No. 3825 July 1, 1938 GYNANDROMORPHS IN DROSOPHILA MELANOGASTER By J. T. PATTERSON Professor of Zoology The University of Texas and WILSON STONE Associate Professor of Zoology The University of Texas ~UllLISHED BY THE UNIVERSITY OF TEXAS AUSTIN Additional copies of this publication may be procured free of charge from the University Publications, The University of Texas, Austin, Texas No. 3825: July 1, 1938 GYNANDROMORPHS IN DROSOPHILA MELANOGASTER By J. T.PATTERSON Professor of Zoology The University of Texas and WILSON STONE Associate Professor of Zoology The University of Texas PUBLISHED BY THE UNIVERSITY FOUR TIMES A MONTH AND ENTERED AS 15EC:OND·C:LASS MATTER AT THE POST OFFIC:E AT AUSTIN, TEXAS, UNDER THE AC:T OF AUGUST 24, 1912 The benefits of education and of useful knowledge, generally diffused through a community, are eNential to the preservation of a free govern• ment. Sam Houston Cultivated mind is the guardian genius of Democracy,and while guided and controlled by virtue, the noblest attribute of man. Itisthe only dictator that freemen acknowledge, and the only security which freemen desire. Mirabeau B. Lamar CONTENTS PAGE INTRODUCTION ----------------------------------------------------------------------------------------------------------5 MATERJAL --------------------------------------------------------------------------------------------------------------------7 DESCRIPTION OF GYNANDROMORPHS________________________________________________________________ 12 1. Half-and-half gynandromorphs --------------------------------------------------------------------------12 A. Approximately bilateral specimens ____________________________________________ ______________ 13 B. Anterior-posterior gynandromorphs -----------"--------------------------------------------15 C. Mixed half-and-half gynandromorphs -----------------------------------------------------17 2. Gynandromorphs about three-fourths male -------------------------------------------------------17 3. Gynandromorphs one-fourth male or less__________________________________________________________ 19 4. Gynandromorphs with broken X-chromosome in male parts ---------------------------22 COMPLEX AND UNDETERMINED CASES ___ _________________________________________________________ 27 FERTILITY OF GYNANDROMORPHS______________________________________________________________________ 32 DISTRIBUTI0 N 0 F MALE PARTS______________________________________ _ ________________________________ 36 1. Cleavage patterns --------------------------------------------------------------------------------------------------37 2. The time of formation of gynandromorphs ·-------------------------------------------------------40 SUMMARY AND CONCLUSIONS------------------------------------------------------------------------------48 LITERATURE CITED --------------------------------------------------------------------------------------------------50 PLATES ---------------------------------------------------------------·-----------------------------------------------------------52 GYNANDROMORPHS IN DROSOPHILA MELANOGASTER J. T. PATTERSON AND WILSON STONE INTRODUCTION The occurrence of mosaic organisms gives an excellent opportunity for studying certain problems in the field of developmental genetics. This opportunity has been greatly increased by the discovery that x-radiation may be used as a means for their production. The term mosaic is applied to any individual which exhibits genetically diverse tissues in its soma, regardless of the causes underlying the production of such tissues. Mosaicism may be due to any one of several causes. Some of the more common of these are the elimination of all or part of a chromosome, somatic crossing­over, unstable translocations, the so-called position effect mottles, somatic mutations, and even germinal mutations if they happen to be of the fractional type. Animal mosaics may be placed under two groups. The first of these includes the sex-mosaics or gynandromorphs. Such individuals usually have two distinct regions, one showing the characteristics of the female and the other having male characters. The two regions are often sharply defined, especially in those species which have a strong sexual dimorphism. Sex-mosaics have been known for a long time and have been reported for a wide range of animal species, but more particularly for the insects. In counter-distinction to the sex-mosaics, are individuals in which the contrasting regions do not show a difference in sexuality. Whatever may be the nature of the causes underlying the formation of these non-sex mosaics, they clearly do not affect a change in the sex of the two kinds of tissue. In general these mosaics result from changes affecting genes or sections of chromosomes which are not primarily concerned with sexual differentiation. The discovery of new cases of mosaics of either kind may lead to the solution of certain difficult problems in biology. A very good illustration of this possibility is to be seen in the recent discovery of two colonies of ants in Trinidad, B.W.I. by Dr. Neal A. Weber. One of these colonies contained more than 4000 gyandromorphs and the other had 164 anomalous individuals. The second colony belongs to a species of fungus-growing ant (Acromyrmex octospinosus) and has recently been described by the late Professor W. M. Wheeler (1937). Among the 164 anomalous specimens were ten sex-mosaics or true gynandro­morphs, 46 female-worker mosaics (gynergates) and one major media mosaic (diploergate) . The last 47 individuals therefore represent caste-mosaics, and since they and the ten gynandromorphs were probably descended from a single dealated mutant female, their mosaicism must have been due to genetic factors. If this could be established by breeding tests, it would definitely settle the controversal question of the origin of caste dimorphism in the social Aculeates, for as Wheeler points out, it would support those who have maintained that this dimorphism is predetermined in the egg (blastogenic) and could not therefore be due, as some have claimed, to dif­ferential larval feeding ( trophogenic) . The first gyandromorph for D. melanogaster was found by Morgan in 1910. Other cases were soon described by several different workers (Duncan, 1915; Morgan, 1915; Hyde and Powell, 1916). They have also been found in D. simulans (Sturtevant, 1921) and in D. funebris (Spencer, 1927, Timofeeff-Ressovsky, 1928). Following their discovery in D. melanogaster, Morgan and his associates began to accumulate data on their occurrence, and in 1919 Morgan and Bridges published the first com­prehensive paper on these mosaics. In this paper the authors gave an extensive discussion of the problem of gynandro­morphs and their origin. They referred to about 100 specimens of Drosophila gynandromorphs, and analyzed in detail some seventy odd cases. They found that a majority of these mosaics were best explained on the basis of the elimination hy­pothesis, which was first suggested by Morgan in 1914. According to this theory, the gynandromorph starts as an XX-zygote or female, but at the first cleavage divi­sion (or a later one in some cases) a daughter X-chromosome is lost or eliminated from one of the resulting nuclei, so that one-half of the fly becomes male of the composition XO. They did not find it necessary to evoke either Boveri's (1888) partial fertilization theory or Morgan's ( 1905) polyspermic theory in order to ac­count for the origin of any of their mosaics. However, they did find some special cases that could not be explained on the basis of simple elimination. They found that these could be explained by assuming, as Doncaster (1916) did for Abraxas mosaics, that the gynandromorph had arisen from a binucleated egg in which the reduced nuclei had been fertilized by different spermatozoa. Other points of interest found in their conclusions are the following: First, that the paternal X-chromosome is eliminated with the same frequency as the maternal X; second, that once the male and female parts are established they become self­differentiating, each developing according to its own constitution; third, that in the same individual mosaic the gonads are always the same, either both ovaries or else both testes. The second conclusion has been slightly modified by more recent dis­coveries (e.g., vermilion eye), and the third conclusion does not hold in certain cases (Dobzhansky, 1931). L. V. Morgan (1929) carried out an extensive experiment for the purpose of de­termining the relative frequency of gynandromorphs and other mosaics due to double fertilization as compared to those due to elimination. She did not obtain any evi­dence for double fertilization in her own experiments, but reviews a number of com­plex composites and offers several suggestions as to their possible mode of origin. These will be considered in connection with the complex cases found in our own material. Sturtevant (1929) has described a large number of gynandromorphs for the claret mutant stock D. simulans. Claret is a III-chromosome recessive, and females homozy­gous for this gene produce among their offspring more than three per cent. of gynan­dromorphs. The most remarkable fact is that this gene causes the elimination t)f chromosomes, and in the case of gynandromorphs it is usually the maternal X that is missing from the male parts. This was shown in crosses in which the paternal X-chromosome carried such mutant genes as yellow, forked or singed. Of the 158 gynandromorphs listed in the tables, 145 showed the paternal sex-linked characters in their male parts, and the female parts were all wild-type. This means that it was the maternal and not the paternal X-chromosome which was absent from the male parts. In the thirteen exceptional specimens the reverse condition was found to exist, since it was the paternal X that was absent. The second part of Sturtevant's paper deals with the question of the distribution of parts in the gynandromorphs and will be referred to in later sections. Ward ( 1936) has presented cytological evidence in support of a very interesting hypothesis to account for the origin of gynandromorphs by claret females of D. simulans. We have no evidence that any gynandromorphs in D. melanogaster are produced by such a mechanism. In 1931 Patterson published a paper on the production of Drosophila gynandro­morphs by means of x-rays. He presented evidence which showed that irradiating germ cells resulted in the formation of gynandromorphs through the elimination of all or a part of one of the X-chromosomes during an early cleavage division. The data contained in that article have been incorporated in the present paper, so that further comments are unnecessary at this point. MATERIAL The material upon which this article is based has been accumulated from time to time since the early part of 1928. At that time the senior author began a series of experiments which had as their object the finding, if possible, of a postulated major sex gene in the X-chromosome of Drosophila. The method of procedure was reh­tively simple, consisting in the main in x-raying flies with normal chromosomes and mating them to individuals which carried a series of recessive genes scattered along the X-chromosome. The results showed that gynandromorphs could be produced in this way, either through the elimination of an entire daughter X, or by breakage and the elimination of a part of the daughter X. The experiments yielded fifty-two gynandromorphs, each of which contained in its male tissue one unbroken and a fragment of another broken daughter X-chromosome. Twenty-six different stocks with marked X-chromosomes were used as the untreated parent, and eight different stocks were employed as the treated parent. The most commonly treated element was the normal X-chromosome, and next to this was the so-called Theta X. . This chromosome is marked with the mutant genes yellow and scute at the left end and carries a short duplication covering these two genes and attached to the right or spindle fiber end. Hence the Theta fly appears wild-type. Homozygous Theta females are not very viable, so it was necessary to carry the stock in balance with the partner X also marked with yellow and scute. The third most frequently used stock is known as scute-8 apricot and in this the X-chromosome has a long inversion, which extends from a point between the loci of achaete and scute to beyond the right of the locus of bobbed. The record of each gynandromorph was entered on a 4 X 6 inch card which had a simple outline drawing of a fly printed on the upper left-hand corner. The male parts of the mosaic were indicated on the diagram by stippling for all except the eyes and tip of the abdomen which were colored solid black. The card had suffi­cient space for entering all essential details which would not appear on the diagram. Copies of these diagrams are used as illustrations for the eight plates. Below each diagram is given the type of X-chromosome (or chromosomes) which was present in the male parts of the mosaic. TABLE 1. CLASSES OF GYNANDROMORPHS X-chromosome lost % male Partial ~male 14 male Total or broken Treated paternal 11 34 36 28 109 A 20 31 Treated maternal Untreated maternal 0 6 5 2 12 11 338 B 41 Broken paternal Untreated paternal 0 11 6 24 49 Broken maternal 6 25 13 5 0 1 2 0 3 ----------:------1-----+-----1------t--­Control paternal 2 8 1 24 35 D 3 7 5 10 25 Control maternal ---+-----------1----1------t-------j------------­E I Complex or 0 3 1 5 9 undetermined Total 24 107 77 127 335 Per cent 7.1 31.9 22.9 37.9 The total number of gynandromorphs obtained was 335, and these are all listed in Table 1. They are arranged under the four classes of three-quarters male, one­half male, one-quarter male and partials. This represents our original classification which was based on the relative amounts of male and female tissue as judged from an examination of the external surface of the fly. Such a classification is subject to a certain amount of unavoidable error, and especially true is this in those mosaics which had male and female parts irregularly distributed or scattered. The specimens are arranged in the table under five groups (A-E, on left of table). The first group (A) gives the gynandromorphs derived from x-rayed sperm, the first line showing the number with treated paternal X missing, and the second line show­ing those with the partner or untreated maternal X absent. The ratio of 109:31 will serve to indicate the effectiveness of x-radiation in producing gynandromorphs. This contrast becomes all the more striking if we include the 49 cases with broken paternal X-chromosomes (group C), which came from the same parents. This would make the ratio 158 :31, and since in spontaneous gynandromorphs the paternal and maternal X-chromosomes are eliminated with about the same frequency (Morgan and Bridges, 1919), about 137 of these gynandromorphs must have been due to the ef­fects of the x-radiation. Group B shows the number of gynandromorphs obtained in the treated female series. One should add to the total of the first line of this group the three cases of broken maternal X-chromosomes (group C), since they were obtained in the same series. This would give a ratio of 36 :41 in favor of the eliminated untreated paternal X-chromosome, but this difference is clearly not significant. The fact that the two X-chromosomes are eliminated with about the same frequency, when female germ cells are treated, is probably due to an indirect effect of the x-rays on the paternal X operating through the cytoplasm (Patterson, 1931). All detectable cases of gynandromorphs carrying an extra fragment of a broken X-chromosome in the male parts have been placed in group C. The great excess of cases from the treated male series in contrast to those from the treated female series (49 to 3) is in part due to the fact that males were more frequently and more heavily treated than females. But the chief reason for this difference must lie in the condi­tion of the chromosome at the time the x-rays are applied. In order to produce a mosaic by breakage, the gametic chromosome must in effect be in the two-strand stage, and it appears from the results that this is less frequently the case in the un­reduced egg nucleus than in the reduced sperm nucleus. The control series (group D) shows a ratio of 35 :25 cases in favor of the elimina­tion of the paternal X. The difference here centers in the partial cases and is due to the use of the scute-8 X-chromosome in an extended series of experiments. This chromosome has the peculiar effect of bringing about the elimination of the partner X in late cleavage stages, and in the series of experiments just referred to scute-3 females were extensively employed. This resulted in producing a preponderance of partial gynandromorphs through the elimination of the paternal X-chromosome. In all other stocks we found that the paternal and maternal X-chromosomes were eliminated with about the same frequency in the untreated series. The complex and undetermined cases are placed in group E. There are only nine of these specimens and they will all be considered in a later section dealing with such cases. Since we shall have occasion to refer repeatedly to Tables 2-8 in the next section it is necessary at this point to explain the nature of the data displayed in them. The records of 213 of the 335 gynandromorphs are included in the seven tables. We have recorded all cases in which the male parts were yellow, forked (usually JS) or singed, or else were combinations of yellow-forked or yellow-singed. These three mutant characters will show on all bristle-bearing parts of the fly whenever the op­posing dominant genes have been removed by the elimination of the non-marked X-chromosome. This makes it possible to identify quite accurately the limits of the male parts as they appear on the surface of the mosaic. For the most part the 122 gynandromorphs not recorded in these tables but listed in Table 1, represent cases in which either the marked X-chromosome had been eliminated or yellow, forked, or singed were not involved, so that it was not possible to determine accurately the exact limits of their male parts. We have recorded for each of the 213 specimens the character of the different or­gans that must be supposed to develop from separate imaginal discs. The determina­tion of such imaginal disc derivitives should be based on a thorough knowledge of their embryology. But up until very recently there has not been available any informa­tion on the development of Drosophila beyond the formation of the blastoderm. In his paper on the distribution of parts in the gynandromorphs of D. simulans, Sturte­vant found it necessary to use as a basis for his study the accounts of development which had been given for such forms of Musca, Calliphora and Rhagoletis. Since then papers by Chen (1929), Strasburger (1935) and Parks (1936) on the develop­ment of Drosophila have appeared. Chen made a special study of the development of the imaginal discs or buds of the cephalic and thoracic regions. He showed that the head of the adult is represented in the mature larva by two pairs of primordia, a large and a small pair. The smaller pair is called the labial buds and gives rise to the probosis and the low~r part of the head. The large cephalic complex is composed of two pairs of bud.s, the antenna! and optic buds. The former produce the antennae and the front part of the head, the latter give rise to the compound eyes and the dorsal part of the head. Chen also studied the development of the thoracic imaginal discs or buds. There are six pairs of these buds, or two pairs for each thoracic segment. In the first thoracic segment, the two dorsal buds give rise to the humeri, and the two ventral buds produce the first pair of legs and propleura. The two dorsal buds of the second thoracic segment give rise to the mesonotum, wings and scutellum, while the two ventral buds give rise to the second pair of legs and the sternopleura. Finally, in the third thoracic segment, the two dorsal buds produce the halteres and metanotum, and the two ventral buds form the third pair of legs and hypopleura. Parks shows that there are eight abdominal segments in the larva, and that each of these segments has four imaginal discs, a dorsal and a ventral disc on each side. According to his account and that of Sturtevant, the fate of the segments and the imaginal discs is as follows. In the formation of the adult segment, each of the dorsal discs of any segment gives rise to half of the dorsal sclerite (tergite), and each of the ventral discs gives rise to half of the ventral sclerite (sternite). The first and second larval segments combine to form the first abdominal segment of the adult, and the last two larval segments give rise to the anal and genital parts in the adult. In the tables (2-8) we have checked forty-two separate structures for each mo­saic. The symbols used are plus ( +) for female, minus ( -) for male and plus­minus ( ±) for structures that show a mixture of male and female characters. They are recorded separately for the left and right sides of the three main regions of the fly. For the head region we have used three areas that might be supposed to develop from the three imaginal discs on each side, as indicated by Chen's Chart 1 (1929, p. 142). For the optic bud derivitive (Op) we have included the compound eye and the dorsal part of the head; for the antenna! bud derivitive (An), the antenna and the front part of the head; and for the labial bud derivitive (Pb), the probosis and the lower part of the head. In recording the parts of the thorax, the dorsal and ventral organs are checked for the first two segments, and the ventral only for the third segment, as follows: First segment dorsal, humerus (H), ventral, first leg (Ll) ; second segment dorsal, wing (W), mesonotum (M) and scutellum (Sc), ventral second leg (L2); third segment ventral, third leg (L3). The dorsal derivitives of the third segment are not included. TABLE 2 ~ ~ i.., c ;:! c ~ ~ ;;· t::::i .g"' ~ ;;:=-­ .... ~ Si: f Cb i. Cb .., ..... ..... The dorsal and ventral parts are checked for each of the first five abdominal seg· ments, but only the genitalia (G) are plotted for the genital and anal region. In gynandromorphs showing a bilateral arrangement in the abdominal region the genital part is marked plus and minus, respectively, for the female and male sides, unless the apparatus is mixed, in which case the plus-minus symbol is used on both sides. For ease of reference there has been placed in the first column on the left of each table a series of letters. This makes it possible for the reader to find at once any specimen cited in the text. Thus specimen 123-a, listed in Table 3, is referred to in the text as T. 3 S. The records for the abdominal regions of the mosaics listed .in Table 7 are omitted because the male parts of these specimens were limited to the cephalic and thoracic regions. The records for the cephalic and thoracic regions for the mosaics listed in Table 8 have likewise been omitted because they contained no male parts. DESCRIPTION OF GYNANDROMORPHS In considering the array of 335 specimens as arranged in Table 1 we shall first describe the half-and-half mosaics, and then take up the other types. By following this method one is able to determine more accurateJy whether a given specimen be­longs to the three-quarters, one-quarter or the partial group of gynandromorphs. The following statement will, we hope, make this point clear. If a fly has male tissue on one side and female tissue on the other, with the line separating the two areas exactly median, then the mosaic is obviously half male and half female and represents a perfect bilateral gynandromorph. If the head and thorax are male and the abdomen female, or vice versa, it may not be so obvious that the mosaic is likewise half male and half female. But a study of many mosaics makes it clear that this is probably the correct interpretation for such anterior-posterior gynandromorphs. On the basis of this statement it follows that if a fly has the male tissue restricted to one side of either the cephalic-thoracic region or the abdomen, it belongs to the one-quarter group of gynandromorphs. In the complementary or three-quarters group the cephalic-thoracic region plus one-half of the abdomen (or else the re­verse) must be composed of male tissue. Any fly which is less than one-fourth male is placed in the partial group. Mosaics with the two tissues mixed or irregularly distributed are more difficult to deal with. It is for this reason that we shall discuss in a later section the question of the relation of the male and female areas in the terms of imaginal discs. 1. HALF-AND-HALF GYNANDROMORPHS It will not be necessary to describe each individual mosaic, but for purposes \)f description they have been arranged into groups. Each group includes all mosaics which showed the same general distributional pattern of male and female parts. This same arrangement will be found in tables 2-8. Thus, in considering the half-and­half mosaics, it will be convenient to establish three groups as follows: A, all speci­mens in which the male and female parts are approximately bilateral in their ar­rangement; B, all specimens which show an anterior-posterior arrangement; C, speci­mens in which the two parts are mixed or irregularly arranged. A. Approximately Bilateral Specimens A large number of gynandromorphs may be placed in this group, although some of these specimens diverge quite widely from the perfect bilateral mosaic. As a matter of fact, a perfect bilateral gynandromorph is a rather rare specimen, and yet flies approximating this condition are not uncommon. There were 24 specimens which without careful inspection would have been classed as perfect bilateral:;. However, when these flies were examined in detail it was found that only six of them fulfilled exactly the conditions of a perfect bilateral sex-mosaic. These are plot ed on Tables 2 (W-Z) and 3 (C, D), and a typical case is shown in Plate I, fig. 1. The inspection revealed that in eight other mosaics the male or female area ex­tended across the median line at some point. Often the extension of the male area across the median line into the female half is counterbalanced by the invasion of a corresponding amount of female tissue into the male half. Consequently, and in spite of the disturbance of bilateral symmetry, the fly is still approximately half male and half female in so far as the external surface is concerned. These devia­tions are clearly brought out in the tabulated data (T. 2 a; T. 3 A, B, E, G-J). The remaining mosaics of this group were likewise bilaterals, but the absence of distinctive gene markers made it inadvisable to list them in the tables. The compensating shift of tissue between the two halves was found to be common among gynandromorphs, and in this paper it will be referred to simply as compen­sation. It gives a very striking contrast where the shift involves the derivitives of two adjacent imaginal discs. Thus in the mosaic illustrated in fig. 2 compensation is seen in the fact that the two halves ( tergite) of the fourth abdominal segment are reversed with respect to the btlateral distribution of the male and female tissues. Fifteen of the mosaics belonging to the next general type of bilateral gynandro­morphs are listed in Table 3, K-Y. In a typical case the head and one side of the body are female, while the other half of the body is male. In eight cases the male side was on the left, and in seven cases it was on the right. In twelve flies the head was entirely female, but in the other three there was a small amount of male tissue present in the cephalic region (T. 3 Q, R. Y). In the matter of compensation, ten of these flies showed male tissue in some form on the female side. A very good example of this is seen in specimens 78-c and 175-a. In these two flies the head and most of the left side of the body were female, but the male area extended across the median line in the region of the scutellum, which was entirely male (fig. 3, Pl. I). In the mosaics of this type the male area is somewhat less than one-half of the surface of the fly. The next type of mosaic is very similar to the one just described, but shows a greater degree of compensation. In these forms the head (or greater part), one half of the thorax, and one-half of the first three or four abdominal segments are female, while the rest of the body is male (T. 4 A). Specimen 110-c, shown i.n fig. 4, is a typical case. Three other mosaics were very similar to this one, except that the female tissue was on the left instead of the right side, and in two of them there was a small amount of male tissue present in the cephalic region. Two other mosaics varied slightly from the type, in that only the fifth abdominal segment was all male (T. 4 C, D). One of these is shown in fig. 60, Plate VII, and another is illustrated in fig. 64. ,.,. """" TASLE 3 t ~ ~· <'ll ... "' ~· ~ ~ e i;:: """ 0­ ....... ~· ~ s· ;::s Fourteen gynandromorphs constitute the last general class of mosaics listed under the bilateral group. They are characterized by having the head composed entirely (one minor exception) of male tissue and with all or the greater part of one side also male. As a matter of fact these mosaics belong to two types rather than to a single class. In one type consisting of five flies the male half of the body fails to reach the tip of the abdomen by one or two segments, so that one or two of the abdominal segments are female in character (T. 4 E-H). Specimen 274-f (fig. 5) is a good example of the type. The type to which these five mosaics belong is, in the main, complementary to the preceding type ( c. f., figs. 4 and 5), and in both types the nature of the compensation is such that the male and female areas are about equal to each other. Of the nine flies belonging to the second type of mosaics with male heads, four were very similar. In two of the four specimens (185-a, 13-y, fig. 71) the left half of the body throughout was male, and in two (161-c, 161-h; T. 4 L, M) the right half was the male side. The other five flies exhibited certain variations in this pat­tern. In specime!'1 113-a (fig. 6, T. 4 N) the head and right half of the body were signed and male, except for the 2nd and 3rd legs which were non-singed and female. Specimen 112-c had an interesting criss-cross pattern; the head, right half of the thorax and the left half of the abdomen were yellow and male, while the other parts were all gray and female (fig. 7; T. 4 0). In specimen 247-1 (T. 4 I) the left half of the body was yellow and male throughout, but only part of the head was composed of male tissue. The left eye was white-lozenge (two of the marker genes) with two small red spots near the center, and the right eye likwise contained a red area; the right antenna was yellow, but otherwise the head was grey. In speci­men 142-f (T. 4 J) the head and most of the left half of the body were forked and male, the 2nd and 3rd legs forming the exception; hut on the right side of the body, the fore-leg and a small area at the tip of the wing were composed of male tissue. The ninth and last specimen of this series (288, T. 4 K) showed the widest variation. In this mosaic the head and left half of the body were yellow and male, including the genitalia; and on the right side the following parts were also male; fore-leg, humerus and 3rd leg. In a majority of the flies of this type the male region is slightly greater than one-half of the surface of the individual. B. Anterior-Posterior Gynandromorphs Eighteen gynandromorphs are placed in this group, but they clearly belong to two complementary types. In the typical specimen of the first type, the head, thorax and all appendages are male, while the entire abdomen is female, and in that of the second type this arrangement of the male and female parts is reversed. Eight mo­ saics belong to the first type and ten to the second. Specimen 269-d showed the typical distribution of the first type (fig. 8; T. 4 P), and a second fly (155-f; T. 4 Q) varied hut slightly in having a red spot on an otherwise white eye and a mixed left fore-leg. Specimen 320-c was male for the head, thorax and appendages, but in addition the left halves (tergites) of abdomiml segments 4 and 5 were smaller and showed male coloration. Another mosaic (R, T. 4 R) was forked and male for the head, thorax, fore-legs and wings, but the ....... 0\ TASLE 4­ ~ ~ ;:s ..... <:! ~ ..., "' ~· ~ f '"t1 ;: O" ~ ~ c;· ;:s other four legs were non-forked and therefore female. The upper surface of the fourth abdominal segment showed male coloration, otherwise the abdomen was female. Specimen 308-c showed the typical distribution, except that the left half of segment 5 had male coloration (fig. 65, Pl. VII) . The anterior half of specimen 272-b had a considerable amount of female tissue, but the abdomen was entirely female (T. 4 S). Finally, the two remaining mosaics were still more variable, as may be seen from the tabulated data (T. 4 T, U). In the second type of anterior-posterior gynandromorphs, as indicated above, the relation of the male and female parts is reversed-the head, thorax and its ap­pendages are female, while the abdomen is male. There were four mosaics which showed this typical arrangement (fig. 9; T. 4 V, W), and two others that were nearly typical, but each had a small patch of female tissue on the dorsal surface of the abdomen (T. 4 X, Y) . Specimens 74 and 273-c also showed the typical distri­bution, except for the genitalia. In 74 the genitalia were double, with a complete male apparatus above, and a female apparatus below, and in 273-c the genitalia were mixed. In specimen 151-a the entire abdomen was singed and male, but in the anterior half the left wing and left postalars and scutellars were also singed and therefore male (T. 4 Z). Specimen 70-c represented the most extreme variant, with the entire abdomen and all six legs showing forked and male (T. 4 a). C. Mixed Half-and-Half Gynandromorphs There were a number of mosaics in which the arrangement of the male and female parts was in the main too irregular to justify placing them in either of the two preceding groups, a'iid yet a study of the record cards shows that some of these belong to the half-and-half group of gynandromorphs, in the sense that the total male area is about equal to one-half of the area of the fly. To attempt to describe even a majority of these flies would require entirely too much space. Five of these mosaics have been selected for the illustrations and appear in figs. 50--52, 54, Plate VI and 63, Plate VII, and eleven are listed in Table 5 (A, B, D, E, G-L, N). If the figures and data are examined it will be observed that while the distribution of the male and female tissues is irregular, yet it shows a distinct tend­ ency toward a bilateral arrangement (figs. 50, 52, 54, 63). This is also true for fourteen other flies which are not illustrated. In four mosaics the distribution is more nearly that of the anterior-posterior type, as will be seen by examining fig. 51. 2. GYNANDROMORPHS ABOUT THREE-FOURTHS MALE A group of sixteen mosaics which were approximately three-fourths male will be referred to in this section. We shall not attempt to describe in detail these specimens, especially in view of the fact that ten of them are illustrated and briefly described in the plates (figs. 10--16, Pl. II, and 59, 61, 62, Pl. VII) , and all except two are listed in Table 2 (A-C, H, J, L, M, 0, P, R-V). It is interesting to note that in twelve cases the abdomen was entirely male. Figures 10--15 of Plate II and 61 and 62 of Plate VII illustrate eight of these flies. Four other gynandromorphs had ab­domens that were male except for the genitalia, which in two cases were female and in the other two mixed (figs. 16, 59). >-' O:> TA&LE S - ~ c::: ;:s ~­ (b ;:i .... ~ ~ ~ ~ ~ O" ~ e c:;· ;:s 3. GYNANDROMORPHS ONE-FOURTH MALE OR LESS We shall describe in this section a few of the many flies which are listed under the two groups in Table 1 as one-fourth and partial gynandromorphs. As a matter of fact the two groups grade into each other, so that it was not easy to classify a considerable number of the recorded cases. In figures 19-49, Plates 111-V, are shown 31 of these mosaics. These were selected with the view of illustrating the more common patterns found in our material. In one of the common types of one-quarter mosaic one side of the anterior half of the fly is male. A typical case is shown in fig. 19, and other specimens which approximate this condition are listed in Table 7 (A-E, G-1). Another common type is illustrated in figs.20 and 21. In these mosaics the head is entirely male, and the greater part of one side of the thoracic region is also male. Of the two flies illus­trated, one has a mixed fore-leg on the male side (fig. 20), and the other has a female 3rd leg on the male half (fig. 21). A third common type has a female head with the male tissue restricted mainly to one-half of the thorax. These flies are clearly less than one-fourth male, except in two cases which had male tissue on the abdomen. Specimen 384-f represents one with male tissue on the abdomen (fig. 22). On the basis of our description of the half-and-half gynandromorphs, one-half of the abdomen should represent about one-fourth of the fly. We have therefore placed under a fourth type a series of mosaics in which the abdomen was about half male and half female, and in which the arrangement of the two parts was usually more or less bilateral. However, there were many variations from this standard type. In fact only seven specimens had the male area occupying exactly one side of the abdomen (fig. 23), and in only three of these were the genitalia half-and-half; the other four had female genitalia. In two other specimens the abdomen was exactly half male, but the genitalia were entirely female, a condition counter-balanced by the presence of male tissue in another region (figs. 24, 26). There were four gynandromorphs with one half of the abdomen and all of Eeg­ment 5 composed of male tissue, and another similar specimen had, in addition, :l male patch on the scutellum (fig. 25). There were four flies with half of the abdo­men and all of segments 4 and 5 included in the male area, and four others in which the 2nd and 3rd legs were included in the male region (fig. 27). In Table 1 there are listed 127 mosaics which were classified as partial gynan­dromorphs. They constitute a group in which the total male area was estimated to be less than one-fourth of the external surface of the fly. From the very nature of these mosaics, the male parts are frequently represented by two or more male areas which are often widely separated by intervening female tissues. It is a fact worth recording, and one which will be commented on later, that in none of these gynan­dromorphs did we find two or more male areas of different phenotype. In Plate IV are shown thirteen specimens with the male area of less than one­fourth of the fly. There were fourteen mosaics with the male tissue limited to the head region (T. 7, X-Z, a-k), and in none of these cases was the entire head male. Figures 28-32 illustrate five of these partials. In eight flies the male area was limited to one side of the head (figs. 31, 32; T. 7, d-k). Male tissue was found in the cephalic and thoracic regions in eleven specimens. Two of these were male for the right half of the head and all or a part of the right fore-leg (figs. 33, 34) and five others are listed in Table 7 (n-r). There were only two partials which had male tissue in the head and abdominal regions (T.6, R,T). There were fifteen partial gynandromorphs with the male tissue restricted to the thoracic region and its appendages. In five of these the male area showed on one side of the thorax only (T.6 U-W, T.7 s,t). In each fly it occupied one-half of the dorsal surface, including the sternopleura in two cases and not including the humerus in a third case. In all five mosaics the legs were entirely female in character, but in four flies the wing on the male side was of the male-type, although in one case it was mixed yellow and gray (fig. 36) . Three other mosaics had male tissue on the thorax but in each case it was less than one-half the surface (T.6 X,Y; T.7u) . Tn five mosaics the male tissue was limited to a single wing. One small yellow female (200-a) had an imperfect sex-comb on each fore-leg, and one gray female (141-a) had a tiny gray sex-comb on the right fore-leg. The latter was fertile, but the former was sterile and may have been a female-type intersex. Seven mosaics showed male tissue on various parts of the thoracic and abdominal regions. Two of these are illustrated and described in figs. 38 and 40 (T.6 O,P). Three specimens were similar in that two separate male areas were located on the thorax and one on the abdomen. The first of these had a male fore-leg and male­type wing on the right side, and rotated male genitalia. The second specimen had a male fore-leg and male-type wing on the left side, and male genitalia. The third mosaic (T.6 Q) is illustrated in fig. 70, Plate VIII, where it is briefly described. Specimen 314-e had the yellow male parts as indicated in Table 6,S. Mosaic 269-a was a gray female with a tiny gray sex-comb on the right fore-leg and double geni­talia, female above and male below and both gray. A single partial gynandromorph had male tissue in all three regions; left eye, left fore-leg and mixed genitalia. There were 77 gynandromorphs with the male tissue· located at or near the posterior end of the abdomen. Most of these were clearly partials, but a few belonged to the border-line between one-quarter and partial specimens. The main distributional patterns under which they will be described are illustrated in figs. 41-49 of Plate V. In one specimen the three posterior segments were yellow and male, although the genitalia were mixed (T.8 0, fig. 41). Another mosaic (714-c) had tergites 2-5 on the right side yellow and male, but all other parts of the abdomen were gray and female including t:he genitalia. Figure 42 illustrates a fly with the left half of segment 3 and both sides of 4 and 5 showing male characteristics (T.8 Q). Mosaic 301-b was yellow and male throughout the left half of segments 3-5 (T.8 P). There were six mosaics in which segments 4 and 5 showed male coloration (fig. 43). In four of these the male area was limited to the dorsal surface or ter­gites, and hence the ventral side and genitalia were female, while in the other two the entire segments were composed of male tissue, including the genitalia (T.8 R). Figure 44 shows one of two mosaics with one-half of the 4th segment and all of segment 5 composed of male tissue; both had male genitalia (T.8 S). One of the most common patterns found among these partial gynandromorphs is shown in fig. 45. There were no less than twenty-one flies which showed this pattern; fourteen with the male area located on the right side and seven with it on the left side (T.8 T-V). TA6LE G ~ ~ 5 ~ .... 0 ;! .;: 0 ~ s· t::i .... 0 .g"' ;::.-. ~ .,, ~ 5 ~ ~ .,, ~ .... t-.:1 ..... A total of eleven mosaics showed male coloration for one-half of segment 4 (fig. 46), and here the male area was located on the right side in nine cases (T.8 W) and on the left side in only two cases. The male tissue seemed to be limited to the tergite and did not involve any other part of the abdomen. One fly showed a criss-cross pattern of the male area for the 4th and 5th segments (T.8 X, fig. 47). The 5th abdominal segment was male in five gynandromorphs (fig. 48) but the genitalia were mixed in two of these (T.8 Y), male in two others and female in one (T.8 Z). In another large group of twenty flies one-half of the 5th segment showed male coloration (fig. 49); ten on the right side and ten or the left (T.8 a-e). Finally seven mosaics had male genitalia, but otherwise they were female (T.8 f-h). Three of these flies were found in the control or untreated series and in each case the paternal X had been lost. Of the four from the X-rayed series, two had the treated X and two the untreated X missing. A very interesting fact has developed with reference to this group of gynandro­morphs. Thirty-eight, or about one-half of them, were obtained in a series of experi­ments in which one of the X-chromosomes employed in the combination contained a long inversion. The stock carrying this inversion is known as scute-8. From the character of the male area it was noticed that almost without exception this area appeared to have lost the partner X-chromosome {marked with yellow in all cases), and this was true no matter which parent furnished the scute-8 chromosome, and the loss was also independent of the x-ray treatments. The results obtained were as follows: x-rayed scute-8 apricot females crossed to yellow males yielded ten mosaics in which the male area was gray, and none in which it was yellow, that is the un­treated paternal X was absent. X-rayed scute-8 apricot males crossed the female~ homozygous for yellow gave six mosaics with gray male areas, and none with yellow areas. In the control or untreated series, scute-8 apricot females crossed to yellow males gave eighteen mosaics with gray areas and two with yellow. Scute-8 apricot males crossed to yellow females yielded two mosaics, both with gray male areas. These results show that in thirty-six out of thirty-eight mosaics, which had received the scute-8 chromosome, the male area retained this element, and that on the basis of the elimination theory the partner X-chromosome had been lost. Moreover, what­ever may have been the underlying cause of their production it can not be attributed to the effects of x-rays. Many of these mosaics with the male tissue restricted to the tip of the abdomen were due to the loss of the yellow-bearing X-chromosome, and hence the area showed the black male coloration. This in part accounts for the reason why so many specimens of this type were detected and recorded. Such small male areas occurring in other regions of the fly could easily be overlooked. 4. GYNANDROMORPHS WITH BROKEN X-CHROMOSOME IN MALE PARTS In this section we shall describe in more detail the type of gynandromorph which carries in its male parts a broken daughter X-chromosome, in addition to one com­plete X. As shown in Table 1, there were fifty-two such mosaics, of which forty­nine were obtained from treated fathers and only three from treated mothers. In the treated male series, there were six specimens of the three-quarters class, twenty­five of the half-and-half class, thirteen of the one-quarter class and five of the partial class. In the treated female series, one was half-and-half and the other two were one-fourth male. In every case the broken X had been treated with x-rays, and the records show that seven different types of X-chromosome were involved, as fol­lows: twenty-seven normal X's, fifteen Theta X's, six scute-8 X's, and one each of scute-1, eosin-signed, eosin-miniature and forked-bar. These mosaics are therefore produced irrespective of the type of X-chromosome used. The fact that larger num­bers of them were obtained from treated normal and Theta-X chromosomes is be­cause these two chromosomes were the ones most frequently employed in the experi­ments. The general nature of the breaks which produce the extra X-chromosome frag­ment found in the male parts of these gynandromorphs has been discussed elsewhere (Patterson 1931, 1938). A brief statement covering this point will therefore be suffi­cient. There are at least two types of chromosomal aberrations produced by X-rays which are responsible for the formation of these duplication fragments. In some cases the fragment has been produced, apparently, by a single break in one of the daughter chromosomes, resulting in the loss of the section lying distal to the break and the retention of the proximal section with its spindle-fibre attachment. In other cases it has been produced by two breaks, resulting in the elimination of the section lying between the points of breakage and the retention of the two ends, which unite to form the duplication. The type of fly developing from such aneuploid zygotes will depend upon two factors; first the condition of the gametic chromosome at the time of treatment, and second, the location of the section of the chromosome which has been eliminated. If at the time of treatment the gametic element is in effect in the single strand stage, the deficiency created by the e~imination will be transmitted to all of the descendant cells. This will result in the production of either hyperploid males or hypoploid females, but not of mosaic flies. As to which of these two types of individuals will arise will in turn depend upon the section of the X-chromosome that has been eliminated. Our results show that whenever the middle section is lost, the zygote if viable will develop into a hyperploid male, but if sections from either end are eliminated, the zygote if viable will develop into a hypoploid female. If at the time of treatment the gametic element is in effect in the two strand stage, and only one of the two strands is broken, the deficiency created will be trans­mitted to only a part (usually a half) of the descendant cells, due to the segregation of the affected and unaffected daughter chromosomes at a cleavage division. Here again the type of fly produced will depend on the section eliminated, but in any event it will be a mosaic, a sex-mosaic or gynandromorph if the middle section has been lost, a mosaic female if certain other sections have been eliminated. We have selected for special consideration eleven of the fifty-two gynandromorphs, and these appear in figs. 59 to 69 where they are described in the legends. The diagram of the X-chromosome and the duplication fragment beneath each figure gives our interpretation of the composition of the male parts of the mosaic. The upper line represents the untreated marked X, while the lower one shows the frag­ment with the deficient section indicated by a broken line. A diagram for the female parts is not given, but in each case it would show two unbroken daughter chromo­somes, the untreated marked X and the treated and usually unmarked X. TABLE 7 SPECIMEN ~60-b A g 381-a 319-a c lll-b D 285 E F 345-a 192-d G 152-c H 711-c I 311-a J K 317.-i 187-b L 312.-c M w N 340-d 0 p 384-d Q. 336-a 301-k R s 297 T 12.3-b u Z02-b v 301-i w 114-e x 19-w y 191.-a 303-f l 313-a a 316-c b 40 c d 306-a 38S-a e f 40-e g 16S-d 114-d h i 118-a j 156-a le 24-e I 301-j m 2.44-h n 2.14-c 0 329-a 44 p q 302.-c r 7.14-g s 31S-d 169-a t 184-b II HE.AD LEFT RIGHT An Pb Op An Op - - - + + - - - + + + + ± + + - - ± + + - - ± ± - + + ± + + - - + + + - - + + + - - + + + + + + ± + - + + + ± - - - - - - - - - - - - - - - - - - - - - - - - ± - - - ± + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + - - - - ± - - - -+ - - + + + - - + + + - - + ± + - - - ± ± - - - + + - - - + + - - - + + - - - + + - - ± + + + + + ± + + + + + ± + + ± + + - - - - - - - - - - SI S2 S3 SI Sl 53 D y D v v D v D v v ~ ~.. H LI w M Sc L2 L3 H LI w M Sc L2 L3 u + + ------+ + + + + + + 1 + + + -----+ + + + + + + G + + ------+ + + + + + + s + ----+ ± + + + + + 8 - -----+ +--+ + + + + 9 + -± -± -+ ++ + + + + + + 3 - + + + + + --+ + - -- --8 - + + + + + --+ + - -- --8 - + + + + + -±+ + - -- -+ 6 + + + + + + --+ + - -- --6 - + + + + + + ++ + - -- + + s - -----+ +--+ + + + + II - -± -± -+ +--+ + + + + II - -----+ +-+ + + + + + IQ - --± --+ +--+ + + + + II - + -+ + + ± ---- -- + + II - --± ± + + +--+ + + + + 11 + -+ ---+ +--+ + + + + 4 + + ----+ +--+ + + + + 4 + + + + + + --+ + - -- --s + + + + + + --+ + - -- -+ 4 + -----+ +--+ + + + + 5 + --+ --+ +--+ + + + + s - + + + + + + ++ + + + + + + 6 - + + + + + + ++ + + + + . + + 5 + + + + + + + ++ + + + + + + 3 - + + + + + + ++ + + + + + + s - + + + + + + ++ + + + + + + 4 - + + + + + + ++ + + + + + + 6 + + + + + + + ++ + + + + + + 3 + + + + + + + ++ + + + + + + 3 + + + + + + + ++ + + + + + + 3 + + + + + + + ++ + + + + + + 3 + + + + + + + ++ + + + + + + 3 + + + + + + + ++ + + + + + + I + + + + + + + ++ + + + + + + 1 + + + + + + + ++ + + + + + + 'L - + -+ + + -++ + + + + + + 8 - + --± + + ++ + + + + + + 1 + -+ + + + +--+ + + + + & + + + + + + + ++ + + + + + + 3 + + -+ + + + +--+ + + + + 4 - + + + + + + ±+ + - + + + + & - + + + + + + ++ + + + + + + 3 + -+ ± --+ ++ + + + + + + 'Z + -+ ---+ ++ + + + + + + 2. + + --± -+ ++ + + + - + + 3 - + + + + ± -- - + + - + + + + - - + + + - - + + + + + + + + + + + + + + + + + + THORAX Pb LEFT RIGHT Specimen 386-c (fig. 59, T.2 0) was three-fourths male in which the male parts showed yellow scute garnet and pleated of the untreated X, but not forked, while the female quarter was wild-type. The locus of forked was therefore covered. We inter­pret this to mean that the treated daughter X had been broken between the loci of pleated and forked with the section y-pl eliminated and the fragment retaining .its own spindle-fiber attachment. Specimen 305-b (fig. 60, T.4 C) was a half-and-half mosaic showing compen­sation, with female parts wild-type. The male half showed only the mutant charac­ter forked-5; three of the other mutant markers were covered, but the locus of garnet may not have been covered; since the head w&,s not involved in the male area, one could not be certain. Our interpretation is that the duplication fragment had been produced by a deletion which eliminated a section including the locus of forked. Specimen 29-t (fig. 61, T.2 J) was about three-fourths male, with the male parts showing the mutant characters miniature and forked-5. The loci of yellow and cross­veinless were covered by the duplication. The locus of white was also probably covered, since it lies beyond crossveinless in the long inversion of the scute-8 chromosome, which was the treated element. The female quarter was gray and the eyes were heterozygous apricot. The duplication in this case was produced by a deletion which eliminated the m-/5 section. Specimen 387-i (fig. 62, T.2 T) is another mosaic that was three-fourths male. The female quarter was wild-type, but the male parts showed the mutant characters garnet, pleated and forked. Yellow and scute, the other two markers, were covered by the duplication. The fragment was produced by a deletion which removed at least as much as the g-/ section of the treated X. Specimen 253-j (fig. 63, T.5 E) was a mixed half-and-half gynandromorph with the male parts showing the mutant characters forked, fused and carnation, and with yellow, scute and miniature covered. Since the locus of miniature is covered it is probable that the loci of yellow and scute are also present in the fragment, even though they are covered by the small duplication attached to the spindle-fiber end of the Theta X. We therefore conclude that this fragment resulted from a deletion which eliminated the /-car section. Specimen 303-i (fig. 64) was about half male and half female with both parts gray. The limits of the male area were determined by the smaller bristles and smaller size of the right side, and by the male genitalia and coloration of the tip of the abdomen. Moreover, the right eye had a garnet spot and the right wing was miniature. The other three mutant markers, yellow, scute and forked-5, were covered by the duplication, which was produced by a deletion which eliminated a section including the loci of miniature and garnet. Specimen 308-c (fig. 65) was an anterior-posterior gynandromorph, with the an­terior half male and showing the mutant characters scute, miniature and garnet; yellow and forked-5 were covered by the duplication. The abdomen was nearly all wild-type female. The extra fragment in the male parts of this mosaic was pro­duced by a long deletion which eliminated the sc-g section of the treated X. Specimen 302 (fig. 66, T.6 A) was about one-fourth male. The male parts showed the mutant characters furrowed, garnet and forked, but miniature was covered by the duplication; the fw-f section was therefore eliminated. Specimen 306-a (fig. 67, T.7 d) was a partial gynandromorph which was nearly all wild-type female. However, the left half of the head was smaller, had forked-5 bristles and a small garnet, male-type eye, and the absence of the first and second orbital, postvertical and ocellar bristles on this side showed that the locus of scute was deficient. We therefore conclude that the fragment in the male area of this fly was produced by a long deletion which eliminated the sc-f section. Specimen Z (fig. 68) was a mosaic with the right half of the head and body and the upper surface of abdominal segments 4 and 5 showing male characteristics. The right wing and part of the left were miniature. The left eye was heterozygous for eosin and bar, but the right eye was hemizygous bar and light e6sin. Either the treated y w f B X-chromosome had been broken between the loci of miniature and bar and the entire left-hand section lost, as shown in the diagram, or else a deletion eliminating the middle region of the chromosome had occurred. TA8Ll 8 Specimen 190-c (fig. 69) was slightly greater than one-fourth male. The male parts, as revealed in the figure, showed the mutant characters crossveinless and minia­ture, but the locus of forked was covered, and since it is a broken Theta X, either the y-m section was lost, as indicated in the diagram, or else a deletion eliminated the middle section of the treated X-chromosome. The eleven flies just described were selected with the view of exhibiting a sample of gynandromorphs which carried duplication fragments of the X-chromosome ::n their male parts. This group raises several interesting questions concerning the nature of mosaic formation, but we shall defer a discussion of such questions until the last section of the paper. COMPLEX AND UNDETERMINED CASES A few mosaics were found that could not be explained on the basis of the simple elimination theory. Three of these were complex cases, and six were mosaic males. In addition to these mosaics there were five others, all belonging to the gynandro­morphic series, in which it was not possible to determine which of the two X-chromo­somes was absent from the male area. We shall first describe the three complex cases. Specimen 219-a. Obtained in a cross between x-rayed Theta females and y sc m J car males. It is shown in fig. 72. The male half was gray and included the entire head and the left half of the thoracic and abdominal regions, except the fore-leg and 5th segment, and all of the scutellum. The left wing was male-type and the left half of segment 4 showed the black male coloration. The right side was yellow-scute and female and the external genitalia were female. It was mated but proved to be sterile. The male half of this fly must have received the Theta X from the mother, and the female half must have received two yellow-bearing X-chromosomes. It can be ex­plained best on the theory that the egg from which it arose was binucleated. On the basis of this assumption, one reduced nucleus received the Theta X and the other nucleus received a yellow scute X. Fertilization of the first nucleus by a Y-bearing sperm would produce the gray or male half of the fly, while fertilization of the sec­ond nucleus by an X-bearing sperm would account for the yellow-scute or female half. Double elimination would require that the paternal X-chromosome was lost from the male portion of the mosaic and that the Theta fragment was broken off the maternal X-chromosome in the female section. Specimen 13-y. This gynandromorph was obtained in a cross between x-rayed scute-8 apricot females and y w cv m JS males. It is illustrated in fig. 71, Plate VIII. The male half was gray and scute-8 and included all of the head and exactly the left half of the thoracic and abdominal regions. The eyes were hemizygous apricot or male-type. The right or female half was yellow and forked-5 and the right wing showed yellow crossveinless and miniature. The mixed genitalia were gray and male on the left and yellow and female on the right. Since the genitalia were abnormal, breeding tests could not be made. In this mosaic the male half received the scute-8 maternal X, while the female half evidently contained two y w cv m JS paternal X-chromosomes. This case can be explained on the basis that the fly arose from a single nucleated egg, provided non­disjunction and elimination occurred at the first cleavage division. On this assump­tion, one nucleus would have to receive two paternal daughter X-chromosomes only, while the other nucleus would have to receive a single maternal X. It can also be explained on the basis of a binucleated egg, but it would be necessary to assume that a non-X non-disjunctional egg nucleus was fertilized by a non-disjunctional two-X sperm. Specimen B. Obtained in a cross between yellow females with attached X-chromo­somes and Theta males, both parents untreated. The fly was mostly yellow female, but there were three separate gray regions which were male in character. These three regions were the left half of the scutellum, the inner third of the right wing and the right half of segments 4, and 5, including the genitalia (fig. 70). The yellow female regions must have received the attached X-chromosomes from the mother, while the gray male regions must have received the Theta X from the father. The simplest explanation for the origin of this mosaic is the theory of binucleated egg. One reduced nucleus with the attached X-chromosomes fertilized by a Y-bearing sperm will account for the yellow parts; the other nucleus with a Y-chromosome fertilized by an X-bearing sperm will explain the scatter gray male parts. The case may also be explained on the bases of a single nucleus fertilized by an X-bearing sperm, if double elimination is assumed to occur. The failure to determine which X-chromosome was missing from the male area of five mosaics was due in part to the location of this area and in part to inadequate gene markers. A brief description of one case will make this point clear. The male area of specimen 73-b was restricted to the right half of tergite 4 which was black. The fly came from a cross between miniature females and white forked males. Since the eye where white could show was not involved in the male area, the only marker upon which to depend for the determination was forked, which happened to be a weak allelomorph that did not affect the setae of this region of the abdomen. The elimination of either X would account for this case. Among the many different types of mosaic flies obtained in the different experi­ments were six mosaic males. These flies were entirely male, but each showed two contrasting areas, the character of these being determined by the type of X-chromo­some present in the tissue. In four cases the nature of the two areas revealed that one must have received a paternal X and the other a maternal X-chromosome. In the fifth case both areas had received a maternal X, but one of them contained in ad­dition a fragment of the treated paternal X. In the sixth case the reverse of this was true, for both areas contained a paternal X, and in addition, one carried a frag­ment of the treated maternal X. These flies are not true gynandromorphs, but on account of their unusual interest it seems worth while to describe them and to offer some suggestions concerning their possible origin. Specimen 4-x. Obtained in a cross between y w cv m /5 females and scute-8 apricot males, both parents untreated. The contrasting areas were gray versus yellow, the latter represented by stippling in fig. 73. The eyes were hemizygous apricot; the head, outer one-third of the thorax, proximal part of left fore-leg, ]st and 2nd legs on the right and the anterior margin of each wing were all gray. The inner two-thirds of each wing, a broad band along the dorsal surface of the thorax, distal part of left fore-leg, 2nd and 3rd legs on the left, right 3rd leg and the entire abdomen were all yellow and forked-5, both wings showing on their yellow parts the mutant characters crossveinless and miniature. This fly was fer­tile, for when mated to yellow females with attached X-chromosomes the cross gave twelve females like the mother and ten y W CV m JS males. The gray part of this fly received the scute-8 apricot X from the father, and the yellow part, which included the entire abdomen, received the marked X from the mother. Since the fly was fertile, the yellow region must also have contained a Y­chromosome. Among possible explanation that one may offer to account for this mosaic, the following are perhaps the most plausible: (1) It arose from a binu­cleated egg, in which one nucleus contained a r w cv m JS X, and the other, through non-disjunction or elimination, did not possess an X-chromosome. The first nucleus fertilized by a Y-bearing sperm would produce the yellow region, and the second nucleus fertilized by an X-bearing sperm would produce the gray region. (2) On the basis of a single nuclear egg, it would be necessary to assume that one pronucleus had an X and a Y by non-disjunction and the other pronucleus had an X, and it would be necessary further to assume that at the first cleavage division a paternal X was eliminated from one daughter group of chromosomes and a matern;tl X from the other group. Specimen 163-a. Obtained in a cross between Theta females and r w sn m B males, both parents untreated. It was entirely yellow, but the contrasting areas were non-scute singed versus scute non-singed, the former represented by stippling in fig. 74. The eyes were white and male-type bar; the head, left half of thorax throughout and left half of abdominal segment I through the anterior part of 3 were all non-scute and singed, the left wing showing the mutant character minia­ture; all other parts of the fly were scute and non-singed. This male was fertile, and when mated to yellow females with attached X-chromosomes gave twenty-three females like the mother and nineteen yellow scute males. The head and left side received the marked paternal X, while the rest of the fly received the yellow scute maternal X (partner of the Theta X) . The yellow scute part also contained a Y-chromosome. The same suggestions offered for the preced­ing mosaic will also explain this case. Specimen 1S7-a. Obtained in a cross between x-rayed Theta females and r w cv sn males. A yellow fly, with the contrasting areas scute non-singed versus non-scute singed. The head was non-scute and singed with small white male-type eyes; all other parts of the fly were scute and non-singed. This male was sterile. The head region received the marked paternal X and the rest of the fly received the yellow scute maternal X. Since the fly was sterile it might be concluded that a Y-chromo­some was not present, but inasmuch as the progeny derived from x-rayed parents are often sterile, such a conclusion does not necessarily follow. We would there­fore offer the same suggestions as for the preceding cases. Specimen 333-a. Obtained in a cross between r SC m g JS females and x-rayed Theta males. The contrasting areas were gray versus yellow, the latter represented by stippling in fig. 75. Both eyes were garnet, but the left half of the head, the entire thorax with appendages and the right half of abdominal segments 1-4 were all yellow and forked-5, the wings also showing miniature. The right half of the head and the left half of the abdomen plus all of segment 5 were gray non-scute and forked-5. This male was sterile. The yellow region of this fly received a marked maternal X, as did also the gray region, but the latter received in addition a fragment of the treated X, with its duplication covering the loci of yellow and scute (see diagram below fig. 75). On the basis that it arose from a binucleated egg, each reduced nucleus contained a maternal daughter X, and one nucleus was fertilized by a Y-bearing sperm (yellow region) and the other by a sperm carrying the fragment of the broken X (gray region). On the other hand, if it arose from an egg with a single reduced nucleus, then the egg would have to have been fertilized by a sperm carrying a broken and an unbroken daughter X-chromosome, the latter having been eliminated at the first cleavage division. Specimen 290. Obtained in a cross between yellow scute-4 females and x-rayed apricot males. The eyes were male-type and red, and the head and thorax, except the left 3rd leg, were gray and non-scute; the abdomen and left 3rd leg were yellow scute-4. This fly was sterile. The same suggestions as for the preceding mosaic will explain this case. Specimen 205-c. Obtained in a cross between x-rayed Theta females and y sc m car males. The contrasting areas were gray versus yellow. The right half of the thorax, including fore-leg and the wing on that side, was gray, and non-scute, the wing also being miniature; the rest of the fly showed all four of the recessive markers. This fly died before it had been tested for fertility. Because of its similarity to the last case, the same suggestions for its origin may be offered. There have been other cases of mosaic males reported in the literature. Morgan and Bridges described two mosaics which were entirely male, but their specimens differed from most of ours in that the X-chromosomes found in the contrasting regions were both derived from the mother (c.f., our 333-a) . They offered the alternative assumptions of binucleated and mononuclear eggs as a basis for their origin. Bonnier (1928) has described what appears to be one of these cases, although he was not able to demonstrate conclusively that the smaller area, representing one· half of the head, was male rather than female in composition. He discusses several possible explanations, and among these is the suggestion that the smaller area may have arisen from a fertilized polar body nucleus of the type described by Huettner (1924). Hyde and Powell (1916) have reported a case which has some bearing on the interpretation of these complex cases. This fly was a gynandromorph and not a mosaic male. This fly received from its mother an X-chromosome carrying the gene for blood, and a second X from its father carrying the ~ene for eosin. The entire thorax and abdomen were female, and according to expectation the eyes should have been compounds, lighter than blood but darker than the eyes of eosin females. Instead, one eye showed blood and the other was male-type eosin. This case has been explained as double elimination by L. V. Morgan, with the paternal X lost from one side and the maternal X from the other. Two other types of flies obtained in the experiments are also of interest. One of these is the aberrant male, which carries one complete X and the fragment of another X broken by x-rays. Such males are X-hyperploids, and during the course of the experiments a total of 252 was recorded. The other type is the mosaic female, which is X-hypoploidy. A large number of these were also secured, but an account of either of these types does not come within the scope of this paper. However, there is one mosaic female, which has been referred to in another paper (Patterson, 1938), that should be more fully described. Specimen 304-c. Obtained in a cross between y sc m g /5 females and treated wild-type males. The contrasting areas were gray versus yellow, the latter repre­sented by stippling in fig. 76. Most of the head and all of the right half were gray and female; the back of the left eye showed a patch of garnet facets. The entire left. half of the thorax and abdomen was y sc m but bristles were not forked; the left wing was miniature and notched; the left fore-leg did not have a sex comb; and this side was also female, including the genitalia, which were gray on the right side and yellow on the left. The fly mated and laid about twenty eggs which failed to develop. She was weak and upon becoming moribund was carefully dissected. The internal sex apparatus was found to be normal, each ovary containing eight or nine well developed eggs. Possibly the best explanation to account for the origin of this mosaic is as follows: The gray half received a paternal and a maternal daughter X-chromosome, while the yellow half inherited a complete maternal daughter X and the right-hand portion of a broken paternal daughter X, the section from yellow to garnet being lost (see diagram at bottom of fig. 76) . This would mean that one daughter chromosome had been broken at some point lying between the loci of garent and forked. The notched wing would indicate that the yellow half carried a deficiency. Certainly, an appeal to somatic crossing over as an explanation of the origin of such half-and­half mosaics can not be advanced, for as Huettner (1924) has shown the chromo­somes of the pronuclei enter the first cleavage spindle as separate groups, and an opportunity for mingling and pairing of homologous chromosomes is not present until the second cleavage division. That somatic crossingover may occasionally lie at the basis of the formation of certain mosaic females is indicated by the condition found in our specimen 27. This mosaic came from a cross between a yellow white female and al). x-rayed eosin singed male. It had a white area near the top of the left eye and directly opposite to this was a gray singed area of' about the same size. The twin spot thus formed could have arisen by somatic non-disjunction or by somatic crossingover. Goldschmidt (1937) has reported a case which could only have come from double fertilization. However, L. V. Morgan {1929) in discussing the exceptional com­posites known up to that time, came to the conclusion that double elimination was a more probable explanation than double fertilization where either could be the explanation. Both explanations have been offered for these cases of composites; in some the one seems more likely, while in others the alternative explanation would imply a less complicated action. These cases merely serve to indicate that peculiar irregularities very occasionally occur in fertilization and meiotic and mitotic mechanisms. It is difficult to relate the simple eliminations, breakages, deletions and these more complicated cases to the remarkable series described in corn by Jones (1937), although obviously several of the phenomena are identical. As the cases under consideration were gynandromorphs, or mosaic males, and not mosaic females (except the single case cited above) , somatic crossing over (Stern, 1936) as an explanation of such cases does not need to be discussed further. FERTILITY OF GYNANDROMORPHS The production of offspring by a gynandromorph will depend upon certain conditions. Among the more important of these are its sexual responses, the location of its male parts, and especially the correlation between its gonads and the reproductive tract. The sexual reactions of a mosaic are important because if it is indifferent to other flies and refuses to mate, obviously offspring will not be produced, even though its gonads might be able to produce functional gametes. Both Sturtevant ( 1915) and Spelt (1931) have studied the courtship behavior of males and females. Sturtevant first showed that the behavior in court ship was elaborate and quite different in the two sexes. He thus laid the foundation for the study of the sexual behavior of gynandromorphs. He tested six such mosaics and found that of two flies which were all male except for female genitalia, one behaved as a male and the other reacted to both males and females. One fly which had sex-combs on the fore-legs and one male-type wing gave both reactions. Of the three remaining flies, one was male for one side of the head, one was a bilateral mosaic and the third was female except for the male genitalia. These three flies were all courted by males, but none of them showed any certain indication of male behavior. Two other references to tests made on the sexual responses of gynandromorphs may be cited. Morgan (1915) has described a mosaic which was male for one side of the head and thorax and for most of the abdomen, including the genitalia. When placed with mature unmated females it failed to court them. Duncan (1915) has reported on the behavior of three gynandromorphs. The first was a bilateral mosaic throughout with abnormal genitalia. It was courted but refused to mate. The second fly was a lateral mosaic in which the male parts were restricted to the posterior part of the thorax and to the abdomen, and had mixed genitalia. It was also courted by males but would not mate. The third mosaic was bilateral for the abdomen and had mixed genitalia. It was courted by a male as long as the male remained in front of it. Except for a few casual observations, we did not attempt to make a study of the sexual reactions of the gynandromorphs which we obtained. However, Dr. Meta Suche did make such a study on twenty-six of our specimens in 1932. These results have not been published, but with her permission we are able to give here a summary of her observations and conclusions. The group of gynandromorphs tested by her were obtained from a cross between scute-8 apricot males and y w cv m /5 females, and the reciprocal cross. In each case the scute-8 apricot flies were x-rayed. The group of mosaics represented various combinations of male and female parts, from specimens three-fourths male to individuals with the male tissue limited to the genitalia or to a small area on the head. Fourteen of these are listed in our tables and six are illustrated in the plates. In quoting below Dr. Suche's tentative conclusions we have made no important changes in the wording of her original manuscript, except to interpolate, wherever possible, references to our table and figure numbers. "l. There is no definite correlation between the extent of the external female tissue on a gynandromorph and its reaction to males, nor the reaction of males to gynandromorphs. The one case of a gynandromorph with abdomen and genitalia wholly male, which was not courted, may be significant, but one case is scarcely sufficient to draw conclusions upon. Gynandromorphs with the entire abdomen, including genitalia, of female tissue were courted no more vigorously than gynandromorphs with abdomen part male; nor did they respond more readily to males. Bilaterally divided gynandromorphs were courted without discrimination on male and female sides. "2. The sex of the genitalia has no effect on the reactions of males to gynan­dromorphs or of gynandromorphs to males. Of the eleven gynandromorphs with female genitalia (T.3 B,M; T.5 F; T.6 T; T.7 k,X; fig. 34) all were courted; only two mated, four definitely avoided males (figs. 28, 58), and one actually threatened a male (T.3 B). Of the ten with male genitalia (T.2 K,V; T.3 W,Z; T.6 C; T.8 D); fig. 71), only one was not courted (fig. 13), four avoided males and one permitted mounting. All five gynandromorphs with mixed genitalia were courted (T.2 J; T.5 P; fig. 61). Of these, two avoided males and one responded positively. "3. The sex of the genitalia has no effect on the reaction of gynandromorphs to females. Of the two gynandromorphs which gave male response, one had female genitalia (fig. 58) and the other male (fig. 13). "4. The sex of the head has no effect on the reaction of males to gynandromorphs. Males made no distinction between male and female sides of the head. "5. The sex of the head may determine the reaction of the gynandromorph to females. Both gynandromorphs which gave male reaction had male tissue on the head. Specimen 10--m, which gave a slight male reaction, was of male composition throughout the right half of the head and thorax and the abdomen was entirely male. Specimen 25--a which gave a pronounced male reaction, had male tissue confined to two-thirds of the right eye and the right half of the abdominal segments 4 and 5. These two cases are in keeping with Whiting's observations on Habrobracon which led him to conclude that the sex of the head determines the sexual reaction of a gynandromorph. But the limited number of positive cases and the indifferent response of seven other gynandromorphs with the head part male, makes a similar conclusion as yet unwarranted in the case of Drosophila." Perhaps this brief review of the studies which have been made on a limited number of gynandromorphs will be sufficient to show that the sexual behavior of these mosaics is an important factor in determining their reproductive capacity. But their behavior pattern is in part determined by the nature of their structural composition. A mosaic which is all female, with only male genitalia, does not have the right kind of sensory receptors for carrying on the elaborate male courtship so necessary for successful mating. Nor can such a fly be bred to a male, even though it is persistently courted and possesses functional ovaries, because of the mechanical difficulties of copulation. The same would be true for mosaics with mixed genitalia. A mosaic which is completely male for the abdomen is usually sterile, irrespective of its sexual behavior, because the composition of its male tissues is XO. The lack of correlation between the gonads and the rest of the sex apparatus is another frequent cause for the failure of gynandromorphs to produce offspring. Duncan (1915) seems to have been the first to suggest that the gonads of gynandro­morphs in D. melanogaster always belong to the same sex, regardless of the somatic condition of the abdomen. This was, however, only a tentative conclusion, which he based on an examination of five specimens. The fact that the two gonads were always the same he thought could be explained by supposing that the gonads were always derived from a single embryonic cell, which in a mosaic must be either male or female producing. Four years later, Morgan and Bridges (1919) confirmed Duncan's conclusion and added twenty additional cases. They state that an examina­tion of the gonads in Drosophila gynandromorphs has shown that in every case the gonads are the same, regardless of the somatic distribution of the male and female parts of the mosaic. These conclusions were in line with the conventional explanation of the origin of the gonads in Diptera, especially in such forms as Miastor and Chironomus in which Kahle and Hasper, respectively, had shown that the germ cells arose from a single polar nucleus. But Huettner ( 1923) soon reported that the germ cells of D. melanogaster had a polynuclear origin. He showed that at the time the cleavage nuclei begin to migrate to the surface to form the blastoderm, from five to eleven nuclei appear simultaneously in the posterior polar plasm, where they give rise to the primary germ cells. In the meantime Sturtevant (1921) had discovered several gynandromorphs in D. simulans, and one of these was unusual in that it possessed certain dual parts in the sexual organs. Later, Sturtevant (1929) described, for the claret mutant stock of this species, four other gynandromorphs which had both ovarian and testicular tissue in the reproductive organs. This was followed by the work of Dobzhansky (1931) on gynandromorphs arising in this same stock. He reported on 171 specimens and found that 96 had ovaries, 46 had testes, and 29 had one ovary and one testis. As his Table 1 shows the correlation between the sex of the genitalia and that of the gonads is far from being complete. But the correlation between the sex of the external genitalia and that of the genital ducts is nearly complete, because as he shows these two sets of organs develop from the same imaginal disc. We did not attempt to make detailed dissections of the gynandromorphs we found, although a number of them were examined sufficiently to determine the nature of the gonads. In one specimen a mixed condition of the genital ducts was found (714-e; T.7 W). This mosaic had received from its mother a y sc m g X-chromo­some and from its father a treated normal X. The left half of the thorax with all three legs was yellow and male and the lefi wing was mixed, the yellow part being mmiature. The fly was thoroughly tested for fertility, but proved to be sterile, whereupon it was dissected by Mr. Walter Burdette. He found that the sex apparatus was of the female type throughout with small undeveloped ovaries. However, there was attached to the right oviduct a glandular organ which was similar in every way to a paragonium of the male. From this brief survey of the conditions which influence reproduction in gynandro­morphs, it is easy to see why so many of them fail to produce offspring. A large number of our specimens were mated, but many of them died within a short time, thus failing to give an adequate test. There were fifty-eight which lived long enough for us to determine whether or not they were fertile. Of this number, eighteen produced offspring and forty were completely sterile. It is probably true that many of those which died within a short time would have been found to be sterile had they lived long enough to allow for a thorough test. In the fertile group the external genitalia were without exception of the normal female type. As to the extent of the male area, four were one-fourth male, and the others belonged to various types of partials. Two of the one-quarter mosaics were alike, each having the right half of the head and thorax with legs and wing composed of male tissue. Specimen 152-c (T.7 H) gave twenty-four females and twenty-six males, but the number produced by the second fly was not recorded. The other two one-quarter mosaics were somewhat similar to each other, each having the right half of the abdomen male, but specimen 78-b had the entire tip male and the ventral surface was nearly all female (T.8 K). It laid two eggs which developed into pupae, but flies did not emerge. The other S,Pecimen (115-b; T.8 H) was highly fertile, producing thirty-seven females and forty-one males. There were two partial gynandromorphs which were identical in pattern in having one side of the thorax and wing composed of male tissue. The first of these (145-f; T.6 V) gave thirteen females and thirteen males, and the second (169-a; T.7 t) produced sixteen females and eighteen males. The only other specimen with the male tissue limited to the thoracic region was 139-b, which had one male-type wing. It was fairly fertile, yielding twenty-four females and thirteen males. Three partials were male for part of the head. Specimen 156-a (fig. 32; T. 7 j) gave twenty females and twenty-five males, and specimen 714-d (T.7 h) produced twenty-nine females and twenty-seven males. In the third specimen the left eye was small and white and the left wing was male-type. It gave four females and six males. Eight fertile partials had the male tissue restricted to the abdominal region, and in each case it was confined to the dorsal (tergites) surface. In four of these flies the male area was restricted to the male. coloration of one side of tergite 4 (fig. 46), and these gave the following number of offspring: Thirty-six females and twenty­seven males; twenty-eight females and eighteen males; twenty-two females and twenty-three males; eleven females and four males. In three specimens tergites 4 and 5 on one side showed male coloration (fig. 45) , and these produced the following number of offspring: Thirty-five females and thirty-four males; twenty-nine females and nineteen males; twenty-three females and nineteen males. Finally, there was one fly which showed yellow male coloration on the 5th tergite (fig. 48 shows example) and gave thirty-seven females and forty-one males. The character of the mosaics making up the sterile group is of equal interest. Twelve of these had male genitalia, and three had mixed genitalia. The three with mixed organs were all bilateral mosaics. Of those with complete external male genitalia, three were male for the entire abdomen, two were bilaterals, four were one-fourth male of the abdominal type, and three were partials. Two of the partials simply had male genitalia, while the third fly had, in addition to this, male coloration for the 5th and for half of the 4th segment. Three of the mosaics with female genitalia were three-fourths male. Of the six that were half-and-half, three were bilaterals (two seen in figs. 7 and 72), two were anterior-posterior with abdomen mostly or entirely female, and one was dorso­ventral. Nine flies were one-fourth male with the male area located on the head and thorax in four cases, on the abdomen in three cases, on the head and abdomen in one case and on all three regions in one case. There were seven sterile partials. Three of these had all or part of the head male and a small amount of male tissue on the thorax, two had one male-type wing and two had small male areas at or near the posterior end of the abdomen. A comparison of the fertile and sterile groups of gynandromorphs will serve to emphasize certain points discussed in the first part of this section. In the first place, only those flies were fertile which possessed female genitalia. Mosaics with mixed or completely male genitalia were always sterile, regardless of the size of the male area. On the basis of the elimination hypothesis, this result is to be expected since the composition of the male tissue must be XO, unless by chance a Y-chromosome is received from one of the parents. In the second place, the size of the male area in the fertile mosaics was limited. In no case did it exceed one-fourth of the surface area of the fly, and was less than this fraction in fourteen cases. In contrast to this condition, nine sterile mosaics had a male area that was equal to one-half of the surface, or even a greater portion. There were also nine specimens in which the male area was as much as a quarter of the surface. It is not easy to understand why some of the partials with female genitalia should have been sterile. In several cases flies with comparable male areas were sometimes fertile and sometimes sterile. Thus, by the way of illustration, of two flies each with one male-type wing, one was fertile and the other was sterile. Again, of five flies showing male for one side of tergite 4, four were fertile and one was sterile. It does not seem likely that the failure of such partials to reproduce could be due to a maladjustment of the sexual reactions alone. It seems more probable that it is due to a lack of correlation between the external features and such internal parts as the gonads. DISTRIBUTION OF MALE PARTS In this section we shall analyze the frequency and distribution of the male parts of the mosaics on the basis of the imaginal disc from which they were derived. Sturtevant used 36 groups of parts to represent the imaginal discs. We have used 38 such parts. The two additional ones represent the external genitalia. The disc derivitives used have already been described in the first section of the paper. The character of these imaginal discs for 213 suitable specimens is shown in Tables 2-8. The calculations were based on 211 of these cases. Specimens B and 247-1 were not included, because at the time the calculations were made their records were temporarily misplaced. In the last vertical column is given the number of male imaginal discs for each mosaic, as determined from an examination of the record card. Mixed discs were classed either as male or female, depending upon which type of tissue predominated. It should also be pointed out that although the wing, mesonotum and scutellum are listed separately, yet they all come from one imaginal disc, the second dorsal thoracic. 1. CLEAVAGE PATTERNS Both Sturtevant and Parks have concluded that the first cleavage patterns in Drosophila are indeterminate. In agreement with this theory, there are a number of cases with two separate areas of male or female tissue present (Table 12). In addition, there were 104 mosaics in which one or more imaginal disc derivatives were composed of both male and female tissue. There is little to add to Sturtevant's discussion of the problem, except to point out that in several of our cases there is an area of female tissue located in male tissue which had developed from one imaginal disc, and that was not adjacent to other female tissue. This must mean that the imaginal disc may be of mixed male and female cells, as neither late elimination nor somatic non-disjunction will explain such cases. TABLE 9. RELATION OF FOUR DISCS FOR EACH ABDOMINAL SEGMENT (DVVD, CLOCKWISE). 1 2 3 4 5 Total DVVD DVVD DVVD Segments ++++ +++­ ++-+ 114 I 3 2 117 3 2 I 1 12H 2 -·---­ 117 5 2 119 7 2 587 20 9 ~ DVVD DVVDDVVD DVVD ::[ ++-­ +-++ +-+­ +-­ 79 3 4 I 3 77 4 I 2 -H 1 1 ­ -4-·1 ·­ 75 0 3 4 70 0 5 2 66 1 7 ~JJ --­ 14 367 3 23 I 20 There are several ways by which to measure the amount and distribution of male and female tissue in these mosaics. Thu!> in Table 12 there are 33 cases which involve two male or female areas, separated by at least one intervening imaginal disc of the opposite sex. This gives a ratio of 11 cases of separted female tissue to 22 cases of separated male tissue and this is of interest in comparison with the ratio of female to male tissue in these gynandromorphs. Table 13 shows the com­parative amount of male and female tissue in terms of imaginal discs. The ratio is two female to one male disc in all cases. This ratio holds also for the cases where only the 20 discs forming the abdomen were considered. Here 200 out of 211 gynan­dromorphs did not have mixed discs on the abdomen. In this group there were 63 cases of gynandromorphs where the abdomen was all female as compared to 20 with the abdomen all male (Table 16). Therefore, part of the difference between the 11 cases of two areas of female tissue to 22 cases of two areas of male tissue lies in the comparative size of the female and male tissue (where isolated tissue of the opposite sex could be found), as well as in the fact that possibly not all gynandro­morphs result from a single elimination. However, these data can not be taken to prove that the daughter chromosome of an X which had the sister strand eliminated at the first division is more liable to be eliminated later than any other X-chromo­some in the complex. If elimination at two somatic divisions is responsible for the male areas in any of these cases, the descendants of the same homologue were in­volved in each elimination. The fact that there are small isolated patches of female tissue within male tissue produced by one imaginal disc proves that the discs may be of mixed origin. There are a few cases which might represent twin spots due to dossing over and non­disjunction with an inversion as suggested by Stern. These are rare as compared to other types. Furthermore, the fact that two separate areas of male tissue are always (so far as could be determined) of the same genotype suggests that they are not of independent origin. Sturtevant (1929) supposed, from analogy to other forms (Snodgrass 1924), that the dorsal larval ectoderm differentiated directly in place from the blastoderm. Parks (1936) subsequently showed that the dorsal ectoderm of the larva is formed from the upgrowth of cells from the ventral germ band. The imaginal discs, which are later to give rise to the ectoderm of the adult, probably arise from the group of cells thus derived from the ventral germ band. If this is true, the tergite and sternite pair on one side of the abdomen should show a very high positive correlation. This correlation may be measured, as there is one imaginal disc for each tergite and each sternite on the side of any abdominal segment. The four discs of the segment are probably more nearly regular in arrangement and equal in size than those which form the exoskeleton of the anterior half of the body. Moreover, their history and development are simpler. We shall consider the five major abdominal segments as representing the product of twenty imaginal discs, even though the first segment may be the result of the fusion of two larval segments. This apparently has made no difference in our data. Furthermore, there is no reason to believe that the initial distribution of cells that are to form the germ band is different in the anterior and posterior portions of the egg, except that the cells which are to form the gonads have a special history (Buettner, 1924; Parks, 1936). In addition the adult abdomen and genitalia represent imaginal discs from 8 of the 12 larval segments. The distribution of male and female tissue can be measured fairly accurately on the abdomen with the gene markers used, yellow and forked-5. In addition, in 200 out of 211 gynandromorphs, the imaginal discs which formed the abdomen seemed to be composed of unmixed male or female tissue (Table 16). Table 9 has been compiled from Tables 2-8 to show the relation between the four discs in any one abdominal segment with all segments with mixed discs omitted. There is no marked difference between any of the segments. Only those cases where two of the four imaginal discs of a segment differed from the other two can be used to determine the dorsal-ventral correlation. Of the 393 segments of this type, 367 had the dorsal and ventral discs on the one side of one sex and the dorsal and ventral disc of the other side of the opposite sex, 23 had the two do,rsal discs alike and three had the dorsal disc on the right side similar to the ventral disc on the left side. Thus, in 93.5 per cent. of the cases the dorsal and ventral discs on one side were alike as compared with 6.5 per cent. of the cases where they were different. These facts prove conclusively that the dorsal and ventral discs on one side are more closely related than either of the two discs across the median line. As the dorsal and ventral discs have this high positive correlation ( .935), of necessity there must be a much higher positive correlation in the anterior-posterior direction than across the medial line, otherwise checkerboard patterns would be frequent, but these seldom occur. TABLE 10. SPECIMENS WITH TWO OR FOUR ISOLATED DISCS (*MIXED CASES) The correlation anterior-posterior or across the median line can be measured in several ways. When two, three or four contiguous imaginal discs are involved, and these are not adjacent to similar tissue, their distribution will show the strength of correlation in an anterior-posterior versus a transverse-median direction. Table 10 shows the relations. Eight cases are positively correlated anterior-posteriorly, and five cases across the median line. Two of the latter ( 70-b, 338-a) represent small spots of female tissue isolated within male tissue on the mid-dorsal line. They must be considered pertinent, however, unless they should be regarded as a remnant of larval ectoderm which redifferentiated in situ to form the adult exoskeleton. Speci­men 9-c is peculiar in that tergite 4. on the left is similar to tergite 5 on the right. This case should not be classified as simply plus correlation across the mid-line. In Table 11 are listed those· cases where one-half or less of the abdomen is male or female, and this part is bounded anteriorly by tissue of the opposite sex. Here there are seventeen cases showing positive anterior-posterior correlation, nine with cross correlation, and four mixed. Each of these measurements of selected samples shows the strongest correlation anterior-posteriorly, which supports the general im­pression from all of the data. TABLE 11. MOSAICS WITH ISOLATED PART OF ONE-HALF OR LESS OF ABDOMEN INVOLVED (*MIXED CASES) Anterior-Posterior 8, 305-d, 172-a, 301, 115-b, 333-c, 301-b, 280-b, 226-c, 110-d, 712-d*, 33-a, 370-b*, 714-a, 37, 78-a, 25-a 17 Anterior-posterior and across median line 78-b, 264-b, 9-c, 71 4 Across median line 374-d, 303-n, 18-b, 110-b, 70­b*, 338-a*, 338-b, 300*, 717-d 9 There are only fifteen cases (Table 10) of two contiguous discs isolated in tissue of the opposite sex. The positive correlations in these cases are dorsal-ventral six, anterior-posterior 5, across 3 and mixed (diagonal) one. These numbers are too small to allow satisfactory deductions as to the true relations. However they do, together with the high dorsal-ventral correlation and the absence of checkerboard patterns, indicate a very high anterior-posterior correlation. This agrees with Sturtevant's conclusion. The picture that Sturtevant drew of the development of the claret gynandromorphs needs only be modified to include the upward growth of the ectoderm from the ventral germ band, as Parks has shown. 2 .. THE TIME OF FORMATION OF GYNANDROMORPHS Gynandromorphs could result from the loss of one X-chromosome from the pair of homologues, either at some specific division or at different cell divisions. The evidence is in favor of the hypothesis of Sturtevant to explain the claret mosaics that the gynandromorphs under consideration are the effect of loss of the whole or part of an X-chromosome in some one of the first six to eight divisions. Some very few of our own cases may have come from the elimination of a chromosome in early imaginal disc cells. Our data do not indicate that all gynandromorphs arise from the elimination of a chromosome at a specific division. However, there is considerable reason to believe that elimination at the first cleavage division is relatively much more frequent than at any other. The results of all of the different measurements of the relative amounts of male and female tissues can be explained by the hypothesis that there is a high rate of elimination at the first cleavage division. A.f:ter this division the rate is very much lower, and the frequency of different size areas of male tissue is proportional to the number of possible divisions where loss could occur. That is, if the gynandromorphs resulted from the loss of an X-chromosome at the first division, the amount of male and female tissue should be nearly equal, the number of times tissue of one sex was separated into two areas should be the same and the number of times the genitalia were male and female should be equal {mixed cases may be considered both ways). On the average there are two female to one male disc in the mosaics (Table 13). There were 22 cases with two separate male areas and 11 cases with two separate female areas (Table 12). TABLE 12. TWO SEPARATE AREAS OF ONE GENOTYPE Two male areas B, R, 23-a, 25-a, 37, 41-g, M, 87, 110-e, 188-a, 231-b, 274--e, 300, 302, 312-d, 314--e, 340-f, 370-b, 384-f, 386-e, 388-b, 717-d 22 Two female areas 33-a, 55, 60, 155-f, 165-c, 171-a, 295, 333-c, 338-a, 386-c, 717-e ll In Table 14 are summarized certain data from Tables 2-8. Here are listed the types of external genitalia found in the 211 mosaics. The numbers show that there were 108 female, 51 male, and 52 mixed. It is interesting to note that when male and female tissues are adjacent to the genitalia, these are mixed in 44 cases, female in 33 cases and male in 14 cases. Of the 44 mixed cases, 41 had male and female parts similar to the respective sides of the abdomen, and in only three cases were the genitalia mixed on both sides. In the cases where the abdomen was half.and-half and the parts arranged bilaterally, 29 had the same right·left arrangement of genitalia, three were mixed cases, 14 had female genitalia and nine had male genitalia. There is therefore a very high positive correlation between the external genitalia and the adjacent part of the abdomen, and this correlation was even higher when the proximal abdominal tissue was only male or only female. There is another way to determine the time of elimination of the X-chromosome. This may be done by measuring the number of imaginal discs that were male, out of the possible 38, and then to determine the ferquencies of the various combinations. These data were considered in two ways. First, by counting the total male parts for all of the mosaic (in this case the mixed imaginal discs were considered as the nearest whole number) as determined by an examination of the data on the record cards. Second, by considering the imaginal discs on the abdominal region, omitting the eleven mosaics with mixed discs. The data are listed in Tables 15 and 16. TABLE 13. RATIO OF THE NUMBER OF FEMALE TO MALE !MAGIN AL DISCS All cases Cases with no mixed discs Cases with no mixed discs, abdomen only Number Female Male Ratio female to male 2:1 2:1 2:1 2ll 107 200 5329 2675 2613 2689 1391 1387 Mr. Walter J. Burdette kindly made the calculations given below for tables 15 and 16. To present the data in Table 15 in a more suitable form for visual comparison the smoothed curves as plotted in fig. 1 were derived by averaging by fives twice. The data for the frequency-distribution of male discs on the abdomen were less irregular and therefore the curves shown in fig. 2 were plotted directly from the data in Table 16. It is necessary to test the data presented in Tables 15 and 16 to see if there is any significant difference between the control and experimental data. The test used is the t-test devised by Fisher (1936) for determining if significant differences exist in small samples. The t-test requires no initial assumption as to the nature of the population from which the sample is drawn. Tables 17 and 18 show that there is no significant difference between the means of control and experimental groups, either when the entire body or only the abdomen is considered. 11 IZ ', ',, 0 o 3 G 11 II 18 11 14 Zl JO J3 11 1; 18 11 14 11 10 Jl HAU DISCS ALL MOSAICS TREATED NORMAL MALES IZ11 ,~ --­ - 4 0 0 3 6 11 II 18 11 14 17 JO :13 /"!ALE OISCS INVERSIONS TREATED FEMALES 11 ~Ill ~ '<10 ::! ~ "8 '• ·'; ' \~1' ·;­' : \ ,,' ... ' ' ~ -..... -, ... -­ ~ 0 0 3 i; 11 1; 18 ll 14 n 30 n O 0 l 6 9 IZ II 18 11 14 11 30 l~ 1'1AL£ DISCS AfAlf l)IS(S BROKEN X·CHROMOSOME ALL TREATED MALES FIGURE 1 :-DISTRIBUTION OF MALE IMAGINAL DISCS TABLE 14, GENITALIA WITH VARIOUS DISTRIBUTION OF TISSUE ON ABDOMEN Adjacent abdomen female Adjacent abdomen mixed Adjacent abdomen male Female IMixed J Male Female IMixed IMale ~~ale IMix~IMale Genitalia 73 4 4 33 44 14 2 1 I 33 I TABLE 15. FREQUENCY-DISTRIBUTION OF MALE IMAGINAL DISCS ' Disc Control X-rayed X-rayed IInversion Mosaics . Broken IX-rayed males + males X all X Females 1 :: __ 1_"__1_ -----1--1___ J ~ --T-1==~ ----31-1--- I I ~,_____ -------}----I I I ___ 30__1_____1__1_ __,__1____ 1 --~_ I_____ 29 2 2 I 3 I 1 1 2s ____2__.__,__2____-,I---3 _ 1 1 ------1---­ 27 l 3 2 1 4 -------1-----1----1·----------1----1----+-----­ 26 3 1 2 3 1 ------1-----1---------1-----1-----1-----1------­ 25 3 1 1 3 1 -------r----1----r------r-----;-----1----i-------­u 1 l -------+-----1---.-----1------1-------;--,_________,_______ 23 l 2 l l -----1··---------------1-----+--·--·--------­ 22 1 2 1 6 3 -----------1-----------1-----1------1---+-----­ 21 4 2 1 1 6 ______"_____ ------+---+----r---------.--1------+-------­ 20 1 3 3 1 6 1 1 ------------1------1----,----1-------+-----­19 2 10 4 5 13 4 2 -----1---------{-----1------;-----l----<-----­ 18 l 2 2 3 I -----1----1--------~----+------'--------1-----­ 1 ---~-:--1----~---1---:-----:-----2-.·-1: --H--+--~---­ -----··­ 15 l 4 _J 2 2 5 I 14 1 10 1--3---+--2---­ 11-2J--­ _ __ 13 --4--~ 6 ·~ ,-I i--1?---'~--2·---+-----­ ____1_2 1 j__7___ _!____ ----__7_ --'-----+------­ _ _____n___~-­ 1 6 2 ~--~---r-6_____ 10 -----------1 9 : ~ ~-; I 2 8 1 7 2 l 5 4 7 2 2 7 5 2 1 10 1 6 10 8 _ ___5___ -------l--6--+--6----1------1---8---·,--~---_2____ -----:--1---1--1--1-;--~--~===:~==~:--1~1--~--+--:---­ -~----r------;----t----­ 2 4 4 4 13 5 1 1 3 3 2 9 5 ------1-----j------·----l-------r-----------t--------­Total 30 U3 86 W 211 32 48 10 10 10 O 0 I g 5 n ~ ~ W ffALl DISCS TREATED NORMAL MALES 18 19 10 1Q 10 0 0 I 1 l 4-5 6 1 a 9 10 II ll ll l~ 1; lo 17 IS 19 10 /'1Al.l DISCS BROKEN X-CHROMOSOME ALL TREATED MALES FIGURE 2 :-DISTRIBUTION OF UNMIXED MALE IMAGINAL DISCS ON ABDOMEN There is no significant difference in the ratios of partials ( 4-6 imaginal male discs) to other types of gynandromorphs in this data. The test used is to find the absolute difference in the probabilities of control and experimental partials and to divide that value by the standard deviation of the groups considered together. The quotient is greater than two when significance exists. There is some indication that in several cases the difference in distribution in experimental and control groups might pos­sibly become significant with additional data. However, in the data there is no significant difference in the distribution of the different types of gynandromorphs Gynandromorphs in Drosophila Melanogaster :1.5 classified as to male and female parts occurring under natural conditions and under the influence of experimental irradiation. In all cases the distribution of the frequencies of male and female parts may be interpreted as the result of two effects: First, a very much higher relative frequency of elimination at the first division, and, second, the effect of elimination at subsequent divisions in proportion to the number occurring, that is, there is equal (or nearly so) chance for elimination at any division. On this basis there is one chance for an X-chromosome to be eliminated at the first division, two chances at the second, four at the third, eight at the fourth, etc. The size of the male area will be one-half as large from the loss of a chromosome at the second division as it will be at the first, but there are two chances to one it will happen. These chance relations seem to hold after the first division. The fact that the first division is especially liable to eliminate the chromosome is particularly noticeable for the abdomen, as plotted in fig. 2 from data in Table 16. The fact that an even number of discs of tissue of each sex is more often the rule is doubtles!' due to the very high positive dorsal­ventral correlation between the imaginal discs. TABLE 16. MALE IMAGINAL DISCS ON ABDOMEN Discs Mosaics Controls Broken I x-rayed x-rayed x-rayed males all X +males females all -20--t----20--~---3--,---5--rl--1-o-~l:--2--1 -4 ---1____ 19 I I I --i-a,____ -1---~1--­ --~---::~~.'.~~=-~--_-:_~:-1==-~:=._-==---=---=~~::----;=_ --.~~~~-=1 14 5 1 2 1 1 1 1 I 2 --1-3--11--2--1-----j---1--;I-----2-1-1 I 2--12-1---·-8--1---2--1--3---..-·2---r·--1-,---5-­ 11 ---11-·-2--l--------l·-·:1------<----1 2 -------1--------1-------,----·--.------.-----------­ 10 55 11 6 ++1~4 __~1 9 2 I 1 1--­ :-----.1--1 a a 2 1 I --r-1__5____ 1 1 ---~-------~-=--=--=··,=--· ======:==+--------:~~:.:.=~~1----:----~=·==--i·--4-___ =====:==:=-~~:+------\-H---1-\:--~--­ --==1 --· ---:--.--1-:--1---~---1-------~--lli-~--:--­ ___o 6_3__~--35_0_-t-____~1r---1443 1.-1__ __l___ 7 __ 4_5___ Total 200 32 ~ 126 The total number of gynandromorphs, and each fraction measured separately, show a sharp peak for the type with abdomen half male and female, indicating that the first division produces more sex-mosaics than any other. The increasing fre­quency of gynandromorphs with male areas of decreasing size apparently follows the geometric progression expected on the basis of chance elimination at successive cell-divisions. It has been pointed out before that the results on amount of male tissue could be explained most simply by these last two factors. The analysis for the 20 abdominal segments is particularly favorable to this interpretation. The data for all 38 imaginal discs of the mosaics are much less definite. However it is open to the same interpretation. It is difficult to explain the differences between these two sets of data but it may be due to the differences in origin of the abdominal, head and thoracic regions which have already been discussed. TABLE 17. TOTAL IMAGINAL DISCS; SIGNIFICANCE TESTS Although the consideration of the whole of the mosaics does not give any measure of phenotypic males and females from the same process that forms mosaics, the data for the abdomen do. The number of all female abdomens ( 63) is much greater than the number of all male abdomens (20), as is to be expected if gynan­ders are produced at divisions other than the first. The cytological data of Huettner and Parks show that the cleavage pattern is indeterminate, the first cleavage is about equally frequently anterior-posterior as across the median line. The frequency of gynanders with the abdomen half male and half female is about twice that of gynanders with the abdomen all male (total mosaics, Table 16) which is in accord with these cytological findings. TABLE 18. ABDOMNAL IMAGINAL DISCS; SIGNIFICANCE TESTS Number of Individuals (n) Control 30 All 200 Mosaics All Treated Males 126 Treated Normal 80 Males Treated 44 Females Broken X­ 32 Chromosome Ari th­ metic Mean 8.50 6.93 6.83 6.58 6.93 8.63 CT 6.24 7.76 8.18 8.30 8.05 7.11 P> 0.20 0.30> P> 0.20 0.30> P> 0.20 0.40> P>0.30 l.OO> P> 0.90 SUMMARY AND CONCLUSIONS Our data support the conclusions of Sturtevant, Huettner and Parks that cleav­age is indeterminate in Drosophila. However, there exists a definite bilateral sym­metry of the two types of tissue in many of these mosaics. Gynandromorphs and mosaic females with a broken daughter X-chromowme show that if there is a primary female sex-factor in this chromosome, it must lie in the garnet-forked interval. This is in agreement with the other published data recently reviewed by the senior author (Patterson, 1938). A large majority of the mosaics have resulted from the elimination of the whole or a part of a daughter X-chromosome. A large number of these mosaics, derived both from control and treated material, were the result of elimination at the first cleavage division. As far as could be determined from the data, irradiation brings about an increase in the frequency of mosaics of all classes. This conclusion fol­lows from the failure to detect a significant difference between the means of the control and experimental data. Also the classes of mosaics with four or less male imaginal discs are as frequent among those coming from irradiated males as among those found in the controls. We know that irradiation of males has a differential effect on the elimination of the paternal and maternal X-chromosomes (Table 1). But this is not the case for that portion of these mosaics which had six or less male imaginal discs. Exclu­sive of those cases involving inverted chromosomes, the ratio here was ten to nine for the paternal and maternal X-chromosome, respectively. In spontaneous gynandromorphs, or those derived from irradiated females, the paternal and maternal X-chromosomes are eliminated with about the same fre­quency. This is obviously different from the results obtained from treated males. Nevertheless, the various classes of mosaics occur at about the same relative fre­quency, whether from the treated male, treated female, or untreated control series. Most of the gynandromorphs in our collection can be accounted for on the basis of simple elimination, as pointed out above, but the three complex cases and six mosaic males can not be thus explained. They require the assumption of either double fertilization, double elimination, or some other complicated cause. Their possible modes of origin have been discussed in a previous section. A careful anaiysis of these cases raises a number of interesting questions. All nine cases may be supposed to have arisen either at the first division, or from two reduced egg nuclei each fertilized by a spermatozoon. These cases fall into two groups and in the first the two types of tissue are about equally divided. This group includes gynandromorphs 13-y and 219-a, with 23 and 19 male imaginal discs respectively, and mosaic males 290, 4-x and 163-a, with 23, 26 and 17 discs respectively of one of the genotypes. In the other group, one of the genotypes is restricted to one-quarter or less of the external surface. This group includes gyandromorph B with slightly over 5 male imaginal discs, and mosaic males 157-a, 333-a and 205, with 6, 11 and 2 discs respectively of one of the two genotypes. It is this condition found in the second group of cases which raises the question of how much of the external surface of the mosaic will he represented by the descendants of each of the first two cleav­age nuclei. On the basis of these cases with such small numbers of imaginal discs of one genotype, one may expect to find almost any ratio between two tissues arising from the first two nuclei. Moreover, we have little or no evidence showing that cromoso­mal eliminations occur at different cleavage divisions in the same embryo. One can therefore explain the origin of most gynandromorphs with more than half the surface male, and a corresponding group with less than half the surface male, on the assumption that segregation of the two genotypes occurred at the first cleavage division. The frequency of distribution of male imaginal discs on the abdomen may be interpreted in the same way, especially in view of the fact that the group with all male abdomen must have arisen from an anterior-posterior cleavage pattern. If this method of interpretation is permissible, some two-thirds of our gynandro­morphs must have resulted from elimination at the first cleavage division. The variation from half-and-half in this class of mosaic must be due to the type of sampling of nuclei and to the mode of origin of the imaginal discs which ultimately form the exoskeleton of the adult fly. LITERATURE CITED Bonnier, G., 1928. A Drosophila mosaic, probably due to dispermic fertilization. Journ. Genet., 19 :257-260. Boveri, Th., 1888. Uber partielle Befruchtung. Sitz.-Ber. d. Ges. £. Morph u. Phys., Munich, IV. Chen, Tse-yin, 1929. On the development of imaginal buds in normal and mutant Drosophila melanogaster. Journ. Morph., 47 :135-200. Dobzhansky, T., 1931. Interaction between female and male parts in gynandromorphs of Droso­phila simulans. Roux' Arch., 123 :719-746. Doncaster, L., 1916. On some gynandromorphous specimens of Abraxas grossulariata. Pro. Camb. Phil. Soc., 18-79. Duncan, F. N., 1915. A note on the gonads of gynandromorphs of Drosophila amelophila. Amer. Nat., 49 :455-456. Fisher, R. A., 1936. Statistical methods for research workers. Sixth Edition. Oliver and Boyd, London. Goldschmidt, Richard, 1937. A gynandromorph in Drosophila produced by double fertilization. Cytologia, Fujii Jubilee Volume: 78-79. Huettner, A. F., 1923. The origin of the germ cells in Drosophila melanogaster. Journ. Morph. a. Phys., 37 :385-423. Huettner, A. F., 1924. Maturation and fertilization in Drosophila melanogaster. Journ. Morph. a. Phys., 39:249-265. Hyde, R. R., and H. M. Powell, 1916. Mosaics in Drosophila amelophila. Genetics, 1 :581-583. Jones, D. F., 1937. Somatic segregation and its relation to atypical growth. Genetics, 22:484-522. Morgan, L. V., 1929. Composits of Drosophila melanogaster. Carnegie Inst. Wash., Publication 399:223-296. Morgan, T. H., 1905. An alternative interpretation of gynandromorphous insects. Science n.s., 21. Morgan, T. H., 1907. The cause of gynandromorphism in insects. Amer. Nat., 41 :715-718 . .Morgan, T. H., 1914. Mosaics and gynandromorphs in Drosophila. Proc. Soc. Exp. Biol. and Med., 11:171-172. Morgan, T. H., 1915. The infertility of rudimentary winged females of Drosophila amelophilia. Amer. Nat., 49 :240-250. Morgan, T. H., and C. B. Bridges, 1919. The origin of gynandromorphs. Carnegie Inst. Wash., Publication 278:1-122. Parks, H. B., 1936. Cleavage patterns in Drosophila and mosaic formation. Ann. Entomol. Soc. Amer., 29:350--392. Patterson, J. T., 1931. The production of gynandromorphs in Drosophila melanogaster by x-rays. Journ. Exp. Zoo!., 60:173-211. Patterson, J. T., 1938. Aberrant forms in Drosophila and sex differentiation. Amer. Nat., 72:193-206. Payne, F., and M. Denny, 1922. A gynandromorph in Drosophila melanogaster. Amer. Nat., 46 :383-384. Snodgrass, R. E., 1924. Anatomy and metamorphosis of the apple magot, Rhagol,etis pomonella. Walsh. Journ. Agriculture Research, 28:1-36. Spencer, W. P., 1927. A gynandromorph in Drosophila funebris. Amer. Nat., 61 :89-91. Spelt, G., 1932. Gibt es eine partielle sexuelle Isolation unter den Mutationen und der Grund­ formen von D. melanogaster? Ztschr. f. lndukt. Abstam-u-Vererbungs., 60:63--83. Stern, C., 1936. Somatic crossing over and segregation in Drosophila melanogaster. Genetics, 21 :625-730. Strasburger, E. H., 1935. Drosophila melanogaster Meig. Eine Einfuhrung in den Bau und die Etwicklung. Verlag v. Julius Springer, Berlin, pp. 1-60. Sturtevant, A. H., 1915. Experiments on sex recognition and the problem of sexual selection in Drosophila. Journ. Animal Behav., 5:351-356. Sturtevant, A. H., 1921. Genetic studies on Drosophila simulans. II. Sex.Jinked group of genes. Genetics, 6 :43-64. Sturtevant, A. H., 1929. The claret mutant type of Drosophila simulans; a study of chromosome elimination and cell·lineage. Ztschr. f. wissensch. Zoo!., 135 :323-356. Timofeeff-Ressovsky, N. A., 1928. Gynandromorphen und Genitalien-Abnormitaten bei Drosophila /unebris. Arch. Entw.-mechan., 113 :254-266. Ward, Helen, 1936. Cytologic studies on abnormal development of eggs of the claret mutant type of Drosophila simulans. Genetics, 21 :264-281. Wheeler, W. M., 1937. Mosaics and other anomalies among ants. Harvard Univ. Press, Cam· bridge, pp. 1-95. PLATE I Fig. 1 (304-b) . Cross: y sc mg /5 females to x-rayed wild-type males. Right half throughout was gray and female; left half showed all five recessive characters and was male. Mixed geni­talia. Loss of treated X. Fig. 2 ( 386--d). Cross: r sc g pl f females to x-rayed wild-type males. Left half was gray and female, except tergite 4 which was y f; right half showed all recessive characters and was male, except tergite 4 which was gray. Mixed genitalia. Loss of treated X. Fig. 3 (175-a). Cross: Theta females to y w cv sn m males, untreated. A yellow fly, with head and left half (except scutr.llum) yellow and female; right half of thorax, all of scutellum, and right half of abdomen were yellow signed and male; right wing cv and m. Mixed genitalia. Fly received from its mother the y sc X, which was absent from male parts. Fig. 4 ( 110--c). Cross: yell"w white females to eosin singed males, untreated. Head, right half of thorax and right half af abdominal segments 1-3 were all gray and female; left half of thorax and ·abdomen plus all of segments 4 and 5 were yellow and male. Male genitalia. Loss of paternal X. Fig. 5 (274-f). Cross: y sc m g /5 females to x-rayed Theta males. Left half of thorax and of abdominal segments 1-3 aud all of 4 and 5 were gray and female; head, right half of thorax and of abdominal segments 1-3 showed all five recessive characters and were male. Female genitalia. J,oss of treated X. Fig. 6 ( 113-a I. Cross: Yellow white females to eosin singed males, untreated. Left half of thorax and abdomen, and ,;ternopleurals, 2nd and 3rd legs on right were gray non-singed and female; all oth"'l" parts were singed and male. Mixed genitalia. Loss of maternal X. Fig. 7 (112---eJ. Cross: Yellow white females to eosin singed males, untreated. Left half of thorax and nght half of al.idomen were gray and female; head with white eyes, right half of thorax and left half of abdomen were yellow and male, except the ventral side of segment 5 which was gray. female genitalia. Loss of paternal X. Fig. 8 (269-d). Cross: y ~c w females to x-rayed wild-type males. Head with white eyes and thorax with all appendagts yellow scute and male; entire abdomen was gray and female, with female genitalia. Loss of treated X. Fig. 9 (344-b). Cross: y w lz m females to x-rayed wild-type males. Head and thorax with all appendagef' were gray and female; entire abdomen was yellow and male, with male genitalia. Loss of treated X. t 6 Y SC m g y w we sn 4 ysc 'IV y w li: l!\ 7 8 9 PLATE II Fig. 10 (110-h). Cross: yellow white females to eosin single males, untreated. Right half of thorax with appendages was gray and female; all other parts were yellow and male, with white eyes and male genitalia. Loss of paternal X. Fig. 11 (314-i). Cross: y w cv m females to x-rayed wild-type males. Entire abdomen, upper left half of thorax, left orbital bristles, and left antenna were yellow and male; all other parts were gray and female, including left sternopleurals. Male genitalia. Loss of treated X. Fig. 12 (332-g). Cross: y sc m g JS females to x-rayed wild-type males. Head, except garnet spot on left eye, and right half of thorax with appendages were gray and female; all other parts were male and showed the five recessive characters. Male genitalia. Loss of treated X. Fig. 13 (10-m). Cross: x-rayed scute-8 apricot females to y w cv m fS males. Part of the head and left half of the thorax except 3rd leg, were gray and female; all other parts were male and showed the five recessive characters. Male genitalia. Loss of treated X. Fig. 14 (196-h). Cross: x-rayed wild-type females toy w m J B males. Left half of head, both antennae, left half of thorax except fore-leg were all gray and female; the right fore-leg and eye was mixed, hut all other parts of the fly were male and showed the five recessive characters. Male genitalia. Loss of treated X. Fig. 15 (314-k). Cross: y w cm m females to x-rayed wild-type males. Head mostly gray female, except for white spot on left eye and yellow left antenna ; the ~eft legs and following bristles on left side were gray and female, notopleurals; supraalars and sternopleurals, left wing mixed and showed gray and y cv m; all other parts of the fly were yellow and male. Male genitalia. Loss of treated X. Fig. 16 (60). Cross: yellow white females to x-rayed wild-type males. Head and left fore· leg were gray and female; all other parts were yellow and male, except for mixed genitalia. Loss of treated X. Fig. 17 (33-a). Cross: scute-8 apricot females to y w cv m JS males, untreated. The follow· ing parts were gray and female: head, all legs, left half of abdominal segments 4 and 5, outer margin of left wing, right humerals, presutural, notopleurals, and stemopleurals on both sides; all other parts of fly were male and showed the recessives y cv m JS. Male genitalia. Loss of maternal X. Fig. 18 (326-h). Cross: y w cv sn m females to x-rayed wild-type males. Right half of abdomen and all of left half except segments 1 and 2 were gray and female, with female genitalia; all other parts of the fly were male and showed the five recessive characters. Loss of treated X. y W CV m ysc y 1\1 § rs t I ' 10 tl 1'2. m f B Y w cv m y " ll 14 15 r ':' c,v s,n 111 r '! c.v l!l f:;, 11 18 PLATE III Fig. 19 092-d). Cross: x·rayed Theta females to y w cv m f B males. Left half of head with heterozygous bar eye, left half of the thorax with appendages and the entire abdomen were gray and female; right half of the head with whitP. male-type bar eye, and right half of the thorax with appendages were male and showed all five recessive w and back of right eye were gray; all other parts were gray and female. Loss of treated X. Fig. 31 (40-e). Cross: x-rayed scute-8 apricot females to y w cm m IS males. Small white left eye and left half of head yellow forked-5; all other parts gray and female. Loss of treated X. Fig. 32 (156-a). Cross: x-rayed Theta females to y w sn lz males. A yellow fly with three­fourths of right eye white lozenge and with bristles adjacent to white area singed. Loss of treated yellow scute X of Theta mother. Fig. 33 (274-g). Cross: y sc mg IS females to x-rayed Theta males. Right half of head with small garnet eye and distal part of right fore-leg were yellow forked and male; all other parts female. Loss of treated X. Fig. 34 (4-dd). Cross: scute-8 apricot females to y w cv m fS males, untreated. Right half of head and apricot eye smaller and right fore-leg with sex-comb; all other parts female. Loss of paternal X. Fig. 35 (329-a). Cross : ClB/ y w lz m females to x-rayed wild-type males. Left eye w lz, left half of head smaller and yellow, left sternopfourals yellow; all other parts gray and female. Loss of treated X. Fig. 36 (315-d) . Cross: y w cv sn m females to x-rayed wild-type males. Upper side of left half of thorax was y sn, left wing mixed showing y m sn; all other parts gray and female. Loss of treated X. Fig. 37 (40-d). Cross: x-rayed scute-8 apricot females toy w cv m fS males. Right supraalars y JS, right wing y m cv JS; all other parts gray and female. Loss of treated X. Fig. 38 (41-g). Cross: y w cv m JS females to x-rayed scute-8 appricot males. Left sterno­pleurals, postalars, scutellars, left half of abdominal segment 5 and proximal part of 2nd leg were all y JS, outer margin of left wing y cv m; all other parts gray and female. Mixed geni­talia. Loss of treated X. Fig. 39 (279-a). Cross: y sc m g JS females to x-rayed Theta males. The following parts were yellow forked-5: both fore-legs with sex-combs, proximal part of left 2nd leg, presutural, 2nd notopleural, anterior supraalar, outer postalar; all other parts were gray and female. Loss of treated X. Fig. 40 (714-f). Cross: y sc m g females to x-rayed wild-type males. Left 3rd leg and left half of abdominal segments 1-3 were y sc and male; all other parts were gray and female. Loss of treated X. '2.8 19 y W CV DI f~ y w y w lz m y W CV Ill f5 y w sn lz y w cv m r5 :t w cv sn 11 37 rs I PLATE V Fig. 41 (171-a). Cross: Theta females to r w sn m B males, untreated. A yellow bar fly with last three abdominal segments r sn and male; all other parts female and non-singed. Geni­talia mixed. Loss of maternal r SC X. Fig. 42 (374-d). Cross: r sc mg sd f females to x-rayed J B males. Left half of abdominal segment 3 and all of 4 and 5 were r sc J. with male genitalia; all other parts gray female. Loss of treated X. Fig. 43 (303--n). Cross: r sc m g JS females to x-rayed wild-type males. The 4th and 5th abdominal segments were y sc JS and male with male genitalia; all other parts gray female. Loss of treated X. Fig. 44 (18-b) . Cross: scute-8 apricot females to r w cv m JS males, untreated. Right half of abdominal segment 4 and all of 5 were r JS and male, with male genitalia; all other parts gray female. Loss of maternal X. Fig. 45 (280-b). Cross: x-rayed bar females to r sc m g /S males. Right half of abdominal segments 4 and 5 were r sc JS and male; all other parts 'Were gray and female with female genitalia. Loss of treated X. Fig. 46 (~n). Cross: scute-8 apricot females to r w cv m JS males, untreated. Right half of abdominal segment 4 had black male coloration; all other parts female. Loss of paternal X. Fig. 47 (9-c). Cross: scute-8 apricot females to r w cv m JS males, untreated. Left 4th and right 5th tergites were r JS and male; all other parts were gray · and female. Loss of maternal X. Fig. 48 (ll-h2). Cross: r w cv m JS females to x-rayed scute-8 apricot males. Fifth abdom­inal segment had black male coloration with male genitalia; all other parts gray female. Loss of untreated X. Fig. 49 (4-b). Cross: x-rayed scute-8 apricot females to r w cv m JS males. Left half of 5th abdominal segment was black and male with mixed genitalia; all other parts gray female. Loss untreated X. m B y w '" 41 42 4.3 y VT CV Ill f5 ysc g r5 44 45 PLATE VI Fig. 50 (276-a). Cross: r sc ct v m g females to x-rayed Theta males. A highly mixed gynan· dromorph with garnet eyes (vermilion not revealed). For details of the distribution of the male and female parts, see Table 5, D. Loss of treated X. Fig. 51 (213-a). Cross: r sc m car females to x-rayed Theta males. Following parts were yellow scute and male: head with carnation eyes, all six legs, left humerus, part of right half of thorax; right wing yellow miniature, left wing mixed; all other parts gray and female. Loss of treated X. Fig. 52 (294). Cross: yellow scute-13 females to x-rayed bar males. Left half throughout was yellow and male, with non-bar eye; right side had male tissue on head, humerus, scutellum and wing which was mixed. Mixed genitalia. Loss of treated X. Fig. 53 (302---i:). Cross: r sc m g /5 females to x-rayed wild-type males. Right half (}f head with garnet eye, distat part of right fore-leg, humerals, presutural, 2nd notopleural (first absent) were r /5 and male; outer margin of right wing y m; all other parts gra.y and female. Loss of treated X. Fig. 54 (274-e). Cross: r sc mg /5 females to x-rayed Theta males. Right eye mostly garnet; on right side, part of head, all three legs, thorax except humerus, and 1st abdominal segment were r sc /5 and male, as was also the left half of abdominal segments 3-5; right wing mixed. Geni· talia male; all other parts gray female. Loss of treated X. Fig. 55 (64). Cross: yellow white females to x-rayed wild-type males. Left eye has white spot; left half of thorax, except humerus, and appendages and right half of abdominal seg· men ts 3-5 were yellow and male; all other parts were gray female. Genitalia male. Loss of treated X. Fig. 56 (312-d). Cross: r sc m g /5 females to x-rayed wild-type males. Left eye mostly garnet and right eye with garnet spot; upper surface of head and entire abdomen were· r sc /5 and male, with male genitalia; all other parts were gray female. Loss of treated X. Fig. 57 (714-a). Cross: y sc m g females to x-rayed wild-type males. Left eye garnet with garnet spot on right eye; following parts were yellow scute and male: left half of head, right verticals; three left legs, all five tergites on left side of abdomen; left wing mixed; all other parts were gray female, including genitalia. Loss of treated X. Fig. 58 (25-a). Cross: x-rayed scute-8 apricot females to y w cv m fS males. Half of right eye white; right orbitals and a&S()Ciated hairs were y /S; right tergites 4 and 5 were y /5 and male; all other parts were gray female, including genitalia. Loss of treated X. Ysc y sc1.3 so SI 53 f5 Y W CV Jll 5G PLATE VII Fig. 59 (386-c). Cross: y sc g pl /5 females to x-rayed wild· type males. Left half of head, left half of thorax with its appendages, and entire abdomen except genitalia showed y sc g pl and were male; all other parts were gray and female. Genitalia mixed. Loss of treated X to left of locus of forked. Fig. 60 (305-b ). Cross: y sc mg /5 females to x-rayed wild-type males. A gray fly, hut the left half of thorax with its appendages and left half of abdomen and segment 5 showed /5 and were male, with male genitalia; left wing male-type but not m; all other parts were non-forked and female. Loss of part of treated X. Fig. 61 (29-t). Cross: y w cv m /5 females to x-rayed scute-8 apricot males. A gray fly, but the right half of the thorax with 2nd and 3rd legs and the entire abdomen were f5 and male; outer three-fourths of right wing was m and non-crossveinless; all other parts were non-forked and female. Loss of part of treated X. Fig. 62 (387-i). Cross: y sc g pl f females to x-rayed wild-type males. A gray fly. Right eye garnet, with garnet spot on left eye; part of head, right half of thorax with appendages and entire abdomen were forked and male; right wing pleated; all other parts non-forked and female. Loss of part of treated X. Fig. 63 (253--j). Cross : y sc m f /u car females to x-rayed Theta males. A gray fly, with following parts forked and male: left half of head, left humerus, posterior dorsocentral, posta­lar and scutellars, left 1st and 2nd tergites, and entire right half of abdomen. Left eye car and left wing fused but not miniature; all other parts were forkl'd and female Genitalia mixed. Loss of /-car section of treated X. Fig. 64 (303-i) . Cross: y sc m g /5 females to x-rayed wild-type males. A gray fly with following parts male: part of right eye garnet, right half of thorax exclusive of legs, right half of abdomen and segments 4 and 5, and genitalia; right wing m; all other parts were female. Loss of middle section of treated X. Fig. 65 (308-c). Cross: y sc m g /5 females to x-rayed wild-type males. A gray fly, with garnet eyes, head and thorax had small bristles, fore-legs with sex-combs, both wings minia­ture, left tergite 5 had male coloration; scute revealed but all bristles non-forked; all other parts female, including genitalia. Loss of sc-g section of treated X. Fig. 66 (302) . Cross: m fw g f females to x-rayed Theta males. Right eye g fw; right side of head and part of left, part of right half of thorax and all three legs, right half of segment 5 were forked and male; right wing male-type but not m; all other parts non-forked and female. Genitalia mixed. Loss of fw-f section of treated X. Fig. 67 (306-a) . Cross : y sc m g /5 females to x-rayed wild-type males. A gray fly, with left half of head showing sc and /5 and small garnet eye; all other parts non-forked and female. Loss of sc-f section of treated X. ysc g pl t ysc Jll f r5 I I I ' + + 59 + GO t GI + 1------r­ t t 1-?----1---1109 wa -X ysc g pl f ysc m ! f~ car y so m f fS ,, A I I I IGZ .:,Y_B.._!l_~'t'4I ____ -~ 0'+ ... + + 63 ~------1-i.--­ i..1...j-------·---1--­ y sc 11 a: f,5 m !wa: f' • • :!'. . I 1 '( I 'If IC Jll Jr f5 65 11 GG 61 ­ I "' ~---_---i-1---Y._s,__c_ __,_+_1__ _ _ --~ ..±.j ________ • -1-­ PLATE VIII Fig. 68 (Z) . Cross: x-rayed y w f B females to eosin mm1ature males. A gray fly, with right eye narrow bar; right half of head, thorax and appendages and right half of abdomen and segments 4 and 5 above all showing smaller ·bristles and male; right wing and inner margin of left showed miniature; all other parts were female, including genitalia. Loss of part of treated X. Fig. 69 (190-c). Cross: x-rayed Theta females to y w cv m f B males. A gray fly, with left postalar and scutellar bristles and bristles on left 2nd and 3rd legs all smaller; left wing cv m; left half of abdomen and segment 5 smaller and male with male genitalia; all other parts were female. Loss of y-m section of treated Theta X. Fig. 70 (B) . Cross: Yell ow females with attached X-chromsoomes to Theta male, untreated. A yellow female with the following gray male parts: left scutellars, inner half of right wing, right half of abdominal segments 4 and 5 and genitalia. Male parts received Theta X. Fig. 71 (13-y). Cross: x-rayed scute-8 apricot females to y w .;v m. /5 males. Complex gynandromorph; see text for description. Fig. 72 (219-a). Cross: x-rayed Theta females to y .sc m. f car males. Complex gynandro· morph; see text for description. Fig. 73 (4-x) . Cross: y w cv m /5 females to scute-8 apricot males, untreated. Mosaic male; see text for description. Fig. 74 (163-a). Cross: Theta females to y w sn m B males, untreated. Mosaic male; see text for description. Fig. 75 (333-a). Cross: y sc m g /5 females to x-rayed Theta males. Mosaic male; see text for description. Fig. 76 (304-c). y sc m g /5 females to x-rayed wild-type males. Mosaic female; see text for description. we m yw CV m ~ B GB ----------i--!-­ G9 -------~ a 6 \{ f er f c It &C _r_s_f___, Ysc 7_1_ __.__._ Y, r c,v f t5 71 l y w SI\ JI! B p:c.,. ·¥ fs • t Sf Ill rs I I & I I I I Ir s,o Ysc Jll g r5 f I I I I I 74 1S I ------i-1 7f, ~ ::::J