females. These facts give the needed opportunity for testing the assump­tion that live spermatozoa are not the active agent in causing the reaction to occur. The arizonensis/ mo.iavenses F1 males were tested with seven different types of females, and failed to mate with those of hamatofila, mulleri, and peninsularis. The results obtained with the four other types of females are tabulated in Table 3. In the backcross with mojavensis females, T ABLE 3 Sterile F, males from the arizonensis/ mojavensis cross mated to mojavensis, aldrichi, arizonensis, and buzzatii females r-· : ­ Types of Dissected females at Character of reaction mass Illustrated D. moiavensis ----------------------­ n. moiavensis____ __________________ _ D. mo iavensis_____________________ n. mo iavensis________________________ D. moiavensis ----------------------­ . . n. mo1avens1,s ·----------------------­ D. mo iavensis ·----------------------­ n.moiavensis________________ __ ____ . . n. mo1avensis·---------------------­ n. moiavensis·----------------------­ n. moiavensis-----------------------­ n . . . mo1avensis-----------------------· D. arizonensis ----------------------­!>. aririt!i nensis ----------------------­ n . . . anzonensis ----------------------­!>. arizonensis ---------------------"­!>. arizonensis·----------------------· n.arizonensis_________ _______ _______, n. rddrichi ---------------------------­!>. huzzatii ----------------------------· D. buzzatii ------------------------­ 0 minutes 12 minutes 1 hour 2 hours 3 hours 5 hours 10 hours 16 hours 22 hours 24 hours 26 hours 30 hours 0 minutes 15 minutes '10 minutes 35 minutes 23 hours 24 hours 24 hours 15 minutes 26 hours beginning to enlarge well-formed. opaque maximum size densely opaque no change no change no change no change clearing mass reduced to 1h mass reduced to 14 vagina normal bev.inning to form well-formed large, opaque maximum size mass reduced to 1;,, mass reduced to %. % r,lP,ared well-formed % cleared Pl. X, Fig. 40 Pl. XI, Fig. 41 Pl. XI, Fig. 42 Pl. XI, Fig. 43 Pl. XI, Fig. 44 Pl. XII, Fig. 45 Pl. XII, Fig. 46 twenty-three successful conulations were obtained. The females were dissected at regular intervals, and the results obtained from a represent­ative series of twelve of these are listed in the table. From the end of coitus to the twenty-sixth hour, the history of the insemination reaction Studies in the Genetics of Drosophila closely parallels that of homogamic matings, both in the sequence of events and in the behavior of the reaction mass. The vagina begins to enlarge immediately following copulation and reaches its maximum size in about an hour. Within two hours it has become densely opaque and remains so through the sixteenth hour (Figs. 40-42). Soon after this, the vagina begins to clear and an indistinct reaction mass occupies the vaginal pouch (Fig. 43). Some time between the twenty-sixth and thirtieth hours it becomes cl~ar and normal. In these backcrosses the chief difference in the history of the insemination reaction between pure homogamic matings and those from the sterile males is the time element. It takes about three times as long for the vagina to clear in the latter type of mating. It was more difficult to obtain backcross matings between the sterile males and arizonensis f em:ales. Only six successful copulations were secured, and the results from these are listed in Table 3. Four of the females were dissected within thirty-five minutes after mating had oc­curred. The behavior of the insemination reaction during this time was very similar to that of the matings with mojavensis females for the corre­sponding period. One female was dissected after twenty-three hours, and the vagina had been reduced to about one-half maximum size. Another female was examined after twenty-four hours, and in this one the vagina had decreased to about one-fourth the volume of the maximum size. From this evidence, the vagina would probably have been normal by thirty hours. In the outcross of the sterile males to aldrichi females, only a single successful copulation was obtained. The female was dissected at twenty­ four hours, and at this time the vagina contained a reaction mass in the vaginal pouch of the heterogamic type (Fig. 45). Two successful. copu­ lations occurred in the outcross matings between the sterile males and females of buzzatii. One female was examined after fifteen minutes, and it had a typically enlarged vagina. The other was dissected at twenty-six hours. The reaction mass was located in the vaginal pouch, and was also of the heterogamic type (Fig. 46). Attempts to backcross the sterile F 1 males from the mulleri/ mojavensis cross proved to be still more difficult. After many attempts, two matings with mulleri females were obtained on the thirteenth day after the emergence of these males. The first successful mating lasted but thirty seconds and the vagina was swollen and slightly opaque when dissected out at twenty minutes. The second mating lasted thirty-seconds, and the female was dissected one hour later. Her vagina was larger than that of the first female, and had a mottled appearance. Every effort to obtain matings with the mojavensis females completely failed. Since the semen of the F1 hybrid males is sperm-free, the results from the tests with these males demonstrate conclusively that the presence of live spermatozoa are not necessary for the induction of the insemina­ tion reaction. This does not exclude the possibility that degenerative products from the abnormal testes may constitute a part of the semen, The University of Texas Publication Not much is known about the exact composition of the semen in Dro­sophila. According to Nonidez (1920), the paragonia constitute a pairof accessory glands which secrete into the vas deferens a dense sticky · PLATE XI. · Figures 41 to 43 belong to the series begun in Fig. 40, and represent . different stages of the insemination reaction. Fig. 41, at ten hours. Fig. 42, at sixteen hours. Fig. 43, at twenty-two hours. Fig. 44, from cross of D. arizonensis female with one 6f these sterile males, twenty-three hours after insemination. Studies in the Genetics of Drosophila fluid with floating refractive granules. This fluid becomes mixed with the spermatozoa from the testes and forms the liquid part of the ejaculate. PLATE XII. Fig. 45, from mating of D. aldrichi female with sterile Fi male, dissected at twenty-four hours. Fig. 46, from mating of D. buzzatii female with sterile Fi male, dissected at twenty-six hours. Figures 47-52 are D. buzzatii females from the buzzatiil arizonensis cross remated at different intervals with D. buzzatii males. Fig. 47, female remated at 168 hours and dissected thirty minutes later. Fig. 48, same cross dissected at one hour. The University of Texas Publication The sexual behavior of the sterile males used in this series of tests was very different from that of the fertile males of the parent species. For ten days after emergence, these males were indifferent to the presence of females, and it was not until the thirteenth day, or later, that the matings listed in the table were obtained. In contrast to this, the fertile males began courting as early as the fifth day, and always carried on a vigorous courtship when placed with females. The duration of coitus was somewhat shorter for the hybrid males, averaging two minutes and sixteen seconds as compared with three minutes and thirteen seconds for the fertile males. MATING HABITS OF FEMALES WHICH HAD PREVIOUSLY COPULATED One question which has arisen in connection with these tests on mem­bers of the mulleri subgroup, is whether a female which had been insem­inated would copulate again, and if so, whether the insemination reaction would occur a second time. There is considerable evidence to show, both from this and other laboratories, that in a species in which the insemina­tion reaction does not occur, the female will mate two or more times, and in some species at .frequent intervals. We found, however, that in a species in which a strong reaction occurs, the female usually does not ;remate for a long time. It was further found, both after homogamic and heterogamic matings, that the females do not immediately respond to the vigorous courtship of the males. A special test was made of homogamic matings of D. buzzatii. A series of ten females which had just copulated for the first time were isolated, and each female was exposed to five males of her own kind at intervals extending from the fourteenth to the 135th hour after the initial copulation. Each exposure lasted from one to two hours at the following periods: 14, 20, 43, 61, 85, 114, and 135 hours. During this entire series of tests the flies were kept under constant observation while exposed, and not a single copulation took place, although the females were courted vigorously by the males. It may not be safe to conclude that the mated females of all members of this group of species never mate a second time. If they do, it must be after a considerable period of time. We also made tests with females from the heterogamic buzzatii/ arizo­ nensis cross. Females which had been seen mating with arizonensis males were used. A limited series of inseminated females were tested to ari­ zonensis males, but we were never able to secure a second mating with this type of male. Much better success was achieved when buzzatii males were used. A total of forty-seven females was isolated immediately following coitus, and then tested singly to the buzzatii males. The earliest time at which these females would accept the males was about 120 hours after the end of the fir.st mating. Dissections revealed that forty-six of the females still had the old reaction mass present in the vaginal pouch, and one did not, and that thirty-four had been reinseminated. The fail­ Studies in the Genetics of Drosophila ure of the other thirteen specimens to become inseminated is not clear. It is possible that the presence of the old reaction mass adversely affected the insemination act. The data from a representative sample of twenty-seven specimens are tabulated in Table 4. Female No. 5 is the one which did not have an old reaction mass. Incomplete copulations are not infrequent in Drosophila, especially in heterogamic matings, and this is probably the explanation of this case. The time at which the second mating occurred in the several females is given in hours in the column on the time of mating, while the time of dissection is given in minutes or hours following the second mating. Dissections were begun at fifteen minutes and continued through the eighth hour. They were discontinued at eight hours because by this time the vagina is practically normal, except for the presence of the old reaction mass. TABLE 4 D. buzzatii females from the buzzatii/arizonensis cross remated to buzzatii males Time of Dissected Reaction mass Female remating at old new Illus trated 1 144 hours 15 minutes present pr_esent 2 168 hours 15 minutes present present 3 120 hours 20 minutes present present 4 120 hours 20 minutes present present 5 6 144 hours 168 hours 30 minutes 30 minutes absent present present present Pl. XII, Fig. 47 7 144 hours 30 minutes present present 8 144 hours 32 minutes present present 9 144 hours 35 minutes present present 10 11 124 hours 124 hours 1 hour 1 hour present present present present Pl. XII, Fig. 48 12 125 hours 1 hour present present 13 144 hours 1 hour Qresent present 14 144 hours 1 hour present present 15 144 hours 1 hour present present 16 144 hours 1 hour present present 17 160 hours 1 hour present present 18 120 hours 2%, hours present present 19 20 21 22 23 24 25 120 hours 144 hours 144 hours 168 hours 168 hours 168 hours 168 hours 3 hours 3%, hours 6 hours 7 hours 7% hours 7112 hours 7%, hours present present present present present present present present present present present present present present Pl. XIII, Fig. 49 Pl. XIII, Fig. 50 Pl. XIII, Fig. 51 26 27 168 hours 168 hours 7 5/6 hours 8 hours present present present cleared Pl. XIII, Fig. 52 A study of the specimens from this close series of dissections made it possible to follow the history of the insemination reaction which re­sults from a second copulation. In general, the reaction undergoes the same series of changes as occur after the first insemination. Within thirty minutes, the new reaction material is formed and is beginning to surround the old reaction mass (Fig. 47), and by the end of one hour it has practically engulfed this object (Fig. 48). The vagina is densely The University of Texas Publication opaque and has become greatly enlarged. This condition continues for the next two or three hours, during which time the vagina has become somewhat larger. There then follows the elimination process of the new reaction material. This is evident in the specimen illustrated in Figure 49, which was dissected three and three-fourths hours after copulation. The old reaction mass has again become visible. The new reaction mate­rial continues to drain out during the next few hours (Figs. 50, 51). By the end of eight hours, only a small trace of this material is present at the lower end of the vaginal cavity, and the vagina has returned to its original size, while the old reaction mass becomes sharply outlined in the vaginal pouch (Fig. 52). The point of greatest interest here is that the old reaction mass from the original .heterogamic mating does not materially affect the history of the insemination reaction from this second or homogamic mating. The course of the second reactiQn, both in the sequence of the changes and the time of duration, very closely parallels the corresponding changes and time which follow an original homogamic mating of D. buzzatii. DISCUSSION AND CONCLUSIONS The main purpose of this article was to present any evidence obtained from a study of the insemination reaction which might have some bear­ing on the question of speciation in the niulleri subgroup of species. In any attempt to evaluate this evidence, one must keep in mind that evolu­tionary divergence is rarely, if ever, brought about through the operation of a single isolating mechanism. It is usually the total effect of several of these mechanisms which brings about a more or less complete separa­tion of two diverging forms, by reducing, or entirely preventing, the exchange of genes between them. It was, therefore, not surprising to find that at least five such mechanisms were functioning between the several members of this subgroup. We have shown in the preceding article that sexual isolation is a common mechanism in preventing matings between tlie eight members of the mulleri subgroup. Of the ,fifty-six possible crosses, thirty failed to show a single insemination among the one-hundred females examined by dissection for each of these crosses. In addition to this, there were several other crosses which yielded but a few inseminations, some as low as a single specimen. Sexual isolation obviously represents one of .the most important mechanisms in the evolution of Drosophila. We have shown in previous publications that the spermatozoa may becoine inactivated, or even killed, when introduced into the reproduc­ tive tract of an alien female. Much evidence of this phenomenon was observed in the present investigation. It was clear from a study of the results from certain crosses that the spermatozoa soon become inactive, and may eventually disappear. For example, in one cross (hamatofila! mojavensis) dissections showed that twenty-four hours after copulatio~ there were many motile spermatozoa present in the ventral receptacle, Studies in the Genetics of Drosophila but later dissections revealed that most of the sperm had become immotile, or had disappeared altogether. Gametic rnortality constitutes a second type of isolating mechanism. PLATE XIII. Continuation of the series begun with Fig. 47, Plate XII. Fig. 49, same cross, female remated at 144 hours and dissected at three and three-fourths hours. Fig. 50, same cross, female remated at 168 hours and dissected at seven hours. Fig. 51, same cross and same age, dissected at seven and three-fourths hours. Fig. 52, same cross and same age, dissected at eight hours. The University of Texas Publication In crosses between arizonensis 'i? X buzzatii t and mulleri 'i? X buz­zatti c; the larvae frequently perish. In the first cross they all die during mid-larval stages, and in the second a majority of the zygotes die in the pupal stage, with only a few of the adults emerging (Patterson 1942). Zygotic mortality, therefore, represents a third type of isolating mechan­ism in the mulleri subgroup. Reference to Table 1 of the preceding article shows that only ten of the fifty-six possible interspecific crosses gave hybrids of any kind. More­over, most of these hybrids were found to be completely sterile. Only the following combinations yield fertile off spring: I, the mulleri 'i? x mo­javensis c; cross gives fertile females and sterile males; 2, the arizonen­sis 'i? x mojavensis 8 cross likewise gives fertile females and sterile males; 3, the reciprocal cross, mojavensis 'i? X arizonensis c; , gives fertile males and females. Hybrid sterility represents a fourth type of isolating mechanism. With this array of isolating mechanisms functioning in the mulleri subgroup (and there are probably others), it is somewhat difficult to determine precisely just what proportion of the sum total of isolation in this group should be assigned to the effects of the insemination reaction. It is clear from the experimental results that this reaction constitutes a potent isolating mechanism in the interspecific crosses of the mulleri­like forms. Among thirteen such crosses which did not yield hybrids, 557 inseminated females were found in 2,600 dissected specimens (21% ) . In one cross (buzzatii/ arizonensis), a total of 159 of the 200 dissected females had been inseminated (78%). It is safe to assume that, in the absence of a reaction mass from the vagina, many of these females would have produced offspring. All of these females did have a reaction mass in the vagina, and motile sperm were present in the ventral recep­tacles of sixty-two per cent of them (Table 2). Their failure to produce offspring was not due to the absence of motile sperm in all specimens, but rather to their inability to lay fertile eggs in the presence of the reaction mass. The insemination reaction, in addition, adversely affects · reproduction in fertile crosses, by reducing the number of hybrids. Thus, in the mojavensis/arizonensis cross, which is the most fertile one of the entire series, only seventy-five per cent of the females produced hybrids, although ninety-three per cent of them had been inseminated. In the absence of workable paleontological material of Drosophila, it is not possible to work out a satisfactory phylogeny for the different forms. However, the several species belonging to the mulleri subgroup must have evolved from a common stem, and their morphological fea­tures, together with the experimental results now available, do make it possible to indicate some of their relationships to one another. From their taxonomic features alone, mulleri, arizonensis, aldrichi, and moja­vensis are more closely related to one another than to the other four species. This conclusion is supported by the results obtained in the cross-fertility tests. Seven of the ten crosses producing hybrids were Studies in the Genetics of Drosophila restricted to members of these four species, and one member of each of the other fertile crosses also belonged to this group (Table 2). The two most closely related of these four species are arizonensis and mojavensis, for their reciprocal crosses yield by far the largest number of fertile hybrids of any combination. In addition, arizonensis females are sometimes inseminated by buzzatii males, but the larvae die. Next in closeness of relationship would be mulleri and aldrichi. These two species are very similar morphologically, but have developed complete reproduc­tive isolation, and for this reason are able to maintain their identities in the large area of their overlapping distribution ranges. Fertile crosses occur between mulleri females and the males of both arizonensis and mojavensis, and in addition, this species shows its relationship to two other members of the subgroup by virtue of hybridization. Its females are sometimes inseminated by buzzatii males, and although most of the zygotes die in the pupal stage, yet a few adult flies emerge. Its females are also occasionally inseminated by hamatofila males, producing a few sterile hybrids. In all nonfertile crosses the duration of the effects of the insemination reaction is longer than in fertile crosses. This is due to the deleterious effects caused by the prolonged retention of the reaction material. Or to express it in another way, the sooner this material is eliminated from the vagina, the better are the chances that hybridization will follow. The only observed exception to this rule was the mojavensis/ hamatofila cross, but we have already commented on that case. The remaining crosses listed in Table 2 represent nonfertile combina­tions, that is, crosses in which the females have been inseminated but do rtot produce offspring. As to what value should be placed on these data in attempting to estimate a degree of relationship is difficult to determine. There is very little reliable experimental evidence to show whether hybridization occurs between two forms belonging to different species groups, and there is still less relating to the question of insemina­tion. Nevertheless, it seems reasonable to assume that any two species in which a large number of females had been inseminated in nonfertile heterogamic matings would be more closely related than two other forms in which only one or a few females had been impregnated. The following attempt to indicate possible relationships from the data on nonfertile crosses is based on this assumption. The nonfertile crosses involve, in the main, members of the second group of species, namely, buzzatii, hamatofila, peninsularis, and ritae. Of these four species, buzzatii is probably the most closely related to mem­bers of the first group. It is morphologically more similar to them and has the same metaphase chromosome pattern (Wharton 1943). Females of this species were inseminated in nonfertile crosses by the males of three other species (arizonensis, hamatofila, mojavensis), and gave a total of 310 impregnations in 600 dissected females (51.6%). The males of this species inseminated females of four other species, two represent­ing fertile crosses (mulleri, arizonensis), and two nonfertile crosses The University of Texas Publication (hamatofila, peninsularis). The total number of alien females inseminated in these four crosses was only five in 1,000 dissected specimens (0.5%). On some grounds D. hamatofila might be regarded to be more closely related to the first group of species than buzzatii. But it is not as similar morphologically as buzzatii and has a different metaphase chromosome pattern (Wharton). Females of this species were inseminated in non­fertile crosses by males of four other species (arizonensis, buzzatii, mo­javensis, ritae). The total number of inseminations was thirty-three in 800 dissected females (4.1 o/o). The males of this species inseminated females of the other seven species, with 258 inseminations in 1,400 dis­sected specimens (18.4o/o). Only one of the crosses was fertile. Of the two remaining species, D. peninsularis has the same metaphase chromosome pattern as hamatofila, but otherwise does not resemble that species. Its females were inseminated by males of three other species (arizonensis, buzzatii, ha1natofila), with a total of seventy-two insemina­tions in 600 dissected specimens (12%). All three crosses were nonfertile. The males of this species failed to inseminate females of any of the other seven species. The inseminations with ritae occurred only in the reciprocal crosses with hamatofila. Ten ritae females were inseminated by hamatofila males (5 % ) , and a single hamatofila female was insem­inated by a ritae male ( 0.5 % ) . Both crosses were nonfertile. These two species have reached the highest degree of divergence of any !fiembers of the mulleri subgroup, and both have developed complete reproductive isolation. The species ritae is further separated from all other members of the subgroup by morphological differences, and by a different type of metaphase chromosome pattern (Wharton 1943) . The nature of the insemination reaction is a question which remains to be discussed. It was pointed out in the preliminary paper that this reaction follows almost immediately after the introduction of the semen into the cavity of the vagina, and is visibly expressed in the form of a rapid swelling of that organ. This increase in size is due to a secretion of fluid by the cells of the lining membrane of the vagina. Evidently this is caused by a response of this membrane to the presence of foreign material, presumably proteins. This would imply that the membrane is hypersensitive to the proteins. Moreover, its hypersensitiveness must be inherited, since the reaction occurs after the first copulation and without any previous sensitization. In any event, the insemination re­action in its manifestations does resemble certain features of some im­munological reactions. Unfortunately there is very little known about hypersensitivity and anaphylaxis in insects, and any further speculations must await future investigations. REFERENCES Abraham, R. 1934. Z. Parasitenk., 6:559-591. (Behavior .of sperm in cimex.) Anderson, R. C. 1945. A study of the factors affecting fertility of lozenge females of Drosophila melanogaster. Genetics, 30: 280-296. Cragg, F. W. 1920 and 1923. Ind. J. Med. Res., 8:32-79; 11:449-473. (Behavior of sperm in bed bug.) Studies in the Genetics of Drosophila Handlirsch, A. 1900. Verhandl. Zool-Bot. Ges. Wien, 50:105-112. (Fate of sperm insects.) Nonidez, J. F. 1920. The internal phenomena of reproduction in Drosophila. Biol. Bull., 39 :207-230. Pantel, J. 1910. Recherches sur les Dipteres a Larves Entomobies. I. Caracteres parasitiques aux points de vue biologique, ethologique et histologique. La Cellule, 26 :27-216. Patterson, J. T. 1942. Isolating mechanisms in the genus Drosophila. Biological Symposia, 6 :271-287. Patterson, J. T. 1946. A new type of isolating mechanism in Drosophila. Proc. Nat. Acad. Sci., 32: 202-208. Snodgrass,.. R. E. 1935. Principles of insect morphology. McGraw-Hill, New York, N.Y. Townsend, C. H. T .. 1911. Review of work by Pante! and Portchinski on reproduc­tive and early stage characters of Muscoid flies. Proc. Ent. Soc. Wash., 13: 151-170. Wharton, Linda T. i943. Analysis of the metaphase and salivary chromosome mor­phology within the genus Drosophila. Univ. Tex. Pub!., 4313 :282-319. Wigglesworth; V. B; 1939. · The principles of insect physiology. Dutton and Com­pany; New York, N.Y. IV. THE INSEMINATION REACTION IN INTRASPECIFIC MATINGS OF DROSOPHILA MARSHALL R. WHEELER* INTRODUCTION The insemination reaction, first described by Patterson (1946) and discussed more fully by him in the preceding article of this bulletin, pro­vides one of the most striking examples of an isolating mechanisro operat­ing to reduce or prohibit the production of hybrids between species which are not deterred from mating by sexual or psychological isolation. Patter­son has amply demonstrated, in the papers cited, the effectiveness of this barrier to gene exchange in heterogamic matings, i.e., between males and females of different species. An unusual aspect of the reaction phe­nomenon, however, is its occurrence in homogamic matings, i.e., between individuals of the same species. The presence of such a reaction in intra­specific matings forces one to the conclusion that it is a normal conse­quence of insemination and may have some useful function in insemination, fertilization, or oviposition. The extreme form of many heterogamic reactions, then, can best be interpreted as an aberrant condition caused by the deposition of semen in the reproductive tract of an alien female. Early in the insemination studies it was noticed that the extent and appearance of the reaction in homogamic matings varied widely among the different species. In some forms the development of the reaction mass followed the pattern described by Patterson for members of the mulleri subgroup of species. In these forms the vagina of the female swells to three or four times its normal size and is filled for many hours with a dense, opaque reaction mass. In other forms, however, it is much less extreme, while in still others it appears to be weak or entirely absent. As will be pointed out later, it is now believed likely that a reaction is present in all Drosophila species, but, due to its varied expression, is not visibly apparent in some forms. A preliminary report on this phase of the insemination studies was presented by Patterson (1946) in which he described briefly the course of the reaction in homogamic matings of one of the members of the mulleri subgroup of species and indicated the presence or absence of a reaction in the thirty-five species which had been investigated up to that time. It is the purpose of the present paper to present the results of a more thorough study of the insemination reaction in homogamic matings in the seventy~eight species available for study. Descriptions of the appear­ance of the vaginae following insemination will be given for the various species, including selected plates of photomicrographs of typical prepara­tions. Finally, possible functions of the reaction in intraspecific matings will be discussed. *Predoctoral Fellow of the National Research Council. Studies in the Genetics of Drosophila MATERIALS AND METHODS Most of the species used in this investigation were maintained as stocks in this laboratory. The writer wishes to express his appreciation to the following persons for additional stocks which they sent to us for use in this and other experiments: to Dr. A.H. Sturtevant, for stocks of D. dun­cani, ritae, cordata, elliptica, emarginata, rectangularis, micromelanica, and nigromelanica; to Dr. Th. Dobzhansky, for a stock of D. persimilis; and to Dr. H. T. Spieth, for stocks of D. sucinea and fumipennis. To determine the character of the insemination reaction, females were dissected and examined at varying intervals after copulation. The ideal has been to place a single virgin female in a vial with about five males, all of known age, and observe mating so that a timed series of dissections could be. secured. This procedure was relatively easy with many species once the proper age for mating had been determined. . In certain instances where it seemed too difficult to obtain matings in the manner outlined above, the stock culture bottles were observed from time to time in an effort to detect a pair in copula. When such a pair was found it was isolated by inserting a test tube into the culture bottle, thus imprisoning the pair. After completion of mating the flies were removed by inverting the apparatus and shaking the mated pair down into the tube. This method had an obvious disadvantage in that one could not know whether the female so obtained was virgin prior to mating. A variation of this technique consisted of placing virgin flies of assorted ages together in a bottle and removing mated pairs by the test tube method. Finally, we have been unable to devise any method of obtaining matings with a few species. In order to secure any sort of result with these forms we have been forced to dissect females from the stock culture tubes at various times in the hope of finding one inseminated recently enough to afford some indication of its reaction tendencies. At the risk of stating the obv~ous, it might be pointed out that positive results obtained in this way are fairly reliable, while negative results are inconclusive. The techniques employed in dissecting and photographing the repro­ ductive organs have been discussed in detail in the preceding article by Dr. Patterson. The plates accompanying this paper, numbered consecu­ tively in the two papers, will be discussed in the following section dealing with the individual species. DESCRIPTIONS OF THE INSEMINATION REACTIONS In the following section the nature of the insemination reaction Will be given for two species of the genus Chymomyza, one species of Scap­tomyza, and sixty-seven species of Drosophila. The eight members of the mulleri subgroup of species, namely, D. mulleri, aldrichi, arizonensis, mojavensis, buzzatii, hamatofila, ritae, and peninsularis, have been dis­cussed by Dr. Patterson in the preceding article. The various species will be treated in the order outlined by Sturtevant (1942). The University of Texas Publication The age of the flies when mated and the approximate average length of copulation will be given for each species, as well as descriptions of the female reproductive tracts at various intervals after copulation. Unless otherwise stated the time of dissection is recorded as the elapsed time from the completion of mating. Discussion of the .type of reaction concerned in each· case will be deferred to the following section of the paper. GENUS CHYMOMYZA Czerny Chymomyza amoena Loew. Matings were secured with flies from 1 to 5 days old. The average length of copulation was about 14 minutes. Dissection at 30 minutes revealed a large sperm mass in the cavity of the vagina which was not enlarged except at the postero-ventral angle, the latter swollen in the form of a small pouch. The ventra~ receptacle contained a moderate number of sperm while the spermathecae were so filled as to appear solid. At one hour the mass in the pouch was smaller and surrounded by motile sperm while the main cavity of the vagina was empty. At one and one-half hours there still remained a moderate sized mass in the pouch and there appeared to be a small mass stringing posteriorly to the ovipositor. This is indicative of the manner in which the females are believed to expel the excess sperm and fluids. In this specimen a small number of motile sperm was still visible around the mass remain­ing in the pouch. Chymomyza procnemis Williston. Matings were secured with flies 6 days old. The average length of copulation was about 28 minutes. The vagina of a female dissected. 15 minutes after mating appeared normal in all respects, being free of semen and sperm. The ventral receptacle contained a moderate number of sperm and the spermathecae were quite dense with them. A female dissected at 30 minutes extruded a small "wad" of non-motile sperm from the ovipositor upon dissection. This specimen was similar in all respects to the previous one. GENUS SCAPTOMYZA Hardy Scaptomyza graminum Fallen. One mating was secured with flies 2 and 3 days old. The length of copulation was one minute and 55 seconds. A dissection of the female 30 minutes after mating revealed motile sperm in the ventral receptacle but none in the vagina which was of normal size. The spermathecae were torn off and lost in dissection. Studies in the Genetics of Drosophila GENUS DROSOPHILA Fallen Subgenus Hirtodrosophila Duda. Drosophila duncani Sturtevant. Matings were secured with flies 8, 9, 10, and 13 days old. The average length of copulation was about 15 minutes. The earliest dissection per­formed was of a female in copula, exact duration unknown: The vagina was full of semen and was already beginning to exhibit a posterior pouch. The ventral receptacle and spermathecae contained many sperm. In a dissection at 20 minutes the vagina appeared still further enlarged and the pouch was more pronounced. The entire lumen of the vagina was filled with the mass. The typical appearance of a specimen dissected at 30 minutes is shown in Fig. 53 (Pl. XIV). The reaction mass was rather soft and was easily forced through the broken end of the oviduct. Specimens dissected at one and one-half hours after mating showed little change from the earlier condition. The mass, however, was more firm and was not expelled under pressure. The latest dissection observed was of a female from the stock culture. In this specimen the vagina and pouch were much smaller and were only about half filled with the reaction mass, revealing a moderate number of motile sperm remaining in the cleared portion of the lumen. Motile sperm were observed in the stalk and lumen of the parovaria in this female-a situation not normally oc­curring in members of the genus. Subgenus Pholadoris Sturtevant. Drosophila victoria Sturtevant. Matings were secured with flies 6, 7, and 8 days, old. The average length of copulation was about 43 seconds. The earliest specimen examined was . dissected 3 minutes after copulation. A large part of the semen was expelled from the ovipositor upon dissection. The vagina contained a large reaction mass, granular in appearance and containing many motile sperm. No sperm were visible, however, in the receptacles. The vagina of an individual dissected at 30 minutes appeared quite enlarged and was filled with a dense, opaque mass containing motile sperm. No sperm were seen in the spermathecae at this time although the ventral rece:IT-· tacle appeared to be :filled with these cells. ~n a specimen dissected at one hour the sperm around the edges of the mass were still motile while those within the mass seemed to have ceased all activity. Two specimens dissected 3 hours after mating revealed but little change. The mass was a little smaller, leaving a cleared area in the dorsal region. Motile sperm were still evident in this portion. In one specimen the spermathecae contained motile sperm while these receptacles were still empty in the other. The individual illustrated in Fig. 55 (Pl. XIV) was dissected at about 5 hours after mating. By this time the mass had developed a rather definite border and was clearly outlined as a dense, opaque mass. The University of Texas Publication Motile sperm were still present, principally in a pocket dorsal to the mass. Dissections at 6 and 7 hours showed no appreciable change but at 8 hours the mass was smaller · and was beginning to lose its opaque quality, although still rather sharply defined. At 10 hours the size of the PLATE XIV.. Vag~nae f1:om homogamic matings of four species of Drosophila. Fig. 53, D_. dunci;ini at th1~ty mmutes. J'.'ig. 54. D. victoria at five hours. Fig. 55, D. sucin3a at thirty mmutes. Fig. 56, D. fumipennis at one hour. Studies in the Genetics of Drosophila 83 . reaction mass seemed unchanged but the loss of opacity was becoming pronounced. By 12 hours this clearing had proceeded still further so that the mass was becoming crystalline in appearance. In this specimen some of the reaction material seemed to be passing from the main body of the mass tci the ovipositor. Motile sperm were still present in the vaginal cavity dorsal to the mass. The latest dissection performed was at 15 hours and 15 minutes after mating. The reaction mass was practically clear by this time but was still in evidence due to its discrete outline. Neither motile nor non-motile sperm were visible in the vaginal cavity at this time. Subgenus Dorsilopha Sturtevant. Drosophila busckii Coquillet. Matings were secured with flies 1 and 2 days old. The average length of copulation was about 3 minutes. In a specimen dissected at 20 minutes both the ventral receptacle and spermathecae were teeming with sperm. The vagina was filled with an emulsion-like mass containing many motile sperm. There was a slight enlargement of this organ when compared with that of a virgin female. Dissections at 30 minutes and at one hour showed no obvious change, the vagina in both instances containing a fair amount of the emulsion-like material with many motile sperm. In a specimen dissected at 2 hours after mating the mass appeared to have consolidated into an irregular, indefinitely defined mass in the central area of the vagina, leaving a large cleared area around its edges in which motile sperm still persisted. Subgenus Sophophora Sturtevant. I. saltans group Drosophila sturtevanti Duda. A pair of flies observed copulating in the stock bottle was isolated and the female dissected 30 minutes after the completion of mating. The vagina appeared normal in size and did not contain any sperm or semen. The ventral receptacle was teeming with sperm. Drosophila rectangularis Sturtevant. A pair of flies observed mating in the stock bottle was isolated and the female dissected 30 minutes after the completion of copulation. The vagina was not enlarged but was filled with an emulsion-like mass con­taining many sperm, most of which appeared non-motile. The ventral receptacle and spermathecae contained a moderate number of sperm. Pressure on the cover slip forced most of the sperm mass out through the broken end of the median oviduct revealing a large number of sperma­tozoa which were quite active in the saline solution. Attempted matings with flies 1'2, 13, and 14 days old were unsuccessful. The University of Texas Publication Drosophila prosaltans Duda. Matings were secured with flies 5, 6, and 9 days old. The average length of copulation was 18 minutes. Two specimens were dissected one hour after the c9mpletion of mating. The vagina appeared normal in size and was not dense and opaque although it was filled with an emulsion-like material. No sperm were observed in the vagina of either specimen, but in both instances a large mass of motile sperm was observed moving through the median oviduct and emerging into the saline mounting me­dium. The ventral receptacles and spermathecae were filled with sperm. Drosophila cordata Sturtevant. A pair of flies observed copulating in the stock bottle was isolated and the female dissected 30 minutes after the completion of mating. The vagina appeared swollen and rounded, not particularly elongate, and was filled with an emulsion-like mass. Near the anterior end of the vaginal cavity was a dense brown mass with motile sperm around its edges. The ventral receptacle and spermathecae were both filled with active spermatozoa. Drosophila elliptica Sturtevant. A pair of flies observed mating in the stock bottle was isolated and the female dissected 30 minutes later. The vagina was somewhat enlarged and elongated, the entire cavity filled with a dense, opaque, granular mass. A large pocket of motile sperm was visible dorsally near the base of the oviduct. The ventral receptacle was teeming with sperm but none were visible in either spermatheca. Drosophila emarginata Sturtevant. Attempted matings with flies 10 to 14 days old were unsuccessful. A pair of flies observed in copula in the stock bottle was isolated and the female dissected 30 minutes after the completion of mating. The vagina was not swollen but was rather elongate and contained two rather dense masses separated by emulsion-like material. Each of these masses was surrounded by small numbers of sperm. It is likely that copulation was interrupted by the isolation procedure so that insemination was disturbed midway ill' the process. This would account for the presence of two masses in the vaginal cavity. The ventral receptacle and spermathecae contained a moderate number of sperm;· 2. willistoni group Drosophila willistoni Sturtevant. A pair of flies less than 24 hours old was observed in copula. The female was isolated and dissected 30 minutes after copulation. The vagina did not appear to be enlarged and, although the sperm mass filled about half the vaginal cavity, it did not appear dense or opaque. The sperm mass was almost clear, luwing a slightly granular appearance and contained Studies in the Genetics of Drosophila some motile sperm. No spermatozoa were observed in the ventral recep­tacle or spermathecae. Drosophila equinoxialis Dobzhansky, Attempted matings with flies one day old were unsuccessful. These flies were then left together overnight and the females dissected early the following morning. Two inseminations were secured. In both specimens the ventral receptacle was teeming with sperm while the spermathecae contained only a few. The vagina was filled with semen, partly con­solidated into a granular, semi-formed mass containing a few sperm. The remainder of the lumen contained relatively clear fluid with many motile sperm. The vaginae of these specimens were not noticeably enlarged. Drosophila nebulosa Sturtevant. Matings were secured with flies 7 and 8 days old. The average length of copulation wa,s about one minute and 30 seconds. In a specimen dis­sected 15 minutes after copulation the ventral receptacle and spermathecae were full of sperm while the vagina contained no sperm or semen. In a specimen dissected at 30 minutes, however, a mass of motile sperm was .observed in the median oviduct and motile sperm were emerging from the broken end of this organ. The vagina was likewise free of semen in a specimen dissected at one hour, but a specimen dissected at one and one-half hours revealed a few sperm along the dorsal and ventral edges of the vagina as well as a pocket of motile sperm in the oviduct. A speci­men dissected at two hours appeared normal but had an egg in the vaginal cavity. In none of these individuals did the vagina seem to be enlarged. D~osophila sucinea Patterson and Mainland. One mating was observed in a bottle of previously unmated flies of mixed ages. The female was isolated and dissected at 30 minutes. The vagina was not noticeably enlarged but contained a moderate amount of granular material in the ventral half. This material did not appear dense or opaque and contained no visible spermatozoa. Both the spermathecae and ventral receptacle were teeming with sperm. This individual is illustrated in Fig. 55 (Pl. XIV). Drosophila fumipennis Duda. Two matings were observed in the stock including one complete copu­lat.ion which lasted 5 minutes and 30 seconds. One female was dissected with the ovaries attached 30 minutes after mating. The vagina was slightly enlarged and was completely filled with the sperm mass, part of which had moved into the median oviduct. A number of sperm were seen at the base of the eggs in the lateral oviduct. The ventral receptacle was filled with sperm while only one of the spermathecae contained sperm. The second female was dissected at one hour. The vagina was a little enlarged and was filled with a rather clear, granular material, motile The University of Texas Publication sperm being seen only near the anterior end at the base of the oviduct, within which a small mass of sperm was visible. The ventral receptacle and both spermathecae were teeming with motile sperm. This specimen is shown in Fig. 56 (Pl. XIV). In this and the previous figure (Fig. 55, D. sucinea) an additional struc­ture has been included in the dissections. This organ, whose function is completely unknown at present, has been observed only in members of the willistoni species group. It is typically a clear-walled, bladder-like sac composed of two lobes and a narrowing neck which seems to be attached to the base of the ovipositor. Since it invariably remains in position when the intestine is removed from the specimen it must have no connection with the digestive tract, and, conversely, since it remains fastened to the base of the ovipositor it is quite possible that it is related in some way to the reproductive tract. 3. melarwgaster group Drosophila melanogaster Meigen. Matings were secured with flies of various ages. The average length of copulation was about 10 minutes. The earliest dissection performed was 15 minutes after the completion of mating. The vagina was not enlarged but was filled with a sperm mass, large numbers of motile sperm passing through the oviduct and emerging from the broken end of this organ. The ventral receptacle and spermathecae contained many sperm; In a specimen dissected at 30 minutes the sperm remaining in the vagina were largely non-motile. At one hour the vagina was free of sperm and semen in one specimen, while at one and one-half hours another specimen still retained a moderate number of sperm in the vaginal cavity. The vagina of a specimen dissected at 3 hours appeared normal in all respects. A specimen dissected 6 and one-half hours after mating contained an egg in the vagina. Since there were no eggs on the food of the vial in which she had been kept, this would represent the first egg laid by this female after mating. A small droplet on the surface of the food was mounted in saline and observed under the microscope. It was found to be a mass of non-motile sperm, thus affording direct evidence that these elements are expelled by the female in some homogamic matings. Drosophila simulans Sturtevant. A pair of flies observed in copula in the stock was isolated and the female dissected 15 minutes after the completion of mating. The vagina was not enlarged and was free of semen and sperm, although the ventral receptacle and spermathecae were filled with these elements. It might be of interest to record here the results of dissections of two melanogaster females inseminated by simulans males. Neither specimen showed an en­larged vagina or a reaction mass present after 24 hours exposure. Studies in the Genetics of Drosophila Drosophila ananassae Doleschall. Matings were secured with flies 2, 3, and 4 days old. The average length of copulation was about 4 minutes. The only accurately timed dissection was made 30 minutes after the completion of mating. All the semen appeared to be in the cavity of the vagina, no sperm being visible as yet in the ventral receptacle or spermathecae. The sperm in the vagina were mostly motile. Near the posterior end, in the region of the ovipositor, a part of the semen was formed into an irregular, crystalline-appearing mass surrounded by granular material. This posterior "plug" was later seen in several other specimens. It seems likely that the first portion of the ejaculate contains the motile sperm in its carrier fluid while the por­tion ejaculated near the conclusion of copulation is sperm free and soon forms a gel-like mass which forces retention of the semen for some time, thus affording the spermatozoa ample time to reach the receptacles. This plug is probably softened in time by the secretions of the vagina , and is expelled, along with the excess semen, or it is possible that muscle contractions eventually force the plug out through the posterior 0rifice and expulsion of the semen follows. 4. obscura group Drosophila pseudoobscura Frolova. Matings were secured with flies 2, 3, 4, 5, and 6 days old. The average length of copulation was about 5 minutes. Mayr (1946) gives the median duration as 6 minutes and 15 seconds. Several specimens dissected 15 minutes after copulation did not have enlarged vaginae, but were filled with semen containing motile sperm. No sperm were visible, however, in the ventral receptacles or spermathecae. The pressure of the cover glass forced most of the sperm mass into the oviduct. Additional speci­mens were dissected at 30 minutes, one, one and one-half, and two hours. In each case a mass of sperm was forced into the oviduct upon dissection or shortly thereafter. The spermatozoa seem to move very slowly into the receptacles in this species. Only 10 to 15 sperm were visible in the ventral receptacle of one individual dissected at one ho'ur, and about two dozen were seen in this organ in the one and one-half hour specimen. The two-hour specimen revealed a large number of sperm in the ventral receptacle but only a moderate number in the spermathecae where they were observed mainly in the stalks. Drosophila persimilis Dobzhansky and Epling. A pair of flies observed mating in the stock bottle was isolated and the female dissected 30 minutes later. The vagina was very slightly enlarged and was filled with granular material which was neither dense nor opaque. No sperm were visible in the mass or in the receptacles. Pressure was placed on the cover glass to expel the contents of the vagina revealing a large number of motile sperm in the saline mounting fluid. It seems likely that at this time the secretions from the female tract necessary to The University of Texas Publication stimulate the spermatozoa into motility were insufficient. This theory would similarly explain the long time required for the sperm to enter the receptacles in the closely related D. pseudoobscura. Drosophila afjinis Sturtevant. Matings were secured with previously unmated flies of various ages mixed together in a stock bottle. The vagina of a specimen dissected at 30 minutes was torn in dissection allowing the contents to exude into the saline solution. A large number of motile sperm were observed in an extremely small quantity of granular material. Very few sperm were visible in the ventral receptacle, none in the spermathecae. In a second specimen, dissected at one hour, the vagina was clear and normal except for a small pocket of motile sperm in a slight pouch. No sperm were visible as yet in the spermathecae. Drosophila azteca Sturtevant and Dobzhansky. One mating was secured with previously unmated flies of mixed ages, and the female was dissected 30 minutes later. The vagina was not en­larged but was filled with a granular material. A large sperm mass was expelled through the oviduct immediately following dissection. No sperm were seen in the vaginal cavity although both the ventral receptacle and spermathecae were teeming with spermatozoa. Drosophila tolteca Patterson and Mainland. One mating was obtained by mixing previously unmated flies of various ages in a culture bottle. The dissection was made at one hour. The vagina was not noticeably enlarged but was filled with the sperm mass. The contents were granular and relatively clear and contained motile sperm throughout. Part of this mass had migrated into the oviduct. The ventral receptacle contained a moderate number of sperm while none were visible in the spermathecae. Subgenus Drosophila. Fallen. 1. quin.aria group Drosophila transversa Fallen. Matings were secured with flies 5 and 7 days old. The average length of copulation was about 9 minutes and 30 seconds. The earliest dissection was performed 15 minutes after mating. The vagina was about three times normal size and was quite dense and opaque. Motile sperm were observed around the edges of the reaction mass but only non-motile sperm were visible within the mass. Both the ventral receptacle and spermathecae were teeming with sperm. A specimen dissected 30 minutes after mating was similar in most respects to the previous one, and is illustrated in Fig. 58 (Pl. XV) . Drosophila munda Spencer. A series of matings was obtained with stock flies by transferring them to fresh food and placing them before a bright light. Copulation times Studies in the Genetics of Drosophila were recorded for two of these matings, the average time being about 3 minutes and 40 seconds. Several specimens dissected 30 minutes after mating were alike in having a very large vagina filled with the dense, opaque mass. A few motile sperm were visible around the edges of the mass. The ventral receptacles contained a large number of sperm but PLATE XV. Vaginae from homogamic matings of three species of Drosophila. Fig. 57, D. munda at thirty minutes. Fig. 58, D. transversa at thirty minutes. Fig. 59, D. subquinaria at one hour. The University of Texas Publication only a moderate number were seen in the spermathecae. A specimen dis­sected 30 minutes after mating is shown in Fig. 57 (Pl. XV). A pecu­liarity of this species is the tendency for the tissues of the vagina, just anterior to the ovipositor, to stretch when the dissection is being made. This is clearly shown in the illustration. A specimen dissected at one hour revealed that the contents of the vagina had cleared somewhat, having lost much of the opacity present in earlier . specimens. By this time the spermathecae seemed quite solid with sperm. Drosophila quinaria Loew. Two pairs of flies were observed mating and were isolated. The com­plete copulation time, obtained for one of them, was 7 minutes and 30 seconds. The first specimen was dissected 30 minutes after mating. The vagina was greatly enlarged and filled with a very dense, opaque reaction mass. A few motile sperm were visible around the edges of the mass, while the ventral receptacle and spermathecae contained many sperm. The second specimen was dissected one hour after mating. The vagina appeared about normal in size and contained but a small amount of the granular mass near the base of the ovipositor. A few motile sperm were seen in a pocket along the edge. Both the ventral receptacle and sperma­thecae were teeming with sperm. The position of the remainder of the mass near the ovipositor suggests that females of this species expel the semen and excess sperm as has been previously suggested for several other species. Drosophila subquinaria Spencer. Matings were secured with flies 7 and 8 days old. The average length of copulation was about 8 minutes and 45 seconds. The vagina of a speci­men dissected 30 minutes after copulation was extremely enlarged and filled with a very dense, opaque reaction mass. Motile sperm were observed around the edges of the mass and many non-motile sperm could be seen within the mass. The ventral receptacle was filled with sperm while the spermathecae contained only a moderate number of these cells. In another specimen dissected at 30 minutes a part of the mass was moving into the oviduct and motile sperm were seen emerging from the broken end of this organ. A specimen dissected one hour after mating burst when the cover slip was added. The vaginal contents were granular and quite dense. No motile sperm were seen. The last dissection was performed one and one-half hours after mating. The following notes, made at the time, adequately describe this preparation : "Vagina exceed­ingly enlarged, perfectly round and too dense and opaque to see through. The ventral receptacle and spermathecae are teeming with sperm and many sperm are emerging from the oviduct, the latter also somewhat opaque. The most extreme reaction yet seen." A specimen dissected at one hour is shown in Fig. 59 (Pl. XV). Studies in the Genetics of Drosophila Drosophila suboccidentalis Spencer. Matings were secured with flies 7, 8, and 9 days old. The average length of copulation was about 10 minutes. The first specimen, dissected at 30 minutes, was peculiar in several respects. Copulation lasted only two minutes and the female appeared quite agitated during the entire time. The vagina was somewhat enlarged and rather opaque. Its surface had an unusual appearance, being extremely rough and knobby looking, and possessed a pouch at the distal ventral end. No sperm were visible in the preparation. It is probable that this was a malformed individual and that complete copulation was prevented by its abnormalities. A second specimen was dissected one hour after mating. The vagina was quite large, rounded, and filled with a dense, opaque mass which slowly oozed through the ovipositor while under observation. Although only non-motile sperm could be detected within the mass in the vaginal cavity, many motile sperm were seen in the saline mounting fluid as this mass was extruded. Both the ventral receptacle and spermathecae were teeming with sperm. A last specimen, dissected 3 hours after mating, appeared normal in all respects. The only sperm seen were in the receptacles. Drosophila innubila Spencer. Matings were secured with flies 8 and 9 days old. The only complete copulation recorded lasted 3 minutes. The vagina of a female dissected 30 minutes after mating was quite large, dense, and opaque. Motile sperm were visible along the edges of the mass while the sperm seen within the mass appeared completely non-motile. Both the ventral receptacle and spermathecae contained sperm. The reaction mass of a specimen dis­sected one hour after mating was so dense that one could not see through it. A part of the mass had moved into the median oviduct. Drosophila subpalustris Spencer. A pair of flies observed copulating in the stock bottle was isolated and the female dissected 30 minutes after the completion of mating. The vagina was greatly enlarged and filled with the dense, granular mass. Sperm were visible in the ventral receptacle and spermathecae. 2. guttifera group Drosophila guttifera Walker. Matings were secured with flies 4 and 5 ®ys old. The average length of copulation was about 9 minutes. In a specimen dissected at 30 minutes the vagina was slightly enlarged and contained some granular material but was not noticeably opaque. Motile sperm were present in this organ as well as in the ventral receptacle and spermathecae. In one specimen dissected at one hour the vagina was not particularly enlarged or opaque. A few motile sperm were present as scattered bunches in the cavity. In another specimen, also dissected at one hour, the vagina was definitely enlarged and showed a slight tendency toward pouch formation. This The University of Texas Publication preparation is shown in Fig. 60 (Pl. XVI). The granular nature of the reaction mass is easily seen. The vagina of a specimen dissected at one and one-half hours was noticeably smaller but still enlarged. The opaque mass still filled the lumen and contained non-motile sperm. 3. pinicola group Drosophila pinicola Sturtevant. This species, considered by Sturtevant (1942) as the most primitive species in the genus, was unavailable for study. This is indeed unfortunate since it would have been most interesting, from an evolutionary viewpoint, to compare the effects of insemination in this species with the results observed in the more highly evolved forms. 4. virilis group Drosophila virilis Sturtevant. Matings were secured with flies of various ages, the youngest being 7 days old and the oldest·31 days old. The average length of copulation was about 2 minutes and 55 seconds. It is very difficult to describe the sequence of events following insemination in this species because the contents of the vagina escaped through the ovipositor or the broken end of the oviduct in nearly every dissection performed. Dissections have been made at 8, 20, 25, 30, 35, 40, and 50 minutes after mating and in every case the vagina appeared to have been enlarged and filled with an emulsion-like mass containing either motile or non-motile sperm. By 20 minutes both the ventral receptacle and spermathecae contained large numbers of sperm. In some of the individuals the expulsion of the mass · took place when the flies were etherized. -In a few cases the mass extruded from the ovipositor retained its shape, indicating that it was not com­pletely fluid. Because of the indeterminate nature of the reaction in this species and since matings were very easy to secure, it was decided to try remating in order to determine if later inseminations called forth a reaction more severe than that of the first mating. Accordingly females were dissected 30 minutes after a second mating and 30 minutes after a fifth mating. The earliest time at which remating could be accomplished was 5 hours and 15 minutes after the first. Rematings at 24-hour intervals were, in general, not difficult to secure. The results of the dissections of females who had mated a second ti~e were definitely inconclusive. No difference in the severity of the reaction could be detected. Similarly, dissections after the fifth mating in as many days resulted in a reaction not noticeably more severe than the first. Finally, an attempt was made to determine if the age of the flies used would influence the severity of the reaction. When 31-day-old females were placed with 4-day-old males no matings were obtained. When 4-day­old females were placed with 25-day-old males some courting ensued but; once again, no matings wer~ obtained. One mating was secured when a Studies in the Genetics of Drosophila female 7 days old was placed with males 25 days old. Dissection, how­ever, revealed that the female had not been inseminated. Two matings were secured with 31-day-old females and 7-day-old males. The first of these was dissected 30 minutes after mating. Early in dissection the -~-----____J PLATE XVI. Vaginae from homogamic matings of four species of Drosophila. Fig. 60, D. giittifera at one hour. Fig. 61, D. americana at thirty minutes. Fig. 62, D. montana at thirty minutes. Fig. 63, D. lacicola, at thirty minutes. The University of Texas Publication sperm mass was expelled through the oviduct and, although irregular in shape, retained its form when in the saline solution. The second mating was dissected one hour later. The notes recorded at the time are as follows: "Vagina very large, rounded, and filled with an opaque mass but containing only a few sperm. The ventral receptacle and spermathecae have a moderate number of sperm and there are a few pockets of sperm in the vagina near their openings. Generally speaking, the reaction mass is too dense to see through." Although the evidence is slight, it is quite possible that the reaction in older females is more severe than in younger females. Drosophila americana Spencer. A pair of flies observed mating in the stock bottle was isolated and the female dissected 30 minutes after the completion of copulation. The vagina was enlarged about three times its normal size, and, as may be seen in Fig. 61 (Pl. XVI), showed a moderate-sized pouch. The entire cavity of the vagina was filled with a granular, dense, opaque mass with a few motile sperm visible along its edges. Part of the mass had moved into the oviduct and motile sperin were emerging from the broken end of this organ. The ventral receptacle and spermathecae contained many spermatozoa. Drosophila americana texana Patterson, Stone, and Griffen. Two pairs of flies observed mating in a stock bottle were isolated and the females were dissected at 30 minutes and at one hour, respectively. In the specimen dissected at 30 minutes the vagina was seen to be notice­ably enlarged while dissecting but before the operation was completed a large mass of sperm and granular material escaped through the ovi­duct, leaving the vagina about twice its normal size. It was still filled, however, with granular material but did not appear particularly dense. A small pocket of motile sperm was visible near the base of the ventral receptacle which was filled with sperm. Both the stalks and the bodies of the spermathecae contained sperm. As the first dissecting needle was placed in the thorax of the specimen dissected at one hour a large dense mass was extruded from the ovipositor. It retained its shape in the saline solution in which the dissection was being made. The vagina appeared only slightly enlarged and was quite clear except for a small mass of non-motile sperm near the ovipositor and a few motile sperm along the ventral side of the vaginal cavity. A few non-motile and a single motile sperm were seen around the mass which had been expelled. Drosophila montana Patterson and Wheeler. Two pairs of flies observed mating in the stock bottles were isolated and the females dissected at 30 minutes and at one hour after mating, respectively. The vagina of the specimen dissected at 30 minutes was about twice the normal size of this organ·but was probably larger since a part of the sperm mass was expelled through the oviduct. The vagina Studies in the Genetics of Drosophila was filled with a granular, dense, opaque material except near the oviduct where some motile sperm were visible. There appeared to be slight tend­ency for pouch formation although it was not as pronounced as in D. ameri­cana. Both the ventral receptacle and spermathecae were teeming with sperm. This specimen is illustrated in Fig. 62 (Pl. XVI). The second specimen was dissected at one hour after mating and showed no obvious changes from the earlier dissection. Drosophila lacicola Patterson. Matings were secured with previously unmated flies of various ages placed together in a stock bottle. A specimen dissected 30 minutes after mating is shown in Fig. 63 (Pl. XVI). The vagina was slightly enlarged but showed no tendency for pouch formation. There was a moderate amount of granular material in the cavity of the vagina. Many non­motile sperm were visible in the mass while motile sperm were seen along the ventral edge. A fair number of sperm were present in the ventral receptacle but none were seen in the spermathecae. A specimen dissected at one hour had practically returned to normal. The vagina contained a small amount of granular material and many non-motile sperm. No sperm were seen in either the ventral receptacle or the spermathecae of this individual. 5. testacea group Drosophila putrida Sturtevant. Matings were secured with flies 4 days old. The average length of copulation was about 29 minutes. The vagina of a specimen dissected at 30 minutes was greatly enlarged, being at least three times normal size, and was completely filled with a very dense. opaque reaction mass. No motile sperm were visible in the mass, possibly because it was too dense to see through clearly. The ventral receptacle was teeming with sperm and the spermathecae contained a moderate number. While under ob­servation the pressure of the cover glass forced most of the mass out through the oviduct liberatin~ a small number of motile snerm. The appearance of the extruded mass was peculiar. This material retained its linear order as it emerged, forming a coiled and twisted ribbon in the saline solution. The specimen illustrated in Fig. 64 (Pl. XVII) was dissected at one hour after mating. The exceedingly dense, balloon-like vagina, filled with opaque material, was likewise typical of specimens dissected 2 and 3 hours after mating. A dissection at 4 hours revealed a normal vagina in most respects, no mass or sperm being visible in this organ. 6. tripunctata group Drosophila tripunctata Loew. Matings were secured with flies 6 to 11 days old. The average length of copulation was about 48 minutes, the longest successful copulation recorded being 82 minutes. The earliest dissection performed was imme­ The University of Texas Publication diately following a copulation of 65 minutes. The vagina appeared nor­mal in all respects, containing no sperm or mass. Both the ventral receptacle and spermathecae contained large numbers of spermatozoa. Additional dissections were made at the following intervals after mating: PLATE xyn. Vaginae from homogamic matings of four species of Drosophila. Fig. 64, D. putrida at one hour. Fig. 65, D. tripunctata at thirty minutes. Fig. 66, D. sub­funebris at one hour. Fig. 67, D. macrosvina at thirty minutes. Studies in the Genetics of Drosophila 5, 10, 15, 20, 25, 30, and 45 minutes, and at one, one and one-half, and 4 hours. In none of these specimens was there any sign of vaginal enlarge­ment. The appearance of a specimen dissected at 30 minutes after mating is shown in Fig. 65 (Pl. XVII). Eighteen eggs were observed in the vial in which the 4-hour specimen had been kept. Only three of them failed to hatch. This is of particular interest since attempts to get rematings with this species were unsuccess­ful until at least six days after the first mating. It is also of interest to compare the copulation times of these two matings. The length of the first mating was 47 minutes, that of the second, 82 minutes. This is ap­parently the reverse of the situation found by Mayr (1946) for D. pseudo­obscura, in_which later matings were usually shorter than the first. Drosophila cro.cina Patterson and Mainland. One mating was secured with flies 9 days old. The vagina of the female, dissected 30 minutes after mating, appeared normal in all respects. Both the ventral receptacle and spermathecae were teeming with spermatozoa. 7. funebris group Drosophila funebris (Fabricius). Matings were secured with flies 6, 7, 8, and 10 days old. The average length of copulation was 16 minutes and 33 seconds. Several dissections were made 30 minutes after mating. The vagina was two to three times its normal size, rounded, and filled with a granular, opaque mass. A few motile sperm were visible around the edges of the mass and a large pocket of them was invariably observed in the median oviduct. Both the ventral receptacle and spermathecae were teeming with sperm by this time. It might be of interest to record here the observation of a successful copulation of a pair of flies of which the male had been thoroughly ether­ized less than two hours previously. Drosophila subfunebris Stalker and Spencer. Matings were secured with flies 9 days old. The average length of copu­lation was about 23 minutes and 30 seconds. The vagina of a specimen dissected 30 minutes after mating was definitely enlarged but showed no tendency for pouch formation. The entire cavity was filled with a large opaque, emulsion-like mass. No sperm were visible, however, until the contents had been expelled into the saline solution by pressure on the cover slip. A few active sperm were then seen in the mounting fluid. Specimens dissected at one hour after mating showed no apparent change from the previous one. Such a dissection is shown in Fig. 66 (Pl. XVII). Drosophila macrospina Stalker and Spencer. Matings were secured with flies I '.l days old. The average length of copulation was about 35 minutes. A specimen dissr~cted at 20 minutes appeared to have had an enlarged vagina but nearly all the mass passed out through the oviduct upon dissection. One specimen dissected at 30 The University of Texas Publication minutes had expelled the entire contents of the vagina before observation. Another dissection made at 30 minutes is shown in Fig. 67 (Pl. XVII). · The vagina was definitely enlarged and filled with a dense reaction mass. A few motile sperm were visible along the edges of the niass. A last specimen, dissected at 40 minutes, had cleared and the vagina appeared normal. 8. repleta group Drosophila repleta Wollaston. Several matings were observed in a stock bottle in which had been mixed previously unmated flies of various ages. The males of five of the seven copulations observed failed to inseminate the females. The appearance of a specimen dissected at 30 minutes is shown in Fig. 68 (Pl. XVIII). The vagina was but slightly larger than that of a virgin and contained a moderate amount of granular material. This mass was relatively clear, not dense and opaque, and contained motile sperm in clumps scattered throughout the vaginal cavity. The ventral receptacle and spermathecae contained many spermatozoa. A specimen dissected one hour after mating was similar in most respects to the former, the only. observable change being the presence of greater numbers of sperm in the receptacles. Drosophila neorepleta Patterson and Wheeler. One mating was observed in a stock bottle of 12-day-old flies. Additional dissections were made of females selected ·at random from this bottle. A specimen dissected 30 minutes after mating is shown in Fig. 69 (Pi. XVIII). The vagina was somewhat enlarged and showed a definite tendency for pouch formation. The vaginal cavity was nearly filled with a granular mass in which many sperm were visible. Part of the sperm mass moved into the oviduct while under observation. Both the ventral receptacle and spermathecae were teeming with sperm. Several dissections of the fe­males selected at random appeared to have been made at a time greater than 30 minutes after mating. Most of the vaginal sperm mass was con­centrated in the pouch' where their motility was easily observed. Drosophila species (= limensis Pavan in press). Females from the stock culture were dissected at random since no observed matings were secured. Several of these specimens appeared to have been inseminated fairly recently. The vaginae were slightly enlarged and filled with a relatively clear granular material. Many sperm "'ere present in the vaginal cavity, those along the ventral side being highly motile. Drosophila canapalpa Patterson and Mainland. Two pairs of flies were observed mating in the stock bottle and isolated. A specimen dissected 30 minutes after mating had a slightly enlarged vagina with no evidence of pouch formation. The sperm mass, which . Studi~s in the Genetics of Drosophila filled the entire vaginal cavity and part of which had moved into the median oviduct, was granular and dense and contained many sperm which appeared· only slightly motile. The ventral receptacle and spermathecae contained a moderate number of sperm. The vagina of a specimen dis­seCted at one hour had returned to normal and was cleared, except for a small pocket of non-motile sperm near the ovipositor. Drosophila melanopalpa Patterson and Wheeler. Two matings were observed in the stock culture bottle. Additional dissections were made of females chosen at random from this bottle. A specimen dissected at 30 minutes did not appear to have been inseminated. The vagina of a specimen dissected at 40 minutes was of normal size and contained no visible mass of any kind. The only evidence of a recent insemination was the presence of motile sperm in the basal section of . .the ventral receptacle. There were many sperm in the mid-section of this organ .but none were visible in the spermathecae. The only informa­tion obtained from the females taken at random from the culture bottle was the observation that most inseminated females contained but a small number of visible sperm in their receptacles. Drosophila hydei Sturtevant. . Matings were secured with files 11 and 13 days old. The average length ~f copulation was about 4 minutes. The vagina of a specimen dissected 15 minutes after mating was about of normal size with a small opaque area at the anterior end. There were a few motile and many non-motile sperm in the vaginal cavity and both the ventral receptacle and sperma­thecae were teeming with·them. Dissections 30 minutes and one hour after mating revealed no signs of enlargement and no tendency for reac­ tion mass formation. Drosophiia mulleri Sturtevant. Drosophila aldrichi Patterson and Crow. Drosophila arizonensis Patterson and Wheeler. Drosophila mojavensis Patterson and Crow. Drosophila buzzatii Patterson and Wheeler: Drosophila hamatofila, Patterson and Wheeler. Drosophila, ritae Patterson and Wheeler. Drosophila peninsularis Patterson and Wheeler. The eight species listed above, all members of the mulleri subgroup of species, have been discussed in detail by Dr. Patterson in the preceding article of this publication. May it suffice here to point out that homogamic inseminations in the first five members of this group are followed almost immediately by a striking enlargement of the vagina, followed by the development of a very dense, opaque reaction mass which persists for several hours. Each of these species shows a definite pouch when enlarged. D. hamatofila and D. ritae, however, are characterized by an elongated swelling of the vagina, the latter similarly filled with a dense reaction mass. The last species, D. peninsularis, develOps the least extreme reac­ The University of Texas Publication tion mass of t:~1e group, the vagina exhibiting only a small amount of swelling and the mass never becoming as dense and opaque as in the other members of this subgroup. -' PLATE XVIII. Vaginae from homogamic matings of four species of Drosophila. Fig:·68, D. repleta at thirty minutes. Fig. 69, D. neorepleta at thirty minutes. Fig. 70, D. mercatorurn at one hour and thirty.minutes. Fig. 71, D. hexastigrna at one hour and fifteen minutes. -.. .. Studies in the Genetics of Drosophila Drosophila mercatorum Patterson and Wheel~r. Matings were secured with flies 5 days old. The average length of copu­lation was about 3 minutes. A specimen dissected 30 minutes after mating had a greatly enlarged vagina consisting mainly of an immense pouch, the'entire cavity being filled with an emulsion-like mass. No sperm were visible in the mass or in the spermathecae although the ventral receptacle was filled with them. Dissections at one, one and one-half, and two hours after mating revealed no observable change from the condition described above. A specimen dissected at one and one-half hours after mating is illustrated in Fig. 70 (Pl. XVIII). Drosophila mercatorum pararepleta Dobzhansky and Pavan. One mating was observed in the stock culture bottle. Additional females from this bottle were selected at random for dissection. The mated female was dissected 30 minutes after the completion of copulation. The vagina was quite enlarged with a rounded pouch, the entire cavity being filled with an emulsion-like material. The ventral receptacle was teeming with sperm but none were visible in the spermathecae. Several of the stock females which were dissected were quite similar to that described above while others were apparently very late stages in which the vagina had practically cleared. Drosophila meridiana Patterson and Wheeler. Matings were secured with flies 10 days old. The average length of copulation was about 2 minutes. The vagina of a specimen dissected 10 minutes after mating was moderately enlarged, granular, and slightly opaque, and exhibited a small pouch. A few motile sperm were seen in the pouch area. A part of the granular mass was observed in the oviduct. The ventral receptacle was teeming with sperm while only the stalks of the spermathecae contained sperm. A specimen dissected at 30 minutes was much like the previous one, but the entire contents of the vagina were expelled through the oviduct before observation under the microscope. The vagina of a last specimen, dissected at one hour, was practically cleared, no mass remaining in the vagina or pouch. Drosophila anceps Patterson and Mainland. Matings were secured with flies 7, 10, and 11 days old. The average length of copulation was about 5 minutes and 30 seconds. The appearance of a specimen dissecte~ at 30 minutes is shown in Fig. 73 (PL XIX). The vagina was somewhat enlarged with a slight pouch. The contents were fairly clear near the ovipositor but were darker and more dense in the pouch where a few motile and many non-motile sperm could be seen. A dissection at one hour revealed very little change. The vagina of a specimen dissected at 3 hours was much smaller, however, and the pouch was greatly reduced in size. The dark area of the mass was likewise smaller, the lighter material remaining as before. As a 4-hour dissection was being performed a semisolid mass was extruded from the ovipositor. On a slide in saline this mat.i;lrial was seen to consist largely on non-motile sperm with a few motile sperm around the edges. The size of the ·vagina of this specimen was normal when examined. Drosophila hexastigma Patterson and Mainland. · · Two matings were secured by mixing previously,unmated flies of various ages in a culture bottle. The length of copulation was 6 minutes and 30 seconds for one timed mating. Additional females were selected at ran­dom from the stock and dissected. In a specimen dissected · at one hour the sperm had not yet entered the ventral receptacle but were quite active along the ventral side of the vagina, the dorsal side being filled with a very dark reaction mass which extended into a rather elongate pouch. A specimen similar to this, dissected at one hour and 1'5 minutes after mating, is shown in Fig. 71 (Pl. XVIII). The vagina proper was slightly enlarged and contained a small amount of the reaction material but the greatly enlarged pouch was .filled with this exceedingly dense~ elongate mass. A small "wad" of this material was observed in the ovi­positor indicating that the female was in the act of expelling this _material. Drosophila gibberosa Patterson and Mainland. Matings were secured with flies from 10 to 14 d~ys old. The average length of copulation was about 8 minutes and 24 seconds. The vagina of a specimen dissected 30 minutes after mating was much enlarged, form­ing a· pouch, and the entire cavity was filled with an erlremely dense, opaque, granular mass teeming with motile sperm. A few sperm were visible in the basal part of the ventral receptacle am~, althoug~ one sperma~ theca contained no sperm, the other was about half filled with these cells. The vagina was still more enlarged at one hour, but at two hours it had begun clearing a little and was definitely smaller. . Th~ appearance of this dissection is shown in Fig. 72 (Pl. XIX). No motile sperm were seen in or around the mass but a moderate number were visible iri the ventral receptacle. Specimens dissected at 5 and at 23 hours after mating appeared normal. Several matings were secured in which 'the males failed · to inseminate. In one instance a copulation lasting 6 minutes wqs fol-. lowed one minute.later by a second copulation with another male lasting · 2 minutes. Dissection of the female revealed that she had not been in-. seminated, suggesting that the failure of mating was due somehow to the fem ale and not to the males concerned. 9. robusta group Drosophila robusta Sturtevant. Matings were secured with flies 20 and 21 days old, although, in. mass matings, insemination occurred with flies 11 and 13 days old. The average length of copulation was about one minute. The vagina of a specimen dissected 30 minutes after mating did not appear to be enlarged but con­tained a rather small granular mass in the lumen in which some non-'"motile sperm were visible along the edges. Both the ventral .receptacle and Studies in the Genetics of Drosophila spermathecae were teeming with sperm. Dissection at one hour revealed very little change, the granular mass being relatively clear. No sperm were visible in the cavity of the vagina. By one and one-half hours the vagina appeared to be entirely normal. I ·1 I I 13 14 PLATE XIX. Vaginae from homogamic matings of three species of Drosophila. Fig. 72, D. gibberosa at two hours. Fig. 73, D. anceps at thirty minutes. Fig. 74, D. carbonaria at thirty minutes. 10. melanica group Drosophila melanica Sturtevant. One mating was secured with flies 16 days old. Copulation lasted 3 minutes and 45 seconds. This specimen, dissected at 30 minutes, is shown in Fig. 75 (Pl. XX). The main body of the vagina was only slightly enlarged but a pronounced pouch was present. Before the photograph could be taken most of the mass in the pouch escaped through the oviduct and can be seen in the illustration around the broken end of this organ. The reaction mass was relatively clear except in the pouch where many sperm were visible. Both the ventral receptacle and spermathecae con­tained many sperm. Additional dissections of stock females chosen at random revealed various stages of clearing. Apparently the mass in the main cavity of the vagina clears first; this is followed by a decrease in density of the mass within the pouch, and eventually the loss of this mass. Drosophila melanica paramelanica Patterson. Dissections were made of females selected at random from the stock culture bottle. One specimen was observed with a reaction of about the same severity as that observed in D. 1nelanica described above. This indi­vidual is illustrated in Fig. 76 (Pl. XX). The vagina exhibited a pro­nounced pouch filled with a granular reaction mass which was not par­ticularly dense in appearance. There were many motile sperm around its edges. The mass extended into the lumen of the vagina where more motile sperm were observed. A small narrow band of this material was seen leading toward the ovipositor as if expulsion of the mass was in progress. Both the ventral receptacle and spermathecae contained many sperm. Drosophila nigromelanica Patterson and Wheeler. One mating was observed among previously unmated flies of various ages. Dissection was performed at 30 minutes. The vagina was mod­erately enlarged and rounded with a slight pouch at the base of the ventral receptacle. The entire vaginal cavity was filled with a rather clear granular mass, motile sperm being·visible only in a pocket in the pouch area. Many sperm were visible in the ventral receptacle and in the stalks of the spermathecae. · Drosophila micro1nelanica Patterson. A single copulation was observed in a vial containing two-day-old un­mated flies. The female was isolated and dissected at 30 minutes. The vagina was enlarged in the form of a pouch which was filled with a granular, clear reaction mass in which many sperm were observed. Al­though the ventral receptacle was teeming with sperm, none were visible in the spermathecae. This specimen is shown in Fig. 77 (Pl. XX). Dis­sections of two-and three-day-old flies which had not been separated since emergence, revealed that insemination occurs very early in the life of this species although the ovaries are still quite immature and these Studies in the Genetics of Drosophila females do not begin laying fertile eggs for some time. This condition, which has also been observed in several other species, has been described by Stalker (1942) for certain members of the virilis species complex as follows: " . . Females of some Drosophila species pass through two P LATE XX. Vaginae from homogamic matings of four species of Drosophila. Fig. 75, D. melanica at thirty minutes. Fig. 76, D. rnelanica paramelanica from stock. Fig. 77, D. micromelanica at thirty minutes. Fig. 78, D. polychaeta at thirty minutes. The University of Texas Publication distinct stages in their adult development. In the first stage they will mate, and can be inseminated, but are unable to lay any eggs due to the fact that their ovaries are undeveloped. In the second stage they can both mate and oviposit." 11. polychaeta group Drosophila polychaeta Patterson and Wheeler. The peculiar mating behavior of this species made it very difficult to know when copulation had occurred. The males seldom courted the females before mounting but usually walked up from behind and forcibly jabbed at.the female with the tip of the abdomen. Many times they managed to secure a grip on the wings of the female and remain in position for several seconds without actually accomplishing penetration. Six-and 7-day-old flies, for example, were set up in matings and watched intently for two hours, during which time numerous jabs at the females were made. When the females were later dissected none had been inseminated. A copulation lasting 35 seconds was observed in a vial of 13-day-old flies and the female was dissected 30 minutes later. Insemination had been successful but all the sperm were in the ventral receptacle, the vagina being free of sperm or semen. This individual is shown in Fig. 78 (Pl. XX). In another instance, a 7-day-old female was placed with several 6-day-old males. There were numerous jabs by the males and then an apparent copulation which lasted 20 seconds. This was followed by more jabs and then another copulation lasting about 15 seconds. The female was isolated and dissected one hour after the first copulation. Insemination had been accomplished but no sperm were visible in the vagina or spermathecae while many were observed in the ventral re­ceptacle. 12. carbonaria group Drosophila carbonaria Patterson and Wheeler. Matings were secured with flies 3 and -4 days old. The average length of copulation was about 26 minutes. Several pairs observed mating in the wild around exuding sap of the mesquite (Prosopis sp.) were cap­tured and the females dissected. A specimen dissected at 30 minutes after mating is illustrated in Fig. 74 (Pl. XIX). The reaction mass was con­fined almost entirely to the elongated pouch, the remainder of the vaginal cavity appearing free of sperm and semen. Both the ventral receptacle and spermathecae contained many sperm. Other dissections at 30 minutes resulted in expulsion of most of the mass. The pouch of a specimen dissected at one and one-half hours was considerably smaller but still contained a fair sized reaction mass. 13. cardini group Drosophila cardini Sturtevant. Matings were secured with fli~s 4 days old. The average length of copulation was about 23 minutes and 45 seconds. A specimen dissected at Studies in the Genetics of Drosophila 30 minutes is shown in Fig. 80 (PL XXI). The vagina appeared to be a little larger than that of a virgin but contained no observable semen. A very few non-motile sperm were visible along the ventral edge of the cavity. Sperm were present in both the ventral receptacle and sperma­thecae. Another specimen dissected at 30 minutes and one dissected at one hour appeared to have had some granular material in the vagina which escaped to the outside during dissection. In both cases a few motile sperm were observed in the saline solution. Dissections at one and one­half and at two hours showed no further ev?dence of an obvious reaction. 14. immigrans group Drosophila immigrans Sturtevant. Matings were secured with flies 3, 4, 5, and 9 days old. Considerable variation was found in the duration of copulation, the shortest recorded being 1'4 minutes, the longest, 64 minutes, with an average of about 30 minutes, although Sturtevant (1942) gives the average as 53 minutes. A dissection 4 minutes from the beginning of mating revealed that a small amount of granular material had been deposited in the vagina but no sperm were visible as yet. A dissection 10 minutes from the beginning ·of mating revealed many motile sperm in the vaginal cavity although none were seen in the receptacles. A specimen dissected immediately following a copulation of 30 minutes had an enlarged vagina filled with an emulsion-like material in which large numbers of motile sperm were visible. , The vagina appeared rounded and somewhat elongated but showed no observable tendency for pouch formation. A specimen dissected one ·hour after a copulation of 36 minutes had a riearly normal vagina, this organ being approximately the size of that of a virgin and containing only a minute amount of granular material. All of the receptacles were ·filled with sperm. The vagina of a specimen dissected two hours after a 37-minute copulation was very large, h?wever, and was filled with the reaction mass, motile sperm being observed only along the edges of the mass. A dissection 4 hours after mating once again revealed a vagina of normal size, free of all semen and sperm. Additional dissections were performed at varying intervals after mating in an effort to determine if there was any relation between the length of copulation and the dura­tion of the reaction. Only conflicting results have been secured to date so that a discussion of this question must await further study. 15. macroptera group 16. alagitans group 17. rubrifrons group No members of the above groups were available for study. 18. guarani group Drosophila guarani Dobzhansky and Pavan. A pair of flies was observed mating in the stock bottle and the female isolated. The dissection was made at 30 minutes with the ovaries attached. The Univers'ity of Texas Publication The vagina was slightly enlarged and was filled with a rather clear granu­lar material in which a few motile sperm were visible. There was a large mass of sperm in the median oviduct and some of these sperm were observed to have migrated into the lateral oviduct. The ventral receptacle contained a moderate number of sperm and the spermathecae were filled with these cells. Twelve females, 7 days old, were exposed for 16 hours 8 81 PLATE XXL Vaginae from homogamic matings of three species of Drosophila. Fig. 79, D. pallidipennis at one hour. Fig. 80, D. cardini at thirty minutes. Fig. 81, D. subbadia at thirty minutes. Studies in the Genetics of Drosophila to males of the same age and then dissected. Only two were found to have been inseminated and no evidence of a reaction was found. Drosophila subbadia Patterson and Mainland. One mating was observed in a vial of previously unmated, 10-day-old flies. The female was isolated and dissected 30 minutes later. The vagina was not enlarged and contained no semen or sperm. Both the ventral receptacle and spermathecae were teeming with sperm. This specimen is illustrated in Fig. 81 (Pl. XXI). Species unclassified as to groupin[J Drosophila parachrogaster Patterson and Mainland. One mating was secured with flies 22 days old. The length of copula­tion was 30 minutes and 25 seconds but it is likely that this is longer than the normal since the first part of the mating seemed to be faulty. The female was dissected 30 minutes after the completion of copulation. The vagina was nearly three times its normal size and appeared to be elongate and rounded with no tendency for pouch formation. Both the ventral receptacle and spermathecae were teeming with sperm. Before this specimen could be photographed at least half of the vaginal contents were expelled through the oviduct. Many motile sperm were then observed in the saline mounting fluid. Drosophila pallidipennis Dobzhansky and Pavan. Matings were secured with flies 4 to 6 days old. The average length of copulation was about 30 minutes. Dissections made at 30 minutes and at one, one and one-half, and two hours after mating were alike in nearly all respects. The preparation shown in Fig. 79 (Pl. XXI), dissected at one hour, is typical of this species. · The vagina was exceedingly large and showed no tendency for pouch formation. The entire cavity was filled with a very dense, opaque reaction mass with a few motile sperm visible along the edges. Some of this mass and the contained sperm had moved into the oviduct. Both the ventral receptacle and spermathecae contained many sperm. In the specimen illustrated one spermathecal stalk was doubled at the tip and possessed two bodies. Drosophila pallidipennis centralis Patterson and Mainland. Matings were secured with flies 2 and 3 days old. The average length of copuJation was about 32 minutes. Dissections at 30 minutes, and at one, and one and one-half hours after mating showed that the develop­ment of the reaction in this form follows very closely the pattern of the typical species discussed above. The dissection at one hour was of particular interest since the male did not appear to have delivered any spermatozoa in the ejaculate. The reaction developed, however, in the usual fashion, indicating that the presence of sperm is not a requisite for the formation of the insemination reaction. DISCUSSION In the present section it is proposed to discuss briefly the .sequence of events following insemination in view of the varied expression of the insemination reaction and to present a classification of the species and species groups based on the type of reaction they show. Finally, an attempt will be made to show some possible functions of the reaction in homogamic matings. An attempt to interpret the mechanisms involved in a reaction between the semen of the male and the reproductive tract of the female should logically begin with an inspection of the origin and development of the semen and contained sperm since it is obvious that there can be no in:­semination reaction without one or both of these-products. The only other alternative would be the possibility that the phy;:;ical act of copulation, per se, might inaugurate the reaction. The fact that many copulations, particularly those with old males, fail to result in ejaculation and, simi-. larly, fail to result in the development of the typical reaction would tend to show that strictly mechanical processes are incapable of provoking this phenomenon. Dr. Patterson has shown in the preceding article that when known sterile males are mated to females which would normally develop the reaction following copulation the typical form of the insemination re­action follows even though no mature sperm are delivered. Further evidence that sperm are not a requisite for the development of this reaction is presented by the dissection of a D. pallidipennis centralis female, described in the previous section of this paper, in which it was observed that the male had apparently failed to deliver any sperm in . the ejaculate, but, nevertheless, the typical reaction developed. According to Nonidez (1920), who studied the pass~ge of the sperma­tozoa from the time of their formation to ejaculation, the liquid part of the ejaculate is composed of the secretions of the paragonia. He describes this material as a dense, sticky fluid in which float abundant refractive granules of varied size. The similarity of this description to that of the material observed in the vaginal cavity of many species immediately following copulation is strikingly obvious. There can be but little doubt that the material seen in these vaginae at this time is composed largely of the paragonial secretions and the spermatozoa. The emulsion-like material observed in certain species, e.g., D. mercatorurn, merely repre­sents a specific modification of the type of secretion. It seems likely, then, that any observed reaction is largely a i::,eaction between these secretions and the female reproductive tract and, further,. more, that the severity of the ·visible reaction is determined by those changes in the paragonial secretions which are observable. In some species, e.g., D. polychaeta, one is unable to detect any change, the ejacu­late disappearing from the vaginal cavity with remarkable rapidity. In other forms, members of the obscura group, for example, this material disappears much less rapidly, while in species like D~ ananassae it seems Studies in the GenetiCs of Drosophila to harden into a gel-like mass. In the most extreme cases, e.g., in mem­bers of the mulleri subgroup, the vagina becomes tremendously swollen and the contents become consolidated into a dense, formed mass which may remain in the vagina for many hours. It is difficult to explain the relatively sudden swelling of the vagina in these latter forms. Patterson (1946) concluded that the introduction of the semen into the vagina was followed almost immediately by a reaction of the mucous membrane which secreted a relatively large amount of fluid into the cavity, thus bringing about the characteristic swelling of this organ. This explana­tion adequately explains the increase in size but should not be interpreted as resulting in a dilution of the ejaculate but rather as resulting in a further reaction which is evidenced by the consolidation of the material into a formed, persistent mass. In view of the many observed species differences with respect to the age at mating, the duration of copulation, the speed with which the sperm migrate into the receptacles, etc., it is not surprising that the observed severity of the insemination reaction should also vary among the dif­ferent forms. Such variation is apparent among the species described in the preceding section of this paper. Table 1 represents an attempt to classify the species under consideration into three groups based on the type of insemination reaction they exhibit. Generally speaking, class 1 is composed of those forms in which no reaction is apparent; class 2 con­tains those in which a slight or moderate reaction develops; and class 3 contains those in which the severe enlargement of the vagina and the TABLE 1 Classification of seventy-eight species with respect to the type of insemination reaction present Class 1 Class 2 Class 3 C. procnemis S. graminum D. sturtevanti D. rectangularis D. nebulosa D. melanogaster D. simulans D. tripunctata D. crocina D. repleta D. melanopalpa D. hydei D. robusta D. polychaeta D. cardini D. guarani D. subbadia C. amoena D. busckii D. prosaltans D. cordata D. elliptica D. emarginata D. willistoni D. equinoxialis D. sucinea D. fumipennis D. ananassae D. pseudoobscura D. persimilis D. affinis D. azteca D. tolteca D. neorepleta D. "limensis" D. canapalpa D. duncani D. victoria D. transversa D. munda D. quinaria D. subquinaria D. suboccidentalis D. innubila D. subpalustris D. guttifera D. virilis D. americana D. a. texana D. montana D. lacicola D. putrida D. funebris D. subfunebris D. macrospina D. mulleri o._ aldrichi D. arizonensis D. mojavensis D. buzzatii D. hamatofila D. ritae D. peninsularis D. mercatorum D. m. pararepleta D. meridiana D. anceps D. hexastigma D. gibberosa D. melanica D. m. paramelanica D. nigromelanica D. micromelanica D. car bonaria n. immigrans D. parachrogaster D. pallidipennis D. p. centralis dense, opaque nature of the mass are easily observed. Obviously, attempts · to classify the borderline cases can only reflect personal opinion since there are no definite rules that may be relied upon in such cases. Where considerable doubt existed as to the proper category to which certain forms belonged they have been placed in the lower one of the two classes concerned. In addition to the seventy species considered in this paper, the eight members of the mulleri subgroup have been added in order to include all species for which information was available. The major distinction between members of class 1, in which no reaction of any sort can be observed, and members of class 2, in which only a slight reaction can be observed, is largely the speed with which the ejacu­late disappears from the vaginal cavity. Certainly, if a female of class 1 is dissected at the appropriate time during copulation the vagina will be seen to contain a certain amount of ejaculate. Frequently, however, this material is so slight in amount and disappears so rapidly that by the end of copulation no signs of it are left. On the other hand, a like amount of material present in the vagina of a female of class 2 persists for a longer time and is visible after copulation. Furthermore, in none of the members of class 1 does there appear to be a consolidation of the ejaculate to form an observable persisting mass. One cannot say, however, that the pres­ence of the ejaculate in the tract of the fema1e does not stimulate secre..: tions by the female organs nor exert some influence on the vagina that may affect fertilization or oviposition. The members of class 1 form a rather heterogeneous group. The tripunctata, robusta, polychaeta, cardini, guarani, and part of the repleta species group of the subgenus Drosophila are included here as well as parts of the saltans and melanogaster groups of Sophophora. Sturtevant (1942, p. 29) separates the saltans group into subgroups on the basis of their mesonotal markings. With the exception of D. prosaltans, which has doubtfully been placed in class 2, the members of this group which are here considered as belonging to class 1 occur in the same subgroup, while the members of the second subgroup have been placed in class 2, thus adding evidence that the division is a natural one. Class 2 consists of those forms which are intermediate between species showing no observable reaction and those in which a violent and persistent reaction develops. Most of the species included in class 2 exhibit a cer­tain amount of ejaculate in the vaginal cavity at least thirty minutes after mating but do not show any obvious enlargement of this organ. Frequently the ejaculate seems to gel, as in D. ananassae, and this formed mass may be retained for several hours before expulsion. With the exception of C. amoena, D. (Dorsilopha) busckii, and certain members of the repleta section of the repleta group of the subgenus Dro­sophila, class 2 contains those members of the subgenus Sophophora not included in the first class, i.e., the willistoni (except for D. nebulosa) and obscura groups, and the remainder of the saltans group. Studies in the Genetics of Drosophila Class 3 contains those forms in which enlargement of the vagina is usually accompanied by the development of a dense reaction mass. Quite frequently the posterior portion of the vagina enlarges to form a pouch. This may be accompanied by enlargement of the entire vagina, as in D. mercatorum, or the enlargement may be confined to the pouch, as in D. carbonaria. A few forms included in this class, e.g., D. anceps, fail to develop the dense, opaque mass, this material appearing clear and crystalline when viewed under the microscope. Inspection of the members included in class 3 reveals that all but two belong to the subgenus Drosophila, the exceptions being D. (Hirtodro­sophila) duncani and D. (Pholadoris) victoria. It is of interest to compare the hypothetical relationships of the mem­bers of the genus as suggested by Sturtevant (1'942) in the light of the insemination reaction. Two possible evolutionary interpretations may be considered : (a) independent origin of the character at various times in different phylogenetic lines, and (b) a single appearance of the character early in evolution followed by modification along the various lines. It is important to remember here that, as was pointed out by Patterson (1946), if the insemination reaction arose as a mutation, irrespective of any selective value it might have to the species as a whole, it would spread throughout the population. As a consequence, once such a mutation arose in an evolutionary line it would tend to remain. This does not imply that such a mutation could not occur several times, and, quite pos­sibly, in different ways, but it does imply that once the reaction character developed in a form it would remain in the lines of descendance even though modified by further changes. With respect to Drosophila phylogeny, then, certain lines of descent, as interpreted by Sturtevant, seem to fit this pattern very well, as, ·for example, the virilis-quinaria-testacea-guttifera groups. It seems probable, however, that the origin of the subgenus Dorsilopha (D. busckii) from this stem is doubtful. Similarly, a study of the supposed lines of descent through the virilis-tripunctata-funebris-repleta groups sugg.ests that the tripunctata group is out of place in this series and that what is generally termed the repleta group is probably composed of several independent species groups, the hydei-repleta portion having evolved along one line, the mulleri-mercatorum portion along another. The arrangement of the groups of Sophophora fits the pattern quite well while the origin of the Hirtodrosophila from this stem requires the assumption of an independent mutation for this character. Several possible functions of the insemination reaction in homogamic matings may be mentioned. It has been pointed out that the lack of art observable reaction in members of class 1 does not necessarily mean that the only function of insemination is sperm delivery. The comments of Patterson ( 1946, p. 207) are especially pertinent: "It may have the effeCt here of preparing the reproductive tract ·for the fertilization mechanism The University of Texas Publication which is to follow. It should be pointed out, in this connection, that even in forms which show no visible reaction there still may be a change in the mucous membrane which has the same effect." The retention of the ejaculate in members of class 2 appears to be correlated in some instances with the slower movement of the spermatozoa into the receptacles. It seems logical that in these forms the carrier fluid is retained until most of the sperm have had time to reach the receptacles. Evidence is presented that after a certain length of time this material is expelled by the female along with the excess sperm. This expulsion (in D. melanogaster, for example) may account in part for the discrepancy in the number of sperm deposited by the male of this species in relation to the number utilized by the female as determined· by -Kaufmann and Demerec (1942). They found that each fertile egg layed accounted for twenty to thirty sperm cells whereas numbers greater than seven or eight per egg were rarely seen, the conclusion being that large numbers of sperm were squandered during the laying of the first eggs. It now appears likely that expulsion by the ) female of the excess sperm may account, at least in part, for this apparent squandering. In D. ananassae the reaction material seems to form a plug at the posterior orifice, effe~tively preventing expulsion for some time. It is suggested that secretions of the female tract may eventually soften this plug and allow its expulsion or it may be forced out due to muscle contractions. Among members of class 3 it has been frequently observed that motile sperm were visible around the edges of the mass while only arent and having females with the formula 2A :XXYY and males with the formula 2A :XYYY. THE TWO STRAINS OF DROSOPHILA PENINSULARIS In 1942 we described a new species from Florida under the name Dro­sophila peninsularis (Patterson and Wheeler 1942). Last May, 1946, Dr. Harrison D. Stalker of Washington University, St. Louis, sent us a culture of a new geographic strain of this species from Tarpon Springs, Florida, where he had collected twenty-nine specimens. We are grateful to Dr. Stalker for his generosity in sending this stock with his permission to use it in any way we might desire. A preliminary examination of these flies confirmed Stalker's identifi­cation that they belonged to the species D. peninsularis. A more detailed study of this stock revealed that the strain from Tarpon Springs showed certain differences when compared with the type strain from Lake Mc­Kethan, a stock of which was still available in the laboratory. Specimens The University of Texas · Publication of the new strain have a lighter body and eye color than those of the Lake McKethan strain. Another difference observed concerns the nature of the spermathecae stalks. In the strain from Tarpon Springs these stalks are separated, as in most species of Drosophila, with their points of origin from the vagina usually some little distance apart. In contrast to this condition, these stalks of the Lake McKethan strain are fused, and arise from a common point on the wall of the vagina. The phenotypic expres­sion of this character is variable and shows different degrees of fusion of the two stalks. On the basis of these observations, it was decided to carry out a series of cross tests with the two strains to see if there was any genetic difference between them. Five different types of crosses were carried out with the two strains, and the results obtained are recorded in Table 1. The five types of crosses were, controls, P/s, F1 inbred, first backcrosses, and the inbred progeny of these backcrosses. The letters M and T are used to designate the Lake McKethan and Tarpon Springs strains, respectively. In the table the pedigree of each type of hybrid is indicated as follows : The original symbols show the female first. Thus, MT is the hybrid from the cross Lake McKethan 'i' X Tarpon Springs ~ , and TM indicates the reciprocal cross. For the hybrid progeny of the backcrosses, the symbol (MT) M is the offspring of an F1 hybrid female crossed to a Lake McKethan :male, while the symbol M (MT) represents the offspring of a Lake McKethan female crossed to an F1 hybrid male. TABLE 1 Crosses between Lake McKethan (M) and Tarpon Springs (T) geographic strains of Drosophila peninsularis . Crosses Matings 'i' ~ Percentage fertile Sex ratios 'i' 8 Number per culture Controls MxM T X T 89 35 47.8% :52.2% 48.7% :51.3% 26.1 25.6 P1 crosses MXT TXM 76 52 47.8% :52.2% 47.1 % :52.9% 24.2 21.4 F1 inbred MTxMT TMXTM 70 73 44.5% :55.5% 53.9% :46.1% 69.8 53.8 MT x M 51 49.7% :50.3% 30.6 MTXT 68 44.6% :55.4% 33.6 TMXT 76 44.2% :55.1)% 21.2 First TMxM 56 49.2%:5M% 29.5 backcrosses MxMT 67 46.7%:53.3% 37.2 MXTM 45 50.4% :49.(1% 35.2 TxMT 16 50.2% :49.1)% 26.3 T XTM 32 50.7% :49.:}% 29.2 Inbred progeny of first backcrosses (MT)M X (MT)M (MT)T X (MT)T (TM)M X (TM)M (TM) T X (TM) T M(MT) X M(MT) M (TM) X M (TM) T(MT) X T(MT) 47 35 60 40 44 65 40 50.0% :50.0% 50.7% :49.3% 51.5% :48.5% 49.7%:50.3% 50.3% :49.7% 52.4% :47.6% 50.9% :49.l% 43.3 34.5 61.0 33.1 38.0 47.6 32.8 T(TM) X T(TM) 50 51.0% :49.0% 39.0 Studies in the Genetics of Drosophila In addition to the types of crosses and matings, the table also shows the fertility percentages, the sex-ratios, and the number of offspring per culture for each type of mating. The fertility percentage is based on a check of one-hundred tested pairs, and the number per culture is based on the average number of offspring from a sample of twenty-five fertile pairs. The control crosses gave fertility percentages of 89 and 35 for the M X Mand T X T matings, respectively. The fertility percentage of 35 for Tarpon Springs seems low for homogamic matings, but this result is not unique. In the pallidipennis-centralis tests the homogamic matings of centralis also gave a fertility percentage of only 35. It is a rather common experience to find strains of Drosophila which breed well in mass cultures, but fail to do so in pair mating tests. This was found to be true for several strains of members of the virilis group (Patterson and Griffen 1944). The sex-ratios for the two control crosses are similar, each with slightly more males than females, and the average number of offspring is also similar for the two crosses. The P1 crosses gave per­centages of fertility of 76 for the M X T mating, and 52 for the T X M mating. The sex-ratios and the number of offspring per culture are similar to those for the control crosses. The two F1 inbred crosses gave very similar fertility percentages, and both showed a much higher average per culture than either the control or P1 matings, indicating hybrid vigor. The fertility percentages of the eight backcrosses varied from 16 to 76, with a general average of 51.3. The ~eneral average per culture was 30.4. Both of these averages are higher than the corresponding ones for the control or P1 matings. The fertility percentages of the eight inbred tests varied from 35 to 65, with an average of 47.6. The general average of the number per culture was 4l.l. The difference in the fertility percentages of these two sets of matings is not significant (51.3 vs. 47.6), but the difference in the aver­age per culture is (30.4 vs. 41.1). While the sex-ratios for the different matings vary, yet none of them deviate significantly from a 1 :1 ratio. TABLE 2 Dissection records of males and females from sterile matings Males Females 2 Crosses (; motile sperm non-motile or immature sperm inseminated offspring MT x MT 9 1 5 none TM X TM 2 8 0 none TM X T 2 8 4 none TM x M 7 3 7 none M X TM 10 0 0 none T x TM 5 5 0 none (MT)M X (MT)M 0 10 0 none (TM ) M X (TMJ M (TM)T X (TM )T 10 10 0 0 4 5 none none M(TM) X M(TMJ 9 1 4 none T(TMJ X T(TM) 10 0 2 none The Univ ersity of Texas Publication In order to determine why so many of the hybrid crosses resulted in sterile matings, samples of ten pairs of suc}l matings from each of eleven different crosses were dissected. The results obtained from these dissec­tions are recorded in Table 2. Of the two F1 inbred crosses, the MTX MT matings had nine males with motile and one with immature sperm. Five of the females had been inseminated, but, like all of the inseminated females found in this series of dissections, none produced offspring. The TM X TM matings had two males with motile and eight with immature sperm. None of the ten females had been inseminated. Four samples of the eight different backcrosses were dissected. Two of these represent backcrosses of TM females to the parent types of males, and two represent TM males backcrossed to the parent types of females. In the TM X T matings two of the males had motile and eight had non-motile sperm. Since the ventral receptacles of four of the females coptained non-motile sperm, and only two of the males had motile sperm at the time the dissections were made, two of these females must have mated before the sperm had become inactive in the testes. In the TM X M matings seven of the males had motile and three had non-motile sperm. Seven of the females had been inseminated, the receptacles of four con­taining motile sperm. In the M X TM matings all of the males had motile sperm, but none of the females had been inseminated. In the T X TM matings five of the males had motile and five had non-motile sperm. None of the females had been inseminated. Five different types of matings from ·the inbred tests of the progeny of the backcrosses were examined. In the first of these, the (MT) M X (MT) M matings, all ten males had immature sperm, and naturally the females contained no sperm in their ventral receptacles: In the four remaining types of matings, all of the males had motile sperm except one. In the M(TM) X M(TM) . matings a single male had immature sperm in its testes. The results obtained in the dissections of the flies from fertile matings show that the sterility was due to two or more causes. In some of the crosses it was the result of sexual isolation. This is evidently true for matings in which none of the females had been inseminated, although most of the males contained many motile sperm in their testes. The matings M X TM and T X TM belong in this category. A second cause of sterility is seen in crosses in which the process of sperm formation had not been completed. In such males spermiogenesis proceeds to the point at which the strands of immature spermatozoa are formed, but stops just short of motile-sperm formation. Good examples of this type of sterility are found in the TM X TM matings, in which eight of the ten males had immature sperm, and the (MT) M X (MT) M matings in which all ten males had immature sperm. In two matings only one male out of ten had such sperm. These were MT X MT and M(TM) X M (TM). Finally, a few cases were observed in which a single cyst had succeeded in producing motile sperm, while the rest of the testis contained only immature stages. A third possible cause of sterility is the inactivation Studies in the Genetics of Drosophila o! the sperm in the ventral receptacle of the female. Several cases of this type were found, two of which were seen in the TM X T and TM X M matings. The two geographic strains were tested for sexual isolation by the multiple choice method. Five females of each strain were mated to a single male of one of the strains. The results from these tests are pre­sented in Table 3. The Lake McKethan males showed a distinct preference for their own kind of female. The x2 value and the isolation index both indicate a significant difference between the homogamic and heterogamic matings. In the reciprocal cross, Tarpon Springs males showed a slight preference for Lake McKethan females, giving a negative isolation index, but the x2 value of 4.0 and the isolation index of -0.19 do not indicate a significant difference. We have found two crosses in which males of one species preferred the alien females to their own. Thus, texana males prefer americana females, and buzzatii males prefer arizonensis females. The same was found to be true for several different strains among mem­bers of the virilis group (see Article I). Negative isolation indices have been reported for geographic strains of D. willistoni (Dobzhansky and Mayr 1944), and for those of D. prosaltans (Dobzhansky and Streisinger 1944). TABLE 3 Number of female dissected (n) and per cent inseminated (%) in crosses between the two strains of D. peninsularis Iso­Homogamic Heterogamic lation x2 Females Males n % n % Index McKethan, Tarpon______________________ _ McKethan 100 50.0 102 26.4 11.6 0.31 Tarpon, McKethan______________________ ___ Tarpon 101 28.7 102 42.1 4.0 -0.19 THE FUSED MUTATION As stated above, the strain from Lake McKethan contains a mutant factor which causes a fusion of the spermathecae stalks. This character is variable in expression. In some specimens the proximal ends of the stalks are united, forming a V-shaped figure; in others the union involves varying lengths of the proximal ends, giving Y-shaped figures ; in still others the two stalks are fused throughout their entire lengths, with the spermathecae occupying a terminal position. In contrast to this condition, the stalks of the Tarpon Springs strain arise at separate points in the wall of the vagina, as is the usual condition in most species of the genus Drosophila. In Table 4 are listed the conditions of fused in seven different female genotypes. An analysis of these data shows that non-fused is dominant, or nearly so, over fused. The lack of complete dominance may be due to the presence of heterozygotes in the Tarpon Springs strain. This is indi­cated by the fact that three of the one hundred dissected females of this The University of Texas Publication strain showed the character fused. Since the original flies of the two strains were collected at points only about thirty-six miles apart, crossing between the two strains might well have taken place. The data also reveal that fused is not sex-linked, and that only one or a few factors are involved. TABLE 4 Determination of the character fus ed in different female genotypes Type of female non-fused fused V-shaped Y-shaped Terminal M___________________________________________ __________ r __________ _____________________________________________ MT_____________________________________________________ TM_____________________________________________;____ (MT) M____________________________________ (MT) (MT) ______________________________________ (TM) (TM) ----------------------------------­ 0 97 68 83 50 67 48 6 2 21 13 22 12 34 90 4 1 0 11 0 4 0 26 2 19 2 17 1 DISCUSSION AND CONCLUSION The two geographic strains of D. peninsularis add another member to the growing list of pairs of similar Drosophila forms which have attained some degree of reproductive isolation without at the same time under­going marked morphological differences. This pair of strains may be compared with the three other cases cited in the introductory section of this article. Of the four cases, the willistoni-equinoxialis pair has reached the highest degree of divergence, and, having developed complete repro­ductive isolation, the two forms represent full species. Next in .line would be the pallidipennis-centralis pair, in which the F1 males are completely sterile, and most of the males of the succeeding backcross generations are also sterile. The members of this pair have been ranked as subspecies. Close to this case would be the mercatorum-pararepleta pair. The recipro­cal crosses of the two forms produce viable hybrids which show hybrid vigor, but the F 2 crosses show a great reduction in fertility, and the F 2 flies are less fecund and viable than the F1 -individuals. These forms have also been ranked as subspecies. The two geographic strafos of peninsularis show at least three different types of isolating mechanisms operating in the direction of divergence, but they have not reached the taxonomic level of subspecies. The two members of each of the four pairs may be compared with reference to their metaphase and salivary gland chromosomes. Accord­ing to Dobzhansky (1946), the metaphase chromosome patterns of willistoni and equinoxialis are alike, but the latter species was not favor­able .for the study of salivary gland chromosomes. Patterson and Dob­zhansky (1945) found that the metaphase patterns of pallidipennis and centralis were similar. A study of the salivary gland preparations of the F1 larvae showed that synapsis was complete. The only difference in gene arrangement of the two forms is a long inversion in one of the Studies in the Genetics of Drosophila autosomes. Wharton (1944) has shown that the metaphase patterns of mercatorum and pararepleta differ. The females of both forms have in common two large V's, a pair of long rods, and two pairs of short rods, but pararepleta has a pair of dots, whereas mercatorum has a pair of small V's instead. The male patterns are similar, except that the short Y chromosome is absent in mercatorum. She suggests that the condi­tion in mercatorum has been brought about by a fusion or translocation of the short Y to a dot-like chromosome of a pararepleta-like form, thus forming a small V-shaped element. Both forms show five long euchro­matic arms and the dot-like element in salivary gland preparations. Dr. W. S. Stone kindly made preparations of brain cells of the Tarpon Springs strain, and of the salivary glands of the F1 hybrid larvae of the two strains of peninsularis. He reports that the metaphase patterns of the two forms are alike, and that the pairing of the salivary chromosomes is complete, with no apparent differences in gene arrangement. The degree of divergence exhibited between the members of each of these four pairs of forms must have been brought about by genetic fac­tors, rather than by chromosomal differentiations. This seems evident from the fact that very little difference in metaphase patterns and gene arrangement could be detected. Even in the mercatorum-pararepleta forms, in which tlie short Y and dot-like chromosomes had fused to form asmall, V-shaped element, the change does not represent a gene rearrange­ment of any great magnitude. Some of the more important components of speciation in Drosophila are found among the four cases under consideration. We have shown, however, that two of these components, morphological differentiation and divergence of gene arrangement, are not involved in these forms, at least to any great extent. Sexual isolation, which is one of the most common isolating mechanisms in Drosophila, is almost complete between willistoni and equinoxialis, and although interspecific inseminations some­times occur, hybrids are never produced. Sexual isolation between pallidipennis and centralis is very weak and is restricted to only one of the two possible reciprocal crosses. It has developed between the two strains of peninsularis, as is shown by the results from the dissections of the flies from sterile matings, and from the sexual preference tests. Thus far, no critical experiment has been carried out to determine whether sexual isolatiOn has developed between the members of .the mercatorum-pararepleta pair of subspecies. The development of hybrid sterility is also of common occurrence in Drosophila. It is not involved in the willistoni-equinoxialis crosses, for the reason that hybrids are never produced. The F 1 hybrid males from the reciprocal crosses of pallidipennis and centralis are completely sterile, and a large majority of those from succeeding backcrosses are also sterile. There is no conclusive evidence that the F1 flies from the reciprocal crosses between mercatorum and pararepleta are sterile, but Wharton observed a striking reduction of fertility in nearly all F 2 crosses. There is evidence The University of Texas Publication that some of the hybrids from crosses of the two strains of peninsularis are sterile. In the case of male hybrids the sterility is due to a failure of gamete formation. F The failure of spermatogenesis to produce functional gametes has been observed in the hybrids of several different forms. It occurs in the 1 sterile hybrids of pallidipennis and centralis. In these males the first and second meiotic divisions are abortive, and this results in the formation of groups of abnormal spermatids which finally degenerate to form granu­lar masses (Patterson and Dobzhansky 1945). This case was worked out by Dobzhansky, who previously had reported (1934) a similar degen­erative process in the hybrids between weak strains of D. pseudoobscura and D. persimilis. A somewhat similar degenerative process has been observed in the sterile male hybrids of the two strains of peninsularis. In this case the degeneration does not begin until near the end of sper­miogenesis, and results in the formation of bundles (strands) of non­motile sperm. A very similar phenomenon has been found by Stone in sterile F 2 males between members of the virilis group (see article IX in this publication). Another mechanism which adversely affects reproduction in Drosophila is the inactivation of the spermatozoa after they have been introduced into the receptacles of alien females. This was first established in crosses between members of the virilis group (Patterson 1942; Patterson, Stone, and Griffen 1942). It has frequently been observed in crosses between the two strains of peninsularis. A similar phenomenon has been reported by Oliver and his students in D. melanogaster. In this case, however, the loss of sperm motility in the female is associated with one of the multiple effects of the mutant lozenge gene (Oliver and Green 1'943; Oliver and Anderson 1945; Anderson 1945). The geographic relationships of the two members of each of the four pairs of similar forms are not altogether clear. Dobzhansky is uncer­tain as to whether willistoni and equinoxialis are wholly allopatric or sympatric in parts of their distribution ranges. The question is not too important, since they represent full-fl.edged species. In the case of pallidi­pennis and centralis, not too much is known about their distribution ranges. The former was taken at two places in southern Brazil, and the latter at one place in the state of Vera Cruz, Mexico. From the results of cross tests it seemed best to regard them as subspecies. Of the third pair, mercatorum is widely distributed over southwestern United States, and its range extends down through Mexico and Central America to as far south as Lima, Peru, while pararepleta has been reported from south­ern Brazil. These two forms are also regarded as subspecies. The pair of geographic strains of peninsularis came from Florida, where the species has been collected at eight different places. The experimental tests show that these two forms have not diverged to the subspecific level. Various degrees of evolutionary divergence are found among pairs of similar, closely related forms in the genus Drosophila. At the upper level Studies in the Genetics of Drosophila are forms like the melanogaster-simulans and willistoni-equinoxialis pairs which represent full-fledged species, no longer capable of gene exchange. At the middle, or subspecific, level are several pairs in which gene ex­change between the two members would be possible wherever their distribution ranges might overlap. Among these are the pallidipennis­centralis and mercatorum-pararepleta pairs. Finally, at the lower level are geographic strains in which gene exchanges can take place rather readily, but which show the very beginning of evolutionary divergence. Such a case is represented by the two strains of peninsularis from Florida. REFERENCES Anderson, Ray C. 1945. A study of the factors affecting fertility of lozenge females of Drsophila melanogaster. Genetics, 30 :280-296. Dobzhansky, Th. 1934. Studies on hybrid sterility. I. Spermatogenesis in pure and hybrid Drosophila pseudoobscura. Z. Zellf. mikr. -Anat., 21 :169-223. Dobzhansky, Th. 1946. Complete reproductive isolation between two morphologically similar species of Drosophila. Ecology, 27 :205-211. Dobzhansky, Th., and C. Pavan. 1943. Studies on Brazilian species of Drosophila. Biol. Facul. Fil. Cien. Letr. Univer. Sao Paulo, No. 36, Biol. Geral., 4:7-72. Oliver, C. P., and M. M. Green. 1944. Heterosis in compounds of lozenge alleles in Drosophila melanogaster. Genetics, 29 :331-347. Oliver, C. P., and R. C. Anderson. 1945. The effect of rematings on the fecundity of an infertile mutant female. Amer. Nat., 79 :89-94. Patterson, J. T. 1942. Isolating mechanisms in the genus Drosophila. Biol. Symp., 6:271-287. Patterson, J. T., and Marshall R. Wheeler. 1942. Description of new species of the subgenera Hirtodrosophila and Drosophila. Univ. Tex. Puhl., 4213 :67-109. Patterson, J. T., W. S. Stone, and A. B. Griffen. 1942. Genetic and cytological analysis of the virilis species group. Univ. Tex. Puhl., 4228:162-200. Patterson, J. T., and A. B. Griffen. 1944. Relationships of Drosophila montana and D. lacicola to other members of the virilis group. Univ. Tex. Puhl., 4445:193-211. Patterson, J. T., and Th. Dobzhansky. 1945. Incipient reproductive isolation between two subspecies of Drosophila pallidipenni.s. Genetics, 30 :429-438. Wharton, Linda T. 1944. Interspecific hybridization in the repleta group. Univ. Tex. Puhl., 4445: 175-194. VI. A STUDY OF THE ISOLATING MECHANISMS FOUND IN DRO­ SOPHILA ARIZONENSIS AND DROSOPHILA MOJA VENSIS WILLIAM K. BAKER* INTRODUCTION From a complete study of only one stage of any dynamic process, for example, evolution, it is impossible to determine the past interactions of events leading to this particular stage; neither is it possible to predict · accurately the characteristics of a future stage in the process, nor the future interactions of mechanisms observed to be operative during a particular stage. Obviously, a thorough knowledge of a single process can only be obtained by a study of all stages in the course of events. Because of the time limitations placed upon experimental investigations of evolu­tion, it is impossible to study more than "one" stage of this process in a particular form. In order to partially circumvent this difficulty, studies .are made of different stages in the evolutionary process in forms where the pattern of evolution is thought to be comparable, i.e., closely related species. The present investigation is an attempt to determine some of the inherent factors existing at the present time in different populations of Drosophila arizonensis which might serve to isolate these strains, as well as an attempt to discover the nature of isolating mechanisms which are effective between this species and its very close relative, Drosophila mojavensis. MATERIALS AND METHODS The collection records of this laboratory (Patterson and Wagner 1943) indicate that the bulk of the arizonensis population is concentrated in the region of the Sonoran Desert which covers southern Arizona and most of the state of Sonora, Mexico. Two other apparently isolated populations of this species exist in the state of Tamaulipas near the east coast of Mexico and in the state of Guerrero on the southern coast. Fur.: ther collections along the Pacific Coast of Mexico will be necessary in order to determine whether there is a connecting link between the latter population and the large population located in the Sonoran Desert. The distribution of mojavensis is not as well known. Spencer (1941) indi­cates that it is common in the Mojave Desert and Death Valley regions of southern California, and Sturtevant (see Patterson and Wagner 1943) reports that this species has been found in the part of the Sonoran Desert comprising Baja California. Collections in southwestern Arizona have failed to reveal a zone of overlap in this region between these two species. *Pre-doctoral Fellow of the National Research Council. Studies in the Genetics of Drosophila In the experiments to be described, the stocks listed below were utilized. Along with the geographical origin, the symbol for each stock, as used in the tables, is given . . D. arizonensis Patterson and Wheeler, 1942. A-Tucson, Arizona As-San Bernardino, Arizona Am-Magdalena, Sonora, Mexico Ag-Guaymas, Sonor:;i, Mexico At-Rio Purificacion, Tamaulipas, Mexico D. mojavensis Patterson and Crow, 1940 (redescribed by Patterson and Wheeler, 1942). Md-Mesquite Springs, Death Valley, California Mc-Chocolate Mountains, just east of Salton Sea, California. In order to determine qualitatively the type of isolating mechanisms, if such should exist, which would be effective in curtailing gene exchange between the different strains of arizonensis, the following procedure was used in the experiments to be discussed. Pair matings of, virgin flies from three to six days old were made in shell vials containing a banana­ . Karo-yeast-agar medium which was inoculated with a heavy suspension of yeast. Fourteen days after the matings were made any vials in which both parents were not living or vials which were contaminated with mold were discarded. At this time the vials were examined under the stereoscopic microscope, if necessary, to determine the presence or absence of larvae. Within forty-eight hours from the time of this examination, the females of the pairs which did not produce any progeny were dis­sected and their reproductive tracts examined under the compound micro­scope for the presence of spermatozoa. If less than thirty pairs of a particular cross did not produce any progeny, all of the females were dissected; however, if more than thirty pairs produced no offspring, dis­sections were made on a sample of at least twenty-five females. Twenty days after matings were made, ten vials were selected at random from the pairs which had produced offspring and the number of imagines per · vial was recorded. The same general procedure was used in obtaining the data on the hybrid arizonensis-mojavensis crosses, as presented in Table 4, with · these exceptions : virgins from six to eight days old were used and dis­sections were made of all females which did not produce progeny; these dissections were made on the fourteenth day after making the matings. I:p these crosses the reproductive tracts of the females were examined to determine the presence of the "insemination reaction" (Patterson 1946) as well as the presence of motile or immotile sperm. From the preceding paragraphs it is evident that these experiments were designed to measure the following three quantities: (1') the num­ber of females fertilized-a measure of sexual isolation, (2) the rela­tion between the number of females fertilized and the number producing progeny-a measure of the effect of isolating mechanisms acting be~ween the time of insemination and the hatching of larvae from the eggs, (3) the number of imagines produced in twenty days-a partial measure of the fecundity of the cross. MATING RESULTS In Table 1 are recorded the results of intraspecific matings of the five strains of D. arizonensis. The first column of this table indicates the cross being tested. In the second column is recorded the number of vials in which both parents were living fourteen days after the particular mating was made. The percentage of these pairs which produced off­spring in the fourteen-day period is given in the third column. The fourth column is a tabulation of the percentage of the females tested which were inseminated. This includes the females producing offspring as well as the number of females, as estimated by the sample of non­f ertile females dissected, which were inseminated but did · not produce progeny. Recorded in the last column is the average number of adult progeny produced within the twenty-day pe~iod after the matings were made. In none of the crosses being reported were any significant devia­tions in the sex ratio of the offspring apparent; therefore, in the fifth · column the number of males and females produced are lumped together. Tables 2 and 3 are arranged in the same manner as Table 1. TABLE 1 Fertility Relationships in P, Crosses of D. arizonensis % 'i''i' Estimated No. Pairs Producing %'i''i' No. Progeny Na1·11re of Cross Tested Progeny Inseminated per Vial A'i' X A & ..·-············· 152 78 84 42 As 'i' X As &-----------··· 87 26 37 39 Am 'i' X Am & ___________ 175 75 88 23 Ag 'i' X Ag & -------·-----· 141 55 62 38 As 'i' X A & -----------···--152 68 74 39 A 'i' X As & ________________ 128 54 70 36 Am 'i' X A & --------------165 60 69 66 A 'i' X Am &--------··----134 63 71 62 Ag 'i' X A & ---------------72 32 43 A 'i' X Ag & ---------------112 70 75 60 At 'i' X A & ----------------56 70 82 30 A 'i' X At J ________________ 59 86 90 22 As 'i' X Am &------------· 148 34 48 27 Am 'i' X As & ------------· 164 As 'i' X Ag & --------------124 Ag 'i' X As &--------------95 As 'i' X At & ______________ 50 At 'i' X As &______________ 54 Am 'i' X Ag &---·--------123 Ag 'i' X Am &------------117 Am 'i' X At & ------------· 69 At 'i' X Am &*---------­Ag 'i' X At & -------------· 59 At 'i' X Ag & --------------41 66 33 32 56 72 71 37 72 86 59 77 13 35 31 47 29 74 12 88 16 76 39 50 6 77 26 15 90 32 88 24 *This cross was partially contaminated with mold. Progeny data were taken from mold-free vials. Studies in the Genetics of Drosophila An examination of Table 1 discloses the fact that in many of these intraspecific crosses, namely, As X At, As X A, A X At, Ag X At, Ag X A, Ag X Am, and As X Am, the percentage of females producing progeny differs in the reciprocal crosses. The differences in the first three of the crosses mentioned may not be real; however, in the last four cases the differences are certainly significant. The percentage differences in the reciprocal crosses of Ag X A, Ag X Am, and As X Am are evidently caused by sexual isolating mechanisms since the variations in the per­centage of females fertilized are in the same direction and of the same general magnitude as the differences in the percentage of females pro­ducing offspring. However, in the crosses As X A, A X At, and Ag X At, if the differences are real, they must be attributed mainly to isolating mechanisms which are effective after insemination has taken place. This is also the case in the crosses between the As and Ag strains since, in the reciprocal crosses, practically the same percentage of females pro­duce progeny although more of the females are inseminated if the As strain is used as the male parent. The cross As X At produces a condi­tion intermediate between the two mentioned cases; i.e., both types of isolating mechanisms are effective to about an equal degree. It is worthy of note that there is no apparent correlation between the strain used as the female parent and the fertility of the cross. For example, when the Ag strain is used as the female parent in crosses to A or Am, the number of progeny produced is reduced in comparison with crosses using this strain as the male parent; however, the cross Ag~ X At 'i> % 'i> Cj> lnsem-No. Pairs Producing %'i>'i> inated Showing Nature of Cross Tested Progeny Inseminated Reaction Mass A'i>X AJ _________________. 79 (3) 118 76 67 Mc Cj> X Mc J -----------­ 119 4 100 32 (112) A'i> x Mc _______________ (20) 115 49 66 25 73 93 (74) Mc 'i> X A J ---------------124 2 AMc 'i> X A J ____________ 113 33 86 29 (55) A Cj> X AMc J ------------Males Sterile AMc 'i> X Mc J ----------153 75 97 28 (32) Mc 'i> X AMc J ---------Males Sterile McA 'i> X A J ____________ 108 37 75 22.5 (40) A 'i> X McA J ____________ 108 0 51 40 (45) (26)McA 'i> X Mc J ---------135 77 97 19 (84) Mc 'i> X McA J ---------110 0 86 89 AMc 'i> X AMc J ______ Males Sterile McA 'i> X McA J ______ 59 0 59 85 (33) A Mc 'i> X Mc J _________ 89 0 79 62 (66) McA 'i> X AMc J ______ Males Sterile The University of Texas Publication of non-fertile, inseminated females on which the percentage figures are based. The male hybrids produced from the cross arizonensis female to mo.fa­vensis male are sterile (Crow 1941). Smears of the testes of these males reveal that mature and motile spermatozoa are never formed. However, Patterson (1946) has demonstrated that these males copulate with females of either parent species and that the presence of the sperm-free semen in the vagina causes a typical reaction mass. Another graphic demonstra­tion of this same behavior of sterile hybrid males was uncovered when males from the reciprocal of the above-mentioned cross were mated to mojavensis females, i.e., the cross Mc <;? X McA ~ . Among the ninety­seven suitable dissections of the Mc females, it was found that forty­nine of them contained sperm somewhere in their reproductive tract, usually in the ventral receptacle. However, thirty-five cases were found in which the vagina of the female was swollen and granular (typical of the reaction mass of the insemination reaction), but a careful inspection of all parts of the reproductive tract of each female failed to reveal the presence of a single sperm. In order to determine if the McA males which were mated to these particular females were producing motile sperm, testes from fifteen of them were smeared and examined. It was found that thirteen of the males contained no motile sperm in their testes. Of the other two males, only a single small group of motile sperm was,found in the one, while in the other there was an abundance of motile sperm. In. this last case transfer of sperm during copulation had evi­dently not been successful. In this connection, it was noted, during dis­section of the females of this particular cross, that in most of the cases where females contain sperm, only a relatively small number were present. Therefore, in studies of sexual isolation where sterile males may be involved, the presence or absence of sperm in the female is not a reliable cri~erion of whether copulation has taken place. For example, in the cross being discussed, Crow (1942), prior to the discovery of the insem­ination reaction, reports that 56% of Md females were fertilized; while these tests, using the Mc strain, indicate that 86 % were inseminated. If, in the present experiments, the insemination reaction is disregarded as evidence of previous copulation, a figure of 51 % which is comparable with Crow's data would be obtained. · Actually the sexual isolation is probably even less than reported because the reaction mass caused by these sterile males had probably cleared "in some cases and these females were recorded as not inseminated. Although none of the pair matings invplying McA males produced progeny in the fourteen-day period, all of these crosses will produce some offspring if mass matings are used. The fact that the majority of these hybrid males are fertile indicates the presence of still other mechanisms operating to reduce the production of progeny. An examination of the results tabulated in Table 4 indicates that there is but a weak correlation between the variations in the percentage of Studies in the Genetics of Drosophila females inseminated and the variations in the percentage of females pro­ducing progeny. Superficially this would apparently provide evidence that sexual isolating mechanisms play a relatively minor role in limiting the fertility of these crosses. However, the possibility that sexual isolating factors may be involved is evident from the results of crosses in which a large percentage of the females which were inseminated contained a reac­tion mass (see the last column of Table 4). Patterson (unpublished) has found that in the cross Mc.....,... mbpalustris 5 + * ;<;? ­sperm ­ ''' 0 sperm ­ 5 +;'i? -sperm ­ 5 ++;'i?-sperm ­ Cop. sperm+ Cop. sperm* J + ;'i? ­sperm ­ 0 sperm ­ J + ;'i?­sperm ­ J + -; 'i? ­sperm- I "ti ~ sperm ­ sperm- sperm ­ sperm- sperm ­ sperm- sperm ·­ sperm+ sperm ­ sperm ­ O".,........ trans versa J +;'i? -sperm ­ 0 sperm ­ 5 +;'i? -sperm ­ J + ­;'i? ­sperm­ 0 sperm ­ 0 sperm ­ 0 sperm ­ 0 sperm- Cop. sperm+ 5 +;'i? -sperm ­ (') ~ .,....... munda 0 0 J + ;'i?­ 5 + ­ ;'i? ­ 5 + ­ ;'i?­ . Cop. 5 + ;'i?-* 0 5 ++ ;'i?-* Cop. 0 ;;s sperm ­ sperm ­ sperm ­ sperm- sperm ­ sperm+ sperm- sperm- sperm ­ sperm+ 0 =No interest on the part of either sex. 'i? -= Females indifferent to advance of males. 5 + = Males showed interest, began courting. 'i? -* = Females antagonistic to advance of males. 5 + + = Males very interested, persistent courting. sperm -=No sperm in female reproductive tract. 5 + * = Males attempted to copulate without success. sperm + -= Sperm present and active in female tract. J + -= Males showed doubtful interest. sperm * = Sperm present but inactive in female tract. Studies in the Genetics of Drosophila suboccidentalis C;1 X occidentalis ii gave the highest percentage of fer­tility encountered in any of the interspecific crosses which were made. Seventy-nine out of a possible one hundred vials were fertile and produced an average of 85 flies per vial. These hybrids also are one hundred per cent fertile, and the F2 hybrids show fertility. occidentalis C;1 X palustris ii yielded six female and four male hybrids from mass matings. The male hybrids were completely sterile but the female hybrids were fertile to both parental types of males. occidentalis C;1 X subpalustris ii produced eighty-four hybrids from mass matings; only the female hybrids showed fertility. innubila C;1 X occidentalis ii hybridized in pair matings only after many repetitions of the standard pair-mating test. One vial eventually yielded three offspring, one female and two males. These hybrids were tested by inbreeding a.nd backcrossing and were completely sterile. However, this test is insufficient. suboccidentalis t;? X palustris ii produced fertile hybrid females and sterile hybrid males. The female hybrids .showed fertility only when crossed to palustris males and were sterile when crossed to suboccidentalis males. palustris C;1 X subpalustris ii . Nineteen out of the first one hundred counted vials produced offspring for an average of 35.1 flies per fertile culture. These hybrids were fertile in all crosses made. The F2 hybrids also showed some fertility. subpalustris C;1 X palustris ii . This cross was sixteen per cent fertile and yielded 6.3 flies per fertile vial. The hybrids were fertile on inbreed­ing and on backcrossing to all parent types. The F 2 hybrids also were fertile. palustris C;1 X innubila ii yielded forty-nine hybrids from mass matings. This cross did not go in pair matings. The hybrids were fertile when inbred or when crossed to innubila males or females, but the fertility was very slight in each case. Within the limits of this experiment the hybrids were incompatible with palustris. transversa C;1 X innubila ii produced hybrids which were completely sterile in all tests. The intraspecific crosses subquinaria (1210.5) 9 X subquinaria (1202.3) ii and subquinaria (1202.3) C;1 X subquinaria (1210.5) ii showed two and three per cent fertility respectively. Oniy female hybrids from the last cross showed fertility in backcrosses, and all hybrids were sterile when inbred. Out of the ninety possible cross-matings eighteen showed fertility and seventy-two produced no hybrids. The University of Texas Publication CYTOLOGICAL RESULTS Larval ganglionic tissue from both the parent stocks and the F1 hybrids were examined with the use of ·the acetocarmine smear technique. The results of the study are in complete agreement with those of Wharton (1943). The descriptions given below summarize the metaphase con­figurations which were observed in the several stocks and hybrid classes. Drosophila quinaria has two V-shaped autosomes, two J-autosomes, two dots, and long rod-shaped X-chromosomes with proximal constric­tions. The Y-chromosome is a J-shaped element with a constriction approximately at the center of the long arm. Drosophila subquinaria has eight short rods, long rod-shaped X-chromo­somes, and two dots. The Y-chromosome is a small V. Four species, occidentalis, transversa, palustris, and subpalustris, have identical metaphase configurations, showing eight medium-length r.ods, two long rod-shaped sex-chromosomes, and two dots. The male and female brain smears showed the same configuration. Drosophila innubila males and females exhibit eight medium-length rods, long sex chromosomes with proximal constrictions, and two dots. Drosophila suboccidentalis has eight medium-length rods, long rod­shaped X-chromosomes with proximal constrictions, and two dots. The Y-chromosome of the male is a long simple rod without the proximal constriction which is seen in the X-chromosomes. Drosophila munda males and females have four large V-shaped auto­somes, two rod-like sex chromosomes, and two dots. Figure 1 shows the metaphase configurations of nine female hybrids and of the corresponding male and female parents; the hybrid chromosome configurations are as follows : quinaria-subquinaria hybrids. The female hybrids had one large V-shaped autosome, a J-shaped autosome, four short rods, two dots and two long rods, one of which had a proximal constriction. The V, J, one dot, and the long rod with the constriction were contributed by the quinaria parent; the remaining elements were from subquinaria. subquinaria-munda hybrids. The configuration of the female hybrid showed two large V's, four medium-length rods, two long rods, and two dots. The munda parent had contributed the two V's, one long rod, and one dot. The four medium-length rods, one long rod, and one dot were from the subquinaria parent. munda-occidentalis hybrids. The metaphase configuration of these hybrids showed the two large V-shaped autosomes from munda, four medium-length rod-shaped chromosomes from occidentalis, and two long­rods and two dots contributed by both parents. munda-suboccidentalis hybrids. These hybrids had metaphase plates like those described immediately above except for a constriction at the Studies in the Genetics of Drosophila proximal end of one of the long rods. This represents the X-chromosome from suboccidentalis. suboccidentalis-occidentalis hybrids. The larval brain smears of female hybrids showed four pair of medium-length rod-like autosomes, two long rods (one with a proximal constriction), and two dots. The larvae were selected with particular care since only female larvae would be diagnostic. The hybrid male larvae would be indistinguishable from male larvae of suboccidentalis since they both would have the same metaphase configura­tion if suboccidentalis were used as the female parent. If occidentalis were used as the female parent the male hybrid larvae would have a metaphase configuration like that of occidentalis. In the reciprocal crosses involving occidental is and suboccidentalis the female hybrid larvae, how­ever, would have an X-chromosome from each parent. Since the X-chromo­somes of the two species can be distinguished from each other, that of suboccidentalis having a proximal constriction, it is possible to distinguish hybrid female larvae from female larvae of either parent stock by examin­ing the brain smears (see Figure 1) . transversa-innubila hybrids. Male and female larvae showed in the metaphase plates four pairs of medium-length rods, two long rod-like X-chromosomes (one with a proximal constriction), and two dots. The long rod with the proximal constriction was from the male parent, innubila. palustris-innubila hybrids. Male and female hybrid larvae showed the same configuration as that described above for the transversa-innubila hybrids. The long rod with the constriction again came from innubila. The metaphase configuration of hybrid larvae from crosses involving occidentalis, palustris and subpalustris are not useful in checking hybridi­zation since these three species have identical metaphase plates. Hybrid larval salivary gland chromosomes were used instead. These showed in each case numerous inversions not found in the salivary chromosomes of the parent species. The metaphase configurations of hybrid larvae from the crosses sub­occidentalis X palustris, subpalustris X munda, and innubila X occi­dentalis are not given since too few larvae were produced to obtain both brain smears and adult flies. The adult flies proved by their appearance to be hybrid. The salivary gland chromosomes of the species studied in this group are not well adapted to exacting analysis since they tend to remain knotted into small masses when the smears are made. Inversions were found within all the stocks used in this investigation. This made the analysis of the inversions found in the hybrids so difficult that it will be left for later work. It should be noted, however, that many more complicated inversions were found in the hybrids than in the parent stocks. The sali­vary gland chromosomes of F1 larvae from the crosses between palustris and subpalustris are a tangled mass of poorly synapsed strands with many complex inversions in each chromosome. This is in agreement with The University of Texas Publication the observations :reported by Spencer (Hl40). In all .crosses the F1 hybrid salivary chromosomes were similar to those described above in that they revealed numerous inversions and very poor synapsis. It is not possible at this time to state whether these complicated configurations are the result of overlapping, tandem, or included inversions; on the other hand, it is obvious that none is simple in nature. FEMALE PARENT ~~ .. \( H QUINARIA )) .. (( u MUNDA )) .. (( ii MUNDA )) .. (( u MUNDA ,, ~ ,, .. ~ ~u~ OC C IDE NTA LIS ''I/ ~··" u TRANSVERSA ,,// ~-·~ u f'ALVSTRIS FIGURE I MALE PARENT ~~ c;;::?.. Q 0 c:::::::> "07 ~if'~ SUBQUINARIA ~I/ ~··~ ~~ SUBQUINARIA ~\ // c7. 0 ~ 0 (/'~~~ OCCIDENTALIS ~fl/ 00 ~ ~~~~ SUBOCCIDEl'HALI S ~~I/ ~ "" ~ (f/7~~~ SU BO CC I DEN TA LIS \.\fl/ d', ~ 00 (f;7 M INNUBILA \.\ ~ ~~1~ INNUBILA FEMALE HYBRID ~~ ..~ u )) /J ··~( u )) .. (( '~ )'O• /I( (j' ~ H <~I/ ?". "' ~ ~H~ ~f/ ~o·~ I~ ~'~ _,,, .. (;1 rn"' ~ Studies in the Genetics of Drosophila DESCRIPTION OF HYBRIDS Some hybrids obtained in this investigation were intermediate in appear­ance between the parent forms, others were mixtures. ·Only those that were most striking in appearance will be dealt with here. transversa-innubila hybrids. Th~ abdomen of transversa is yellow with dark spots on each tergite. The eye color is an orange-red and the wings have clouded cross veins. Drosophila innubila has a yellowish-grey abdo­men with tergites four, five, and six a dark brown. Its eye color is a dull red and the wings are not clouded. The hybrids are larger than either of the parent species but have a body shape more like that of transversa with yellowish-grey abdomens and with dark spots on each tergite. The eye color is intermediate between an orange-red and a dull red. The wings have .very slight clouds on the cross veins, lighter than those of ,transversa but darker than those of innubila. palustris-innubila hybrids. Drosophila palustris has dark clouds on the cross veins and at the apex of the second, third~ and fourth longitudinal veins. In innubila the wings are clear. The abdomen of palustris is a dark brow~ish-grey with three broad yellow stripes running lengthwise. The abdomen of innubila is a yellowish-grey with narrow bands interrupted in the mid-dorsal line on tergites two and three; tergites four, five, and six are dark brown. The eye color of palustris is red, and that of innubila is a dull red. The .eye color of the hybrid is intermediate, the wings are clear, and the abdomen is more lil, X occidentalis J produced hybrids that were fer­tile when inbred. The male hybrids, however, were sterile when back­crossed to either parent type. The cross munda <;> X suboccidentalis J seems to be another case of unbalance similar to the preceding one, except that the hybrid males were fertile when backcrossed to munda females. The cross of occidentalis females to palustris or subpalustris males pro­duced fertile female and sterile male hybrids. In the cross suboeciden­talis <;> X palustris J the female hybrids were slightly fertile if back­crossed to the palustris parent type. The male hybrids were completely sterile. The cross transversa <;> X innubila J gave sterile hybrids, as shown by the tests of five hundred twenty-eight hybrids. The cross palustris <;> X innubila J produced hybrids that were slightly fertile when inbred and when backcrossed to innubila in both directions, but completely sterile when backcrossed to palustris in both directions. It is possible that, since only forty-nine hybrids were obtained, the crosses were not sufficiently large to show any slight fertility that might exist. The crosses occidentalis <;> X suboccidentalis J and suboccidentalis <;> X occidentalis J were 35% and 79% fertile, respectively. The hybrids from both crosses showed heterosis and were 100% fertile when inbred. The F2 hybrids were also fertile. Table 4 indicates that there is little if any sexual isolation present between these species. In the first cross there is some gametic mortality which reduces the fertility of that cross but not sufficiently to make the cross sterile. Morphologically these species are very similar, but cross fertility tests show that they are slightly different. The difference is illustrated by the fact that occidentalis crosses with suboccidentalis, palustris, subpalustris, munda, and innubila, while sub­occidentalis crosses with occidentalis, palustris, and munda only. The crosses palustris <;> X subpalustris J and subpalustris <;> X palus­tris 3 showed 19 o/o and 16% fertility, respectively. This low fertility is not surprising when it is noticed that the controls are quite low, that for palustris being 27 % and for subpalustris 38% . The hybrids from the crosses were 19 % and 82 % fertile, respectively, when inbred; this fer­tility percentage was obtained by testing only twelve pairs of flies, since no others were available. The F 2 hybrids in both cases were fertile. . From an examination of Table 4 it can be seen that there is some sexual isola­tion present but it is not sufficient to separate these forms completely. IV. GAMETIC MORTALITY Gametic mortality was found to be effective in reducing the fertility in crosses between D. virilis <;> and D. americana J (Patterson, Stone, and Griffen 1940). The americana sperm were inactivated and perhaps killed in the reproductive tracts of the virilis females so that only the eggs laid during the first day after copulation were fertilized. In a Studies in the Genetics of Drosophila comparable study in the quinaria group no hybrids were produced from the crosses subpalustris '? X occidentalis o and palustris '? X occiden­talis o, although copulation was observed. Upon the dissection of the females it was found that the sperm in each case were inactivated in the reproductive tracts of the females. In the cross occidentalis '? X subocci­dentalis o the first eggs which were laid after copulation produced larvae, but when the females were dissected at the end of three days the unused sperm in the genital tracts were immotile. In the reciprocal cross the sperm were still motile at the end of the three-day period. SUMMARY 1. Only fifteen of a possible eighty-eight interspecific crosses tested have proven fertile. The most important cause of this lack of cross­fertility is sexual isolation. 2. Seven out of the fifteen interspecific hybrid combinations were sterile on inbreeding, and only twenty-seven of the sixty possible back­crosses were fertile. Therefore, the interspecific hybrids usually had a much reduced reproductive efficiency. 3. Two cases are noteworthy if verified by further investigation. These are munda '? X occidentalis o, and perhaps munda '? ·X suboccid-entalis o. In these two cases the hybrid males did not fertilize the parent ·type females, although they were fertile with their hybrid sibs. These cases represent the origin of a cross sterility due to a new combination of genes present in these hybrid males. 4. The metaphase chromosome configurations of these species vary quite widely and fusion must have been an important factor in the change in chromosome number (e.g., occidentalis and munda). The chromosome patterns, including constrictions, are the properties of the chromosomes themselves for they persist in the hybrids. In addition to fusions, there are many inversions present in this species group. 5. Numerous phenotypic differences exist between these species. The hybrids show that both blending and dominant relations exist between the several gene systems which determine these phenotypes. BIBLIOGRAPHY Crow, J. F. 1942. Cross Fertility and Isolating Mechanisms in the Drosophila Mulleri Group. Univ. Tex. Publ., 4228 :53-67. Dobzhansky, Th., and P. Ch. Koller. 1938. An experimental study of sexual isolation in Drosophila. B. Z., 58:589-607. Griffen, A. B. 1942. Relationships in the melanica species group. Univ. Tex. Puhl., 4228 :68-73. Mainland, G. B. 1942. Genetic relationships in the Drosophila funebris group. Univ. Tex. Publ., 4228:74-112. Patterson, J. T. 1942a. Drosophila and Speciation. Science, 95:153-159. Patterson, J. T. 1942b. Interspecific hybridization in the genus Drosophila. Univ. Tex. Publ., 4228:7-15. The University of Texas Publication Patterson, J. T. 1942c. Isolating mechanisms in the genus Drosophila. Biol. Sympo­sium, VI :271-287. Patterson, J. T. 1943. The Drosophilidae of the southwest. Univ. Tex. Puhl., 4313: 7-216. Patterson, J. T., and J. F. Crow. 1940. Hybridization in the mulleri group of Dro­sophila. Univ. Tex. Puhl., 4032:251-256. Patterson, J. T., Wilson Stone, and A. B. Griffen. 1940. Evolution of the virilis group in Drosophila. Univ. Tex. Puhl., 4032:218-250. Patterson, J. T., Wilson Stone, and A. B. Griffen. 1942. Genetic and cytological analysis of the virilis species group. Univ. Tex. Publ., 4228: 162-200. Patterson, J. T., and R. P. Wagner. 1943. Geographical distribution of species of the genus Drosophila in the United States and Mexico. Univ. Tex. Puhl., 4313: 217-281. ' Spencer, Warren P. 1940. Subspecies, hybrids and speciation in Drosophila hydei and Drosophila virilis. Amer. Nat., 74:157-179. Spencer, Warren P. 1942. New species in the quinaria group of the subgenus Dro" sophila. Univ. Tex. Puhl., 4213:55-66. Stalker, !I. D. 194.2. Sexual isolation studies in the species complex Drosophila virilis. Gen~tics, 27:238-257. · Stalker, H. D., and Warren P. Spencer. 1939. Four new species of Drosophila with notes . on the funebris group. Ann. Ent. Soc. Amer., 32:105-112. Sturtevant, A. H. 1921. The North American species of Drosophila. Carnegie Insti­tute of Washington Puhl., 301. Sturtevant, A. H. ' 1940. Genetic data on Drosophila afjinis with a discussion of the relationships in the subgenus· Sophophora. Genetics, 25 :337-353. Sturtevant, A. H. 1942. The classification of the genus Drosophila, with descrip­tion~ pf nine new species. Univ. Tex. Publ., 4213 : 6-51. Sturtevant, A. H., and Th. Dobzhansky. 1936. Observations on the species related to Dros1/phila afjinis, with descriptions of seven new forms. Amer. Nat., 70 :574-584. Wharton,·Linda T. 1942. Analysis of the repleta group of Drosophila. Univ. Tex. Puhl., 4228 :23-52. W:!iarton, Linda T. 1943. Analysis of the metaphase and salivary chromosome morphology within the genus Drosophila. Univ. Tex. Puhl., 4313 :282-319. . VIII. THE SPECIES RELATIONSHIPS IN THE VIRILIS GROUP WILSON S. STONE AND J. T. PATTERSON Sturtevant (1942) raised a question as to the species status of the five members of the virilis species group known at that time. He classifies Drosophila virilis as the type species with americana, texana, nova­mexicana, and montana as additional subspecies. Patterson, Stone, and Griffen (1942) disagreed with Sturtevant on this classification and re­garded all five members of the group as separate species. Additional evidence leads us to propose certain revisions of this group. .·· .· There is_an overlap zone where americana and texana bccur, which we know extends through Tennessee, Arkansas, and into Texas (Phtterson and Wagner 1943). In the fall of 1946, Mr. Marshall R. Wheeler and Mr. William K. Baker collected a number of females of the virilis group at Morrilton, Arkansas. These females were separated from the males as soon as they were trapped, and each female was placed in a separate container to establish a stock. In case the female . was not fertile, she was mated to males of a virilis stock with the following markers in the major autosomes: ·Rounded (R, 2 chromosome), cinnebar (en, 3), plexus (px, 4), peach (pe, 5). In case the captured female was fertile:. the F1 females were mated to the R en Px pe males. The F 1 males fr~m each cross to the marker .virilis stock were testcrossed individually to en px pe females. This allows us to determine in the next generation the segre­gation of the X chromosome and each of the major autosomes. ·The small sixth chromosome was not followed in these experiments. The effective­ -ness of this test depends on the chromosome architecture of these forms (see Wharton 1943, for illustrations of the metaphase chromosome. con­'fig1:1rations of the known species of this group). Drosophila virilis has the primitive chromosome configuration for Drosophila, i.e., five pairs of rods and a pair of dots. In this species the X and Y and the four major autosome pairs are rod-shaped chromosomes, indistinguishable in the ordinary metaphase configuration. In americana Stalker (1940), and Patterson, Stone, and Griffen (1940) reported that females have two pairs of V-shaped chromosomes and a pair of rods, while the males have three V-chromosomes and four rods. The pair of V's common to both male and female are fused 2 and 3 chromosomes. The other pair of V'.s in the female are the fused X and 4 chromosomes, while the male has one V representing the fused X and 4 chromosomes and two :free rods which are the unfused Y and the other 4 chromosome. Therefore, americana is a species with a multiple Y complex (the Y and free 4 chromosome) . Patterson, Stone, and Griffen (1940) described the ordinary meta­phase chromosome pattern of texana as one pair of autosomal V's, three pairs of rods (one of these is the X-Y pair), and a pair of dots. In that publication the autosomal fusion was described as linking chromosomes The University of Texas Publication 3 and 4 and was so referred to in later publications. Further investi­gation has revealed that this is not correct for a number of texana stocks, and that the autosomal fusion involves chromosomes 2 and 3 as in americana. Segregation tests with texana (stock number listed) were used to determine this association, as follows : The texana stock used was from Georgetown, Texas (Stock 84.7). P 1 84.7 ~ X R(2) cn(3) px (4) pe(5) ~. F1 84.7/Rcnpxpe ~ X cnpxpe~. F 2 R en px pe = 48; + = 54; R en px = 40; pe = 51; R en = 51; px pe = 46; R en pe = 45; px = 41. Similar results were obtained with strains of texana from New Or­leans, Louisiana (1128.10), from Caddo Lake, Texas (1593 B.6a), and from Lake McKethan, Florida (1148.9). In these cases also R and en segregated together, indicating that chromosomes 2 and 3 are fused. We desired to determine whether americana and texana hybridized in nature. The segregation test outlined above allows us to check this pos­sibility. Stocks of americana, where the X and 4 chromosomes are fused, will show this linkage : If americana females are crossed to R (2) en (3) px(4) pe(5) males and the F1 males testcrossed to cn(3) px(4) pe(5) females, all the F 2 females will be non-plexus ( 4), and the F 2 males will be plexus(4) for that reason. As the autosomes 2 and 3 are fused, R(2) and cn(3) will segregate together. On the other hand, as texana does not have this X-4 fusion, plexus ( 4) will occur at random in both sexes. In case a hybrid female between texana and americana is tested, one half the F1 males will have the texana linkages and the other half the ameri­cana linkages. Such hybrids were present in the collection from Morrilton, Arkansas. The stock 1586 (8) .3e from a single fertile female proved to be hybrid. The F 2 progeny from test of individual F1 males were. (1) males= 4 Ren px pe; 4 Ren px; 5 px pe; 6 px. females= 5 Ren pe; 4 Ren; 6 pe; 6+ (americana linkage). (2) males= 3 Ren px pe; 3 Ren px; 2 px pe; 6 px; 3 pe; 2+. females = 1 R en px pe; 5 R en px; 4 px pe; 5 px; 1 R en pe; 2 Ren; 5 pe; 5-f-(texana linkage). (3) males = 5 R en px pe; 4 R en px; 7 px pe; 5 px. · females= 5R en pe; 4 Ren; 4 pe; 3+ (amerioana linkage). I One test was made with a female which was virgin when captured 1591 (12) la. She was mated directly to Ren px pe males. The F 2 progeny were: (1) males= 7 Ren px pe; 6 Ren px; 9 px pe; 7 px. females= 6 Ren pe; 7 Ren; 6 pe; 9+ (americana linkage). (2) males= 2 Ren px; 1 px pe; 2 px; 2 Ren pe; 1 pe; 1+. females = 1 R en px pe; 2 R en px; 1 px pe; 2 px; 1 R en pe; 2 Ren; 3+ (texana linkage) . Studies in the Genetics of Drosophila (3) males = 8 R en px pe; 9 R en px; 11 px pe; 11 px; 7 R en pe; 9 Ren; 10 pe; 13+. females= 8 Ren px pe; 7 Ren px; 10 px pe; 12 px; 8 R en pe; 9 Ren; 9 pe; 11+ (texana linkage). This female proved to be heterozygous for texana (free X and 4 chromo­somes) and americana (fused X and 4 chromosomes). Therefore she was a hybrid captured in nature. The first stock tested may have been estab­lished either from a hybrid female, or from hybridization in nature. Several other stocks tested proved to be only americana or texana in so far as linkage tests could show. We could not distinguish several pos­sible types of back cross hybrids by this method, so we cannot determine the history of these strains. There is very little comment necessary on these results. All the ameri­cana, texana, and hybrid stocks tested showed linkage between chromo­somes 2 and 3, so this probably is the autosomal fusion present in these two forms. Strains of texana and americana cross readily with each other in the laboratory. These tests show that they hybridize under natural conditions in the overlap zone of their respective distribution areas. We do not have quantitative data on the extent of such hybridi­zation. There exists certain cytological information on the hybridization in the overlap zone between texana and americana. These two forms differ in their gene sequence, due to inversions in the X and 4 chromo­somes. Warters (1'944) demonstrated that certain stocks, which had only texana fusions, had the americana gene order in one or the other chromosome where their gene orders differ. These stocks were from the overlap zone of these two forms, and certain of them must have come from natural hybridization. These two forms are, therefore, best classified as subspecies, which in the main occupy different areas of distribution and cross where these areas join or overlap. These forms are sufficiently alike morphologically that it is almost impossible to distinguish on that basis. We cannot determine the status of D. novamexicana until a group of representative stocks are available for testing. A discussion of the origin and relations of novamexicana must be deferred until these relationships are tested. We therefore suggest that the known relationships and differences can justify only subspecific rank, and these two forms should be designated D. americana americana and D. americana texana. The revision of the virilis group would then be as follows: I. D. virilis Sturtevant II. D. montana Patterson and Wheeler III. D. lacicola Patterson IV. (a) D. americana americana Spencer (b) D. americana texana Patterson, Stone, and Griffen V. D. novamexicana Patterson-undetermined relationship to tex­ana and americano The University of Texas Publication This group is composed of four distinct species which can be separated on the basis of their morphology, and which genetic tests show to be separate species. The status of the fifth member remains undetermined. REFERENCES Patterson, J. T., Wilson S. Stone, and A. B. Griffen. 1940. Evolution of the virilis group in Drosophila; Univ. Tex. Puhl., 4032 :218-250. Patterson, J. T., Wilson S. Stone, and A. B. Griffen. 1942. Genetic and cytological analysis of the virilis species group. Univ. Tex. Puhl., 4228: 162-200. Patterson, J. T., and R. P. Wagner. 1943. Geographical distribution of species of the genus Drosophila in the United States and .Mexico. Univ. Tex. Puhl., 43i3: 217-281. Stalker, Harrison D. 1940. Chromosome homologies in two subspecies of Drosophila virilis. Proc. Nat. Acad. Sci., 26:575--578. Sturtevant, A. H. 1942. The classification of the genus Drosophila, with descrip­tions of nine new species. Univ. Tex. Puhl., 4213:5-51. Warters, Mary. 1944. Chromosomal aberrations in wild populations of Drosophila. Univ. Tex. Puhl., 4445:129-174. ·· Wharton, Linda T. 1943. Analysis of the metaphase and salivary chromosome mor­phology within the genus Drosophila. Univ. Tex. Puhl., 4313: 282-319. IX. GENE REPLACEMENT IN THE VIRILIS GROUP WILSON S. STONE This paper reports preliminary investigations of the ability of a group of genes (one or more chromosomes) from one species to replace eff ec­tively the equivalent group of genes in a second species. This work also includes some measurements of the effect of the chromosomal differences that exist between these species on the fertility of their hybrids. MATERIALS AND METHODS These tests were made with three members of the virilis group, Dro­sophil,a virilis Sturtevant, Drosophila americana americana Spencer, and Drosophila americana texana Patterson, Stone, and Griffen. The species relationships have been discussed elsewhere in this volume. The virilis stocks used were Mexico City (1341.13) and two marked stocks: (1) Rounded (R = 2 chromosome); cardinal (cd = 4 chromo­some)' and peach (pe = 5 chromosome); (2) a sto(!k of cd pe for test­crossing heterozygotes. The Anderson, Indiana, stock of americana was used. Three strains of texana were tested, the one from Georgetown, Texas (84.7); a second from Newton, Texas (841.10); and a third from New Orleans, Louisiana (1128.10). The experimental procedure consisted of crosses between texana (or americana) and R cd pe. The F 1 hybrids and F2 segregants were crossed to cd pe. As an alternative, certain of the F2 combinations were crossed to the unrelated Mexico City stock so that the effect of stock lethals could be minimized. Egg counts and hatch counts were made in certain crosses to determine the effect of the recombination on reproductive efficiency. The differences in chromosome architecture that exist between these species must be considered in analyzing the results of these tests. Dro­sophila virilis has the primitive chromosome complex for Drosophila of five pairs of rods and a pair of dots. This species will be used as the base in describing differences that occur between these forms. In texana and americana chromosomes 2 and 3 are fused to form a large autosomal V in place of the two rods in virilis. In addition there is a medium size inversion in 2, involving about one-third of the chromosome. Chromo­some 5 in americana (and usually in texana) differs from virilis by a small inversion near the free end of the chromosome and a large inversion nearer the centromere, involving about two-thirds of the chromosome. The X in texana has a double inversion in the proximal half and ameri­cana has an additional overlapping inversion extending further toward the free end of the chromosome. Finally, the X and 4 chromosomes are fused in americana although the Y and 4 are not, so that americana has a complex Y chromosome. The University of Texas Publication In making reciprocal crosses, the direction of the initial cross deter­mines the X and Y chromosomes of the F1 and F 2 hybrids; for example, the X -4 fusion and lack of Y -4 fusion leads to several differences in reciprocal crosses involving americana. (See Patterson, Stone, and Griffen (1940) for Griffen's chromosome maps and analysis of the inversions.) RESULTS Hybrids between virilis and texana or americana are obtainable if suffi­cient crosses are made even though the egg hatch is quite low and sexual isolation decreases the amount of cross fertilization. These experiments have not indicated a great differential in viability between combinations, for if the F1 males from matings of americana or texana to R cd pe are testcrossed to cd pe, the different possible recombinations occur with nearly equal frequencies. Also, the fertile F 2 males of the different com­binations produced the different possible F 3 classes in the expected ratios. Therefore, the viability of the different recombina,tion classes does not seem to be markedly changed by the differences in association. It should be noted that all the tests in Tables r and 2 hold the foreign chromo­ . some(s) in the heterozygous condition, and, therefore, at minimum possible selective disadvantage. Table 1 shows that all female F 2 recombinations from all reciprocal crosses between americana or texana and virilis are fertile, i.e., produced offspring. Even though some of the tests involve small numbers, the number fertile (F) is always larger than the number sterile (S). In fact, in most cases the fertility of these F 2 recombinations seemed normal. In contrast to these results, the data for male recombinations were not uniform. In the crosses where americana and texana were used as the P1 female all F 2 male combinations were fertile, with little, if any, reduc­tion below normal. In reciprocal P 1 crosses, the F 2 males showed that the complementary Y-autosome combination previously reported was necessary for fertility (Patterson, Stone, and Griffen 1940, 1942). If the texana or americana Y-chromosome is present, the 2-3 chromosome fusion and the 5 chromosome from texana or americana must also be present for normal fertility. Occasionally combinations without all complementary chromosomes produce offspring. In Table 1, 15 out of 17 4 without 2-3 chromosome of texana and 1 out of 153 without either 2-3 or 5 chromo­some were fertile. In Table 1, the semisterile males which produced offspring are starred. However, none of these males produced more than one or two progeny. The egg hatch which will be discussed later indicates that only a few sperm are functional. Certain of these hybrid male combinations have been dissected and their descriptions follow: In all cases, the sperm pump appeared normal but the paragonia appeared to be immature. In the classes without the 5 chromosome of texana or americana (the 2-3 fusion might or might not be present) the testes were small with many large abnormal cells and cell debris; usually TABLE 1 Fertility Tests of F2 Recombinations texana or americana autosomes ; see cross for X chromosome 2-3, 4, 5 . 2-3, 4 4, 5 4 2-3, 5 2-3 5 0 Phenotype of F2 = pe R R pe cd cd pe R cd R cd pe + F s F s F s F s F s F s F s F s R cd pe 'i? X Anderson & ; (1) Fi & X cd pe 'i?; F, 'i? X cd pe & Anderson 'i? X R cd pe & ; (2) Fi & X cd pe 'i? ; F2 'i? X cd pe & Anderson 'i? X R cd pe & ; (3) Fi & x~cd pe 'i? ; F" & X cd pe 'i? R cd pc 'i? X New Orleans & (4) Fi & X cd pe 'i?; F2 'i? X cd pe & R cd pe 'i? X Newton & ; (5) F, & X cd pe 'i? ; F, 'i? X cd pe & R cd pe 'i? X Georgetown & (6) Fi & X cd pe 'i? ; F, 'i? X cd pe & (7) Total F, 'i? testcrossed R cd pe 'i? X New Orleans & (8) Fi 8 X cd pe 'i? ; F, 8 X cd pe 'i? R cd pe 'i? X Newton 8 ; (9) Fi8 X cd pe 'i?; F, 8 X cd pe 'i? R cd pe 'i? X Georgetown 8 (10) Fi 8 X cd pe 'i? ; F2 8 X cd pe 'i? (11) Total, 8 = texana Y testcrossed New Orleans 'i? X R cd pe 8 (12) Fi 8 X cd pe 'i?; F, 'i? X cd pe 8 New Orleans 'i? X R cd pe 8 (13) Fi 8 X cd pe 'i? ; F2 8 X cd pe 'i? 15 92 16 32 22 70 26 24 9 2 18 0 3 4 7 4 2 8 12 56 ID ~ 17 50 0 0 0 2 4 2 1 6 9 ~ 23 34 11 23 62 ~ ~ 22 81 3* 6* 0 0 5 1 0 1 2 27 27 16 17 36 18 w 26 104 0 0 0 7 20 1 10 41 10 5 M 2 2 ~ 0 7 13 9 14 55 11 27 25 0 54 36 6 2.5 5 1 57 14 0 68 9* 70 0 106 66 7 18 6 36 7 18 1 15 2 26 2 27 4 20 5 13 5 12 3 31 3 17 0 25 28 1 C'I'.) ~ R.""·{<:) Cl> 61 10 35 io 27 4 ~· ~ ;;:r­(1:> 15 1 26 6 27 0 ~ ~ 10 6 4 ~ 32 6 0 ~ 37 3 3 ~ {<:) ~ ""·~ Cl> 0- 78 11 116 12 130 6 tl 0 0 22 65 5* 1* 33 43 l* 0 38 90 ~ Cl> 0 'ti ;;:r­""· ~ 0 18 0 13 0 25 0 105 6* 89 1 * 153 38 8 16 . 4 6 12 39 4 17 1 23 8 ....... Cl') ~ The University of Texas Publication the sperm bundles were abnormal, with a few long thick strands, but occasionally some strands approached the immature normal in appear­ance although no sperm or motility was observed. In the cases with the 2-3 fusion absent (but 5 present), the testes were more nearly normal in size and shape. The sperm bundles were more often abnormal with beaded sperm but a few were nearly normal and there was some motility. In no case did they appear normal, with the usual masses of motile sperm. Obviously chromosome 5 has much more effect on the fertility of these males (the 1 male classed as R cd pe which produced an offspring may have been a mutation, a misclassified R cd male or have been mated to a non-virgin female). However, the 2-3 fusion is necessary for normal fertility for if it is missing, at best only a few sperm are functional. The egg hatch counts deserve some comment. The eggs were collected from one to ten pairs of flies-usually over five pairs-to reduce the effect of extrinsic factors on fertility. Certain cross.es have not been made and these are left blank. Others where no fertile crosses were obtained are indicated by 0. The egg counts for the semisterile males were usually made from a pair or pairs that had produced at least one larvae from an unknown number of eggs previously laid. In these cases, the number of such offspring (and tests also as only one offspring per pair was obtained) is indicated as (+ 2), (+ 4), and (+ 1) with the number of eggs subsequently counted without further offspring indicated. These males, therefore, produced a very few short-lived functional sperm. There were certain complications due to genes in the related virilis marked stocks R cd pe and cd pe. Some few females were sterile because they produced abnormal eggs without filaments which failed to hatch. This made the hatch low in some crosses, especially affecting those where the number of eggs was low. · Despite these complications the effect of gene replacement and chromo­somal differences on these hybrids can be determined. The 4 chromosome of texana (and 6, which may be present or absent in any one of these combinations) has no major detectable gene rearrangements as compared to virilis. It can replace the virilis 4 at least heterozygous without ill effects. For example, in ·an unlisted cross of R cd pe c;i X Georgetown ~ ; F1 ~ X cd pe; F2 R pe <;> X Mexico~, there :.were 1,359 hatched from 1,411 eggs, or 96.3 % . This is a normal hatch, despite the fact that the texana 4 chromosome is heterozygous. Other similar crosses included in Table 2 numbers 6, 7, 8, (14) indicate that the 4 chromosome does not reduce the egg hatch below the normal if present in either male or female parent in the heterozygous condition. It should be noted that the R cd pe class is in effect a control even though ·the 6 chromosome may be heterozygous. The 5 chromosome of texana or americana which has two inversions in comparison to the virilis 5, does not have much detrimental effect if heterozygous. The results with the semi-sterile males, of course are the exception, but this is the Y-autosome complementary action, not the effect TABLE 2 E_o;,;g Counts with F, Recombinations t i!> X cd pe 'i'; F, 'i' X cd pe i!> 362 --= 56.5% 641 135 --=38.8% 348 451 --=81.4% 554 626 -=78.9% 793 231 -=30.0% 777 117 -=28.3% 413 459 -=79.7% 576 475 --=65.1% 730 (2) Anderson 'i'X R cd pe i!> F, d X cd pe 'i'; F, 'i' X cd pe i!> 597 -=50.3% 1188 394 -=64.5% 611 85 -=35.0% 243 177 -=77.6% 228 (3) Anderson 'i' X R cd pe i!> F, 6 X rd pe 'i'; F, 'i' X Mexico i!> 402 --= <;47 -12.4% 577 --=59.9% 953 (4) Anderson 'i'X R cd pe i!> F, i!> X cd pe 'i'; F, i!> X cd pe 'i' 428 -=60.0% 713 346 -=57.8% 599 319 --=63.3% 504 656 --=78.0% 841 (5) Anderson 'i'X R cd pe i!> F, i!> X cd pe 'i' ; Fz d' X Mexico 'i' 542 -=65.1% 832 571 -=65.3% 875 433 -­=91.5% 473 613 --=95.9% 639 (6) R crl pe 'i'X F, i!> X cd pe New Orleans i!> 'i'; F2 'i' X cd pe i!> 9 --= 20.0% 46 233 --= 43.1% 540 132 -=58.4% 226 570 -=84.3% 676 281 -=49.3% 570 19 -=19.0% 101 569 --=83.1% 685 430 -=91.9% 468 (7) R F, cd pe 'i'X 6 X cd pe Newton 6 'i'; Fo 'i' X cd pe 6 253 -=52.2% 4-84 56 -=52.8% 106 625 --=86.6% 722 529 -=87.1% 607 190 -=53.8% 353 125 -=42.4% 295 526 --=85.5% 615 386 -=86.0% 449 (8) R F, cd pe 'i'X Newton 6 6 X crl pe 'i' ; Fz 'i' X Mexico 6 347 -=45.2% 767 366 --= 42.1% 869 916 -=68.1% 1346 734 -= 905 81.1.% 355 -=54.7% 649 264 -=41.6% 635 407 -=91.3% '!16 796 --=91.6% 859 (9) R cd pe 'i'X F, i!> X cd pe New Orleans 6 'i'; Fo J X cd pe 'i' 125 -=.39.4% 317 0 0(+2) ---=1.±% 333(+) 0 70 -=40.7% 172 0 0(+4) --=1.±% 648(+J O(+ll -­=1±% 165 (+) (10) R cd pe 'i'X F, J X cd pe Newton i!> 'i'; F, J X cd pe 'i' 165 -=50.3% 388 0 0 0 0 0 0 (11) R cd pe 'i' X Newton J F, i!> x cd pe 'i'; F2 i!>X Mexico 'i' 237 -­=57.5% 412 0 806 3 --= 2.1% 437 0 610 8rn -=64.6% 1301 0 1152 0 984 0 1278 (12) New Orleans 'i' X R cd pe J F, 6 X cd pe 'i'; Fz 'i' X cd pe J 2c5 -=50.4% 166 541 -=49.3% 1097 345 -=61.9% 577 470 -=79.1% 594 235 --= 36.5% 643 260 -=46.1% 564 440 -=66.0% 667 247 -­=77.9% 317 (13) New Orleans 'i' X R cd pe J F, 6X cd pe 'i'; Fz 'i'X Mexico J 103 -=30.2% 341 198 --=38.7% 511 234 --= 77.5% 302 826 -­=46.0% l797 385 -­=57.7% 667 79 -=69.3% 114 163 -=61.5% 265 (14) New Orleans 'i'X R cd pe J F, J X cd pe 'i'; Fz J X cd pe 'i' 61 -=34.3% 178 80 -­=35.5% 225 179 -=76.4% 234 77 -=78.4% 84 120 --=44.8% 268 253 -=55.5% 474 14 -=42.4% 33 127 -=78.4% 162 (15) New Orleans 'i'X R cd pe J F, J X cd pe 'i'; Fz J X Mexico 'i' 463 -=57.7% 802 792 -=48.8% 1523 365 -=61.7% 592 Studies in the Genetics of Drosophila of the 5 as such (see the data with the virilis Y-chromosome). If, how­ever, we consider those combinations which have the 2-3 fusion of americana or texana, .:e find that the egg hatch in general is much below the normal or that of gene combinations without the fusion. There are two possible explanations. One is that the gene combination is less viable. The second and more probable explanation is that the 2 chromo­some inversion and the 2-3 fusion cause abnormalities in meiosis. A similar situation has been investigated by Brown (1940) who demon­strated the general relation between crossing over and disjunction that exist in such cases. There is not enough data on the X-4 fusion of americana to determine its effect. Only a few data are available on crosses back to the americana P1 type. These suggest certain restrictions on the ability of certain chromosomes to replace others. The tests are as follows: P 1 Anderson , Studies in the Genetics of Drosophila In the descriptions, the ten alleles are placed in seriation from the most normal to the least normal in appearance. Each description is based upon visual examinations of etherized flies and upon studies of the impressions of the eyes. The least effect on facets is observed in lz97. Males of this genotype have a roughened area due to irregular facets in the posterior part of the eye and sometimes a wettish appearance in this area because of the fusion of a few facets. The eyes are normal in shape and size. Females lz37 are almost normal in appearance. In the isogenic strain the roughened area appears more regularly than in cul­tures which are not isogenic. The arrangement of facets in lz and lzBS is somewhat similar although lzBS is the less normal in appearance and also the more variable. Both genotypes are rougher than normal. In general small areas in which the facets are irregular in shape and the hairs are unevenly placed are scattered over the eye. Other facets have a normal appearance. Some of the areas of irregular facets have a wettish, warty appearance caused by fused facets and occasionally these fused areas are pimply because of a few facets which are rounded instead of hexagonally shaped. The warty areas vary in size and appear in the impressions as relatively large spots which resemble molten glass. Especially in lzBS these areas may on occasion approach the extremeness of lzY and cover most of the eye, with only small islands of normal facets being present. If the fused area is large, the appearance is that of a blister on the eye. In size and shape, lzBS eyes are smaller than normal but are rounded except for a flattened posterior margin. In lz the eyes vary from an ovoid to an almond shape even in isogenic cultures. The rounded eye in lz11 is almost normal in size and has an overall glossy appearance. Areas of fused facets are surrounded by islands and streaks of normal facets. The hairs on the eyes are irregular in position. Some of the fused areas resemble blisters and may have a slight pimply appearance. The characteristic glossiness of the eye is caused by the several streaks and areas of fused facets which give the appearance of molten glass. This allele differs from lz and lzBS in that it causes more blistery areas which tend to cover more of the surface of the eye. A more extreme variation of the eyes occurs in lz34. They are narrow and pointed at the dorsal and particularly at the ventral portion. Most of the facets are fused into irregularly shaped, glass-like blisters which vary in size. Some of these areas have the smoothness of glass but may have a few pimples scattered over the surface. Others, especially the large areas, have a creased central region which gives the appearance of erosion. Between the blisters are found one to three rows of normal facets with approximately five facets in a row. Although the number and size of the blisters vary, most of the eye is covered with the glass­like blisters which cause the eye to have a glazed appearance. The other five alleles have eyes which are narrower than normal. In each of them true facets fail to form and the whQle eye has a glass-like The University of Texas Publication surface. Two of these alleles, lz9 and lzY4 have a roughish, wettish ap­pearance due to the presence of pimples on the eye and to a few scattered hairs. The pimples which seem to be single rounded facets are irregular in number and position, but they are never frequent enough to repress the general appearance of a glass-like surface. The pimples and hairs occur more frequently in lz3 than in lzY4 and the two usually can be distinguished phenotypically by the rougher surface of lz9 eyes. In the three alleles, lzs, lzsB and lz36, the surface of the eye is a smooth, hairless sheet. When light is reflected on the eyes, pimples are recognizable on some lzs and lz•8 females, but in the impressions of the eyes no pimples or facets are observed. Of these three alleles, lz96 is the smoothest. COLOR OF THE EYES The characteristic red-brown color of the wild-type eye of Drosophila is due to the amount and distribution of two pigments, red and yellow, in the pigment cells of the ommatidia. Yellow pigment predominates in the outer cells of the ommatidia and red pigment in the lower cells. Most of the ten lozenge alleles affect both pigments of the eye. It is not possible to seriate the alleles because of the variations·in concentra­tion of the pigment. One group of the alleles can be graded from normal to extreme dark, sepia-like color of the eye, but the others have a yellow­ish, straw-brown color associated with a deeper, web-like arrangement of red and a marginal rim of darker pigment. In each of the alleles gene e which causes the dark ebony body-color intensifies the red but does not suppress the brownish color of the eyes. All examinations were made on white paper. Except for a wettish and sometimes shiny-red appearance in the small rough area, lz37 flies are uniformly normal in color. In lz flies the eyes are normal except for flecks of brown and sometimes of shiny red in the areas of fused facets. If these areas are numerous the whole eye has a dark color, or sometimes the color is a sharp red. The color in lz88 varies according to the irregularity of the facets. In some flies the eyes have normal areas with spots of shiny red where the blisters of fused facets occur. In others the color in some areas and at times over the whole eye is an opaque, muddy-looking clay-brown, somewhat similar to lzg. However, the eye seems to have more red than brown in it as though the red pigment was normal and the brown was decreased in spots. In isogenic cultures, lzg ordinarily has no normal color in any part of the eye. The color is best described as a venous or dark blood-red with wettish, shiny spots occurring in some areas. Two of the alleles are much darker than normal but the red pigment is concentrated in irregular fashion. A dark maroon-red color with occasional wettish spots characterizes lz84 • Variegation is observable in some eyes more than in others. In the heavier pigmented eyes, small, facet-size, yellowish spot~ ~re observed around which is a heavy aggregate Studies in the Genetics of Drosophila of red color, the whole appearance suggesting that the primary pigment cells surrounding the cone are almost pigmentless and the secondary pig­ment cells have a heavy concentration of granules. Other eyes have yellowish areas of varied sizes with streaks and spots of the darker red pigment throughout the lighter areas, somewhat resembling the web­like appearance in the facetless alleles except that lz94 has more red pig­ment. The margin of the eye in lz34 is always heavily pigmented but the streaks extending from it into the center of the eye prevent the appear­ance of a rim. A dark, sepia-like color is found in the facetless eyes of aged lzY4 flies. In contrast to lzu and lz94, the eyes are darker and seem more brownish in color although both red and brown pigments are observable in the eyes. In very young flies the color is not evenly distributed. A heavy and wide band of pigment is found at the margin of the eye and in the central por­tion, yellowish areas irregular in shape and size have streaks of dark color penetrating them. These streaks often are wide near the margin and come to a point nearer the center of the eye. As the flies increase in age, the yellowish areas become lost to view and the whole eye be­comes darker. No true rim of pigment is present because of the heavy streaks of pigment extending from the marginal areas. The other four facetless alleles have a greatly decreased amount of pigment in the eyes. The two spectacles, lzs and lz88, are identical. Their color is dilute, resembling a very thin layer of dried blood. On close inspection the color is found to be an even yellowish, pale straw-brown color with a deeper web-like distribution of soft, red-brown which gives the appearance of a fine stipple. A narrow darker rim encircles the eye. Eyes darken with age due to darkening of the red-brown pigment hµt the pale yellpw color is constant. Small variations in the depth of color are observed. In lz9 the color is heavier than in lzs because of an increased amount of red distributed more evenly over the eye. The yellowish color is only a little darker than in lzs. In older flies the color becomes a rusty brown. A rim is observed only in those flies which have a lighter center but ordinarily the color of the whole eye in lz9 is more nearly that of the rim of lzs. The center of the eye in lz36 is tannish with the very fine web-like distribution of red pigment. This has the lightest eye of all the alleles and in older flies the color approaches lzs but the rim persists. EFFECT OF ALLELES ON BROWN PIGMENT According to Ephrussi and Herold (1944) the recessive gene bw (brown) removes or suppresses all the red and also decreases the amount of brown pigment. Placing bw into the genotypes with the lozenge alleles, therefore, will show how these alleles affect the brown pigment which remai:p.s in the presence of bw. All except one of the lozenge alleles have an effect upon the brown pigment, and of these 9 all except one causes a definite decrease in the amount of pigment. The University of Texas Publication The allele with the most nearly normal facet arrangement, lz97, has no observable effect upon the brown pigment. The brown color is uniformly distributed over the eye. In lz, however, there is an increased concentra­tion of pigment but the opaque color appears more red than brown. In lz flies with the rougher eyes due to facet irregularities and in older flies, the color increases in redness. The distribution of the pigment over the eye is uniform. Three of the alleles have a slightly reduced quantity of brown pig­ment. In lza the eyes have a foamy or frosted brown color and a glazed appearance. The rougher areas have a darker color, somewhat reddish in appearance. Age also darkens the color. The pigment is reduced over all the eye so that it is uniformly distributed. Although the facets are more regular than in lza, there is a greater reduction of brown pigment in lzBs. The eyes of lz88 are faded in color, with less brown pigment in central areas than in the marginal regions, although some color is present throughout the eye. In older flies, the depth of color increases especially in the marginal regions to a reddish tinge. Even less brown pigment is found in eyes of lz94, the allele of these three which has the least regular facets. The eyes have a light, dirty, smeared brown appearance. This color is traceable to the irregular distribution of the pigment which ranges from very light areas to small reddish spots. Age increases the color of the eyes and causes the smeared appearance to disappear, but the increased color does not make the eye as dark as lza bw or lzBS bw. Five of the alleles have a greatly reduced amount of brown pigment in the eyes. In lz9 and lzY4 the distribution is more or less uniform over the whole eye. In color, lz3 bw is only slightly lighter than lz9 bw+, and is uniformly pigmented with a heavy straw-brown color. A marginal rim is found only in the very young flies in which the central area has not become fully pigmented. The eye color in lzu4 bw flies is just a little lighter than that in lz9 bw flies. In general the pigment is uniformly distributed although occasionally a fly will have a small, yellowish area in the central portion of the eyes. This allele, lzY4, which is sepia-like and therefore the darkest in color when both pigments are present, very closely resembles spectacle (lz•), one of the alleles least normal in color, when bw is added to the genotype to suppress the red pigment, the main difference being that lzY4bw does not have the rim of heavier color characteristic of lz•. The least amount of brown pigment is found in the three alleles which have the smoothest, glass-like eyes. In lz• bw and lz•B bw flies, the eyes have a tannish-brown central area not much lighter than that of the lz• bw+. In the central area of the eye can be observed small spots of pigment more or less uniformly scattered over the eye. A narrow mar­ginal rim of heavier color is present, this rim being narrower in the pres­ ence of bw than with bw+, and in lz•8 bw the rim is narrower in the anterior margin than in the posterior margin. Color increases in in­ Studies in the Genetics of Drosophila tensity with age, more so in lz•B bw than in lz• bw. Allele lz96 with bw differs from lz• bw primarily by having smaller spots of pigment in the central area of the eye. There spots resemble fine stippling and are uniformly distributed over the eye. The resulting color is a light tan center with a thin rim of heavier color. Although lz36 bw is the lightest of all the alleles in the tests with bw, with age the color of the eyes is as dark as that in aged lz• bw. EFFECT OF ALLELES ON RED PIGMENT The recessive gene st, scarlet, almost entirely suppresses the brown pigment and does not interfere with the red (Ephrussi and Herold). By adding st to the genotypes of the lozenge aTleles, therefore, it is possi­ble to determine the effects of the various alleles upon the red pigment of the eye. All except one of the ten alleles decrease the quantity of red pigment and some of them also cause very irregular distribution of the pigment. In many of the alleles black facet-size spots are found in the eyes. Three of the alleles seem to have little effect upon the red pigment, as determined by visual inspection. A uniform scarlet color not differing from st flies is found in lz37 st. In lzu and lzBS the color is distinctly scarlet but the whole eye is lighter and more shiny than in lz97 st flies. The dis­tribution of the pigment is not uniform although variegation is not very striking. In combination with st, lz usually causes a light, scarlet color with wettish or shiny spots on the eye, but in some of the flies the whole eye has a uniform orangy appearance although definitely not an orange. This color seems to be due to a decreased quantity of pigment over the whole eye. The red pigment is rarely concentrated enough in lz94 to cause a scarlet eye color. Ordinarily the eyes have an orangy tint, more orange than the lightest lz st but less so than lzu4 st. In the lz34 orangy eyes, the red pigment occurs in numerous spots and blotches which are irregular in . shape and varied in size. These aggregates of pigment are scattered over the eye and sometimes seem to be attached by thin lines of red pigment. Whether the small spaces between the spots of pigment are colorless or are yellow can not be determined by inspection. The whole combina­tion of the eye seems to be responsible for the color, with the greater concentration in a few eyes leading to a brilliant scarlet. A marginal rim rarely occurs, and when it does the inner margin is very irregular and the rim varies in width. The eyes of lzu4st flies are bright red-orange in color. The red pigment is distributed in irregular, anastomosing streaks which extend over the eye. Ordinarily the streaks are not wide although they may have points at which the aggregation of red pigment causes an irregular spot of some size. The colorless areas between the streaks are larger than those The University of Texas Publication in lz94. The pigment is always heavier along the margin of the eye causing the appearance of an orangy rim which however is very irregular along its inner margin. , The red pigment in lz3 st causes the eye to have a creamy-pink color. The small amount of red pigment that is present occurs in small spots which are distributed uniformly over the eye except in the marginal rim. The rim is very narrow especially in the dorsal margin of the eye. The other three alleles, lzs, lzsB and lz96, have no observable red pig­ment in the presence of st. With each allele the color is whitish although it is not the same white as that produced by w. These alleles with st behave in the same manner as homozygous st bw. It would seem, there­fore, that if st has no effec! upon red pigment, either of these three alleles suppresses the development of all the red pigment which their pigment cells should have in the presence of st. INTERACTION WITH VERMILION The sex-linked recessive gene v also controls the development of brown pigment and leaves only red pigment. Although one of the lozenge alleles, lz•B, cannot be tested with v because that gene cannot be crossed into the inversion associated with lzsB, the other alleles have been tested. The main difference between the combinations of the lozenge alleles with st and with v are found in the more extreme alleles and the positions of aggregates of pigment with lz94 and lzY4. However, the v combinations have more pigment than do the st combinations. In the lz97 v flies, the color is a uniform vermilion. This color also occurs in lzu v except that shiny spots occur on the eye. Ordinarily the red pigment is uniformly distributed but on occasions a small colorless spot occurs near the center of the eye, and the area around such a color­less spot is vermilion in color. The combination lz v has vermilion color with shiny (wettish) spots scattered over the eyes. Occasionally lighter vermilion areas occur in the center of the eye. Black facet-like spots are often observed in the eye, the number varying from one to as many as twenty. Considerable variation is found in the lzBS v combination. Many are vermilion in color although shiny spots and small areas of decreased pigment often occur. Sometimes several black facet-size spots are scat­tered over the eye. In some other flies a relatively large area and some­times smaller twin areas occur in which pigment is entirely absent. Surrounding this region, the red pigment is aggregated into a thick, opaque wide marginal rim covering approximately one-third of the eye on the anterior and posterior margins and a lesser width on the dorsal and ventral portions. This rim, if it may so be called, is a dark, muddy­gumbo color. In the three alleles lz97, lz, and lzu with v and in those lzBs v flies which are vermilion, the color as determined by visual inspection is the same. Studies in the Genetics of Drosophila However, Green (1947) has studied the absorption of the red pigment and determined that the quantity differs for the various alleles. · Both lz94v and lzY4 v have colors similar to the st combinations. The color in lz94 v is orangy. Red pigment is always present in the marginal area but is distributed irregularly in the central portion. In.many eyes small, facet-size clear but not colorless islands are surrounded by con­ centrations of red pigment. These give the impression of an absence of color in the primary pigment cells with as much as normal pigment in the other cells. Sometimes the clear yellowish areas are larger and streaks of red pigment of various widths, from fine lines to heavier streaks, extend over the eye in all directions, crossing and joining with other streaks. A rim as such is observable only if many thin streaks occupy the central portion of the eye. A heavy orange or tangerine color occurs in lzY4 v. Streaks of red pigment extend over the eye from a margin of concentrated pigment. The streaks vary in size and shape sometimes being wide near the margin and ending in a point nearer the center of the eye, although fine lines radiate out from these and cross or join with other lines. Where red pigment is not observable, the color is light yellow. Only a small amount of red pigment is present in lz9 v, the color being a pinkish cream and having more red than that observed in lz9 st. The red pigment is distributed over the eye as a fine stipple but distribution is slightly irregular. A thin orange rim is present around the margin of the eye. Only a very small amount of red pigment is present in lz• v and lz96 v. The pigment is uniformly distributed over the eye as a very fine stipple and gives a creamy color to the eye. A thin rim is present around all the eye. In lz96 v the posterior ventral portion of the rim is· wider than the other parts. The combination lz•B v has not been observed. DISCUSSION AND CONCLUSIONS The ten lozenge alleles under study affect the structure of the eyes, as measured by facet irregularities and size and shape of the eyes. All except one of the alleles also cause an alteration in the quantity of both the red and brown pigments in the eyes. The distribution of the pig­ments varies markedly with the different alleles. SERIATION OF INDIVIDUAL EFFECTS Normal facet arrangement is not characteristic of any of the ten alleles. Their effects on facets can be seriated from almost normal arrangement to complete absence of facets (Table I), but some of the alleles duplicate each other. The only deviation from normal in lz97 is a small area of irregularly shaped facets in the posterior part of the eye. Small warty areas of rough and fused facets occur in lz, and in lzBS the areas are larger and occur more frequently. An increased number of areas of <+ '"'! 0 0 ....., ::r '"'! ~ ::r 0 ....., ....., s:: II:> 0 -:i ct> s:: ...... s::: rJl <+ <+ ct> ....., i:l '"'! OQ ::r ::r Q.. 00 II:> Q.. rJl ::r ct> ct> n ct> ....., ~ ~ Q.. ~ s ~~ ~ ....., ...... 0 ct> ct> ct> II:>~ g.~~ ...... &r- TABLE 1 '1 ct> ct>!"~ ct> &i" <+ ~ .,..... .,..... a.s. Relative Order of Effects of Lozenge Alleles on Surface and Pigment of the Eyes 2'" ~OQ ~~ < g; >~~!\:> ~ ~ Brown Red Pigment o..::::::::r~::!~ ...... "-3 ct> ct>. -,_; ~ ~ Eye Pigment with st with v Color of (1) s· ~ '"'! ...... °" "' Allele Surface Amt. Dist. Amt. Dist. Amt. Dist. eye <+wsl-3 ~ n11:> 0 ~II:> it ~~ ~ OQ ~ Wild-type_______________________ _ 100 100 + 100 + 100 + 100 II:> "' :>;"'ct> II:> rJl 0 ~· OQll:>ct>'<'"'lct>w (1) l z3 7-------------------------------95 100 100 100 100 ,_.~ ct>ct><+w ~ + + + ~ Q.. st rJl ~. ::r '< .... lz_ 80 120 75 90 ± 105 ""' lz~~----------------------------+ + 1f ...... ro::rsro11:> ~ ,_.~ 11:> .....·.ro'O l zg --------------------------------75 80 90 90 110 ...... ~ct><>-''<'O@ '< ct> !:: ct> ct> ~ 50 90 90 95 120 II:> ct> .• rJl II:> l z34 ------------------------------+ + rJl '"'! rJl rJl ...... '"'! "-3 25 70 60 70 140 s:: ct> 0 i:l s-§ (1) l z3 -----------------------­ :4, ...... ;:;s .... n ~ lzY~-------------------------------10 40 + 10 + 15 + 60 II:> Q.. s::: ct> ::r ::r ct> ~ nct>OQ>ol\:>11:>· l zS ---------------------------------9 35 + 40 50 170 ct> i:l ::r ...... <+ < ""' ~g:· Sstrol-3 25 0.1 1 ~ lzS~-----------------------------4 + + + 30 ...... II:> 'O ct> II:> ::r' O"' .,..... st ,_. 1-3 (D' ;::; • OQ ct> 4 25 0.1 No test No test 30 .... 3 20 0.1 1 25 .,.... l z36 ___ ----------------------------­+ + o ..... ::r-w ,_......, ~ s:: i:l ct> ....., II:> s:: ~ ------------------------------- + + + II:> I\:> N rJl .... <+ll:>>o~nroro c ::r''O ...... o,.ct> Q..O.. ;_::;LEGEND. The alleles are placed in order of severity of effect on facets from normal to least normal. Brown pigment was tested ll:>>oS e; by using bw. The "Color of eye" is based upon amount of color. The numerical values are not percentages but merely show ..... ct> ...... 0 ct> II:> I\) '1;1):=.,nw'O'"'! the relative degree of difference from normalcy for the specific category in the column. "Amt." refers to quantity of pig­ Y''"'!ronw>oct> ment; "Dist." to uniform distribution, if, + or irregular or variegated, if -. II:> rJl II:> ct> ct> II:> II:> i:l ~. '< I\) rJl l=lnworo'"'ln O..!'C~l:lwll:>o I\) ?-j~'"'!~< ,_.1--10,_.roroct> st i:l <+ '< f{5 • '1 OO"O::r'S1-3S ~g_o-~e::~g ::r' ::r ct> ct> ct> 0 ct> Studies in the Genetics of Drosophila small pimples seem to be present when light is reflected on the eye, pimples are not observable in impressions of the eyes. The smoothest glass-like eye is found in lz36 flies. It is not possible to use the same seriation fol' eye color that is used for facet irregularity. One group of alleles can be graded from normal to increased dark color but in four of the alleles the color is considerably reduced (Table I). The color in lz97 is normal. Brown flecks in the eyes cause a slightly darkened color in lz. Eyes of lzBs vary in color from that of lz to a clay-red and sometimes to a darker color. The color in lzu is brownish and resembles venous blood. A heavy maroon-red is char­acteristic of lz34 flies. The darkest eyes are found in lzY4, the color being sepia-like. In contrast to those alleles, color is reduced in lz9 to rusty, reddish brown. Both lzs and lzsB are yellowish straw-brown with an underneath, web-like scattering of red and a darker marginal rim; and lz96 is similar to them except that the center is a little lighter. When the pigment components are followed separately, linear seria­tion of the effects upon the quantity of pigment can be made. The brown pigment is increased sufficiently to give a reddish color to lz. It is normal in quantity and distribution in lz37 and decreased in amount in lzu, lzBs, and lzs4, in that respective order. Distribution of the brown pigment is uniform over the eyes in lzu but somewhat variegated in the other two. Both lz9 and lzY4 which are similar for facet arrangement are also simila·r for quantity and distribution of brown; both have a dilute color but lzY4 is slightly lighter. Allele lzY4, which is the darkest of the series when both pigments are present, becomes similar to one of the spectacle series when only the brown pigment is present. The other three alleles have only a very small amount of brown pigment. The alleles vary in the amount and distribution of red pigment which they produce and can be seriated according to their effects from prac­tically normal to almost complete absence of the pigment. As is true with the brown pigment the seriation of the effects on red pigment does not follow exactly that of facet effects (Table I). In general the same degree of effect occurs whether st or v is used to remove the brown pigment, but slightly more pigment is present with v than with st in the genotype. No alteration of red pigment is observed in lz97• The three alleles lzu, lzBs, and lz cause some reduction in the amount of red pigment. With st, the greatest reduction is likely to occur in lz, the reduced but evenly distributed pigment causing the eye to have an orangy appear­ance. The other two alleles have approximately the same degree of redness and do not differ greatly from st. With v, the three alleles are alike as determined by visual observation, but differ in the order lzu, lzBs, lz as shown by the absorption spectra of extracts of the eyes (Green 1947) ; this is just the reverse of the order expected if the seriation were the same as that for facet arrangement. With lz94 and lzY4 the red pigment is reduced and is irregularly distributed in spots, streaks and The University of Texas Publication blotches of heavy red concentrations surrounded by clear areas of varied sizes and shapes. Both the amount and distribution give an orangy tint to the eyes of lz84 and a definite orange color to lzY4. Allele lz9 has a creamy color with st and pinkish center and orangy rim with v, due to the greatly reduced a~ount of pigment and i.ts even distribution over the eye. The three alleles lzs, lzsB, and lz86 have very little red pigment. When combined with st, their color is whitish, although not the same white as that caused by w, and only a small amount of color (creamy) is observed when the alleles are combined with v. One, lzsB, cannot be tested with v. RELATIONSHIP OF SERIATIONS Comparisons of the seriations indicate that the effects on facet dis­order and on pigment production are interrelated to some extent. Dif­ferences in the behavior of duplicates and certain exceptions to linear seriation, however, sugg~st that some ~egree of independence between the two effects exists. A relationship between the effects on the facets and pigmentation is suggested by the tendency for certain of the alleles to have the same severity of expression for each trait. In lz97, only a small area of the eye is irregular and the red and brown pigments are normal in quantity and distribution. Normal facets occur infrequently in lz84. Both pig~ ments in flies having this allele are irregularly distributed and greatly reduced in amount. The three "spectacles" which have facetless, hair­less, glass-like eyes also have very little brown pigment and practically no red pigment in the eyes. Lack of ·complete relationship is suggested, however, when an attempt i& made to fit the linear seriation of one effect to that of another effect. In a comparison of facet effects and color of the eyes, one group of alleles, lz37, lz, lz8s, lzY, and lz84, can be graded in that order from normal facets through increasing irregularity and in the same order from normal through increasing darkness of color. The two alleles zzs and lzY4 which almost duplicate each other and are next after lz84 in severity of effect on facets break the relationship. Although lzY4 is even darker than lz84, the opposite is true for lz3 in which the pigment is greatly reduced in amount. This difference in color between lz9 and lzY4 is pri­marily one of a difference in quantity and distribution of red pigment. The other three alleles (the spectacles) occur in a series with lz3; that is, the more abnormal the facets, the greater the reduction in pigment. If one follows the brown and red pigments separately, the alleles can be seriated in approximately the same order as they are for facet irregu­larity, although here too striking exceptions occur. Only lz, lzBS and lzY which follow that order for increasing facet irregularity fail to follow the same general gradation for brown pigment production. All three alleles are rougher than lz87 but lz has a heavier concentration of brown Studies in the Genetics of Drosophila pigment and both lz88 and lzY have less brown pigment than lz97, with lz88 having the least amount and the more irregular distribution. As determined by visual inspection lz11 and lz88 have the same quantity of red pigment and lz with its more normal facets has a reduced amount of the pigment. Also lz9 and lzY4 which are similar for facet arrangement differ considerably in quantity and distribution of their red pigment. The other alleles have the same seriation for facet arrangement that they have for production of red pigment. Some alleles fail to follow the same pattern of seriation for the pro­duction of the two pigments. In lz•B and lz•, lz36, and lz3 the brown is not decreased as much as the red. The brown in lz is increased and the red decreased in quantity. The two alleles lz34 and lzY4 cause considerable variegation of the red pigment as well as a decrease in its quantity, but with the brown pigment variegation is observed only in lz34 and that is less so than with the red. Allele lz88 has less effect on red pigment than it has on the brown. In the lozenge alleles all of the effects reported here involve the eyes and therefore might be expected not to show very much independence. Nevertheless the lack of common seriation and also the behavior of some alleles which duplicate each other for certain effects indicate that there is some independence of the effects. Alleles lz3 and lzY4 are almost identical for eye surface and for their effect on brown pigment, but they differ greatly for effect on red pigment. In lz3 only a small amount of red is present and this is evenly distributed causing the eye to have a pinkish tint. On the other hand lzY4 has the red pigment ag­gregated in streaks scattered over the eye in such a manner that the color is an orange or tangerine. To a lesser extent lz and lz88 are alike for facet irregularity although lz8 S is slightly the less normal, These two have greater differences in their production of brown and of red pigment when the latter is studied with st. It is also known that lz• and lz•B which are identical in their effect on facets, brown pigment and red pigment production do not behave alike in compounds with some of their alleles (011ver 1945; Cummings 1946). One of them, lz•, has a comple­mentary effect with certain alleles which interact with lz•B to produce a blending of the trait. INTERPRETATION OF EYE· COLOR The gene may control the production of pigment by one of two methods or by a combination of the two. Decrease in pigment may be secondary to cellular disarrangement or it may be caused as a primary effect of the gene. Evidence that some ommatidia are pigmentless even though the cellular arrangement is normal, has been reported by Casteel (1929). Johannsen (1924) has also reported genie effects on pigment which are quite independent of structural arrangement of the ommatidia. Some of the ommatidia in the lozenge alleles may likewise have normal cellular The University of Texas Publication structure but lack pigment or have a decreased amount of pigment. Under those circumstances the effects on facet arrangement and color may be independent of each other, thus being similar to a condition re­ported in maize. The R allele in corn may have its effect on the aleurone, seedling or pericarp and the effect on one may be independent of the others (Stadler and Fogel 1945; Stadler -1946). Some of the alleles, however, affect more than one character although the degrees of expres­sion may not be the same (Fogel 1946). The irregularity of the facets and of pigment production may both be directly related to irregularities in the ommatidia. Waddington and Pilkington (1943) reported for lzs, one of the alleles studied here, that the retinulae and cone cells are in disorder. Primary pigment cells are present and have produced pigment as well as a flat cornea. Basal pig­ment cells are present at the base of the ommatidia, but the secondary pigment cells are absent or disorderly arranged. Although it is unwise to assume that each allele of lz acts in a manner identical to that of lzs, it is nevertheless conceivable and very probable that with the alleles facet irregularities are caused by cellular disarrangements and these disturbances may include the loss and irregular arrangement of some of the pigment cells. A gene control of cellular differentiation would therefore secondarily cause abnormal facet and pigment production. The color of the eyes in the lozenge series of alleles is determined by the quantity and distribution of the two pigments, but another important factor is the contrast developed as a consequence of the distribution, in which one color intensifies an accompanying color. The small amount of red in lzs, lzsB and lz96 is not observable as such in the presence of st although the eye does not have the colorless appearance that is observed when gene w is present. When these alleles are present with bw+ and st+, they have a very light straw-brown color and the red pigment can be detected as a fine web-like stipple underneath. The orangy appear­ance of lz st is apparently the result of a dilution of the red pigment over all the eye, but in lz94 st it is due to the reflection of light through the clear areas around which the red pigment is aggregated in spots and fine streaks. When both pigments are present in lz94, the contrast made between the smeared, somewhat irregularly distributed brown and the extremely variegated red pigment causes a dark maroon-red color. In lzY4 st the reflection of the light around the heavy streaks of red pigment causes an orange color. The brown pigment in the eye is regularly distributed but is reduced to a rusty brown color. These two pigments give a contrast which together with the altered reflection of light causes an eye color that is as dark as sepia. The flecky color in lz seems to be associated with the heavy concentration of brown pigment and the irregular distribution of red in the eye although no area lacks red pigment all together. In lz88 v the occasional fly with the wide, dark rim and colorless center of the eye apparently has a concentration of Studies in the Genetics of Drosophila red pigment through which little if any light can be reflected. The re­sulting color is that of muddy gumbo. In part the colors of the eyes as well as facet irregularity seem to be secondary effects of a gene action which causes irregular cell differentia­tion in the ommatidia as shown by Waddington and Pilkington for lz•. Another effect of the alleles, reduced fertility, is related to some extent to abnormal genitalia. The degree of effect of the zzsr gene on the genitalia is associated with a similar degree of effect on fertility (Anderson 1945). In lzu, the abnormal genitalia are due to lack of proper cellular differen­tiation in the genital disc, and this is possibly true for the other alleles. There is thus a suggestion that these four effects are all related to one primary action of the gene, the control of cell differentiation. However, seriation of the alleles for their effects upon one trait is not always the same as that for another. Although the exceptions may be less extreme for color and facet arrangement, all of the alleles except lz37 are identical in that they cause total absence of the spermathecae and parovaria. Moreover, some of the alleles that are phenotypically identical do not interact in the same manner when in compounds with other alleles. These facts suggest that some of the alleles differ in their production of the effects and that the gene at this locus produces more than one physio­logically independent effect. SUMMARY Ten lozenge alleles were studied for their effects upon facet irregularity and pigmentation of the eyes. The two pigments were studied separately by inserting bw to remove the red pigment and st and also v to remove the brown. The alleles can be seriated in their effects on facet irregularities, rang­ing from almost normal to total absence of facets. Some duplicate alleles occur. For color of the eyes, one group of the alleles can be graded from normal up through increasing darkness to sepia and another series grades down to a dilute brown. Production of brown and of red pigment can be seriated from normal through decreasing quantity to almost total absence of pigment. Attempts to fit the several seriations of effects show that in general they are in agreement but that striking exceptions occur. Some alleles which duplicate each other for one or more effects differ greatly for another. It is suggested that in part the facet irregularities and the production and distribution of brown and red pigments are interrelated and that both are secondary effects of the genie control of the structure of omma­tidia. It is also suggested that in some of the alleles the effects on the eyes are results of independent controls by the gene. Color of the eye is caused by quantity and distribution of pigment, but also by contrasts of the two pigments in which one intensifies the other. The University of Texas Publication REFERENCES Anderson, R. C. 1945. A study ef the factors affecting fertility of lozenge females of Drosophila melanogaster. Genetics, 3() :280-296. Bridges, C. B. and Brehme, K. S. 1944. The Mutants of Drosophila melanogaster. Carnegie Inst. Wash. Pub. 55~. Casteel, D. B. 1929. Histology of the eyes of X-rayed Drosophila. J. Exp. Zoo!., 53 :373-385. Cummings, K. 1946. A study of a series of multiple alleles at the lozenge locus in Drosophila melanogaster. Thesis for M.A. Univ. of Minn., June, 1946. Dobzhansky, T. and Holz, A. M. 1943. A re-examination of the problem of manifold effects of genes in Drosophila melanogaster. Genetics 28:295-303. Ephrussi, B. and Herold, J. L. 1944. Studies of eye pigments of Drosophila. I. Methods of extraction and quantitative estimation of the pigment components. Genetics, 29:148-175. Fogel, S. 1946. Gene action and histological specificity of pigmentation patterns of certain R alleles. Genetics, 31 :215-216. Glass, H. B. 1934. A study of dominant mosaic eye-color mutants in Drosophila melanogaster. Amer. Nat. 68:107-114. Gottschewski, G. 1936. Quantitative and qualitative unterscheide innerhalb einer Allelenreihe bei Drosophila melanogaster. Zoo!. Anz. Supp., 9:104-112. Green, M. M. 1947. Red eye pigment in the lozenge allelic series of D. •melanogaster. Genetics, 32:89. Griineberg, H. 1938. An analysis of the "pleiotropic" effects of a new lethal mutation in the rat (Mus Norvegicus). Proc. Roy. Soc. London B, 125 :123-144. Griineberg, H. 1943. Congenital hydrocephalus in the mouse, a case of spurious pleio­tropism. J. Genetics, 45 :1-21. Johannsen, O. A. 1924. Eye structure in normal and eye-mutant Dro.sophilas. J. Morph. and Physiol., 39 :337-349. Mohr, 0. 1932. On the potency of mutant genes and wild-type allelomorphs. Proc. Sixth Int. Congress Genetics, 1:190-212. Oliver, C. P. 1945. Four lozenge alleles phenotypically alike which react differently with their other alleles. Genetics, 30 :16. Oliver, C. P. and Anderson, R. C. 1945. The effect of rematings on the fecundity of an infertile mutant female. Amer. Nat., 79 :89-94. Oliver, C. P. and Green, M. M. 1944. Heterosis in compounds of lozenge alleles in Drosophila melanogaster. Genetics, 29 :331-347. Stadler, L. J. 194.6. Spontaneous mutation at the R locus in maize. I. The aleurone­color and plant-color effects. Genetics, 31 :377-394. Stadler, L. J. and Fogel, S. 1945. Gene variability in maize. II. The action of cer­tain R alleles. Genetics, 30: 23-24. Stern, C. 1943. Genie action as studied by means of the effects of different doses and combinations of alleles. Genetics, 28 :441-475. Waddington, C. H. and Pilkington, R. W. 1943. The structure and development of four mutant eyes in Drosophila. J. Genetics, 45:44-50. Wright, S. 1932. Complementary factors for eye-color in Drosophila. Amer. Nat., 66:282-283.