i i ru o m o If A.T1TRAL SmWN MPT. CONTRIBUTIONS TO THE GENETICS DROSOPHILA MELANOGASTER. I. THE ORIGIN OF GYNANDROMORPHS. BY T. H. MORGAN and C. B. BRIDGES. II. THE SECOND CHROMOSOME GROUP OF MUTANT CHARACTERS. BY C. B. BRIDGES and T. H. MORGAN. III. INHERITED LINKAGE VARIATIONS IN THE SECOND CHROMOSOME. BY A. H. STURTEVANT. IV. A DEMONSTRATION OF GENES MODIFYING THE CHARACTER "NOTCH." BY T. H. MORGAN. . PUBLISHED BY THE CARNEGIE INSTITUTION OF WASHINGTON WASHINGTON, 1919 CARNEGIE INSTITUTION OF WASHINGTON PUBLICATION No. 278 PRESS OF GIBSON BROS, INC. WASHINGTON, D. C. CONTENTS. PAGE. I. The Origin of Gynandromorphs. By T. H. MORGAN and C. R. BRIDGES ......... 1 Frequency of Occurrence of Gynandromorphs ............................. 10 Relative Frequency of Elimination of the Maternal and Paternal Sex Chromosomes .................................................. 11 Distribution of Segmentation Nuclei as deduced from Distribution of the Characters of Gynandromorphs ................................. 12 Starting as a Male vs. Starting as a Female ................................ 12 Cytological Evidence of Chromosomal Elimination .......................... 13 Earlier Hypotheses to explain Gynandromorphs ............................ 17 The Origin of the Germ-cells in Flies .................................... 22 Courtship of Gynandromorphs ........................................... 22 Phototropism in Mosaics with one White and one Red Eye ................. 23 Sex-limited Mosaics .................................................... 24 Somatic Mosaics ....................................................... 26 Somatic Mutation ...................................................... 27 Mosaics in Plants ...................................................... 32 Classification and description of Gynandromorphs of Drosophila ............. 33 Gynandromorphs approximately bilateral ................................ 35 Gynandromorphs mainly female ........................................ 41 Gynandromorphs mainly male .......................................... 48 Gynandromorphs roughly "fore-and-aft" ................................. 51 Gynandromorphs produced by XXY females ............................. 53 Gynandromorphs of complex type ....................................... 55 Special Cases ...................................................... 57 Gynandromorphs with incomplete data .................................. 70 Drosophila Gynandromophs previously published ......................... 72 Gynandromorphs and Mosaics in Bees .................................... 74 Gynandromorphs in Lepidoptera ......................................... 81 Other Insects ........................................ 94 Spiders ............................................. 94 Crustacea ........................................... 95 Molluscs ............................................ 97 Echinoderms ........................................ 98 Vertebrates ......................................... 98 Fishes .............................................. 98 Amphibia ........................................... 99 Reptiles ............................................ 101 Birds ............................................... 101 Mammals — Man ..................................... 106 Is Cancer a Somatic Mosaic? ............................................ 108 Is the Freemartin a Gynandromorph? .................................... 109 Summary ............................................................. Ill Literature Cited ...................................................... 116 II. The Second Chromosome Group of Mutant Characters. By C. B. BRIDGES and T. H. MORGAN ................................................ 123 Introduction ........................................................... 125 Chronological list of the II Chromosome Mutations .................... 126 Map of Chromosome II ............................................. 127 Speck ................................................................. 128 Olive ................................................................. 135 Truncate .............................................................. 136 Truncate Lethal ................................................... 138 Snub ............................................................. 140 Truncate Intensification by Cut ..................................... 143 in IV CONTENTS. PAGE. II. The Second Chromosome Group of Mutant Characters — continued. Black 144 Balloon 148 Vestigial 150 Blistered 155 The Semi-Dominance of Blistered — Free- Vein 158 Jaunty 160 A Mutating Period for Jaunty 161 Curved 164 Purple 169 The Differentiation of Purple by Vermilion — Disproportional Modification . 170 No Crossing Over in the Male 174 The Inviability of Vestigial — Prematuration, Repugnance, Lethals 177 The Purple "Epidemic" — "Mutating Periods" 178 Balanced Inviability, Complementary Crosses 181 The Variation of Crossing Over with Age 183 Coincidence 188 The Relation between Coincidence and Map Distance 188 Special Problems Involving Purple — Age- Variations, Coincidence, Tempera- ture-Variations, Cross-Over Mutations, Progeny Test for Crossing-Over 193 Strap 200 Arc 202 Gap 208 Antlered 211 Dachs 216 Streak 222 Dominance and Lethal Effect of Streak, Parallel to Yellow Mouse 223 Comma 228 Morula 230 Female Sterility of Morula 230 Apterous 236 Cu i and Cn r 239 Cream 11 239 Trefoil 244 Cream b 245 Pinkish 247 The Double-Mating Method 248 Plexus 251 Limited 254 Confluent 255 Confluent Virilis 257 Fringed 257 Star 259 Lethal Nature of the Homozygous Star 260 Crossing Over in the Male 263 Nick 273 Vestigial-Nick Compound 275 Dachs-Lethal 277 Dachs-Deficiency? 278 Balanced Lethals 279 Squat 283 Lethal Ha 286 Telescope 291 Second Chromosome Modifiers for Dichsete Bristle Number 293 Dachsold 294 The Construction of the Map of the Second Chromosome 297 Summary of Available Data on Crossing Over in the Second Chromosome 298 Constructional Map 302 Working and Valuation Map 303 Bibliography t. 304 CONTENTS. V PAGE. III. Inherited Linkage Variations in the Second Chromosome. By A. H. STURTEVANT. 305 Introduction 307 "Nova Scotia" Chromosome 307 Tests of Cross-Overs 312 Right-hand end of Nova Scotia Chromosome (Cn r) 313 Left-hand end of Nova Scotia Chromosome (Cn i) 316 Homozygous Cn r 319 With Cn i 319 Without Cn 321 No Tests of Homozygous Cn i 322 Tests Showing No Crossing-Over in Males 322 Constitution of the Nova Scotia Stock 322 Another Second-Chromosome Linkage Variation 324 Comparison with Results obtained from Cm 325 Discussion 327 Summary 330 Appendix 331 Literature Cited 341 IV. A Demonstration of Genes Modifying the Character " Notch." By T. H. MORGAN 343 Variation of Notch 346 The Problem 347 Condition of Stock before Selection 348 Selection of Females having Notch in one Wing only 349 Selection of Somatically Normal Winged Females that are Genetically Notched Females 350 Duplicate Selection Experiment 355 Localization of the Gene for Notch 358 The Indentification of the Modifying Genes 361 Short Notch 364 First Test 366 Second Test 367 Third Test 368 Fourth Test (for fourth-chromosome modifiers) 369 Recombination of Bent and Short Notch 370 Crosses between Short Notch and Other Stocks 371 Short Notch by Star Dichaete 372 Short Notch by Eosin Ruby Forked 372 Classification of Types of Notch 376 Aberrant Notch Wings 379 Deformed Eyes 379 Little Eyes 381 High Sex-ratios Caused by Lethals 381 Other Characters that Look Something Like Notch 382 Gynandromorph; Notch Eosin Ruby 384 Summary 387 I. THE ORIGIN OF GYNANDROMORPHS. BY T. H. MORGAN AND C. B. BRIDGES. With four plates and seventy text-figures. I. THE ORIGIN OF GYNANDROMORPHS. BY T. H. MORGAN AND C. B. BRIDGES. INTRODUCTION AND GENERAL DISCUSSION. The sharp distinction into two kinds of individuals, males and females, characteristic of so many animals, is occasionally done away with when an individual appears that bears the structures peculiar to the male in some parts and to the female in other parts of the body. Such an individual may show not only the secondary sexual differences (either sex-limited or sex-linked) of male and female, but gonads and genitalia of both kinds as well. We speak of these as gynandromorphs. The union of the two sexes in a single individual shows how far the characteristics normally associated with one sex alone are compatible with the presence in another part of the same body of somatic structures and reproductive organs of the opposite sex. In a word, how far each is independent of sex hormones. But the chief importance of these rare combinations lies in the opportunity they furnish for analysis of the changes in the hereditary mechanism of sex determination that makes such combinations possible. This evidence is chiefly derived from gynandromorphs that are also hybrids. Such individuals may combine not only male and female sex differences, but the character- istic racial differences as well. Whether gynandromorphs arise more frequently in hybrids or whether it is only that their detection is easier under such circumstances will be discussed later. The occurrence of hybrid gynandromorphs offers at any rate a unique opportunity to discover the method of origin of such kinds of individuals. In hybrid gynandromorphs the differences that are shown may be due to genes carried by the sex chromosomes. Most of the gynandro- morphs of Drosophila belong to this category. In many cases, how- ever, especially in other insects, it is not known whether the differences shown by the hybrid gynandromorph are due to the sex chromosomes or to other chromosomes, either because the ancestry of the gynandro- morph is unknown or because the method of inheritance of the gene is unknown. There are, however, some very rare cases in Drosophila in which the characters involved are probably autosomal and the individual, while showing its dual parentage hi different parts of the body, is not a sex-mosaic. It may be convenient to designate such types as mosaics, while the sex-mosaics may be designated by the more special term gynandromorphs. In our work on Drosophila melanogaster (ampelophila) a large number of gynandromorphs and mosaics have appeared, and since the first THE ORIGIN OF GYNANDROMORPHS. description of a few of them was published we have continued to keep records of their occurrence. Others, too, working with our mutant types have found them, and a few have been described by Dexter, Duncan, and Hyde. We soon realized that they occurred with suffi- cient frequency to make it possible to devise experiments of a sort to furnish the long-sought criterion as to the most common method of their occurrence. It is this evidence on which we wish now to lay chief emphasis. The ordinary gynandromorph is an animal that is male on one side of the body and female on the other. The reproductive organs, gonads, and ducts may or, hi bees at least, may not show a corre- sponding difference. A typical case of a gynandromorph that is bilateral, at least superficially, is represented in plate 1, figure 1. For a long time it has been recognized that bilateral gynandromorphism is only one kind of abnormal distribution of the sex characters; even in the classical case of the Eugster bees (see p. 74) other distributions of the characters were recorded. In the fly represented in plate 3, fig- ure 2, the upper part of the abdomen is female, but the lower side of the abdomen, notably the external genitalia, are male. In the individual represented in plate 3, figure 5, the left anterior side of the head is male, the right fe- male, while the left posterior parts of the body are female, the right male. Other cases will be described later in which even more irregular and complex distributions of male and female parts exist. B Before discussing TEXT-FIGURE 1. these and other cases in detail, it may be well to give three of the most recent interpreta- tions of gynandromorphism resting on a chromosomal basis and the criteria by which the validity of each has been tested. In 1888 Boveri suggested that on rare occasions a spermatozoon, on entering the egg, might be delayed in its penetration to the vicinity of the egg-nucleus, and the latter might meanwhile have begun to divide, so that the sperm-nucleus came to unite with only one of its halves. In consequence, two kinds of nuclei would be produced in the embryo (text fig. 1 A). The nuclei that come from the sperm plus the half egg-nucleus would be diploid. If, as in the bee, one nucleus stands for the male and two for the female, it follows in such cases that all those parts of the body whose nuclei are derived from the THE ORIGIN OF GYNANDROMORPHS. 5 single (haploid) nucleus would be male, all those from the double (diploid) would be female. Moreover, if the two differ in one or more characters, the male parts of the gynandromorph should be expected to be like the mother, i. e., maternal, and the female parts should be paternal if the paternal characters involved are dominant. The pos- sibility of testing Boveri's hypothesis was pointed out by one of us (Morgan) in 1905, and a test case was apparently furnished by a hybrid gynandromorph of the silkworm moth described by Toyama. The result was not in harmony with Boveri's hypothesis, but since the relation of one or of two nuclei to sex was not then known for moths, the case is not decisive, as will be shown more at length later. On the other hand, Boveri's discovery of some preserved specimens of the original Eugster gynandromorph bees and his analysis of their hybrid characters seemed to show that the condition of these bees was com- patible with his theory. This evidence will also be taken up more fully later. We may anticipate our account of hybrid gynandromorphs of Drosophila and state that they furnish direct evidence against Boveri's hypothesis, for these flies at least. In 1905 Morgan suggested an alternative hypothesis based on the fact that more than one spermatozoon had been found to enter the bee's eggs. Should one only of these sperm-nuclei unite with the egg-nucleus, the combination would give rise to the diploid cells of the embryo, while if a second (or a third, etc.) sperm-nucleus should develop it would give rise to haploid cells in the rest of the embryo (fig. 1 B). On this view the haploid cells should be paternal and pro- duce male parts, and the diploid cells maternal and produce female parts, which is exactly the reverse relation in regard to parental origin of the male and female parts from that expected on Boveri's hypothesis. A decision as to which view is correct might be reached in any special case in which sex-linked characters enter from the paternal and maternal sides. As will be shown later, some of the evidence from the Drosophila gynandromorphs is incompatible with this hypothesis of Morgan. A third hypothesis that grew out of the work done in this labora- tory was published in 1914 by Morgan, based on evidence from the Drosophila cases. On this view the gynandromorphs are due to an elimination of one of X chromosomes, usually at some early division of the segmentation-nuclei. Rarely, in consequence of a delay in the divi- sion of one of the X chromosomes, one of the daughter-halves fails to reach its pole and is lost in the mid-plate or in the cell- wall (fig. 1 c). As a result, the embryo comes to carry two kinds of nuclei, one kind containing one X and the other kind two X chromosomes. The critical evidence in favor of this interpretation is found in the presence on both sides of the gynandromorph of other mutant characters whose genes are not in the X chromosomes, but in autosomes. If, for example, 0 THE ORIGIN OF GYNANDROMORPHS. the mother contains a mutant gene in one of her autosomes and the father contains its normal allelomorph, it is expected, on Boveri's view, that the male side of the gynandromorph should show this maternal autosomal character, even though recessive. But on the hypothesis of chromosomal elimination, both sides of the gynandro- morph should show the same autosomal characters. Conversely, if the cross is so arranged that a recessive mutant autosomal gene enters from the father's side, then, on Morgan's earlier view of polyspermic fertilization, the male side of the gynandromorph should show this recessive mutant character; but on the elimination hypothesis both sides should show the same (dominant) autosomal characters. It may now be shown by critical examples that the hypothesis of chromo- somal elimination will cover nearly all of the cases of Drosophila, and is therefore preferable to either of the other two, even although in special cases either of these two other ways of producing gynan- dromorphs may be realized. A few additional cases have been found that call for still other interpretations. The critical cases are as follows: A yellow white male was mated to a female pure for the recessive autosomal genes for peach eye- color, spineless body, kidney eye-shape, sooty body-color, and rough eyes. A gynandromorph was found (plate 1, fig. 1) that was male on one side, as shown by his shorter wing, sex-comb on the foreleg, and the shorter bristles characteristic of the male (the body was also slightly bent to the smaller male side), and female on the other side, as shown by the converse characters to those just given. The gynandromorph possesses on both sides all of the characters dominant to the five recessive autosomal factors that came in with the sperma- tozoon. On Boveri's explanation, the male side should have a yellow body-color and a white eye, because their two genes are carried by the maternal nucleus, while the female side should show the normal characters of the wild fly, as is the case. The absence of yellow body-color and white eyes on the male side rules out his explanation. On Morgan's hypothesis of polyspermy, the male side that comes from one or more supernumerary sperms should show the five auto- somal recessive characters brought in by each sperm, which is not the case, and the female side should show the normal characters, as it does. The absence of the five recessive characters on the male side rules out this explanation also. On the theory of chromosomal elimination the gynandromorph started as an ordinary XX female — one X carrying the genes for yellow and for white, the other carrying their normal allelomorphs, viz, genes for gray and for red. Either of these chromosomes might be the one to be eliminated, i. e.} at some division either one of the yellow white daughter chromosomes failed to reach one of the daughter cells, or one of the gray red daughter chromo- somes failed. If the former, the male side should get only the gray PLATE 1 : ( GYNANDROMORPHS OF DROSOPHILA THE ORIGIN OF GYNANDROMORPHS. 7 red chromosome, and show the corresponding characters, which in fact it does. If the other chromosome had lost one of its halves at the critical division, the male side should be yellow white, which is not the case. Evidently, then, it must have been a yellow white daughter chromosome that was lost in this case. In regard to the five autosomal characters, it is clear that since both male and female sides show all the dominant characters, both sides of the body received the autosome that bears their genes. This hypothesis thus covers the facts in the case. Sections of the abdomen showed abnormal gonads that appeared to be testes. Another gynandromorph is drawn in plate 1, figure 2. It, too, came from this same cross of a yellow white male by a female of a race with the same five recessive characters. It is not a bilateral TEXT-FIGURE 2. gynandromorph, but more nearly an anterior-posterior combination. The abdomen is male, and since the forelegs bear no sex-combs, some at least of the anterior end is female. One wing is male; at least it is shorter than the one on the opposite side, which is presumably female. As in the last case, the fly shows only the characteristics belonging to the normal allelomorphs of the five recessive autosomal factors. The analysis here is the same as above. Another gynandromorph, drawn in text-figure 2, arose from a cross between a male that was heterozygous for the two dominant autosomal genes for star eyes and for dichaete bristles and a female that was notch 8 THE ORIGIN OF GYNANDROMORPHS. ("short" type) . The mother had one sex chromosome with the dominant gene for notch and another sex chromosome that had the normal allelo- morph of notch and also a gene for eosin eye-color. The gynandromorph was male on one side, with an eosin eye (with a red fleck in it), a sex- comb, and a short wing on that side, and female on the other side with a red eye, no sex-comb, and a longer wing. The genitalia were male. The gynandromorph arose by the fertilization of an egg containing the sex chromosome bearing the eosin eye-color (because had the other maternal X chromosome been present one of the wings, or both, would have shown the notch character). In this case it was the X chromo- some from the father that was eliminated, since the male side shows the eosin eye-color of the maternal sex chromosome. Boveri's ex- planation will not fit this case, even though the male side shows a maternal character, viz, eosin eye, because that side is dichsete, hence contains dominant factors from the paternal autosome. Morgan's hypothesis of polyspermy will not fit this case, for the male side should have red instead of eosin eye-color, since red was brought in by the sperm. On the hypothesis of elimination, it is apparent that one of the daughter halves of the normal X chromosome was lost; the cells of both sides got the regular autosomal groups, for dichaete came from the father. The father was heterogyzous for star, and it must have been one of his gametes without star, but with dichsete, that fertilized the egg. Here again neither of the earlier explanations fits the case, but the third hypothesis covers it. Another gynandromorph was described in "Mosaics and Gynandro- morphs in Drosophila" in 1914. It was the first case discovered in which the presence of an autosomal factor made it possible to decide which of the three explanations was the correct one. A yellow white female was crossed to a male that carried a recessive autosomal gene for ebony body-color. The gynandromorph was preponderantly male on one side and female on the other. Both eyes were red and the body-color was gray (or possibly heterozygous ebony) on both sides. Here Boveri's explanation fails, because the male side should have been entirely maternal, therefore yellow and white; and Morgan's earlier explanation fails, because the male side was not ebony. On the elimination hypothesis a maternal yellow white daughter chromo- some was lost; hence both sides had red eyes and not yellow body-color, and both sides received the same normal autosomes. This cross, in which a yellow white female was mated to an ebony male, was carried out extensively (January to May 1914) and 6 more gynandro- morphs were found. However, in order to discriminate between partial fertilization and polyspermy on the one hand and elimination on the other only those cases are diagnostic in which the male parts come from the father and show at the same time autosomal parts from the mother. THE ORIGIN OF GYNANDROMORPHS. 9 Another gynandromorph (obtained by Sturtevant), text-figure 3, came from a mother that had in one second chromosome the genes for Ciri and for curved, and in the other the genes for black and for vestigial. She may have had a third chromosome gene for crossing- over. The father was homozygous for black, purple, curved, plexus, speck, all in the second chromosome. Brothers and sisters were as expected; the black curved crossing over was 28 per cent. The fly was black and showed no trace of purple, vestigial, curved, plexus, or speck. It was male on the left side, female on the right, except for head bristles. The genitalia were male. The fly was sterile. Unless the egg were a double cross-over for black vestigial curved, which is unlikely, it contained a black vestigial bearing chromosome. The sperm contained the five sec- ond-chromosome genes. Since the male parts showed none of these sec- ond-chromosome char- acters, except black, although all the rest ex- cept purple might have been visible, it is highly probable that the male parts contained both sec- ond-chromosomes. The result shows at least that the theory of chromo- some elimination is a more probable explana- tion than partial ferti- lization or multiple ferti- lization, and the result TEXT-FIGURE 3. would be conclusive if the possibility of double crossing-over were rejected. Another case (found by Sturtevant, 4079 C, Oct. 31, 1917) occurred in a cross betweem a male with a normal X chromosome and pure for the second-chromosomal genes for black, purple, and curved, and a forked female that was heterozygous for the second-chromosomal genes for black, purple, and curved. The gynandromorph (plate 1, fig. 3) had a short wing on the left side, but the left foreleg was not male. The abdomen had the male banding and genitalia and contained two testes. No forked bristles were found in any part of the body. Elim- ination of one of the forked-bearing maternal X chromosomes left the wild-type X chromosomes to determine the character of the male parts. 10 THE ORIGIN OF GYNANDROMORPHS. The gynandromorph must have received the normal second chromo- some from its mother (since normal autosomal characters only ap- peared) and a second chromosome from its father with the three recessive genes. Since neither male nor female parts show these recessive genes, two second chromosomes must have been present in all the nuclei, both in the male and in the female parts. FREQUENCY OF OCCURRENCE OF GYNANDROMORPHS. In general, we have no record of the frequency of the occurrence of gynandromorphs. They are found from week to week, their number being roughly in proportion to the number of flies passing under observation, and also in proportion to the care with which the flies are scrutinized in detail. On four occasions, however, the frequency of their appearance was recorded. In the first case (in 1914) a cross, involving yellow flies, white-eyed and eosin-eyed flies, and wild-type flies, seemed to give gynandromorphs more often than usual. It is to be noticed that the striking color differences of eye and body in this combination would, as a rule, make it easy to detect hybrid gynandromorphs, and their frequency may have been due to this fact. In all 32 gynandromorphs were found in a total of 42,409 flies, or 1 in 1,325. Duncan, in 1915, made a careful examination of hybrid flies and found 3 gynandromorphs in a total of 16,637 flies, or 1 in 5,500. All flies were so thoroughly scrutinized that probably most of the gynan- dromorphs that occurred were found. The third set of observations was made on material that was chosen because, in addition to sex-linked factors, autosomal genes were present, which should give an answer to the three contrasted hypotheses de- scribed in the preceding pages. In all, 2 gynandromorphs were found in a total of 4,979 flies. A fourth record made by Sturtevant also involved autosomal as well as sex-linked characters. Forked females were mated to males with normal bristles. The female was heterozygous for the second- chromosome genes, black, purple, curved; the male homozygous for the same genes; 3 gynandromorphs were found in about 24,000 offspring. Taking all these results together, the observed ratio is 1 gynandro- morph in 2,200 flies. Whenever the chromosomal elimination occurs at an early stage in development, or when the color or structural difference involved is striking, the gynandromorph is more likely to be found than when the contrary conditions are present. If elimination occurs late in develop- ment the region affected may be so small as to escape detection. It seems probable, therefore, that such irregularities may be more frequent than the figures given above indicate. THE ORIGIN OF GYNANDROMORPHS. 11 It is a curious fact that practically all of the mosaics of Drosophila involve the sex chromosomes. It is true that the differences in the sexes are so marked that individuals partly male, partly female, could easily be detected on this basis alone. On the other hand, the mutant characters that are sex-linked are not more striking than are those of autosomal mutants. The almost complete absence of the latter kind of mosaics in our cultures shows very positively that elimination is very infrequent in these chromosomes, or, if it occurs, that an individual or part with only one autosome is less likely to survive than an individual with one X chromosome. Until this question is settled it can not safely be concluded that the sex chromosomes suffer elimination more than do the autosomes. The fact that autosomal non-disjunction has not yet been observed in Drosophila, though looked for, lends support to the view that variations in autosomal number are either rare or are fatal. RELATIVE FREQUENCY OF ELIMINATION OF THE MATERNAL AND PATERNAL SEX CHROMOSOME. It might have been supposed a priori that delay in the unraveling of the chromosomes of the sperm might be the most frequent cause of the elimination of chromosomes. As a matter of fact, the evidence shows clearly that the maternal X is as likely to be eliminated as the paternal. For example, we find on looking through our records that in 15 cases the maternal X chromosome and in 15 cases the paternal chromosome must have been the one eliminated. There were 16 cases hi which from the nature of the cross or of the result it could not be determined which one was eliminated. In the above estimation we also have left out of account all cases that were entirely male, or for which special explanations are called for. There can then be no doubt but that elimination is somehow connected with the nature of the X chromosomes themselves, such as slowness in dividing or in reaching the poles of the spindle, and that elimination is not due to delay in the development of either pronucleus. An examination of the gonads in Drosophila gynandromorphs has shown hi every case that the two gonads are the same, i. e., both are ovaries or both are testes. Even in bilateral types the two gonads are alike. Duncan found this true for the few cases that he sectioned. This number was, however, insufficient to establish the rule, but we can now add about 20 other cases to the list. There can remain no doubt that the gonads are alike, regardless of the way in which the male and female parts are distributed on the surface. The results are in accord with the early formation of the germ-cells in Diptera and probably mean that both gonads are derived from one and the same cleavage nucleus. 12 THE ORIGIN OF GYNANDROMORPHS. DISTRIBUTION OF SEGMENTATION NUCLEI AS DEDUCED FROM DIS- TRIBUTION OF THE CHARACTERS OF GYNANDROMORPHS. If the first division of the segmentation nucleus corresponds with the right and left sides of the embryo, and if chromosomal elimina- tion is more common at this tune or more easily detected, we should expect most gynandromorphs to be roughly bilateral. We have found that this is the most frequent type. If the first division were in the antero-posterior direction and elimination were frequent at this time, we should expect to find some gynandromorphs with the anterior end of one sex and the posterior end of the other sex. This type also is fairly frequent. If the first division were dorso-ventral we might expect correspond- ing gynandromorphs, but, although more difficult to detect, they appear almost never to be of this kind. If the second division were a time of elimination we would expect quadrants instead of halves. Such cases are known. The striking fact about the gynandromorphs is that large regions of the body are involved. Granting that later differences would be less easily detected, in certain organs at least, the results are so em- phatically in favor of large parts of the body being involved that we think it highly probable that the elimination is most frequent in the first division. The difficulty of reaching a decision is greatly increased when it is recalled that from the ventral plate of the embryo the serosa is formed by a folding upward of the sides of the plate. How much of the ventral ectoderm is carried in this way to the dorsal surface is not known. Should it replace the dorsal covering derived from the segmentation nuclei (that goes then into the serosa which is later thrown off), the results for ectodermal organs are restricted to the regions on each side of the ventral plate. The mesoderm also grows from the ventral to the dorsal surface, and presumably mesodermal dorsal structures have come from ventral material. A further complication arises in connection with the imaginal plates out of which many adult organs are produced. Unless the exact origin of their cells is known, it is not possible to safely conclude at what time the early elimination takes place. STARTING AS A MALE VERSUS STARTING AS A FEMALE. The evidence recorded in the preceding pages is analyzed on the basis that the gynandromorph starts as an XX individual, or female, and that the male parts arise by the elimination of an X from one of the cells. The evidence from hybrid combinations shows very clearly that practically all of our gynandromorphs have started as XX individuals, as 19 are more female, 14 nearly equal, 6 more male. THE ORIGIN OF GYNANDROMORPHS. 13 There are, however, other theoretical possibilities that should be noticed, for it is possible that gynandromorphs may sometimes arise in other ways. In fact, one or two of those we describe may be ex- plained in the following way : An X egg fertilized by a Y sperm (a regular male), might later become partly female, i. e., gynandromorph, through somatic non-disjunction, both daughter X's remaining in the same cell at some early embryonic division. Parts descended from the XX Y cell are female; the other (Y) cell would presumably die. If such a process occurred at the first division and all of the yt>lk was later occu- pied by the viable XXY cells, the embryo would become entirely female, although containing only sex-linked genes from the mother, and might be mistaken for a case of 'primary non-disjunction.' A non-disjunctionally produced egg containing a Y chromosome or an egg without a sex chromosome fertilized by an X sperm might also, starting as a male, produce a purely paternal female or female parts (mosaic) through somatic non-disjunction. If non-disjunction occurred at a late division a proportionately smaller part of female tissue would be formed and the regular male cells formed earlier would give male parts — i. e., the individual might be more male than female. There are no cases where these explanations only will apply, but a few cases accounted for by chromosome elimination may be also explained in one or the other of these ways, viz, that the gynandro- morph started as a male. CYTOLOGICAL EVIDENCE OF CHROMOSOMAL ELIMINATION. The most important case of chromosomal elimination involving one of the sex chromosomes, and therefore most like the case of gynandromorphism in Drosophila, has been described in Ascaris (Rhabditis) nigrovenosus by Boveri and by Schleip. In this nematode there is a hermaphroditic generation that lives in the lungs of the frog. Eggs and sperm are produced at the same time in the her- maphroditic gonad. The full number of chromosomes is the same in the early oogonia and spermatogonia. This number is reduced to half in the egg and also in the sperm at the reduction division, but while all the eggs are alike, there are two kinds of spermatozoa, one containing one less chromosome than the other. This loss of one of the chromosomes in one-half of the sperm-cells is apparently brought about as a regular process by the failure at reduction of one member of the paired sex chromosomes to reach the pole. It is caught at the division plane or else remains near that plane and disappears. This process differs however, from what we suppose to occur in eliminating a sex chromosome in Drosophila when a gynandromorph is produced in that an undivided X is lost. Whether in Ascaris this process occurs in all the cells at a given division or is somewhat irregular is not certain, and can only be determined by a fuller knowledge of the ratio of males 14 THE ORIGIN OF GYNANDROMORPHS. to females that result. Boveri thought, from the evidence obtained, that the loss of one chromosome at this time is a constant phenomenon. If so, it differs in this regard from the rare occurrence of elimination in Drosophila. In the group of aphids and phylloxerans a process occurs that has at least a certain analogy to elimination. When the male-producing egg, which is smaller (in the latter group) than the female-producing egg, throws off its single polar body, one sex chromosome is eliminated from the egg, although the autosomes divide equationally at this time. This elimination is not due to loss of a daughter chromosome, because it is preceded by a sort of synaptic union and disjunction of the chromosome in question. Here the lagging of one whole chromosome in the middle part of the spindle, and its failure to reach the outer pole in time to become incorporated in the nucleus of the polar body, furnishes a certain resemblance, at least, to the elimination process. In one species, P. fallax, there are four sex chromosomes, two of which are eliminated from the male-producing egg, as described above. There remain, then, two sex chromosomes in the male. When the Sperms are produced these two do not act as mates when the other chromosomes (autosomes) pair and segregate, but both pass together to one pole. The daughter cells that get them become the functional female-producing spermatozoa; the other cell that lacks them de- generates. Here, then, although two sex chromosomes are present, they both pass to one pole. This behavior is quite unlike the results produced by chromosomal elimination. In one of the aphids Morgan found a cyst in which, owing apparently to the failure of the autosomes to pair before segregation, an irregular distribution of the chromosomes took place, including an erratic dis- tribution, somewhat imperfect, it is true, of the sex chromosomes also. This unusual and irregular occurrence might lead to complica- tion in the distribution of the sex chromosomes in the next generation, if such sperm were to become functional, and furnish a parallel case to the phenomenon of primary non-disjunction that Bridges has described in Drosophila. In Drosophila there takes place on rare occasions an erratic distribu- tion of the sex chromosomes, either in the male or in the female, that has been called primary non-disj unction. Occasionally, both sex chromo- somes are eliminated in the polar body, leaving in the egg the haploid number of chromosomes, but not a sex chromosome. If such an egg is fertilized by a female-producing sperm containing one X chromosome, an XO male results. The male, lacking the characteristic Y chromo- some of the normal male, nevertheless resembles a normal male in all respects, except that he is sterile. Conversely, in other cases, both X chromosomes may remain in the egg. Such an egg does not develop if it is fertilized by a female-producing sperm giving it three X's, but THE ORIGIN OF GYNANDROMORPHS. 15 if such an egg is fertilized by a male-producing Y-bearing sperm, it produces a female XXY, that is like a normal female in its somatic characters; but such a female, owing to the presence of three sex chromosomes (XXY), gives rise to the phenomenon of secondary non-disjunction to be described presently. Similarly in the male, primary non-disjunction may take place in the formation of the spermatozoon. If at the reduction division the X and Y chromosomes, that normally pass to opposite poles, should pass to one pole, a spermatozoon would result from one of the daughter cells that contains both an X and a Y, and such a sperm fertilizing an X-bearing egg would give rise to an XXY female that would exhibit secondary non-disjunction. The other daughter cell without X or Y also produces a functional sperm. In these cases of primary non-disjunction an irregular distribution of the sex chromosomes leads to unusual types of sex-linked inheritance, but not to gynandro- morphism or to mosaics. In secondary non-disjunction, owing to the presence of three sex chromosomes, any two of which may form a pair, there is left one chromosome without a mate. Genetic analysis shows that the un- paired chromosomes, in some cases one of the X's, in others the Y, may either pass out of the egg at maturation or remain in the egg. Aside from this irregularity there is not much in the process that is akin to the kind of chromosomal elimination postulated for gynandro- morphs, since the processes underlying the two phenomena are prob- ably quite different. These cases furnish exceptions in regard to genetic behavior and furnish important evidence bearing on the deter- mination of sex, but do not lead to the kinds of effects seen in the pro- duction of gynandromorphs, except when the non-disjunction occurs at a cleavage stage, as already explained. As stated, Boveri based his hypothesis of gynandromorph produc- tion on an earlier observation that he had made with the sea-urchin eggs. He found that occasionally the egg-nucleus began to divide before the sperm-nucleus had fused with it. In consequence, the sperm-nucleus fertilized, as it were, only one-half of the egg; i. e., it approached one of the two daughter nuclei, and later became incorporated with that one. In consequence, all the nuclei descending from this fusion had the diploid number of chromosomes, while the nuclei descending from the single daughter egg-nucleus had only the haploid number. In the sea-urchin it has not been found possible to raise plutei to maturity; hence the effect of this partial fertilization on sex could not be determined. Boveri's application of this evidence to gynandromorphs of the bee was purely theoretical, since at that time the genetic evidence, that has since become available, did not exist. At about the same time Herbst carried out some experiments with sea-urchin eggs that enabled him to produce a large number of em- 16 THE ORIGIN OF GYNANDROMORPHS. bryos in which a process similar to that just described took place. The unfertilized eggs were stimulated to parthenogenetic develop- ment by placing them in sea- water containing a little valerianic acid. After a few minutes the eggs were returned to sea-water and sperm added. The sperm-nucleus did not penetrate in many cases until the egg nucleus had begun to divide and then, as in Boveri's case, it often united with one of the daughter nuclei. In neither of the cases is there any elimination of single chromosomes, but in a more general sense the earlier group of paternal chromosomes was dislocated in that it failed to reach its normal destination. The extremely important experiments that Baltzer made with sea- urchin eggs resulted in demonstrable cases of elimination, but here of whole undivided chromosomes. For instance, when the eggs of Strongylocentrotus are fertilized with the sperm of Sphcerechinus, it is found at the first division of the egg that while some of the chromo- somes divide and the halves move to opposite poles, other chromo- somes remain in place, or become scattered irregularly between the two poles of the spindle. They appear later as irregular granules and show signs of degeneration, ana although remnants of them may persist for a while, they take no further part in the development. The maternal egg-nucleus contained in this case 18 chromosomes and likewise the paternal sperm-nucleus. Hence, after union and division, 36 chromosomes should go to each pole of the segmentation spindle if all divided. Baltzer found, however, only 21 chromosomes at each pole, which means that 15 chromosomes have failed to behave normally, and it is probable that these are derived from the paternal nucleus. Three chromosomes only of the latter, on this interpreta- tion, take part in the division. In consequence, the nuclei of the embryo contain almost exclusively maternal chromosomes, and it is significant that the larvae are largely or entirely maternal in char- acter. It is true that we have no evidence to show at present that the larvse of these sea-urchins differ in only one or more Mendelian factors. It would be very surprising if such were the case, yet the results show at least so great a preponderance of maternal characters that we must infer that the three surviving paternal chromosomes produce no marked difference. The reciprocal cross gave a different result. When the eggs of Sphcerechinus are fertilized by the sperm of Strongylocentrotus, division of all of the chromosomes takes place normally and 36 are found at each pole. The pluteus that develops shows peculiarities of both paternal and maternal types. The difference between the two crosses is probably due to the observed differences in the behavior of the chromosomes. In the first case, the lagging and subsequent degenera- tion of certain chromosomes may be spoken of as a sort of elimination, although the causes that bring it about must be supposed to be of a THE ORIGIN OF GYNANDROMORPHS. 17 different kind from those involved in Drosophila when a half of a single chromosome fails to reach its normal destination. EARLIER HYPOTHESES TO EXPLAIN GYNANDROMORPHS. Dalla Torre and Friese (1897) and Mehling (1915) have reviewed the earlier attempts to account for gynandromorphs. Donhoff (I860) suggested that gynandromorph bees arose from eggs with two yolks, one of which was fertilized, the other not; one began to form a worker, the other a drone, both fusing into one individual later. A second interpretation based on Dzierzon's theory was also suggested, viz, that the egg contains the male potentiality, the sperm the female poten- tiality. In fertilized eggs the latter influence usually predominates. In the gynandromorph, one of these influences predominates in one region, the other in other regions. In 1861, Wittenhagen suggested that a queen that produces gynandromorphs has reached a higher stage of fertility which causes male parts to arise even after fertiliza- tion. Menzel (1862) made several guesses, such as that delayed fer- tilization of the egg leads to irregular distribution of the mass of the sperm material with consequent disturbance in the development. Later (1864) he suggested that abnormal organization of the oviducts, leading to delay in passing of the egg, interferes with the sperm, so that the egg no longer has the possibility of producing a complete female, except in certain regions of the body. Von Siebold (1864) thought that insufficient fertilization is re- sponsible for the appearance of gynandromorphs. He assumed that a definite number of spermatozoa are necessary to produce a female. When from any cause an insufficient number of sperms is present, the egg can not develop a female, or a male, but an intermediate type. According to Cockayne (1915, p. 117), Scopoli (1777) suggested that a gynandromorph of Phalcena pini might have arisen through the fusion of two pupae lying in one cocoon. Donhoff' s suggestion (as above) of two yolks in one shell that fused is a somewhat similar view, and Wheeler in 1910 made a like suggestion, viz, that two eggs (fer- tilized?) fused at a very early stage, one a male-producing, the other a female-producing. Such a process will not apply, however, to most of the cases in Drosophila, because the evidence shows that the eggs are normally not of two kinds. The male alone produces two kinds of gametes. The sex-linked characters in hybrid gynandromorphs show very clearly that the results are not due to the fusion of two eggs, but to a different sort of process. In the bee also it appears that there is only one kind of egg, and that the female sex is determined by the fertilization of the egg; the male comes from the unfertilized egg. On the other hand, there are several cases in Drosophila which can not be explained by simple chromosomal elimination, but which can be explained on the assumption that the egg had two nuclei. Here 18 THE ORIGIN OF GYNANDROMORPHS. the appeal is made to a binucleated egg in order to account for the distribution of the sex-linked characters, but only indirectly for the sex differences hi the gynandromorph. The different sexes in the two parts are due to fertilization of the two nuclei by male and female producing sperm respectively. The presence of two nuclei in these eggs is easily explained as due to the fusion of two oogonial cells or else by an oogonial nuclear division without cytoplasmic division. The conditions existing at the completion of the last oogonial division are particularly favorable for such a union, for at this stage from a collection of cells (presumably all alike) the most favorably situated turns into the egg and the others into nurse-cells very intimately con- nected with the egg-cell. This view, while similar to Wheeler's, puts a different emphasis on the facts, for here the presence in the eggs of two nuclei does not directly account for the different sex of the parts of the gynandromorph (for this difference is due to the two kinds of sperm that have entered), but explains the distribution of the sex- linked characters in the hybrid gynandromorphs. On the other hand, Wheeler's idea is that two eggs in themselves determined as male and female fuse bodily, i. e., side by side, to give rise to male and female parts respectively. His view would be more nearly realized in the case of moths where the female is the heterozygous sex, and consequently a binucleated condition can be utilized directly to ex- plain not only the difference of sex in the gynandromorph (one nucleus retaining a Z and the other a W chromosome), but also the autosomal mosaics, as in the cases described by Toyama. Arnold Lang suggested another possibility in 1912, viz, that an egg that had developed parthenogenetically to the stage when the first two nuclei were formed might be fertilized by a female and a male producing sperm, each sperm uniting with one or the other of the two egg-nuclei. As a result one half should be male, the other half female. The hypothesis will not apply, however, to the bee — the forms whose parthenogenetic process of development would seem to best fit such a view — because only one kind of sperm is supposed to be produced. Double nuclei should produce female parts. The explanation will also obviously not apply to such cases in Drosophila as those in which the male half shows maternal recessive factors. De Meijere (1910-11) has offered certain suggestions concerning the origin of gynandromorphs. He starts from the old idea that each individual, male or female, contains within itself the characters of the opposite sex. He thinks that this holds for the gametes as well as for the somatic cells. Darwin held a similar view and thought that this was true not only for the primary sex-cells (sperm and eggs) but for the secondary sexual characters as well. To-day, however, it is clear that such a statement, at least in regard to the established cases of sex determination by means of sex factors, calls for a more definite THE ORIGIN OF GYNANDROMORPHS. 19 pronouncement as to the sense in which the phrase is employed; otherwise it is little more than a play on words. For instance, when one X chromosome is present the individual is a male, which means that one X plus all the rest of the cell makes a male, and when two X's are present, these two plus all the rest of the cell make a female. In what sense can such a statement be twisted to mean that each such combination contains in a latent condition the opposite condition? Compare the facts with a similar chemical situation and the absurdity of the inclusion hypothesis is evident. Maltose has the formula C^HjKjOn and glucose the formula CeH^Oe- One is twice the other minus one H2O. To state that maltose contains glucose latent or that glucose contains maltose latent is obviously absurd, yet this does not differ much from the view that each sex contains the opposite one in latent form. De Meijere thinks that gynandromorphs can be explained in "that the activation of the opposite sex (opposite to the one already under way) has started in, relatively later, after all the parts have taken on their definite positions; many of the parts have gone too far in the first direction, i. e., they are too old, but those that have not may be turned aside and produce the oppo- J K site results. ' ' 1 This view is offered to account for mosaics of sex char- acter. The bilateral gynandro- morph, he supposes, owes its origin to the above changes having taken place very early, even at the first division. De Meijere thinks ap- parently of effects being produced by external factors of some unknown kind rather than internal ones connected with a sex mechanism. His idea is too vague to be of use and too remote from present-day knowledge about sex determination to call for extended criticism. Arnold Lang, accepting the same general conception of sex and expressing what he believed to be the real relations by means of the formulae that Goldschmidt had advocated, offered another possible interpretation of gynandromorphs that is superficially exactly like the theory of chromosomal elimination which the results in Drosophila show to hold for this insect. In fact, Lang's view, if divested of the unnecessary encumbrance of De Meijere's conception and of Gold- schmidt's formulae, is then identical with the theory of chromosomal elimination. For example, Lang represents the fertilized egg (one that will give rise to a female) by the scheme shown in text-figure 4. The primary sex characters for the male are M carried by a pair of 1 See Goldschmidt's view in respect to the rate of development of male and female organs in the intersexes of the gipsy moth. TEXT-FIGURES 4 AND 5. 20 THE ORIGIN OF GYNANDROMORPHS. chromosomes that also carries the factors for the secondary sexual char- acter A. The primary sex character of the female is represented by F, carried by a second pair of chromosomes, and the secondary sexual character by G, both as before, carried in the same chromosome. In other words, the two pairs of sex chromosomes are (FG) (FG) and (MA) (MA) for the female, and, for the male, (FG) (fg) and (MA) (MA). Lang suggests that a loss by mutation takes place in females (as above) in the sense that one FG disappears and may now be repre- sented by (fg). The resulting division is shown in text-figure 5. The mutation causes the sex-balance in the cell on the right side to turn into a male, while that of the left remains a female. Lang appears to mean that the "mutation by loss " is the loss of a daughter chromosome. If we ignore the special interpretation of sex employed by Lang and borne out by his formulae, his view has several points in common with the hypothesis of chromosomal elimination. It should be noted, however, that there are also differences in the application of Lang's and the present interpretation, when the question of the sex-linked factors is involved, because the two X chromosomes represented by Lang by FG, FG carry many other genes, besides those for sex, even some for secondary sexual characters. Which of these comes to expression in the hybrid gynandromorph depend on which FG is elim- inated and not on the resulting change in balance (epigenetic effects) between the FG's and the MA's. Furthermore, Lang's scheme in- volves the relation between two pairs of chromosomes (four in all) while in the actual case of Drosophila only one pair is needed to account for all the facts. Cockayne, in 1915, announced independently the same view of elimination that Morgan had published the year before. He had found several halved gynandromorphs, all of which showed the specific characters of both parents on both sides. Both parental nuclei must therefore have contributed to both sides. He points out that since the division into male and female parts sometimes coincides with other characters the latter must be carried by the sex chromosomes. Doncaster, in 1914, described binucleated eggs in Abraxas, each nucleus giving off its two polar bodies and each being independently fertilized. He suggests that gynandromorphs might arise from such eggs, but did not obtain any in the particular lines that showed such binucleated eggs. The two gynandromorphs in Abraxas that Don- caster described later (1917), and which are considered here on page 85, he did not attempt to explain by this condition. The gynandromorphs of Drosophila have been from the time of then* first appearance in our cultures, about 8 years ago, a subject of general interest and discussion, especially by Muller, Sturtevant, Bridges, and Morgan. Their relation to the gynandromorphs in bees and to the theories of the origin of the latter has been constantly THE ORIGIN OF GYNANDROMORPHS. 21 under discussion. The critical evidence that shows that they were not due to separation of whole maternal and paternal nuclei was first obtained and published by Morgan (in 1914). Prior to that time Bridges (1913) had published an account of two hybrid gynandro- morphs, and had suggested that they were due to somatic non-dis- junction. By this term it was meant at the time that at an early embryonic division of a female the two daughter halves of one of the X chromosomes did not disjoin from each other to pass, as normally, into sister cells, but were included in the same cell, the other cell not receiving its half. The non-disjoining X was assumed to divide normally and the result was an X cell developing into male parts and an XXX cell developing into female parts. This hypothesis served to explain all the facts known at that tune. Soon, however, it was established (Bridges, 1916) that XXX individuals are unable to sur- vive, and this brought into question the conclusion that the female parts of gynandromorphs were XXX. This difficulty was later avoided by the assumption of " etimination " (earlier called "mitotic dislocation," Morgan, 1914). As already stated, this meant that one of the daughter X's was caught by the mid-plate and prevented from taking its place in either nucleus. There is another class of gynandromorphs (including here four cases) in which another procedure may account for the results. Primary equational non-disjunction occurred, as evidenced by the presence in each of the four gynandromorphs of two X chromosomes from the mother, one of these being a non-cross-over and the other a cross-over X, as is usual for XX eggs produced in this fashion. This XX egg was then fertilized by an X sperm, giving an XXX individual. This XXX zygote is prevented from dying and at the same time converted into a gynandromorph by the occurrence of somatic reduction at the first or a very early embryonic division. In each of the four cases the male parts of the gynandromorph were derived from one of the two maternal X's, which suggests that the essential feature of this somatic reduction is the active separation of the two X's that came from the mother and the passive inclusion of the X from the father with one or the other of them. There have been other cases which may support this view, cases in which XX eggs equationally produced have been fertilized by Y sperm, and then the two X's have likewise reduced, with the result that each cell gets one X, and the entire individual is converted into a male which is a mosaic of different parts clearly marked by the character corresponding to the two dif- ferent X's. The difficulty with this view is that it assumes that reduction can take place between two X's at a cell division without the X's themselves splitting, although all of the other chromosomes do so at this tune — a situation for which no support is given by cytology. It is to be noted in this connection that all cases that appear 22 THE ORIGIN OF GYNANDROMORPHS. to belong to this category are also explained by the assumption that the egg started with two nuclei, and in the description of cases both of these views are given as alternatives. THE ORIGIN OF THE GERM-CELLS IN FLIES. In several species of flies (Miastor, Chironomus, Calliphora) it is known that the germ-cells of the ovary or testis arise from a single cell at an early stage in the cleavage. In Miastor, for instance, when the four first-formed nuclei divide, one of the eight daughter nuclei moves to one pole of the egg, where it becomes surrounded by the peculiar protoplasm of this pole and subsequently pinches off from the surface of the egg. From this single cell by later division arise all of the germ-cells. A similar process has been described for other species of flies. If this holds also for Drosophila it follows that all of the germ-cells must be either eggs or sperm, regardless of whether the somatic parts are male or female. On the other hand, if the germ- cells hi Drosophila and in the bee are formed as in some of the other insects, i. e., in the beetle Calligrapha described by Hegner, where 16 cells simultaneously reach the polar field, it would be possible for some of the cells to have descended from one of the first two segmentation nuclei and some from the other. In such a case, if the first-division figure underwent elimination, both ovaries and testes might appear in the same individual. In butterflies and moths, where many gynan- dromorphs have been dissected, several cases in which both testes and ovaries occur are known. This is also the case in bees. A difference in the time of isolation of the germ-cells in these groups and in Dro- sophila may account for the difference in the results. COURTSHIP OF GYNANDROMORPHS. Sturtevant's paper on sex recognition and sexual selection hi Dro- sophila gives a full account of the rather elaborate courtship of this fly, in which the behavior of the two sexes is quite different. The re- actions of an animal, male on one side female on the other, or of one that had a female head and a male abdomen, might be expected to furnish interesting conclusions as to the relative importance of the sense-organs versus the reproductive organs in the behavior during courtship. Sturtevant tested 6 gynandromorphs. One was male throughout, ex- cept the genitalia, which were female. It behaved as a male. Sections of the abdomen showed one abnormal egg present. Another had 2 sex- combs, right and left, and the right wing was shorter than the left. The abdomen was female. She produced at least 1 egg. Sections of the abdomen showed 2 large eggs and a degenerate ovary present. She courted and was courted, thus giving both reactions. A third was THE ORIGIN OF GYNANDROMORPHS. 23 male, except the genitalia, which were female. Sections showed an abnormal testis near posterior end. It courted and was courted. Sturtevant records observations on three other gynandromorphs tested for sexual behavior: "None showed any certain indications of male behavior, but all were vigorously courted by males. Of these three gynandromorphs the external characters were as follows: (A) All female, except one side of the head, which was male; (B) female on one side of the whole body, male on the other side; (C) female, except the genitalia, which were male." Duncan describes the behavior of a bilateral gynandromorph. Its mating instincts were found to be indifferent. It was courted by males but would not court females. The gonads were both testes with ripe sperm. In a second gynandromorph, the eyes were female, but the forelegs had sex-combs; one wing was long (female); the ab- domen was male type, but the genitalia were half male, half female. Two ovaries were present. The fly was courted " assiduously" by males but would not mate. A third gynandromorph was without sex-combs on the forelegs, the wings were the same length, but the abdomen was male on one side, female on the other, as were the external genitalia also. Mature sperm were present in both testes. This fly was anteriorly female and posteriorly half male and half female. A normal male courted this gynandromorph when in front, but did not copulate with it. The gynandromorph drawn in text-figure 34 was tested by one of us (Morgan, T. H., Amer. Nat., 1915, p. 246). One side of the head and thorax is male, the other side female. The abdomen is pig- mented above as in a male and there is a penis below. When put with mature unmated females it did not court them, although it was quite active. Attempts to breed from gynandromorphs have been often made. It was not to be expected that those in which the genitalia were mixed would successfully copulate. Those with female abdomen have more often given offspring. Since, as explained elsewhere, the gynandro- morphs with male abdomen would not be expected to be fertile (be- cause the XO combination has been shown to be sterile), the frequent failure to obtain offspring from such males is in accordance with ex- pectation. On the other hand, an occasional fertile male gynandro- morph occurs. In these cases the combination was known or suspected of being XXY, the presence of the Y chromosome making the male (XY) fertile. PHOTOTROPISM IN MOSAICS WITH ONE WHITE AND ONE RED EYE. On several occasions it has been observed that when a mosaic had one red and one white eye it circled to the red side. This behavior is expected from observations by McEwen on the light reaction of flies 24 THE ORIGIN OF GYNANDROMORPHS. from white-eyed stock. He showed that these flies respond much less actively to light than do red-eyed flies. In these red-white mosaics the red eye, giving a stronger positively phototropic reaction, turns the fly toward that side. Of course, if the fly turns toward a single source of illumination, such as a window or artificial light, the red eye will soon pass into its own shadow as the fly turns, and the con- dition on the two sides may become balanced, unless the general illumination from the wall of the room, for instance, is still stronger than the influence of the window's light on the white eye. In order to avoid this complication the fly should be kept on a vertical surface held at right angles to the light, when its circus movements are not interfered with by the opacity of its own body. Since the male side of the body, including the legs, is generally smaller than the female side, and since the male side is the one that has the white eye, there is a chance that the movements toward the red side are against the stronger action of that side. This complica- tion was, however, not realized in all the cases in which circling occurred, but since in several of them the legs on the right and left sides were the same it is practically certain that the results are largely, if not entirely, due to the difference in stimulus from the two eyes. SEX-LIMITED MOSAICS. By a sex-limited character (in contradistinction to a sex-linked) we mean a character that is peculiar to one or the other sex, but is not necessarily transmitted by means of a gene in the sex chromosome. Such a character is shown by a stock called white tip, in which the pigment bands are absent from the last segments of the abdomen in the female but not in the male. In this stock a gynandromorph arose (text-fig. 6), male on the left side and female on the right. On the male side the black tip to the abdomen is present, although here, as in the stock itself, it is not as black as in the wild type. On the female side the abdomen has a white end. In this case elimination of a sex chromosome produced the gynan- dromorphous condition, and since in this stock the female parts are different from the male, owing to a factor presumably not in the sex chromosome, the right side of the gynandromorph also shows this peculiarity, owing to its femaleness. A similar case appeared (No. 2864, Jan. 1915) in a cross between a faint-band female and a star faint-band male. Faint-band is a sex- linked character which appears only in the female. All of the flies of the above cross were pure faint-bands; but while the females were characterized by abdominal bands in which both chitinization and pigmentation were weak and by short, slender, and irregular bristles throughout, the males could not be distinguished from wild males in appearance. The gynandromorph was completely bilateral, the THE ORIGIN OF GYNANDROMORPHS. 25 right side being male, with sex-comb, smaller eye, wing, etc., and the right side of the abdomen with male coloration. The genitalia were half-and-half also. The interesting feature was that throughout the female left side the bristles were weak and irregular and the bands "faint, " while the male right side was entirely wild-type in appearance. Another striking case appeared (on March 23, 1916) among the offspring of a pair, the female of which was heterozygous for the sex- limited character ("side-abnormal") and the father was pure for it. The character "side-abnormal" is sex-linked in inheritance and sex- limited in appearance, being seen only in females. In this mutant the bands of the abdomen of the female are "abnormal" at the sides, i. e., while the mid-dorsal part of the band is normal the ends of the band where they come around the side are cut away irregularly to ragged points and the color is etched with white splotches in the dark. The ventral plates are much smaller and are irreg- ularly rounded. In the male all parts are as in the wild flies. This gynandromorph (3806) showed a nor- mal male right half of the abdomen and a female left half, with all the characteristics of the side-abnormal character. The ven- tral plates were full and normal in the male parts and small and irregular in the female parts. Other evidences of maleness were present — a sex-comb on the right foreleg and a smaller right wing. Elsewhere in the text we have described several other cases involv- ing characters both sex-linked and sex-limited. Thus in gynandro- morph 7530, page 46, the male eye on the right showed marked devel- opment of the character facet, as in the normal facet male, while the female left eye, also facet, could hardly be told from wild-type, as is usual in facet females. All gynandromorphs involving eosin eye-color show the light type of eosin in the male eyes and the dark type in the female eyes. TEXT-FIGURE 6. 26 THE ORIGIN OF GYNANDROMORPHS. SOMATIC MOSAICS. Somatic mosaics can be accounted for by autosomal elimination in the same way that gynandromorphs are accounted for by X-chromo- somal elimination. Somatic mosaics might also be expected to arise from binucleated eggs and to be as often found as are gynandromorphs with the same origin. As a matter of fact, we have found only one certain case, which is less than expected on the latter view. The case is as follows : The grandmother was spineless (third-chromosome recessive) and the grandfather was spread (another third-chromosome recessive). The daughters and sons were wild-type. A pair of these gave a 2 : 1 : 1 : 0 ratio, as expected, because of no crossing over in the male. One of the granddaughters (No. 561, Oct. 3, 1914, text-fig. 7) was a mosaic of spineless and not-spineless. The left side of the thorax and abdomen and the left wing and the middle and last left leg were spineless. The rest of the female (including all of the head and left foreleg) had long bristles and hairs of the wild type. Simple elimination of the third chromosome from the spread parent would explain this case were it not that the existence of an individual lacking an autosome is doubtful, because none have as yet appeared through autosomal non-disjunction. On the alternative view of a binucleated egg, one nucleus contained the spineless third chromosome, the other a spread-bearing chromosome; both nuclei were fertilized by X sperm bearing the spineless X chromosome, and gave the female spineless on the left side and wild-type on the right side. The fact that the overwhelming number of hy- brid mosaics are gynandromorphs, involving there- fore the sex chromosome, can not be explained as due to failure to discover autosomal mosaics if they occurred. In most of our cases these would be just as striking as in the cases where the sex chromo- somes are involved. Evidently some peculiarity in the separation of the halves of the sex chromosomes makes the elimination of one of the daughter halves more probable than in the case of other chro- mosomes. Such a supposition is, of course, in harmony with the pecu- liar behavior of the sex chromosome at the reduction division of the male, at least when it lags on the spindle. On the other hand, when it does divide, as in the female, no such peculiarity is recorded, and it is this reduction, rather than the former one, that we need for com- parison. TEXT-FIGURE 7. THE ORIGIN OF GYNANDROMORPHS. 27 SOMATIC MUTATION. That mutation may take place in somatic cells comparable to the mutation process in the germ-tract can not be doubted. The bud- sports long familiar to botanists probably furnish in some instances examples of this sort; but the best authenticated cases are the modern ones that have been analyzed by recognized genetic methods. Few examples are known to zoologists; the monsters, freaks, and duplica- tions that are frequently found are generally due to environmental effects on the embryo. If somatic mutation occurs in only one chromosome of a pair, as seems to be the case with germinal mutations, the immediate result will not be seen except when the mutation is dominant. In the case of mutation in the germ-tract, a recessive gene in one chromosome of a pair may likewise not have opportunity at first to express itself, but if it is carried to one of the offspring it will there become multiplied and get into daughters and sons (or in hermaphroditic species into pollen and ovules). Chance union of the gametes that contain the mutated chromosomes will later bring even the recessive genes to expression. It is more probable, therefore, that recessive mutations will appear in the sexually reproducing species more readily than in those with vegetative reproduction, except where the latter are already heterozygous. The same comparison may be made between parthenogenetic species and sexual ones. In the former, a recessive mutation appearing in one chromosome of a pair will have no oppor- tunity to show effects, and the line may be lost by chance alone. Preservation will be favored only if the heterozygous state has an ad- vantage over the original form. Sexual reproduction, therefore, has the advantage that every recessive mutation will have a far better chance of showing itself as a character modification and, if beneficial, of being preserved by natural selection. In fact, if it could be shown that a preponderant number of recessive mutations have furnished the material for evolution, it might possibly appear that we had some hint as to how the process has come to be such an almost universal method of propagation. On the other hand, dominant mutations might nourish, as well by the one as by the other method. The best authenticated case of somatic mutation in plants is that described by Emerson, who has brought forward convincing evidence that in corn a gene for certain types of variegation (striped seeds) mutates not infrequently to a gene for uniform-colored grain. The gene for medium variegated "mutates much more frequently than that for very light variegation. " By crossing plants from the mutated grains to pure recessive types Emerson has shown that when the mutation occurs it involves only one member (at a time) of the pair of allelomorphs in question. In these cases the mutation takes place in cell lines (subepidermis) that may ultimately contribute both to 28 THE ORIGIN OF GYNANDROMORPHS. the germ-tract and to the soma. Through the former, inheritance becomes possible, through the latter the effects of the mutation be- come visible only on the plant in which the mutation took place. There are other mutative changes in corn that Emerson describes in which the effect is only in the epidermal cells; hence, while it becomes visible in the plant in which it has taken place, it is not inherited, since the germ-tract does not come from this part of the plant. In the course of our work on Drosophila a few flies have appeared with characters which seem to have arisen by somatic mutation. If, as there is reason to suppose, the mutation changes that gave rise to them appeared in only one chromosome, the change must either have been dominant or, if recessive, in the single X chromosome of the male. Since visible mutations in the sex chromosome have been shown to be at least four tunes as frequent as dominants in all of the chromosomes together, the chance that these sporting characters are dominants is smaller than that they are recessive and in the sex chromosome. In support of the latter is the fact that nine out of ten of the sporting characters look like known sex-linked genetic char- acters, and more important still is the fact that all the cases so far found are males. (1) One of these somatic sports is shown in plate 1, figure 4. The right side of the body is pale, almost white. The history of this fly is as follows :* One of the X chromosomes of the mother contained the genes for lethal 7 and for forked, the other X the genes for yellow and for white. The X chromosome of the father carried the genes for yellow and for white. The fly was a yellow white forked male throughout, but the right side of the thorax, the right wing, and the right side of abdomen were pale, almost white, as shown in the drawing. Testes were present, with sperm. The pale light side is clearly due to somatic mutation, since no such pale body-color was present in the cross or was known elsewhere. Whether the mutation oc- curred in the X (if recessive) or in an autosome (if dominant) is undeter- minable, since the fly was not bred. (2) In another case (II 108, Oct. 21, 1913), the left side of the body, at least for a middle section, is brown in color, looking like the double recessive yellow black (text-fig. 8). The fly had the following history: Some F2 blacks from the cross of black by jaunty (both second-chromo- some) were inbred in an attempt to secure the double recessive black jaunty. One of the F3 black males had the left side of its thorax and abdomen, left wing, and left legs colored like the double recessive yellow black. It was at once assumed that mutation to yellow had occurred in the early embryo in the cells which gave rise to the left side. A test was made to see whether the germ-cells carried the mutant gene. The mosaic male was outcrossed to a black female and gave only black offspring (M 69, black 9 27, black d" 23). Three pairs and a mass-culture of these FI flies were inbred and gave a total ^o. 2493; November 20, 1915. THE ORIGIN OF GYNANDROMORPHS. 29 of 152 black 9 and 147 black cT, with no yellow-black offspring. Evidently, then, the testes came from a cell which had not mutated. While the "brown" color of the mosaic was like that produced by yellow acting with black, it is possible that the mutant gene was not the yellow already known, but a new yellow. (3) Among the grandchildren of the last somatic sport a fly was found with a wing of an unusual type (text-fig. 9). This wing was about half the usual length and had almost exactly the form of min- iature, but there was none of the dark color normally present in miniature wings. This wing seems to have been a new mutant type, the mutation having occurred in the early embryonic cells of the fly. There have been quite a number of such occurrences, some, as in the present case, giv- ing striking differ- ences. (4) A fly ap- peared in vestigial stock (August 13, 1912) with one normal wing (text- fig. 10). It was described as a case of somatic atav- ism. An alterna- tive view is also possible, viz, that a somatic muta- tion occurred else- where, i. e., in an- other chromosome or in another region of the second chromosome, of such a sort that it neutralized the effect of both genes for vestigial. In the cells containing this mutant gene the conditions for normal wings are again restored. (5) and (6) Two further cases of mutation in the male were found by Sturtevant (not published); both were males throughout; one had forked bristles on one side of the body, although there were no forked flies in the immediate ancestry. The other had a dark body-color on part of the thorax, there being no sex-linked dark body-color in the pedigree. Neither fly was tested. (7) In a stock pure for red eyes, miniature wings, and yellow body- colo r a fly appeared with all the characters of its race except that one eye was white with a fleck of red at its posterior edge (text-fig. 11). TEXT-FIGURE 8. TEXT-FIGURE 9. 30 THE ORIGIN OF GYNANDROMORPHS. Since there was no white in the stock, the white eye must have come by mutation and possibly by mutation to a sex-linked white-eyed gene. (8) In a mating in which both parents were pure bar-eyed flies a male appeared (1917) (text-fig. 12) in which both eyes were round and in addition one eye was three-quarters white, and the other had a fleck of white in it. A germinal mutation in the mother of bar to round eye must have taken place, as shown by the fact that when the fly was bred it produced only normal-eyed offspring. Since this male was normal, it must have come from the union of a Y-bearing sperm and an X egg. Since the bar gene is carried by the X chromo- some, it follows here that mutation must have occurred in one sex chromosome of the mother. It is significant in this connection to call attention to the fact that bar-eye not infrequently mutates (reverts) to normal, as May has clearly proven. The other change to white was due to a somatic mutation. (9) In stock pure for black and for miniature and impure for white and for red eyes a male appeared that had one white eye (text-fig. 13). It might appear here that simple elimination in a heterozygous female would account for the white eye, but if the fly arose in this way the rest of it should be female. Double elim- ination will, however, give a result of this kind, i. e., a red X is lost from one half and a white X from the other side, leaving both parts male, one red, the other white. If, on the other hand, the fly started as a red-eyed male and dislocation occurred, so that most of the fly had an X, the other part a Y chromosome, the expectation, based on the evidence from nondisjunction, would be that the male part would die. However, it might be claimed that the evidence applies to the fly as a whole and not to the survival of a small part of the body, which might very well be capable of living. But we should expect the absence of X to carry other consequences hi its train besides loss of eye-color, so that this explanation seems improbable. A third explan- ation is that of somatic mutation. It is not possible to decide between the assumption of double elimination and that of somatic mutation. (10) A somewhat similar case is shown in the male figured in plate 1, figure 5. Its ancestry is not now a matter of record, but probably it arose in red-eyed bifid stock that we had at the tune. If so, double elimination is excluded and the fly must have arisen by mutation in the sex chromosome. TKXT-FIGURE 10. THE ORIGIN OF GYNANDROMORPHS. 31 It is a matter of great interest to find that all the ten cases of somatic mutation that we have recorded in Drosophila have been males. The significance of this was not appreciated until the material had been sorted out for other purposes. It probably means that a recessive somatic mutation takes place in the sex chromosome and shows at once in a male in those parts of the body whose cells contain the mutant gene because the male has only one sex chromosome. Should a reces- sive mutation occur in the X chromosome of a female its effect would not appear in the soma because the normal allelomorph would conceal it. It is interesting to apply this point of view to certain results hi Lepi- doptera in which mosaics or gynandromorphs have been recorded that carry in parts of the body characteristics that are known to occur, although rarely, in varieties of sports of the species. TEXT-FIGURE 11. TEXT-FIGURE 12. TEXT-FIGURE 13. Among these a number have been described with one half of the body of one species and the other half of a varietal type of the same species. In some cases the variety is so rare that there might seem to be no question of a hybrid cross involved, since this in itself would be rare, and that both this and a later mosaic condition result is beyond reasonable probability. An alternative view would be that of somatic mutation. If such were the explanation we should expect the indi- vidual to be female and the mutation to have occurred in the single Z chromosome. In the cases brought together by Cockayne, in which the same individual is partly one species, partly a variety (1915, pp. 87-90), there are about 10 such cases recorded as females, 2 as males; in 12 cases no sex is stated by Cockayne. If further examination of the original sources shows as high a percentage of females as in the recorded cases, the evidence is in favor of the interpretation suggested above. The males call for another interpretation, and each such case will need special examination. 32 THE ORIGIN OF GYNANDROMORPHS. These cases are not to be confused with mutation in the germ- tract, where, in a sense, the reverse situation is realized, for while in Drosophila the mutation of a sex-linked character in one female chromo- some appears immediately in one (or more) of her sons, the mutation itself occurred first in the female. Conversely, in moths, if a germ- tract mutation took place in the male it would show immediately in one or more daughters. The well-known case of Abraxas grossulariata may be taken to show why mutation taking place in a male is expected to show first in the female and not in the male offspring. The genetic evidence for Abraxas indicates that the female has one sex chromo- some, the male two. The aberrant form lacticolor is found occasion- ally in nature and is always female. A mutation to lacticolor in a Z chromosome of the male would give rise to a daughter if this sperm fertilized a not-Z egg that would at once show the sex-linked character lacticolor. MOSAICS IN PLANTS. The cause of variegation in plants is too involved and obscure1 to attempt to discuss in this connection. On the other hand, the occurrence of bud-sports is generally recognized as due to somatic mutation which may include the germ-tract also. The frequent occur- rence of bud variation in the cultivated forms of the foliage plant Coleus has recently been studied by Stout, who has obtained from a single plant (and its clones) a number of types differing both in color and form of the leaves. The cultivated varieties have arisen through hybridization. Three interpretations suggest themselves as possible. Elimination of the chromosomes of the hybrid might account for the results, but no information as to the chromosomes in the different types is available. If any of the colors are due to cytoplasmic plastids, their irregular distribution might also be responsible for the result. Thirdly, the change might be due to a mutation. If the types studied are complex hybrids with one or more heterogeneous pah's of chromo- somes, a change in one gene of one chromosome might bring about directly a visible change in the color. Until more critical Mendelian work is done it is not possible to reach any plausible or even probable conclusion. It might be possible to analyze the results more closely if we knew what kinds of offspring arise from the original plant and its varieties. Owing to the complex nature of the plants this pro- cedure offers difficulties. A few facts are given by Stout. He states that "plants grown from seed give wide variations .... Many of the types that had appeared as bud variations appeared also in the seed progenies. " Winkler produced mosaics by grafting tomato and nightshade, which are now supposed to be due to a combination of the tissues of the 1 Except in the case of Pelargonium and of Mirabilis, where Baur and Correns have shown that the mosaics are caused, in some instances at least, by plastid assortments. THE ORIGIN OF GYNANDROMORPHS. 33 two plants — the epidermis of one species and a core of the other species. The mosaic shown in Cytisus adami, a hybrid resulting from grafting Cytisus purpureus and Laburnum vulgare, seems also to be due to a similar sort of combination. In animals mosaics have been produced in hydra by King by grafting pieces of a deep-green race on a light one, and by Whitney by destroying the green pigment of one indi- vidual and grafting pieces of it onto a normal green hydra. In tad- poles combinations of different species caused by grafting have been made by Born, Harrison, Morgan, and others. A result strictly comparable to the periclinal chimseras of plants has been reached by grafting a piece of the tail of one species on to the amputated stump of another species. As the new tail grows the skin of the stock is carried out over the core derived from the graft, and as a result an organ is formed with an outer layer of one and a core of another species. The mosaic seeds of corn that are striped with red and white have been shown by Emerson to arise through a mutation in the gene for striping. The "half-and-half" mosaic grains that have been recorded by Correns (1899), Weber (1900), East and Hayes (1911), Emerson (1915), and Collins (1919) have been variously accounted, for — re- calling the different interpretations that have been advocated for gynandromorphs in animals. Emerson (1915) reviews these theories and advances the explanation of somatic mutation. It seems not improbable that elimination will account for those mosaics in which the triploid endosperm nucleus is involved. CLASSIFICATION AND DESCRIPTION OF GYNANDROMORPHS OF DROSOPHILA. The main group includes the gynandromorphs that are adequately explained by chromosomal elimination. It is subdivided according to the type of gynandromorph into: (1) those approximately bilateral, (2) those mainly female, (3) those mainly male, (4) those in which the type is largely "fore and aft," and (5) those in which the mother was known to have been an XXYfemale, but in which simple elimination is sufficient to account for the results. Another group (6) includes those in which the distribution of parts is irregular. These types are only approximations and by no means mutually exclusive; it is often somewhat difficult to decide to which type a specimen belongs. The highly interesting group of special cases (7) is undivided, though it calls for three or four different genetic explanations, based, however, on special modes of distribution of the sex chromosomes. In the Appendix are included those cases in which our records are incomplete as to parentage or in which the specimen has been lost, >so that the description is sketchy. This group contains many of the very early gynandromorphs. To this subdivision is added a brief review of previously published gynandromorphs in Drosophila. 34 THE ORIGIN OF GYNANDROMORPHS. Within each subdivision the arrangement of the cases is according to the order of discovery, that is, by date, except that the colored figures are taken out of order and described first in each group. Each case is known by a number, which is usually that of the culture bottle in which the gynandromorph was found, but in some cases letters or small numbers are used, which, however, correspond to the bottle in which the specimen is preserved or the order in which the descriptions were first arranged. The date, the finder, and the type of illustrations are also indicated on the number line. The information on each case is then given in the order, Parentage, Description, and Explanation. In many of the cases the explanation is followed by a diagram showing at the left the two X chromosomes of the zygote, which at the same time represent the female parts of the gynandromorph, and at the right the single X that is left after elimina- tion, which gives the constitution of the male parts. In case somatic reduction was involved the leftmost set of chromosomes represents the initial condition of the zygote, and the other two sets to the right the resulting two conditions, whether male or female. A knowledge of the order and the relative spacing of the genes along the chromosome is indispensable, and we have therefore made a list of the sex-linked mutants mentioned, with their symbols and the approximate locus of each : Mutant. Symbol. Locus. Mutant. Symbol. Locus. Sable duplication . . . s 0.0 Club Cj 16.7 Lethal 6 u — 0.004 Cut ct 20.0 Yellow y 0.0 Tan t 27.5 Lethal 7 1? 0 3 Vermilion v 33.0 White w Miniature . m 36.1 Eosin we Lethal 9 1, 38.0 Blood w» 1.1 Sable . . s 43.0 Cherry ... wc Garnet g 44.4 Notch N 1 Rugose rfl 45=*= Facet fa > 2.6 Lethal 4 14 49.0 Bifid b« 6.3 Rudimentary r 55.1 Ruby TO 1 Forked f 56.5 Claret r&c \ 7.0 Bar B 57.0 Crimson rb-figure 20 (diagram). Parentage. — The parentage of gynandromorph F is unrecorded, though it is probable that it was found in a wild stock. Description. — The fly was a completely bilateral gynandromorph having on the right side a sex-comb, shorter wing, shorter bristles, and smaller parts in head, thorax, and abdomen. The coloration of the abdomen was male at the tip on the right side, but female in the remainder. The genitalia were entirely female. The abdomen contained a fully developed pair of ovaries and she produced many offspring which were all wild-type. Explanation. — Elimination of one X occurred in a normal female zygote; whether this X was maternal or paternal is indeterminable. 40 THE ORIGIN OF GYNANDROMORPHS. No. 380. March 28, 1916. A. Weinstein. No diagram. Parentage. — The mother carried the genes for ruby and forked in one X and the genes for eosin and sable in the other X. The father was eosin-bar. Description. — The gynandro- morph was about half and half, the left side being mainly male and the right female. The left side of the head and thorax and the left wing were smaller and the left foreleg bore a partly double sex-comb. The left eye was eosin bar of the male type. The genitalia were double pos- teriorly; there was a penis with claspers and anterior to the right of this an ovipositor and female-type anal prominences. The abdomen was female in coloration, except at the tip on the left side, which showed the male banding. The right eye was red and of the broad hetero- zygous bar female type. Explanation. — A ruby forked X egg was fertilized by an eosin bar X sperm. Elimination of the maternal ruby forked X oc- TEXT-FIGURE 21. curred. Mf B We B No. SSO1122AAA7344512 Selection Experiment. January 18, 1917. T. H. Morgan. Plate 4, Figure 1 (diagram). Parentage. — The mother was notch, having therefore one X chromosome with the dominant gene for notch; the other X carried the recessives eosin and ruby. The father was likewise eosin ruby. Description. — The gynandromorph was male on the right side, except for spots of red (female) in the eosin ruby eye of that side. The coloration of the abdomen was male throughout. The genitalia were mainly male, but showed female parts. The left side was mainly female, having a red eye and a notch wing of slight type. No gonads were found in the sections examined, but it is probable that there were very rudimentary ovaries. r* Explanation. — An egg bearing the gene for notch was fertilized by an X sperm with the genes for eosin and for ruby. Elimination of a maternal X chromosome left the male parts to be determined by the paternal eosin ruby X. N we W PLATE 3 \ la 3. Normal 3a. Normal E. M. WALLACE Pinx GYNANDROMORPHS OF DROSOPHILA THE ORIGIN OF GYNANDROMORPHS. 41 No. 29. February 11, 1918. T. H. Morgan. Text-figure 21 (drawing). Parentage. — Both mother and father were eosin. Description. — The gynandromorph was bilateral, except that the entire head was female, having eosin eyes of the dark homozygous eosin color. The left side was male, having a sex-comb, shorter legs, shorter bristles and wing, and a smaller left side to the thorax and abdomen. The coloration of the abdomen was half and half, but there appeared to be a pair of ovaries and female genitalia. Explanation. — An egg containing an X chromosome with the gene for eosin was fertilized by an X sperm carrying eosin. Elimination of either X gave the nearly bilateral gynandromorph. we we GYNANDROMORPHS MAINLY FEMALE. No. C2C19. January 1914. E. M. Wallace. Plate 2, Figure 3 (colored drawing). Parentage. — The mother was white-eosin compound, having the genes for yellow and white in one X and eosin in the other X. The father was eosin. Description. — All of the fly was gray and female, except the upper right half of the thorax and the right wing, which were yellow and male. Both eyes were eosin, of the dark type of the homozygous eosin female. Well- developed ovaries were present on both sides. Mated to a yellow white male this gynandromorph was fertile and produced white-eosin females 70; eosin males 42; yellow white-eosin females 53; yellow eosin males 58. Explanation. — An egg with a cross-over X containing the genes for yellow and eosin was fertilized by an X sperm with a gene for eosin. Elimination of one of the latter left the maternal X to produce the male parts on the upper right side of the thorax. ywe ywe vcl* No. 5137. September 4, 1916. C. B. Bridges. Plate 2, Figures 5 and 5a (colored drawings). Parentage. — The mother had one chromosome with the genes for vermilion, sable, garnet, and forked, and the other X with the genes for vermilion and for bar. The X-bearing sperm carried the genes for eosin and for miniature wings. Description. — The mosaic was entirely female, except for a patch of eosin in the left eye. The eosin part of the eye was round and light eosin (male type) while red bar both above and below (very slight amount below). The right eye was red bar. The whole abdomen was full of eggs. Explanation. — An egg with the X chromosome carrying the genes for vermilion and bar was fertilized by a sperm carrying the genes for eosin and miniature. Elimination of a maternal chromosome took place, leaving the 42 THE ORIGIN OF GYNANDROMORPHS. one X with eosin and miniature genes to produce the male parts, which in this case affected visibly only a part of the left eye. v ' B we m we m No. M, 114. January 23, 1914. C. B. Bridges. Text-figure 22 (diagram). Parentage. — One of the X chromosomes of the mother contained the gene for eosin, the other the gene for bar. The father was white bar. Both parents were heterozygous for the autosomal recessive gene "whiting," which is a specific modifier of eosin. Description. — The gynandromorph was somewhat more than half female. The left side of the gynan- dromorph, except for the head, was male, with sex- comb, smaller bristles half-thorax and wing. In col- oration the abdomen was male on the left and female on the right. The genitalia were entirely female. The head had heterozygous bar (female) eyes, which were white-eosin compound (female) in color. A pair of ovaries was present. Explanation. — An egg containing the X chromo- some with the gene for eosin was fertilized by the X sperm with the genes for white and for bar. Either chromosome may have been the one to suffer elimination; which one it was could not be determined, since the head did not show male parts. No. 438. August 16, 1914. C. B. Bridges. Text- figure 23 (diagram). Parentage. — One X chro- mosome of the mother con- tained the genes for eosin and for vermilion, the other X the gene for white eyes, the male carried the gene for eosin. Description. — The left side of the gynandromorph was largely male with a white eye (containing a fleck of white-eosin), a sex-comb, and a shorter wing. The right side was female, with a white-eosin com- pound eye, no sex-comb, and a longer wing. The abdomen was banded like a female. When bred as a female the fly gave the classes expected for a white-eosin compound. No sections were made. Explanation. — An egg containing the X chromo- some with the gene for white was fertilized by an X sperm carrying the gene for eosin. The latter — the paternal chromosome — suffered elimination, leaving the white-bearing X to produce the male side. The color of the eye on the female side was white-eosin compound, which is the expected result for the two X's involved. TEXT-FIGURE 22. The X chromosome of TEXT-FIGURE 23. W W we THE ORIGIN OF GYNANDROMORPHS. 43 No. 922. December 16, 1914. C. B. Bridges. Text- figure 24 (diagram). Parentage. — One of the X chromosomes of the mother contained the genes for eosin and for vermil- ion, the other X the gene for forked. The X chromo- some of the male carried the genes for white and for bar. Description. — The fly was female throughout (with- out sex-combs) and possessed white-eosin heterozy- gous-bar eyes, except that the tip of the abdomen on the left side was banded like a male. Below there was a normal penis and male armature. In sec- tions an ovary was found on one side, nothing on the other, Explanation. — An egg containing the X chromo- some with the genes for eosin and for vermilion was fertilized by the X-bearing sperm with the genes for white and for bar. Elimination of either X chro- mosome would account for the male parts at the tip of the abdomen. we TEXT-FIGURE 24. or w B w No. 925. December 18, 1914. C. B. Bridges. Text-figure 25 (diagram). Parentage. — The mother was club, carrying in one X the sex-linked gene club, and in the other X lethal 2 which is a deficiency for club. The X sperm of the father carried the genes for eosin and for miniature. Description. — The only male part was the right eye, which was eosin (male type) in color, except for a fleck of red (female). The fly was fertile as a female when mated to a wild male and produced: No. 1117; wild type females, 101; eosin miniature males, 55; miniature male, 1; eosin males, 9. Explanation. — An egg with an X bearing the gene for lethal 2 was fertilized by an X sperm with the genes for eosin and miniature. Elimination took place in one of the maternal chromosomes, leaving the paternal X with eosin and miniature to form the male parts, viz, the right side of the head (in part). The rest of the mosaic was female; hence both wings were wild-type. Cl m we m No. 1010. December 20, 1914. C. B. Bridges. Text-figure 26 (diagram). Parentage. — One X chromosome of the mother carried the genes for yellow and for white, and the other X the gene for lethal 6. The X chromosome of the father carried only wild-type genes. 44 THE ORIGIN OF GYNANDROMORPHS. Description. — The left side of the thorax of the gynandromorph was male, with a sex-comb and a shorter wing. Both eyes were red and female. The abdomen and genitalia were female. The body-color was wild-type through- out. Sections showed ovaries on both sides. Explanation. — An egg containing the lethal 6 X chromosome was fertilized by a wild-type X sperm. Either one of the X chromosomes being eliminated would account for the result. If the paternal X were eliminated the male parts would be lethal 6, and hence it is more probable that the maternal X was eliminated. or No. 1808. July 7, 1915. C. B. Bridges. Text-figure 27 (diagram). Parentage. — The mother was pure for the second-chromosome recessives purple, curved, and speck. The father was heterozygous for the dominant star (eyes). No sex-linked mutant characters were present. TEXT-FIGURE 25. TEXT-FIGURE 26. TEXT-FIGURE 27. TEXT-FIGURE 28. Description. — The gynandromorph was female throughout, except for the abdomen, which had male coloration on the left side and was twisted to the left. A perfect penis was present. The eyes were star. The male parts could not have shown the recessive second-chromosome characters, even had they been present. No testes or ovaries were found, but there was a genital tube with pointed cells like abnormal spermatozoa. THE ORIGIN OF GYNANDROMORPHS. 45 No. 5238. September 23, 1916. C. B. Bridges. Text-figure 28 (diagram). Parentage. — One of the X chromosomes of the mother carried the genes for vermilion eye-color and for bar eye, the other X the gene for forked bristles. The X chromosome of the father carried the genes for eosin, vermilion, and forked. Description. — The fly was mainly female, but is exceptionally interesting from the peculiar description of the male parts, which constitute a very narrow stripe running through the middle of the left eye and along the left side of the thorax, including the wing. The left eye was eosin vermilion in color in the male parts and red in the female parts, both above and below the eosin vermilion. These female parts were heterozygous for bar and the red portions above and below were therefore characteristically narrow, while the eosin-vennilion part was not-bar and projected forward, so that the male stripe could be traced forward to the normal margin of the round eye. The male part of the thorax could likewise be traced by means of the forked bristles, of which there were three anterior to the wing, one above, and none below. The wing itself was included in the male region and was smaller and had forked marginal bristles. There was no sex-comb on the left side. Explanation. — An egg containing an X chromosome with the gene for bar was fertilized by the eosin vermilion forked sperm. A maternal X suffered elimination, leaving the eosin vermilion forked X to produce the male parts. B We V f We V f No. 2. September, 1917. T. H. Morgan. Text-figure 29 (drawing). Parentage. — The fly appeared in "selected notch" stock in which, in each generation, red-eyed notch females were bred to eosin ruby males. Description. — The right eye was red, the left partly red, partly eosin ruby, with a very irregular boundary-line; other- wise the fly was female. Explanation. — An egg with a gene for notch wing was fer- tilized by an X sperm bearing eosin and ruby. Elimination of one of the maternal X's left a part of one side of the head with the eosin ruby X. The wings, although not showing notch, must have contained the gene. Since less than half TEXT-FIGURE 29. of the notch flies in this selected stock showed the notch character, its absence here is not difficult to explain. N We rb We 7*6 No. 477. October 31, 1917. D. E. Lancefield. Text-figure 30 (drawing). Parentage. — One of the X chromosomes of the mother had a bar gene, the other a gene for forked. The father was bar. Description. — The head was small, with round eyes and forked bristles. The thorax and wings seemed to be female. No sex-combs present. The abdomen was entirely female, with eggs inside, but she did not breed. 46 THE ORIGIN OF GYNANDROMORPHS. Explanation. — An egg containing an X with a gene for forked was fertilized by a bar X sperm. The paternal X with bar was eliminated, leaving the head male and forked. TEXT-FIGURE 30. No. 71. October 23, 1917. E. M. Wallace. Text-figure 31 (drawing). Parentage. — Pure stock of bar. Description. — A female was observed that had a short left wing. Closer ex- amination showed that the bristles on that side of the thorax and head were shorter and that the left side of the head was slightly contracted and the eye smaller. It is probable that the left side of head and thorax (dorsally) were male. No. 7530. August 18, 1917. C. B. Bridges. Text-figure 32 (diagram). Parentage. — One of the X chromosomes of the mother carried the gene for facet eye and the other X the gene for notch wings (dominant). The X chromosome of the father carried the gene for facet. Description. — The left side of the gynandromorph was male, with a shorter wing, sex-comb, and smaller eye, whose markedly faceted eye was character- istic for that character as it appears in the male of the mutant type. The female side had a faceted eye of the female type, which is far less marked. The abdomen was banded as in the female, but below a penis was present. Testes were found on both sides, with an abundance of sperm. Explanation. — An X egg-carrying facet was fertilized by the X sperm- carrying facet. Elimination of either occurred. The gonads were formed from a male cell. Very frequently a male-appearing abdomen contains ovaries; only very rarely does a female-type abdomen contain testes. THE ORIGIN OF GYNANDROMORPHS. 47 No. Xi. January, 1914. E. M. Wallace. No diagram. Parentage. — This gynandromorph arose in a mass-culture whose parents were yellow white females and eosin males. Description. — The gynandromorph was largely female. The male parts were yellow and included the left dorsal side of the thorax with the shorter wing and the left side of the abdomen. These parts were all smaller, bore TEXT-FIGURE 31. TEXT-FIGURE 32. smaller bristles, and the left half of the abdomen had male-type coloration. The genitalia were female. The female parts throughout were wild-type in body-color, including especially the left legs and all the head. There were no sex-combs. The eyes were both white-eosin compound. Explanation. — A yellow white X egg was fertilized by an eosin X sperm. Elimination of the paternal X occurred. y w y w 48 THE ORIGIN OF GYNANDROMORPHS. GYNANDROMORPHS MAINLY MALE. No. GiAb2Ca2. March, 1914. E. M. Wallace. Plate 2, figures 6 and 60 (colored drawings). Parentage. — The mother was a yellow white female, a daughter of gynandro- morph G^AbsC. The father was an ebony (third-chromosome) male. This mating was part of the second of the tests specifically designed to show the absence of elimination of autosomes in the production of gynandromorphs. Description. — The gynandromorph was mainly male, with only the head and genitalia female. The color of the entire thorax, abdomen, legs, and wings was yellow, and correspondingly the bristles of these parts were brown. These yellow parts were male, as proved by the sex-combs on both forelegs, by the small (male) size of the bristles, of the thorax, and particularly of the abdomen, and by the male coloration and shape of the abdomen. However, the genitalia were an exception, for the anal prominence and the ovipositor were purely female in structure and bore black spines which showed that the body-color was wild-type. The head was entirely female, as proved by its large size, the wild-type color with black bristles, and by the red eyes. Thus the head and genitalia — the two ends of the fly — were female and all the region between was male. Explanation. — An egg carrying the genes for yellow and white was fertilized by sperm carrying only wild-type genes in the X. Elimination of a paternal X occurred and subsequent shifting isolated a female cell which gave rise to the genitalia. The absence of ebony proves that the third chromosome did not undergo elimination. y w y w No. X2. February 1914. E. M. Wallace. Text-figure 33 (drawing). Parentage.-j-Gynandromorph X2 appeared in a mass-culture, the mothers of which carried yellow and white in oneX and eosin in the other; the fathers were yellow-white. Description. — The gynandromorph was mainly male. The female parts were confined to the abdomen, which had female coloration on the left side and apparently male on the right. The abdomen was twisted to the right, which also suggests that the right side was male. However, the genitalia re- versed this relation, the right side being largely female, with an anal prominence of female type; the left side was male and there was a median penis. The abdomen was of large size and a pair of ovaries could be clearly seen within. The thorax and head were entirely male, as evidenced by their size and the type of bristles and the pres- ence of sex-combs on both forelegs. "^fc/ A B The eyes were both whiteand the body- TEXT-FIGURE 33. color was yellow throughout. Explanation. — A yellow white X egg was fertilized by a yellow white X sperm. Elimination of either X occurred. An alternative explanation is that the egg was fertilized by a Y sperm giving a yellow white male. Somatic non-disjunction resulted in a cell with both daughter X's present, and this gave rise to the female parts. THE ORIGIN OF GYNANDROMORPHS. No. 3. February 1915. T. H. Morgan. Text-figure 34 (drawing) 49 Parentage. — The mother was rudimentary and the father bar. Description. — The gynandromorph was about three-fourths male. The right halves of the head and of the thorax were female, being larger in size, having larger bristles and a larger wing, which was wild- type, and no sex- comb. The right eye was heterozygous bar (female). The left eye was bar of the male type and the left halves of the head and of the thorax were male. The left wing was smaller, but not rudimentary. The abdomen seemed en- tirely male, with a normal penis. This gynandromorph was tested as to sexual behavior and was found to pay no attention to mature virgin females. An account of this gynandromorph and the drawing have been previously published. (Morgan, Am. Nat., V. 49, p. 240, April 1915.) Explanations. — An egg with a rudimentary X was fertilized by an X sperm carrying bar. Elimination of the maternal rudimentary X occurred. Some of the female cells were lost in cleavage, so that the individual is prepon- derantly male. TEXT-FIGURE 34. TEXT-FIGURE 35. No. 2317. November 2, 1915. C. B. Bridges. Text-figure 35 (drawing). Parentage. — One X chromosome of the mother carried the genes for rudimen- tary wing and fused veins, and the other X the gene for bar. The X chromo- some of the father carried the genes for vermilion eye and forked bristles. Description. — The left side of the gynandromorph is male, with sex-comb and rudimentary fused wing, the left side of the abdomen is male, but the genitalia are female. The ocelli on the head are like those of fused, and the head is therefore male. No sections were made. 50 THE ORIGIN OF GYNANDROMORPHS. Explanations. — An egg containing the rudimentary fused X was fertilized by the vermilion forked sperm. A paternal vermilion forked X was elim- inated, leaving the other rudimentary fused X to produce the male side, while the female side contains both original X's, namely, rudimentary fused and vermilion forked, and is accordingly wild-type. No. 3272. February 10, 1916. C. B. Bridges. Text-figures 36 and 36a (drawings). Parentage. — One X chromosome of the mother carried the genes for sable, garnet and also "sable-duplication" at zero. The other X carried the genes for eosin and for miniature. The father was eosin-miniature. Descriptions. — The entire abdomen was apparently male in shape, banding, and genitalia, though it is not known whether testes or ovaries were present. The right side of the thorax was smaller in size and bore smaller bristles and TEXT-FIGURE 36a. TEXT-FIGURE 36. TEXT-FIGURE 37. a smaller (male-type) miniature wing. All right legs were male and both fore- legs bore sex-combs. The right eye (see drawing) had a streak of male tissue ("light" eosin color) running forward completely through. Above and below this streak the tissue was female ("dark" eosin color) There is one other curious feature — the left foreleg as well as the right bore a sex- comb. The head, except the male streak, the right side of the thorax with its miniature wing, and the two rear legs were female Explanations — An egg bearing the eosin miniature non-cross-over X was fertilized by the X sperm carrying eosin and miniature Elimination of one of these X's (either maternal or paternal) was followed by shifting of cleavage nuclei or by shifting of the anlage in the formation of the pupa. THE ORIGIN OF GYNANDROMORPHS. 51 GYNANDROMORPHS ROUGHLY "FORE-AND-AFT." No. II 139. January 12, 1914. C. B. Bridges. Text-figure 37 (diagram). Parentage. — The mother was black (second chromosome), but carried only wild-type genes in her X chromosomes. The father was a bar not-black male. Descriptions. — The fly was heterozygous bar in both eyes and female through- out, except for the external genitalia, which were male (penis), and the coloration of abdomen. Sections showed that a pair of ovaries was present. Explanations. — An egg with a wild-type X was fertilized by the X sperm with the gene for bar. Since the male parts did not involve the eye, it can not be determined whether they arose from cells carrying the bar (paternal) or the wild-type (maternal) X. The fly did not show black in the male parts, but since the male region was so small and also normally dark-colored, this case could not be accepted as proving that the elimination did not affect the autosomes, as is proved in several later cases, especially devised for that purpose. or B B No. 1813. July 5, 1915. C. B. Bridges. Text-figure 38 (diagram). Parentage. — One X chromosome of mother carried the genes for forked and for cleft (wing) ; the other X only wild-type genes. The father was forked. Descriptions. — The head, thorax, wings, and legs were female. The ab- domen had the male coloration and a normal penis. The (poor) sections TEXT-FIGURE 38. TEXT-FIGURE 39. TEXT-FIGURE 40. showed that at least one ovary was present. The wings were not cleft and the male parts showed no forked spines. Explanations. — The egg contained the wild-type X and was fertilized by the X sperm carrying forked. The paternal X was eliminated, leaving the male parts wild-type. t 52 THE ORIGIN OF GYNANDROMORPHS. No. 2204. October 5, 1914. C. B. Bridges. Text-figure 39 (diagram). Parentage. — One X of the mother carried the gene for eosin, the other the genes for vermilion and forked. The father was bar. Description. — The gynandromorph was of the "fore-and-aft" type. The abdomen was of the male shape, with male coloration on the left side and partially male on the right. There was a normal penis. The eyes were heterozygous bar (female) and the head, thorax, legs, and wing were female. Sections showed that email ovaries were present. Explanations. — Since the male parts were not forked, the egg probably carried the eosin X. The X sperm carried bar. The eyes were female and there is no criterion as to which X was eliminated. An alternative explana- tion assumes a vermilion forked X in the egg, and subsequent elimination of this same X to give the not-forked male parts. No. X. August 1916. A. Weinstein. Text-figure 40 (drawings of wings). Parentage. — The mother had the genes for eosin, ruby, and forked in one X and for fused in the other. The father was probably eosin ruby forked. Description. — The gynandromorph was largely male anteriorly and female posteriorly. The head was entirely male, with eosin ruby eyes. There were sex-combs on both forelegs, which means that the ventral part of the thorax was male. The left dorsal part was also male, having a small wing which was fused. The right dorsal part was female with a large wild-type wing. The abdomen and genitalia were female. Explanations. — The egg carried a cross-over eosin ruby fused X and the sperm an eosin ruby forked X. Elimination of the paternal X occurred. we rb fu we rb /„ we rb f No. 4614. January 22, 1918. A. H. Sturtevant. Text-figure 41 (diagram). Parentage. — One X of the mother carried the genes for eosin, vermilion, and forked; the other X carried only wild-type genes. One of the third- chromosomes carried the recessive genes for sepia, spineless, kidney, sooty, and rough ; the other was wild-type. The father was a bar male from stock. Description. — Except for the wings, the gynandromorph is divided antero- posteriorly. The right wing was slightly larger than the left and may have been female. The other wing and the remainder of the thorax was male. There were sex-combs on both forelegs. The head was entirely male, with eosin-vermilion eyes and forked bristles. The thorax and legs had also forked bristles. The abdomen was female, both in banding and in shape. The genitalia were female, but slightly abnormal. Tested as a female she proved sterile. None of the third-chromosome recessives showed in any part, either male or female, of the gynandromorphs. Explanations. — An egg containing the non-cross-over eosin vermilion forked X was fertilized by an X sperm carrying bar. The paternal X was eliminated, producing the anterior male parts. The absence of the recessive third-chromo- some characters in the male parts proves that the elimination of the X was independent of the third chromosome. We V f We V f B THE ORIGIN OF GYNANDROMORPHS. 53 GYNANDROMORPHS PRODUCED BY XXY FEMALES. No. N 2. December 12, 1912. C. B. Bridges. Plate 3, Figures 1 and la (colored drawings). Parentage. — The mother was an XXY female homozygous for white and heterozygous for the third-chromosome mutant pink. The father was red- eyed, and also heterozygous for pink. Both parents were exceptions produced by secondary non-disjunction. Description. — The fly was a completely bilateral gynandromorph, male on left side, female on right. The male side was smaller, with sex-comb, the genitalia half and half. The fly was unable to breed as a male or as a TEXT-FIGURE 41. TEXT-FIGURE 42. TEXT-FIGURE 43. TEXT-FIGURE 44. female. The abdomen was large and evidently contained a pair of ovaries. The fly was figured in Heredity and Sex, page 163, and the origin given in Journ. Exp. Zool., 1913, page 597. Explanations. — A regular X egg carrying the gene for white was fertilized by an X sperm carrying the wild-type allelomorph red. One of the maternal X's, bearing the gene for white eye, was eliminated. The white-eye character therefore does not appear on either side. As both parents were heterozygous for pink, the fly may have come from third chromosomes bearing normal genes only, or one of them may have had the gene for pink, so that! the gynan- dromorph is heterozygous. w No. N 3. November 30, 1912. C. B. Bridges. Plate 3, Figures 2 and 2a (colored drawings). (See fig. 17.) Parentage. — The mother was an XXY female, carrying white in both X chromosomes. The father was a wild male. 54 THE ORIGIN OF GYNANDROMORPHS. Description. — The gynandromorph was entirely female, except for the tip of the abdomen below, where a perfectly normal penis and male genitalia were found. The anal prominence and the parts immediately surrounding the genitalia were also male. The posterior ventral plate was male type, being broad, rounded, and hairless. No. 1221. February 2, 1915. C. B. Bridges. Text-figure 42 (diagram). Parentage. — The mother was an XXY wild-type female, one of whose X chromosomes carried the gene for eosin, the other X only wild-type genes. The father was bar. Description. — The gynandromorph was bilateral, except for the head, which was entirely female, with red bar eyes of the heterozygous type. The right side of the thorax dorsally was male, with shorter bristles and very small wing (abnormal). There were no sex-combs. The right side of the abdomen was male in coloration, and the genitalia were almost entirely male. There was a pair of testes with ripe spermatozoa. The two halves of the thorax failed to come together and the male and female parts were unfused. Explanations. — The egg carried the eosin X and may or may not have con- tained a Y. The sperm was the X sperm carrying bar. Elimination of either X occurred. It is possible that the spina bifida condition may have been a result of the gynandromorphism. No. 1892. July 19, 1915. C. B. Bridges. Text-figure 43 (diagram). Parentage. — The mother was a wild-type XXY female, which was an ex- ception from "high" non-disjunction. One X carried the gene for eosin, the other the genes for vermilion and forked. The father was bar. Description. — The fly was female throughout, except that the left side of the abdomen, especially at the tip, showed male coloration and the genitalia were entirely male. The eyes were heterozygous bar (female). Explanation. — An egg with one X (either) and with or without a Y (even chance) was fertilized by an X sperm carrying the gene for bar. Elimination of either X occurred. No. 7673. October 16, 1917. C. B. Bridges. Plate 4, Figure 3 (drawing). Parentage. — The mother was an eosin-eyed XXY exception from a special strain of "high" non-disjunction (He), which had arisen by equational non- disjunction from the regular high strain. One of her X chromosomes carried the gene for eosin and the other the genes for eosin and forked. The father was bar. Description. — The male parts of the gynandromorph constituted the entire head, which had eosin eyes of the male type and forked bristles; the left side of the thorax, which had forked bristles, was smaller, with a male-size wing and a sex- comb and a slight patch of male tissue at the tip of the ab- domen, but on the right side, not the left. The abdomen was twisted, as it usually is when bilateral, but since the bristles were not forked, the male parts, if any, must have been below or internal. The right wing was abnormal. Explanations. — An X egg carrying the genes for eosin and forked was fertil- ized by the X sperm carrying the gene for bar. Whether or not a Y was present in the egg is not known (chances even). Elimination of a paternal bar X occurred. we f we f B PLATE 4 GYNANDROMORPHS OF DROSOPHILA THE ORIGIN OF GYNANDROMORPHS. 55 No. 5485. October 18, 1916. C. B. Bridges. Text-figure 44 (diagram). Parentage. — The mother was an XXY female, one of whose chromosomes contained the genes for yellow and for white, the other X the gene for lethal 7. The X chromosome of the father carried the genes for yellow, claret, vermilion, and forked. Description. — The gynandromorph was a yellow female, except that three- quarters of the right eye was white in color and male, the remainder, which was a perfect quarter sector of the eye, being red and female. Sections showed normal ovaries to be present. Explanations. — An egg containing the X chromosome with the genes for yellow and for white was fertilized by the X sperm with the genes for yellow and the three other recessive genes named above. Elimination of a paternal chromosome occurred, leaving the yellow white X to determine the character of the male parts, viz, the right eye, except for a triangular area of female tissue. y r* v f y w y w GYNANDROMORPHS OF COMPLEX TYPE. No. 487. November 27, 1917. D. E. Lancefield. Text-figure 45 (drawing). Parentage. — The mother was an XXY female homozygous for eosin and miniature. The father was a wild male. Description. — The distribution of male and female parts was very complex. The entire head was female, as evidenced by its large size and by the color of both eyes, which was eosin of the dark female type. The right dorsal part of the thorax was female, as shown by its large size and the large size of the bristles and of the wing, which was also wild-type and not miniature. The only other female parts seemed to be the left ventral part of the thorax, including left legs, since left foreleg carried no sex-comb. The other two sectors of the thorax — the right ventral and the left dorsal — were male, as proved by the smaller size of the parts themselves and of their bristles, and even better by the presence of a sex-comb upon the right foreleg and of a miniature wing of male size upon the left side. As the head was entirely female, the abdomen seemed to be entirely male, except that the armature seemed slightly different in the two sides of the penis. Explanations. — An egg carrying an eosin miniature X (whether or not a Y also is unknown) was fertilized by the X sperm carrying only wild-type genes. Elimination of a paternal X occurred. The segmentation nuclei descended from this same pair of male and female cells were distributed in a regular but complex pattern. No. 941. December 15, 1914. C. B. Bridges. Text-figure 46 (diagram). Parentage. — The parentage is somewhat uncertain, probably as follows: The mother had one X with eosin, notch, tan, and vermilion, and the other X wild-type. The father was eosin tan vermilion. Description. — The gynandromorph was about half-and-half, but rather complex in the distribution of male and female parts. The head was large, therefore probably female. The eyes were alike and vermilion. The right wing was a typical notch (female) but was only doubtfully larger than the left. The abdomen was female in coloration anteriorly but male posteriorly. 56 THE ORIGIN OF GYNANDROMORPHS. The genitalia were largely male, but had female parts on the right side. A pair of rudimentary ovaries were present. There were sex-combs on both forelegs, so that the ventral side of the thorax was entirely male. The fly was tan throughout. Explanations. — An egg containing a cross-over chromosome with the genes for notch, tan, and vermilion was fertilized by an X sperm carrying eosin, tan, and vermilion. Elimination of the maternal X was followed by shifting of the cleavage nuclei. N t we t t No. 983. December 20, 1914. C. B. Bridges. Text-figure 47 (diagram). Parentage. — One of the X chromosomes of the mother carried the genes for white and for bar, and the other X the gene for eosin. The father was miniature. Description. — The separation of the sex-characters is very complex. The dorsal parts of the thorax and the wings are, from their size, female; the lower TEXT-FIGURE 45. TEXT-FIGURE 46. TEXT-IIGURE 47. part of the thorax, from the presence of sex-combs on both forelegs, is male. The abdomen is female on the left half and male on the right. The genitalia are female. The abdomen contained a pair of ovaries as seen through the body-wall and in sections. The fly was sterile. The head was entirely male, with white eyes, not-bar. THE ORIGIN OF GYNANDROMORPHS. 57 Explanations. — A cross-over X carrying the gene for white but not for bar was present in the egg, which was fertilized by the X sperm carrying the gene for miniature. Elimination of this paternal X left a cell with the white X to determine the male parts. In the early cleavage there must have been extensive shifting of the nuclei to produce the observed mosaic of female and male parts. w w m No. 16240521114. Selection Experiment. August 2, 1916. T. H. Morgan. Plate 4, Figure 2 (drawing). Parentage. — The gynandromorph arose in a "selected" notch stock in which the female carried notch in one X and eosin and ruby in the other. The father was eosin ruby. Description. — The gynandromorph was "quartered," being male in the anterior left section and also in the posterior right section, and female in the two other sections. The left eye was mainly eosin ruby, but had a small section of red (female) pushed in from the rear. The left side of the thorax was male, as evidenced by the sex-comb and the shorter wing. The right side of the abdomen had male coloration above and below, and the genitalia were male on the right side and female on the left. The abdomen seemed to have a pair of ovaries when examined, but the sections made later were too poor to confirm this. The right eye was red and the right wing notch. Explanations. — An egg carrying the gene for notch was fertilized by a sperm carrying the genes for eosin and ruby. Elimination of the maternal notch X occurred at the first division, leaving the eosin ruby paternal X to determine the character of the male parts. The products of the second division rear- ranged themselves so that sister cells took part in the development of opposite sides of the body. This is only a little more extreme than the usual rear- rangement and shifting of parts (see patch of red in left eye). N We Tb We Tb SPECIAL CASES. The following cases were brought together because they could not be explained simply by the theory of elimination. Analysis showed that in each of these cases there were present two different chromo- somes, both derived from the mother. Non-disjunction obviously offered an explanation for this fact. But the application of this hypothesis required the additional assumption of "somatic reduction" to explain the gynandromorphism. This means that at an early division the two X's derived from the mother separate without division. On the other hand, if we assume for these cases that both sex chromosomes leave a daughter half at the mid-plate (double elimination) the assump- tion just stated is avoided. Until further explanation is obtained these two interpretations may be given as alternatives. 58 THE ORIGIN OF GYNANDROMORPHS. Doncaster's observations on binucleated eggs of Abraxas, where both nuclei underwent separate reduction and fertilization, offer a simpler explanation. On the other hand, it should be pointed out that there should have been at least as many autosomal mosaics as sex-linked mosaics produced by fertilization of binucleated eggs of heterozygous mothers; and this does not seem to be the case. No. B. 90. June 17, 1912. C. B. Bridges. Text-figure 48 (drawing). Parentage. — This gynandromorph appeared in F2 from the cross of rudimentary female to white miniature male; that is, the mother (Fj female) carried rudimentary in one X and white and miniature in the other; and the father was a rudimentary (F^ male. Description. — The individual seemed to be male through- out. Both eyes were red. Sex-combs were present on both forelegs. The right wing was long, and though slightly deformed, was undoubtedly wild-type. The left wing was a typical and perfect miniature rudimentary wing. The abdomen was entirely male, and when mated to a vermilion female the fly bred as a male, producing abundant offspring. Several pairs of the wild-type daugh- ters and vermilion sons of this mating were bred and all produced red and vermilion in equal numbers, both in males and females. That is, the gynandromorph bred as a wild- type male carrying no mutant genes. Two of the F2 pairs are given as samples: TEXT-FIGURE 48. Wild-type 9 Wild-type d" Vermilion 9 Vermilion cf B. 98.1 B. 98.2 45 26 33 16 32 30 43 33 The drawing has been previously figured in Zeit. f. ind. Abst. und Verer., 1912, p. 324. Explanations. — Simple elimination fails to explain this case, because the characters of the fly, as well as its genetic behavior, show that it received two different X chromosomes from its mother. For instance, miniature and rudimentary were both present in the left (male) wing, which proves that the X contained in these parts came from the mother and that crossing-over in the mother must have occurred. Since the right wing was wild- type, its cells must have contained a wild-type X, which likewise could only have come from the mother. The FX and F2 offspring of the gynandromorph showed that he had such a wild-type X in the testis, which presumably came from the same kind of cells as those of the right side. The offspring also show that the gynandromorph had not received an X sperm from the father, which would have given rudimentary offspring. Therefore the right side, at least, must have come from a Y-bearing sperm, as further proved by the fact that the gynandromorph was fertile as a male (males without a Y being sterile). On the view that the gynandromorph came from an egg with two nuclei, a simple explanation of the result may be given. Before reduction, each of the postulated nuclei must have had one white miniature X and one red rudimentary X chromosome; after crossing-over and reduction in each, one THE ORIGIN OF GYNANDROMORPHS. 59 nucleus contained a white miniature rudimentary cross-over X, and the other nucleus a wild-type cross-over X. Each nucleus was fertilized by a Y-type sperm, proof of which for the right side has been given; proof for left side is as follows: The left wing is miniature as well as rudimentary, and since TEXT-FIGURE 49. TEXT-FIGURE 50. TEXT-FIGURE 51. the X of the father did not carry miniature, this left side could not have contained a paternal X and must therefore have contained a paternal Y chromosome. One of the cross-over chromosomes was white as well as miniature and rudimentary; but since the eye on the side with miniature was red, we may suppose that all of the head came, as is very often the case, from cells from one side only, namely, the right, which was here carrying red; or this cross- over chromosome may have come from double crossing-over, and in this case it would have carried red. Left side. Right side. W m m or -x -Y On an alternative view that both of these X's were in a single nucleus, the following assumption seems necessary. An XX egg was produced by reduc- tional primary non-disjunction (see Bridges, 1916), preceded by crossing-over, so that one X contained white miniature rudimentary and the other was the complementary X containing only wild-type genes. This XX egg was then fertilized by a Y sperm. That the individual was entirely male with no female parts can be explained either by double elimination or somatic reduction at the first division of the zygote ; that is, one member of each pair was caught by the elimination plate, 60 THE ORIGIN OF GYNANDROMORPHS. so that each of the two first daughter cells had but one X and these different from each other. Zygole. Left side. Right side, w m r w m r No. I 92. August 16, 1913. C. B. Bridges. Text-figure 49 (diagram). Parentage. — One of the X chromosomes of the mother carried the genes for vermilion and for fused and the other X the gene for bar. The father was vermilion fused. Description. — The gynandromorph was completely bilateral, except for the genitalia, which were female. The left side was male, as evidenced by the smaller size throughout, by the sex-comb, and by male coloration on the abdomen. The left eye was bar of the male type. The right side was female in every part, and was chiefly remarkable in that its large wing was fused. The eyes were both red, not vermilion. The right eye was round, not hetero- zygous bar. A pair of ovaries was found in the sections. Explanations. — On the assumption of two nuclei in the egg, one nucleus after reduction contained a non-cross-over bar X chromosome, and this nucleus fertilized by a Y sperm gave the bar male left side, with bar eye; the other nucleus after crossing-over and reduction contained a cross-over fused X chromosome, wh ch nucleus fertilized by the vermilion fused X sperm gave the female right side with fused wing: Left side. Right side. B fu On the alternative view that both X's from the mother were retained after reduction in the same nucleus of the egg, the case is difficult, but may be accounted for in the following way: Since the left side is male throughout and shows the bar eye-character (of male type), this side must have come from a non-cross-over X of the mother. But this bar X is not represented at all on the right side, as proved by the round eye, which, although female, is not even heterozygous for bar. That the right side is female requires that two X's be present, and the fact that the wing is fused requires that both carry the fused gene. A non-cross-over vermilion fused X must have come from the mother along with the bar X. The egg, then, was an XX egg pro- duced by primary non-disjunction which was equational, since the bar X was a non-cross-over and the fused X a cross-over chromosome (Bridges, 1916). This XX egg was fertilized by an X sperm carrying the genes for vermilion and for fused. It is known that XXX zygotes are unable to hatch as adult flies (Bridges, 1916), but since neither the time nor the mechan- ism of their elimination is known, it is possible that if double elimination or somatic reduction followed soon after fertilization the life of the XXX individual would be saved, but at the price of becoming a gynandromorph. Two of the X's, in this case the paternal vermilion fused and the maternal THE ORIGIN OF GYNANDROMORPHS. 61 fused cross-over X, remained in one cleavage cell which gave rise to the not- vermilion not-bar fused female right side. The other X, the maternal non- cross-over bar X, passed into the other daughter cell and gave rise to the not-vermilion bar not-fused left side. Zygote. V Left side. Right side. V B -X .x B No. 937. December 17, 1914. C. B. Brdges. Text-figure 50 (diagram). Parentage. — The grandmother was a wild-type XXY female carrying the genes for eosin and vermilion in one X and in the other only wild-type genes; the grandfather was white bar. By equational non-disjunction an XXY eosin daughter was produced which carried eosin and vermilion in one X and eosin in the other. This female was out-crossed to a vermilion male and produced among the sons a mosaic. Description. — The mosaic, as in the case B 90, was male throughout, but the left eye was eosin (of the male type) and the right eye was eosin vermilion. The male was fertile when bred to a vermilion female, giving wild-type daughters and vermilion sons (No. 1116). One of the wild-type daughters out-crossed to a forked male gave eosin and vermilion as the main classes of the sons. Explanations. — On the hypothesis of a binucleated egg, one nucleus after reduction contained an eosin vermilion X and the other nucleus an eosin X. Since no eye-color corresponded to the X sperm of the father, and since the individual was male throughout, both of the egg-nuclei must have been fert Sized by a Y sperm, which is further shown by the fertility of the male. Left side. Right side. W< On the view that a single nucleus was present, the following situation de- velops: Since the right eye showed both eosin and vermilion, the mosaic must have contained the eosin vermilion X of the mother. Since the other eye showed eosin (not vermilion) , this X must have been the other or eosin X of the mother. That is, both X chromosomes of the mosaic came from the mother by means of an XX egg produced through non-disjunction. The ver- milion X of the father was not present at all, as proved by the fact that the left eye of the mosaic was eosin (not red) and male (not female) , and by the breeding-test, which showed that the gonads carried only the eosin X. The sperm was not the X sperm of the father, but the Y sperm, as further indicated by the fertility of the male. As in case B 90, there must have been double elimination or somatic re- duction, so that one cleavage-cell received the eosin X and a Y, and the other 62 THE ORIGIN OF GYNANDROMORPHS. the eosin vermilion X and a Y. The gonads developed from an eosin cell as shown by the FI and F2 results of his breeding test. Zygote. Left i>ide. Right side. We V W' V w* we No. 1333. February 19, 1915. C. B. Bridges. Text-figure 51 (diagram). Parentage. — The mother was a wild-type XXY female, carrying the genes for eosin in one X and for vermilion and forked in the other. The father was bar. Description. — The fly was female throughout, except for the left eye which was round (not bar) and red (not eosin or vermilion). The eye has been examined repeatedly at different times since the mosaic was on hand, and the eye is undoubtedly, not-bar and is of the right size for a male. The right eye was heterozygous bar. There were no forked bristles present around the left eye elsewhere. The female was mated to a sable forked male and pro- duced: No. 1555 — forked females, 11; forked bar females, 0; bar females, 18; wild-type female, 1; vermilion forked males, 18; bar males, 6; vermilion bar males, 3; forked males, 2. Explanations. — On the hypothesis of a binucleated egg, one nucleus after reduction contained a cross-over wild-type X and the other a non-cross-over vermilion forked X chromosome. The former fertilized by a Y sperm gave rise to the wild-type (male) left eye; the latter fertilized by a bar X sperm gave rise to the rest of the fly. Left side. Right side. V f B The following alternative possibilities may be considered: The simplest possible explanation is that this is a mosaic or somatic mutation — that the bar gene in the cell that gave rise to the left eye reverted to not-bar, or to an allelomorph which gives a small round eye. If, as is more probable, this mosaic is a gynandromorph arising by chromosomal disturbance, the ex- planation is like that for No. I 92, i. e., the egg arose by equational non-dis- junction and contained a non-cross-over vermilion forked X and a cross-over wild-type X. This egg probably did not contain a Y, as evidenced by the lack of exceptions among the sons of the mosaic, and as is possible in accordance with the assumption of equational non-disjunction, for equational non-dis- junction, even when occurring in a female with a Y, is probably always primary. One eye was clearly heterozygous bar; hence it is known that the XX egg was fertilized by an X sperm carrying the gene for bar. This XXX zygote would ultimately die, unless at an early stage the XXX condition was cor- rected by reduction or elimination. Double elimination or somatic reduction in a cleavage-cell would save the individual, but turn it into a gynandromorph. The other X chromosome, wild-type, passed into the sister cell and gave rise THE ORIGIN OF GYNANDROMORPHS. 63 to male parts, which, because of the lateness of the occurrence, or from shift- ing of nuclei, constituted but a small part of the gynandromorph. Zygole. V Left side. Right side. V f -x .x B B No. 2349. November 3, 1915. C. B. Bridges. Text-figure 52 (drawing). Parentage. — The mother was from a strain of high non-disjunction, but was known to be XX and not XXY. One X carried the genes for vermilion and forked, the other X the gene for bar. The father was a vermilion forked male. Description. — The gynandromorph was largely male. The female parts included the left legs, which were without a sex-comb and had forked bristles. The female parts throughout had forked bristles and could therefore be readily traced. All three left legs were forked and female to the mid-ventral line. A very narrow strip of female tissue ran diagonally forward from above TEXT-FIGUHE 52. TEXT-FIGURE 53. the middle left leg to the shoulder, being chiefly marked by one large forked bristle and several smaller ones. Most of the left side of the head bore forked bristles, including the left antenna, the dorsal region to the left of the line in the diagram, a small zone of tissue around the eye to the rear, and the region below the eye including the oral bristles. The left eye was red (not vermilion) and round (not bar or heterozygous bar — the small nick seen in the drawing of the eye seems to be an artifact) . The abdomen was male type, the genitalia 64 THE ORIGIN OF GYNANDROMORPHS. were half and half, the left half bearing a female type anal prominence with forked bristles. Sections showed that ovaries were present with well-devel- oped eggs, which account for the large size anteriorly of male-type abdomen. The male parts, as distinguished by the normal bristles, included the whole dorsal surface of the thorax, the two wings, which were of equal size (male), the abdomen (except for half the genitalia), all the right legs, and somewhat more than the right side of the head. The right foreleg bore a sex-comb. The right eye was bar (male type), not vermilion in color. Explanations. — On the theory that two nuclei were present in the egg, one nucleus contained after reduction a cross-over forked X, the other nucleus a bar X chromosome. The former fertilized by a vermilion forked X sperm gave rise to the female parts on the left side, the latter fertilized by a Y sperm gave rise to the left male side, with the bar eye, etc. Left side. Right side. f B v f On the assumption of a single nucleus in the egg a possible explanation is as follows: The male parts show the character bar, and since bar was present only in the mother, they are known to have been derived from the maternal bar X, which was a non-cross-over X, since the eye did not show the character vermilion. The female parts were forked, but since the eye was not vermilion, one of the forked X's must have been a cross-over between vermilion and forked. Crossing-over takes place only in the female and not in the male, wherefore this X also is known to have come from the mother. One cross- over and one non-cross-over X is the general rule for eggs produced by primary equational non-disjunction. The other forked X must have come from the father and therefore carried the gene for vermilion; but vermilion is recessive and its effect is hidden by the normal allelomorph in the cross-over X from the mother. The gynandromorph, as in cases 192 and 1333, started as a XXX zygote which was saved from death and at the same time converted into a mosaic by double elimination or somatic reduction at the first cleavage division. Zygote. Left side. Right side. B f B v f No. 4241. May 15, 1916. C. B. Bridges. Text-figure 53 (diagram). Parentage. — One of the X chromosomes of the mother carried the genes for lethal 7 and ruby eye-color, the other X the genes for yellow, eosin, and forked. The X chromosome of the father carried the genes for yellow, eosin, and forked. Description. — The right side was male throughout, except that in the head the female part (bordered by the dashed line in the diagram) extended nearly to but did not include the right eye. The right eye was smaller and ruby. THE ORIGIN OF GYNANDROMORPHS. 65 The two anterior bristles above the right eye and all the bristles below it were forked, agreeing with the forked bristles present throughout the rest of the right side on legs, thorax, and abdomen. A sex-comb was present on the right fore-leg and all male parts were smaller. The right side was female throughout, with normal bristles and a red eye. The genitalia were half- and-half. In the sections of the abdomen an ovary could be identified on one side, less certainly on the other. The body-color of both male and female parts was wild-type throughout, with no yellow. Explanations. — On the theory of the binucleated egg, one nucleus contained an X with the genes for lethal and ruby, the other nucleus a cross-over X with the genes for lethal, ruby, and forked. The former fertilized by X sperm with yellow eosin forked, produced a female left side with only wild-type characters; the latter, fertilized by a Y sperm, gave the ruby forked male side. Left side. Right side. y w< f I, rb f i rb An alternative view based on a single nucleus is as follows: Simple elimina- tion fails to explain the case, since the male parts that are forked are not- yellow and not-eosin, as might be expected, but instead were ruby. Since ruby was present only in the mother, the male parts must have come from a lethal 7 ruby forked cross-over chromosome produced by the mother. That the other X chromosome of the zygote was not the yellow eosin forked X of the father is proved by the not-forked character of the female parts. It seems certain that both X chromosomes of the zygote came from the mother, that is, that the egg was a non-disjunctional XX egg. This must have been by primary non-disjunction, since the pedigree is fully known and no other exceptions were produced. Another fact points to the same conclusion, namely, that these X's were both cross-overs, and, as Bridges has shown, both X's of secondary exceptions are always non-cross-overs. What occurred, then, y we f was crossing-over in the — '—. — ! — : '• female between the li rb loci ruby and forked. Owing, perhaps, to some entanglement in the process of crossing-over, the chromosomes were unable to separate in time for the reduction division and both were retained in the egg. This egg, containing a yellow eosin X and a lethal 7 ruby forked X, was fertilized by a Y sperm. At the first segmentation divisions one of these maternal yellow eosin chromosomes was eliminated, giving a gynandromorph whose male parts were lethal 7, ruby, and forked. Left side. Right side. li rb f IT rb f y we f A very interesting point in connection with gynandromorph 4241 is the fact that a male part, which must be assumed to have the lethal 7 gene, was able to live when associated with the not-lethal partner in the gynandromorph. This is, however, understandable when the nature of the action of the lethal 7 66 THE ORIGIN OF GYNANDROMORPHS. gene is considered. Normally, the males which possess the lethal 7 gene not only begin development, but continue often to the full-grown larva stage. The immediate cause of their death at this late stage is the development of one or more black granules which cause death either because they are them- selves toxic or because their substance is derived from an excessive malforma- tion of organs essential to the further development of the fly. On the first view, either the half amount of toxic body was insufficient to prevent the gynandromorph from continuing its development, or the corresponding normal parts of the female side counteracted this toxic effect; on the second view, the lack of the essential organ on the male side was supplied by the normal organ of the female side. No. 3674. August 9, 1917. A. H. Sturtevant. Text-figure 54 (diagram). Parentage. — One of the X chromosomes of the mother carried the genes for cut, vermilion, and for forked; the other X the gene for rugose. The father was rugose forked. Description. — The male parts of the gynandro- morph constituted the left side of the thorax and abdomen, as indicated by the smaller size, the sex- comb, the smaller wing, and the male coloration of that side of the abdomen. Testes were found in the abdomen. The left wing showed the character cut. The bristles of all male parts were wild-type. The bristles of the female parts were forked, and this character forms the most useful index of the divi- sion-line. The entire head, including the bristles above and below and around the left as well as the right eye, was 'forked. The first and second leg on the right side were forked, the third was not forked and presumably therefore male. The bristles on the right wing and on the right side of the abdo- men were forked. There was no sex-comb on the right side. The right wing was of female size and not cut. Both eyes were red (not vermilion) and also not rugose. Explanations. — On the theory of two nuclei, one nucleus contained a cross-over X with the gene for cut and rugose, the other an X with the genes for cut vermilion forked. The former nucleus was fertilized by a Y sperm to produce the left side, the latter nucleus by an X sperm with rugose and forked to produce the female right side. Left side. Right side. TEXT-FIGURE 54. ct rg ct f rg f An alternative explanation on the assumption of a single nucleus follows: The mechanism in this case must be essentially the same as in cases I 92, 1333, and 2349 already given. The female parts were entirely forked, there- fore two forked chromosomes were present. One of these could have come from the father, whose X was rugose forked; the other X could have come from THE ORIGIN OF GYNANDROMORPHS. 67 the mother, and could have been the npn-cross-over cut vermilion forked X. Neither cut nor vermilion would show in the female parts, since they would be recessive to their wild-type allelomorphs in the rugose forked X; and like- wise rugose would not show, for it would be recessive to its wild-type allelo- morph in the cut vermilion forked X. Both of these X's could not have come from the father, for in that case both eyes would have been rugose. One of the forked X's therefore came from the mother. The left wing was cut, and since cut was present only in the mother, this X also must have come from the mother. Since the cut side did not show forked, this cut X must have been a cross-over anywhere between cut and forked. Thus we see that the egg contained two X's which were different, one being the non-cross-over cut vermilion forked X and the other the cross-over cut X, which is the normal condition of XX eggs produced by primary equational non-disjunction. This XX egg was fertilized by the rugose forked X sperm of the father, giving an XXX zygote. At the first segmentation division, double elimination or somatic reduction occurred, thereby enabling the fly to survive, but only at the price of becoming a gynandromorph. The paternal (rugose forked) and one of the maternal X's (cut vermilion forked) entered one cell, from which developed the female right side, which showed only one mutant char- acter, namely, forked. The cross-over maternal X (cut vermilion? rugose? not-forked) entered the other cell and gave rise to the male left side, showing the mutant character cut only. It should be noticed that in all four of these cases it has been the paternal X and one of the maternal X's that have come together into the female part, and that the male part was in each case maternal. This suggests that the essential feature of the reduction is the active separation of the two X's which ab- normally came from the same individual and the passive inclusion of the paternal X in the same cell with either separated maternal X. Zygote. Left side. Right side. Ct f Ct f Ct rg Ct .x _x rg f rg f No. 7730. October 24, 1917. C. B. Bridges. Text-figure 55 (drawing). Parentage. — The mother was a wild- type regular XX female (from a strain of high non-disjunction) carrying the genes for eosin and forked in one X and only wild-type genes in the other. The father was bar. No exceptions were produced other than the fol- lowing gynandromorph. Description. — The gynandromorph was almost entirely male. All parts, except the head, were male and had forked bristles. The head was mainly female, having straight bristles and a red (not-bar) eye on the right side, and on the left side a division-line which ran forward through the eye. Above this line, which was perfectly clean and TEXT-FIGURE 55. 68 THE ORIGIN OF GYNANDROMORPHS. sharp, the eye was red and the bristles wild-type; below the line the eye- color was eosin (male type) and the bristles were clearly forked. It is pos- sible that the not-eosin, not-forked part, described above as female, was really male, in which case the fly would be a male mosaic. The fly, bred as a male to an eosin-crimson female, produced 143 wild-type daughters and 151 eosin crimson sons. Two pairs of these were inbred and produced: No. Eosin 9 Eosin crimson 9 Eosin crimson of Eosin forked cf Eosin c? Eosin crimson forked of 8056 68 83 41 44 26 32 8057 75 84 45 46 24 25 Explanations. — On the assumption that two nuclei were in the egg and that the fly was entirely male, one nucleus contained an eosin forked X and the other nucleus a wild-type X; each nucleus having been fertilized by a Y sperm, the former gave rise to the eosin forked male parts and the latter to wild-type male parts. In case the red parts were female the corresponding nuclei must have come from an XX egg produced by primary non-disjunction, and likewise fertilized by a Y sperm. Left side. We f If male. Right side. we If female. On the view that only one nucleus was present in the egg, the possible explanation is as follows: On the assumption that the wild-type parts were female, both of the X's present must have come from the mother, since the eye was not-bar, as would have been the case if the X of the father were present. Moreover, that the sperm was the Y sperm is known from the fact that the mother had no Y to contribute and yet the fly was fertile. The egg was there- fore XX, one X being eosin forked and the other wild-type, and was pro- duced by direct primary non-disjunction. Elimination of the wild-type X occurred and the male cell gave rise to most of the body, including the gonads. Left side. Right side. we f On the assumption that the wild-type parts were male, the zygote must have had the same origin as above, but double elimination (or somatic re- duction) occurred, so that one cell received a single eosin forked X and the other a single wild-type X. Left side. Right side. ive THE ORIGIN OF GYNANDROMORPHS. No. da, 62, c. April 1914. E.M.Wallace. Plate 3, figure 4 (colored drawing). Parentage. — The mother was a white-eosin compound female, carrying the genes for yellow and white in one X and eosin in the other. The father was an ebony male, used to show the lack of elimination of the third chromosome. Description. — The mosaic was entirely female. The right side of the thorax, the right wing, and the right legs were yellow in color, while the rest of the female, including all of the head and abdomen, was gray. Both wings were of the same size and there was no size inequality in bristles or other parts. There was no sex-comb on the yellow right foreleg. Explanations. — On the view that this gynandromorph arose from a bi- nuclealed egg, it must be assumed that one of these nuclei must have contained two yellow white bearing X's that arose through equational non-disjunction; the other nucleus contained (as the offspring showed) a yellow white chromo- some. The former nucleus fertilized by a Y sperm gave the yellow parts of the fly (not including right side of head, which is gray red) ; the latter nucleus fertilized by wild-type X sperm (from the ebony male) gave the left side of the fly, including all of head and abdomen. Left side. Right side. W y w If the mosaic had arisen from a yellow white egg fertilized by the X sperm of the ebony male (whose X chromosome carried only wild-type genes) it would have been easy to explain the case as simple elimination were it not that the yellow parts were unmistakably female, which is impossible without the additional hypothesis of a succeeding somatic non-disjunction. It was next supposed that the mechanism of the production of the mosaic had been double somatic non-disjunction, that the two daughter wild-type X's had gone into the same cell, giving rise to the wild-type left-side female parts, and that the two daughter yellow white X's had both been included in the other cell, which gave rise to the yellow female parts on the right side. On this hypothe- sis the offspring (disregarding ebony) should correspond to those of a pure yellow white female or of a pure wild female. In fact, however, the offspring correspond to those of the original zygote when the mosaic was mated to a yellow white brother: yellow white females, 106; yellow white males, 103; wild-type females, 117; wild-type males, 107; yellow males, 2; white male, 1. A possible escape from this dilemma is to suppose that the non-disjunction took place after the first division and that the normal cell was the one which gave rise to the germ-cells. This mosaic would then be triregional — the ab- domen and gonads heterozygous for yellow and white and representing the original zygote, the right side of the thorax pure yellow white, the left side of the thorax and the head pure wild-type. Another type of explanation is that in the normal XX zygote somatic muta- tion to yellow occurred in the wild-type chromosome, so that the yellow part contains a mutant yellow X and the maternal yellow white X. Or somatic deficiency for the yellow locus occurred in the wild-type X, so that the yellow parts are haploid for yellow, and like the normally haploid male show yellow, while these parts are female because the sex gene is situated in some other part of the X than the yellow-deficient region. 70 THE ORIGIN OF GYNANDROMORPHS. GYNANDROMORPHS WITH INCOMPLETE DATA. A fly was figured by Morgan (Zeit. f. i. Abst. u. Ver. vn 1912, fig. 3) with one long wing and one miniature wing (text-fig. 56). Its history has been lost, but it is recorded in a paper giving crosses that involve miniature wings. The fly was probably a gynandromorph. No. 28. February 11, 1918. T. H. Morgan. Text-figure 57 (drawing). Parentage. — Uncertain; probably the gynandromorph appeared in a stock of "serrate" extracted from a cross of dichaete (carrying serrate) to short notch. TEXT-FIGURE 56. TEXT-FIGURE 57. Description. — The gynandromorph was largely female, the entire head and the right side of the thorax with the right wing and legs being male. The sex-combs seemed to be only half as large as that of a normal male. No. M. February 1912. E. M. Wallace. Text-figure 58 (diagram). Parentage. — The ancestry is unknown. Description. — The gynandromorph was largely female; the male parts being the right dorsal half of the thorax with its wing, which were yellow in color and of smaller size. No. H. July 1913. Text-figure 59 (diagram). Parentage. — Ancestry unknown. Description. — The fly was yellow and female to all appearances, except tip of abdomen, which was male. A penis was present. Sections showed one abnormal testis and one broken one. THE ORIGIN OF GYNANDROMORPHS. 71 No. G. May 1914. T. H. Morgan. Text-figure 60 (diagram). The ancestry is unknown. The gynandromorph was mainly female. The fly was gray with red eyes. There were no sex-combs and the wings were equal in length. The tip of the abdomen had male banding. A pair of ovaries were seen through the body-wall and eggs were found in section. No. N. September 1916. T. H. Morgan. Text-figure 61 (diagram). Parentage. — The ancestry is uncertain; probably the gynandromorph came from the notch stock. If so, the mother carried notch in one X and eosin in the other, and the father was eosin. Description. — The gynandromorph was largely male. The entire abdomen, the right half of thorax, with wing and legs, were male. The division TEXT-FIGUBE 58. TEXT-FIGUHE 59. TEXT-FIGURE 60. between male and female in the head ran through the right eye, which was light eosin (male) below and dark eosin (female) on the dorsal half. No. O. January 1912. Text-figure 62. The ancestry not recorded. It had one red eye (right) and one white eye with a red fleck in it (left). On the left side there was a sex-comb. The wings were equal in length and apparently female. The abdomen and genitalia were entirely female. Poorly developed eggs were seen in section. No. P. December 1913. Text-figure 63. The ancestry of this gynandromorph is not recorded. It has a sex-comb and short wing on the right side. The abdomen was mostly female, but showed some male parts in the genitalia. 72 THE ORIGIN OF GYNANDROMORPHS. No. X. July 15, 1916. Text-figure 64. Parentage. — Ancestry unknown. Description— The head was small and apparently therefore male. The eyes were eosin ruby. Sex-combs were present on both sides. The abdomen was female. This is apparently an antero-posterior mosaic. TEXT-FIGURE 61. TEXT-FIGURE 62. TEXT-FIGURE 63. No. 110. December 12, 1915. A. Weinstein. No diagram. The mother had the genes for eosin, ruby, and forked in one X chromosome, and the genes for fused in the other. The father was bar. The eyes of the gynandromorph were bar (homozygous or heterozygous?); the wings were abnormal; the abdomen was female, with female genitalia somewhat abnormal. DROSOPHILA GYNANDROMORPHS PREVIOUSLY PUBLISHED. There are several references to cases where white spots were found in the eyes of Drosophila, sometimes in cases where the gene for white eyes might have been present (Amer. Naturalist, XLVXII, Aug. 1913, p. 509; Morgan, Science, xxxm, Apr. 1911, p. 534). No. I. J. S. Dexter, 1912. Biol. Bull. August 1912. Parentage. — The mother carried the gene for yellow and white in one X and only wild-type genes in the other. Description. — Although the individual is described as a female, it is more likely that the yellow white right side was male and the wild-type left side female. This female was found to be sterile, which agrees better with the assumption that the right side was male, since mosaics which are entirely or even more than half female usually are fertile. Explanations. — A yellow white X egg was fertilized by a wild-type X sperm. Elimination of the paternal X occurred. THE ORIGIN OF GYNANDROMORPHS. 73 In the early stages of non-disjunction, C. B. Bridges (Journ. Exp. Zool, Nov. 1913) found several gynandromorphs, two of which (N2 and (Ns) have been figured in Heredity and Sex (p. 163) and refigured here. Breeding-tests were tried on all these and it was shown that some were fertile, and further that the gynandromorphs were not due to an inherited condition. It was pointed out (p. 600) that such mosaic forms can be explained as due to somatic non-disjunction and also even to XXX zygotes. Again, in the experiments on "Dilution Effects on Certain Eye Colors" (Morgan and Bridges, J. E. Z., Nov. 1913), about a dozen gynandromorphs were recorded, most of which were sterile, but those which bred (as females) behaved genetically as did their regular sisters; that is, they showed no trace in their gonads of the effect of the bodily division. This was especially strik- ing in one case where the head was entirely white, yet in which the offspring showed eosin (pp. 44, 51). No. I. F. N. Duncan. (See Am. Nat., vol. 49, p. 455, 1915.) Parentage. — The father had white eyes, the mother was wild-type. Description. — The fly had on one side a red eye, long wing, no sex-comb, and female abdomen. On the other side white eye, short wing, sex-comb, and male abdomen. Courted by males but would not court. Two testes with ripe sperm. Explanations. — Elimination of a maternal X chromosome explains the results. No. II. F. N. Duncan. Plate 3, figure 5 (colored drawing). Parentage. — The male grandparent was cherry club vermilion, the female wild-type. The mother was heterozygous for the above genes. The father was wild-type. Description. — The fly had a cherry left eye and red right eye. Sex-comb on left foreleg only. Right wing shorter than left. Abdomen largely female, more female left, more male right. Contained two testes with immature sperm. Explanations. — An egg containing a cross-over cherry X was fertilized by an X sperm. Elimination of a paternal chromosome followed by an irregular distribution of the nuclei with one sex chromosome explains the results. No. III. F. N. Duncan. Parentage. — Same origin as No. II. Description. — The fly had red eyes and sex-combs, left wing longer, abdomen male. Genitalia half male, half female. Was courted but would not mate. Two ovaries with ripe eggs. Explanations. — Elimination of either a maternal or a paternal X chromosome will explain the result. No. IV. F. N. Duncan. Parentage. — Same origin as No. II. Description. — Both eyes red, no sex-combs, wings same length. Abdomen and genitalia male on one side, female on other. Was courted. Two testes with mature sperm. Explanations. — It is not possible to determine which X chromosome was eliminated. No. I. 1915. Hyde and Powell. Genetics, 1, 1916, p. 580 (colored diagram). Parentage. — The mother was pure for blood, an allelomorph of white. The father was eosin, another allelomorph of white. 74 THE ORIGIN OF GYNANDROMORPHS. Description. — The mosaic was female except for the head, which was entirely male. The left eye was eosin (male type) and the right blood (i. e., it was not eosin-blood compound). Explanations. — A blood X egg was fertilized by an eosin X sperm. If at some cell division in the future head region of the very early embryo somatic reduction occurred, that is, if the eosin X went into one cell and the blood into the other, neither dividing, both cells would produce male parts with the eosin and blood type eyes. The result may, however, be explained in another way, viz, both chromosomes divided, but in an early cell division double elimination occurred. One daughter chromosome from each X was caught by the elimination plate, and the remaining X's were left, one in each cell. No. II. 1916. Hyde and Powell. Parentage. — The mother had white eyes and wild-type wings; the father had red eyes and truncate wings (second chromosome). Description. — The gynandromorph had a white eye and a truncate wing on the left side and a red eye and wild-type wing on the right side. The fly was female in other parts and when mated to a white-eyed brother produced : red females, 75; white females, 70; red males, 65; white males, 65. Explanations. — An egg containing a white-bearing X was fertilized by a red X sperm. Elimination of a maternal X left the male parts with the white X. The appearance and disappearance of truncate are so erratic that in this case no safe conclusion can be drawn from the appearance in only the male side. One might suppose that the male and female sides, differing in their X chromosomes, also differ in a sex-linked modifier for truncate. GYNANDROMORPHS AND MOSAICS IN BEES.1 The domesticated bees have furnished many cases of gynandro- morphs, both in hives supposedly pure and in hybrid communities. An excellent review of the recorded cases is given in Miss Mehling's paper of 1915. The earliest description is said to be that of Lau- bender in 1801. Lefebure in 1835, Donhoff in 1861, Smith in 1862, and Menzel in 1862 described gynandromorph bees. Widespread interest in the subject was aroused by the discovery of many gynandro- morphs in the stock of an apiarist, Herr Eugster, in Constance. Menzel first reported on this occurrence. It was, however, von Sie- bold's account of the Eugster gynandromorphs (1868) that brought the subject to the general attention of zoologists. He gave not only a description of many of these bees, but dissected them also, and de- termined the correspondence or lack of correspondence between the internal sexual organs and the external sex characters. In this hive there was a queen of the yellow Italian race o£ bees (Apis ligustica) fertilized by a drone of the darker German race (Apis mellifica). Her sons were Italian, which is the expectation for this combination. After the death of the queen, another queen, of "dark color" was present in the stock. She also produced some gynandromorphs. 178 species of Hymenoptera in which gynandroiaorphs have been described are listed by Enderlein, including Tenthredinidse, Braconidse, Proctotrupidse, Ichneumonidse, Formicidse, Mutil- lidae, Crabonidse, Scoliidse, Pompilidae, Vespidse, and Aphidse (11 families in all). THE ORIGIN OF GYNANDROMORPHS. 75 Several other descriptions of gynandromorphs in bees have been published (see Mehling, p. 174, and Dalla Torre and Friese, 1898). There are certain facts in connection with sex determination in the bee that are almost unique and give an unusual interest to the situa- tion. The queen has the double set of chromosomes which is reduced to the single or haploid number in the ripe egg, after two polar bodies are extruded. If the egg is fertilized it gives rise to a female (queen or worker), but if the egg is unfertilized it produces a male (drone). The male has only the single set of chromosomes. One set of chromosomes, then, produces a male, two a female; but whether sex is the result of special genes carried by one or two sex chromosomes has not been determined. Corresponding with the single (haploid) number of chromosomes in the male, the spermato- genesis shows certain special features. The preparation for the first division takes place, and only a small non-nucleated piece of proto- plasm is pinched off at one pole of the cytoplasmic spindle. Prepara- tion for a second division follows and the chromosomes separate into two groups, but the cytoplasmic division is very unequal and only one of the nucleated cells that results becomes a functional spermato- zoon. That at the second division an equational division of the chromosomes occurs is probable, for in the closely related wasps the second division takes place normally (according to Mark and Cope- land) and two spermatozoa are formed, each with the single number of chromosomes. Since in the male the haploid number of chromo- somes must be supposed to be present, it might have been anticipated that his nuclei would be half the size of those in the corresponding parts of the female, as happens in the sea-urchin egg when haploid and diploid nuclei occur in different regions of the same embryo. An examination of this relation by Miss M. Oehninger has shown, how- ever, that no such difference is present; hence what might have been a means of determining the constitution of the male and female parts of the gynandromorph is lacking. Von Siebold found the male and female characters combined in many different ways in his gynandromorph bees, much as we find them in Drosophila. In some cases one side was male, the other female; or the anterior end might be like that of one sex and the posterior like that of the other; sometimes different regions of the same organ, such as an eye, leg, or antenna, might contain both male and female regions. The normal worker has a sting, the male is without this organ. In the gynandromorphs the sting was present if the abdomen was like that of the worker, but absent when the abdomen was like that of the male. No definite relation was found between the super- ficial characters of the abdomen and its contained gonad. Testis and ovary might even be combined into one organ. Externally the male genital apparatus might be present and ovaries and oviducts exist inside. 76 THE ORIGIN OF GYNANDROMORPHS. Von Siebold attempted to account for the gynandromorphs by the assumption that an insufficient number of sperm entered the egg, so that part of it lacked sufficient quantities of the male element. It is true that recent discoveries (Nachtsheim) have shown that more than one spermatozoon does usually enter the egg, but we can not explain the results on this basis in the sense intended by von Siebold. Von Siebold gave no clear account of the varietal character of the male and female parts of the gynandromorph (see Boveri, p. 286), and in consequence it is not possible from his account to determine whether the male parts were like those of the Italian or German parent. It remained for Boveri, after 47 years, to attempt to make out, from the alcoholic remains of some of von Siebold's bees, the character of these parts. In order to better present here the conditions in the gynandro- morph, copies of Mehling's figures of the head of the normal bees are reproduced in text-figure 64, a, of the drone, b of the worker, and c TEXT-FIGURE 64. of one of the gynandromorphs of von Siebold's bees. The com- pound eyes of the drone are enormous in comparison with those of the worker, and meet above at the top of the head. The three simple eyes are, hi the drone, forward near the middle of the "face", but on the top of the head in the worker. In the latter there is a tuft of long hair on the top of the head. In the gynandromorph copied hi figure 64 the same approximate differences in the size of the compound eyes is seen on the male and on the female side. Two of the simple eyes on the drone side are low down, while the third, on the worker side, is at the top of the head, where a small tuft of hair is also present. The face of the worker is darker than that of the male, and the same difference in color is seen in the gynandromorph. The antennae are larger in the drone, and this difference, too, is manifest in the gynandromorph, as is also the difference hi size of the jaws. Many other differences THE ORIGIN OF GYNANDROMORPHS. 77 as striking as these are found in other parts of the body and come out equally well in the gynandromorph. Miss Mehling shows that the male and female parts may sometimes, however, be so intimately combined that a particular organ, such as a leg, may seem, on super- ficial examination, to be a blend of the two. A minute examination shows, however, as a rule, that such an organ is a piecework or mosaic of male and female characters. It will be recalled that Boveri's hypothesis appealed to the phenom- enon of partial fertilization. A belated sperm, sometimes failing to fuse with the egg nucleus before the latter divides, comes to combine with only one half of the latter. As a result, one of the first two segmentation nuclei contains only the maternal daughter nucleus, the other the combined maternal daughter nucleus and the entire sperm nucleus. The application of the results to the gynandromorph bees is obvious. If these are due to partial fertilization, then we should expect the male side to be like that of the mother's race — the Italian bee — because its haploid chromosome group came directly from the Italian mother's egg. Vice versa, the female side should show hybrid characters, or the Italian character if the Italian race dominates completely the German race. If the latter, both sides would then be alike and racially indistinguishable. Morgan's sug- gestion of polyspermy leads to the following explanation: If under unusual circumstances one (or more) of the spermatozoa should develop, the parts supplied by its nuclei would be haploid, hence male, while the other parts resulting from the combined nuclei would be female. The expected characters of the two parts of the gynandro- morph would be the reverse of those called for by Boveri's hypothesis, for the male parts should be paternal on Morgan's view, maternal on Boveri's. The decision lies, therefore, in the character of the male parts of the gynandromorph. Boveri examined von Siebold's bees, some of which had been preserved in Munich, in order to get an answer to this problem, and reached the conclusion that the male parts are maternal. Hence the answer was in favor of his own hypo- thesis. We may now proceed to examine this evidence in detail and then see whether the hypothesis of chromosome elimination may not fit the facts as well as either of the alternative views. After nearly 50 years in alcohol the Eugster gynandromorphs had lost so much of their color that a comparison with the racial pigmenta- tion as seen in the living bees was impossible. Only after extracting the superficial pigment and dissolving away the external parts of fresh individuals of the two parental races was it possible for Boveri to make any reliable comparisons. Even then only a few individuals were available, because "an vielen Exemplaren das Abdomen nahezu farblos ist." Boveri confines his account to four specimens and in these takes only the head and abdomen into account. 78 THE ORIGIN OF GYNANDROMORPHS. The difference in the coloration of the heads of the males of mellifica and ligustica is in the prepared skeleton very slight, but Boveri thinks that the male parts of the gynandromorphs' heads are colored more like the same parts of the ligustica. Their abdomens show more strik- ing differences, not only in the relative amount of deeper pigment, but in the pigment pattern as well. A comparison of the distribution of the pigment of the male side of the gynandromorph with the sides of the males of the two races seemed to him to show again that the closer match is with the ligustica type of pigmentation. The deep pigmenta- tion of the ventral surfaces of the two races, especially in the males, offers more positive differences, especially as to color-pattern. The comparison shows here that when the abdomen of the gynandromorph is male its deeper color is more like the ligustica type, except when in places the male parts include or are replaced by female areas. Despite the fact that the comparisons that Boveri gives rest on rather a slender foundation, the evidence, so far as it goes, is clearly in favor of his interpretation of the nature of the male parts of the gynandromorphs. The well-known accuracy and carefulness of Boveri's work prejudices one strongly in favor of his opinion. Boveri's evidence would seem, then, to settle the case in his favor were it not that another account appeared just before the publication of Boveri's paper (which he cites at length), based on observation of living1 material — an account that leads to exactly the opposite con- clusion from that reached by Boveri. Engelhardt described (1914) some gynandromorphs in which he stated that the male parts are dark brown (paternal) and the female reddish yellow (maternal). These gynandromorphs came from an Italian mother and a father belonging to a local (einheimischen) race. The case is parallel to the Eugster bees, and the only room left for doubt is the nature of the local race. The local race of the northern Caucasus, whence the evidence comes, is probably, according to Boveri, Apis mellifica ramipes. Here there is some more recent evidence that is important. Quinn has shown (1916) that when a yellow Italian queen is crossed to a gray drone of the Caucasus race the daughters (hybrids) and the drones are yellow like the Italian. This result indicates that the material used by von Engelhardt was suitable for giving differences in the gynandromorphs that could be used to distinguish the character of the male parts. It follows that von Engelhardt's results support Morgan's and not Boveri's hypothesis. Since these views deal with paternal or maternal nuclei as wholes, it is immaterial whether the factor differences are carried by the sex chromosomes or by some other chromosomes, but when the third view comes up for consideration the question of which chromosome pair is involved is of vital moment. Let us see, then, how the hypothesis 1 Or at least not alcoholic. THE ORIGIN OF GYNANDROMORPHS. 79 of chromosomal elimination applies to von Siebold's and von Engel- hardt's gynandromorphs. In the first place, it is important to understand that there is no conclusive evidence that the racial difference here involved has anything to do with one particular chromosome, or even with the sex chromosome. In bees the mother transmits her characters directly to her sons, as is the case in sex-linked inheritance of the Drosophila type, but in bees this form of inheritance is obviously due to the fact that the male develops directly from the unfertilized egg, hence must inherit all the maternal characters, whether in the sex chromosome or in the autosomes. The special sex differences are, of course, due to whatever it is that makes the egg a male or a female. In cases where the queen is heterozygous she may produce two kinds of sons, which is expected if the two races differ in one Mendelian gene,1 but this would hold whether this pair of genes is autosomal or in a pair of sex chromosomes. If the two races differ in more than one pair of genes, more than two kinds of males are expected. The clearest evidence that we have in regard to what a hybrid queen produces is furnished by the recent work of Newell (1914) and Quinn (1916). Newell crossed a yellow Italian queen bee to a gray Carniolan drone. The daughters were yellow like the Italian, showing the dominance of that color. In the reciprocal cross, Carniolan female by Italian drone, the daughters were also yellow, but not as completely so as in the last. Whether this is due to modifying factors of some kind is not known. Quinn, as stated above, used Italian and Caucasus races, crossing both ways, and in both the daughters were the same, viz, yellow like the dominant Italian race. He also found that the FI daughters gave two kinds of drones and two only, which indicates that the factor difference is carried by a pair of chromosomes, but this evi- dence alone does not show that the pair is the pair of sex chromosomes, for any other pair would give in the bee the same result. However, when taken in connection with the gynandromorph results, the evidence becomes somewhat stronger that sex chromosomes are involved. What, then, is the expectation on the elimination view? It is at once apparent that the elimination must involve a sex chromosome, for, otherwise, there is no reason to suppose that an autosomal dif- ference would at the same time be associated with a difference in sex. In other words, the elimination hypothesis can apply here only if the chromosome that determines sex is the same chromosome that causes this racial difference. Elimination of one of the sex chromosomes that carries the factor for mellifica would produce a cell containing only the ligustica-be&Tmg chromosome, and all parts descending from that cell would be both 1 It has been pointed out that the exceptions recorded by Cu6not may be due to drones coming from hybrid workers. (Morgan, 1909o, Am. Nat., XLIII.) 80 THE ORIGIN OF GYNANDROMORPHS. male and ligustica. This is the result which Boveri thinks is shown by the Eugster gynandromorphs. Conversely, if the ligustica chromo- some were lost, all parts containing the descendants of this nucleus would be male and mellifica. This is the result that von Engelhardt claims to have found in his gynandromorphs. Thus both results are expected on the hypothesis of chromosomal elimination; each is equally possible. Boveri's hypothesis of partial fertilization explains only one case; Morgan's former view of sperm-nuclear development will explain only the other; the hypothesis of elimination will explain both, and for this reason is at present to be preferred. Moreover, since it is demonstrably the way in which gynandromorphs are produced in Droso- phila, this hypothesis is more general than either of the earlier views. There is a further implication in these cases of hybrid gynandro- morphs in bees that can now be cleared up. The female parts of the gynandromorph are of hybrid origin. On any view, therefore, these parts are expected to be not necessarily like the mother (unless her character is the dominant one), but hybrid. If the mellifica color is dominant, then on Boveri's views the female side of the gynandro- morph should be mellifica, but according to Newell the Italian yellow color is dominant, hence in half the Eugster gynandromorphs the male and female sides should have the same color. Perhaps this accounts for the astonishing failure on the part of von Siebold to mention the color differences in his gynandromorphs, since the superficial (the racial differences in color) color was often the same on the two sides. If this is the real situation, Boveri must have worked with a deeper color difference, one that is ordinarily not apparent. It is doubtful from his description whether he could determine if the female parts were mellifica or Italian or intermediate. He recognizes the difficulty, for he refrains from making any comparison between the female parts and those of the hybrid workers, but so far as he suggests any com- parison it is with the pure mellifica type. In von Engelhardt's case the male parts are described as darker, hence more like mellifica, while the female parts are described as lighter. Since Quinn shows that the yellow (lighter) color is dominant, the two sides should be different, hence the fact strongly supports von Engelhardt's interpretation. In fact, I do not see how we can avoid the conclusion that von Engelhardt's results are supported by much better evidence than are Boveri's own, if any such comparison must be made. Both are probably right, and the theory of chromo- somal elimination not only accounts for both, but on that theory both kinds of results are expected. If, as here suggested, both the Eugster and the von Engelhardt gynandromorphs are due to chromosomal elimination, it follows that there must have been also other gynandromorphs present that were not color-hybrids, but show the dominant color both in male and in female regions. In fact, these must have been as common as the THE ORIGIN OF GYNANDROMORPHS. 81 hybrid types. How can we account for this absence of all reference to such cases? It is to be recalled that Boveri actually studied only a few cases, stating that others were not sufficiently well preserved to show the hybrid differences between the parts. It should also not be overlooked that the more striking differences in color in the living hybrid bees would draw attention to these, while gynandromorphs colored alike on both sides would be overlooked. A census of all the gynandromorphs occurring under such conditions is necessary before it will be safe to conclude that these reciprocal cases did not occur. Boveri was, of course, only concerned with such cases as showed the maternal character of the male parts, and as such are expected in half of the cases, it would be natural to select these as illustrations of his theory. Until another survey of the entire output in such cases is recorded this test of the correctness of the elimination hypothesis can not be applied. Wheeler (1910) has described a beautiful case of gynandromorphism in a mutillid wasp. The male half of the body is black and winged like the male, while the female half is rich red and wingless. The ants are closely related to the bees, and sex determination appears to be in general the same, although there are some cases, apparently well authenticated, where unfertilized eggs have produced queens and workers as well as males. There were, prior to 1903, 17 cases of gynandromorphs known in ants which were brought together by Wheeler, to which he added, in 1914, 6 new cases. These show the same relations of parts seen in bees and call for no further comment. None were hybrids and furnish, therefore, no evidence for causal analysis. GYNANDROMORPHS IN LEPIDOPTERA. The group of Lepidoptera, including butterflies and moths, has furnished more gynandromorphs than any other group of animals, even more than the single species Drosophila melanogaster, if all butter- flies and moths are taken together. It has been estimated that at least 1,000 cases of gynandromorphs have been recorded for this group.1 Whether they are actually more frequent than in other insects (1906), summing up Schultz's reviews of 1898-1899, states that the 909 gynandro- morphs (and hermaphrodites) brought together by the latter fall within the following species: Species. Indi- viduals. Species. Indi- viduals. Smerinthus populi 67 Lycaena icarus 28 Saturnia pavonia 51 Bombyx quercus 24 Rhodocera rhamni 40 Ocneria dispar 23 Rhodocera cleopatra 34 Bupalus piniarius 16 Anthocharis cardamines . . . Argynnis paphia 33 33 Lasiocampa f asciatella . . . Limenitis populi 15 13 Lasiocampa pini 29 82 THE ORIGIN OF GYNANDROMORPHS. wz qr whether, owing to the striking character of their wings, they have more often attracted attention, is perhaps open to question. The dif- ferences between the coloration of the males and females in some species would at once arrest attention. On the 'other hand, in certain species and in certain hybrid combinations the number of gynandro- morphs is so great that there can be little doubt that their occurrence here is directly related to the specific or to the hybrid nature of the insects. Eleven more gynandromorphs of Argynnis paphia added by Wenke brings the total to nearly 1,000. In regard to the chromosomal background, the situation is the converse of that in Drosophila and in nearly all other insects. The male has two sex chromosomes (text-fig. 65), which may we call ZZ, and the female one, Z, and another called W, corresponding to the Y of Drosophila. The genetic evidence in the case of Abraxas makes this view highly probable, and Seiler has shown in another moth that there is, in fact, such a chro- mosomal difference between the female and the male. As has been stated, in Dro- sophila the female combina- tion XX is the basis for most of the gynandromorphs be- cause the combination al- lows, through the elimina- tion of one of the X's, the formation of parts with one TEXT-FIGURE 65. X which is male. By anal- ogy we should expect in Lepidoptera that the male combination ZZ would furnish the basis for the gynandromorphs of this group, since through elimination of one Z the female condition would arise. The most interesting case in the Lepidoptera is that of a hybrid gynandromorph in the silkworm moth, because here we know the genetic relation of the factors involved. Toyama obtained two bilateral gynandromorph caterpillars whose mother belonged to a race with a striped "zebra" pattern in the caterpillars and whose father belonged to a race with unicolorous white larvae. Experiments show that in general zebra pattern is dominant to white. Neither is sex-linked. The left female side of the gynandromorph caterpillar was zebra, the right side white. If we attempt to analyze this case on the basis of Boveri's or of Morgan's earlier views — views based on the assumption that one or two nuclei determine male and female respectively — and assuming that, as in the bees, the male parts have one nucleus and the female parts the combined nuclei, then the result confirms Morgan's view and not Boveri's. But this interpretation does not wz THE ORIGIN OF GYNANDROMORPHS. 83 get to the bottom of the situation in the light of more recent work, for in moths it now seems probable that one Z sex chromosome (the equivalent in part of one nucleus) makes a female and two a male. There are, then, two kinds of ripe eggs, one with, the other without, a Z, and one kind of sperm, which is Z-bearing. There are six possi- bilities to be considered (see diagrams, text-figs. 66, 67, 68). (1) On Boveri's view (text-fig. 66, 1), if an egg with a W was the kind fertilized, then one half of the maternal segmentation nucleus should have no Z and would probably not develop, while the other cr TEXT-FIGURE 66. half of the egg nucleus, that united with the sperm nucleus, should have one Z and be both female and white. This explanation fails to account for the male sex of the side supposed to be without a Z and for the presence of zebra on that side. (la) On Boveri's view (text-fig. 66, ./a), if an egg with a Z were fertilized by a sperm (bearing Z),then both the male and female sides should be zebra, which is contrary to evidence. 2a TEXT-FIGURE 67. (2) On Morgan's earlier view (text-fig. 67, 2), an egg with a W fertilized by a sperm (bearing Z) should give female parts from the combined nuclei which would be white. The sperm nucleus alone would also give female parts which would be plain. The result is a mosaic, but not a gynandromorph. (2a) On Morgan's view (text-fig. 67, 2a), if an egg with a Z had been fertilized by a Z sperm, all male parts (ZZ) should be zebra. The female parts would be plain, which is again contrary to fact. 84 THE ORIGIN OF GYNANDROMORPHS. (3) On the elimination hypothesis (text-fig. 68, 5), an egg with a W fertilized by a sperm (Z-bearing) should produce a female (ZW). This gives no chance to produce a male side (ZZ) by ordinary elimina- tion. If by somatic non-disjunction (ZZ or no Z) it is not evident that the no-Z part would develop, and if it did, why it should be plain. (3a) If an egg with a Z had been fertilized by a Z sperm (text-fig. 68, Sa), a male (ZZ) would result from which, by elimination of a Z chromosome bearing the white gene, would produce female parts that are zebra and male parts that are also zebra, which is contrary to the actual conditions in the gynandromorph. If the other chromosome should be eliminated, viz, the one bearing the zebra gene, then the male part would be zebra and the female part would be plain. It is evident that this case can not be explained in any of these ways, even though it be assumed that the color-factors are carried by the sex chromosomes. And if we do treat the color-factors as sex- TBXT-FIGURE 68. linked, then they can not be the same zebra-white pair of factors described by Toyama in other crosses which are clearly not sex- linked. To apply the above view tactily takes for granted that the zebra-white pair is not the same pair referred to in other crosses. If it is not, then we are not obliged to assume that zebra is dominant to plain. If plain is assumed to be dominant over zebra, the gynandromorph can be accounted for by Boveri's hypothesis or by elimination. Possibly one might try to find an excuse for such an evasion by pointing out that Toyama states that the two gynandromorphs appeared in a cross between a striped French race with yellow cocoons and a common Japanese race with white cocoons, and that this is not the same cross as that which he described in the body of his paper, where he states that the striped race had white cocoons. On the other hand, both Coutagne and Kellogg, according to Tanaka, have found that striped is dominant to plain, and although I can not find that they have made exactly the same cross as that which yielded the gynandromorph, nevertheless the cumulative evidence is strongly in favor of the view that zebra is both dominant and not sex-linked. It is clear, then, that THE ORIGIN OF GYNANDROMORPHS. 85 we must search for some other kind of explanation for Toyama's gynandromorphs. Fortunately, Doncaster's observation on the eggs of a race of Abraxas gives us a clue to an explanation. Doncaster, as stated on page 20, found occasionally an egg containing two nuclei, each nucleus being about to be fertilized by a separate spermatozoon. Now, if in Toyama's case the zebra mother was heterozygous, one of the two nuclei in question might contain a Z chromosome and an auto- some with a gene for plain color (Z and white), while the other nucleus might contain a W chromosome and an autosomal gene for zebra (W and zebra). Two sperms of the father, each with a white-bearing autosome, each fertilizing one egg nucleus, would give a white male side (Z, white; Z, white) and a female zebra side (W, white; Z, zebra). This seems the most probable interpretation. There is still another possible explanation of Toyama's gynandro- morphs, viz, that the male parts have come from the fusion of nuclei derived from two (or more) spermatozoa. Pairs of such nuclei would give ZZ cells that would be male and paternal. It is true that Herlandt and Brachet find in the frog that sperm nuclei do not fuse in the egg, but they attribute this to the cy tasters that keep them apart. If in the moth (and bee?) the cystasters are less well developed, con- tiguous nuclei might sometimes fuse. Another moth, Abraxas, has been extensively used by Raynor and Doncaster in genetic experiments. The characters in question (gros- sulariataveTSus lacticolor) show sex-linked inheritance and should furnish interesting evidence as to the nature of gynandromorphs in moths. Quite recently Doncaster has reported two gynandromorphs of Abraxas that arose in a cross between these two types. The first case arose in a cross between grossulariata female by lacticolor male. The normal expectation for this cross is: grossulariata males and lacticolor females. There were produced 24 lacticolor females, no grossulariata males, and one gynandromorph that was lacticolor but mixed in certain parts. The absence of males is apparently con- nected with an exceptional chromosomal condition in this family (viz, 55 chromosomal line) of such a sort that all the fertilized eggs lacked a chromosome, the single Z passing out into the polar bodies in all or nearly all cases. The main characters of this gynandromorph are "the right antenna is male, the left female, and the frenulum of the left wing is of the male type and well developed, that of the right male but imperfect. In the external genitalia the chief points are that the uncus, anus, and ovipositor are each divided; the right vulva is not unlike that of a normal male, the left side is abnormal and has attached to it a second anus and half of the ovipositor," etc. Don- caster sums up the chief peculiarities of this moth as follows: " (1) That though predominantly male, it has the lacticolor character which, from its parentage, should be confined to females; (2) throughout the body 86 THE ORIGIN OF GYNANDROMORPHS. \ the right side is male, the left imperfectly developed, a tendency towards the female type The internal genital organs were, as far as is known, imperfectly developed male organs." A theoretical explanation of the case, based on the chromosomal peculiarities of the line, is as follows : Since practically all eggs had but one Z chromosome before polar-body extrusion and lost it at their formation, few males arise as Doncaster has shown, and even if Z should exceptionally remain in a ripe egg it would carry the gene for grossulariata; hence any male coming from it would be grossulariata. Only then by the sperm bringing two Z's into a Z-less egg could a lacticolor male arise. Such an abnormal sperm could arise in any male by primary non-disjunction, or by secondary non-disjunction from a ZZW male, i. e., by the two Z's of the spermatogonia both passing to the same pole at one of the maturation divisions. If this happens, a lacticolor male is expected. The appearance of femaleness in certain parts of the left side must, then, be referred to an elimination of one of the Z's at some early division. Doncaster's second case can be explained as a simple case of chromo- somal elimination. A grossulariata female, by lacticolor male, gave 11 grossulariata males + 11 lacticolor females + 1 gynandromorph whose anterior parts are male (including the wings to some extent), and whose posterior parts are female. Here the normal proportion of males to females, and the expected distribution of color to them, shows that the female was normal as to her chromosomes. If we assume, then, that a Z-bearing egg was fertilized by a normal Z-bearing sperm, the result should be a normal grossulariata female heterozygous for lacticolor. Elimination of one of the paternal Z's would give a grossulariata male in the anterior region coming from the ZZ nuclei and a grossulariata female posterior part coming from the single Z nucleus. The second case is comparable in every way with the cases of Drosophila and allows an extension of the theory of chromosomal elimination to the group of moths, in line with the other critical cases described above. Doncaster's first case must also appeal in part to the same hypothesis, but it is more complicated, since another exceptional phenomenon must have first occurred. This first process gives a lacticolor male when a grossulariata male or no males at all are expected. It is only that elimination later happened to take place in this individual that it comes to be considered in this connection. In other words, there is no necessary connection between the two events, so that the non-disjunction phenomenon does not in reality complicate the elimination explanation. The two are quite inde- pendent. It should be pointed out that such exceptional males due to non-disjunction are known to occur in Abraxas. Another gynandromorph in Abraxas (Tutt, 1897) involves varieties A. ab. suffusa and A. ab. obscura. Since the genetic relation of these THE ORIGIN OF GYNANDROMORPHS. 87 characters to the type grossulariata are not known, nor the parentage of the individual, no analysis of the case is possible. A third aberrant type, nigra, has given a striking bilateral gynandromorph with gros- sulariata (figured by Cockayne). The genetic evidence in regard to this type obtained by Punnett fails to show that the character is a simple Mendelian one, so that this evidence is not available for analysis. The most remarkable mosaics of male and female characters are shown by hybrids of the gipsy moth, Porthetria dispar and japonica. These mosaics have been described by several observers (Wiskott. 1897, Brake, 1907-1910; Brake and C. Frings, 1911 ; Goldschmidt, 1912- 1917; Poppelbaum, 1914). We owe to Goldschmidt not only a most complete account of the hybrids between these two varieties, but of hybrids involving several Japanese local varieties of this moth. In the latter crosses a most astonishing series of mosaics come to light, not as sporadic occurrences, but as regular phenomena of the cross. In his earlier work Goldschmidt called these mosaic forms gynandromorphs, but his later work shows, he thinks, that they are different from gynandromorphs; he now calls them intersex forms. The normal males and females of the gipsy moth differ not only in the characteristic sex differences of this group, but in other secondary sexual differences also. The Japanese varieties show these same sexual differences, though both sexes differ in color and in a few minor points from the European species. Japonica female by dispar male gives equal numbers of daughters and sons that are normal as to sex, but the reciprocal cross, dispar female by japonica, gives normal males and intersex females in equal numbers. These intersexual females from different crosses show a wide range in structure, in color, and in behavior, from almost normal females at one end of the series to forms that externally are about like the normal male. Not only are the wings colored like those of the normal male (with occasional flecks of white like the female), but the antennse, the hair, the size, the genitalia, and the gonads themselves are mosaics of male and female and intermediate conditions also. These relations are more interesting where crosses involving different Japanese races are compared. When a race, Jap. G male is crossed to Jap. K female, all FI daughters are slightly intersexual. When a race, Jap. H female is crossed to Jap G male, the daughters are somewhat more like the males, but the instincts are still female and they attract males. The copulatory organs are so changed in the direction of the male that mating is unsuccessful, and eggs can not be laid, although the char- acteristic hairy sponges are made. When a race, Eur. F female is mated to Jap. G male, the daughters are "more than half-way between males and females." The secondary sexual characters are almost male. The instincts and behavior are about intermediate between those of the two normal races. Males are scarcely attracted 88 THE ORIGIN OF GYNANDROMORPHS. or not at all, and no mating occurs. The copulatory organs show the strangest combinations of the male and female type, but there are still typical but rudimentary ovaries left. When the race, Jap. X female is crossed to Eur. F male, a still higher degree of intersexuality appears. Externally the daughters are "almost indistinguishable from true males." The instincts are entirely male and the moths try unsuccessfully to mate with females. The gonads look like testes, but in sections show a mixture of ovarian and testicular tissue. A step further and the daughters would be transferred into males. The next cross gives this final stage. When Jap. 0 male is crossed to any race of European female, only males are produced, i. e., all the daughters become sons. The reverse picture is given by those combinations in which the intersexes are sons partly changed over into daughters, a condition that Goldschmidt terms male intersexuality. The wings are generally streaked and in the extremest type only a few brown spots appear on the wing-veins. The testis may contain some ovarial tissue, but the changes in the gonads do not appear to run parallel to those seen on the surface. The explanation that Goldschmidt offers for these intersexes is entirely different from the explanation that is demonstrated for the gynandromorphs of Drosophila. He accepts in part the chromosome theory of sex determination and applies it to the present case on the basis that the female is heterozygous for the sex chromosome Mm, and the male homozygous MM. In addition, however, Goldschmidt adds another set of sex-determining factors that he calls FF (inclosing them in brackets), which he locates in the cytoplasm, that is, outside the chromosomal mechanism. These factors do not segregate (the desirability of two F's is therefore not apparent) and are transmitted from the female to her sons and daughters alike. The FF factors stand for femaleness, which apparently includes the eggs, ovaries, secondary sexual characters, and genitalia, in fact, all parts associated with the female. The sex of a given individual is dependent on the balance struck by the activity of the factors Mm and FF, one in the chromo- somes and the other in the cytoplasm. The FF factors are supposed to be located in the cytoplasm because if a certain numerical value is assigned to the egg, this value adheres to the maternal line, no matter which sex chromosomes are introduced from the male side in successive generations. If the factors for female- ness were carried by the male and like other paternal characters in- fluence the cytoplasm, their value would be affected by the kind of males that were employed; but Goldschmidt has shown that his results work out on the assumption that no such effects need be postulated. There is, however, another way in which the inheritance of certain factors along the maternal line may be treated. Goldschmidt has himself admitted this as a possible interpretation, although he has THE ORIGIN OF GYNANDROMORPHS. 89 adopted the cytoplasmic agency. In moths there is present, in certain species, a W sex-chromosome analogue of the Y of Drosophila that is always carried along the female line. If this chromosome carries factors it becomes one of the conditions of the result and the eggs will always be under its influence, and hence differ from the spermatozoa by a constant difference. This assumed difference might account for the fact that in reciprocal crosses the results differ and certain phases of the inheritance consistently follow the egg. There would be no theoretical objection to calling this difference " factors for femaleness." If crossing-over took place between the W and the Z chromosomes, however, this constancy would disappear. Until critical evidence can be obtained, such as the loss of the W chromosome from a line, there is no way of proving or of disproving the cytoplasmic versus W-inheritance hypothesis. In regard to the numerical values that Goldschmidt assigns to M and F, it is obvious that these are from the nature of the case arbitrary, such values being assumed as will give a consistent interpretation. Whether this mode of treatment has the advantage of a quantitative procedure, as claimed, is not so obvious, for the values are simply assigned to the data and are not given by any outside common measure, such as the chemist or physicist uses in quantitative work. If, then, the values are only numerical assumptions, the treatment is not, as Goldschmidt thinks, lifted above the symbolic handling of the problem of heredity, but stands on the same footing as all Mendelian procedure. If the numerical values assumed give consistent results when tested in other crosses where other numerical values have been assigned, there is an undoubted value in handling the problem in this way quite irrespective of the question as to what a quantitative treatment may mean. As stated, Goldschmidt interprets his results as depending on a quantitative relation of the opposing factors for femaleness and for maleness. If the quantitative difference between the factors is suffi- ciently great in one direction the individual is a male ; if in the opposite direction it is a female. If the difference is not sufficiently great either way an intersex develops. If the quantity of the female factors were greater at the beginning a female intersex results; if the quantity of the male factor were greater at the beginning a male intersex develops. Both kinds of intersexes grade in different crosses all the way from nearly normal females to nearly normal males or from nearly normal males to nearly normal females. In each series the sequence in which the characters change towards those of the opposite sex is the reverse of the order in which they develop in the individual. "The last organs to differentiate in the pupa and the first to be intersexual are the branching of the antennae and the coloration of the wings. The first imaginal organ differentiated in the caterpillar and the last hi the series to be changed toward the other sex is the sex-gland. And if we apply 90 THE ORIGIN OF GYNANDROMORPHS. this law even to the minute parts of a single organ, like the copulatory organ, we find it also to apply, as will be demonstrated later. Now, this is the fact which, in connection with the others, enables us to formulate a definite physiological theory of sex-determination."1 Goldschmidt sums up the situation in the following statement: "First, we recognized that the different effects of the same sex-factors in different combinations can be understood only by assuming a quantitatively different action; or, expressed in concrete terms, that the active substances, which we represent as factors, are present in different but typical quantities. Second, we were obliged to assume that these substances are distinct for each sex. Third, we realized that in the action of these substances a time factor is involved, which is definitely proportional to the quantities of the factorial substances. From these facts only one conclusion can at present be drawn: that the sex-factors are enzymes (or bodies with the properties of enzymes) which accelerate a reaction according to their concentration "2 If the nature of the character is dependent on the relative quantities of the male-producing enzyme called andrase and of the female-pro- ducing enzyme called gynase, the question arises how intersexes that are mosaics would ever arise, for there is no obvious reason why the relative concentration should ever change in the course of development as Goldschmidt must assume that it does change. Still less is it clear, when the difference in the concentration is less than a given critical difference (Goldschmidt's definite minimum value e), why the enzyme that starts with a lesser concentration should always overtake the other quantity, no matter which one starts below. Until this critical point is explained all the speculation that Goldschmidt brings to bear on the question only seems to cover up the difficulty rather than to clear it up. Goldschmidt appears to have overlooked this difficulty and sets up the opposite one, viz, that it is difficult to see why every gipsy moth is not an intersex. He meets this supposed difficulty by the consideration of the rate of development of the insect. Whether his answer to this difficulty is valid or not, it does not seem to meet the difficulty which to us seems the real one. Even were it established that many of the changes in embryonic and larval development are due to enzymes — a point that we are far from wishing to dispute — it need not follow that the segregating genes that give rise to them are also these same enzymes. To treat these half-way stages as the genes themselves is at present not without danger, because even if the genes are enzymes it by no means follows that the quantity of the gene is to be measured by the product of the enzyme arising from it. In his latest communication Goldschmidt states his belief that the sex-factors in the different races of gipsy moths are multiple allelo- morphs and compares them to the series of factors that Castle has 1 Goldschmidt. A Further Contribution to the Theory of Sex. (Journ. Exp. Zool., vol. 22, No. 3, April 1917, p. 597). 2 Ibid, p. 598. THE ORIGIN OF GYNANDROMORPHS. 91 found in his series of hooded rats. So far as we know, the conclusion that Castle's series of characters are mainly due to multiple allelo- morphs is far from being established; on the contrary, we are inclined to think that his evidence indicates that he is dealing mainly with a case of multiple factors. Some of the evidence that Goldschmidt himself furnishes for the gipsy moths is perhaps also capable of inter- pretation in the same way. Goldschmidt has shown in some detail that the characters or organs of the intersexes, such as the wings or external genitalia, are mosaics— i. e., relatively large segments or pieces are entirely male or female. In the case of the wings there is no obvious regularity in the mosaic pattern, for the right hind wing may be entirely different from the left hind wing, and the male parts of the right wing do not by any means correspond to the male parts of the left wing, nor does either conform strictly to any underlying structure, such as the veins. In so far, then, as each part is strictly male or female and not a blend of both, the gipsy-moth intersex is like the Drosophila gynandromorph. The results are, however, unlike the Drosophila gynandromorphs in that in the gipsy-moth hybrids the phenomenon must occur very frequently. Baltzer has shown for certain sea-urchin hybrids that when the cross is made one way there is always an irregular (?) elimina- tion of chromosomes, and this result invites at least a comparison with the gipsy hybrids. A solution of the case of intersexes in the gipsy moth «ould probably be reached by the discovery and study of sex- linked characters. Several gynandromorphs of Colias have been described (see Cock- ayne), but of unknown parentage. In the moth Algia tau also several gynandromorphs have been recorded, but the published evidence known to us does not give any clue as to their origin. As has been stated, the great majority of gynandromorph Lepidop- tera are not hybrids, but show the secondary sexual characters of the male on one side and the secondary sexual characters of the female on the other. There are, however, a few gynandromorphs in this group that show racial or specific differences along with the male and female characters. Amongst these only a few have a known ancestry, and amongst these again it is seldom known whether the characters exhibited are sex-linked or not. Even if they are sex-linked the evidence fails to discriminate between a result that depends only on sex-chromosomal differences and a result that depends on a full chro- mosome group. A search through most of the available literature has brought to light only a few cases that bear on the theories that have been already discussed. Nevertheless, it is probable that a more thorough search through the voluminous literature might furnish more of the critical evidence desired. It is not improbable that entomolo- gists who have made varietal crosses may be able to supply some of the needed data. 92 THE ORIGIN OF GYNANDROMORPHS. From the elaborate list of gynandromorphs published in 1896 and 1897 by Schultz, and from the admirable resume by Cockayne in 1915, the following cases have been chosen as the most instructive ones on record. Wheeler (1915) describes a gynandromorph from a cross of Smerin- thus ocellatus by Amorpha populi (hybridus) . The right side is female, the left side male. "The left wings are pinkish, as in ocellatus, while the right wings are entirely gray. The eye-spots of ocellatus are well developed on both wings, as is also the red basal patch of populi. Right antennae like female populi, left like male ocellatus. Right half of body light gray, left half brownish gray.' ' Since both sides of the body show some characters that belong to both parents, it is highly probable that parts of both parental nuclei are present on both sides of the gynandromorph. Briggs (1881) has also described a hybrid gynandromorph showing the characters of Smerinthus ocellatus and populi — right side ocellatus, left side populi. A figure is given, but no description. Whether from the figure it would be possible to determine whether some characters of both parents are present on both sides might no doubt be determined by an expert, but the all too brief text gives no information. Harrison crossed Ennomos subregnaria male by E. quercinaria female, and obtained many hybrids that were " practically the mean of the parents, except that they leaned in the color, both of the head and body and possibly in the general structure of the warts and tubercles, to the male parent." In describing one of these, Harrison says: "At first sight it is merely a male specimen with the left anterior female. Dissection and close examination betray much more interesting characters than that. The genitalia (fig. 4), although nearly so, are not quite purely male; the right lobe of the uncus is replaced by a fully developed right ovi- positor or lobe, while the gathous on the same side is greatly disturbed, and acts as if it were homologous to the female directing rods. In addition, whilst the coloration of both sides of the body is male, the shape of the right wing is female.' ' Harrison points out that "whilst the majority of the characters of the right side were female the color was wholly male." It appears from this description that hybrid characteristics appeared throughout, which indicates that other chromosomes than the sex chromosome were involved on both sides; but since so small apart was distinctly female, it is not entirely clear that the hybrid coloration affected this part too. He states (as above) that "while the majority of characters of the right side were female the color was wholly male.' ' Apparently by male he means hybrid male coloration, and if so the case is instructive. Harrison obtained another aberrant individual from this cross. He states that "the right being exactly that of a normal hybrid, whilst THE ORIGIN OF GYNANDROMORPHS. 93 the left side is pure subregnaria." The text leaves it uncertain whether the individual is a gynandromorph, although this appears not to be the case, for while the former specimen is classified as a gynandromorph, this one is put into a separate paragraph entitled "Asymmetrical specimen." Its genitalia are said to "present the same division of characters as those exhibited externally, as may be seen from figure 8, which shows the furca and the penis, the left side being that of the hybrid, whilst the right is evidently subregnaria, conforming itself to the structure of the left." Harrison explains the result as due to two spermatozoa entering the egg, the nucleus of one of which conjugated as usual with the egg- nucleus, but the nucleus of the other, instead of degenerating, gave rise to the nuclei determining the right side of the body, which would then be pure subregnaria and differs from the hybrid left side, which resulted from the conjugation of nuclei derived from two different species. Insofar as one side is purely paternal, this case is in line with Morgan's hypothesis of multiple fertilization and does not conform to Boveri's view. On the other hand, there is the same cytological difficulty here as encountered in Toyama's case, namely, that in Lepi- doptera the male is the homozygous individual. A single nucleus should give rise, therefore, to a female, but here probably both sides, and certainly the pure subregnaria, side is male. The hypothesis of elimination will not help out here, for even if a quercinaria daughter chromosome was the one lost, the single sex chromosome should give rise to female parts. On the other hand, one of the alternative views suggested above for Abraxas covers this case, viz, the view that an egg had two nuclei or that several sperma- tozoa entering and fusing in pairs gave rise to the male parts. Cockayne (1916) described a hybrid gynandromorph that came from a cross of Amorpha ocellatus male by A. populi female. It was male on the right side and female on the left. Although the wings did not expand, it was evident that on both sides the specific characters were intermediate between the two parents. The insect had neither ovary nor testis, but the external genitalia were male on one side and female on the other. Vasseler described a bilateral gynandromorph of Argynnis paphia in which the left side was male and paphia, and the right side was female and valesina. The latter is a characteristic varietal form. The result can be explained by dislocation of the sex chromosome on the basis that the factor of valesina is sex-linked and that it is recessive. According to Rudolphi,1 a gynandromorph was sent to McLeay from Rio de Janeiro, var. Papilio laodicus on the female side and P. polycaon on the male side. Dr. F. E. Lutz has been good enough to look up 1 Rudolphi, D. K. A., Abh. phys. klass. Konig, Akad. wiss., Berlin, 1825. See Trans. Linn. Soc., XIV, p. 584. 94 THE ORIGIN OF GYNANDROMORPHS. for me the history of the "species" question. The following note I owe to him: "Papilio androgeus is quite variable and, furthermore, shows sexual di- chromatism. Three varieties are accepted: typical androgeus (Colombia to Trinidad, Guianas, Amazon, southward to Bolivia and western Matto Grosso), epidaruus (Mexico to Panama, Cuba, Haiti, and Saint Lucia), and laodocus (Brazil and Paraguay). The name polycaon has been used by authors for each of these forms and has been applied to both males and females. The name laodicus has usually (always?) been applied to the female. It seems probable that the specimen in question was an ordinary gynandro- morph of Papilio androgeus laodicus." Cockayne has discussed at length the evidence showing that gynan- dromorphism is commoner in certain species than in others, and reached the conclusion that this is not due, in several cases at least, to the more striking characters involved, but rather to some peculiar defect in the sex-determining machinery of these species. Moreover, there appears to be good evidence favoring the view that in certain families the number of gynandromorphs is greater than in the race as a whole. The cause of this "inheritance" is obscure. Possibly these are cases of intersexuality rather than of true gynandromorphism. The evidence is more certain that gynandromorphs are more com- mon in certain hybrid combinations than in the pure parent species involved in the cross. Whether such combinations are generally due to the greater liklihood of chromosomal elimination — & view that would seem a priori possible — or to "partial fertilization" or to polyspermy can only be determined when more definite material is obtained that furnishes opportunity for genetic evidence. GYNANDROMORPHS IN OTHER INSECTS. The scarcity of gynandromorphs in other groups of insects is prob- ably due in part to the absence of conspicuous differences between the male and female in such groups as beetles or to certain groups being less collected or observed than others. We have made no attempt to search out in the literature all references to gynandromorphs. Occasional references to gynandromorphs in earwigs, Orthoptera, beetles, and bugs are to be found in the International Catalogue. GYNANDROMORPHS IN SPIDERS. In a recent paper J. E. Hull has brought together the few cases of gynandromorphs in spiders that are known. The best example is that described by Kulezynski that is male on one side and female on the other. Another described by Falconer was also male on one side, female on the other. Another gynandromorph described by the author (Hull) is male and female anteriorly and female and male posteriorly (quadripartite). Three other cases of bilateral gynandromorphs have THE ORIGIN OF GYNANDROMORPHS. 95 been reported, according to Hull, and one or two other gynandromorphs incompletely described. Most of the gynandromorphs in spiders belong to one family. Thus amongst the 232 species of British Linyphiidse there are seven gynan- dromorphs known, while amongst the 377 other species only one. Hull estimates that gynandromorphs are nine times as frequent in the Linyphiidse as in all the rest taken together. Since the male is heterozygous for the X chromosome in spiders the results may have the same explanation as in insects, but since no hybrid-gynandromorphs have been found it is impossible to do more than point out a possible solution. GYNANDROMORPHS IN CRUSTACEA. The frequency of bilateral gynandromorphs in insects is in marked contrast to the almost total absence of such types in the large group of Crustacea. It is true that in the latter there are examples of inter- sexual individuals, but it is not clear whether these come under the same category as the gynandromorph insects or are special cases more like hermaphrodites. It may be of interest to observe in this connection that in the Crustacea no sex chromosomes have as yet been discovered, but it may be replied that this may be due to the well-known difficulties of technique rather than to a real difference. However this may be, there are certain well-ascertained facts about some of the Crustacea suggesting that the condition of hermaphroditism is, so to speak, nearer the surface in the sense that the swing towards one sex or the other in a given individual is brought about more readily by age or environ- mental conditions than in other groups where a change is more difficult because the internal hereditary factor differences prevail over ordinary external or age differences. For example, in the group of cirripeds hermaphroditic species and species with separate sexes exist, as well as species related to hermaphroditic species in which the females have complemental males. It has been suggested that these males are themselves only those arrested females or hermaphrodites that settle down and become parasitic on the larger sessile females; in other words, that these males had the potentiality of becoming females if they had chanced to lead a different existence. There are families amongst the isopods that are hermaphroditic. Certain species of amphipods are said to be males when young, females when older. Eggs have been found at certain stages intermediate in size between the small male-producing eggs and larger female-producing eggs. The transformation of some of the secondary sexual characters of the male into those of the female in certain parasitized crabs has a bearing both on the relation of these characters to the sex-glands and possibly also on the causes that determine sex in the Crustacea. 96 THE ORIGIN OF GYNANDROMORPHS. Giard, and later Geoffrey Smith, have described the changes that take place when crabs are parasitized by Sacculina and other para- sitic Crustacea. When the male spider-crab Inachus dorsettensis is parasitized by Sacculina the abdomen becomes wide like that of the female, and its posterior appendages, that are absent in the male, develop and become somewhat like those of the female. The chelae likewise come to resemble those of the female. The testis, which may not be affected at first, may later degenerate to some extent, and in one case after the parasite had fallen off the regenerating testis produced eggs. It was formerly supposed that the degeneration of the testis might be the cause of the change in the secondary sexual organs, although no such relation between gonad and soma is known to exist in this group; but the work of Geoffrey Smith seemed to him to suggest that the results are directly caused by the parasite itself by stimulating the formation of fatty substance whose presence in the blood may cause eggs to develop and the secondary sexual organs of the female to appear. In other words, "the crab comes to resemble a female because the physiology of its body-tissues has been changed from the male to the female type" (Doncaster). Whatever the explanation may ultimately be found to be, the fact of the change is important. The result falls into line with the other evidence con- cerning sex determination in the Crustacea, viz, that maleness and femaleness are not so fixed by internal genetic factors if such exist, but that the balance may be shifted by other agents as well. A parallel case is known in the Andrenine bees, parasitized by another insect, Stylops. According to Perez, the stylopized males come to resemble in certain respects the females, and inversely the stylopized females the males. The sex-glands are not always affected. If in bees as in moths the secondary sexual characters are independent of the gonads, the effect of Stylops must be either directly on the host or through a change in its metabolism. W. M. Wheeler has described stylopized American wasps of the genus Polistes. No change hi the secondary characters takes place, at least not to any marked extent. The decapods have, as a rule, males and females sharply dis- tinguished, although the females of Gebia major have only ovaries, the males have, behind the testes, ovaries more or less developed. A crab, Lysmata seticaudata, has as a rule both ovaries and testes, with their ducts. Several hermaphroditic crayfish have been described, especially by Faxon (1898) and Hay (1907). One (text-figure 69) had ovaries on both sides, also on the right side a testis (without sperm) and a vas deferens. It had the external characters of a "first form" male except for the openings of the oviducts on the third pair of legs. It appears, at least in the genus Cambarus, that hermaphroditic indi- viduals are females which, "owing to some ambiguity of the formative cells in the embryo, have developed to a greater or less degree the THE ORIGIN OF GYNANDROMORPHS. 97 characters of the opposite sex. The condition is a very rare one and is usually shown in the external organs only." In the Philosophical Transactions for 1734 is a full account of a bilateral gynandromorph lobster by Dr. F. Nichols. The drawings of the external parts show that the animal is female on the right side and male on the left. Dissection showed an ovary with eggs on the right side, and a testis with vas deferens on the left. This case is exactly like the bilateral gynandromorphs of Drosophila, and is the only case known to us of a strictly bilateral type of gynandromorph in the group Crustacea. Olga Kuttner (1909) found a wild individual of Daphnia pulex that had some male characters on one side but had two ovaries. Twelve broods were produced and in nearly every brood some individuals were mixed gynandromorphs, but nearly all were predominantly female. A similar case has been recorded by Banta for Simocephalus vetulus, (1916). In a pedigreed strain there " suddenly appeared" a large number of "sex intergrades — males with one or more female secondary sex characters, females with one to several male characters, and some hermaphrodites with vari- ous combinations of male and female secondary sex characters." The more extreme intersex individuals fail to propagate; others, less modi- fied, reproduce. By propagating from female intergrades mixed broods of males, females, and intergrades are obtained. The noteworthy point here is that the intergrades are mosaics rather than blended forms of the two sexes. GYNANDROMORPHS IN MOLLUSCS. The molluscs, like the Crustacea, contain a number of hermaphro- ditic species, but there are also species with separate sexes. Here, too, cytological study has failed as yet to demonstrate sex chromo- somes. One species of Crepidula is male in the juvenile state and female in older individuals, at least when certain external conditions are fulfilled. Gould has recently shown that when young males are placed in the vicinity of large females the males absorb their testes and genitalia (ducts and penis) and develop ovaries and oviducts. This case recalls in many ways the conditions in Bonellia as described by Baltzer. If the embryos of Bonellia are isolated they become sexual females without showing the male stage. If, however, the embryo, when ready to settle down, comes to rest on the proboscis of a female it develops into a rudimentary male. A few embryos in cultures may show intermediate or rather hermaphroditic conditions. The cirri- peds referred to above appear, according to one interpretation, suscep- — oviduct —testis -/---vas.def TEXT-FIGURE 69. 98 THE ORIGIN OF GYNANDROMORPHS. tible of similar modifications, according to whether they remain free or become parasitic on a female. The conditions in Crustacea and molluscs seem to show that, in some cases at least, the animals are essentially hermaphrodites and that external conditions and age are important factors in determining the sex of the individual. These cases recall the phenomena shown by many flowering plants where at one stage or under certain conditions the male organs develop, under other conditions the female organs. If in such cases sex-determining genes are present, their influence may be readily overcome by external agencies or by age itself, which is in a sense a condition in which some part of the body (through its output) acts as an external agent to other parts. GYNANDROMORPHS IN ECHINODERMS. Cue* not (1898) and Delage (1902) described each a mature starfish (Asterias glacialis) that had small patches of testis (with sperm) in the ovary, and Buchner (1911) has recorded a similar case. Herlandt has recently described a sea-urchin (Paracentrotus) that had three normal and one "dark" testes and a large ovotestis with functional ova and sperm. Artificial fertilization with the products of this ovo- testis was successful, and the larvae were normal. Since Tennant has shown for the sea-urchin that the female is homozygous and the male heterozygous for the X chromosome, these cases can be easily explained on the hypothesis of elimination of an X chromosome — the resulting parts being male. GYNANDROMORPHS IN VERTEBRATES. The group of vertebrates shows as a rule sharp separation into two sexes, but the evidence relating to the factors involved is often so little known that the group as a whole is difficult to handle. In one subdivision, the birds, the female is the heterogametic sex with regard to sex chromosomes, while in mammals, certainly in man, it is the male that is heterogametic. The contrast here is the same as that in insects, where the moths resemble the birds and the flies man. In some of the lower groups there are evidences of hermaphroditism or of transi- tory sex conditions. It becomes necessary, therefore, to take up the different groups independently. GYNANDROMORPHS IN FISHES. Myxine, according to Cunningham and Nansen, is male when young and later becomes female. In the young the anterior portion of the testis is male, the posterior female; the testicular part atrophies after it has functioned as a testis. But the later results of the Schreiners indicate that while young Myxine is a true hermaphrodite as far as the histological structure of the glands is concerned, it is not so func- THE ORIGIN OF GYNANDROMORPHS. 99 tionally. They believe that any one individual after passing through this stage becomes definitely either male or female, although certain individuals remain sterile, neither alternative being realized (quoted from Caullery, Les Problemes de la Sexualite, 1913, p. 53.) A gynandromorph was described in 1914 by Vayssiere and Quintaret in one of the sharks, Scylliwn stellare. The left pelvic fin was female, the right male with a well-developed clasper. An ovary and both ovi- ducts were present. On the right side there was a testis, with normal male ducts on this side only. Miss Ruth C. Bamber described (1918) a hermaphroditic shark, Scythum cavicula, in which both testes were present. The anterior end of the right testis had ovarian tissue. Normal oviducts were present and the male ducts were well developed. Externally this animal was typically a male with well-developed claspers. Most of the bony fishes have separate sexes, but certain species (Serranus) are true hermaphrodites. (See Shattuck and Seligmann.) Other species give exceptional individuals that have traces of both sexes. Chidester has described a male fundulus with ova attached to the mesentery of the intestine and liver. GYNANDROMORPHS IN AMPHIBIA. The sharp separation into adult males and females is characteristic of the group Amphibia. According to Miss Stevens there is a pair of XY chromosomes in the male of one of the urodeles, but in a frog, Rana pipiens, Swingle states there is only one sex chromosome in the male. Certain species of frogs pass through a stage that appears to be hermaphroditic — at least individuals that later become males may contain in the young tadpole stage large cells that appear to be incipient ova, which later disappear when the spermatozoa are formed. In the adult toad there is a region anterior to the testis proper called Bidder's organ, in which ova-like cells are present. There are a number of observations in the older literature to the effect that well- fed tadpoles produce more females than males, and vice versa, that starved tadpoles give an excess of males. On the other hand, there are other later observations that flatly contradict these conclusions. There are some observations, especially those of King, that show the proportion of males and of females may be determined by treating the eggs (or even the sperm) with certain substances in solution, but whether the change is due to the chemicals injuring one kind of sperm (or of egg) more than the other kind, or whether the change is of a kind to really determine the sex, irrespective of the combinations formed by the germ-cells, is open to debate. The most remarkable observations on Amphibia are those of Richard Hertwig and his pupils, particularly Kuschakowitsch. They show that by delaying the fertilization of the egg there is caused an increase in the number of 100 THE ORIGIN OF GYNANDROMORPHS. males produced. By prolonging the time to the point where the eggs are almost ready to die, all or almost all of the frogs become males. The result, moreover, appears from Kuschakowitsch's results not to be due to selective mortality. Hertwig attempts to explain the results in accordance with his view of nuclear size versus cell size, but the case seems peculiarly ill suited to this interpretation, because the nucleus has dissolved and the chromosomes are already in the meta- phase condition when the eggs enter the oviduct, and it is here that the delay occurs. It is not at all obvious how delay in this condition can have much to do with cell size versus nuclear size. One of us (Morgan, 1913) has suggested that Hertwig's results may be due to a sort of parthenogenetic development in those eggs whose progress is held back. Such a result might be due either to the egg nucleus giving rise to the embryo (the sperm merely starting it, but taking no further part in the development), or to the sperm nucleus becom- ing the functional one, the egg nucleus having disintegrated in the interval. In support of such a view may be cited the observation of Oscar Hertwig and of Gunther and Paula Hertwig on frogs' eggs treated with radium. They interpret certain of their results as due to mononuclear development of the treated egg or sperm. The sex of the resulting larvse was not determined. The recent results of Loeb and Bancroft and of Loeb have shown that frogs' eggs, in- cited to development by Bataillon's puncture method, give rise to males and females in a few cases in which the frog stage was reached. Loeb states (1918) that the parthenogenetic males have the double number of chromosomes. Herlandt describes the parthenogenetic embryos of the frog as arising in such a way that the haploid number of chromosomes at the first division must be supposed to be present, but Brachet states that he has found the diploid number of chromo- somes present. Until further cytological work is done the explanation of the facts remains obscure. Swingle has recently described hermaphroditic stages of the young frog Rana pipiens. Both eggs and sperm are formed in the gonad of some individuals, whereas other individuals have only testes or ovaries, i. e., not mixed. He suggests the possibility that an irregular distri- bution of the sex chromosomes in early oogonial divisions may account for this condition. In one hermaphroditic individual he found 13 chro- mosomes in the spermatocytes, one of which is dumbbell-shaped, and this he thinks is the sex chromosome. In most first spermatocyte divisions the 12 autosomes divide, but the dumbbell-shaped chromo- some goes to one pole. Exceptionally, however, the chromosome divides, one half going to each pole. An irregular division of the kind (or of some other kind), if it occurred at an earlier stage, might give the chromosomal combination that would produce an egg, even in a potential male. THE ORIGIN OF GYNANDROMORPHS. 101 Among the urodeles, la, Valette St. George has described a newt having external male characters and an ovotestis on each side. Among the Anura several cases of hermaphroditism beside those referred to above have been described. Loisel described a frog with the secondary sexual characters of the male. On the right side no gonad was present and on the left the ovary was small and pigmented. It had no ova. This condition suggests that the male character had developed as a result of natural castration, but on the other hand, the two conditions may have had some common cause. Other cases of hermaphroditism in frogs and toads are reported by Spengel, Knappe, Hoffman, and Stephan. GYNANDROMORPHS IN REPTILES. Only two cases are known to me in this group — one a lizard and the other a turtle. Jacquet has described an individual (Lacerta agilis) that was externally a male, but had on each side a well-developed oviduct that was attached to the cloaca at one end and opened into the body-cavity at the other. No ovaries were present, however. Fantham has described a turtle (Testudo grceca) that had the external characters of a male. The concavity of the plastron was less marked than in a normal male. It had on the left side an ovo- testis, and on the right a testis. Two ova were present in the former. Such a condition might, as suggested above for the Crustacea, be imagined to be due to chromosomal elimination, but the effect here was not localized, but extended beyond the ovotestis, since both sets of ducts were present. GYNANDROMORPHS IN BIRDS. The division into males and females is sharply drawn in the groups of birds, although hi some families, as in the pigeons, the external differences (the secondary sexual differences) may be slight, while in other groups, owing to the development of secondary sexual characters, the external differences are very striking. In still other forms the secondary sexual characters appear only at certain seasons of the year and disappear largely at other seasons. The five cases of bilateral gynandromorphs that have been recorded make the group of particular interest in the present connection, while the exceptional conditions shown by certain hybrid crosses of pheasants call for careful analysis, especially in connection with what appears to be at least an analogous condition in hybrids of the gipsy moth. The genetic evidence shows very explicitly that the female is hetero- gametic, the male homogametic. The sex-linked inheritance shown by poultry and canaries is strictly comparable to that in Drosophila, except that in the birds the male has two Z chromosomes (or ZZ) and the female one Z (and possibly also a W, i. e., she is ZW). The cyto- 102 THE ORIGIN OF GYNANDROMORPHS. logical evidence that can be adduced in support of this view is not definitely established. Guyer's account of the ripening of the sperm and eggs in the fowl is as follows: In the male there are 18 chromosomes, including two Z chromosomes. After synapsis there are 9 double chromosomes in the first spermatocytes, all of which except the double Z divide (or separate), 9 going to one pole, 8 to the other. Thus one daughter cell gets both Z's. This cell divides again, the Z's presumably separat- ing, so that two second spermatocytes are produced, each with 9 chro- mosomes (in eluding the Z). These become the functional sperm. The other daughter cell (without the Z's) may divide again, but it, or its products, degenerate. In the female there are 17 chromosomes, including one Z. Pre- sumably after reduction half of the eggs contain a Z and half are without it. The Z-bearing egg fertilized by any sperm (each carries one Z) will make a male with 18 chromosomes, including two Z's; the egg without Z fertilized by any sperm makes a female with 17 chromosomes, including one Z. The scheme will account for the sex- linked inheritance shown by the fowl. All genes carried by the two sex chromosomes of the father will be transmitted to, and shown by, his daughters, because each daughter gets her single sex chromosome from her father. If the male carries dominant genes in his sex chromo- somes, both daughters and sons will show the corresponding dominant characters, etc. It is important to observe here that while this mechanism gives the same results as to sex and sex-linked inheritance as the mechanism described by Seiler for moths, the actual process by which the two end-results are reached are quite different in the male, although presumably the same in the female. In the moth the reduction has been worked out both in the male and female, while in the bird only in the male. Five cases of gynandromorph have been described in birds, four of which were bilaterally halved.1 Poll described a bullfinch that had a testis on the right side, and this side had the red color on the breast characteristic of the normal male; on the left side there was an ovary, and the left side of the breast was gray like the normal female. (See frontispiece in Doncaster's book on The Determination of Sex.) Weber gives a full account of a finch, Fringilla ccelebs, that had the adult male plumage on the right side and that of the female on the left side. The left side contained an ovary, the right a testis. Weber states that Cabanis (Journ. fur Ornithologie, XXII, 1874) describes a "Dompfaffen" (Pyrrhula vulgaris) that was a bilateral gynandromorph — on the right side male, on the left female. The bird 1 Several mixed cases in hybrid pheasants and in Tetrao testrix'ha.ve been omitted here, as well as references to "hermaphroditic" fowls. THE ORIGIN OF GYNANDROMORPHS. 103 was not dissected. He records, apparently also on the authority of Cabanis, another bilateral gynandromorph in the species Colaptes mexicanus. Here, curiously enough, the right half was female, the left male, but Weber suggests that possibly the bird had the adult male plumage on the left side, while on the right the plumage was juve- nile; in other words, the bird was a male, but with the full plumage only on one side, and that the left side, which normally contains the ovary. Brandt states that Lorenz found in the markets of Moscow, in the course of fifteen years, three male Tetrao tetrix with female plumage; one of these had a testis on one side and an ovary on the other. Bond has described a pheasant with the plumage of the left side preponderantly male, that of the right side preponderantly female. On the left side there was an ovary, and this is the normal position of the ovary in birds. It contained both ovarian and testicular tissue. There was no trace of a gonad on the right side. In the last three cases there is no stated correspondence between the external and the internal division of the sexes. Setting aside the two rather doubtful cases (that of Cabanis and the uncertain reference to Lorenz's case), there remain the two well- established cases of Poll and Weber, where dissection was made, and Bond's case, that is like the last, but not so clear, since the ovary contained also testicular tissue. It is very difficult to explain these cases by chromosomal elimina- tion, even if the male and female plumage differences were supposed to be due to two or one (Z) chromosomes in the parts affected. Start- ing as a male with two Z chromosomes, if one were lost at an early division one half of the bird would be female, Z, and the other male, ZZ. This possibility could be established only by finding a bilateral gynandromorph in a hybrid that was heterozygous for sex-linked factors. Such factors have been described for pigeons (Cole and Staples-Browne) and for doves (Strong, R. M.,and Riddle), for canaries, and for fowls, but no cases of gynandromorphs hi them have yet been met with in which these characters were involved. An attempt to bring the avian results in line with the Drosophila runs counter to the evidence from gonadectomy, since it assumes that the differences involved are due directly to the chromosomal composition of the male and the female parts, and are not due to ovarian extract, which, in poultry and ducks at least, has been shown to suppress in the female her potentiality of developing the full cock plumage. It may be interesting to review briefly this situation, since Goodale's results with ducks show that the relation of the plumage to the gonad is not so simple as appeared at first. It has long been known in poultry that the removal of the testes does not interfere with the development of the secondary sexual plumage of the cock. In color, shape, and size of the feathers the capon 104 THE ORIGIN OF GYNANDROMORPHS. is very similar to the normal cock. The comb and wattles, however, are greatly reduced in size and have a pale color, being relatively deficient in blood. The influence of castration on the spurs is not clear, for they may be well developed in the capon and even in the hen. The influence of the ovary on the plumage of the hen has long been suspected to be important. Old hens in which the ovary had ceased to function were known to develop cock feathering, and the same result was said to follow if the ovary became diseased. But much uncertainty existed in regard to this evidence until Goodale, by care- fully planned and thorough work, showed that when the ovary was removed from young birds they developed the complete plumage of the male. In the race of Leghorns the cock is red with plumage like that of the wild Gallus bankiva; the hens are brown. After spaying, the hens develop the complete male plumage. The spurs develop more fully than in the normal female of the Leghorn race. When pieces of the ovary of a Leghorn hen were inserted in the body-cavity of a Leghorn capon, the latter developed only the female plumage. In domesticated ducks (Rouen and Mallard) there are two molts. The drake molts in June and assumes his summer plumage, which is more like that of the female than is his other so-called nuptial plumage. The nuptial plumage develops during the autumn molt. If the testes are completely removed after the autumn molt the male retains his nuptial plumage even through the summer molt. Goodale finds that in normal birds, when the summer plumage reaches its highest stage of development, sexual activity diminishes or disappears, and few or no sperms are present. It is at this time then that the drake de- velops his nuptial plumage, as removal of feathers shows. In other words, it is the summer plumage (the one that is more like the female) that develops when the sexual organs are at the highest development, while the nuptial plumage develops when the sperms are not being produced in the testes. It appears, then, that the nuptial plumage is not influenced by the testicular condition, while the female-like plumage may possibly be due to the inhibitory effects of the testicular secretions. In other words, the case is somewhat like that of the Sebright, in which the presence of the active testis suppresses the potential cock feathering of the male. These results do not appear to furnish any solution of the problem of bilateral gynandromorphs in birds, because the chief difficulty remains so long as any internal secretion, whether ovarian or testicular, determines in an individual the character of its plumage. Any theory of bilateral gynandromorphs in birds must be prepared to offer some explanation as to why the ovarian extracts do not suppress in them the male feathering on the male side. Two more or less plausible answers can be given at present. One of them is that in certain THE ORIGIN OF GYNANDROMORPHS. 105 species of birds the male plumage is not affected by ovarian secretions, as it is in poultry and in ducks, but is due directly to genetic factors that act effectively in the male but not in the female. It ought to be comparatively easy to find this out for each race by means of gonodectomy. The other possible explanation is that although in a bird genetically male (ZZ) on one side and female (Z) on the other, the secondary sexual characters would be female; yet if the ovary should become diseased or old and its secretions diminished, a point might be reached where the secretion could no longer hold in check the full develop- ment of the male part. The bilateral gynandromorph in birds would on this view represent only a transient stage. In point of fact, none of them have been kept alive for any length of time, so that we do not know that they would hold their superficial peculiarity. An alternative to this view that the secretions were insufficient because of disease or age is to suppose that the ovary is abnormally small from accident or heredity. In this case the gynandromorph stage would be more permanent. Such birds would be expected in all cases to have an ovary, or at least to have some traces of one, unless the species resembled Mallards or Sebrights, where the testis influences the plumage. The results that Riddle has reported concerning intersexes in hybrid pigeons do not call for detailed review here, since the phenomena recorded relate largely to behavior. Riddle believes that under " conditions of overwork" a female produces eggs, some of which are male-producing, others female-producing, as shown by mating such females to the males of their own species when equal numbers of males and females are produced. But such overworked eggs, if fertil- ized by a male of a different genus, produce predominantly female birds. The result, however, is not attributed to the male, or to the cross, but to some change in the egg that causes a reversal of the sex tendency. The only case that Riddle has reported in which the color inheritance is given, so that one can follow the sex-linked heredity in connection with the abnormal sex ratio, is that recorded in the Naturalist for 1916.1 The first 17 doves were 5 male to 12 female doves; the second 17 doves were 4 males to 13 females; the last 17 doves were 2 males to 15 females. The cross was made between Streptopelia alba male and St. risoria female. As R. M. Strong had previously shown, the expectation here is for dark sons and white daughters. Since the reciprocal cross gives all dark offspring, the factor involved is sex- linked and not merely sex-limited. Riddle obtained only dark males and white females, except two that were dark (one being questioned by himself). Strong also found a few dark exceptions, as did also Staples Brown. As Bridges has shown, these exceptions can be 1 Reproduced and expanded in the Journal of the Washington Academy of Sciences, June 1917. 106 THE ORIGIN OF GYNANDROMORPHS. explained by non-disjunction. They are too few in any case to affect Riddle's argument based on the sex-ratio. It follows, then, that Riddle's results, instead of showing that some females started as males, show exactly the reverse, since the genetic history shows that all his females must have had the genetic chromosome constitution character- istic of the female and have gotten it in the usual way. GYNANDROMORPHS IN MAMMALS-MAN. Cases of true hermaphroditism or gynandromorphism in mammals and in man are extremely rare. From the meager evidence it is not clear whether the cases reported belong under one or the other head, but there are, as far as we know, very few if any cases of strictly bilateral gynandromorphs. The secondary sexual differences, while not so marked as in some other groups, are yet sufficient, one would suppose, to make a bilateral type clearly evident. Goldschmidt has suggested that intersexes occur in man of the kind shown by the gipsy moth. So far, at least, there is no positive evidence to show that such individuals occur more frequently in racial crosses in man than within the race, but the human races are themselves so mixed in origin that this point may not have any critical value for the subject. A priori, it is equally possible that the intersexual individuals, if genetic ones exist, may be due to autosomal differences that affect the normal instincts rather than to differences in the sex genes themselves. It is not claimed, I believe, that the actual sex-organs themselves are involved, but rather secondary sexual characters and instincts whose relation to the sex mechanism are in man entirely obscure. According to Rudolphi, there is a record by Schlumpf (Arch. f. Thierheilkunde, Zurich, 1824, pp. 204-206) of a calf externally like a male, but in place of the scrotum there are present the udders with the usual number of nipples. The uterus had only one horn and funnel, and an ovary fastened to right side of "der Leiden." To one kidney (left) was attached a small testis. Rudolphi also describes a seven weeks' old child that lived about three months that had a hypospadic penis and in the right scrotum a testis, but no testis in the left. There was a uterus whose left upper end was connected with a Fallopian tube attached to which was an ovary. On the right side the uterus ended blindly and there was neither Fallopian tube nor ovary present. Two very similar cases, one by Gautier (1752) and one by Pinel, are referred to by Rudolphi.1 According to Pick (1914), Sauerbeck admits only 7 cases of her- maphrodites in mammals as certain and complete, 5 for swine, and 2 for man (Salens, 1899, and Simon, 1903), to which number are added 3 1Rudolphi (1825) refers to two supposed cases of hermaphrodites in fowls, which he very properly questions. THE ORIGIN OF GYNANDROMORPHS. 107 B TEXT-FIGURE 70. cases for mammals (roebuck and goat) and 5 for man as very probable. To these Pick adds a later case by Uffreduzzi (1910) and one by Gudernatsch (1911). The 5 cases found in swine by Pick are of unusual interest. The external genitalia were entirely female or nearly so. Within the abdomen were uterus and fallopian tubes (text-fig. 70) . In four of the cases an ovotestis was present on each side in the normal position of the ovary. In the fifth case an ovary with a small piece of testis was on one side and a testis on the other. The conditions here suggest at least that the results are not due to chromosomal elimination, al- though such an interpretation might be given. If, for instance, the gonads arise at an early stage from a single cell in which an anterio-posterior division occurred and the later mass of cells was subsequently separated into right and left parts, the condi- tions found might be realized. There is, however, a possibility that here, as in cattle, a union between the chorions of the embryo in the uterus might have brought about a more perfect freemartin than develops in cattle when such a union occurs. (See Lillie.) Ritter described a pig in which on the right side a testis was present and on the left an ovary. (Verh. phys. Med. Gesell. Wiirzburg, XIX, 1886). Harman more recently (1917) describes a cat with a testis on one side and an ovary on the other. Neurgebauer has given a large number of cases in man in which testes and ovaries have been described in the same individual and in which the genitalia show many anomalous relations. Amongst the large number of human hermaph- rodites described there are probably a considerable number of authentic cases where parts of both male and female genitalia were combined in the same individual, but writing as late as 1911, Guder- natsch states that hermaphroditism in the sense of separate ovaries and testes has not been demonstrated in man. He describes a case of an individual with female external genitalia and an abnormal testis hi the right inguinal canal. The proof that hermaphroditism, so-called, in man is produced in the same way as gynandromorphism in Drosophila can not be furnished at present, because there is no probability of the difference in chromo- some number being determined by histological study, owing to the 108 THE ORIGIN OF GYNANDROMORPHS. large number of chromosomes, nor is it probable that a case involving sex-linked factors will soon be found. Some of the older writers seem to mean by hermaphroditism the presence of complete sets of both male and female organs, the two systems superimposed on each other. The rather mythical accounts of such cases do not call for serious comment.1 Where the evidence is anatomical and given by trained observers it appears that some, perhaps all, cases are mosaics rather than "hermaphrodites" in the sense of double-sexual individuals. In other words, parts of one and parts of another system are found in the same individual, replacing each other locally. If this interpretation turns out to cover certain cases the theory of chromosomal elimination will suffice at least as a formal explanation of such human abnormalities. Since the human species is both from the genetic and cytological evidence XX hi the female and XO in the male, the same mechanism exists as is found in Drosophila, and if the theory of chromosomal elimination applies here also, human gynandromorphs would be expected in practically all cases to begin as female (XX) and produce male regions by eliminat- ing one X. An examination of the literature shows hi fact a consider- able preponderance of the cases showing more female than male regions, but the evidence is too uncertain to give any serious weight to it. IS CANCER A SOMATIC MOSAIC? Into the difficult and obscure question as to the cause of cancer it is not our business to enter, but a suggestion made by Boveri (in 1902 and 1914) calls for brief notice, since he appealed to a process akin to chromosome elimination as a possible explanation of the phenomenon. Boveri suggested that an imperfect or irregular division of the chromosomal complex might in certain cases produce combina- tions through loss of specific chromosomes that caused the different cells to run wild, so to speak, in the sense that factors that normally inhibit the rate of growth or the suppression of growth in relation to the cell environment are lost. In support of such a view he appealed to occupational cancer-growth, where cancer develops in parts of the body most subject to mechanical injury or pressure. It is known to students of embryology that compression of a dividing cell may inter- fere with the normal distribution of the chromosomes to the daughter cells. At present, however, reference to such possible sources is too uncertain to be of great value, for there are no instances where irregu- larities of this kind are known to give rise to prolific growth processes. The cancer-like or tumor-like growth shown by a mutant race of Drosophila, discovered by Bridges and described fully by Stark, has not been shown to be associated with abnormal distribution of the 1 See comment by Dr. H. L. Ganigues, Medical Record, 1896, p. 725. THE ORIGIN OF GYNANDROMORPHS. 109 chromosomes, although this point has not been sufficiently studied to exclude such a process. On the other hand, it has been shown that the growth in question is caused by a sex-linked Mendelian gene that is inherited strictly, as are all Mendelian sex-linked genes. This mutant lethal race of Drosophila arose as a mutation, presumably in the same way as other mutations. If it is not admissible at present to draw any analogy between this case and that of mammalian cancer, it is conceivable at least that mammalian cancer may be due to re- current somatic mutation of some gene. Such a conclusion would, however, not invalidate the view that cancer is more likely to occur in certain families, or even be inevitable in them, because recurrent mutation in certain genes appears to be more likely than in other genes. But even if this view were maintained the inheritance would be different in kind from the inheritance of ordinary Mendelian genes, because such a view involves a secondary step, viz, the likelihood of a mutation in a race containing the inherited gene in question. The whole problem of the causes of mutation is at present so obscure that a discussion of this possibility is purely theoretical. Added to this is the uncertainty of how cancer is inherited in those races of mice that appear to produce it with great frequency. Important as the work along these lines unquestionably is, the subject is not yet ripe for any positive statement. It may, nevertheless, be worth while to keep in view the possibility suggested above, viz, that what is in- herited in cancer may be a gene or complex of genes in which somatic mutation is of sufficient frequency to give the appearance that a gene for cancer is itself inheritable. IS THE FREEMARTIN A GYNANDROMORPH? It has been suggested that the pair of twin calves, one male, the other a sterile female (the freemartin), together represent a sort of gynandro- morph. This view is based on the assumption (which Lillie has since disproven) that these twins arise from a single egg. Hart (Proceed- ings Roy. Soc. Edinburg, XXX, p. 218) suggested that "the free- martin with a potent bull twin is the result of a division of a male zygote, so that the somatic determinants are equally divided and the genital determinants unequally divided, the potent going to one twin, the potent bull, the non-potent, genital determinant to the free- martin." It is needless to point out that this vague statement can not be brought into accord with embryological evidence, because Lillie's work shows that each individual of the twins arises from a separate egg. In most cases the eggs arise from the two ovaries, and each embryo lies in a different horn of the uterus. Lillie has shown that in those cases where twins are present, one of which is a freemartin, the two chorions and the two allantois have fused at an early stage, and he has demonstrated that there is an 110 THE ORIGIN OF GYNANDROMORPHS. actual vascular connection between the two individuals. There can remain no doubt that the results are due to the establishment of a common circulation. Lillie brings very strong evidence in favor of the view that the freemartin starts as a normal female. The failure of her ovary to develop, he thinks, is due to a sex hormone (see below) that originates in the testis of the male and suppresses the normal development of the ovary. The external genitalia of the freemartin, and to some extent the uterus and ducts, are, as a rule, less affected by the hormone, so that externally the freemartin appears to be a female. Even more remarkable is the fact that the male ducts are sometimes quite well developed and the development of the ovary appears to take in somewhat the characteristic changes seen in the development of the testis. This conclusion is based largely on the results of a histo- logical examination by Miss C. L. Chapin. Lillie is not inclined, how- ever, to lay very much emphasis on this side of the question, because, as he states, the suppression of the ova (and female stroma?) may in itself allow some of the male characteristics to develop to a stage not normally present in the female. In other words, the development of the accessory organs may to a certain extent be under the influence of the gonad. The assumption of a male hormone originating in the interstitial cells of the testis is more problematical. The only fact advanced by Lillie in favor of this interpretation is that in the testis the interstitial tissue develops at an earlier stage than that in the ovary. It is true that there is also some evidence indicating that the interstitial cells of the testis produce some substance that affects the secondary sexual characters of the male. But it may be that other substances in the blood of the male affect the ovary of the freemartin and retard its development. Such substances might also be called hormones, but have no direct relation either to the development of the germ-cells in the testes or to sex determination in any specific sense. If in cattle the male differs from the female by one sex chromosome, it is quite possible that the composition of the blood of the male is different in some substances (or relative proportion of substances) from the blood of the female. The difference, while the product of sex in the sense that all the body-cells of the male differ by one chromosome from the body-cells of the females, might not in any way be connected with sex determination, even although it affected injuriously the develop- ment of the ovary of the young female embryo. Until further evidence is obtained, the source of the ' 'hormone" that affects the freemartin must remain an open question. If the ovary of the freemartin is actually changed to a testis it may be said that the freemartin is a sex mosaic, the external genitalia female and the gonads more or less male. The cause of such a sex mosaic would, then, obviously, be entirely different from the cause of the gynandromorphs of Drosophila. THE ORIGIN OF GYNANDROMORPHS. Ill SUMMARY. (1) (a) The main outcome of this work on gynandromorphs of Drosophila is an experimental demonstration of the principal cause of the regional differences that gives rise to the combinations of male and female in the same individual. The demonstration was made possible by taking advantage of the genetic situation in this material. (6) Many of the gynandromorphs were hybrids of known sex-linked characters, i. e., characters whose genes are carried by the sex chromosomes. (c) By adding to such crosses additional characters whose genes lie in other than the sex chromosomes it has been possible to prove that the male and female parts of the gynandromorph differ by the sex chromosomes alone, i. e., both male and female parts contain the same autosomal group. (d) It was possible, in consequence, to show that these gynandro- morphs are not due to partial fertilization (Boveri), or to polyspermy (Morgan), but to chromosomal elimination (Morgan). Chromosomal elimination means that at an early stage in embryonic development one of the daughter chromosomes of one of the X's fails to pass over to one of the daughter plates, and accordingly gets left out of that nucleus. In consequence, one of the two cells will contain only one X chromosome and produce male parts, while the sister cell with two daughter X chromosomes will produce female parts. The evidence that elimination of this kind takes place rests on cases in which the X chromosome derived from the father contains different sex-linked genes from the X chromosome derived from the mother. (e) A census of the available gynandromorphs shows that a paternal X chromosome is eliminated as often as a maternal X chromosome. (2) A logical consequence of the proof that the gynandromorphs arise through elimination is that they should all start as females, i. e., as XX individuals. If the elimination always takes place at the first division the expectation would be for the male and female parts to be equal; but if at the second, third, or any later division of the nuclei, we should expect to find, on the whole, a preponderance of female parts over male parts. Such is strikingly the case. (3) A second logical consequence of chromosomal elimination is that starting as an XX individual ; the male parts will be XO, and not XY as in the normal male. Now, it has been shown by Bridges (1916) that XO males arising from primary non-disjunction are sterile (although in structure, etc., they are exactly like XY or normal males). The great majority of gynandromorph individuals with male abdomen and testes are infertile, while if the corresponding parts are female the individual is fertile. The few gynandromorphs, fertile as males, are known from other genetic evidence to have come from XXY 112 THE ORIGIN OF GYNANDROMORPHS. mothers or to be themselves XXY zygotes. In such cases, after elimi- nation, the male parts are expected to be XY, and hence an individual of this origin with a male abdomen and testes is expected to be fertile. (4) A striking fact in regard to these gynandromorphs is that the male and female parts and their sex-linked characters are strictly self-determining, each developing according to its own constitution. No matter how large or how small a region may be, it is not interfered with by the aspirations of its neighbors, nor is it overruled by the action of the gonad. (5) Four experiments were made in which suitable material was carefully scrutinized for gynandromorphs. In the 88,000 flies, there were found 40 gynandromorphs, or 1 to 2,200. Since only those that start as females give this kind of gynandromorph, chromosomal elimination may have occurred once in 1,100 individuals. (6) (a) If chromosomal elimination took place at the first division of the segmentation nucleus, a half-and-half gynandromorph is expected (right-and-left or anterio-posterior). Whether dorso-ventral separation is expected for such a division depends on whence comes the material that ultimately reaches the dorsal surface of the fly. (6) If the chromosomal elimination took place at the second-division period in one of the nuclei only a quadrant is expected to be male, etc. (c) The fact that most of our mosaics include large regions of the body may mean that elimination takes place more often during the first or second division, but it may also mean that when smaller regions are involved the gynandromorph would be more often overlooked. (7) (a) Both gonads of the same individual are always alike, i. e., both are testes or both are ovaries, even when the external markings of the abdomen are male on one side, female on the other. This result finds its explanation in the assumption that the germ-plasm of Drosophila, as in some other flies, arises from a single cell. This cell, arising after elimination, must be either a spermatogonium or oogonium. If the cell be the former the sex-linked factor of the germ-plasm must be that of the male-determining X chromosome alone and not show any of the factors contained in the other X of the female parts. Such is the case. (6) Conversely, the ovary of a gynandromorph containing both X chromosomes should produce eggs containing the original X chromo- somal combinations as well as their cross-over combinations. This, too, is the case. (8) It is a striking fact that we have found so few cases of autosomal elimination. The lack of such mosaics may be due to the failure of the ordinary chromosome to lag in division as the X is assumed to do, or it may be that a fly or part of a fly can not exist if one autosome is absent from its complex. That a part may exist with one X chromo- some lost might be explained as due to that condition having been THE ORIGIN OF GYNANDROMORPHS. 113 already acquired by the male. Future work must show whether or not such autosomal mosaics are viable. (9) Courtship has been watched in a number of flies that were partly male and partly female. Many of them are indifferent; some react as males, some as females. (10) In several cases flies that had one white eye and one red eye have been observed to show circus movements. Since the white- eyed fly is less responsive to light than the red-eye fly, the circus move- ments of the gynandromorph with one white and one red eye is what is to be expected. Of course, such cases must be selected so that the legs are not male on one side and female on the other. (11) The general evidence from mutation in Drosophila makes it highly probable that when a mutation occurs it takes place in only one chromosome of the pair. Hence any mutation in somatic tissue, if recessive, would be concealed by the presence of the normal allelomorph in the homologous chromosome. If, however, a mutation should appear in the sex chromosome of the male, even though recessive, its effects might be apparent. It is probably significant that the ten cases here described and supposed to be somatic mutations are all males. (12) Theoretically, at least, there is the possibility that an indi- vidual starting as a male might produce female parts. If at some embryonic division both daughter X's of an XY cell should pass into the same cell, it would be expected to produce female parts. There is, however, a difficulty with the other cell containing a Y chromosome and no X. It would probably die. (13) In addition to the two earlier theories of Boveri and Morgan mentioned above, other theories are critically considered from the point of view of the gynandromorphs of Drosophila. The only other theory besides elimination that we have found necessary to employ in ac- counting for the gynandromorphs of Drosophila, where the genetic evi- dence makes the analysis possible, is the theory of binucleated eggs. (14) In the light of the evidence from Drosophila, both the Eugster- bee gynandromorph and von Engelhardt's gynandromorph can be accounted for on the hypothesis of chromosomal elimination, especially since the work of Newell and Quinn shows that the racial characters involved differ in one Mendelian gene (though not necessarily <3ne in the sex chromosome). However, in both cases, if paternal and maternal elimination are equally likely in both combinations, as many gynandromorphs showing the racial character of only one type are expected as those mosaic for racial characters as well as for sex. Such have not been reported. (15) (a) In moths several gynandromorphs have been reported that were mosaics for paternal and maternal characters well as as for sex. Some of these, starting as males, can be explained by chromosomal elimination. 114 THE ORIGIN OF GYNANDROMORPHS. (6) In Abraxas a factor involved is known to be sex-linked. Two mosaics between A. grossulariata and lacticolor described by Doncaster can be accounted for by chromosomal elimination in one case and by a non-disjunctional sperm and elimination in the other. (c) The two gynandromorphs in silkworms described by Toyama can be explained, genetically, on the basis of two nuclei present in the eggs. Doncaster has found such eggs in Abraxas. (d) Whether the mosaics in the gipsy moth, formed by racial crosses, are due to different sex-factors having different quantitative value, as maintained by Goldschmidt, or due to some other relations, seems uncertain from the evidence so far published. (16) In reviewing the literature it is pointed out that in the Crus- tacea and molluscs there are several cases where an individual is male at one period of its life and female at another, just as some plants pass through similar stages. In such cases the environment, taken in the widest sense, may suppress one sex and develop the other. The influence of the environment is clearly shown in the case of the crabs infected by Sacculina, where the secondary sexual characters are changed ; and in Crepidula, where proximity to another individual effects a change of sex, and in the worm Bonellia, where a similar change is brought about. There is nothing here that is in the least inimical to the view that in other cases, and even in these same groups, there may be genetic factors that determine sex under ordinary or other circumstances. The bilateral gynandromorph of the crayfish (p. 97) may be a case in point. (17) A few cases of bilateral gynandromorphs in birds have been reported. Their occurrence is unexpected because of the known effect of the ovary in suppressing most of the secondary characters of the male. It is suggested that in some species of birds, particular secondary sexual differences are not influenced by internal secretions, hence a gynandromorph condition in the chromosomal composition might show itself in plumage characters. It is also suggested that if a bird showed the female complex in one region and a male complex in another the amount of internal secretion that might inhibit one side might be in- sufficient to inhibit the other. A transient or an abnormal condition of the ovary might make the gynandromorph differences visible. (18) In man and in other mammals a number of cases of gynandro- morphs are known, some of them at least well authenticated. Most of the cases rest on the condition of the gonads and accessory sexual organs. Sex mosaics like those of Drosophila are expected, because the mechanism of sex determination is the same. On the other hand, in the light of Lillie's evidence for the freemartin, other kinds of modifica- tions may be possible. Even in cases where only a single individual is born an earlier connection with an absorbed or aborted embryo might be responsible for an abnormal condition of the sexual organs. THE ORIGIN OF GYNANDROMORPHS. 115 POSTSCRIPT. Professor F. R. Lillie has called my attention to two important papers on freemartins. It appears that Tandler and Keller1 had already published, in 1911, the essential facts relating to the vascular connection between the embryos in utero, leading to the development of the freemartin out of the female member of the united pair. They had also shown that the embryos come from two eggs. Magnussen2 in 1918 has described a considerable number of cases of freemartins. He regards both individuals as having started as males, and compares the usual rudimentary condition of the testes of the freemartin with that of a cryptorchid testis. He adduces no evidence of importance in favor of his view that the twins started as males, while LilhVs evidence is convincing in support of the view that the freemartin started as a female. The most important facts reported by Magnussen are those relating to the histological condition of the testes of the freemartin. Well-developed testes are present in some of the older freemartins, ranging in size from that of a hazel-nut to that of a hen's egg. The vasa deferentia, the epididymus, and notably even the tubular tissue characteristic of the testes were present, but no germ-cells were found. Now the absence of germ-cells from the tubular tissue of the testes of the adult freemartin may be accounted for, as Magnussen does account for it, viz, as due to the "retention" of the testes of the free- martin. This condition would not, however, be expected to hold for the embryonic testes, where in the walls of the testes at birth one would expect to find the germ-cells present. If a critical examination of these stages shows that germ-cells are not present in the tubules of the testes of the freemartin, then the evidence from the freemartin shows not that the sex of the female has been changed, but that under the influence of the blood of the male the accessory organs, as well as the secondary sexual organs characteristic of the male, have developed in the female; while at the same time her own female accessory organs have correspondingly failed to develop fully. This state- ment implies that the critical evidence for sex is the kind of germ-cell produced, while the development of the secondary sexual characters and of the accessory organs of reproduction in the mammal is determined, in part at least, by the germ-cells. It will be recalled that, according to the most recent work in mammalian embryology, the germ-cells originate in or from the region of the intestinal tract, far removed from the final position in the gonad into which they find their way by migration. If, then, it prove that no true germ-cells are found in the testicular tubules of the freemartin, the presence of the "testes," including even the epididymus, and tubules demonstrates only how far the origin of these parts is dependent on something in the male ; but whether this something comes from the germ-cells of the male (directly or indirectly) or is a consequence of the genetic composition of the male is not shown. July 17, 1919. 1 Tandler und Keller, 1911. Ueber das Verbal ten des Choriona bei verschieden-geschlecht- licher Zwillingsgravitat des Rindes, uhd ueber die Morphologic des Genitales der weiblichen Tiere, welche einer solchen Gravitat enstammen. Deutsche tieraerzliche Wochenschrift. (No. 10.) 2 Magnussen, H., 1918. Geschlechtslose Zwillinge. Eine gewShnlich Form von Herma- phroditismua beim Rinde. Archiv. f. Anat. u. Physiol. Anat. Abt. LITERATURE. ANDREWS, E. A. 1909. A male crayfish with some female organs. The American Nat- uralist, XLIII, 1909. ARMBRUSTER, L. 1913. Chromosomen Verhaltnisse bei die Spermatogenese solitar Apidse. (Osmia cornuta) Arch. f. Zellf., XI. BABCOCK, E. B., and R. E. CLAUSEN. 1918. Genetics in relation to agriculture, New York. , and F. E. LLOYD. 1917. Somatic segregation. Journ. of Heredity, VIII, 2. v. BAEHR. 1908. Uber die Bildung von Sexualzellen bei Aphidse. Zool. Anz., bd. XXXIII . . 1909. Die Ovogenese bei einigen viviparen Aphididen und die Spermatogenese von Aphis saliceti. Arch. f. Zellf., III. BALTZER, F. 1909. Die Chromosomen von Strongylocentrotus lividus und Echinus microtubermlatus. Arch. f. Zellf., II. . 1913. Uber die Herkunft der Idiochromosomen bei Seeigeln. Sitzber. d. phys.- med. Ges. Wurzburg. . 1914. Die Bestimmung und der Dimorphismus des Geschlechtes bei Bonellia. Sitzungsb. d. phys.-Med. Ges. zu Wurzburg. BAMBER, RUTH C. 1918. Note on a hermaphrodite dogfish. Proc. Zool. Soc. London, III and IV. BANTA, A. M. 1916. Sex intergrades in a species of Crustacea. Proc. Nat. Acad. Sci., II. BATESON, WILLIAM. 1913. Problems of genetics (see pp. 44-45). Yale University Press. BAUR, E. 1914. Einfuhrung in die experimentelle Vererbungslehre. Berlin. BECKER, EMIL. 1896-97. Uber Zwitterbildung beim Schwein. Verh. d. phys.-med. Ges. in Wurzburg, XXX. BLOCHMANN, F. 1886. Uber die Richtungskorper bei Insekteneiern. Morph. Jahrb. XII. BOND, C. J. 1914. On a case of unilateral development of secondary male characters in a pheasant, with remarks on the influence of hormones in the production of secondary sex characters. Jour. Gen., III. BORING, A. 1909. A small chromosome in Ascaris megalocephala. Arch. f. Zellf., IV. , A. M., and R. PEARL. 1917. Sex studies. IX. Interstitial cells in the reproductive organs of the chicken. Anat. Rec., XIII. BOUIN et ANCEL. 1904. Sur un cas d'hermaphrodisme glandulaire chez les mammiffcres. Compt. rend, de stances de la socie'te' de biolog. BRAKE, B. 1907. Resultate der Kreuzung. zwischen Lymantria dispar. Ent. Zeit., XX. , and C. FRING. 1911. Einige weitere Ergebnisse der Brake'schen Lymantria dispar, Ent. Zeit., XXV. BOVERI, TH. 1888. Uber partielle Befruchtung. Sitz.-Ber. d. Ges. f. Morph. u. Phys., Munch., IV. . 1908-09. Uber Beziehungen des Chromatins zur Geschlechtsbestimmung. Sitz. Phys.-med. Ges. Wurzburg. . 1909. Uber Geschlechtschromosomen bei Nematoden. Arch. f. Zellf., IV. . 1911. Uber das Verhalten der Geschlechtschromosomen bei Hermaphroditismus. Beobachtungen an Rhabditis nigrovenosa. Sitzber. phys.-med. Ges. Wurzburg. . 1914. Un Frage der Entstehung maligner Tumoren. Jena., G. Fischer. . 1915. Uber die Entstehung der Eugsterschen Zwitterbienen. Arch. f. Entw. Organismen, XLI. BRANDT, ALEX. 1889. Anatomisches und Allgemeines liber die sogenannte Hahnen- fedrigkeit und uber anderweitige Geschlecht Anomalien bei Vogeln. Zeitschr. f. wiss. Zool., Bd. 48. BRIDGES, C. B. 1913. Non-disjunction of the sex chromosomes of Drosophila. Jour. Exp. Zool., XV. . 1916. Non-disjunction as proof of the chromosome theory of heredity. Gen., I. BUCHNER, F. 1911. Uber hermaphrodite Seesterne. Zool. Anz., Bd. 38, p. 315. BUTTEL-REEPEN, H. v. 1906. Apistika. Beitr. zur Systematik, Biologic, sowie zur geschichtlichen und geographischen Verbreitung der Honigbiene (Apis mellifica L.), ihrer Varietaten und der ubrigen Apis-Arten. Mitt, aus dem Zool. Museum zu Berlin., III. . 1911. Eine Stacheldrohne, Bienenwirtsch. Zentralbl. Nr. 14. Hannover. 116. LITERATURE. 117 CAULLERY, M. 1913. Les Problemes de la sexualite'. Paris. CHAPIN, CATHARINE I. 1915. A case of hermaphroditism in Spelerpes bilineatus. Biol. Bull., XXIX. CHIDESTER, F. E. 1917. Hermaphrotism in Fundulus heteroclitus. Anat. Rec., XII. COCKAYNE, E. A. 1915. Gynandromorphism and kindred problems. Jour. Gen., V. . 1916. Gynandromorphous Lepidoptera. Trans. Ent. Soc. London. . 1916. The relation between the secondary sexual characters and the gonads and accessory sexual glands in insects. Trans. Ent. Soc. London. . 1916. An intersex of Amorpha populi. Trans. Ent. Soc. London. . 1916. Gynandromorphous Agriades coridon. Trans. Ent. Soc. London. COLE, L. J. 1912. A case of sex-linked inheritance in the domestic pigeon. Sci., XXXVI. . 1916. Twinning in cattle, with special reference to the freemartin. Science, n.s., XLIII. COLLINS, G. N. 1913. Mosaic coherence of characters in seeds of maize. Cir. U. S. Dept. Agr., Bur. PI. Industry, 132, 19-21. COLLINS, J. L. 1919. Chimeras in corn hybrids. Journ. of Heredity, X, 1. CORRENS, C. 1899. Untersuchen tiber die Zenien bei Zea Mays. Ber. d. Deutsche Bot. Ges., 17. . 1901. Bastarde zwischen Maisrassen mit besonderer Beriicksichtigung der Xenien. Billiotheca Botanica 53, 1-161. . 1913. Xenia and the endosperm of Angiosperms. Bot. Gaz. 56, 217-224. CUENOT, L. 1898. L' hermaphrodisme protandrique d' Asterina gibbosa et ses variations suivant les localites. Zool. Anz., Bd. 21, p. 273. DALLA TORRE, K. W. v., u. FRIESE, H. 1898. Bericht des naturwiss. med. Vereins in Innsbruck. XXIV. Jahrg. DANNENBERG. 1913. Stammbaum Fragen der Smer. ocellatus L. und Am. populi L. Gruppe. Zeitschr. f. Wissen. Insektenbiologie, IX. DELAGE, Y. 1912. Quelques experiences et observations sur les Asteries. Arch. Zool. exp. et gen. IX, p. 237. DEXTER, J. S. 1912. On coupling of certain sex-linked characters in Drosophila. Biol. Bull., XXIII. . 1914. The analysis of a case of continuous variation in Drosophila. Am. Nat., XLVIII. DONCASTER, L. 1907. Gametogenesis and fertilisation in Nematus ribesii. Quart. Journ. Mic. Sc., LI. . 1910-11. Gametogenesis of the gallfly, Neuroterus lenticularis (Spathegaster baccarum). Parts 1 and II. Proc. Roy. Soc., B., LXXXII and LXXXIII. . 1910. Note on the spermatogenesis of Abraxas grossulariata (currant moth). Proceed, of the Cambridge Phil. Soc., XV. . 1911. Some stages in the spermatogenesis of Abraxas grossulariata and its variety lacticolor. Journ. Gen., I. . 1912. Note on the chromosomes in oogenesis and spermatogenesis of the white butterfly, Pieris brassicoe. Proc. Cambridge Phil. Soc., XXVI. . 1913. On an inherited tendency to produce purely female families in Abraxas grossulariata. Jour. Gen., III. . 1914. On the relations between chromosomes, sex-limited transmission, and sex determination in Abraxas grossidariala. Jour. Gen., IV. . 1914. The determination of sex. Cambridge. . 1914. Chromosomes, heredity, and sex: A review of the present state of the evidence with regard to the material basis of hereditary transmission and sex- determination. Quart. Journ. Mic. Sc., LIX. . 1916. On some gynandromorphic specimens of Abraxas grossulariata. Proc. Cambridge. Phil. Soc., XVIII. , and G. H. RAYNOR. 1906. Breeding experiments with Lepidoptera. Proc. Zool. Soc. London, I. Also 1907, Nature, LXXVI. . 1908. Sex inheritance in the moth Abraxas grossulariata and its variety lacti- color. Rep. Evol. Com., IV. DODGE, S. C. 1879. Monstrosities among bees (hermaphroditism). Amer. Bee Journ., XV. DONHOFF. 1860 a. Em Bienenzwitter. Bienenzeitung. . 1860 b. Beitr. z. Bienenkunde, I. Uber Zwitterbienen. Bienenzeitung. 118 LITERATURE. DUNCAN, F. N. 1915. An attempt to produce mutations through hybridization. Amer. Nat., XLIX. . 1915. A note on the gonads of gynandromorphs of Drosophila ampelophila. Am. Nat., XLIX. EAST, E. M., and H. K. HAYES. 1911. Inheritance in maize. Conn. A. E. S. Bui. 167. EMERSON, R. A. 1915. Anomalous endosperm development in maize and the problem of bud sports. Zeits. f. ind. Abstam. u. Vererb., XIV, 5. . 1917. Genetical studies of varigated pericarp in maize. Gen., II. ENDERLEIN, G. 1913. Ein hervorragender Zwitter von Xylocopa mendozana aus Argen- tinien. Mit einem Vergeichnis aller bisher beobachteten gynandromorphen Hymenopteren. Stettin, entomol. Ztg., 74-124. v. ENGELHARDT. 1914. Uber den Bau der Gynandromorphen Bienen. Zeit. f. wiss. Insektenbiologie, X. FANTHAM, H. B. 1905. On hermaphroditism and vestigial structures in the reproduc- tive organs of Testudo grceca. Annals and Magazine of Natural History, XVI, 1905. FAXON. 1898. Observations of the Astacidae in the U. S. Nat. Museum, XXI. FEDERLEY, H. 1913. Das Verhalten der Chromosomen bei der Spermatogenese der Schmetterlinge Pygaera anachoreta, curtula und pigra, sowie einiger ihrer Bas- tarde. Zeitschr. f. indukt. Abst. u. Vererb., IX. FLETT, W. 1878-79. Gynandromorphism in the honey-bee. Scot. Natural., IV. Fuss, A. 1912. Uber die Geschlechtszellen des Menschen und der Saugetiere. Arch. f. mikr. Anat., LXXI. GAERTS, J. M. 1911. Cytologische Untersuchungen einiger Bastarde von Oenothera gigas Ber. Deutsch. Bot. Gesellsch., 29, 160-166. GERSTACKER, C. E. 1864. Uber Zwitterbildungen der Honigbiene. Sitzungsber. Ges. naturf. Fr. Berlin. GIARD, A. 1886. De 1'influence de certains parasites Rhizoce"phales sur les caracteres sexuels exte"rieurs de leur h6te. C. R. Acad. Sc., Paris, 103. . 1887. La castration parasitaire. Bull. Sc. D6p. Nord., XVIII. . 1887. Sur la castration parasitaire chez I'Eupagurus bemhardus et chez la Gebia steUata. C. R. Acad. Sc., Paris, 104. . 1888. Sur la castration parasitaire chez les genres Paloemon et Hippolyte. C. R. Acad. Sc., Paris, 106. . 1888. La castration parasitaire (Nouvelles Recherches). Bull. Sc. Dep. Nord., I. GOLDSCHMIDT, R. 1912. Erblichkeitsstudien an Schmetterlingen. I. Zeit. f. ind. Abst. und Vererb., VII. . 1916. Experimental intersexuality and the sex-problem. Am. Nat., L. . 1916. A preliminary report on further experiments on inheritance and de- termination of sex. Proc. Nat. Acad. Sciences, 11. . 1917. A further contribution to the theory of sex. Jour. Exp. Zool., XXII. , and H. POPPELBAUM. 1914. Erblichkeitsstudien an Schmetterlingen. II. Zeit. f. indukt. Abstam. und Vererbungslehre, XI. GOODALE, H. D. 1911. Studies in hybrid ducks. Jour. Exp. Zool., X. . 1916. Gonadectomy in relation to the secondary sexual characters of some domestic birds. Carnegie Inst. Wash. Pub. No. 243. . 1916. A feminized cockerel. Jour. Exp. Zool., XX. . 1916. Further development in ovariotomized fowl. Biol. Bull., XXX. GOTTSCHALK, EDUARD. 1909. Uber einen Fall von Hypertrichosis faciei (Frauenbart) in Verbindung mit Hennaphroditismus verus lateralis. Fortschr. auf dem Gebiete der Rontgenstrahlen, XIII, H. 6. GOULD, H. N. 1917. Studies on sex in the hermaphrodite mollusc Crepidula plana. I. History of the sexual cycle, XXIII. GROSVENOR, G. H., and GEOFFREY SMITH. 1913. The life cycle of Moina rectirostris. Q. J. Mic. Sc., LVIII. GUDERNATSCH, J. F. 1910. Ein Fall von Hermaphroditismus verus hominis. Verh., VIII. Inter. Zool. Kongress, Graz. . 1911. Hermaphroditism verus in man. Am. Jour. Anat., XI. GULICK, A. 1911. Uber die Geschlechtschromosomen bei einigen Nematoden nebst Bemerkungen iiber die Bedeutung dieser Chromosomen. Arch. f. Zellf . GUYER, M. F. 1910. ' Accessory chromosomes in man. Biol. Bull., XIX. LITERATURE. 119 GUYER, M. F. 1916. Studies on the chromosomes of the common fowl as seen in testes and in embryos. Biol. Bull., XXXI. HARRISON, J. W. H. 1917. Studies in the hybrid Bistoninae. II. Jour. Gen., VI. , and L. DONCASTER. 1913. On hybrids between moths of the geometrid sub- family Bistoninae. Jour. Gen., III. HEINRUTH, H. 1909. Ein lateral hermaphroditisch gefarbter Gimpel. Sitz. Ges. Natur- forschender Freunde. HEGNER, R. W. 1909o. The origin and early history of the germ-cells in some chrysomelid beetles. Journ. Morph., XX. . 19096. The effects of centrifugal force on the eggs of some chrysomelid beetles. Journ. Exp. Zool., VI. . 1912. The history of the germ-cells in the psedogenetic larva of Miastor. Science, XXXVI. . 1914a. Studies on germ-cells. I. The history of the germ-cells in insects, with special reference to the Keimbahn-determinants. II. The origin and significance of the Keimbahn-determinants in animals. Journ. Morph., XXV. . 1914. The germ-cell cycle in animals. Macmillan Co., New York. HERBST, C. 1901. Formative Reize in der Tierischen Ontogenese. Leipzig. HERLANOT, M. 1913. Etude sur les bases cytologiques du me'canisme de la parthe'noge'nese exp^rimentale chez les amphibiens. Arch. d. Biol., XXVIII. . 1918. Un cas d'hermaphrodisme complet et fonctionnel chez Paracentroius lividus. Arch, de Zoo. Exp. et Gen., 1, 57. HERTWIG, R. 1912. Uber den derzeitigen Stand des Sexualitatsproblem nebst eigen Untersuchungen. Biol. Centr., XXXII. HOFSTAETTER. 1912. Uber Kryptorchismus und Anomalien des Descensus testiculorum. Klin. Jahrb., XXVI. HULL, J. E. 1918. Gynandry in Arachnida. Jour, of Genetics, VII., 1918. HYDE, R. R., and H. W. POWELL. 1916. Mosaics in Drosophila ampelophila. Gen., I. JACQUET, MAURICE. 1895. Note sur un cas d'hermaphroditisme incomplet chez Lacerta agilis. Bibliogr. Anat., III. 1895. JOICEY, J. J.,and G. TALBOT. 1918. A gynandromorph of Papilio lycophron Hub. Proc. Zool. Soc. London, III and IV. KAHLE, W. 1908. Die Paedogenese der Cecidomyiden. Zoologica, XXI. KELLOGG, VERNON L. 1908. Inheritance in silkworms. I. Stanford Un. Pub. No. 1. KING, HELEN DEAN. 1910. Some anomalies in the genital organs of Bufo lentiginosus and their probable significance. Am. Jour. Anat., X. . 1911. Studies on sex determination in Amphibians. IV. The effects of external factors, acting before or during the time of fertilization, on the sex ratio of Bufo lentiginosus. Biol. Bull., XX. KINGSBURY, B. F. 1909. Report of a case of hermaphroditism (Hermaphroditismus verus lateralis in Sus scrofa). Anatom. Record, III, 1909. KNAPPE, EMIL. 1886. Das Biddersche Organ. Morph. Jahrb., II, 1886. KOSMINSKY, P. 1909. Einwirkung ausserer Einflusse auf Schmetterlinge. Zool. Jahrb., XXVII. . 1911. Weitere Untersuchungen iiber die Einwirkung ausserer Einflusse auf Schmetterlinge. Zool. Jahrb., XXX. KTTSCHAKEWITSCH, SERGITJS. 1910. Die Entwicklungsgeschichte der Keimdriisen von Rana esculenta. Ein Beitrag zum Sexualitatsproblem. Festschrift zum sech- zigsten Geburtstag Richard Hertwigs, II, pp. 61-224. KUTNER, OLGA. 1909. Untersuchungen iiber Fortpflanzungsverhaltnisse und Vererbung bei Cladoceren. Intern. Rev. d. ges. Hydrobiol. u. Hydrographie, II. LANDAU, L. 1913. Diskussion zum Vortrage L. Picks: Uber den wahren Hermaphro- ditismus des Menschen und der Saugetiere. Berliner klin. Wochenschr., XLV. LANG, A. 1912. Vererbungswissenschaftliche Miszellen. Zeit f. indkt. Abs. und Vererb., VIII. LAUBENDER, B. 1801. Einige Bemerkungen iiber die von Herrn Schulmeister Lukas neu entdeckten Stacheldrohnen. Okonom, Hefte XVII. LEFEBURE, A. 1835. Description d'un Argus cdexis hermaphrodite. Ann. Soc. Entom. France, IV. LILLIE, FRANK R. 1916. The theory of the free-martin. Science, n.s., XL V11I, pp. 61 1-613. 120 LITERATURE. LILLIE, FRANK R. 1917. The free-martin. A study of the action of sex-hormones in the foetal life of cattle. Jour. Exp. Zool., XXIII. LOISEL, GUSTAVE. 1901. Grenouille femelle presentant les caracteres sexuels secondaire du male. C. R. de la soci^t^ de Biologie. 1901. LONNBERG, E. 1898. Some biological and anatomical facts concerning Parastacus. Zool. AM., XXI. LUCAS, H. 1888. Note relative a deux abeilles anormales. Bull. Soc. Entom., France. MARK, E. L., and M. COPELAND. 1907. Maturation stages in the spermatogenesis of Vespa maculata. Proc. Amer. Acad. Arts and Sciences, XLIII. MARSHALL. 1884. On certain abnormal conditions of the reproductive organs of the frog. Jour. Anat. and Phys., VIII, 1884. MATAUSCH, IGNAZ. 1909. Gynandromorphic Membracidae. Journal of the New York Entomological Society, XVII, 1909. . 1911. The effects of parasitic castration in Membracidae. Journal of the New York Entomological Society, XIX, 1911. MAY, H. G. 1917. The appearance of reverse mutations in the bar-eyed race of Drosophila under experimental control. Proc. of Nat. Acad. Sc., III. McCLUNG. 1905. The chromosome complex of orthopteran spermatocytes. Biol. Bull., IX. McEwEN, R. S. 1917. The reactions to light and to gravity in Drosophila and its mutants. Jour. Exper. Zool., XXV. MEHLING, E. 1915. Uber die Gynandromorphen Bienen des Eugsterschen Stockes. Verh. Phys.-Med. Gesell. Wurzburg., XLVIII. DE MEIJERE, J. C. H. 1911. Uber getrennte Vererbung der Geschlechter. Arch. Rass.- Ges.-Biol., VIII. MEISENHEIMER, J. 1907. Ergebnisse einiger Versuchsreihen iiber Exstirpation und Transplantation der Geschlechtsdrusen bei Schmetterlingen. Zool. Anz., XXXII. . 1909. Experimentelle Studien zur Soma- und Geschlecht-differenzi'erung. Jena. MEIXNER. 1905. Zur Frage des Hermaphroditismus verus. Zeitschr. f. Heilk., XXVI. MENZEL, A. 1862a. Abnormitat in der Bildung einer Biene. Bienenzeitung. . 18626. Uber Zwitterbienne. Bienenzeitung. . 1862c. Hymenopterologische Beobachtung I. Uber die Geschlechtsverhaltnisse der Biene im allgemeinen und iiber die Befruchtung der Konigin, iiber Partheno- genesis und Zwitterbildung im besonderen. Mitt. Schweiz. entom. Ges., I. . 1864. Tod der Zwittermutter des Eugsterschen Stockes in Konstanz. Bienen- zeitung, XX. MEVES, FR. 1907. Die Spermatocytenteilungen bei der Honigbiene (Apis mellifica L.), nebst Bemerkungen tiber Chromatinreduktion. Arch. mikr. Anat. Entwicklungs- gesch., LXX. MORGAN, T. H. 1905. An alternative interpretation of gynandromorphous insects. Science, XXI. . 1907. The cause of gynandromorphism in insects. Am. Nat., XLI. . 1908. Sex determination in phylloxera and aphids. Science. . 1909a. Are the drone eggs of the honey-bee fertilized? Am. Nat., XLIII. . 19096. Hybridology and gynandromorphism. Am. Nat., XLIII. . 1909c. A biological and cytological study of sex determination in phylloxerans and aphids. Journ. Exper. Zool., VII. . 1912. The elimination of sex chromosomes from the male-producing eggs of phylloxera. Journ. Exper. Zool., XII. . 1913. Heredity and sex. Columbia University Press, New York. . 1914. Mosaics and gynandromorphs in Drosophila. Proc. Soc. Exp. Biol. and Med., XL . 1915. The infertility of rudimentary winged females of Drosophila ampelophila. Am. Nat., XLIX. . 1916. The Eugster gynandromorph bees. Am. Nat., L. , and C. B. BRIDGES. 1913. Dilution effects and bicolorism in certain eye-colors of Drosophila. Jour. Exp. Zool., XV. MXJLLER, H. J. 1917. An Oenothera-like case in Drosophila. Proc. Nat. Acad. Sci., III. NAGEL. 1899. Zur Frage des Hermaphroditismus verus. Arch. f. Gynakol., LXVI11. NACHTSHEIM, H. 1913. Cytologische Studien liber die Geschlechtsbestimmung bei der Honigbiene (Apis mellifica L.). Arch. f. Zellf., XI. LITERATURE. 121 NEUGEBAUER, FRANZ L. 1908. Hermaphroditismus beim Menschen. Leipzig. NEWELL, W. 1914. Inheritance in the honey bee. Science, XLI. NEWMAN, H. H. 1917. The biology of twins. University of Chicago Press. NORTON, J. B. S. 1910. Some unusual tomato variations. Proc. Soc. Hort. Sci. OEHNINGER, M. 1913. Uber Kerngrossen bei Bienen. Verh. d. phys.-med. Ges Wurzburg, XLII. OUDEMANS, J. TH. 1899. Falter aus castristen Raupen, wie sie aussehen und wie sie sich benehmen. Zool. Jahrb., XII. PEARL, R. 1912. A case of triplet calves, with some general considerations regarding multiple gestation in normally uniparous animals. An. Rpt. Maine Agr. Exp. Station for 1912, pp. 259-282. , and FRANK M. SURFACE. 1915. Sex studies. VII. On the assumption of male secondary characters by a cow with cystic degeneration of the ovaries. An. Rpt. Maine Agr. Exp. Station for 1915, pp. 65-80. PICK, LUDWIG. 1914. Ueber den wahren Hermaphroditismus des Menschen und der Saugethiere. Arch, f . mikr. Anat., LXXXIV. POLL, HEINRICH. 1909. Zur Lehre von den Sekundaren Sexual Charactern. Sitz. Ges. Naturf. Freunde zu Berlin. POTTS, F. A. 1910. Notes on the free-living nematodes. I. The hermaphrodite species Quart. Journ. Mic. Science, LV. RIDDLE, OSCAR. 1916. Sex control and known correlations in pigeons. Am. Nat., pp. 385-410 (references to earlier papers of the author here). . 1917. The control of the sex ratio. Journ. Wash. Acad. Sc., VII, No. 11. . 1917. The theory of sex as stated in terms of results of studies on pigeons. Science, n.s., XLVI, No. 1175, pp. 19-24. ROEPKE, W. 1909. Ergebnisse anatomischer Untersuchungen an Standfuss'schen Lepi- dopteren bastanden. Jena., Zeit. f. Naturw., XLIV. RTJDOLPHI, D. K. A. 1825. Beschreibung einer selienen menschlichen Zwitterbildung. Abh. phys. Klass Konig Akad. Wiss., Berlin. SALEN, ERNST. 1899. Ein Fall von Hermaphroditismus verus unilateralis beim Men- schen. Verh. d. Deutsch. Pathol. Ges., II, SAUERBECK, ERNST. 1909. Uber den Hermaphroditismus verus und den Hermaphro- ditismus im allgemeinen vom morphologeschen Standpunkt aus. Frankf. Zeitschr., f. Pathol., Ill, pp. 2, 3, und 4. SCHLEIP, W. 1908. Die Richtungskorperbildung im Ei von Formica sanguined. Zool. Jahrb., Anat., XXVI. . 1909. Die Reifung des Eies von Rhodites rosae L. und einige allgemeine Be- merkungen uber die Chromosomen bei parthenogenetischer Fortpflanzung. Zool. Anz., XXXV. . 1911. Uber die Chromatinverhaltnisse bei Angiostomum. Ber. natur. Ges. Freiburg i. Br., XIX, und Arch. f. Zellforsch., VII. SCHULTZ, O. 1896. Gynandromorphe (hermaphroditische) Macrolepidopteren der palaarktischen Fauna. Illust. Wochenschr. f. Entom., I. SEILER, J. 1913. Das Verhalten der Geschlechtschromosomen bei Lepidopteren. Zool. Anz., XLI. SHATTOCK, S. G., and C. G. SELIGMAN. 1906. An example of true hermaphroditism in the domestic fowl, with remarks on the phenomenon of allopterotism. Trans. Path. Soc. London. v. SIEBOLD, C. TH. 1864. Ueber Zwitterbienen. Zeit. f. wissensch. Zool., XIV. . 1866. Ersatz der abgestorbenen Zwittermutter des Eugsterchen Zwitterstockes in Constanz. Bienenzeit. Jahrg., 1866. SIMON, WALTER. 1903. Hermaphroditismus verus. Virchows Arch., 172. SMITH, F. 1862. Imperfect hermaphrodite of Apis mellifica from Scotland. Proc. Entom. Soc. London. . 1871. Notes on examples of gynandromorphism in aculeate Hymenoptera. Trans. Entom. Soc. London. SMITH, GEOFFREY. 1906. Fauna und Flora des Golfes von Neapel. Rhizocephala. Zool. Sta. Neapel. Monographic, 29. . 1909. Crustacea. Cam. Nat. Hist. . 1910-12. Studies in the experimental analysis of sex. Parts 1-9. Quart. Journ. Micro. Sc., LIV, LV, LVI, LV11, LVIII. 122 LITERATURE. SPIEGELBERG, OTTO. 1861. Uber die Verkiimmerung der Genitalien bei (angeblich) ver- schieden geschlechtlichen Zwillingskalbern. Zeitschr. f. rst. Med., Ill, Reihe, XI. ST. GEORGE, LA VALLETTE. 1895. Zwitterbildung beim kleinen Wassermolch (Triton taenialus). Arch. f. mikr. Anat., XLV. STANDFUSS, M. 1896. Handbuch der Palaarktischen Gross-Schmetterlinge. Jena. STEINACH, E. 1910. Geschlechtstrieb und echt Sekundare Geschlechtsmerkmale als Folge der innersekretorischen Funktion der Keimdriisen. Zentralbl. f. Physiol., XXIV. . 1912. Willkiirliche Umwandlung von Saugetier. Maunehen in Tiere mit aus- gepragt weiblichen Geschlechtscharakteren und weiblicher Psyche. Arch. f. Ges. Phys., 144. STEVEN, P. 1901. De 1'hermaphrodisme chez les Vertebras. Annales de la Facult6 des Sciences de Marseille, XII, 1901. STEVENS, N. M. 1905. Studies in spermatogenesis, with especial reference to the "ac- cessory chromosome." Carnegie Inst. Wash., Pub. No. 36. . 1908. A study of the germ-cells of certain Diptera. Jour. Exp. Zool., V. . 1909. An unpaired chromosome in the aphids. Jour. Exp. Zool., VI. . 1911. Heterochromosomes in the guinea-pig. Biol. Bull., XXI. STRONG, R. M. 1912. Results of hybridizing ring-doves, including sex-linked inheritance. Biol. Bull., XXIII. STOUT, A. B. 1915. The establishment of varieties in Coleus by the selection of somatic variations. Carnegie Inst. Wash., Pub. No. 218. STURTEVANT, A. H. 1915. Experiments on sex recognition and the problem of sexual selection in Drosophtta. Journ. Animal Behav., V. SWINGLE, W. W. 1917. The accessory chromosome of a frog, etc. Biol. Bull., XXXIII. TICHOMIROW, A., und LORENZ. Bull. Soc. Imp. Nat. Moscow, 1894, p. 333, cited by Naumann. Naturgeschichte der Vogel Mitteleuropas, herausgeg. v. Hennicke, III. TouRNEtrx. 1904. Hermaphroditisme de la glande genitale chez la taupe femelle adulte et localisation des cellules interstitielles dans le segment spermatique. Compt. rend, de 1'association des anatomistes, VI6., session, Toulouse, pp. 49-53. TOYAMA, K. 1906. Studies on the hybridology of insects. I. On some silkworm crosses, with special reference to Mendel's law of heredity. Bull. Col. Agric., Tokyo, Imperial University, VII. UFFREDUZZI, O. 1910. Ermafroditismo vero nell' uomo. Archiv. per le Scienze Mediche, XXXIV. . Idem. Giornale della R. Academ. di Medicin. di Torino, XVI, ann. LXXII1, fasc. I, II. . Idem. Archiv. di Psichiatr., XXXI, fasc. VI. VATSSIERE, A., et G. QUINTARET. 1914. Sur un cas d'hermaphrodisme d'un Scyllium stettare L., C. R. Ac. Sci. Paris, 158, pp. 2013-2014. VOSSELER, J. 1893. Uber Zwitterbildung bei Insekten, Jahreshefte des Verein f. Vater- landsche Naturkunde in Wurttenberg, XLIX. WEBBER, H. J. 1900. Xenia, or the immediate effect of pollen in maize. Bui. 22, Div. Veg. Phys. Path., U. S. Dept. Agr., 1^4. WEBER, M. 1890. Uber einen Fall von Hermaphroditismus bei Fringitta ccelebs, Zool. Anz., XIII. WEISMANN, A. 1882. Beitrage zur Kenntnis der ersten Entwicklungsvorgange im Insektenei. Festschrift fur J. Henle, Bonn. WENKE, KARL. 1906. Anatomic eines Argynninis paphia, Zwitters, etc. Zeitschr. f. wiss. Zool., LXXXIV. WHEELER, W. M. 1903. Some new gynandromorph ants. Bull. Am. Museum Nat. Hist., New York. . 1910. A gynandromorphous mutillid. Psyche, XVII. . 1910. The effects of parasitic and other kinds of castration in insects. Jour. Exp. Zool., VIII. . 1911. Insect parasitism and its peculiarities. Pop. Sc. Mon., November. . 1914. Gynandromorphous ants described during the decade 1903-1913. Am. Nat., XLVIII. WHITEHEAD, R. H. 1904. The embryonic development of the interstitial cells of Leydig. Am. Jour. Anat., III. WITTENHAGEN. 1861. Uber Bienencharakteristik und Bienenzwitter. Bienenzeitung. II. THE SECOND-CHROMOSOME GROUP OF MUTANT CHARACTERS. BY C. B. BRIDGES AND T. H. MORGAN. With seven plates and seventeen text-figures. 123 THE SECOND-CHROMOSOME GROUP OF MUTANT CHARACTERS. BY C. B. BRIDGES AND T. H. MORGAN. INTRODUCTION. This paper deals with 39 mutant races whose genes lie in the "second chromosome.' ' This number includes all of those previously described, and gives a complete account of all the second-chromosome mutants found before 1916. The latter have been for the most part used or mentioned in previous papers by ourselves and others. It has been our rule never to hold back any useful mutant type until it had been recorded by its discoverers, and in consequence a number of the char- acters here described for the first time have already been used widely. This applies equally to the third-chromosome mutants, an account of which we hope to publish soon. . In addition to the 39 mutant types here described there remain about 35 others discovered since 1915 which are still to be described. In making our selection for the present publication we have included those which are most essential for future work both in localization of genes and in special experiments. For example, star eye, because of the location of the gene at the extreme 'left' end of the chromosome, of its dominance, and of its other excellent characteristics, is now the most generally useful second- chromosome mutant. Again, purple eye- color is the only workable eye-color in this chromosome. It has been involved in the localization of many of the genes in the chromosome. Its position is at the center of the chromosome, which center shows certain important peculiarities. Furthermore, its location near black gives a working distance suitable for analyzing linkage and coincidence characteristics; the distance between purple and black is short enough to exclude double crossing-over and long enough to exclude the large and uncertain probable error incident to small percentages of crossing- over. A third important mutant type, speck, whose gene is located near the right end of the chromosome, is used as the basis of reference of the genes of that end of the chromosome. Finally, curved is, in addition to its excellent viability, ease of identification, and other useful features, very valuable from its position at the right of the cen- tral group of the most useful and best located genes. The mutants have been discussed in order of their discovery, since this method involves least use of material requiring special explana- 125 126 THE SECOND-CHROMOSOME GROUP tion, and the earlier experiments were relatively simple. In the case of several of the mutants found very early, little more than a review of the extensive work already published has been given. A chronologically arranged list of all these mutants, together with a summary of main points with respect to their origin, locus, etc., is given in table 1. TABLE 1. — Chronologically arranged list of 1 1 -chromosome mutant genes treated in this paper. Mutant Sym- Locus. Date Culture gene. Fig. bol. S' + Primary base found. No. Speck fPl 5, figs. 1 ) • 80 105.1 c+31.6 1910 Mar. — Wild stock . . . Morgan. Olive fig. 73. PI 5 fig 1 1 106 1 ± «» + 1 ± Do Do. PI 6 T' 28.0 S' +28- 0 Do. Black Body-color PI. 5, fig. 2.. PI 7, fig 1 b ba 46.5 105.5 sp+ 0.4 Oct. — Miniature ex- periment. Do. Do. PI 7 fig 2 65 0 pr-t-12 3 Dec — Do Do. Lethal 7" Life IT 2" ±15 1911 Feb — Do Do. Blistered .... Jaunty Venation Wing curvature. . Text-fig. 74. PI 7, fig 3 b9 103 =•= 46.7 »p— 2 =*= 6+02 Nov. 16 Dec 11 Rudimentary stock. Do. A 23 A 34 Bridges. Do. Curved Do. Text-fig 75 73.5 pr+20 8 Dec 24 Do. A 43 Do. Purple Eye-color PI. 5, fig. 8. . 52.7 6+62 1912 Feb 20 t» stock A66 Do. Strap Wing PI 8 65 0 uj ± 0.0 Apr Do Morgan. Arc Wing curvature. PI. 7, fig. 4. 98.4 tp —6 7 May 24 Black-palpi B30 Bridges. Gap Venation Text-fig 76 July 10 stock . b Xo B42 Do. Antlered Wing PI 9 65 0 vg * 0 0 Sept Morgan. /Venation PI. 10 *a« Oct — Do. Streak .... ILegs ; . . . Thorax pattern . Figs. la-Id.. /PI. 5, fig. 5.. } d }Si 15 4 d 13 6 Nov. 22 Nov. 27 Sable experi- ment. C146 C 149 Bridges. Do. Comma*. . . . Thorax marks. . . \P1. 10, fig. 2. Text-fig 79 !At So* 15 1913 Feb. 5 d^pink II 9 Do. Morula Apterous*. . . Eye-facets Wing, balances PI. 10, fig. 3. PI 7 fig 5 mr 166.3 48 5 a+ 7.9 S' +48 5 * Mar. 8 Aug. — peach Xwild . . . M20 Do. Wallace Cnl II crossing-over. . 6 ? Sept. — " Nova Scotia' ' 7 Cur Do. Cnr •ox ? Sept. — Do. 7 Do Cream II*.. . Eye-color PI 5 fig 10 Sept. 15 Lethal 2 stock. Patched* Trefoil* Abdomen Thorax pattern. . PI. 11 /PL 5, fig. 6.. \ "ti " 50 ± Nov. 25 Nov.— cm M68 Do. Cream 6*. ... Eye-color \P1. 8, fig. 1. . PI 5, fig. 11 f ' 22 5 1914 Mar. 10 82 Pinkish* . . . Do. PI. 5, fig. 12. 100. + 6+60 ± July 27 tion. Eosin 6 stock . Do Plexus Limited Venation Abdomen Text-fig. 80. Text-fig. 81 PX 96.2 106.3 »j>- 8.9 mr+? Aug. 20 Sept. 13 Spread stock . . 557 511 Do. Do Confluent*. . . Venation Ct Sept. 23 550 Do Fringed* .... Star Wing Eye-facets Text-fig. 82 Text-fig 83 fr S' 98.0=*= 0 0 6+51.5 St— 15 4 1915 Jan. 20 Feb. 12 tion. Jaunty I 1,042 1,347 Do. Do Nick Wing Text-fig 84 65 0 t<7± 0.0 May 7 tion. Lethal 2 2,012 Do Dachs-lethal . Life dl 29 0 d± 0 0 Oct. 6 S'Xd 2,217 Do Squat* Lethal Ho. . . Wing Life Text-fig. 85 St 35.5 62 7 6-11.0 c— 10 8 Nov. 29 Dec. 2 Non-disjunc- tion. 2,480 2,675 Do. Do Telescope .... Abdomen, wing . . PI. 7, fig. 6. tt 66 5 Dec. 27 'Crooked' ex- 2,735 Do. Modifiers .... Dichaete »»— ? 1916 Aug. 13 periment. Dachsoid Venation Text-fig. 86 1917 Feb. 9 Eosin X 'seple' 2,671 Do. *Tha mutations marked with the asterisk (*) are no longer on hand, having been lost or discarded. OF MUTANT CHARACTERS. Special attention has been given in the treat- ment of the data to bring out the genetic methods employed, and to trace the development of such methods. The part each mutant has played in this development of methods and principles has been given fully, often at the expense of repeti- tion. An effort has been made to evaluate each mutant with respect to its usefulness as a working tool. Some are practically useless, while certain others, from the possession of excellent charac- teristics or favorable location in the chromosome, etc., must be considered in every carefully planned experiment . By such methods of presentation it is hoped to make available not simply a body of data but also a working familiarity with the material. From the various published and hitherto un- published materials on crossing-over in the second- chromosome a summary is given of total data available on amount of crossing-over between va- rious loci (table 140). The cross-over values cal- culated from these data are still further summar- ized and presented in graphic form in the map of the second chromosome which appears as fig. 72. In the construction of this map it was necessary to correct some of these values by aid of informa- tion gained from a study of the amount of double crossing-over and coincidence in the various re- gions of the second chromosome. Details of how this was done will be found in the last section, which deals with the construction and use of this map. The coincidences, and consequently the corrections, are at present only quite rough approximations, and the same is true of the meth- ods of weighting employed in that section. It should be borne in mind that the map is a com- posite picture in which differences in the data from different sources are no longer apparent. Ref- erence to these separate data will show, however, a very surprising uniformity, especially in view of the many conditions now known to be able to cause the amount of crossing-over to vary. A glance at the map shows at the right end (bot- tom) an unusually dense cluster of genes. When TEXT-FIGURE 72. — Map of the second chromosome, giving the locations with reference to star, and the symbols of the mutants whose loci are known. 127 - Star ( S ) - Streak (Sfc) 28.O 29. 0 65.0 665 105.1^ 1O5.5- 106.1? 106.3 Truncate(T) Dachs(d), dachs -lethal (dj) 35.5- - Scjuat(Sq) Black (b) Jaunty (j) 4-8. 5 -j- Apterous !) Morula (m limited ? FIG. 72. 128 THE SECOND-CHROMOSOME GROUP we come to add to the map the loci of the genes as yet unpublished, another such cluster will appear at the left end of the chromosome also. These clusters at the ends may be looked upon, not as due to the genes being here actually nearer together, but to the probability that at the ends crossing-over is relatively less frequent than in the middle part of the chromosomes. The bearing of the information as to the relative frequency of double crossing-over on the conclusion just stated is discussed in the section on " Purple." To the reader who is not especially concerned with the localization of the genes, we should like to call attention to other subjects of very general interest, such as the discussion of modifiers in the sections on purple, the creams, star, and the second-chromosome modifier of the third-chromosome character, dichaete. Another topic of general inter- est is that of autosomal and balanced lethals discussed in the sections on truncate, streak, confluent, star, dachs-lethal, and lethal Ila. A third topic of interest is that of variations in the amount of crossing- over due to specific genes and to such factors as age and temperature that are discussed in the sections on purple, dachs-lethal, and in the summary dealing with the cross-over variations Cm, and C//r. SPECK (sp). (Text-figures 73 a and b, 75 6, and plate 5, figures 1 and 4) ORIGIN AND STOCK OF SPECK. In the course of the early work upon Drosophila in the Columbia Zoo- logical Laboratory a selection experiment was carried out by Morgan upon a race of wild flies that had showed variation in the extent and dark- ness of the shield or trident pattern upon the thorax. In the fourth gen- eration of selection for a race " without" such a trident, there appeared (March 1910) a few individuals with a tiny black speck (plate 5, fig. 4) at the juncture of each wing with the thorax (Morgan, 1910). At first the breeding results obtained with this character were irregular (Mor- gan, 1910). Some of this irregularity may have been due to non-virgin females (24-hour females were used) and to the practice of using mass cultures, though probably more was due to difficulty of classification before familiarity with the characteristics of the mutation had been acquired. A stock pure for the character was obtained, but was set aside in order that more time might be given to the study of the sex-linked eye-color white which had appeared in April 1910. About a year after this (May 1911) it was found that a stock of flies with a dark body-color called "olive" was pure for a character which was taken to be the same as speck (text-figs. 73 a and 73 6). Accordingly, the first and simpler speck stock was discarded and the olive stock was retained. There is some uncertainty with regard to this "olive" stock, but it seems PLATE 5 1. Speck olive 2. Black 3. "With 5. Streak 4. Speck without" 7. Wild-type 10. Cream II. 9. Eosin E. M. WALLACE Pinx SECOND CHROMOSOME MUTANTS OF DROSOPHILA OF MUTANT CHARACTERS. 129 probable that it was derived from a stock different from the line of trident-pattern selection which gave rise to the "with" and to the "without" stocks. The early selection experiments appear to have resulted in four stocks : (1) " without " trident pattern, (2) "without" pure for the original speck, (3) "with " trident pattern, and (4) "olive " stock. The stocks of ' ' with ' ' and of ' ' olive ' ' differed in the character of the trident pattern in that the "with" is a sharp darkening of the trident pattern, while "olive" is a more diffused darkening. Subsequent analysis of the genetic behavior of these stocks showed that the "with" was a simple third-chromosome semi-dominant, while the "olive" was a trimutant stock homozygous for olive II, olive III, which is a third- chromosome recessive mutant not distinguished in appearance from olive II, and "speck," which is a second-chromosome recessive. An old figure (plate 5, fig. 4) bears out our recollection that the original speck was due to a tiny brush of hairs on the side of the thorax above the wing juncture. The new speck is a black pigment spot in the axil of the wing. The original speck seems not to have been dependable in its behavior, while the new speck has behaved with perfect regularity in inheritance. Thus, both the nature of the characters themselves and the nature of the trident mutants associated with the two "specks" lead to the conclusion that they were not the same mutants, as had been too hastily assumed. The new speck from olive stock will be referred to hereafter as speck, while the original speck will be called the old speck. DESCRIPTION OF SPECK. The character speck, as seen with the magnification with which we ordinarily work (about 15 diameters), is due to the presence of a minute but intensely black round speck in the axil of each wing. The speck is clearly seen from above (fig. 73 a) or from a little below the side (figs. 73 6, 75). Under the microscope the speck is seen to consist of a TEXT-FIGURE 73. — Speck : a, side view ; fc, from above. 130 THE SECOND-CHROMOSOME GROUP heavy deposit of pigment in the lips of a spiracle which is situated a little behind and below the juncture of the base of the wing with the thorax (fig. 736). There is also present upon the side of the scutellum and of the thorax both above and below the wing an added faint tinge of pigment or areola which seems to persist after olive has been eliminated from speck. However, speck is rarely seen without a decided olive color, because the gene for olive is very closely linked to that for speck. Since there is always the chance of crossing-over between speck and olive, and since certain other genes give roughly similar pigmentation, it is well to disregard this pigmentation and classify by means of the speck alone, which is in itself a sufficient index of the speck gene. INHERITANCE AND CHROMOSOME OF SPECK. During the fall and winter of 1912 considerable breeding- work with speck was carried out. The results of these experiments, particu- larly of the F2 from the crosses of speck olive by black, contained the answer to the questions as to which chromosome carried the genes for speck and for olive, but the records were not carefully examined until May 1913. Meanwhile, Sturtevant had begun to work with speck (December 1912) and had shown that it was a complete and invari- able recessive and behaved with regularity in inheritance. TABLE 2. — PI, speck olive 9 X wild cf; FI wild- type 9 + FI wild-type cf. 1912. Dec. — . Wild-type. Speck. Olive. Speck olive. 9 c? 9 rf 9 0" 9 cf la 41 63 44 62 38 92 58 13 68 5 49 73 23 38 46 117 75 11 74 5 2 5 1 5 1 8 5 1 1 1 2 1 4 4 2 2 3 37 20 32 38 11 18 7 10 18 8 20 7 39 25 17 11 2 11 15 6 25 32 21 35 32 49 31 8 28 6 25 27 23 17 22 33 16 4 27 5 16 2a 2b 3a 36 4a 46 5a 56 Total . . 484 511 28 20 199 153 267 199 By means of the 2:1:1:0 ratio in the F2 of the cross of curved to speck he showed that speck was a member of the second-chromosome group. (Jan. 13, 1913). In the crosses carried out by Miss Wallace, speck was found to be readily classified and fully viable. Speck olive females were crossed to wild males and 10 F2 pair cultures raised (table 2). The reciprocal cross (speck cf X wild 9 ) was also made to the extent of 5 F2 cultures OF MUTANT CHARACTERS. 131 (table 3). These two crosses gave identical results as to the distribu- tion of both characters and sex, which proved that no sex-linkage was involved. They may therefore be combined and considered together. The total number of flies produced was 2,857, of which 772 or 27 per cent were speck, which is in agreement with the fact that speck is a simple autosomal recessive. TABLE 3. — PI speck olive cf X wild 9 ; FI mid- type 9+^1 wild-type cf. Wild-type. Speck. Olive. Speck olive. 1912. Dec. — . 9 c? 9 c? 9 if 9 d1 a 1 46 67 4 56 29 24 30 bl 44 37 1 35 27 29 19 b2 55 45 1 1 4 24 17 d2 31 25 5 3 11 6 8 11 dl 107 104 1 2 5 4 42 36 Total.. 283 278 11 7 111 66 127 113 Totals of tables 2 and 3 ... 1,556 66 529 706 The inheritance of " olive" was also followed, but since this will be treated in the following section it need only be stated here that the ratio of wild-type to olive was 9 : 7, which indicates two recessive body- color genes giving similar somatic effects and assorting independently of one another. One of these is olive III, (Ill-chromosome recessive) and the other, olive proper (II chromosome, closely linked to speck) . Speck olive males were crossed to black females, and 5 F2 pah* cul- tures (table 4) and 5 more from the reciprocal (table 5) were raised. The significant fact observed in these crosses was that none of the blacks were speck, which means that speck has its locus in the second chromosome. The combined data (disregarding olive) give 1,098 wild-type, 490 black, 466 speck, 0 black speck, which is an approxi- mation to the 2:1:1:0 ratio expected from such a cross. TABLE 4. — PI, speck olive cf X black 9; FI wild-type 9 -f FI wild-type d*. 1912. Dec. — . Wild-type. Speck. Olive. Speck olive. Black. Black speck. 9 d1 9 c? 9 d1 9 c? 9 d1 9 d1 a 1 34 10 8 5 45 36 17 10 5 43 3 2 1 42 11 14 47 20 14 12 11 32 10 29 8 7 34 30 25 12 10 22 14 34 12 17 18 26 38 9 10 25 8 0 0 0 0 0 0 0 0 0 0 a 2 b 1 b2 . c 1 1 Total. . . 102 111 6 1 134 79 108 83 107 90 0 0 132 THE SECOND-CHROMOSOME GROUP TABLE 5. — PI, speck olive 9 X black d"; FI wild-type 9 + Fi wild-type cf. 1912. Dec. — . Wild-type. Speck. Olive. Speck olive. Black. Black speck. 9 Pr Px 8p Ft; 3535 11,985 5,474 45.7 Vestigial speck 1,446 146 462 520 40 178 36.0 27.4 38.5 1913 Mar. 14 Oct. — 1914 May — Sturtevant . . . Do. Muller Zeit. f. i. A. u. Ver. '15, p. 245. Zeit. f. i. A. u. Ver. '15, p. 287. Am. Nat., '16, p. 422. 2,054 738 35.9 Curved speck.. 1,007 223 462 262 71 150 26.0 31.8 32 5 1913 Mar. 10 Oct. 16 1914 May — Sturtevant. . . Do. Muller Zeit. f. i. A. u. Ver. '15, p. 245. Zeit. f. i. A. u. Ver. '15, p. 247. Am. Nat., '16, p. 422. 952 266 27 9 Aug. 24 Bridges 3j>; Pr c sp B.C. ; 452-508. 1915 S' 632 ,062 226 30.5 35.7 July li 1916 Feb. 8 Do. Do. ; B.(_/. nrsts; 1836- 94. Pr csP liiai PrC8p Ft; 3203-8. 10,042 3,037 30.5 Plexus speck . . 327 29 8.9 Feb. 29 Do. Ilia'- PT Px 8v Ft; 3535 Arc speck Blistered speck Speck balloon 2,625 36 462 156 3 2 5.9 8.3 0 4 1914 Oct. 24 Feb. 27 May — Do. Do. Muller a; pr a sv balanced B.C. ; 637-686. *>» 6,; ~L— B.C.; 72.... «» Am. Nat., '16, p. 422. 134 THE SECOND-CHROMOSOME GROUP finished, very large amounts of additional accurate data have been collected upon the cross-over relations of speck with other second- chromosome genes. These data appear in the tables of the sections following this (see especially star). Only one of these later experiments had as the main object the more exact determination of the locus of speck. This experiment (table 7) was a triple back-cross for the three loci — purple, curved, and speck — and gave a curved speck cross-over value of 27.9 per cent on the basis of the 952 flies, of which 266 were cross-overs between curved and speck. The base of reference in the determination and mapping of the locus of speck is curved, which is the nearest locus accurately mapped in relation to black, the primary base of the entire second chromosome. For the sake of convenience a summary of all the cross-over data in which speck is one of the loci involved is given in table 8. In calcu- lating the locus of any mutant one must consider not only this direct- linkage data, but also the whole mass of data on the other loci of the same chromosome, and especially the information upon the amount of double crossing-over and coincidence in the various regions. By this method speck has been mapped at a locus 31.6 units to the right of curved, which is its immediate base of reference, or, referring back to star as the zero-point, speck is at 105.1. VALUATION OF SPECK. Speck is at present one of the most generally useful and used of the second-chromosome mutants, first, because of the perfect accuracy, ease, and speed with which the recessive character is separable from wild-type; secondly, because it can be used in experiments with any of the other second-chromosome characters (including black) without masking effects or confusion in the classification; thirdly, on account of the value of the position of its gene near the right-hand end of the second chromosome, speck being by far the most workable mutant in that general region; and finally, because its viability, its productivity, and its fertility are above reproach, and it is singularly free from such bad habits as getting drowned, or stuck in the food, or refusing to be emptied from the culture bottle, etc., which alienate the affections of the experimenter from certain other mutants. Speck is to be com- mended for students' use, but care should be taken that the character is clearly recognized. LITERATURE OF SPECK. The more important papers referring to speck are: Morgan, 1910, describing its origin and giving the irregular breeding results already commented upon above; Sturtevant, 1915, giving the data of table 6, by means of which the locus of speck was first worked out; and Mailer, 1916, speck being one of the mutants used in the progeny test of the linkage of second-chromosome genes. OF MUTANT CHARACTERS. 135 OLIVE. ORIGIN AND STOCK OF OLIVE. The origin and early history of the stock called "olive," because of the body-color present, is only partly a matter of record, though the account given in the section on speck is substantially correct. To repeat: In the fall of 1909, Morgan started selection on stocks of wild flies that were throwing individuals with extra-dark trident patterns. One line of selection eliminated this variation, and the resulting "without" stock represented the original wild stock before the occurrence of the "with" mutation. It was in this "without" stock that the old speck was found. The selection in the opposite direction isolated the third- chromosome semi-dominant mutation "with" (plate 5, fig. 3). The "olive" stock appears to have resulted from selection carried out on a stock different from that which gave rise to "with." The trident pat- tern of the "olive" stock is dark, but is not distinct, being submerged in a general olive color that suffuses the thorax. The "olive" stock was obtained about May 1910, and had been kept in the stock room about a year when it was noticed that it was pure for speck. CHROMOSOME AND LOCUS OF OLIVE. During the fall and winter of 1912 several crosses were carried out by Miss Wallace with the " olive " stock. " Olive " crossed to wild gave wild-type offspring which were inbred in pairs to give F2. This cross and the reciprocal were both made and gave the same kind of FI and F2 results in the distribution of both characters and sex, showing that no important sex-linked modifiers of body-color were present. The com- bined F2 counts (2,857 flies) gave a very perfect 9 : 7 ratio of gray to "olive" (1,622 : 1,235), showing that the "olive" stock contained two recessive dark body-colors whose genes assorted independently (tables 2 and 3) . A further separation of the " olive " classes into the component two single and the double recessive forms was not attempted. That one of the recessives (olive II) was carried in the second chromosome was proved by the strong linkage olive showed with speck (tables 2 and 3). There were only 2.3 per cent of flies that were speck not-olive. The whole observed distribution corresponds to about 14.3 per cent of crossing-over between speck and olive. It is certain that this value is far too high due to the difficulty in classifying the olive. Our later experience has been that olive is probably less than a unit distance from speck, and probably to the right, which give an approximate locus of 106 when referred to star. Very rarely have we secured speck flies that we consider free from olive. The original stock of speck was probably not-olive, and in a certain experiment made by Sturtevant (1915) it seems likely that the 136 THE SECOND-CHROMOSOME GROUP great difficulty of classification was due to a speck from which olive had been lost by crossing-over. On one other occasion speck has been found which seemed to be without olive. We believe that most of the few flies classified as " speck not-olive" in the F2 of olive by wild were either young flies in which the olive was not yet developed sufficiently to be surely classifiable or were fluctuants extreme enough to cause trouble, neither of the two olives being sufficiently dark or constant in color to be invariably separable from the wild type. Throughout our discussion of speck and olive it has been assumed that the olive color which is nearly always seen in speck flies is due to a separate gene, and while this is probable, the evidence is by no means conclusive. It may be that the two characters are the products of a single gene, and that the supposed cases of "speck not-olive" have been due on occasion to wrong classification of the poor character olive, or to the action of minus modifying genes, or to a new speck allelomorph which differed in this regard from the old. A further careful investigation would be required to settle this question. VALUATION OF OLIVE. Olive II is of no value for further \vork, since the character is not sufficiently distinct from wild-type so that the normal fluctuations in these two body-colors do not overlap, and classification is accordingly both difficult and inaccurate. Olive is subject to a general objection to the usefulness of all faint body-colors, namely, the fact that body- colors take some tune after hatching for their full development, and in cases where the final difference is not great the intermediate stages are unpleasant to work with and a source of error. A further defect in the value of olive is the uncertainty as to whether the character may not be found to be only another expression of the speck gene. Further- more, in case the investigation should prove that two linked but sep- arable genes were involved, the usefulness of olive would not thereby be improved, since olive could be used for no purpose that could not be better met by the use of speck or one or another of the mutations whose genes are in this same region. As far as we have been able to observe, the presence of olive has had no detrimental effect upon the viability or other qualities of speck. TRUNCATE (T). (Plate 6, figures 1 to 6.) ORIGIN OF TRUNCATE. One of the early mutations (beaded, May 1910) had been run for seven generations in stock cultures when a fly appeared (August 1910) whose wings were somewhat obliquely truncated and somewhat PLATE 6 OF MUTANT CHARACTERS. 137 shorter than usual (Morgan, 1911). This fly when bred to his wild-type sisters produced about 10 per cent of offspring whose wings showed a slight or moderate amount of truncation (plate 6, figs. 4 and 5). Some of these truncated flies bred together produced in F2 nearly 50 per cent of truncated. For several months it was impossible by selection to raise the percentage of truncates much above 50 per cent, though there was an increase in the shortness of the wings of the truncates that did occur. But later certain cultures gave much higher percentages, and selection started at this point and continued for about a year estab- lished a stock in which the percentage of truncates was not far from 90 per cent. Above this level it was impossible to maintain gains. Along with this increase in the percentage of truncate individuals, several other changes were observed to be going on. There was an increase in the number of flies which were sterile and gave no offspring; at times about 50 per cent of the pairs were sterile. There was a marked decrease in the productivity of those pairs which did give offspring. These two facts made the stock very difficult to carry on, except in large mass cultures. In those cultures in which the per- centage of truncates was high, the amount of the truncation was great ; i. e., many of the flies had extra-short wings, some wings being even shorter than the abdomen (plate 6, figs. 1 and 2). It was found that if these short-winged flies were used in carrying on the stock, the per- centage of truncates was higher. The flies whose wings were slightly truncated gave under 90 per cent of truncated, while those flies whose wings showed no truncation either gave no truncates or less than or about 50 per cent of truncates. In all of these cultures more of the females than of the males showed the truncate character; that is, truncate is "partially sex-limited" in its expression. At this point Mr. E. Altenburg took up the selection of truncate and confirmed the points just stated. Later, Altenburg and Muller (Mech- anism of Mendelian Inheritances; also mss.) sought to analyze the truncate stock by the method of linkage to find the cause of the con- tinued appearance of normal flies and of the variations in the extent of the truncation. Their tests showed that truncate is primarily due to a dominant gene in the second chromosome. This gene is lethal when homozygous, so that pure-breeding stock can not be obtained. When heterozygous, this gene is capable of producing only a moderate trun- cation and fluctuates in expression, so that most of the flies having the truncate gene fail to show the character at all. But in the selected stock there are certainly two (one in the first and one in the third) and probably more genes whose effect is to increase the amount of trunca- tion, whereby both a greater proportion can be detected and the wings of those which do show truncation are shortened. Such intensifiers could theoretically only raise the amount of trunca- tion to the extent that every fly carrying the truncate gene can be 138 THE SECOND-CHROMOSOME GROUP distinguished from the wild type, and because of the lethal action of the truncate gene this can not exceed two-thirds of the flies. If some of these modifiers could themselves produce a moderate truncation, then the percentage would pass beyond 67, but there is evidence that the modifiers are unable to produce truncation in the absence of the chief gene. TRUNCATE LETHAL. Since at least one-third of the flies must look like wild flies rather than truncates, their non-appearance in the highly selected stock is to be accounted for by the assumption of a lethal gene which is carried in the II chromosome, homologous to that carrying the truncate gene. This condition has not been proved by testing, but the action of autosomal lethals is well understood and a similar case of an auto- somal lethal giving an apparently pure-breeding stock when balanced against an autosomal dominant which is itself lethal when homozygous (beaded) has been demonstrated by Muller. It is now clear what is the meaning of the early history of the truncate stock which was for so long a puzzle. The original truncate fly was heterozygous for the dominant truncate gene. In a cross to wild-type sisters only 10 per cent of the offspring showed the truncate character, though half of them carried the truncate gene. This low power of expression of the character was due to the lack of effective modifiers, most of which were recessive. The first period of inbreeding and selection resulted in the collect- ing of whatever modifiers were present, and by the approach to homo- zygosis allowed more of the recessive ones to show their effect. But by such means the proportion of truncate could not be advanced beyond 67 per cent and was not advanced much beyond 50 per cent. The limit of 67 per cent imposed by the lethal nature of the primary truncate gene must have been broken down by the occurrence (Febru- ary 1911) of the lethal mutation in the chromosome homologous to that carrying the primary truncate gene. The second period of selec- tion then increased the percentage of truncates by increasing the preva- lence of this lethal in the stock. The limit which this second selection approached was the condition in which every second chromosome which was not carrying the truncate gene should carry the lethal. By this means all the non-truncate flies would be eliminated except those saved by crossing-over. If the lethal were 20 units away from trun- cate, measured along the second chromosome, then 10 per cent of not- truncate flies should survive. The very low productivity of the truncate stock observed at this time was the result of the action of the two lethals — all homozygous truncates as well as all homozygous lethals dying as early zygotes; also the modifiers tending to produce infertility when homozygous. While the proportion of truncates was OF MUTANT CHARACTERS. 139 limited at the two values, 67 and 90 per cent, the accumulation of the modifiers continued to make the wings shorter, and the shortest indi- viduals gave the highest percentage of truncates, since they were on the average the ones carrying most modifiers. The increase in the number of completely sterile individuals was independent of the other changes and due to a mutation which occurred early in the selection. Hyde (1914) was able to eliminate this sterility from the truncate stock by breeding for some generations from those families which showed least sterility. But there was no rise in pro- ductivity parallel to the elimination of the sterility. This sterility behaved as a recessive in the FI out-crosses, and in F2 reappeared, but in such proportion and distribution as to suggest that it also was complex. The fact that the main gene for truncate is in the second chromosome was established by Muller and Altenburg through the non-occurrence of cross-overs in back-cross tests of males heterozygous for truncate and black. In back-cross tests of similar females there was somewhat under 25 per cent of crossing-over. A back-cross test showed that there is somewhat over 25 per cent of crossing-over between star and truncate, and this information, in connection with the known distance of about 50 between star and black, showed that truncate is located about midway between star and black and slightly nearer to black. Since the amount of data secured in these tests is not large and because truncate is so elusive a character, the location is not precise, though the locus is probably not far from 28 (see Snub). TRUNCATE REOCCURRENCES. Many of our mutant characters have made their appearance more than once, and occasionally under circumstances which make it certain that there has been a new occurrence of the same mutative process that was responsible for the original appearance of the character. In other cases it is probable that the character is not reappearing because of a fresh mutation, but that the original mutant gene had been intro- duced in some previous cross and has been unsuspectedly present for several generations but has been unable to appear because of the way in which the crossing has been carried on. In still other cases a char- acter appears which resembles very closely another already known character, but the two are the result of mutations in entirely different loci which are often in different chromosomes. Thus, we have at least half a dozen mutant characters of the "genus" pink which are so similar as to be practically indistinguishable in mixtures, yet which are dependent on entirely distinct genes. Characters of the genus truncate are the most frequently recurring of any, with the possible exception of the beadeds. These two kinds of character have both come up in the breeding work three or four times each year. 140 THE SECOND-CHROMOSOME GROUP In most of these cases the character is also "specifically" truncate, and usually due to the original truncate gene rather than to a fresh mutation. In the absence of its usual intensifies truncate may lurk unsuspected in a stock or an experiment for many generations and is difficult to eliminate. For this reason it is practically impossible to be certain in any unexpected case of truncate appearance that there has been a fresh mutation. It is not ordinarily profitable to pay any attention to these appearances of truncate, but in two instances in which, because of the pedigree, there was less than the usual likelihood that the truncate was due to the original gene, tests were made, and in both cases the character was found to be either truncate or else a very similar allelomorph, though these tests could throw no new light on the question of whether the gene were the original or a fresh truncate mutation. SNUB. The first of these tests was made by Muller (unpublished). The second case appeared in the ninth generation of some experiments on "duplication." A cross had been made between a female with the new sex-linked recessive wing-character "cut" and a not-cut male (February 17, 1916, culture 3338, Bridges). All the daughters were expected to be wild-type and all the sons cut; but 9 of the 77 daughters were seen to be slightly truncated, the character being called "snub," while some 18 of the 67 sons were cut with shortened and blunted wing ends (cut snub). One of these cut-snub males outcrossed to a wild female gave about a quarter of the FI flies with the snub or truncate character. This result showed that the character was a dominant, though a "poor" one; that is, not all the flies genetically alike (heterozygotes) showed the character somatically. The snub appeared among the FI males as well as among the females, and this fact showed that the character was non-sex-linked, for had it been sex-linked it could have appeared only among the daughters of the above cross. When snub flies were bred together, the result was usually an approxi- mation to a 2 snub : 1 not-snub ratio, often with the snubs below expec- tation because of the above-noted occasional failure of heterozygous snubs to show the character somatically. This ratio and the fact that it proved impossible to obtain a pure-breeding snub stock suggested that the mutant was lethal when homozygous, as are most of our dominants. In cut-snub stock the approximation to the 2 : 1 ratios was much closer, and it seems certain that the character cut favors the differentiation of snub (see cultures 1 to 10, table 9). We are well acquainted with such intensifiers or modifiers in other cases, and truncate itself was known to be very susceptible to intensification. A few of the cut-snub pairs gave nearly all of the flies snub (see espe- cially 10 and 12, table 9, Morgan), and it seems probable that in these OF MUTANT CHARACTERS. 141 TABLE 9. — The offspring given by pairs of cut snub flies from the stock of cut snub. cases a lethal was present in the homologous chromosome. Autosomal lethals are very plentiful and have been clearly demonstrated in many other cases, though in this case no further test has yet been made of the correctness of this explanation. That the same kind of modifiers were present in the snub stock as were resopnsible for the short-winged types of truncate appeared certain from the rather sharp differences between the cut snubs which occurred in the inbred stock. While most of the cut snubs were of the type described, in which the wing is nearly as long as the normal cut but differently shaped at the tip, certain ones were much shorter and the oblique truncation was very marked. These shorter ones, just as in the case of the selected truncate stock, were most numerous in those cultures in which the expected 2 : 1 ratio was most closely approached. Pair 10 of table 9 showed a ratio of 31 short snubs to 63 of medium or slight truncation to 46 which showed no truncation; this rather close approach to a 1:2:1 ratio suggested that probably only one such modifier was present in the cross. All of the characteristics of snub thus far found have agreed exactly with those of "specific" truncate. If it should be found that the chromosome locus of snub were the same as for the original truncate, then we should conclude that the mutation is specifi- cally truncate. Because of the fluctuating nature of the dominance of snub and its easy modification, a direct linkage experiment offered diffi- culties. A more exact method would be to establish the lethal nature of the truncate-snub compound. This could be done by showing that the FI ratio obtained by crossing truncate by snub was a derivative of a 2:1 instead of a 3 : 1 ratio. The observed ratio of 174 truncate to 132 not-truncate in the FI from this cross would somewhat favor the view that the ratio is 2 : 1 rather than 3:1, and that snub is therefore truncate (table 10, Morgan). But here again the uncertainty that the number actually showing the truncate character would be a close enough approach to the number heterozygous for truncate, so that we could decide whether we were really dealing with a 2 : 1 or a 3 : 1 ratio made the results of such experiments of doubtful value. It was recalled that cut had acted as an intensifier of snub, so that a larger proportion of cut flies showed the snub character than was the case among the not-cut flies. Advantage of this fact was taken by Culture No. Cut snub. Cut. 1 30 18 2 33 16 3 33 15 4 24 13 5 30 15 6 13 10 7 45 13 8 24 15 9 51 25 10 94 46 377 186 11 27 2 12 25 2 52 4 142 THE SECOND-CHROMOSOME GROUP crossing a cut-snub female to truncate males, in which case all the sons were cut. Among these cut sons the ratio of snub to normal was almost exactly 2 : 1 (table 11). The fact that this experiment gave a close approach to the ratio expected if the truncate-snub compound is lethal could not be accepted as proving that theory, because there might still be enough flies failing to show the truncation, so that the ratio is really the normal 3 : 1 ratio. TABLE 10. — F\ ratios obtained from crosses of truncate 9 X snub cf (cultures Ito4)> and snub 9 X truncate d* (cultures 5 and 6). Culture. Truncate- Not trun- No. like. cate-like. 1 11 8 2 51 45 3 50 48 4 5 12 5 43 16 6 14 3 Total 174 132 TABLE 11. — FI ratios obtained from crosses of cut snub 9 X truncate cf1 (pairs). 1917 Jan. Wild- type 9- Truncate- like 9- Cut d*. Truncated- cut cf. A 48 57 38 56 B 39 66 25 62 C 34 71 29 59 D 22 59 31 66 E 28 73 23 . 56 F 39 39 16 35 Total 210 365 162 333 The scheme finally followed eliminated all classification of both truncates and snubs and depended for identification upon the readily classifiable "character star and upon the assumed lethal nature of the truncate-snub compound. By mating star to truncate, flies can be obtained carrying the star gene in one II-chromosome and truncate in (S' 4- the other V+ Two such flies mated together would give a 2 star to 1 not-star ratio, unless homozygous truncate were lethal. But since homozygous truncate is known to be lethal, this ratio becomes modified by the death of most of the not-star flies. A few not- star offspring will survive because of crossing-over in the female S' r r There is a precise relation between the amount of this crossing-over and the number of not-star flies which OF MUTANT CHARACTERS. 143 TABLE 12. Culture No. Stars. Not- stars. 1 199 27 2 264 21 3 300 84 4 187 17 Total 950 149 appear. If x is the percentage of cross-over gametes, and lOOz of the non-cross-over gametes, then the ratio of star to not-star is 200z :x. Obviously, if snub is truncate, then by mating such a star-truncate heterozygote by a similar star-snub heterozygote, one should get this same 200x : x ratio of star to not-star. That is, the occurrence of a ratio of star to not-stars in which the not-stars are less numerous than in the ordinary 2 : 1 ratio would mean that snub is truncate, and by calculation from the observed ratio we can find out how far the locus of truncate is from star. The experiment as carried out gave (table 12, Wallace) four cul- tures in which the ratio was significantly different from 2:1. The total of stars was 950 and of not-stars 149, or a ratio of 6.3 : 1. From this we may conclude that snub is truncate or an allelomorph so similar in its obvious characteristics that to demonstrate a difference would require a refined biometrical study or elaborate 'interaction' tests. In table 12 will be found the ratio of star to not-star given by crosses of females heterozygous for star and truncateto males heterozygous for star and snub. The solution of the proportion 200x : x : 950 : 149 gives a value of 27.1 for x; that is, there is 27.1 per cent of crossing-over between star and truncate. Nearly 1 per cent should be allowed for double cross- ing-over within the distance from star to truncate, or the map distance should be about 28.0 on the basis of this data. This position also agrees with what is known of the location of truncate. INTENSIFICATION OF TRUNCATE BY CUT. That cut intensifies the original truncate in the same manner as it does snub is shown by the results of a test of this point (table 13). When a cut female was crossed to a truncate male the ratio of not- truncate to truncate among the males was 1: 0.79; that is, quite a TABLE 13. — Intensification of truncate by cut — Pi, cut 9 X truncate cf . 1917, Jan. Wild- type 9 . Trun- cate ? . Cut d". Truncated cut cf. I II Total . . 81 66 44 15 61 44 42 41 147 59 105 83 close approach to the expected 1:1; but among the sisters, which were not cut, this ratio was 1 : 0.40, which means that about 43 per cent of the females genetically truncate failed to show the character. The intensification by cut is more extensive than appears at first Wild-type "Slight" " Long' ' "Intermediate* ' . "Short"... 37 96 62 164 52 104 46 50 146 63 THE SECOND-CHROMOSOME GROUP glance, because normally the males show a considerably smaller per- centage of truncate than do the females. The difference in culture II was especially striking. A recent census of the truncate stock, TABLE 14. which has been run for about 5 years under selection not especially rigorous, showed about 17 per cent of wild-type flies (table 14). The truncate flies were of various degrees of shortness, of which the most frequent was that known as "intermediate" (correspond- ing to fig. 4 of Plate 6). The very short truncates were not especially numerous, although in selecting the parents each generation they had been preferred. Table 14 gives the census of truncate stock (May 1917). BLACK (*). (Plate 5, figure 2.) ORIGIN OF BLACK. The first workable body-color mutation, black, was found by Morgan, October 1910, in the F2 of a cross between miniature and wild flies (Morgan, 1911). DESCRIPTION OF BLACK. When black flies are freshly hatched little black color has developed on the body, though the legs and feet are darker than normal. Within a few minutes after hatching the color has deepened so that the head, thorax, and abdomen are a clean, fresh, greenish black, more intense on the thorax than on other parts. This color becomes progressively darker with age. The wings, after expanding, also become much darker, and along each side of the veins a broad band of pigment begins to develop and becomes conspicuous in old flies. SEMI-DOMINANCE OF BLACK. While the fly heterozygous for black is noticeably darker than the wild-type, this separation can not be made completely, although it is occasionally made use of for special purposes (see sections on Jaunty, p. 162, and Apterous, p. 237). Black is generally, therefore, treated as a recessive, and the separation of black from the heterozygote is easy and entirely accurate. CHROMOSOME CARRYING BLACK. Black is a member of the second-chromosome group by definition. As soon as it had been established that the loci for the sex-linked mutations were capable of being mapped in definite positions (Sturte- OF MUTANT CHARACTERS. 145 vant, 1912), this same procedure was applied to the non-sex-linked mutations. The cytological work of Miss Stevens had shown that there were at least three autosomes in Drosophila, and it was expected that a group of linked genes would be found to correspond to each of these. At this time there were only two non-sex-linked mutations — black and pink — whose inheritance had been worked out by Morgan and whose behavior was definitely known to be Mendelian and normal. The next point was to establish the relation of these mutants to each other in inheritance. This was done by raising an F2 from the cross of black by pink. The F2 ratio was quite clearly that of independent inheritance, since it approximated 9 : 3 : 3 : 1, with the double reces- sive present in as large numbers as expected (Sturtevant, February 1, 1912). Sturtevant soon showed by means of back-cross tests that there was no appreciable linkage between black and pink. Provision- ally, black and pink were assumed to be in separate chromosomes — the second and the third. The second chromosome is arbitrarily that chromosome which carries the gene for black, and any other genes that may be found to be linked to black; similarly, the third chromosome is defined as that chromosome which carries the gene for pink, and all other genes found to be members of the linkage group containing pink. Soon after the black pink F2 had given the first autosomal inde- pendence, an F2 between black and curved demonstrated the first autosomal linkage. As soon as this linkage was observed (March 4, 1912) definite plans were made to test the linkage relations (chromo- some and locus) of all the autosomal mutants thus far found, making full use of the back-cross method. (See Bridges and Sturtevant, 1914, p. 205). By the middle of July it was known as a result of these tests that besides black and curved, purple, vestigial, balloon, blistered, jaunty, and arc were in this second chromosome. LOCUS OF BLACK. The locus of black was taken as the base of reference in the mapping of these other genes. Since curved was the first mutant known to be in the second chromosome with black, its locus determined the direction along the chromosome which was to be defined as " to the right' ' (black curved). The loci of all the other mutants just mentioned were later found to lie on the same side of black as curved does, so that black was the locus farthest to the left, and the natural zero-point of the map. Black is now the real base of reference in the mapping of the entire second chromosome, and all other genes are plotted in relation to it, either directly in the case of those genes nearby (dachs, jaunty, purple, vestigial, etc.), or indirectly by being located with reference to certain important genes (star, curved, speck, etc.), whose positions with regard to black have been so well established that they in turn can safely be used as secondary bases. 146 THE SECOND-CHROMOSOME GROUP TABLE 15. — Summary of the cross-over data involving black. I Loci. Total. Cross- overs. Per- cent. Date. By — Reference. Star black .... Streak black . . Dacha black . . Squat black. . . Black jaunty . . Black purple . . Black-?/ /a 1,352 496 865 690 13,104 522 203 315 266 4,944 38.6 40.9 36.4 38.6 37.7 Jan. 11, 1915 Oct. 23, 1915 Oct. 26, 1915 Dec. 22, 1915 Dec. 5, 1916 May — , 1914 Mar. 18, 1913 June 30, 1913 Dec. 10, 1913 May — , 1914 Apr. 3. 1911 May — , 1914 Dec. 12, 1912 Aug. 28, 1913 Jan. 9, 1914 May — , 1914 Jan. 5, 1915 Mar. 5, 1917 Jan. 15, 1916 Sept. 10, 1912 Sept. 10, 1912 June 8, 1913 Dec. 10, 1913 Jan. 9, 1914 May — , 1914 Nov. 17, 1915 Mar. 5, 1917 May — , 1917 Bridges .... Do. Do. Do. Plough. .. Muller Bridges .... Do. Do. Muller Bridges. . . . Muller Bridges .... Do. Do. Muller S' Px'i h B.C.; 1921-'24. o px S' fr; , ;• B.C. ; 2282- 84. o Jr „ S' va • K n B.C. ; table 123, this paper. S' di 0 J. E. Z., '17, p. 147; temperature; S' , B.C.; tables 7 (22°), 8» (27°), 8. (22°), Hi (22°), 17,. Am. Nat., '16, p. 422. d; d b Ft; II 34-36. d/d&B.C.; II 40-11 98r. d;d bvg balanced B.C.; II 114-11138. Am. Nat. '16, p. 422. Sq 16,507 6,250 37.9 462 120 26.0 338 933 4,892 462 82 163 874 77 24.3 17.5 17.9 16.7 6,725 1,196 17.8 82 462 9 1? 11.0 0.2? &q; i U.C. 40*4. b Px Am. Nat. '16, p. 422. pr; b pr B.C.; C 174-11 2. pr; bprc B.C.; Ists; II 58-11 88. pr;bprvg balanced B.C. ; 11141-674. Am. Nat. '16, p. 422. J. E. Z., '17, total 6 pT c B.C. J. E. Z., '17, total b pr vg B.C. hi a: b Ina F2;l2840, '59, '63. b Biol. Bull. '14, p. 197; B.C. 773 3,934 5,001 462 36,622 2,139 38 212 322 26 2,214 214 4.9 5.4 6.4 5.6 6.0 10.0 Plough Do. Bridges.. . . Morgan . . . Do. Sturtevant . Bridges .... Do. Muller Bridges .... Plough Gostenhofer 48,931 3,026 6.2 166 22 13.0 Black vestigial. 3,499 1,268 694 4,892 5,001 462 450 2,139 1,748 632 217 169 806 815 78 99 477 285 18.1 17.1 24.4 16.5 16.3 16.9 22.0 22.3 16.3 Vg Biol. Bull. '14, p. 198; 6 vg B.C. Zeit. f. i. A. u. V.'IS, p. 286; b v, c balanced B.C. d; d b va balanced B.C.; II 114- II 138. pr; b pf vg balanced B.C. ; II 141-674 Am. Nat. '16, p. 422. S' Von; i n B.C. ; table 123, this paper. O Vg J. E. Z., '17, b pfVgB.C. b B.C.; students' records. Vg 20,153 3.578 17.8 OF MUTANT CHARACTERS. 147 TABLE 15. — Summary of cross-over data involving black — continued. Loci. Total. Cross- overs. Per- cent Date. By- Reference. Black curved . . Black plexus . . Black fringed . Black arc 7,419 260 253 402 3,934 223 462 36,622 13,104 1,717 69 66 120 839 63 106 8,598 2,659 23.1 25.5 26.1 29.9 21.3 28.2 22.9 . 23.4 20.3 Jan. 13, 1913 Jan. 13, 1913 Jan. 13, 1913 June 8, 1913 Aug. 24, 1913 Oct. 16, 1913 May — . 1914 Jan. 5, 1915 Dec. 13, 1915 Jan. 1, 1915 July 20, 1915 Apr. 3, 1916 Oct. 23, 1915 Dec. 13, 1912 Aug. 4, 1914 Nov. 4, 1913 Sept. 23, 1914 Oct. 16, 1913 May — , 1914 Mar. 29, 1913 May — , 1914 Sept. 28, 1913 Aug. 4, 1914 B. & Stutt. Do. Do. Sturtevant . Bridges .... Sturtevant . Muller Plough Plough Bridges .... Do. Do. Bridges .... Bridges Do. Bridges. . . . Do. Sturtevant . Muller Sturtevant . Muller .... Bridges Do. b Biol. Bull. '14, p. 209; b c and — B.C. Biol. Bull. '14, p. 208; 6 c F2. b Biol. Bull. '14, p. 212, — B.C. Zeit. f. I. A. u. V.; '15, p. 247; b rig c B.C. pr; b prc B.C.; Ists; II58-II88. Zeit. f. I. A. u. V., '15, p. 247, bos, B.C. Am. Nat. '16, p. 422. J. E. Z. '17, 6 j>r c B.C. totals. S' J. E. Z. '17 , B.C. controls, o c px;bpxB.C.; 1084-'99. S' 62,679 14,237 22.7 1,026 1,352 82 417 576 38 40.6 42.6 46.4 px; — / "• B.C.; 1921— '24. S, «' 6PxBlC-;4( S' 2,460 1,031 41.9 496 798 6,794 211 286 2,951 42.5 35.9 43.4 Sri — r' "7 B.C. ; 2282— '84. 0 Jr a; ba B.C. ;C 172-113. mr; b a mr balanced B.C.; 364 . b bt; -5- B.C.; II 102 . B* Pinkish; b pinkish B.C.; 525-2426. Zeit. f. I. A. u. V. '15, p. 247, b c sp B.C. Am. Nat. '16, p. 422. Zeit. f. I. A. u. V. '15, p. 276; B.C. mr; b mr B.C.; II 93-11 96. mT; b a mT balanced B.C. ; 364. Black blistered Black pinkish . Black speck . . . Black balloon . Black morula . 7,592 3,237 42.6 224 736 223 462 93 371 110 216 41.5 51.4 49.3 46.8 685 326 47.6 1,774 462 857 216 48.3 46.8 2,236 1,073 48.1 755 6,794 353 3,165 46.1 46.6 7,549 3,518 46.6 The zero position in the second chromosome has been delegated to star, which is found to be 46.5 units to the left of black, that is, black is at an approximate position of 46.5 on the map as recast with star as the zero-point. Table 15 gives a summary of all the cross-over data previously published as well as that given in other sections of this paper. 148 THE SECOND-CHROMOSOME GROUP "BROWN" BODY-COLOR. Soon after the discovery of black, there appeared in the black stock a few males whose color is a rich "brown" instead of being a clear, cold, greenish black. These flies were in reality a double recessive, as was shown by the F2 results from the out-crossing of these males to wild females. The other recessive turned out to be "yellow," which is sex-linked. An astounding number of flies (hundreds of thousands) were raised (by Morgan, Wallace, Bridges, and Eleth Cattell) in work- ing out the simple relation of brown to black, to yellow, and to the wild form. A similar interest was shown in the relations of vermilion and pink (the double recessive being called "orange"), these relations being then regarded as highly important from the standpoint of the presence-and-absence theory and the seriation of characters. VALUATION. Black is a mutation of first rank in value, and has been used more extensively than any other autosomal character. Its viability is excellent. With a little practice black can be separated from the heterozygous form with perfect accuracy. There are no other body- color mutations in the second-chromosome that interfere with the classification of black, and in turn black can be used in experiments with any of them (including speck and streak) without masking effects or confusion. Black has been extensively used in .class work in genetics by beginners; in this case the only caution necessary is in the classification of very young flies, since the full black color is slow to develop. Even experienced workers occasionally put back into the culture-bottle the^very young flies (in a cornucopia, if etherized), and then classify them when they are again taken out at the next counting. BALLOON (Plate 7, figure 1.) The mutant character balloon was found by Morgan (November 1910) in a stock culture of truncate flies (Morgan, 1911). This char- acter was first noticed from the fact that certain flies not long emerged from the pupa-case had their wings pumped full of liquid, the two laminae of the wing being separated, except at the edges, to form a balloon (plate 7, fig. 1). As these flies became older these vesicles usually broke or the liquid was resorbed, so that the laminae came together, giving an uneven or blistered appearance to the wing. These wings were held out at a wide angle from the body (plate 7, fig. 1) and this character forms the most quickly recognizable mark of identifi- cation in the separations. The divergence of the wings serves well for a quick and rough preliminary separation, though a more reliable PLATE 7 OF MUTANT CHARACTERS. 149 character is the presence of extra veins in the wings, which should be looked for in checking up whether any balloon flies with only slight divergence have been passed over in the preliminary separation. These extra veins are most plentiful as a plexus about the posterior cross-vein and also between the marginal and second longitudinal veins (plate 7, fig. 1) . The balloon wing is usually considerably smaller than normal and is of a brownish uneven color and of markedly chiti- nous appearance. The balloon flies run about very actively and take short, quick jumps, but are unable to fly. Since balloon appeared in truncate stock it partook of the sterility of truncate. By out-crossing and extraction a stock was obtained which was fully fertile and which was free from truncate. CHROMOSOME OF BALLOON. That the gene for balloon is hi the second chromosome was shown by Sturtevant (April 20, 1912) by means of a cross to curved, which gave in F2 the 2:1 : 1 : 0 ratio typical in Drosophila for experiments with genes in different homologues of the same chromosome pair. LOCUS OF BALLOON. Sturtevant also found that the locus of balloon is very far away from that for black (black balloon cross-over value 48.3), and is in fact in the right-hand end hi the same region with speck. Speck and bal- loon were found to be so close together that all attempts to synthesize the double recessive, speck balloon, failed. Without this double recessive it was impossible to run a back-cross test which would have told on which side of speck the balloon gene is situated and exactly how far distant. By the laborious method of testing individually the offspring of females heterozygous for speck and balloon f — J-r- ) Muller (1916) found that two1 individuals out of a total of 462 represented crossing over between speck and balloon. Balloon is therefore about 0.4 unit away from speck, and to the right as was shown by the other second- chromosome characters tested at the same time. On account of its marked variability, the character balloon has been used by Marshall and Muller (1917) in a study of the question of the contamination of "a gene by its allelomorph when the two are present in the heterozygote. By back-crossing in each generation a male heterozygous for balloon, and for certain other characters used as indexes, to a female which has these index characters but is free from balloon, a stock was carried on for some 50 generations (nearly 3 years), 1 These two cross-overs were inadvertently omitted from the table of page 422 (Muller 1916) and from his summary on page 423. 150 THE SECOND-CHROMOSOME GROUP during which time the balloon gene was constantly maintained in heterozygous condition. If the effect of the not-balloon gene always present hi the homologous chromosome were to render the balloon gene less characteristically balloon-producing, then the balloon stock finally extracted from this long-continued heterozygosis should exhibit a lower grade of the balloon character than that shown by the regular stock of balloon which for some 5 years had been kept homozygous by inbreeding. When the average grade of the individuals of a stock freshly extracted from this heterozygous condition was determined and compared with the like grade determined for the homozygous stock, it was found that the difference from normal of the outcrossed type was not less than the difference of the inbred stock. A comparison of the standard deviations of these two stocks showed that there had been no increase in variability on account of the continued heterozy- gosis. These facts together showed that, in an adequately tested case of character variability, contamination of genes was not operative to a detectable degree. A summary of the linkage data involving balloon is given in table 16. TABLE 16. — Summary of data upon linkage of balloon with other second- chromosome loci. Loci. Total. Cross- overs. Per cent. Date. By- Reference. Streak balloon . . . Dachs balloon. . . Black balloon . . . Purple balloon . . . Vestigial balloon. Curved balloon . . Speck balloon.). . . 462 462 1,774 462 242 231 857 216 52.3 50.0 48.3 46.8 May — , 1914 May — , 1914 Mar. 29, 1913 May — , 1914 May — , 1914 May — , 1914 May — , 1914 May — , 1914 Muller Do. Sturtevant . Muller Muller Do. Do. Do. Am. Nat., 1916, p. 422. Do. Zeit. f.i. A.u.V., 1915, p. 276, B.C. Am. Nat., 1916, p. 422. Am.Nat., 1916, p, 422. Do. Do. Do. 2,236 1,073 48.1 462 462 462 462 218 178 150 2 47.4 38.5 32.5 0.4 VESTIGIAL (vg). (Plate 7, figure 2.) ORIGIN AND DESCRIPTION OF VESTIGIAL. The mutant wing-character now called vestigial was found by Mor- gan (December 1910) in a stock culture of truncate flies (Morgan, 1911). A few flies of both sexes were found which seemed to have tiny scales in place of wings. The size of the vestigial wing in relation to the size of the body, and the characteristic manner in which these wings are held out at right angles to the body instead of lying back above the abdomen, are shown by the figure. The character was at first called "wingless," and this name appeared in the first few publications OF MUTANT CHARACTERS. 151 describing it (Morgan, 1911; Morgan and Lynch, 1912; Morgan, 1912). The name "vestigial" was adopted when it was found that the "scale' ' was the remaining basal portion of the normal wing with the venation characteristic of that region (plate 7, fig. 2). The enormous reduction in the size of the wing is mainly due to the trimming away of the ter- minal and marginal regions of the wing. There is a marked uniformity in the extent of this trimming and in the character of the venation ves- tiges. Most commonly the wing is trimmed away as far as the anterior cross- vein, which in many specimens follows the new margin. The true marginal vein with its characteristic chsetae is entirely removed. The basal parts of all five longitudinal veins are easily recognizable, and have their normal relationship and junctures with one another. Certain small veins at the base of the wing are represented here as in the normal wing. The vestigial wing is ordinarily held out at right angles to the body, probably because of the relative thickness of the posterior margin of the wing. Sometimes, however, the ends of the wings are bent sharply backward. These wings seem to be cut off in a squarer fashion than the normal vestigial wing. It is not known whether this varia- tion has any hereditary significance. The "balancers" of vestigial flies are affected in a way analogous to the wings. The basal segment is little affected, except that it is slightly shorter and smaller. The second segment is much reduced in size and in apparent complexity. The terminal segment shows the greatest reduction, becoming a barely discernible pip (plate 7, fig. 2) instead of the balloon-like segment which is the largest part of the normal balancer (see plate 7, fig. 1). Another constant feature of vestigial flies is that the two rearmost bristles on the scutellum are separated a little wider than normal and are erect (plate 7, fig. 2) instead of turning backwards (plate 7, fig. 3). Occasionally vestigial wings are somewhat longer than is typi- cal and it is probable that this lengthening is more frequent during hot weather.1 The vestigial flies are sometimes inactive, but at other times run about very actively, appearing much like ants. Special care has to be taken with experimental cultures involving vestigial to see that the vestigial flies are all shaken out, since they cling fast to the food or paper in the culture bottle and are exceptionally slow and difficult to get out. The viability of vestigial flies is fairly good, very close approaches to expectation being obtainable when pairs are used and food conditions are favorable. The earlier work showed consider- able deviations from expectation because of failure to recognize the necessity of these conditions. Vestigial flies tend to hatch two or more days later than the not-vestigial flies, and unless the cultures are run full term will give ratios in which the numbers of vestigial are below expectation. 1 Since the above was written, Roberts (J. E. Z., 1918) has strikingly confirmed the fact that high temperature favors the production of wings approaching the wild-type in size. 152 THE SECOND-CHROMOSOME GROUP STOCK OF VESTIGIAL. Since vestigial, like balloon, appeared in the truncate stock, the vestigial flies were often at the same tune truncate, though these could not be distinguished by inspection from the simple vestigials. By out- crossing and extraction a stock was obtained which seemed to be free from truncate (as judged by the absence of truncates among the not- vestigial flies of certain FI and F2 cultures). The balloon mutation which had appeared in the truncate stock just before the occurrence of vestigial was even more difficult to eliminate and occasionally cropped out in the early experiments in which vestigial was used. INHERITANCE OF VESTIGIAL. In out-crosses to wild the vestigial appeared to be completely reces- sive. In F2 the vestigials reappeared in much less than a quarter of the flies, due to the practice of using mass cultures and to the rather poor feeding methods of that time. Reciprocal crosses gave the same results in FI and F2, so that the gene was known to be not sex-linked. Lutz (1913) made a biometrical study of wing-length and found that the wings of flies heterozygous for vestigial are slightly but actually shorter than the wings of wild flies. Also, the ratio of wing-length to femur-length was less, showing that vestigial is not completely reces- sive. It is known in other cases also (see Morgan and Bridges, 1913) that characters that to simple inspection are completely recessive really are influencing the character of the heterozygous individuals. CHROMOSOME CARRYING VESTIGIAL. At this tune the only case of autosomal linkage known in Drosophila was the observation by Bridges that no black curved flies had appeared in the F2 of the cross of black by curved (Bridges and Sturtevant, 1914). Following this, a concerted testing of the linkage relations of all the known autosomal mutations was carried out. One of these tests, made by Sturtevant and by Miss Clara J. Lynch, showed that in the F2 of the cross of black by vestigial no black vestigial flies appeared. Both of these cases were put down as very close linkage, "complete repulsion," since it was not yet known that there is no crossing-over in the male whereby this result would be obtained, even though the crossing-over in the female were very free. That crossing- over actually had occurred was shown by the results of mating some of the F2 black by some of the F2 vestigial flies. In one of these F3 cultures some black flies occurred, which meant that at least one of i j, the vestigial parents had been heterozygous for black, -=- , the b v, 0 Vn OF MUTANT CHARACTERS. 153 chromosome being a cross-over. By breeding together some of these F3 blacks, in F4 black vestigial flies were obtained from which a stock was made that is still running. LOCUS OF VESTIGIAL. By aid of this stock of black-vestigial it was possible to make back- cross tests of the amount of crossing-over. When these experiments were carried out by Morgan it was found that there was about 22 per cent of crossing-over in the female, but none whatever in the male (Morgan, 1912). The principle of no-crossing-over in the male, first clearly demonstrated in the above back-cross, has been found to apply to all cases in both the second and third chromosomes of Drosophila. From the early data of Morgan it was known that the two loci, black and vestigial, were about 22 units from each other. Several errors have been found in the data for crossing-over as first presented (Morgan, 1912). These data, as corrected and considerably added to (Morgan, 1914), show that the locus of vestigial is about 18 rather than 22 units from black. The mapping of vestigial in relation to the other second-chromosome genes was carried out by Bridges (through the use of purple as a sec- ondary base) and by Sturtevant (through use of curved as a base of reference) . Relatively large amounts of data involving the relation of vestigial and other second-chromosome mutants was soon secured. The most useful of these determinations have been the various purple- vestigial data, for purple is the base of reference for vestigial. Table 17 gives a summary of this early data and of all that have since become available. VALUATION OF. VESTIGIAL. Vestigial, while it is not strictly of first rank in usefulness, approaches very close to this standard. In ease and quickness of separation from the wild-type it is unsurpassed. Its position in the chromosome is one which is important and convenient. Enough work has been done using vestigial as material, so that in undertaking fresh work one can count on a sound basis for comparison. Its viability is good under good cultural conditions. The above are the points in its favor; its disadvantages are that it masks all other wing-characters, so that its use in an experiment materially reduces the availability of other wing- characters, some of which, such as curved, are themselves of first rank. Its viabilty is apt to be poor unless careful and experienced attention is given to cultural conditions, and its lateness of hatching and the diffi- culty of getting the vestigials out of the culture bottle also tend to give aberrant ratios to the unwary. 154 THE SECOND-CHROMOSOME GROUP TABLE 17. — Summary of cross-over data involving vestigial and other second- chromosome loci. Loci. Total. Cross- overs. Per cent. Date. By- Reference. Star vestigial . . Streak vestigial Dachs vestigial Black vestigial Purple vestigial Vestigial curved Vestigial speck 450 462 195 164 43.3 35.5 Nov. 17, 1915 May — , 1914 Dec. 10, 1913 May — , 1914 Sept. 10, 1912 Sept. 10, 1912 June 8, 1913 Dec. 10, 1913 Jan. 9, 1914 Nov. 17, 1915 Mar. 5, 1917 May — , 1917 July 16, 1912 July 5, 1913 July 5, 1913 Jan. 9, 1914 May — , 1914 Mar. 5, 1917 Mar. 14, 1913 June 8, 1913 May — , 1914 Mar. 14, 1913 Sept. — , 1914 May — , 1914 Bridges .... Muller « s' Vg ; £ „ B.C. ; table 123, this paper. Am. Nat., 1916, p. 422. a; d b vg balanced B.C.; II 114- II 138. Am. Nat., 1916, p. 422. Biol. Bui., 1914, p. 197: B.C. Vg Biol. Bull., 1914, p. 198; 6 va B.C. Zeit. f. i. A. u. V., 1915, p. 286, b vg c B.C. d; d b vg balanced B.C.; II 114- II 138. Am. Nat., 1916, p. 422. « S' Bridges .... Muller 4,892 462 1,456 129 29.7 27.9 Morgan . . . Do 5,354 1,585 29.6 3,499 1,268 694 4,892 5,001 450 2,139 1,748 632 217 169 806 815 99 477 285 18.1 17.1 24.4 16.5 16.3 22.0 22.3 16.3 Sturtevant . Bridges .... Muller Bridges .... Plough Gostenhofer Bridges .... Do ....Do.... ....Do.... Muller Plough Sturtevant . Do ve • j, „ n B.C. ; table 123, this paper. J. E. Z., 1917; b pr vg B.C. 6 — — B.C.; students records. Vg pr; pr vg B.C.; B 36.1-B 39.2. PT; pr vg B.C. 1st; DA-DH. pr; — B.C. 1st; DI-DO. Vg Pr', b Pr vg balanced B.C., II 141-674. Am. Nat., 1916, p. 422. J. E. Z.; 1917, bprVgB.C. Zeit. f. i. A. u. V., 1915, p. 245, Vg C B.C. Zeit f. i. A. u. V., 1915, p. 247, Vg C Sp B.C. Am. Nat., 1916, p. 422. Zeit. f. i. A. u. V., 1915, p. 245. Zeit. f. i. A. u. V., 1915, p. 287. Am. Nat., 1916, p. 422. 20,153 3,578 17.8 825 2,839 2,335 5,001 462 2,139 79 305 303 539 60 323 9.1 10.7 13.0 10.8 13.0 15.1 13,601 1,609 11.8 856 402 462 75 32 34 8.8 8.0 7.4 Muller Sturtevant . ....Do.... Muller 1,720 141 8.2 1,446 146 462 520 40 178 36.0 27.4 38.5 2,054 738 35.9 OF MUTANT CHARACTERS. 155 BLISTERED (*.). (Text-figure 74.) ORIGIN AND DESCRIPTION OF BLISTERED. The mutation "blistered"1 was found by Bridges (November 16, 1911) in a mixed stock of rudimentary and normal (culture A23) as a igaudinal vein TEXT-FIGURE 74. — Blistered wing and venation. 74a shows a pair of wings with the bend in the fourth longitudinal vein, and extra veins. 746 shows the plexus near the end of the fifth longitudinal vein. 74c shows for comparison a normal wing with the standard terminology for venation and for the cells of the wing. It is probable that 74c is on a larger scale than 74a and 746, though the blistered wing is characteristically smaller than normal. not-rudimentary female which had in each wing a small vesicle in the region of the fifth longitudinal vein just beyond its junction with the posterior cross-vein. This female was bred to a wild-type brother and 1 Mac Dowell ('15)published bristle counts on this stock under its earlier name of "half-balloon." 156 THE SECOND-CHROMOSOME GROUP gave only wild-type sons and daughters, from which it was concluded that the character was recessive and non-sex-linked. The F2 genera- tion gave only a few blistered (about 1 in 10), and these were nearly all females. By mating together the blistered individuals hi pairs a stock was obtained which must have been genetically homozygous for blistered, although only about half of the females and about a quarter of the males showed the character. It had been noticed that the size of the vesicle varied from a very minute bubble to one which covered over half the area of the wing, and that there was not a very close correspondence between the two wings; frequently the blister would appear in only one of the two wings. A closer inspection showed that the wings which did not show a vesicle had a small plexus of veins in the region occupied by the vesicle (figs. 74 a and 6), and it was found that the flies could be quite readily classified by this character, irre- spective of whether they showed the blister or not. At the same tune the results given by breeding from these abnormally veined flies showed that the venation and the blistering were both products of the same gene. A third manifestation of this gene is a sharp bend in the distal end of the fourth longitudinal vein (shown hi text-figure 74, a). TABLE 18. — PI mating, curved cf X blistered 9 ; F\ mating, wild-type 99 + wild-type a" cf . Apr. 30, 1912. Wild- type. Curved. Blistered. Curved blistered. B12a.... 81 18 23 0 B126 62 25 27 0 B12c 58 12 6 0 Total 201 55 56 0 THE CHROMOSOME OF BLISTERED. Little was done with blistered, aside from getting the pure stock just described, until the discovery that black and curved were hi the same chromosome gave a sharp impetus to further study of autosomal linkage. Shortly thereafter (April 3, 1912) blistered was crossed to curved and three F2 cultures were raised (table 18). No curved- blistered flies were found hi the F2. The numbers given in table 18 represent only the first counts from each of these three cultures, and for this reason the number of wild-type flies, which are the first to hatch, is abnormally high, No further records were made of the F2 offspring, because of the suspicion that blistered might not be distinguishable in curved flies, and that the absence of the double recessive might be due to inhibition or masking instead of the supposed linkage. Never- theless, all the F2 flies were examined in the hope of finding a double OF MUTANT CHARACTERS. 157 recessive. When none was found the experiment was discontinued, since it was thought that the chances of finding a double were as good in F2 as in any subsequent generation — an attitude excusable at that time, before the fact of no-crossing-over in the male had yet been discovered. TABLE 19. — PI, blistered 9 X pink cf; FI wild-type 9 and d" . Aug. 10, 1913. Wild type. Blistered. Pink. Pink blistered. M52 M 53 Total 157 146 60 45 58 54 25 13 303 105 112 38 TABLE 20.— PI, blistered 9 X black Oct. 23, 1913. Wild-type 9- Wild-type cf. Free-vein 9 • Free- vein d"1. II 99 27 24 48 4 B.C., Fi free-vein cT X black 9 . Nov. 23, 1913. Non-cross-overs. Cross-overs. Black. Free- vein. Black free-vein. Wild-type. 9 -j— |— ~ ) ' a *ew °^ these F2 blacks should be heterozygous for curved ( — r— - ) and a corre- \ o c / spending few of the curves should be heterozygous for black ( J . The appearance in FS of a few black flies ( -j— - ) showed that at \ o c / least one of the F2 curved flies had been the result of crossing-over. By inbreeding these F3 blacks, black curved flies (25 per cent) were secured in F4. That the absence of black curved flies in F2 was really due to lack of crossing-over in the male was shown by the results of the back-cross tests carried out upon double heterozygous males as compared with like tests of the females. In the tests of the males no cross-overs appeared in a total of 1,066 flies, while in the tests of the females 1,717 OF MUTANT CHARACTERS. 167 TABLE 25. — Summary of cross-over data involving chromosome loci. curved and other second- Loci. Total. Cross- overs. Per- cent. Date. By— Reference. S' Star curved . . . 46.8 ^'> ™ x. . B.C. Ists; 1836-'94. Pr c Sp Sr J. E. Z., 17, , B.C. controls. 19,870 9,123 45.9 Streak curved.. 1,807 462 745 178 41.2 38.5 Nov. 6, 1913 May — , 1914 Bridges .... Muller S't; S' Balanced B.C.; II 103-124. Am. Nat., 1916, p. 422. 2,269 923 40.7 Dachs curved. . 462 145 31.4 May — , 1914 Muller Am. Nat., 1916, p. 422. Black curved . . 7,419 260 253 402 3,934 223 462 36,622 1,717 69 66 120 839 63 106 8,598 23.1 25.5 26.1 29.9 21.3 28.2 22.9 23.4 Jan. 13, 1913 Jan. 13, 1913 Jan. 13, 1913 June 8, 1913 Aug. 24, 1913 Oct. 16, 1913 May — , 1914 Jan. 5, 1915 B. & Strt. . . Do. Do. Sturtevant . Bridges .... Sturtevant . Muller Plough "Plnilfrll Biol-lBull., 1914, p. 209; b c and b -B.C. Biol. Bull., 1914, p. 208; b c Fa. Biol. Bull.,'1914, p. 212; — B.C. C Zeit. f. i. A. u. V., '15, p. 247; $ b va c B.C. pr; b pr c B.C. Ists; II 58-11 88. |Zeit. f. i. A. u. V.; '15, p. 247; 6 c 8P B.C. Am. Nat., 1916, p. 422. J. E. Z. 1917, 6 pr c B.C. controls. S' J. E. Z. 1917, , B.C. controls. 62,679 14,237 22.7 Purple curved . 3,934 1,807 462 711 375 90 18.1 20.7 19 5 Aug. 24, 1913 Nov. 6, 1913 May 1914 Bridges .... Do. Muller .... PT; b pr c B.C. Ists; II 58-1188. S't; S't PT c balanced B.C. Am Mat IQIfi r\ 4.99 952 36,622 182 7,222 19.1 19.7 Aug. 24, 1914 Jan. 8, 1915 Bridges .... Plough sp; pr c sp B.C.; 452-508. J. E. Z. '17, b pT c B.C. controls. S' 593 ,508 117 22.3 19.7 Feb. 8, 1916 Do. S'; B.C. Ists; 1836-'94. Pr c sp In a; bprcF2; 3203-'08. 51,136 10,205 19.9 Lethal Ha curved 94Q In a'' lua c Fa ; 2675, 2840; '59, '60, '63. Vestigial curved. "856 402 462 75 32 34 8.8 8.0 7 4 Mar. 14, 1913 June 8, 1913 May 1914 Sturtevant . Do. Muller Zeit. f. i. A. u. V., p. 245, v9 c B.C. Zeit. f . i. A. u. V., p. 247, va c sp B.C. Am TVnt IQIfi r\ 4.99 1,720 141 8.2 Curved speck.- 1,007 223 462 952 262 71 150 266 26.0 31.8 32.5 27.9 Mar. 10, 1913 Oct. 16, 1913 May — , 1914 Aug. 24, 1914 Sturtevant . Do. Muller Bridges .... Do Zeit. f. i. A. u. V., '15, p. 245; » c sp B.C. | Zeit. f. i. A. u. V., '15, p. 247; b c sp B.C. Am. Nat., 1916, p. 422. sp; pr c Sp B.C. ; 452-508. •*">"" ~F- -* -**&&&*, 632 226 30.5 35.7 Feb. 8, 1916 Do. S'; — B.C., Ists.; 1836-'94. Pr c sP - _ ' liia! PT c sp|F2;[3203-'08. 10,042 3,037 30.2 Curved balloon 462 150 32.5 May — , 1914 Muller Am. Nat., 1916, p. 422. 168 THE SECOND-CHROMOSOME GROUP cross-overs appeared in a total of 7,419 flies (Bridges and Sturtevant, 1914). In the male there was no crossing-over whatever, while in the female there was about 25 per cent. The direction along the II chromosome from black to curved was called "to the right," and curved was therefore mapped at a locus 25 units to the right of black. The position of curved has been more accurately determined by use of intermediate loci than was possible from the rather long black curved interval. Thus the purple curved stock, made by Bridges in preparation for the triple back-cross black purple curved, was used by Mr. W. S. Adkins to run extensive purple curved back-cross tests. These crosses, the data for which are not yet available, gave about 18 per cent of crossing-over between purple and curved, which agrees with the result of the black purple curved back-crosses. Much data has since been collected on this cross-over value, largely incidental to the work on age variation by Bridges and temperature variation by Plough. A total of 51,136 flies included 10,205 cross-overs between purple and curved; the observed percentage of crossing-over was thus 19.9. It is possible that there is considerable double crossing-over within this region, since it is in the middle of the chromosome, where double crossing-over in relation to map-distance has been found to be extraordinarily high. If a coincidence of 70 is assumed, the corrected purple-curved value becomes 21.4 and the locus of purple is 27.6 units to the right of black. The locus of curved was also referred to vestigial by Sturtevant, who ran vestigial curved and black vestigial curved back-crosses. This method has great advantages in mapping the locus of curved, since vestigial is itself accurately mapped in relation to purple and is an intermediate base between purple and curved. Only the not-vestigial back-cross flies can be used in the calculation, since curved vestigials can not be distinguished from the simple vestigial class. Hence the vestigial curved data are too meager as yet (1,720 flies) to be used as the main basis of the location of curved. A summary of the cross-over data involving curved and other second- chromosome loci is given in table 25. VALUATION OF CURVED. Curved is in all respects a mutant character of first rank, both for student use and in special experiments. Its separability from the wild-type is both easy and accurate, even without experience. Its viability is excellent. It causes no trouble through liability to drown- ing or miring, which might have been expected on account of the strongly divergent wings. Its locus is the outpost of the central body of well-mapped genes and it is therefore the base of reference for speck and for all genes near the right end of the chromosome. OF MUTANT CHARACTERS. 169 PURPLE (pr). (Plate 5, figure 8.) ORIGIN OF PURPLE. In a stock kept in the stock-room and supposed to be simply vesti- gial, there was found, February 20, 1912, a single male which had an eye-color much like that of the well-known double recessive vermilion pink. The color of the vermilion pink eye is about that of the pulp of an orange, and the early papers accordingly referred to this double recessive as " orange." The new color was seen to differ slightly from vermilion pink in that it Was of a brilliant ruby-like transparency and lacked the flocculent or slightly cloudy appearance of vermilion pink. This difference seems to arise partly from a difference in the distribution of the pigment. In vermilion pink the pigment looks as though it were mainly in the spaces between the radially arranged ommatidia with a clearer zone just under the surface of the eye. One sees in the vermilion pink eye a light fleck which travels over the eye as it is turned. This seems to be due to a deficiency of pigment hi the deeper parts of the eye and the light fleck is this light center seen through the small group of facets whose axes are in line with our eye. * The pigment in the case of the new eye-color gave the appearance one would expect if it were uniformly distributed or even in solution throughout the eye. INHERITANCE OF PURPLE. This single male with the orange-like eye-color was out-crossed to a wild female, and in FI gave only wild-type males and females (wild- type 9 32, cf 33; reference No., B 1) which showed that the color was recessive. In F2 the orange-like color reappeared, but in addition the sex-linked eye-color vermilion emerged, and also a new eye-color " purple" which appeared equally among the F2 females and males and was therefore known to be an autosomal (not sex-linked) character. It was now evident that the orange-like color resembled the old "orange" (vermilion pink) genetically as well as somatically, for it was proved by this F2 to be a double recessive, vermilion purple, in which purple corresponds to pink. It seems probable that the two eye-color mutations, vermilion and purple, present in the male first found were not of simultaneous or related origin. There was a vague rumor that the vestigial stock had contained vermilion at some tune previous to this discovery. No vermilion or purple was found hi it subsequently, however. DESCRIPTION OF PURPLE. The eye-color of purple flies passes, in its development, through an interesting cycle of changes closely parallel to those seen in the ripen- ing of a sweet cherry. In the pupa the eye is at first colorless, then it 170 THE SECOND-CHROMOSOME GROUP assumes a creamy tone which in turn becomes pinkish, passing pro- gressively through a yellowish pink to pink and to ruby. When the flies hatch, the color is a transparent rather deep ruby. This color rapidly deepens to garnet and then passes on to a purplish tone. The typical purple color at its maximum development — in flies about a day old — while retaining much of its transparency, appears darker in tone than the red of the wild-type, purple being the first of such "dark" eye-colors. As the fly becomes older this " ripe-cherry " color is progressively obscured, apparently by an increase in a flocculent red pigment, like that of the wild fly. The eye-color thus becomes somewhat lighter than red again, though always distinguishable by a lesser opacity and by a light "fleck" in place of the hard dark fleck seen in the wild eye. With extreme old age the color approaches still closer to red, but does not become strikingly darker, as do pink and sepia, for example. In purples of the same age fluctuations in color are not great. Separations are easy if done, as usual, while the flies are mostly under 2 days old, though the climax in the development of the purplish tone offers the most favorable stage. THE DIFFERENTIATION OF PURPLE BY VERMILION— DISPROPOR- TIONATE MODIFICATION. While the difference between the color produced by the purple gene and the color produced by its wild-type allelomorph (red) is distinct, it is neither great nor striking, since in tone purple is first slightly darker and later somewhat lighter than red. However, in classifying the eye-colors in F2 from the cross of vermilion by wild, it was observed that the difference between vermilion purple and vermilion not- purple was not only constant in direction, but also conspicuous in extent. The separability of purple versus not-purple is favored by the presence of vermilion, which may therefore be called a "differentia- tor" of purple. Regarded in the converse relation, namely, the effect of purple on vermilion rather than the effect of vermilion on purple, purple is a much stronger modifier of vermilion than of not- vermilion. Purple may be described as a "disproportionate modifier" of vermil- ion, since from the small amount of its effect on eye-color when acting alone one would not have expected the great effect it produces when acting in the presence of vermilion. This type of intensification — disproportionate modifier and, con- versely, differentiator — stands midway between the normal relation where combination effects are roughly proportional to the separate effects, so that both genes may be called "general modifiers," and the special relation where a given gene, "specific modifier," produces by itself no visible effect whatever, but which gives a more or less marked effect when acting in conjunction with some other gene, its specific base, sensitizer, or differentiator. OF MUTANT CHARACTERS. 171 In order to make full use of this differentiation of purple versus not- purple by vermilion, it is necessary that all flies used in the experiment should be made homozygous for vermilion. This is often inconve- nient, and accordingly only in the early and comparatively simple experiments was this method employed. It was soon found also that the separation of purple from red was not causing any trouble, so that the differentiation in this case has little net advantage, though it is still of interest as being the first example in Drosophila in which inten- sification was recognized and deliberately made use of. THE RELATION OF PURPLE TO PINK. Some of the first purples which emerged in the F2 were crossed to pink to test whether these two eye-colors were allelomorphic or not. Only wild-type FI males and females were produced (table 26), which showed that the purple is not an allelomorph of pink. TABLE 26. — F\ progeny from out-cross of purple. Apr. 17, 1912. Wild- type 9- Wild- type cr, ... B C Ists' 1836-1891 Pr c 8V S' pr _ 1,027 362 413 138 40.2 38.1 Aug. 24, 1916 July 21, 1917 Do. Do. &i o j B.C.; 4999—5110. v, 1 b I 1 1 «, Dec. 11, 1913. Dachs. Black vestigial. Dachs black vestigial. Wild- type. Dachs vestigial . Black. Dachs black. Vestigial. Total. II 120 88 135 12 29 22 26 3 3 318 II 121 51 79 12 15 8 18 1 184 II 122 81 98 10 36 7 21 3 5 261 II 123 87 110 16 43 23 19 3 5 306 II 124 93 147 20 36 15 27 1 5 344 II 126 74 70 12 16 18 17 2 209 Total 474 639 82 175 93 128 13 18 1,622 The triple back-cross was made to the extent of about 1,600 flies in three of the four possible ways (tables 63, 64, and 65) and gave a grand total of 4,892 flies (table 66) of which 3,329 were non-cross-overs, 757 were cross-overs between dachs and black, 689 were cross-overs between black and vestigial, and 117 were double cross-overs. 220 THE SECOND-CHROMOSOME GROUP As soon as the first flies from the triple back-cross had begun to hatch it became evident that the locus of dachs is to the left of black and not to the right. The discovery that streak, a dominant, was so far to the left of black that it gave very free crossing-over had just been made and had led to the most extensive recasting that any of our maps had been subjected to. The location of dachs in the middle of this long gap was therefore very welcome. TABLE 65. — PI, dachs black cf cf X vestigial 99. FI wild type 9 dachs black-vestigial cf cf from stock. d b d \ Vg d b I », d I 1 Vg \ b 1 1 b \Vg Dec. 11, 1913. Dachs black. Vestigial. Dachs vestigial. Black. Dachs black vestigial. Wild- type. Dachs. Black vestigial. Total. II 127 69 96 28 29 26 24 3 5 280 II 128 111 109 23 20 32 31 4 6 336 II 129 78 82 13 19 10 16 3 2 223 II 130 101 141 30 34 18 31 4 8 367 II 131 109 174 26 39 27 33 12 6 426 Total 468 602 120 141 113 135 26 27 1,632 The total amount of crossing-over between dachs and black as calcu- lated from this balanced experiment was 17.9 units, which agreed with t'he 17.7 units found in the simple dachs black back-cross. TABLE 66. — Linkage of dachs black and vestigial with balanced inviability. Dec. 10, 1913. I | I Total. Cross-over values. 1 1 db b vg d Vg d b Vg 1,146 1,113 1,070 239 257 261 220 221 248 33 31 53 1,638 1,622 1,632 16.6 17.8 19.2 15.5 15.5 18.4 28.1 29.5 31.2 d b Vg d b Vg Total 3,329 757 689 117 4,892 17.9 16.5 29.7 Table 67 gives the summary of the crossing-over data including dachs. The calculation of the locus of dachs on the basis of all the data is 17.5 units to the left of black, or referred to star as the zero- point at 29.0. VALUATION OF DACHS. The usefulness of dachs is limited only by its rather poor and erratic viability. In many experiments the viability of dachs is up to par, but in combination with certain other characters it has been unsatisfactory; thus the stock called "r" (dbprcpxsp) is poorly viable and almost useless, while the "TT — " stock, which differs only in the omission of OF MUTANT CHARACTERS. 221 TABLE. 67. — Summary of the cross-over data involving dachs. Loci. Total. Cross- overs. Per cent. Date. By- Reference. Star dachs .... Streak dachs . . Dachs black. . . Dachs purple . . Dachs vestigial. Dachs curved . Dachs speck. . . Dachs balloon. 96 152 1,617 369 211 1,027 31 53 425 112 57 271 32.3 34.8 26.3 30.4 27.1 26.4 Sept. 12, 1915 Sept. 12, 1915 Sept. 15, 1915 Oct. 6, 1915 Nov. 18, 1915 Aug. 24, 1916 May — , 1914 Aug. 24, 1916 Mar. 18, 1913 June 30, 1913 Dec. 10, 1913 May — , 1914 May — , 1914 Aug. 24, 1916 Dec. 20, 1913 May — , 1914 May — , 1914 May — , 1914 May — , 1914 Bridges Do. Do. Do. Do. Do. Muller Bridges Bridges Do. Do. Muller Muller Bridges Bridges Muller Muller Do. Do. - S' Ft, dachs flies; °1' J 2141-2216. Idem, not-dachs flies. S' S'; -7 B.C.; 2146-2305. a di; ^-r F,; 2217-2659. OJ S' di;—^ B.C.; 2460. q. S' Pr B.C.: 4999- 0 ' St d 5110. Am. Nat., 1916, p. 422. _P s/ PrB. C.:4999- 3,472 949 27.3 462 396 45 64 9.7 16.2 ^' Std 5110. d; d b. F2; II 34-11 36. d; db B. C.; II 40-11 98r. d; d b Vg balanced B. C.; II 114-11 138. Am. Nat., 1916, p. 422. Am. Nat., 1916, p. 422. c. s' Pr B. C.:4999- 858 109 12.7 338 933 4,892 462 82 163 874 77 24.3 17.5 17.9 16.7 6,725 1,196 17.8 462 1,027 97 196 21.0 19.1 5ft d 5110. d; d b vg balanced B. C.; II 114-11 138. Am. Nat., 1916, p. 422. Do. Do. Do. 1,489 293 19.7 4,892 462 1,456 129 29.7 27.9 5,354 1,585 29.6 462 462 462 145 231 231 31.4 50.0 50.0 dachs, is entirely normal in viability. In all other respects dachs is of first rank. Since the character used in classification is the number of joints of the tarsus, there is no masking effect possible with any of the other second-chromosome mutants. The recessiveness of dachs is com- plete, its identification is perfect, and the separations are easy and rapid. The locus of dachs is such that it is the most important connecting- link between the left end of the chromosome and the securely estab- lished and well-mapped region from black to the right end of the chromosome. 222 THE SECOND-CHROMOSOME GROUP STREAK (SA). (Plate 5, figure 5, and Plate 10, figure 2.) ORIGIN OF STREAK. In a stock culture from a pair of flies with the mutant called "lop- wing" (culture C 149, November 27, 1912), Bridges found a single female which had a prominent broad, dark streak down the middle of the thorax. STOCK OF STREAK. This female (non-virgin) was mated to several of her brothers and produced many streaks among the offspring. It was assumed that the character was recessive and that some of the brothers had been heterozygous. No FI counts were made and not much attention was paid to the character. Several of the streak individuals were mated together to provide stock. In this F2 culture somewhat more than half (no counts) of the offspring were streak where all had been expected to be streak. This was thought to indicate a "poor" character, which, like truncate, club (wings), and others, shows in a variable proportion of the flies of the same genetic constitution. The stock was carried on in this way for two more generations, when it was decided to throw it away as being too poor to repay further labor. This would have been done had not Morgan seen in this character a bearing on a selection problem which he had been carrying out for over two years on the thorax pattern of "with" flies. In the course of selections for a still darker pattern three notable successes had been obtained, all of which turned out to be simply new mutations (speck, olive, and band) which had occurred in the selected stocks, but which gave no further variability or progress when once the stocks were pure for them. In streak there was an example of a dark thorax character which closely resembled in pattern the darkest of the long selected "bands," though not as dense in pig- mentation, but which had arisen entirely independent of any selection whatsoever. Later the trident mutant "trefoil" (plate 5, fig. 6) likewise arose independent of selection. These independent mutations and the fact that during this same period over a hundred other mutations affecting every part of the body had appeared in Drosophila, left no basis what- ever for the supposition that the selection had had any effect what- ever on either the frequency or the nature of the mutations, or that any other process, aside from the production of three definite mutations, had contributed to the success of the selection. Morgan selected the streak stock for about six months, though not very vigorously, without increasing the intensity of the pattern or the frequency of the streak individuals. OF MUTANT CHARACTERS. 223 DESCRIPTION OF STREAK. The principal characteristic of streak flies is the band of pigment along the thorax and scutellum. This band seems to be rather deep- lying, and is possibly situated in a different layer from that in which the other pigment characters develop. There is considerable variation in both the intensity and the extent of the dark color. In its greatest development it is a solid band, like that of colored figure 5, filling in the entire region between the dorso-central bristles and extending over the entire scutellum. In less-developed types the weakening starts in the region ahead of the dorso-central bristles and is most pronounced between the prongs of the trident pattern, so that an appearance much like typical band is given. (For figures of "with," band, and trefoil, see Mechanism of Mendelian Heredity, p. 206). The intensity of color is never very great and the color may nearly vanish. However, there are other accessory characteristics that aid in the classification. Chief of these is a flattening of the thorax and the appearance of bubbles. Both of these effects seem to be due to an ill development of the underlying muscles. There appear to be present in the thorax large spaces or sinuses filled only with blood and large bubbles. Where there are no bubbles present this condition is not so easy to distinguish, though it may sometimes be made out by slightly pressing the thorax. The wings are apt to droop and to diverge slightly, probably also on account of the muscular condition. DOMINANCE AND LETHAL EFFECT OF STREAK, PARALLEL TO YELLOW MOUSE. The occurrence of the mutation as a single individual — a female — in a pair culture, its immediate reappearance in about half the FI flies after crossing to normal males, and the failure of these FI flies to breed TABLE 68. — Pi, streak 9 X wild cf. July 22, 1914. Streak 9. Streak cf. Wild- type 9. Wad- type cf. 336 23 22 25 28 391 53 69 69 64 393 37 37 48 43 394 76 68 64 60 Total 189 196 206 195 true, found an explanation (rather delayed) in the assumption that streak was an autosomal dominant. Moreover, the fact that the stock could not be made to breed true (continually producing at least a third of the offspring wild-type) and that repeated pair matings as well gave this same result, led to the assumption that homozygous individuals were not produced. This case was seen to be a parallel to 224 THE SECOND-CHROMOSOME GROUP the well-known case of the yellow mouse, and was the first of many to be found in Drosophila, where a homozygous dominant is lethal. Outcrosses of streak by wild gave in FI streaks as approximately half of the flies. The records of these early out-crosses were lost (note- book S II), but similar out-crosses made later illustrate the fact as well (table 68). CHROMOSOME CARRYING STREAK. The next task was to determine in which chromosome the gene for streak is located. This was done by back-cross tests of the male for black (II chromosome) and pink (III chromosome). TABLE 69. — PI, streak 9 X pink cf; FI streak d" X pink 9 of stock. July 7, 1913. Streak. Pink. Streak pink. Wild- type. D 7 90 133 71 117 D 7r 65 50 77 64 Total 155 183 148 181 Streak males heterozygous for pink were produced from the mating of streak female by pink male. Two of these FI males were back-crossed to pink females and produced a total of 667 offspring, 329 of which were recombinations (table 69). The presence of the streak-pink and the wild-type flies as 49.3 per cent of the whole proved that streak was not in the third chromosome, since independent assortment was demon- strated. TABLE 70. — Pi, streak 9 X black cf; FI streak d" X black 9 of stock. July 7, 1913. Streak. Black. Streak black. Wild- type. D 5 19 21 0 0 In the back-cross test of the streak male heterozygous for black no recombination occurred. Every one of the 19 not-black flies was distinctly streak, and likewise none of the 21 black flies showed a trace of streak (table 70). This result was due to the fact that the locus of streak is in the second chromosome and the lack of crossing-over in the male. LOCUS OF STREAK. Immediately following the appearance of the first flies in the pre- ceding back-cross tests of the male, a test of the locus of streak was made by means of the mutant morula, which had itself just been mapped at the right end of the second chromosome. Back-cross tests of FI streak females, from the cross of streak females by morula males, gave a total of 876 flies, of which 405 or 46.2 per cent were cross-overs (table 71). OF MUTANT CHARACTERS. 225 This very free crossing-over between streak and morula indicated that streak was far away from the right end of the chromosome, where the gene for morula had been located. It was thought that the locus of streak was in the neighborhood of black (which was at that time considered the left end of the chromosome) or purple, which was not far from black. It did not seem probable that an accurate classifi- cation of streak and black could be made at the same time, so purple was used instead. TABLE 71. — PI, streak 9 X morula &; FI streak 9 X morula d* of stock. Aug. 24, 1913. Streak. Morula. Streak- morula. Wild- type. II 64 11 8 4 6 II 82 50 47 40 31 II 92 55 55 44 61 II 97 124 121 130 89 Total 240 231 218 187 A three-locus experiment partly balanced for inviability was carried out. A streak female was out-crossed to a purple curved male and FI streak females were back-crossed by purple curved males. Since streak is a dominant, the triple recessive not-streak purple curved was a double mutant form, which was a great saving of labor over the ordinary case, in which the triple mutant multiple recessive has to be made up. TABLE 72. — PI, streak 9 X purple curved cT; B.C., curved d* of stock. streak 9 X purple st s* \ PT C 0ft 1 c st I Pr \ Pr C \ Pr \ 1 1 C Nov. 6, 1913. Streak. Purple curved. Streak purple curved. Wild- type. Streak curved. Purple. Streak purple. Curved. II 103 82 81 45 49 31 28 9 12 II 104 83 91 57 55 27 18 14 19 II 110 46 52 18 23 14 9 8 7 Total 211 224 120 127 72 55 31 38 The result of the back-cross (table 72) was surprising, since it upset the ill-founded notion that black was at the left end of the second chromosome. The cross-over values (streak purple = 36.0, purple curved 24.5, streak curved 45.7) and the double cross-over classes (streak purple versus curved) showed that streak was fully 40 units to the left of purple (allowing for double crossing-over). No trouble in classifying streak was met in this experiment, so that all of the flies are available for the calculation. From the streak purple and the curved flies that appeared in the back-cross just described a PI mating was made for the second type of 226 THE SECOND-CHROMOSOME GROUP back-cross (table 73). This second back-cross gave linkage results which differed only very slightly from those of the first, and in such a way that by combining the two sets of data the deviations of one tend to balance those of the other and more nearly correct values can- be calculated. TABLE 73. — Pi, streak purple 9 X curved d1; B.C., FI streak 9 X purple curved c? of stock. St l ->T st | c St P, 1 c St 1 ! c I Pr 1 1 1 >, | c Dec. 18, 1913. Streak purple. Curved. Streak curved. Purple. Streak purple curved. Wild- type. Streak. Purple curved. II 136 66 58 23 33 10 11 7 5 II 137 29 27 12 16 3 7 7 2 13 26 27 8 12 5 10 2 4 15 24 45 25 29 9 14 5 5 16 54 53 22 32 12 14 8 8 23 24 16 14 8 6 2 1 3 24 23 24 13 7 5 9 2 3 Total 246 250 117 137 50 67 32 30 The combined data gave 1,807 individuals, of which 931 were non- cross-overs, 501 cross-overs between streak and purple, 244 cross-overs between purple and curved, and 131 double cross-overs. The cross- over values calculated from this distribution are streak purple 35.0 per cent, purple curved 20.7 per cent, and streak curved, 41.2 per cent. The coincidence from the first back-cross was 98.0 and from the second 102.0. The coincidence from the combined data was 99.8 1807X131X100 = 99.8^ 632X375 For a section of this great length and involving this particular region the net result was that the occurrence of a cross-over in one section was without effect upon the occurrence of a cross-over in the other section. The locus of streak as calculated from the above data, making allowance for the probable amount of double crossing-over, was 40.6 to the left of purple or 34.7 units to the left of black, which was the locus farthest to the left of those previously determined. This great gap was almost immediately filled by the mapping of dachs at 17.9 units to the left of black, or almost exactly midway between the two. This position of dachs offered a new base of reference for streak and one much more dependable than the remote purple. The data on which a direct determination of the streak dachs distance was made is included in the section on star, for by means of a quadruple back-cross involving star, streak, dachs, and purple the loci of both star and streak were linked up with the portions already mapped. The streak dachs interval was found by this experiment to be about 16.2 units. OF MUTANT CHARACTERS. 227 The calculation of the locus of streak on the basis of all the available data places it at 13.6 units to the left of dachs and 15.4 units to the right of star, star being the zero-point for the chromosome. A summary of the linkage data directly involving streak is given in table 74. TABLE 74. — Summary of cross-over data involving streak. Loci. Total. Cross- overs. Per cent. Date. By- Reference. 396 63 15.9 Aug. 24, 1916 May — , 1914 Aug. 24, 1916 May — , 1914 Nov. 6, 1913 May — , 1914 Aug. 24, 1916 May — , 1914 Nov. 6,1913 May — , 1914 Feb. 23,1914 May — , 1914 May — , 1914 Aug. 24, 1913 Bridges Muller Bridges Muller Bridges Muller Bridges Muller Bridges Muller Bridges Muller Do. Bridges S' pf Streak dachs .... Streak black Streak purple Streak vestigial . . Streak curved . . . Streak blistered. . Streak speck Streak balloon . . . Streak morula . . . 0 ' Skd flies only; 4999-5110. Am. Nat., 1916, p. 422. S'.S' Prg £ Skf 462 396 45 64 9.7 16.2 Sid flies only; 4999-5110. Am. Nat., 1916, p. 422. Si', Si Pr c balanced B. C.; II 103-11 124. Am. Nat.. 1916, p. 422. at. S' Pr B C • St 858 109 12.7 462 120 26.0 1,807 462 396 632 137 114 35.0 29.7 28.8 Std flies only; 4999-5110. Am. Nat., 1916, p. 422. St', Si pr c balanced B. C.; II 103-124. Am. Nat., 1916, p. 422. b • ^* B C • 69 2,665 883 33.1 462 164 35.5 1,807 462 745 178 41.2 38.5 2,269 923 40.7 11 462 462 876 5 242 242 405 45.0 52.3 52.3 46.3 bt Am. Nat., 1916, p. 422. Do. Sf S* B C • mr II 64-11 97. VALUATION OF STREAK. There is only one drawback — but that one very serious — to the use- fulness of streak, namely, the difficulty of separating all the streaks from the wild-type. In most of the experiments conducted by Bridges this difficulty was not great enough to impair the accuracy of the result. However, a streak dachs back-cross was attempted and abandoned, (Si' 7) \ — } the separation is not com- Sk d / plete in all cultures. In these cultures the first separation performed was that of the streak from the not-streak, without regard to the other 228 THE SECOND-CHROMOSOME GROUP character. Among the streaks the other mutant characters should be distributed in the same ratio as among all the flies, so that tolerably accurate calculations could be made using streak flies only, but such ex- periments are inefficient. In the "progeny test" experiment of Muller this difficulty was entirely avoided, since the easily determined presence or the absence of streak from a progeny-test culture was all that was required to classify each parent. Streak is not a character that can be successfully handled without quite extensive experience, and even under the best of conditions there is chance of error. In favor of streak is its location, which is very important as the link between star and the rest of the chromosomes. A favorable location far more than doubles the usefulness of a character, other things being equal. In viability and other features streak is satis- factory. COMMA. (Text-figure 79.) ORIGIN OF COMMA. In one of the F2 cultures from the cross of dachs by pink, there appeared a mutant character called "comma" (culture II 9, February 5, 1913), which consists of a pair of chitinous thickenings on the anterior dorsal part of the thorax (fig. 79). In shape these thickenings are like a pair of commas, lying back to back, with the blunt tails pointing posteriorly, and depressed below the general level of the surface. This character was confined very largely to the females, of which about 20 per cent showed the character; a few of the males also were commas. From the frequency of the commas it was concluded that the character was an auto- somal recessive, which was either very in- viable or, more probably, failed to show in all those flies which were homozygous. In either case the character was markedly sex-limited in the sense that under like conditions far fewer of the males than of the females showed the f* M f\ Y£\ f* I f*T* CHROMOSOME CARRYING COMMA. In the F2 of the dachs pink cross, the commas seemed to be distrib- uted at random among the pinks and the wild-types, while none were seen among the dachs. This was interpreted as meaning that the locus is in the second chromosome. To test this point, commas were out- crossed to pink of the third chromosome and to vestigial of the second chromosome and an F2 mass-culture was raised in each case (tables TEXT-FIGURE 79. — Diagram of "comma." OF MUTANT CHARACTERS. 229 75 and 76). Two or three out-crosses of the pink comma flies were also attempted to secure PI for a back-cross, but these failed. The F2 of the comma by pink gave only 8 commas, all females, among 440 flies. Of these 8, 3 were pink commas, so that comma was proved to be not third-chromosome. The fewness of the commas was partly due to the crowding of the mass-cultures, but more to the fact that the character fails to show on all the flies that are genetically commas. TABLE 75. — PI, comma 9 X pink d". FI, wild-type 99+ FI wild-type cfcf . Apr. 3, 1913. Wild-type. Comma. Pink. Comma pink. 9 ff 9 d? 9 c? 9 d" M 24 147 163 5 55 67 3 The F2 from the cross of comma by vestigial gave 18 females and 1 male with comma, but not one of these was vestigial. While these numbers are not large enough to prove that comma is second-chromo- some, the probablity is high that it is. TABLE 76. — PI, comma 9 X vestigial d". F\ wild-type 9 9 + FI wild-type cf cf. Apr. 3, 1913. Wild-type. Comma. Vestigial. Comma vestigial. 9 ef 9* cf 9 c? 9 wnile the fathers were eosin miniature. The cream males and females which appeared were much paler than cream a, though like cream a they were a light, translucent yellow with little or no pinkish tinge. None of the not-eosin flies were different in color from normal red flies. A careful examination of the stock of eosin miniature failed to show any flies that did not have the standard eosin eye-color, and no lighter eye-color has ever subsequently shown itself in this stock. It is evi- dent that the gene for the modification had been present in the wild- type flies of the lethal 2 stock, but had been unsuspected so long as eosin was not present as a base. The demonstration that the cause of the observed dilution of eosin was a gene behaving in inheritance like the other mutant genes was easily made. INHERITANCE OF CREAM II. One of these cream males was out-crossed to a wild female. Among the F2 flies the creams reappeared, and, as in the parallel case of cream a, the not-eosin flies were all indistinguishable from one another and from wild flies in color. The F2 result resembles that obtained with cream a, except that, as stated, the new cream was considerably paler; and it was further discovered that besides the creams, approximately 50 per cent of the eosin males were intermediates between eosin and this cream, that is, cream II diluted eosin even in heterozygous form, so that the eosin sons were visibly as well as genetically in the ratio 1 eosin : 2 eosin heterozygous for cream II : 1 eosin pure for cream II. The entire ratio, disregarding sex, approximated 12:1:2:1, the 12 being the red-eyed flies. 1This culture was part of a generation which succeeded generation Q, table 22, p. Ill, Morgan, 1914, and which gave results similar to the results of generations J to Q of table 22. OF MUTANT CHARACTERS. 241 STOCK OF CREAM II. From the F2 a few cream males were selected and bred to their sisters, all of which were wild-type in appearance, though a quarter of them were homozygous for the cream gene (not-eosin creams). This mass-culture gave the expected cream females and males, from which a pure-breeding stock was made up. There was a difference in the color of the males and females of this pure stock, the difference being of the same order as the normal bicolorism of eosin. A complete separation of the eosin from the eosin heterozygous for cream had not been attempted in the original F2 culture. In order to observe the heterozygous condition more closely a cream male from the pure stock was out-crossed to an eosin female. The FI flies both males and females (culture M 68, intermediate males 73, intermediate females 88) 1 were lighter in eye-color than standard eosin, though the difference between eosin and these heterozygotes was less than the difference between the heterozygotes and the pure cream. TABLE 89. — Fz offspring from the cross of a cream male to an eosin female. Females. Males. Dec. 6, 1913. Eosin. Hetero- zygous cream. Cream. Eosin. Hetero- zygous cream. Cream. M 77 30 57 29 29 46 25 M 78 19 49 16 14 30 15 M 79 23 43 19 18 34 20 M 95 14 32 11 13 27 13 Total 86 181 75 74 137 73 Among these F2 offspring (table 89) there were six different eye- colors; among the males, the same three that occurred in the original F2, and among the females three colors which, though corresponding genetically to the classes among the males, were darker in eye-color. The cream female is lighter than the eosin male, while the heterozy- gous cream female is somewhat darker than the eosin male. In order from the darkest (a deep slightly yellowish pink) to the lightest (a pale translucent yellow) the six colors are: eosin female, heterozygous cream female, eosin male, heterozygous cream male, cream female, cream male. The females were first separated from the males. Then 1 One of the 88 intermediate daughters had only three segments to her abdomen instead of the usual five. This female (figured by Morgan, 1915, p. 425, text-figure 3a) was the founder of a new type of abnormal segmentation of the abdomen — "patched." The segments were reduced in number (as in the first specimen), or, more typically, were cut sharply into oblique or triangular pieces which were patched together as illustrated in figures 6 to/, of plate 11. This character was recessive, but it generally reappeared in very much less than a quarter of the F2 offspring. The usual causes for such deficiencies are poor viability, partial or complete dependence for realization on the coaction of one or more other genes, or failure to be developed in all the flies fluctuations of pure for the gene, whether from environmental differences or because the normal genetically the character overlap the wild-type. The gene for patched was in the second chro- mosome, as shown by its strong linkage to the cream. 242 THE SECOND-CHROMOSOME GROUP in each sex the pure creams were separated from the others, and finally the more difficult separation of heterozygous cream from eosin was undertaken. The separation of the creams from the other colors is accurate, but the final separation, that of the heterozygous creams from the eosins, must be regarded as only a close approximation. The sharp 1:2:1 ratio (160 : 318 : 148) which was obtained from this separation probably represents among the eosins a small number of the darkest heterozygotes, while the lightest of the pure eosins were likewise classified among the heterozygotes. Probably 10 per cent of the combined eosin and heterozygous cream class overlapped enough so that the separations might or might not correspond to genetic differences. One test of the correctness of the classification of inter- mediates was made. From culture M 79 a heterozygous male and a heterozygous female were selected, and the results (culture M 75) showed that both individuals were of the supposed type. No attempt has been made to secure a stock homozygous for the cream gene but without eosin. The cultures and experiments in which such not-eosin creams must have constituted one-fourth of the wild-type flies prove that such a stock could not be distinguished by inspection from a wild stock. That the action of cream II is specific to eosin was suggested by crosses of cream with vermilion (X chromosome) and with pink (third chromosome). A careful examination of the F2 flies from these crosses showed no dilution of either vermilion or pink by the cream, that is, the double recessives vermilion cream and pink cream (not-eosin) are indistinguishable from vermilion and pink respectively. LINKAGE METHOD OF ANALYSIS FOR MULTIPLE-GENE CASES. The proper method of study for cases of multiple factors or of modi- fiers is by means of linkage experiments, whereby all guesswork as to the number and effect of modifiers can be eliminated. In Dro- sophila such a study is rendered particularly easy by the fact that in the male there is no crossing-over of any of the chromosomes. In con- sequence, if two recessive genes which belong to the same chromosome, e. g., black and vestigial of the second chromosome, enter the cross from opposite parents ("repulsion"), the F2 never shows flies which have both these mutants at the same time. The double recessive class is entirely unrepresented, and the 2 :1 :1 :0 ratio of "absolute repulsion" results. This ratio holds, whatever may be the amount of crossing-over in the female, for the lack of double-recessive sperm prevents the double-recessive eggs from revealing themselves. This ratio is in marked contrast to the 9:3:3:1 ratio, which obtains when the two genes belong to different chromosomes, e. g., curved of the second chromosome and ebony of the third chomosome. The light color cream was known to be eosin plus a recessive modifier which belonged to an autosome group. To find whether this group OF MUTANT CHARACTERS. 243 was that of the second chromosome, a cream male (from pure stock) was out-crossed to a curved female, curved being a recessive wing- character whose gene is known to belong to the second chromosome (Bridges and Sturtevant, 1914). A pair of FI wild-type flies inbred gave the results of table 90. TABLE 90. — F2 from the cross of cream II male to curved female. Feb. 20, 1914. Not-eosin (cf+ 9). Eosin (males only). Wild- type. Curved. Eosin. Eosin curved. Cream II. Cream II curved. 70 155 64 37 14 15 0 Since cream only shows itself where eosin is already present, we may disregard all the flies of culture 70 except those with eosin eyes. These eosin flies are obviously in the ratio 2:1:1:0 which is expected if the cream gene is in the second chromosome, through the flies are too few to prove the point. TABLE 91. — Pi, cream II c? X eosin black 9 . FI heterozygous cream 9 X FI heterozygous cream cT . FI cultures. Mar. 16, 1914. Heterozygous cream 9 9 • Heterozygous cream c? cT • 119 51 15 41 18 329 Total 66 59 Fa cultures. Aug. 3, 1914. Eosin. Eosin black. Cream II. Black cream II. 9 0* 9 c? 9 tf 9 <7 372 50 57 69 48 38 42 79 59 24 18 36 33 15 31 43 17 14 19 43 24 25 19 39 24 0 0 0 0 0 0 0 0 398 399 400 Total 224 218 111 106 100 107 0 0 442 217 207 0 A more efficient experiment than this last was carried out by making all the flies of the experiment eosin, in which case the 2:1:1:0 ratio involved all the offspring rather than only a quarter, as in culture 70. A stock of eosin black was made up (black being a second-chro- mosome mutant) and a female of this stock was outcrossed to a cream II male. The FI and F2 results are given in table 91. All of the FI flies and half the F2 flies were of the intermediate color of the heterozy- gous cream. In the F2 these intermediates were classified along with the pure eosins, so that the cream was treated as though a strict recessive. 244 THE SECOND-CHROMOSOME GROUP The F2 ratio of 442 :217 : 207 :0 is a very ctose approximation to a 2:1:1:0 ratio and proves that the gene for cream is in the second chromosome (cream II). A similar experiment in which cream was crossed to eosin-ebony (ebony being a third-chromosome mutant, see Sturtevant, 1914) gave a typical 9:3:3:1 ratio (table 92), which agrees with the fact that the cream gene is not in the third chromosome. TABLE 92. — F2 from the cross of cream II o* by eosin ebony 9 . Mar. 31, 1914. Eosin. Cream II. Eosin ebony. Cream II ebony. 154 61 21 18 4 161 85 28 35 14 162 134 37 36 10 Total... 280 86 89 28 In order to find the locus of cream within the second chromosome it would have been necessary to run two linkage experiments in which all the flies were eosin; thus, one of these might have been cream II by eosin black and a back-cross of the FI female to black cream males, and the other a similar back-cross in which curved was used in place of black. The amount of crossing-over between black and curved was known to be about 27 per cent. The two values black cream and curved cream which would be found by two such experiments (both values might, of course, be found from a single suitably devised experi- ment) would enable the locus of cream to be calculated with consid- erable accuracy. While much is to be learned of the mechanism of crossing-over from a study of the relative distributions of loci within various regions of the chromosome, yet in the case of cream II it was thought that the compensation would not be worth the effort. Any further use of cream II in other linkage experiments would involve the "eosinization" of all the stocks used. In the case of certain of the later creams, an approximate location of the gene within the chromo- some has been made, but such location was made easier by the dis- covery of certain dominant mutations which were not available at the time the work on cream II was finished. TREFOIL (//). (Plate 5, fig. 6; plate 8, fig. 1.) ORIGIN AND STOCK OF TREFOIL. The character trefoil was found by Morgan about November 1913, and a pure-breeding stock was secured without difficulty. DESCRIPTION OF TREFOIL. The character trefoil is quite variable in the intensity of pigmenta- tion, as is the case with all of the thorax pattern characters. The OF MUTANT CHARACTERS. 245 distribution of the pigment is very definite, however. The scutellum is largely or entirely dark and the base of the trident pattern on the thorax is broader and much heavier, while the prongs are scarcely darkened at all. The characteristic feature of trefoil is the presence of extra basal sections to the trident outside the regular region. These side areas are fully as dark as the central parts and extend forward even farther. Another region that is dark in trefoil but not in the other thorax patterns is a patch behind each eye on the back of the head. These eye-patches and the side-prongs are the main characters used in classifying trefoil. INHERITANCE OF TREFOIL. In out-crosses to wild, trefoil behaved as an autosomal recessive, giving only wild-type flies in FI, and reappearing as about a quarter of the F2 flies. CHROMOSOME CARRYING TREFOIL. F2 from the cross of trefoil to pink gave a 9 : 3 : 3 : 1 ratio, while the corresponding cross to curved gave a 2 : 1 : 1 : 0 ratio, from which it was seen that the locus of trefoil is in the second chromosome. LOCUS OF TREFOIL. It was found that trefoil and black together gave a very dark fly which was distinguishable from black. With some difficulty a triple- recessive black strap trefoil stock was made up to test the locus of tre- foil. This stock was never used, but from the indications met with in its synthesis it seemed probable that the locus of trefoil is not far from that of black, but between black and strap. This was confirmed roughly by a star trefoil back-cross which gave 42.1 per cent of crossing- over (S 34, tf 55, S tf 42, +23). The locus is thus at about 50.0 with reference to star. VALUATION OF TREFOIL. Considerable difficulty was met with in the classification of trefoil from the variability of the character, and for this reason there was no strong incentive to establish its locus or to use it in any way. CREAM b (cre). (Plate 5, fig. 11.) ORIGIN OF CREAM b. An eosin female from a stock of non-disjunction, when mated to a bar male, gave (culture 82, March 10, 1914) among the eosin sons one whose eye-color was as light as that of cream II or cream III. This male was out-crossed to a wild female and in F2 gave creams among the eosin sons but no disturbance of the color of the not-eosin flies (cultures 183, 184, 185). The F2 ratio was again 12 : 3 : 1, as in similar 246 THE SECOND-CHROMOSOME GROUP crosses with other recessive specific dilutors. But the creams (cream b) which occurred in this F2 were not as pale as any of the preceding creams. From the circumstances of the appearance of cream b, viz, that it was observed in the FI of an out-cross and that as a single individual, we should expect it to be a dominant, but as a matter of fact it proved to be a recessive. It seems probable, in explanation, that more creams were actually present in this FI but were overlooked, since attention was distracted by the simultaneous appearance in the same culture of still another mutation (lethal 4), and more especially since the effect of cream b is rather slight. Only occasionally was one of the F2 creams so marked as the grandfather, and the mutation might not have been recognized at all were it not that an extreme fluctuant had attracted attention. Since cream b is recessive, we must suppose that the gene was present in both parent stocks. It could have been present in the bar stock and been undetected because of the lack of eosin, without which it has no visible effect; and the character might readily have been present in the eosin non-disjunction stock and have been passed over as an age variation, since, as we ordinarily see flies from a stock culture, they are of all ages and of all corresponding densities of pig- mentation. CHROMOSOME CARRYING CREAM b. A pure-breeding stock was made up for use in back-crossing. By this time we were in possession of a good second-chromosome domi- nant "star" and likewise of a perfect third-chromosome dominant TABLE 93. — B. C. offspring from the PI mating of an eosin star dichcete male to a cream b female and the back-crossing of the FI eosin star dichcete male to cream b females. Sept. 8, 1916. Non-cross-overs. Cross-overs (in male). Eosin star. Eosin star dichsete. Cream b. Cream b dichaete. Star cream b Star cream b dichaete. Eosin. Eosin dichsete. rr/9 .- 20 19 14 14 25 21 21 15 26 21 18 17 12 26 26 24 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5155\?.- r,,™/9 -. 5409\c?.. Total 67 82 82 77 0 0 0 0 "dichaete," which mutants have now become the most important in their respective chromosomes. By aid of these two dominants it is very easy to determine in a single experiment whether a given mutant is in the second or third chromosome. Thus, in the case of cream b, a stock of eosin star dichsete was made up and used in making a PI cross to the cream. Then FI eosin males which showed both star and dichsete and which were heterozygous for the recessive cream were back-crossed OF MUTANT CHARACTERS. 247 to cream b females of stock. There is no crossing-over in the male of Drosophila, so that if cream b were in the second chromosome not one of the B. C. star offspring should be cream, while half of the dichaete should be cream and half not. If, on the other hand, the cream were in the third chromosome, then none of the B. C. dichsetes should be cream, while the star and cream should assort at random. The experiment proved that the gene for cream b is in the second chromosome (table 93). LOCUS OF CREAM b. An (eosin) star female and a cream b male selected from the B. C. offspring gave in the next generation the amount of crossing-over between star and cream b (table 94) . This value of 22.1 includes some double crossing-over, and the corrected or "map" distance is probably about 22.5. The chances are in favor of the cream b locus being to the right of star, since star happens to occupy the leftmost of the known loci. TABLE 94. — B. C. offspring given in F3 by an eosin star female and a cream b male from table 93. Oct. 20, 1916. Non-cross-overs . Cross-overs (in female). Eosin star. Cream b. Star cream b. Eosin. (Q 36 22 47 28 41 41 39 49 8 12 13 11 8 12 7 15 5593< \" 6532{cf.'::::: Total.... 133 170 44 42 PINKISH. (Plate 5, fig. 12.) ORIGIN OF PINKISH. In the fall of of 1913 a stock of eosin black had been made up with which to test the chromosome group of cream II. In the following summer (July 27, 1914) Bridges noticed that a few of the males were somewhat lighter in eye-color than the others, but seemed chiefly noticeable because of the weakness of the yellow component of the eosin eye-color. The color of the regular eosin male is a pinkish yellow; the color of cream a, II, III, and b is nearly a pure yellow with little of the pinkish tinge, while this new color was somewhat the converse of this and was a pale pink with relatively little yellow. One of these males mated to a sister gave all of the sons of this pinkish color and all the daughters of a similar color, which is, however, much harder to distinguish from standard eosin. It seems that this character is somewhat sex-limited in the same direction as eosin. 248 THE SECOND-CHROMOSOME GROUP Pure stock of the mutation had been obtained at once through the happy selection of a pure pinkish female which had been taken to be simply an eosin female of somewhat lighter eye-color because of being freshly hatched. CHROMOSOME CARRYING PINKISH. Since pinkish appeared in a stock of eosin black, material was on hand to test the chromosome group at once. Accordingly, black pinkish females were out-crossed to eosin males and the F2 eosin females, standard eosin in color, were back-crossed to black pinkish males. In the back-cross cultures half of the flies were not-black, and the not-black pinkish flies were seen to be less markedly "pinkish" in eye-color than the blacks. In the absence of black the eye-color was more nearly like that of the other creams, though the amount of dilution is less than in any of the other creams. In the first two of TABLE 95. — Offspring given by the F\ eosin-eyed daughters from the out-cross of black pinkish females to eosin males, when back-crossed to black pinkish males. Sept. 13, 1914. Non-cross-overs. Cross-overs. Black pinkish. (Eosin). (Eosin) black. Pinkish. 525 70 36 25 24 28 81 29 21 27 24 71 32 24 29 14 24 22 35 31 29 526 2424 2425 2426 Total .... 183 182 170 201 these back-cross cultures (table 95) males and females were classified together. Some question having been raised in regard to the accu- racy of the separation of pinkish from eosin among females, the cross was repeated and the readily classifiable males (last three cul- tures) gave the same result as before. It was seen that the new or cross-over combinations were as numerous (51. 4 per cent) as the original classes, and this independent inheritance was taken to mean that the gene for pinkish is not in the second chromosome. While this was a mistaken notion — the true relation being that the gene is so far away from black that in the female there is entirely free crossing-over — yet it led to the device of the efficient " double-mating" method of ridding a given stock of an undesired recessive. THE DOUBLE-MATING METHOD. If pinkish were in the third chromosome, then the presence of the black in the pinkish stock could be of no advantage, and might be a very serious handicap, since it would prevent the use of all our third- chromosome stocks containing ebony or sooty. The first step in the OF MUTANT CHARACTERS. 249 elimination of black was to mate together some of the not-black pinkish flies of table 95. One-third of the not-black offspring of such pairs should be of the desired kind — that is, entirely free from black. Our task was then to pick out from the mixture of pure grays and grays heterozygous for black some pure gray males. In this special case we were aided by the fact that black happens to be slightly dominant — that is, the heterozygous blacks are somewhat darker than the pure grays. While this difference is not marked enough to be used regularly in classification, it enables us to pick out by inspection a greater proportion of pure grays than we would get by random selection. Four such males were selected as being probably free from black and were mated to eosin females. Into the same bottle with each pair of these flies was put a virgin (red-eyed) black female. The offspring from these two mothers are easily distinguished, since they are eosin-eyed if from the eosin mother and red-eyed if from the black mother. The offspring from the black mother constitute a test of whether the father were free from black, for in this case none of the red-eyed offspring hatching in the double-mating culture should be black, while if the father were heterozygous for black half of the red-eyed offspring should be black. Only one of the four cultures gave black offspring, and this culture was then discarded. . . . The eosin- eyed flies of the other three cultures were all heterozygous for pinkish, and at the same time free from black. By mating together some of these eosin-eyed flies pure pinkish offspring should be obtained as a quarter of the offspring. A more efficient method, and the one actually followed, was to save the fathers and mate them to their eosin-eyed daughters, since in this case half, rather than a quarter, of the progeny should be pure pinkish. TABLE 96. — B. C. offspring given by the FI eosin star dichcete sons, from the out-cross of a pinkish female to a star dichcete male, when back-crossed to pinkish females. Aug. 25, 1916. Non-cross-overs. Cross-overs (in the male). (Eosin) star. (Eosin) star dichffite. Pinkish. Pinkish dichsete. Star pinkish. Star pinkish dichaete. (Eosin). (Eosin) dichaete. ™J9 .. 10 17 22 20 12 13 20 20 10 16 18 21 9 22 21 26 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5029V-. .„--/$ .. 5266< V Total 69 65 65 78 0 0 0 0 In order to show by an actual test that the gene for pinkish is in the third chromosome, it was decided to take advantage of the fact of no crossing-over in the male and to run a back-cross test of a male hetero- zygous for pinkish and for the dominant third-chromosome character 250 THE SECOND-CHROMOSOME GROUP dichsete. It was now realized that the back-cross tests of females heterozygous for pinkish and black had not excluded the possibility of pinkish being in the second chromosome, though they had shown that, if so, it could be only in one or the other end-region and not near black. Accordingly, exactly the same procedure was followed as in the tests for the location of cream b, that is, a pinkish female was out- crossed to a male which had the second-chromosome dominant star as well as dichsete. The FI eosin star dichsete males were then back- crossed to pinkish females. The result showed (table 96) that the gene for pinkish is in the second and not the third chromosome; for, as well as could be judged, none of the star flies were pinkish, while all the not-stars seemed to be pinkish, and also dichsete was present in half of both the star and the pinkish classes. TABLE 97. — B. C. offspring given by a star female from table 96, when back-crossed to a pinkish male. Sept. 23, 1916. Non-cross-overs. Cross-overs. (Eosin) star. Pinkish. Star pinkish. (Eosin.) OT/9 .- 19 26 30 26 19 19 20 16 526v.. Total.... 45 56 38 36 TABLE 98. — Fz offspring given by the FI wild-type females and eosin males, from out-cross of pinkish females to wild males. Oct. 28, 1916. Wild- type. (Not-eosin) pinkish? Eosin. Pinkish. 5678 121 5 76 24 5680 52 2 46 17 5703 63 14 59 18 5704 57 6 52 26 5705 80 3 67 18 Total.... 373 30 300 103 LOCUS OF PINKISH. In the light of this test, and from the fact that there was about 50 per cent of crossing-over between black and pinkish, we could place pinkish in either the extreme left or the extreme right end-region of the second chromosome. Fortunately, one advantage of the test just described is that it left us in possession of females heterozygous for star and for pinkish, and a back-cross test showed (table 97) that there is very free crossing-over between star and pinkish. Pinkish is known, therefore, to be in the right-hand end of the second chromo- some, in the neighborhood of arc, speck, balloon, etc. Had the test given almost no crossing-over between star and pinkish, we should OF MUTANT CHARACTERS. 251 have known that the gene for pinkish was in the left end, but this was not the case. A test as to whether the pinkish gene would have any visible effect in the absence of eosin showed (table 98) that in a very small per cent- age of the flies homozygous for pinkish there is a very slight dilution. This dilution is, however, so slight that rarely could one be sure that the effect observed is due to dilution rather than to the slight normal fluctuation of the red. PLEXUS (Text-fig. 80.) ORIGIN OF PLEXUS. The venation character "plexus" was found by Bridges in a stock culture of the third-chromosome recessive spread wings (August 20, 1914, culture 557). Fully 10 per cent of the spread flies showed the plexus venation. DESCRIPTION OF PLEXUS. The most striking feature of plexus is a rather tangled knot of extra veins near the distal end of the fifth longitudinal vein and the posterior cross-vein, and another such knot near the distal end of the fourth longi- tudinal vein, with an extra vein running near the mar- gins of the wing and con- necting the two (see text- figure 80). Several other small sections of extra vein are often present in various parts of the wing, most of them lying free in the cells, but some being branches of or connected to the regular veins . These veins are all sharp and clear, without indefiniteness and discoloration such as char- acterize the extra veins of balloon. There is a very characteristic bend forward in the fourth longitudinal vein before it reaches the marginal vein. In less extreme individuals the connecting vein may be quite absent and the knots much reduced. The branch from the posterior cross- vein running parallel to the fifth longitudinal vein and also the bend in the fourth vein are the most persistent of all the characters. TEXT-FIGURE 80. — Plexus venation. 252 THE SECOND-CHROMOSOME GROUP INHERITANCE AND CHROMOSOME OF PLEXUS. One of the plexus spread males was out-crossed to a black female and produced in FI only mid-type flies, showing that the character was recessive. Several F2 cultures were raised from pairs of the FI flies, and in these plexus reappeared as about a quarter of the F2 indi- viduals. The plexus venation was present equally among the F2 females and males, from which it was known not to be sex-linked. Likewise plexus was known not to be third-chromosome from the free recombination of plexus and spread among the F2 individuals. This distribution of plexus and spread might have been thought due to very free crossing-over between spread and plexus instead of the random assortment, were it not that the black and plexus appeared in the typical 2:1:1:0 ratio, which showed that the locus of plexus is in the second chromosome. From the F2 cultures black and plexus flies were crossed to each other, with the two-fold purpose of obtaining the double-recessive form and of elimination the third-chromosome recessive spread. LOCUS OF PLEXUS. The double recessive black plexus was easily obtained, and a back- cross test was made of the amount of crossing-over in the female between these two loci (table 99). The back-cross cultures furnished 1,026 flies, of which 417 or 40.6 per cent were cross-overs. This very free crossing-over located plexus in the region of arc, if its locus were to the right of black, or at a point even farther to the left than streak if it were to the left of black. TABLE 99. — PI, black plexus 9 X wild cf ; B. C., type 9 X black plexus cf. I wild- Jan. 1, 1915. Black plexus. Wild- type. Black. Plexus. 1084 107 84 68 42 1085 58 58 47 51 1086 100 111 62 59 1099 48 43 45 43 Total.... 313 296 222 195 A means of easily testing these alternatives was soon afforded by the discovery and location of "star," which proved to be a dominant mutation whose locus is some distance to the left of streak. A star black plexus back-cross was made by testing the FI star daughter, from the mating of a star female by a black plexus male, to black plexus malqs (table 100). Four such pair cultures gave 1,352 offspring, of which 233 were double cross-overs, 343 simple cross-overs between black and plexus, 289 single cross-overs between star and black, and 487 were non-cross-overs. Plexus gave 42.6 per cent of total observed OF MUTANT CHARACTERS. 253 crossing-over with black and 46.7 per cent with star, from which it was known that the locus of plexus is to the right of black. A com- parison of the combined black plexus cross-over value of 41.8 with other values involving black indicated that the locus of plexus was closest to arc, but probably not quite as far to the right, since black and arc gave 42.6 per cent of observed crossing-over. TABLE 100.— PI, black plexus cf X star 9 ; B. C., FI star 9 X black plexus cf . S' S' 1 b px S' 1 Px S' | b 1 b Pi 1 b I \ 1 Px July 20, 1915. Star. Black plexus. Star black plexus. Wild- type. Star plexus. Black. Star black. Plexus. 1921 81 81 42 48 43 52 32 21 1922 65 41 24 42 45 42 36 28 1923 57 52 35 39 42 40 29 43 1924 . . . 72 38 22 37 36 43 23 21 Total.... 275 212 123 166 166 177 120 113 The loci of plexus and arc were found to be so close together that it was too hard a task to obtain the plexus arc double forms by any simple method. The plexus speck double was obtained fairly easily. TABLE 101. — Summary of all crossing-over data on plexus. Loci. Total. Cross- overs. Per cent. Date. By- Reference. 1 152 632 46 7 Tiilv 90 1Q1 >J p • ^' B C • 1921-'24 82 39 47 6 Anril 3 IQIfi Do b PX Sf- ^' B C • 4044 b px Black plexus . 1,026 417 40.6 Jan. 1, 1915 Do. px; b pxB. C.; 1084-'99. p • ^' B C • 1921 '24 1,352 576 42.6 July 20, 1915 Do. b PX 38 46.4 Apr. 3, 1916 Do. fl; b piB-C-.4044- 2,460 1,031 41.9 Purple plexus. Plexus speck.. 344 327 164 29 47.7 8.9 Feb. 29, 1916 Feb. 29, 1916 Do. Do. II la; PT Px sp F2; 3535-'53. II la; pT Px sp F2; 3535-'53. Only one linkage experiment involving the plexus speck cross-over value is available, and this gave 8.9 per cent of crossing-over (table 135). The locus of plexus is therefore about 3 units to the left of arc, since arc gave 5.9 per cent of crossing-over with speck. This position makes plexus of great importance, since it can serve as a new base of reference for speck itself. At present speck is located by reference to curved, which is too remote to give an accurate measure 254 THE SECOND-CHROMOSOME GROUP of the interval, especially since the amount of double crossing-over and of coincidence in this region are known only by inference from other experiments, there being no intermediate locus by means of which direct calculation could be made. Arc could be used for this purpose, but is unsuitable because of a probable confusion in classi- fication with curved. Plexus, on the other hand, being a venation character only, can readily be used with curved, and its position is more favorable than arc, since it more nearly divides the gap. Preparations were made to make an extensive experiment which should give data on the plexus speck distance as well as on several others throughout the length of the chromosome. But as yet this has not proceeded further than the synthesis of the multiple recessive needed (d b pr c px sp), and of the two parent stocks necessary for an "alternated" experiment (S' b c sp and d pr pj. A summary of the crossing-over data imvolving plexus is given in table 101. The locus of plexus is about 8.9 units to the left of speck, or at 96.2. VALUATION OF PLEXUS. Sturtevant has reported trouble in the classification of plexus in certain crosses, and it is not certain that all the plexus individuals can be separated from wild-type where the variation is great. In none of the experiments here reported was difficulty encountered. LIMITED. (Fig. 81.) In carrying out the black arc morula back-crosses (culture 513, September 13, 1914), Bridges noticed that there was present a charac- ter somewhat similar to abnormal abdomen, except that its main effect was evident on the ventral surface of the abdomen and to a slight extent on the side. The chitinous ventral plates on the abdomen, instead of being full size with rounded edges and many regularly arranged small hairs (as in fig. 81), were reduced often to half the size by an irregular erosion of the edges. The color also was etched. The hairs were very few in number, and those irregularly arranged and directed. The dorsal plates where they bent around to the ventral side were affected in the same manner at their ends, though not so TEXT-FIQUHE si.— Limited strikingly. This character was almost entirely limited to the morula flies and was distributed in such a way that its locus was certainly second- chromosome and probably to the right of morula, though the counts were not made carefully enough to be sure of this. OF MUTANT CHARACTERS. 255 The black arc morula stock was found to be showing the limited band character in all or nearly all of the morula flies, and this condition has been maintained for some five years, which means that the linkage is very close. It is in fact not entirely certain that limited may not be found to be still another effect of morula itself. CONFLUENT (Q). ORIGIN OF CONFLUENT. In culture 550 (which was part of the tests of the method of trans- mission of the ability to produce exceptions by non-disjunction), Bridges found a single male which had thickened, knotted veins in the wing (September 23, 1914). The vein most thickened was the second longitudinal opposite the anterior cross-vein, but especially at the tip, where it was confluent for quite a space with the marginal vein. In addition both cross-veins were thickened and irregular. The wing as a whole was slightly smaller than usual and the fly seemed rather sluggish. INHERITANCE OF CONFLUENT. This male was out-crossed to a wild female and in FI produced nearly half of the offspring with confluent veins (culture 592, table 102). From this FI result confluent was known not to be sex- linked, since the characters appeared in half the FI males as well as females, instead of in none of the males and all of the females, as it would have done had it been a sex-linked dominant (like bar). TABLE 102. — Confluent cf (heterozygous) X wild 9 . Oct. 5, 1914. Wild- type 9 • Wild- type c?. Conflu- ent 9. Conflu- ent cf. 592 24 20 17 11 627 52 33 38 47 628 42 41 37 33 901 41 37 36 46 1038 10 11 12 14 Total.... 169 142 140 151 Some of the FI confluent males were again out-crossed to wild females, and all the cultures of similar character (table 102) gave a total of 600 flies, of which 291 or 48.5 per cent were confluent. Thus the viability of confluent is not bad, though the sterility is very high. Very many such matings failed, and this was especially true of the confluent by confluent matings. The confluent by confluent matings gave consistently about two- thirds of the flies confluent and one-third wild-type. This suggested that, like streak, confluent was lethal when homozygous. 256 THE SECOND-CHROMOSOME GROUP The stock was run by mass-cultures of confluent by confluent, and these also gave the same percentages of confluent. The stock was maintained (with difficulty) for two years by this method and there was no indication of an increase in the percentage of confluent, though no counts were made. These facts prove that homozygous confluents either die, as supposed, or else play only a negligible role through steril- ity if they occasionally survive. TABLE 103. — PI, confluent c? X purple curved speck 9 or X sepia peach ebony 9 . Oct. 23, 1914. Wild- type 9 • Wild- type d" . Conflu- ent 9. Conflu- ent cf. 629 6 3 6 8 732 10 5 16 11 716 5 8 5 2 717 40 25 28 17 Total 61 41 55 38 CHROMOSOME OF CONFLUENT. Confluent males were out-crossed to purple curved speck females (cul- tures 629 and 732, table 103) and to sepia-peach-ebony females (cultures 716 and 717), as PI matings for male back-cross tests to deter- mine the linkage relation of confluent to the second and to the third chromosomes respectively. The tests were made with great difficulty, owing to the sterility and low productivity of confluent. The second- chromosome tests gave a total of 71 offspring, none of which was a cross- over between confluent and any of the three second-chromosome loci (table 104). The third-chromosome tests gave recombination between confluent and the third-chromosome loci, although as soon as the results of the second-chromosome tests became apparent these third-chromo- some counts were discontinued. TABLE 104.— B. C., confluent FI d" (from taUe 103} X purple curved speck 9 . Nov. 26, 1914. Conflu- ent. Purple curved speck. 802 11 2 840 2 3 841 3 5 1149 6 19 1150 12 8 Total 34 37 VALUATION OF CONFLUENT. The very low productivity and high sterility of confluent made it evident that there was little use to be obtained from the mutant in spite of its dominance, good viability, and perfect separability. The OF MUTANT CHARACTERS. 257 determination of the locus was not made, though this would have been done had the stock not died out because of its low productivity and sterility. CONFLUENT VIRILIS. Metz (C. W. Metz, Journ. Gen., 1916, p. 591) found in the species Drosophila virilis a mutation which was a very striking counterpart of confluent of D. melanogaster in all respects, save that it was neither so sterile nor so non-productive. The character of the venation was practically the same in the two cases, though in confluent D. melano- gaster the venation may liave been a trifle thicker and knottier in the affected regions. Confluent D. virilis was a dominant which gave 1 : 1 ratios upon inbreeding, precisely as did confluent D. melanogaster. There is no doubt of the completely lethal effect of confluent virilis when homozygous, and in confluent melanogaster the only indication that an occasional homozygote may survive is the fact that 1 out of 10 of the flies successfully tested. by Metz gave a 27 :0 ratio of confluent to wild-type, instead of the 18 : 9 ratio expected. The other 9 flies tested by Metz were all heterozygous, as had been all those worked with by Bridges. It is possible that this 27 : 0 ratio was the result of a balanced lethal condition such as obtains in truncate, snub, beaded, and other stocks. The fact that several of the mutations secured in D. virilis (or other species) seem parallel in appearance and inheritance with the known mutants of D. melanogaster is of great interest as an indication of the basic similarity of the two systems of genetic materials. FRINGED (fr). (Text-figure 82.) ORIGIN OF FRINGED. In the F2 from a cross of the sex-linked wing-character ''jaunty I" to wild (culture 1042, January 20, 1915), Bridges found that about a quarter of the flies of both sexes were showing an irregular distribu- tion of the hairs on the marginal vein of the wing. The margin showed spots entirely denuded of hairs or with only weak hairs, while the remaining hairs were frayed and irregular in directions. The wings also were slightly smaller, a trifle discolored, and occasionally divergent. CHROMOSOME CARRYING FRINGED. One of the "fringed" males was out-crossed to a black female and produced in F2 the typical 2:1:1:0 ratio that showed that the locus of fringed is in the second chromosome (table 105). From the F2 black and fringed inbred a stock of black fringed was obtained in F4. A similar attempt to obtain a fringed speck double-recessive stock from the F2 of the cross of fringed by speck (table 106) failed entirely. 258 THE SECOND-CHROMOSOME GROUP LOCUS OF FRINGED. The black fringed stock, in combination with the recently mapped dominant star, offered a means of locating the position of fringed. A three-locus back-cross was started by mating a black fringed male to a star female and back-crossing the FI star females by black fringed males. The three back-cross cultures (table 107) gave a total of 496 flies, of which 153 were non-cross-overs, 133 single cross-overs between star and black, 141 single cross-overs between black and fringed, and TEXT-FIGURE 82. — Fringed wing-margin. 70 were double cross-overs. The black fringed cross-over value was 42.5, which places fringed at practically the same locus as arc, which gave 42.6 as the black arc cross-over value. To determine the locus more closely than this would require fringed- speck or fringed arc back-crosses, which have not been made. The only cross-over data on fringed are the values calculated from the star black fringed back-cross above, viz, S'b = 40.9, bfr = 42.5, S'fr = 55.2. OF MUTANT CHARACTERS. 259 In the spring of 1915 Morgan also found fringed, probably through use of the same wild stock from which it originally came in the cross to jaunty I. TABLE 105. — PI, fringed d" X black 9 ; FI mid-type 9 + FI wild-type d" . Feb. 13, 1915. Wild- type. Black. Fringed. Black fringed. 1361 216 100 96 o 1362 150 52 56 0 Total .... 366 152 152 0 TABLE 106. — Pi, fringed fr \ b 1 1 1 fr Oct. 23, 1915. Star. Black fringed. Star black fringed. Wild- type. Star fringed. Black. Star black. Fringed. 2282 36 33 24 28 20 26 13 15 2283 33 12 14 20 27 16 2 11 2284 24 14 19 28 19 33 16 13 Total 93 59 57 76 66 75 31 39 STAR (5'). (Text-figure 83.) ORIGIN OF STAR. In an experiment by means of which it was proved that the excep- tional sons produced through secondary non-disjunction are them- selves unable to transmit the power of producing further secondary exceptions (Bridges, 1916, p. 44), an eosin sable forked male was found which had an eye of the moruloid type and which was very similar in appearance to the sex-linked mutation " facet" (culture, 1347, February 12, 1915). INHERITANCE OF STAR. It was assumed that this character was sex-linked, since it had ap- peared in a single male in a pair culture, as is usual with sex-linked mutations. For this reason the matings for F2 were made without 260 THE SECOND-CHROMOSOME GROUP examining the character of the FI flies, and it was not until the F2 began to hatch that it was realized that the other alternative was correct — that "star," as the character was called, was an auto- somal dominant. Two of the F! pairs gave in F2 no star whatever (1627, 1629), while a third pair (1628, table 108) gave stars among both males and females to the extent of half the flies (52 per cent). The fact that half the flies were stars showed that this culture came from a heterozygous dominant and a wild- type FI pair. That star was an autosomal dom- inant was proved by the sister cultures which gave no stars; had star been sex-linked all the FI females would have been star and hence every F2 pair should have given results like those of culture 1628. These facts were confirmed by the results of further tests of star males; for star males out- crossed to wild females gave in FI stars to the extent of half the flies (table 109, 337 stars in a total of 683, or 49.3 per cent), and the stars were evenly distributed among the males and females. Had star been sex- linked, none of the males but all of the females should have been star. TABLE 108. — PI, star cf X wild 9; FI pair (Fifties chosen at random). TEXT-FIGURE 83. — Star eye, showing the arrangement of the facets and hairs. Com- pare with the normal con- dition shown in plate 10, figure 3c. Mar. 12, 1915. Wild- type 9 . Star 9. Wild- type cf . Star cf . 1628 44 39 33 45 TABLE 109.— Pi, star cf X wild 9 . Mar. 27, 1915. Wild- type 9 . Wild- type cf . Star 9. Star cf . 1719 129 136 129 115 1914 2 8 2 4 1915 2 7 a 2 1916 2 6 a 7 Total.... 34 6 33 7 LETHAL NATURE OF THE HOMOZYGOUS STAR. At the same time that the male out-crossed tests were made, a few pairs of star female by star male were mated. In the next generation, which corresponded to an F2, the flies in one culture (1739, table 110) were exactly two-thirds star and one-third wild-type, which is the typical yellow-mouse ratio that had already been met with in Dro- sophila in the case of streak. The other culture (1740, table 110) gave nearer to a 3 to 1 ratio. Further matings were necessary to be sure OF MUTANT CHARACTERS. 261 which ratio was really present. These further cultures left no doubt that the ratio was really the 2 : 1 ratio corresponding to a dominant lethal when homozygous. The total number of stars in such cultures was 766, which is 67.3 per cent o£ the total number, where 66.7 per cent are expected according to the lethal assumption. TABLE 110.— FI star 9 + FI star rf1. Apr. 9, 1915. Wild- type. Star. 1739 58 117 1740 25 71 1877 46 91 1878 29 77 2025 30 62 2026 71 139 7454 100 166 7455 12 43 Total 371 766 TABLE 111. — PI, star cf X peach sooty 9 . Mar. 31, 1915. Wild- type 9 • Wild- type d". Star 9. Star cf . 1730 48 39 43 47 B C., Fi star 9 X peach sooty cf1. Apr. 12, 1915. Not-star. Star. pv e* Pv p* e' P» 1 e* 1 «* Peach sooty. Wild- type. Peach. Sooty. Star peach sooty. Star. Star peach. Star sooty. 1745 19 6 16 4 6 18 11 3 17 4 10 27 3 1 4 2 6 9 1 3 8 1 3 14 16 6 15 8 5 17 6 6 16 8 4 19 5 2 2 2 3 10 5 2 4 3 2 5 1746 1747 1748 1749 1750 Total 69 72 25 30 67 59 24 21 The history of the stock likewise proved the lethal nature of the homozygote; for star flies were inbred for many generations in mass- culture without giving any closer an approach to a pure-breeding stock. Likewise none of the star flies selected for out-crossing on very numer- ous occasions ever proved to be homozygous; all gave the 1 : 1 ratio typical of a heterozygous dominant. Lately, much more vigorous tests have conclusively proved that star is lethal when homozygous. 262 THE SECOND-CHROMOSOME GROUP CHROMOSOME CARRYING STAR. To test the relation of star to the third chromosome, a star male was out-crossed to the double-recessive peach sooty (peach is an allelo- morph of pink, and sooty an allelomorph of ebony) . In FI the flies were, as expected, half stars and half wild-type (culture 1730, table 111). Some of the FI star females were back-crossed by peach sooty males (table 111). With two loci as far apart as peach and sooty were known to be, there was no need to run a male back-cross test, since the female test must readily reveal linkage to one or to the other of these two loci if the tested gene is in the third chromosome. As a matter of fact, there was complete independence of star and peach (52.6 per cent of recombination) and also of star and sooty (50.4 per cent of recombi- nation). Peach and sooty gave 27.3 per cent of crossing-over, which is a trifle higher than the usual value. LOCUS OF STAR. Since the locus of star was proved not to be in the third chromosome, the chances were about 50 to 1 that its locus was in the second chro- mosome. This probability was so great that an extensive experiment was planned and started without the relation to the second chromosome having been previously tested. This experiment was the quadruple TABLE 112. — PI, star 9 X purple curved speck cf. June 28, 1915.' Star speck. Wild- type. Star. Speck. 1806 45 51 40 50 1807 42 47 43 51 Total.... 1808 87 98 83 83 95 101 back-cross of star and purple curved speck, which was to serve several purposes. In the first place, it was to give an accurate measure of the amount of crossing-over between curved and speck, which was very important, since up to that time only a relatively small amount of data was available on this value whereby the locus of speck and with it the entire right end of the chromosome was mapped in relation to the rest; in the second place, it was to establish the locus of star, which, as was then realized, might prove to be the most useful of all the second- chromosome characters. These linkage values were both to be con- trolled and linked up by means of the accurately mapped loci purple and curved. The third purpose was to test more thoroughly the extent and nature of the change of crossing-over with age in different broods and in different regions of the second chromosome, but more especially the relation between this change and the change in the amount of coincidence (see Bridges, 1915). OF MUTANT CHARACTERS. 263 When the FI flies from the cross of star by purple curved speck began to hatch, a surprise was met with in that half of the flies were speck in two cultures (1806, 1807), but not in the third (1808, table 112). It had not been noticed before that there was any speck in the star stock; and it is not clear how speck came to be there, since no cross to speck had been made, and so far as known none of the stocks concerned in the history of star had contained speck even as an impurity. However, this circumstance gave an immediate test of the linkage relation of star and speck, since these two cultures constituted a star speck back- cross test of crossing-over in the female. The two cultures gave 369 flies, of which 184, or 49.9 per cent, were cross-overs. While this value is that corresponding to any locus as far from speck as black, or any other to the left of black, it is also the result one would obtain if star were not in the second chromosome at all. This possibility caused such concern for the experiment already planned that imme- TABLE 113. — F\, star speck d* X purple curved speck 9. July 19, 1915. Star speck. Purple curved speck. Star curved speck. Purple speck. 1913 60 73 1 0 diately a black-cross test of crossing-over in the male between star and purple curved speck was carried out. This male test proved that star is actually in the second chromosome, since of the 134 flies (cul- ture 1913, table 113) 133 were non-cross-overs, as opposed to the very free crossing-over in the female test. CROSSING-OVER IN THE MALE. But one fly was a cross-over, since it was distinctly a star curved speck male, while all other flies were either star speck or purple curved speck. This cross-over fly occurred on the sixth day of the counts of a pair culture, so that there is no possibility of overlap of generations; and no opportunity for contamination, since no possibility of star curved flies had ever existed in any other or previous culture. That classification and pedigree were both as recorded was proved by a test TABLE 114. — Cross-over star curved speck -llM »^ >H 1-1 ^ r-li-H rH^HINCOCOCO'-ITHrHr-lrtrt t) *> "-1 C « 02 S 3 ft ft o » i-ieOC<3'-i(NCO«)'-iC<) PL, g S1 O Oft Ok w OB Ob C« Cw O» Ob Oa O fl» O> w O» fl» XXXXXXXOOXX XXXXXXXXX OF MUTANT CHARACTERS. 267 3 o H i-HOiCi-iiCOOliCCOiCTfiON. ot^O(Nt>tor-aoa5coi-icoto (N--i W W-H«^ i-l •* (N (N (M n l-H i-< 1-1 a| ii u • O i— 1 u ft &2 A 00 jd § a 02 Xt>-COOiCOCOO5OiCMi-(CO 3 CO 1C 1-1 0 S-s.^ •S & fe * ii O, u O) ^O ^COOiOi— tcD^fCOOCCO Tt< i— 1 CO IN ^H o £ SQ A, 00 u TJ . ? ^ fe S 5& •^O CO IN CO n cc u £ _a| t-xicck»'-(CO'- W ft ^^ OlXOOl^CO^H-^INt^CO^X (N^i-iC^ CO -a • S-D.S-S -S >- C a> ffi §.§& XOt^lNXiCi-i^iNi-iQOiCCO i-i(N (N (MCOCO(Ni-i(M 1^ CO (N t> ic co i-i to 5s A 09 U £ 05 T3 • &1« 53ft PL, 5 oo O^i-i^MWCOCO-HOooOO* TfiCO'-HC^i-'i-i-C «D t- o CO «c 1-4 9 1-1 r-T CO >. •-5 • ~O G 13 p (Nco^icffit^ooasoTficcct- apooaowmaoaowammecM O Oi O^ Ci Oi Oi O5 C5 C5 O O O O "5 1 Fourths. (N •<* CO Tf 1C *C 1C ot~-c5Oi-i O* Oft Oft CB Co w 9 Cv CB O Thirds. - prB C Sk flies only 389 86 22 1 Oct 20 1916 Do Std ' 4999-5116.' o/ c , . ° , B C • 5593+5824 Star truncate 549 149 27.1 May — , 1917 Wallace crb Snub; p. 143, this paper. Star dachs. . . 96 152 1,617 369 31 53 425 32.3 34.8 26.3 on A Sept. 12, 1915 Sept. 12, 1915 Sept. 15, 1915 Oct 6 1915 Do. Do. Do. Do di; — F2, dachs flies; 2141- d 2216. idem, not-dachs flies. S'; — B.C.; 2146-2305. d rli- "PV 9917 Ifi^Q 211 1,027 57 271 27.1 26 4 Nov. 18, 1915 Aug 24 1916 Do. Do. dl dl; ^— B.C.; 2460. di gt. S' PT g £ • 4999-Slin, Sid 3,472 949 27.3 Star black. 1,352 522 38 6 Jan. 1, 1915 Bridges p • ^' B C • 1921 '24. 496 203 40 9 Oct 23 1915 Do. bpx f • ^' B C ' 2282 '84 865 315 36 4 Oct. 26, 1915 Do. b fr M; — B C • table 122, 690 266 38.6 Dec. 22, 1915 Do. b vgn this paper. di • ^' dl B C • 2^79-70X5. 13,104 4,944 37.7 Dec. 5,1916 Plough. b J. E. Z., '17, p. 147; tempera- ture; — B.C.; table 7 (22°), be 16,507 6,250 37.9 86(27°), 86(22°), Hl(22°), 173. Star trefoil . . Star apterous 154 205 65 88 42.2 42.8 Aug. — , 1917 Nov. 18, 1916 Morgan Bridges ap; ^—F2; p. 239, this paper. Star purple. . 6,766 3,010 44.5 July, 11, 1915 Do. S'- B C Ists- 1836 1,027 413 40.2 Aug. 24, 1916 Do. Pr c sp 1894 £/. S' Pr B C • 4999 5110 362 138 38.1 July 21, 1917 Do. 1 Sid 0 Sf; B.C., construction; 8,155 3,561 43.7 OF MUTANT CHARACTERS. 271 TABLE 122. — Summary of all cross-over data involving star — continued. Loci. , Total. Cross- overs. Per cent. Date. By— Reference. 'i-il 450 iqe; 4.0 q Nov 17 1915 D/ Van- B C • table 123 fi 7fifi Q 1 p. A 4fi 8 July 11 1915 Do b v(in this paper. Q/ S'' B C Ists' 1836 13,104 5,959 44.4 Dec. 15,1915 Plough prcsp 1894> J. E. Z.; '17, p. 147; tempera- Of tures; — B.C.; tables 7 be (2Z°), 86(27°), 86(22°), Hi 19,870 9,123 46.9 (22°), 17,. Star telescope 531 236 44.4 May 21, 1916 Bridges <«; — B.C.;4632-'35. <» S' T)T' R O • 1OO1 >r>A AQfi 632 O7/I ^ o n/»+ 97 IQI ^ T)n 6 p* .C' /?•• B P • °r> + c? / 74 28 0 0 2487... / 9460 + * C 8' + o \ 53 88 58 66 0 16 0 41 2458 + 4l ' S' + o 62 71 + + ^ If a fly were carrying " dachs-deficiency' ' in one second chromosome and star in the other and were out-crossed to dachs, then half the off- spring should be dachs, since these flies should carry the dachs gene in one second chromosome and in the other no normal gene to oppose its action. When the test was made half of the offspring were dachs and half were not (table 127). In appearance these dachs flies f — j were indistinguishable from the dachs flies of regular stock. The dachs flies were distributed according to the usual linkage relations of star and dachs. One pair failed to give dachs offspring (2458), corresponding to the crossing-over that occurs normally between star and dachs whereby a certain proportion of the star descendents are not heterozygous for the lethal. BALANCED LETHALS. It was obvious that the stock had to be carried on as a recessive autosomal lethal — that is, by mating together two flies each heterozy- gous for the lethal. The most advantageous method of doing this (Of \ r ), dt/ since in this case advantage could be taken of the fact that most of the flies which would be homozygous for not-lethal would at the same time be homozygous for star and hence be eliminated. Most of the stars in each generation would continue to be heterozygous for the lethal. 280 THE SECOND-CHROMOSOME GROUP Only those which resulted from crossing-over between star and the ocus for dachs would fail to carry the lethal. If these two loci had been closer together, then fewer such cross-overs would occur and selection could be correspondingly relaxed. In the case of a pure breeding stock of " beaded," Muller found that there was an autosomal lethal in the not-beaded third chromosome, and very close indeed to the locus of the beaded allelomorph, so that no selection at all was needed. This principle, first used consciously in carrying on the stock of dachs-lethal, was called by Muller "balanced lethals" as worked out by him in the analysis of the beaded stock. Muller has shown that this principle has wide application, and may solve some of the knotty problems of the genetics of Oenothera, such as pure- breeding heterozygotes, the continual production of rare forms called mutants (which by this principle are due to crossing-over rather than to a fresh occurrence of the imitative process), and also the appearance of twin hybrids from certain crosses. It was quickly recognized that the dachs-deficiency explanation was alternative to that of a simple recessive autosomal lethal occurring in a locus close to that of dachs, the recessive dachs gene being present and unchanged, but prevented from giving rise to the dachs character, because all (or nearly all) of the homozygous dachs flies were also homozygous lethal, and hence never appeared as adults. All of the three parallels to forked-deficiency were equally explainable on the linked lethal view. A possible method of distinguishing between the two conditions was offered by the appearance or non-appearance of dachs flies upon inbreeding. If the phenomena were due to dachs- deficiency, then, of course, no dachs could ever appear, since the lethal effect involved the dachs locus itself. But if two separate and dis- tinct loci were involved — dachs and a neighboring lethal locus — then by crossing-over between them dachs should reappear. For this reason a most careful count was kept of the early stock cultures, which were run by the method of inbreeding. For four generations this was continued (table 126) and not a single dachs fly appeared among the flies. Besides the pair cultures recorded in table 126 (which were necessary in order to avoid all danger of losing the stock by crossing- over between star and the lethal), many other mass-cultures were raised for the purpose of giving full opportunity for dachs to reappear. These were not counted, since the composition of the parents was of two sorts and the ratios correspondingly confused. Approximately 5,000 flies were examined, however, without finding any dachs. The appearance of a dachs fly would have established the linked lethal view; but the non-appearance of such flies did not prove the deficiency view, but only that if a linked lethal were present its locus was extraordinarily close to that of dachs. Such an appearance of a dachs fly would be parallel to the appearance of certain Oenothera "mutants," according to the application made by Muller. OF MUTANT CHARACTERS. 281 There was another possible method of distinguishing between these views, which was tried. It had been found that the occurrence of the forked-bar deficiency had distributed the linkage relations in the first chromosome in a definite way. All crossing-over in the region between forked and bar was eliminated, as proved by direct tests with forked and bar, and likewise by tests of the decrease in the amount of crossing- over from that which nominally occurs between the nearest loci on either side, namely, rudimentary and fused. In the case of dachs- lethal there was no other gene known to be included in the deficient region itself, so that direct tests were impossible; and even worse, there were no loci close enough to dachs to give a measure of the decrease unless it were very marked. It was thought possible that a rather extensive disturbance might be initiated by a relatively short deficiency, since the shortened chromosome might well prevent perfect synapsis for a much longer region because of the " pucker." The only practical but unsatisfactory test that could be made was through black, which was the nearest workable locus to the right, and star, which was the only locus to the left that could be used without inaccuracy. If the dachs locus were deficient it could still be controlled by means of the haploid dachs flies produced by testing with dachs. Thus the proposed experiment involved a female carrying a star dachs-defi- cient second and a black second chromosome, to be tested by means of a dachs black male. Star and dachs were put in the same chromo- some because that method was far easier, and also because the recip- rocal back-cross (- — cf X6 9 ) offered a means of carrying on the dachs-lethal stock with no opportunity for crossing-over, since the only heterozygote was the male, in which sex no crossing-over occurs. The stock had been run with star and dachs-lethal in opposite second chromosomes. To get them into the same chromosome a female (Sf \ - ) was out-crossed to a dachs male. The star dachs cross- di/ over offspring contained a chromosome of the desired composition (SI' H \ — ^ — J. To be sure of retaining this chromosome, and not getting a plain star dachs chromosome by crossing-over, a male of this type was used in the next step, which was an out-cross to black (culture 2613). All of the star offspring of this cross were of the desired com- position ( - * J . Stock was started by mating such males to black females and repeating each generation (table 128). No special pains need be taken to see that the females are virgin in this stock, which is an advantage not possessed by the similar selected stocks of the sex- linked mutants where the heterozygote has to be the female and virgin. 282 THE SECOND-CHROMOSOME GROUP The females back-crossed to dachs black males produced the results shown in table 129 (cultures 2679, 2680, 2681). The crossing-over between star and dachs was found to be normal, but the crossing-over between dachs and black was the lowest ever encountered (outside recognized linkage variations), being only 11.1 per cent, while the mean calculated from 6,725 other flies was 17.8 per S' dl TABLE 128. — Cultures of dachs-lethal stock, c? X 6 9 Dec. 19, 1915. Star. ' Black. 2655 55 48 2656 12 8 2657 135 151 2761 148 163 Total .... 350 370 cent. This same experiment was repeated in 1917 (cultures 7083, 7085, table 129) and the same result was obtained, since the crossing-over between dachs and black was only 10.2percent. The lowest regular value found for dachs black was 16.7 (found by Muller in his progeny tests.) It would seem that these values, which represent a decrease of 39 per cent from the standard values, are sufficiently aberrant to prove that the case is not to be explained by a simple lethal, linked to dachs. TABLE 129.— PI, star dachs-lethal 9 (— ~r)x black — , ) \+ + + +l c ' /fe++\ ' few black flies would be produced I - ) ; likewise the few curved \6 I c / 6 I c / flies corresponded to crossing-over between the lethal and curved. If this were the explanation, then most of the wild-type flies should be of the same constitution as the FI flies ( - - ] together should repeat the F2 result. \+ + +/ and when mated 288 THE SECOND-CHROMOSOME GROUP Seven pairs of F2 (or more likely F3) wild-type flies gave offspring (table 132). Of these pairs, two (2840, 2863) proved to have both parents of the original constitution ( - - J. Two others (2861, 2864) produced only small wild-type offspring, from which it was evident that at least one parent of each had carried neither black nor curved, and had come from the non-crossover chromosome which was alter- native to the b 1 c chromosome. The remaining three pairs contained various cross-over chromosomes. Thus, 2859 came from b + + b I c nocn e + I C b I C b + C b I C 2860from -- #+ -- $,2862from 9+ cf. The mother of 2859 contained two cross-over chromosomes ( - - — ) \+ I c / and must therefore have been an F3 individual, as more were expected to be. LOCUS OF LETHAL Ha. The composition of these flies is important, since from their off- spring calculations of the amounts of crossing-over between the lethal and both black and curved were made. The first culture in which the lethal appeared (2675, table 131) was of the type most advantageous for the study of the crossing-over relations because of being in the form of a back-cross for black and curved. Unfortunately, no counts had been taken on black, so that only the lethal curved value could be calculated. Both the curved and the not-curved classes produced by this cross are composite, the curved class being composed of two cross- overs and a non-crossover class (2x+ri), while conversely the not- curved class is 2n+x. The solution of the equations gives x =—3.4 and n = 60.7. The fact that x has a minus sign is easily accounted for by probable error, and only means that the loci of the lethal and curved are close enough together so that a small deviation of the classes gives an apparently impossible cross-over value. In cultures 2840 (and 2863) the composition of the not-curved class is 3n-j-2x and of the curved class simply x. The not-curved flies totaled 327 and the curved 14, from which x = 14 and n = 99.7. The lethal curved cross-over value is thus 12.3. Likewise the black lethal cross-over value is 9.8. Culture 2859 furnishes data on the lethal curved value according to the equations from which, the cross-over value is 11.4, comparable with the 12.3 just found from 2840+2863. The condition in 2859 with respect to black OF MUTANT CHARACTERS. 289 and lethal is the reverse of that from lethal and curved, since black and the lethal are in different homologues. The equations in this case are from which the black lethal value is 20.3. Culture 2860 gave x = 5, n = 20 and a lethal curved cross-over value of 20. The foregoing cultures have given in the case of black and lethal a total value of n of 144.7 and of x of 21.7, so that the cross-over value is 13.0. Likewise for lethal-curved, n = 227.1, a: = 21. 6, and the cross- over value is 8.7. STOCK OF LETHAL lla. It was foreseen that stock of this lethal, which was called lethal I la (the "II" designating the chromosome, and the "a" denoting the first of this series), could not be run advantageously by means of cul- tures such as were made from the wild-type flies of the original F2. A stock could be maintained by taking advantage of the fact of no crossing-over in the male, if males carrying the lethal in one second chromosome and some recessive in the other were back-crossed in each generation to females homozygous for this same recessive. An effi- cient scheme for obtaining such a stock was devised as follows: The original F2 flies appeared only in the three classes: wild-type, black, and curved. The constitution of any given wild-type fly could not be told by inspection, but the case was different with the blacks and curveds. Their father had carried black, lethal, and curved all in the same second chromosome, and since there is no crossing-over in a male, every sperm which carried one of these genes carried all. Therefore every black fly in the F2 should have the black lethal curved second chromosome of the father. The maternal second chromosome is in this case a cross-over chromo- some, carrying black not-lethal, and not-curved. If such a male — "•£-— ), were crossed to a curved female, the offspring should be of 6 I c / two kinds — wild-type ( — ^^ ) and curved ( - ), and every one of \++c/ \++c/ these curved flies should carry the lethal in the manner required. In- stead of simple curved, a black purple curved plexus speck (TT — ) female was used in this out-cross — the object being to combine the TT — stock with the lna stock so that no separate TT — or llla stocks need be maintained. Three such out-crosses of F2 black males to TT— females were made. Two of these gave only black offspring (culture 2865, 155 blacks; culture 2866, 182 blacks), and no black curved, which showed that the males had been cross-overs belonging to the F3 generation. To check up this analysis, five F2 cultures were raised from one of them 290 THE SECOND-CHROMOSOME GROUP (2866), and, as expected, gave no lethal result (table 133, not separated for plexus). The third culture, 2867, gave 40 black curved and 41 black flies and seemed to be the type to be used. Accordingly several of the black TT ~~~ curved males, supposed to be of the constitution were out-crossed b I c to TT— females to give the required stable stock. It was well that three F2 cultures were raised at the same time, for these test cultures TABLE 133. — Pi, black cf X black purple curved plexus speck 9 ; F\ black 9 + Fiblack cf. Feb. 8, 1916. Black purple curved speck. Black. Black purple. Black curved speck. Black purple curved. Black speck. Black purple speck. Black curved. 3203 24 109 4 13 11 20 2 5 3204 23 214 4 8 22 16 3206 27 167 9 11 21 23 3207 28 171 10 15 15 15 1 5 3208 38 181 21 16 23 28 3 Total 140 842 48 63 98 98 3 13 (table 134) proved that the male was not carrying lethal and had in b + c fact been an F2 double cross-over of the constitution . After this failure to secure stock from flies of the original F2, the attempt was repeated successfully with black flies of the derived culture (F2) which had the same constitution as the original F2, but in which flies of the succeeding overlapping generations were known not TABLE 134. — PI, black cf X black purple curved plexus speck 9 / F\ black curved 9 + FI black curved cf . Feb. 7, 1916. Black purple curved speck. Black curved. Black purple curved. Black curved speck. 3168 34 124 18 14 3169 33 123 21 23 3201 15 73 9 8 Total 82 320 48 45 to be present. Two such black males by TT — females gave as expected black and black curved offspring (3112: 6 = 81,6 c =67; 3197: 6 = 67, 6 c = 53). A stock from this source was turned into the stock-room after tests had shown that it was carrying the lethal. The tests (table 135) consisted of inbreeding some of the black curved males and females, which proved to have the required constitution ( ^r ^* Sp } . \b+lnc++/ OF MUTANT CHARACTERS. 291 These tests brought out the interesting point that in crosses in which a recessive and a lethal are in opposite chromosomes ( J , the percentage of the recessives in F2 is a constant value 33.3, irrespective of how much or how little crossing-over there is between the two loci. Thus, the recessives purple, plexus, and speck are at very different distances from the locus of the lethal, yet the percentages of appearance of all these is practically the same (pr = 37.0, pr = 34.0, sp = 35.3, though slightly higher than the expected value of 33.3. TABLE 135. — The offspring from pairs of black curved flies from Z//a stock /bpT +cpx*p [All flies were black curved and showed also the subdivisions in the table.] Feb. 29, 1916. Purple plexus speck. "Wild- type." Purple. Plexus speck. Purple plexus. Speck. Purple speck. Plexus. 3535 21 53 18 18 3 3 2 3551 17 54 15 15 1 1 4 1 3552 24 65 22 18 1 3 2 3553 20 46 21 19 3 2 2 Total. . . 82 218 76 67 8 9 g 3 TELESCOPE (0- (Plate 7, figure 6.) ORIGIN OF TELESCOPE. In determining the locus of the sex-linked mutant "crooked bristles" (locus 38.0, allelomorph of furrowed) several back-cross tests were made of females carrying vermilion, crooked, sable, and garnet in one X, and only wild-type genes in the other. One of these tests gave slightly less than a quarter of the flies with "telescope" abdomens (2735, December 27, 1915). DESCRIPTION OF TELESCOPE. The abdomens of the "telescope" flies tend to retain the drawn-out appearance that freshly hatched flies have. The segments are slightly separated from one another instead of overlapping. The pigmenta- tion and chitinization of the abdomen remain weak and there is a wet (glazed) appearance to the entire surface of the body. The wings droop at the sides and diverge, this character being very useful for identifi- cation. Each band is sunken at the middle, with slightly raised edges. INHERITANCE OF TELESCOPE. That this character was not sex-linked was seen at once, since in the culture in which it was first found it appeared in females as well as males and showed no linkage to any of the four sex-linked characters 292 THE SECOND-CHROMOSOME GROUP present. A mass-culture of the telescope females and telescope males which showed none of the sex-linked character was made (2867). This culture failed, probably because of sterility rather than from cul- tural conditions, since some of these same flies remated to purple, in or- der to start a second-chromosome linkage test with telescope failed to produce offspring (3120, 3121, 3122). A second mass-culture of telescope (2906) produced a very few flies, from which a successful out- cross of a telescope female to a male from the pink spineless stock (third-chromosome recessive) was made (3213; 3214 sterile). Several FI pairs were started, of which one (3503) produced offspring. These offspring represented a 9 : 3 : 3 : 1 ratio of telescope and pink (+ 188, ts 61, p 44, tsp 12; disregarding spineless), which proved that telescope was not in the third chromosome. This culture furnished (March 4, 1916) one of the most valuable autosomal characters, hairless, a third-chromosome dominant which is fully viable (though lethal when homozygous), which is easy of clas- sification, which does not mask any other third-chromosome character, and whose locus is advantageous. After the discovery that telescope was not in the third chromosome it was thought certain that it was in the second, so experiments were planned on that basis. By means of the dominant "star," at least a rough approximation of the locus was possible. Accordingly several out-crosses to star were made en masse (3848, 3849, 3854), of which 3854 alone produced offspring. In view of the sterility so far encountered it was thought best not to attempt back-crosses, but to raise F2 which involves the mating only of not-telescope flies. The first tests were to check up the assumption that telescope was second- chromosome by means of a male test. This was done by pairing the FI star males ( - - 1 and FI wild-type females ( — )• The telescope \~t~ tg / \ tg / offspring, then, constitute a back-cross test. The result proved, as expected, that the telescope is second-chromosome, for none of the back-cross telescopes were star (table 136). TABLE 136. — PI, telescope 9 9 X star cf d"; FI star TEXT-FIQTTRE 87. — Constructional map of second chromosome, giving bases of reference and indicating various cross-over values used in calculating mean position for each locus. The locus of limited is either the same as that of morula, which is possible, or is slightly to the right. The black fringed value of 42.5 is almost the same as the black arc value of 42.6, so that we may place fringed at 98.0. The locus of dachs-lethal is probably the same as that of dachs, 29.0 (dachs-deficiency) ; but if this is not the case, then the locus is so close to that of dachs that the interval is negligible. Squat gave 11.0 per cent of crossing-over with black and can there- fore be mapped at 35.5. Lethal Ila gave a value of 13.0 with black, and a value of 8.7 with curved. In the first case there were 166 flies, indicating a position of 59.5; and in the second case 249 flies, indicating a position at 64.8. The mean position is thus 62.7. The position of telescope is known only from the star telescope back-cross value of 44.4, which indicates a locus at about 66.5 (C = 100). The loci of the various mutant genes with respect to black as a base of reference have just been found. In some regards it is more con- venient to renumber these loci so that the left-most (star) is taken as OF MUTANT CHARACTERS. 303 the zero-point and the others have consecutive numbers in a single series. The map made in this way has already been given on page 127. In using such a map one should keep in mind both the locus as given and the manner in which that locus has been established, since this largely determines not only the accuracy but also the significance of any particular location. A type of diagram which is capable representing fully the relationship 10 ; - 15.4 zo fat- 38.0 Z9.O 40 50 60 70 80 90 - - 35.5 -465 -46.7 48.5 50.0 ± - - (62.5 65.0 66.5 - - 73.5 of of each locus to the other loci is given in text-figure 87. This diagram could be further elab- orated by making the heaviness of line correspond to the accuracy of data, and by giving, besides the final, the reference-base positions. Thus pr vg gives a locus of vestigial at 6+18.0, while the corrected black vestigial data indicate a locus of vestigial at 6+18.9, while the locus actually given in the diagram is the mean position for ves- tigial at 18.5 The type of map which is in daily use in our laboratory is that given in text- figure 88, in which the loci are further classified according to the value of the character, etc. Thus, the mutants of first rank in value are made conspic- uous and insured first consideration by being lined up at the extreme left edge of the space. The mutants nearly as good, but whose usefulness is restricted in one or another respect, are spaced next in order. Still further to the right are those whose loci are not well estab- lished or whose characteristics are such that they are useful only in experiments of a very special nature. At the ex- treme right are the mutants no longer available, because the stocks have been lost or discarded. This type of map can be kept subject to continuous changes in the valuations or the locations of the different mutants by drawing the map-scale on a soft board and mounting the symbols for each mutant on the head of a thumb-tack. 100 no 96.2 -98.0 -98.4 TEXT-FIGURE 88. — Working and valua- tion map of the second chromosome. The loci mapped at the left margin rep- resent the most valuable mutants, those farther to the right progressively less useful. Those next the right margin are mutants no longer extant. BIBLIOGRAPHY. BRIDGES, CALVIN B. July, 1915. A linkage variation in Drosophila. J. E. Z., 19: 1-19 , Jan.-Mar., 1916. Non-disjunction as proof of the chromosome theory of eredity' Genetics, 1: 1-52, 107-163. , Sept., 1917. Deficiency. Genetics, 2: 445-465. , and A. H. STURTEVANT, Apr., 1914. A new gene in the second chromosome of Dro- sophila, and some considerations on differential viability. Biol. Bull., 26: 205-212. HYDE, R. R. Dec., 1913. Inheritance of the length of life in Drosophila ampelophila. Report, Indiana Acad. of Sci., 113-123. , July, Aug., Oct., 1914. Fertility and sterility in Drosophila ampelophila. J. E. Z., 17: 141-171, 173-212, 343-372. JENNINGS, H. S. May, 1917. Modifying factors and multiple allelomorphs in relation to the results of selection. Am. Nat., 51 : 301-306. LIFF, JOSEPH. Feb., 1915. Data on a peculiar Mendelian ratio in Drosophila ampelophila. Am. Nat., 49: 97-120. LUTZ, F. D. Feb., 1913. Experiments concerning the sexual differences in the wing-length of Drosophila ampelophila. J. E. Z., 14: 267-273. MACDOWELL, E. C. July, 1915. Bristle inheritance in Drosophila. J. E. Z., 19: 61-98. MARSHALL, WALTER W., and H. J. MULLER. Feb., 1917. The effect of long-continued heterozygosis on a variable character in Drosophila. J. E. Z., 22:457-470. METZ, CHAS. W. Nov., 1914. An apterous Drosophila and its genetic behavior. Am. Nat., 48: 675-692. , and B. S. METZ. Mar., 1915. Mutations in two species of Drosophila. Am. Nat., 49: 178-189. MORGAN, T. H. May, 1910. Hybridization in a mutating period in Drosophila. Proc. Soc. Exp. Biol. and Med., 7: 160-162. , Mar., 1911. The origin of nine wing mutations in Drosophila. Science, 33: 496-499. , July, 1912. Heredity of body-color in Drosophila. J. E. Z., 13: 27-45. , Nov., 1912. Further experiments with mutations in eye-color of Drosophila. Jour. Acad. Nat. Sci., Phila., 321-350. , Nov. 22, 1912. Complete linkage in the second chromosome of the male of Dro- sophila. Science, 36: 4-6. , Apr., 1914. No crossing-over in the male of Drosophila of genes in the second and third pairs of chromosomes. Biol. Bull., 26: 195-204. , July, 1914. Two sex-linked lethal factors in Drosophila and their influence on the sex-ratio. J. E. Z., 17: 81-122. , July, 1915. The r61e of the environment in the realization of a sex-linked Men- delian character in Drosophila. Am. Nat., 49: 345-429. , Oct., 1916. A critique of the theory of evolution. Princeton Univ. Press, 195 pp. , and C. B. BRIDGES. May, 1916. Sex-linked inheritance in Drosophila. Carnegie Inst. Wash. Pub. No. 237, 1-87. — , and CLARA J. LYNCH. Aug., 1912. The linkage of two factors in Drosophila that are not sex-linked. Biol. Bull., 23: 174-182. MORGAN, STURTEVANT, MULLER, and BRIDGES. Oct., 1915. The mechanism of Men- delian heredity. Henry Holt & Co., 262 pp. MORGAN, T. H., and S. C. TICE. Apr., 1914. The influence of the environment on the size of the expected classes. Biol. Bull., 26: 213-220. MULLER, H. J., Apr., May, June, July, 1916. The mechanism of crossing-over. Am. Nat., 50: 193-221, 284-305, 350-366, 421-434. PLOUGH, H. H. Nov., 1917. The effect of temperature on crossing-over in Drosophila. J. E. Z., 24: 147-209. STURTEVANT, A. H. Jan., 1915. The linear arrangement of six sex-linked factors in Drosophila as shown by their mode of association. J. E. Z., 14: 43-59. — , 1915. The behavior of the chromosomes as studied through linkage. Zeit. f. i. A. u. V., 13: 234-287. WEINSTJEIN, ALEXANDER. Mar., 1918. Coincidence of crossing-over in Drosophila melan- ogaster (ampelophila). Genetics, 3: 135-172. 304 III. INHERITED LINKAGE VARIATIONS IN THE SECOND CHROMOSOME. BY A. H. STURTEVANT. 305 INHERITED LINKAGE VARIATIONS IN THE SECOND CHROMOSOME. BY A. H. STURTEVANT. INTRODUCTION. The data presented in this paper demonstrate the existence of two genes that influence the amount of crossing-over in the second chro- mosome of Drosophila melanogaster (ampelophila) -1 These two genes were both found in the same female, that came from a stock collected in Nova Scotia. Each of the genes, in females heterozygous for it, decreases the amount of crossing-over in the region in which it lies. One of them (the other has not been tested) produces no appreciable effect on crossing-over in females homozygous for it. These results are both paralleled by the effects produced on the third chromosome by a gene in that chromosome. The latter case is discussed briefly. An account is also given of a race in which the amount of crossing- over in one region of the second chromosome is increased. This last case is not yet fully worked out. NOVA SCOTIA CHROMOSOME. The two loci, vestigial and speck, usually show about 37 per cent of crossing-over, as appears from the summaries here presented by Bridges and Morgan. In September 1913, the writer mated a wild female, of a fresh stock collected by Miss E. M. Wallace at Liverpool, Nova Scotia, to a vestigial speck stock that had been used in making the crosses reported by the writer (1915), and had in those crosses given the usual result. A single FI female from this mating was mated back to three vestigial speck males of the above stock to pro- duce culture 7 of this paper. The result of this mating was 55 wild- type offspring and 44 vestigial specks — no cross-overs (see Appendix). Two of the wild-type daughters were mated to vestigial speck brothers, to produce cultures 68 and 69. These produced 2 cross-overs among 136 and 2 among 120 offspring, respectively. The same type of mat- ing was repeated in the next generation, in cultures 104, 105, 106, 110, 113, and 114. In 104 and 105 great difficulty was experienced in classifying speck (the only time I have ever noticed such a difficulty with this character), and the two cultures were unfortunately dis- carded without any attempt being made to see wherein the difficulty 1A preliminary note on this case has already been published (Sturtevant, 1917). It has also been discussed by Morgan, Sturtevant, Muller, and Bridges (1915), Muller (1916), and elsewhere. 307 308 INHERITED LINKAGE VARIATIONS lay. The other four cultures gave again few or no cross-overs; and this type of mating was carried on for two additional generations with the same result (see Appendix). It is evident that in every case the tested female has at least a part of the " wild-type' ' second chromo- some present in the female of culture 7 and derived from the Nova Scotia stock. That this chromosome is really responsible for the result has been shown in several ways, as follows: A wild-type female from 69 was mated to 4 black curved speck males of an unrelated stock. The FI'S were wild-type and speck in approximately equal numbers, as would be expected. Except for the rare cross-overs, all the not-speck flies should have carried the Nova Scotia chromosome; and all were heterozygous for black, curved, and speck. Two such wild-type females were back-crossed to black curved speck males (cultures 171 and 172). They gave similar re- sults, which, when added show the following relations: 1 2 sp 419 Total 440. 20 Here we have the same reduction of curved speck crossing-over that has already been observed for vestigial speck, which includes the curved speck region, and also a reduction of the black curved cross- ing-over. Experiments exactly analogous to these have been carried out with curved speck, black purple vestigial arc speck, black purple curved, star black purple curved speck, black purple curved morula, star black plexus, and other stocks, always with the same result — greatly reduced crossing-over when the Nova Scotia chromosome is present (see table 1). In several of these cases the chromosome in TABLE 1. — Tests of females with one original Nova Scotia chromosome. Loci. 0 1 2 3 4 1,2 Total. S' b pr c 8p 222 0 0 1 0 0 223 S' bv~. . 384 0 12 0 396 b Pf Vg Of Sp b pr C 1,083 9,422 0 26 10 82 0 0 1 0 0 3 1,094 9,533 2,108 1 42 0 0 1 2,152 6 c Sp 419 20 1 0 440 b mr 272 5 277 bba 1,607 104 1,711 vg Sp 1,183 4 1,187 C Sn. . 1,171 1 1,172 question was transmitted through males, instead of females, as above, but this did not in any way affect the result. Two wild-type females from 110 were mated to black balloon males of an unrelated stock. All the offspring were, as expected, wild-type in appearance. One half of them should have contained the "Nova Scotia' ' chromosome, the other half should have received the vestigial speck chromosome, and therefore should have given the usual result IN THE SECOND CHROMOSOME. 309 TABLE 2. for black balloon (48 per cent). 12 of these females were back-crossed to black balloon males, and gave the two types of results shown in table 2. The experiments described above demonstrated that the unusual result is produced when the Nova Scotia chromosome is present; the black bal- loon result and other similar ones show that offspring of individuals bearing such a chromosome may give the usual result, these evidently being the off- spring that do not receive the chromo- some in question. Table 1 shows the results obtained from females bearing one Nova Scotia chromosome. Since there is here a total of only about 1.5 per cent crossing-over between star and speck, it follows that we have almost certainly been dealing throughout with a second chromosome derived entirely (or at least all of it between star and speck) from the original Nova Scotia stock. In culture 193 a female heterozygous for curved and speck and for the Nova Scotia chromosome was mated to a curved speck male. A speck female, produced as the result of crossing-over between curved TABLE 3. — Tests of females with one Nova Scotia chromosome, the speck end of which has been replaced. Culture. No. No. of Offspring. b ba cross-overs. p. ct. 201 131 3.8 203 203 7.9 204 201 9.0 206 242 6.6 207 145 3.4 209 186 2.7 210 256 5.9 211 347 6.9 202 303 41.3 205 386 54.1 208 224 44.6 212 402 49.3 Loci. 0 1 2 3 4 1, 2 Total. 277 0 5 0 0 1 283 b pr c 478 1 6 0 485 b Sn. . . 565 7 572 and speck (and therefore bearing the original Nova Scotia chromo- some, minus its speck end) , was mated to a curved male of stock. Two daughters of the latter culture (in 283 and 284) gave 1 cross-over be- tween curved and speck in 505 offspring. The results obtained with this Nova Scotia chromosome, from which the speck end had been removed, are shown in table 3. Evidently the speck end of the chromo- some is not responsible for the unusual results. As will be shown below, the Nova Scotia chromosome was ultimately separated into two parts, the separation-point being between purple and vestigial. Tests (see table 16) were made of females in which both parts were present, but were each united to parts of "normal" chromosomes. Culture 778 was of this nature; and 786 and 787 contained daughters of 778 in which the original Nova Scotia chromo- 310 INHERITED LINKAGE VARIATIONS some had been reconstructed by crossing-over. Table 4 presents the results from these two cultures. The combined data from tables 1, 3, and 4 are summarized in table 5. Figure 1, second line, is a map based on this table. TABLE 4. — Tests of females with one reconstituted Nova Scotia chromosome. Loci. 0 1 2 3 Total. b pr c Sp 549 0 6 0 555 0.0 0.0 sp -I S'b pr vgc sp o.b' 0:2 i!s 1.4 S'b \.Pr -F h 0.0 0.5 vgc sp pr.-.----' sp 56.3 4Z.4 48.6 •-."•.. 50.2 50.3 S'b pr c sp 00 0.'2 31 3.2 O.O 0.3 0:4 b pr -1—4- 47.8 FIG. 1.- -Maps based on table 25. The first corresponds to the first column of the table, the second to the second column, etc. The star black plexus data given here show an unexpectedly high percentage of crossing-over between . black and plexus. The data should perhaps not have been included, as there is reason to believe that another gene affecting crossing-over may have been present (see below, p. 324). For this reason plexus has not been entered on the map in figure 1. The black balloon percentage (6.1) is unexpectedly higher than black speck (1.1). As may be seen from the account here given by Bridges and Morgan, speck and balloon certainly give less than 1 per cent crossing-over in ordinary flies. Unless some complication is here present, balloon must be to the right of speck, and the speck balloon region must give more crossing-over than usual in the presence of a Nova Scotia chromosome. IN THE SECOND CHROMOSOME. 311 Table 6 shows the results obtained from second broods,1 produced by females containing a Nova Scotia chromosome. These data were TABLE 5. — Tests of all females bearing one Nova Scotia chromosome (region from star almost to speck) . Loci. 0 1 Total. Percentage. S' b 619 o 619 0 0 S' pr. • • 223 o 223 0 0 S' c 221 1 223 0 4 S' px... 384 12 396 3 0 S' «„.. 222 1 223 0 4 6 pr 14,293 33 14,326 0 2 b Va. . 1,362 16 1,378 1 2 be 13,203 185 13,388 1.4 b Vi. . 384 12 396 3 0 b mr 2,381 48 2,429 2 0 6 8n. • • 3,117 51 3,168 1 6 bba 1,607 104 1,711 6.1 J)r Vo. . 1,362 16 1,378 1 2 Vr C. . . 12,807 141 12,948 1 1 2,109 43 2,152 2.0 pT Sp 2,133 23 2,156 1.1 tin Sn. . . 2,560 5 2,565 0.2 c mr 2,152 0 2,152 0.0 99m, , 2,892 3 2,895 0.1 collected in order to find out if there is a change in the linkage value as a female grows older. The percentages are so small, however, that a comparison with first broods can give no significant result. The small percentages also make impossible a satisfactory study of coincidence in tables 1, 3, and 4. TABLE 6. — One original Nova Scotia chromosome, second broods. Loci. 0 1 Total. Percentage. S' b 222 0 222 0 0 S' pT 221 1 222 0 4 S' c 219 3 222 1 3 S'«B... 219 3 222 1 3 b pr 2,107 9 2,116 0 4 be 2,084 32 2,116 1 5 b mr 936 12 948 1 3 b So. . 219 3 222 1 3 pr c 2,087 29 2,116 1 4 937 11 948 1 2 220 2 222 0 9 c mr 948 0 948 0 0 C 8n. . 430 0 430 0.0 Culture 69, referred to above (p. 307), contained a female of the ... ,. Nova Scotia ~ , . . . constitution — , mated to a vg sp male. One vestigial male, v, sp produced by crossing-over and possessing, presumably, only the ex- 111 First brood" and "second brood" are terms applied to the offspring produced when a female is kept in one bottle for 9 or 10 days (first brood) , and then transferred to another bottle for a second period of 9 or 10 days (second brood). The division is an arbitrary one, and doea not correspond to any "rhythm" in the production of eggs. 312 INHERITED LINKAGE VARIATIONS treme speck end of the Nova Scotia chromosome, was mated to a speck female of stock. The not-speck offspring produced were — - — , with a small piece of the Nova Scotia chromosome opposite sp Four females of this constitution were tested, in cultures 166, 167, 168, and 169, by mating to vestigial speck males. The results are shown in table 7. TABLE 7. 0 1 8P Vg + 1g Sp Total. 166 85 60 46 24 215 167 112 92 42 46 292 168 48 62 46 30 186 169 89 98 41 45 273 Total 334 312 175 145 966 - f. 64 ^ i * r 32 ' } Percentage . . . 3 J.I These data give approximately the usual value for vestigial speck crossing-over, and therefore agree with the data previously presented in indicating that there is no effect on crossing-over produced by the extreme right-hand end of the Nova Scotia chromosome. TESTS OF CROSS-OVERS. When the earlier experiments with the Nova Scotia chromosome were carried out, only that part of the chromosome from black to speck or balloon was studied. Numerous tests of cross-overs were made, in order to find out what part of the Nova Scotia chromosome was responsible for the unusual ratios. The result obtained was that that part of it that lies to the left of a point between purple and vestigial gave approximately the ratios found in ordinary stocks; but that part of it that lies to the right of this point between purple and vestigial and to the left of a point between curved and speck gave what seemed at first to be the same ratios as those given by the whole Nova Scotia chromosome. It was therefore concluded that the peculiarity was due to one gene, located between purple and speck. But it soon appeared that this right-hand end was giving a little more crossing- over in the black-curved region than was the whole Nova Scotia chromosome. By this time star had become available, and it was now found that star and black gave no cross-overs in the presence of the whole chromosome, but gave the usual 40 per cent in the presence of the right-hand end. Tests were then made of the left-hand end again, and star and black were found to give no cross-overs, while IN THE SECOND CHROMOSOME. 313 black-purple gave a greatly reduced value. It therefore follows that the original Nova Scotia chromosome contained two factors: On i, located to the left of purple,1 which makes star black 0.0, and reduces black-purple. Cn r, located between purple and speck,1 which greatly reduces the whole purple speck region. RIGHT-HAND END OF NOVA SCOTIA CHROMOSOME (C//,). Culture 171 contained a female with an original Nova Scotia chromo- C C some and a black curved speck chromosome — — — — , mated to o c sp a black curved speck male. It is included in tables 1 and 5. A black female, produced by crossing-over, must have had the right end of the Nova Scotia chromosome, but not the left ( — )• \o c Sp/ This female, in culture 226, was mated to stock curved speck males, and produced 146 offspring without a cross-over between curved and speck. A wild-type daughter, r , was mated to stock black c sp purple curved males (culture 277). Among 105 offspring, 3 were cross-overs between black and curved. One of the three, a wild-type C II r daughter, 7 — ~ — , was mated, in culture 318, to stock black ' o pr c' purple curved males. Wild-type daughters, of the same constitu- tion, were again mated to black purple curved males in cultures 354 and 355. The results were as shown in table 8. TABLE 8. Culture. 0 1 2 1, 2 Total. 318 227 13 3 0 243 354 211 11 4 0 226 355 184 11 2 2 199 Total 622 35 9 2 668 Percentage .... 93.1 5.2 1.3 0.3 A number of other cultures (see table 9) were made in which this same right-hand end of the Nova Scotia chromosome was tested. All were descended from 226, 277, and 318. The results are included in tables 10 and 11. All these cultures agreed in showing that for the purple-to-speck region we have a result not very different from that given by the whole Nova Scotia chromosome; but for the black purple region the result is not very different from the usual one. 1 Tests made more recently, in connection with studies of Cm, n (see below), indicate that Cm is probably to the left of black, and that Cur is certainly to the left of plexus. 314 INHERITED LINKAGE VARIATIONS Numerous other tests have been made of the right end of the Nova Scotia chromosome. Table 9 gives a list of the different cross-overs tested, together with the cultures derived from those sources. In addition, there are a number of cultures (including all those in which the character star was tested) in which the origin of the CH r segment is uncertain, because it has been passed through females homozygous for CUT from different sources (see below). TABLE 9. — Tests of right-hand end of Nova Scotia chromosome. Culture in which cross-over occurred. Loci between which cross-over occurred. Cultures representing this piece of the Nova Scotia chromosome (see Appendix). 171 be 226, 277, 318, 351, 352, 353, 354, 524 N PT Vg 355, 377, 379, 380, 404, 416, 417, 430, 431, 432, 433, 446, 448, 449, 450, 467, 468, 469, 470, 471, 472, 473, 480, 495, 496, 500, 511, 512, 513, 517 547 524 N Pr Vg 545, 546, 569, 622, 686, 687, 719, 685 pT c 721, 723, 724, 725, 763 707, 708, 709 These cultures all contain only that part of the Nova Scotia chromo- some that lies to the right of a point between black and curved (171 series), purple and vestigial (two 524 N series), or purple and curved (685 series). Since they all agree in the results produced, we may conclude that the gene responsible for these results is located some- where to the right of purple. In dealing with the original Nova Scotia chromosome we found that removal of the speck end made no difference in the ratios given. It was therefore to be expected that the right-hand piece would show the same relation, i. e., that CUr is between purple and speck. The following data show that such is, in fact, the case. Culture 546, of the second 524 N series, contained a female of the r , mated to black purple arc speck males. constitution ar s. By double crossing-over, a female was produced of the constitution — - — - . This female, in which the extreme speck end of the b pr ar sp old Nova Scotia chromosome had been lost, was mated to black purple vestigial arc speck males, in culture 570. All later cultures of the second 524 N series received their C//r from this female. They gave essentially the same results as the other C//r cultures, and have therefore been included in tables 10 and 11. The fourth line of figure 1 represents a map based on table 11. IN THE SECOND CHROMOSOME. 315 TABLE 10.— C//P. Loci. 0 1 2 3 4 1,2 1,3 2,3 2,4 3,4 1,2,3 Total. Sf b pr c Sp 805 657 86 17 1 3?, 11 2 1 1 2 1,615 Sf b PJ 8p . 420 297 *>6 10 9 3 2 0 767 198 10 a 0 0 0 0 0 0 0 0 210 101 16 0 o 0 0 o 0 117 700 39 ?6 o 0 0 0 0 765 b T>9 C 8n . 943 59 1? 0 1 0 o 0 1,015 b Vr C 7,009 427 103 10 7,549 b pr Sp 888 86 61 11 1,046 b ft* 578 25 603 141 15 0 0 156 be 790 65 855 b mr 690 44 734 Pr c 609 6 615 C Sn. . 146 0 146 Second broods. S' b Pf c Sp 435 292 33 16 0 9 5 3 0 0 2 794 S' b pr Sp 163 139 11 2 3 ?, 0 0 320 b pT c 830 70 ?0 8 923 234 50 3f> 4 323 b Vn Sn . . 217 5 0 0 222 TABLE 11.— C/Jr. Loci. 0 1 T. Percentage. Loci. 0 1 T. Percentage. S'b.... 1,371 1,011 2,382 42.4 Second S'pr... 1,275 1,107 2,382 46.5 broods. S' c.... 851 764 1,615 47.3 S'b.... 662 452 1,114 40.6 S' sp... 852 763 1,615 47.2 S' pr... 630 484 1,114 43.5 b pr 12,843 844 13,687 6.2 S' c.... 452 342 794 43.0 b Vg 440 43 483 8.9 S' 8p... 452 342 794 43.0 be 10,920. 879 11,799 7.4 bpT 2,390 192 2,582 7.4 6 mr. .. . 1,390 109 1,499 7.3 b Vg. . . . 217 5 222 2.3 b sp . . . . 4,470 456 4,926 9.2 be 1,562 155 1,717 9.1 Pr tig 325 2 327 0.6 b Sp 1,498 161 1,659 9.7 Pr c . . . . 11,368 191 11,559 1.6 pr c 1,668 49 1,717 2.9 Pr mr. . . 739 26 765 3.4 Pr Sp. .. 1,370 67 1,437 4.7 PfSp. . . 4,634 136 4,770 2.9 Vg Sp. . . 222 0 222 0.0 Vg 8P.. . 483 0 483 0.0 c sP. . . . 794 0 794 0.0 c mr. . . 765 0 765 0.0 c sp. . . . 2,773 3 2,776 0:1 The second-brood data are not very conclusive. Star-black is perhaps lower than in first broods; but the black purple curved com- binations, for which there is more adequate data, all give a slight in- crease. As is shown by the data presented by Plough (1917), more exact methods are needed in studying this problem. An experiment with C/jr is now planned in which the females will be transferred every two days. Until data from such an experiment are available further discussion would be out of place. 316 INHERITED LINKAGE VARIATIONS LEFT-HAND END OF NOVA SCOTIA CHROMOSOME (Cm). Culture 678 was derived from a female with an original Nova Scotia chromosome and with black, purple, curved, and morula in its mate, mated to four black purple curved males. The culture is included in the totals given in table 1. It produced three curved flies by cross- ing-over. These flies must have had the left-hand end of the Nova Scotia chromosome, up to a point between purple and curved; but the right-hand end of the Nova Scotia chromosome had been lost. One of them, a male, was mated to a black purple female that had the right-hand end of the Nova Scotia chromosome (C//r). A wild- type daughter was mated to black purple curved morula, and gave results that will be discussed below (see table 16). A curved-morula son, that must again have had the left-hand end of the Nova Scotia chromosome, was mated to a similar black purple female; and a wild- type daughter was once more back-crossed to black purple curved morula in culture 752. This time, however, it was decided to get the influence of the left-hand piece of the Nova Scotia chromosome with- out the presence of the right-hand end. A curved morula male from 752 was accordingly mated to a star black female of an unrelated 8' b stock. A daughter, of the constitution with the end of c mr' the Nova Scotia chromosome opposite star and black, was mated to black purple curved speck males, in culture 776. The result was 106 non-cross-overs, 0 cross-overs between star and black, and 38 cross-overs between black and curved. Further tests with descendants of 776, in which the same piece of the Nova Scotia chromosome was present, gave the results shown in table 12 (culture 776 itself is included). TABLE 12.— Cne. Loci. 0 1 2 3 1,2 2,3 T. *S'bcsP... 1,006 0 455 649 0 141 2,251 S'bc 200 0 54 0 254 b Pr c Sp . . . 91 0 35 68 0 6 200 bprsp 138 1 109 1 249 Second broods. *S'bcsp... 576 0 216 337 0 41 1,170 S'bc 96 0 11 0 107 b pr c 8P . . . 137 0 36 -82" 0 6 261 * Done with H. H. Plough, and already reported by him (Plough, 1917). Clearly we have here the same results for star purple as in the case of the original Nova Scotia chromosome; but for purple speck we have nearly the usual result. Q Q Culture 259 contained a female of the constitution r — b pr vg ar sp (original Nova Scotia chromosome), and two stock 6 pr vg ar sp males. IN THE SECOND CHROMOSOME. 317 One vestigial (arc) speck male was produced by crossing-over. He must have had the left-hand end only of the Nova Scotia chromo- some, — - . He was mated to a black female of stock, b pr vg ar sp \ TT 7 V d 9 and two wild-type daughters, — r— ~ — - — ? , were tested, in cultures 328 and 329. They gave the results shown in table 13. TABLE 13. 0 1 2 1, 2 Culture. Total. Vg Sp 6 + bVy &p Vg 6 ,, Sp 6 Vg 328 77 75 28 5 79 63 8 6 341 329 44 40 13 7 31 43 o 1 181 121 • v 115 ' 41 „ 12 - -J 110 *• , 106 rr •" 10 s r 7 •> 522 Total 236 53 216 17 Per cent. . 45.2 10.2 41.4 3.3 These data have been added to those of table 12 in the summary for Cji i, table 14. The corresponding map is shown in the third line of figure 1. TABLE 14. — C/jr, summary. Loci. 0 1 Total. Percentage. S' b 2,505 o 2,505 0 0 S' c 1,855 650 2,505 25 9 S' 80.. 1,147 1,104 2 251 49 0 b pr 447 2 449 0.5 b r<7. . 452 70 522 13.4 6c 2,014 691 2,705 25.4 6 So. . . 1,497 1,476 2,973 49.7 pr c 159 41 200 20 5 pr Sp 236 213 449 47 4 Vn ST, . . 289 233 522 44 6 C 8». . 1,587 864 2,451 35 3 Second broods. S'b 1,277 o 1,277 0.0 S' c 1,009 268 1,277 21.0 S' sp 617 553 1,170 47 3 b pr 261 o 261 0 0 be 1,228 310 1,538 20 1 b Sn . . 760 671 1,431 47.0 Vr C. . . 219 42 261 16 1 143 118 261 45 2 C 8n. . . 965 466 1,431 32 6 The values obtained from second broods (table 14) run consistently a trifle lower than the corresponding values for first broods. The differences are small in every case; but since all are in the same direc- tion, and since females with neither C// , or Cn r show an age change 318 INHERITED LINKAGE VARIATIONS in this same direction (Bridges, 1915; Plough, 1917), the decrease is probably significant. More exact methods (see Plough, 1917) are necessary for obtaining clear-cut data on this point, as has already been stated. Here, as also in the case of CIIr, wherever reliable information regarding coincidence is available, the value is not far from the one found in females that contain neither Cm nor C//r (see Bridges TABLE 15.— Cl1' Loci. 0 1 2 3 4 1,2 Total. S' b pr c 8p 659 0 1 24 1 0 685 b pr c 226 2 10 0 238 b Pf C Tflf 631 0 1?! 0 643 Second cultures. S' b Pf c 8p. 497 0 0 11 1 0 509 TABLE 16.— CIIr Loci. 0 1 T. Percentage. S' b 685 0 685 0.0 S' Vr . . . 684 1 685 0.1 S' c 660 25 685 3.6 S' VT .. 659 26 685 3.7 b V* . 1,563 3 1,566 0.2 be 1,517 49 1,566 3.1 b TTlf 631 12 643 1 9 b Sp 659 26 685 3.7 pr c 1,520 46 1,566 2.9 631 1? 643 1.9 660 25 685 3.6 643 0 643 0.0 C 8n . 684 1 685 0.1 Second broods. S'b 509 0 509 0.0 S' pr 509 0 509 0.0 S' c 498 11 509 2.2 S' Sp 497 12 509 2.4 6 Vr 509 0 509 0.0 be 498 11 509 2.2 b 8n. . 497 1ft 509 2.4 pT c 498 11 509 2.2 497 191 509 2.4 C S-n . 508 1 509 0.2 and Morgan, 1919). But in no case does this include a region in the " sphere of influence" of the cross-over gene present; for in all such regions the percentage of crossing-over is too small to give statistically reliable results. Cni Cultures 677 and 678 both contained females, 6 pr 'in (678 was also heterozygous for mr), mated to b pr c males. A cross-over female, IN THE SECOND CHROMOSOME. 319 fr Pr Cnr , ft_- . CHi c mr 7 - , from 677 was mated to a cross-over male, T - o pr c ' o pr c from 678. The resulting wild-type offspring were 7 - —^ — -• A 0 Pr WJr female of this constitution, in culture 713, gave 4.3 per cent crossing-over between pr and c, none between 6 and pr or between c and mr. The Q same general method was followed in making up seven other - —7; — W/r females. The same C//j c chromosome was present in all these females, but Cn r from different sources was used. The results from these females are given in tables 15 and 16 and the fifth line of figure 1. These data agree with those obtained from Cm CIIr, except that pr c shows a slight rise (from 1.1 to 2.9). Owing to the statistical difficulty of handling such small ratios it is not possible to say whether this difference is significant or not until more data can be collected. The point is of interest in its bearing on the mechanism of the action of GUI and CIIr, but must be left unanswered for the present. Q The second brood data here presented for ni — are entirely Cllr inadequate for the purpose of detailed comparison with first broods. They do, however, show an increase for pr c over the C/// C//r second broods (from 1.2 to 2.2). HOMOZYGOUS C//r. We have seen that heterozygous CIIr greatly decreases crossing- over in the region from purple to speck, but does not appreciably affect the region from star to purple. The data now to be presented show that homozygous CIIr gives a value for purple speck that is very close to that found in "normal' ' flies, but again does not influence the region from star to purple. It has so far not been found possible to obtain a chromosome con- taining C// r with vestigial or curved, since heterozygous C// r practi- cally prevents all crossing-over in the region in which these three genes are located. For this reason none of the data on homozygous C//r deal with loci between purple and speck. HOMOZYGOUS C//r WITH Cni. In the course of the experiments with CIIr a chromosome of the constitution Sf b pr CIIr sp was obtained.1 When males with this Pr (-s 1 1 r &-n 1 This chromosome was derived from culture 570 (see p. 3 14) , in which was a female r - • A cross-over in this female gave & b pr Cn T sv chromosome. This chromosome, or a derivative C// T of it, since it had perhaps been passed through a r - ^ - female, was placed opposite star 0 Pr I/// T 8 8v in the females of cultures 696 and 699a. The chromosome referred to above (Sf b pr Cu T sp) was produced by crossing-over in these females and was kept intact thereafter by breeding from males heterozygous for it, in which no crossing-over occurred. /& _ \ I ~~T - 7; - I i \ o pT Cu r sp / 320 INHERITED LINKAGE VARIATIONS 'Hi c II r chromosome were mated to females of the constitution o pr c ' the star not-black offspring must have been of the constitution — 79 — -^r^- — -. Nine such females were tested by mating to W/J Wlr b pr sp males and gave the results shown in the first line of table 17. The data in the second row were obtained in the same way, except that no star had been put in the b pr CH r sp chromosome. The third line represents the offspring of a female (culture 340) of the constitu- tion C II r — - , produced by mating a male — pr c to a female IIr — . The males in 340 were black purple vestigial arc c Sp speck; since no purples were produced the female must have received a b CIIr chromosome from her father; and since she gave 45.2 per cent crossing-over between black and speck, instead of the 9.0 char- acteristic of Cnr females, she must have received from her mother II I TABLE 17.- m Loci. 0 1 2 3 1,2 1,3 2,3 1,2,3 Total. Sfbpr8p... 1,047 154 3 o 0 138 942 0 1 1 0 2 1,995 293 b Sp . 144 119 263 TABLE 18.— Cm Cnr Loci. 0 1 Total. Percentage. S'b.. . 1,989 6l 1,995 '0.3 S' pr. . 1,989 61 1,995 1O.B S' 8p. . 1,048 947 1,995 47.5 b pT. . 2,285 3 2,288 0.1 b sp.. . 1,351 1,200 2,551 47.0 PT8p. . 1,204 1,084 2,288 47.4 1 These cross-overs are very doubtful. None of them were tested ; and there is a small per- centage of error in classifiying star flies. Similar apparent cross-overs were obtained in working with Ctli, but all were shown, when tested to see if star was really present or not, to be wrongly classified. Table 18 and the sixth map of figure 1 summarize the data from these three series of experiments. No second-brood data are avail- able; and the star to purple region gives so few cross-overs that coinci- dence can not profitably be studied. It is, however, very remarkable that all three cross-overs between black and purple were also cross- overs between purple and speck. More data is needed before we can be sure this is a significant result, since purple and speck themselves cross over so frequently (47.4 per cent). IN THE SECOND CHROMOSOME. 321 Comparison of table 18 with table 14 will show that the results given by —^ — - and by Cm are almost if not quite the same. 'II r That is, homozygous Cn r gives the same result as no C// r. HOMOZYGOUS Cn— WITHOUT Cm. pr Crir . . Cni CII A, . A female b pr c mr was mated to a male L , produced by crossing-over in 69 la Crrr S r\ r/ (a cross-over from 699, q. v., b Pr CIIr ,. see above. Six wild-type daughters were tested by mating to b "D C 171 b pr c sp males. Four gave the expected result for —^r " females; W/r Sp and two (745 and 748) gave no curved offspring, so that they must have been r ~ — - — - . b pr CUr Females 885 to 888 contained a S' b CIIr chromosome derived from the - — ^ — - experiments and a pr CIlT sp chromosome derived from a stock culture that came from culture 570 (q. v., p. 314). It is quite possible that some or all of these females carried another gene affecting crossing-over (C///, n — see below) ; but the results have ft— TABLE 19.- Loci. 0 1 2 3 • 1,2 1,3 2,3 1,2,3 Total. 6 pr Sp 295 9 161 6 471 b 8». . 77 74 151 S' b PT 8pl • • 303 192 16 242 2 156 12 4 917 TABLE 20. Loci. 0 1 Total. Percentage. S' b 573 354 927 38.2 S' pr S'sp b Pr 551 473 1,349 376 454 49 927 927 1,398 40.6 49.0 3.5 b 8r>. . . 851 698 1,549 45.0 Pr «p 817 581 1,398 41.5 been included because they are the only ones available for Sf in the presence of homozygous Cn r. Other work done with Cin, n makes it probable that this gene would not seriously affect any region except that from purple to curved; and the purple speck values for this experiment agree with those from 745 and 748. Therefore the two results are probably comparable. Both are included in tables 19 and 20 and the last line of figure 1. *May contain Cm, //. 322 INHERITED LINKAGE VARIATIONS For this combination also no second-brood data is available. Coinci- dence seems to be of approximately the value that is usual, but can be satisfactorily studied only in the series that may have C///, //. ri The •^r~L ratios are clearly not very different from those obtained Wlr with the "usual' ' second chromosome. NO TESTS OF HOMOZYGOUS Cm. No tests were made of females homozygous for Cm, because it was hoped that a cross-over would occur that would give a Sf Cm chromosome, and thus make possible a test of the region in which GUI is located. A few attempts were, it is true, made to get a pure stock of Cm', but no careful records were kept, and these attempts were all unsuccessful. Recent tests show that there is now a lethal gene in the Cm chromosome that is being studied, so that it will probably be impossible to obtain homozygous Cm- It is not certain whether this lethal represents a recent mutation or not. TESTS SHOWING NO CROSSING-OVER IN MALES. Very few counts have been made from heterozygous males; but no crossing-over in males has been assumed throughout the work, and has been depended on frequently in keeping stocks and in producing many of the more unusual combinations of GUI and C//r. These matings have never produced flies that seemed to result from crossing- over in males, and have always given in later generations results that are consistent with the view that such crossing-over does not occur. Taking this evidence in connection with the counts given below (table 21), and with the evidence that shows crossing-over not to occur in males of Drosophila in any of the chromosomes under any known circumstances,1 we may safely conclude tnat Cm and C//r do not cause exceptions to the general rule. CONSTITUTION OF THE NOVA SCOTIA STOCK. The original Nova Scotia female had in her second chromosome two factors for decreased crossing-over. It would be of some interest to find out whether or not this condition was widespread in the stock from which she came. Unfortunately the original stock was lost before it was discovered that two factors, instead of one, are responsible for the result. The following tests are therefore not entirely satisfactory. Three females, from the Nova Scotia stock, were mated to curved speck, and 4, 4, and 1 daughters, respectively, were back-crossed to curved speck. Only a few offspring were counted from each, but enough to show that all 9 females were giving at least 20 per cent of 1 Except the curious case of "somatic crossing-over" recorded by Muller (1916). IN THE SECOND CHROMOSOME. 323 crossing-over. It follows that Cn r was not present. Three females from Nova Scotia stock were crossed to black vestigial, and daughters TABLE 21. — Tests for crossing-over in males. Culture. Crossover constitution. Mutant genes. Non- cross-overs. Cross- overs. Total. 108 109 621 624 338a 3386 423 424 741a 7416 718 Cn i Crir 37 61 49 52 0 0 86 113 Cn i Cn r Vg Sp Vg Sp 98 101 0 199 Cn T b 2 6 7 5 0 0 9 11 Cn r Cn T Pr Sp b Cn T Cn I Cn T Sp b 8 12 0 20 5 58 143 81 8 58 146 77 0 0 0 0 13 116 289 158 Cn T Cn i Cu r Sp b Cn r Cn i Cn T Sp Cnr Cn i Cn T b 8p Cn r Cn i Cn T b Sp S' sp 287 289 0 576 108 85 112 87 0 0 220 172 Cn T Cm Cnr S' Sp Cn r Cnr, or CIIT Cn r S' Sp 193 88 919 102 0 0 392 190 TABLE 22. Mother. Culture. 6 vg percentage. No. of offspring. N 230 8.5 258 N 231 9.6 270 AT 232 10.7 140 F 233 13.9 202 F 234 11.9 168 F 235 16.2 191 F 236 12.9 295 F 237 13.7 269 F 238 20.0 135 B 239 12.3 235 B 240 23.0 161 324 INHERITED LINKAGE VARIATIONS were back-crossed to black vestigial (cultures 230 to 240 inclusive). The results are shown in table 22. None of these females had both Cm and C//r, but it is possible that one of the factors may have been present, especially in the off- spring of female N. These tests show only that the Nova Scotia stock was not homozygous for C// r, and probably not for C// j. No other stocks from northern localities have been tested, so that it is impossible to even guess whether or not these factors occur frequently in Nova Scotia or neighboring regions. ANOTHER SECOND-CHROMOSOME LINKAGE VARIATION. Cultures 733 and 734, referred to elsewhere, contained females of the constitution 777 — 7 ~. As was pointed out above, they o o px gave an unexpectedly high percentage of crossing-over for black and plexus. Culture 812, descended from the same culture that pro- LJ h 77 (* 9 duced females 733 and 734, contained a female 79 „ P . This ^IIl W/r female produced 72 offspring, of which none were cross-overs between Sf and pn or between c and sp, but 11 were cross-overs between pr C C and c. Later descendants of 812, of the constitution r -, b pr c ' gave this same increased value for pT c without any increase for b pr. But it was found impossible to fix this increased value, which fluctu- ated between the expected value (less than 1 per cent) and 20 to 30 per cent. Several selection experiments have been carried out in an effort to get a stock that would constantly give the high value, but without success. The most recent of these experiments has now been carried through 23 generations of brother-sister matings, always breeding only from those pairs that gave the "high" value for pr c. Yet, in the fifteenth generation, occurred a culture that gave only 1 cross-over among 130 offspring, and in the twenty-third was a culture that gave y^-g- = 3.4 per cent. The latter value, while slightly higher than is usual for Cm Cn r, is much lower than the 20 to 30 per cent now given by most of the "high" selected cultures.1 The nature of this case has not yet been worked out in detail, though culture 812 was counted in December 1915, and the problem has been worked at continuously since that tune. The following points now seem fairly certain, though they must still be checked and extended. (1) The "high" value is due, in large part, at least, to a dominant gene. 1 The Cm has apparently been lost, by crossing-over, in part of this experiment. But since the values given above are too high for heterozygous Cnr, the discussion given is not affected. IN THE SECOND CHROMOSOME. 325 (2) This gene is not in the second chromosome at all, but in the third. (3) The third chromosome gene is linked to a gene that is lethal when homozygous. This is the reason the very high values could not be fixed. (4) This gene, called C///, //, also causes an increase in pr c cross- ing-over in Cn r females. Its effect on females of different constitu- tions with respect to CH i and Cn r is not yet clear. (5) Cj/j, j/, when heterozygous, reduces the amount of crossing- over in the third chromosome. Its effect in this respect is similar to, but not identical with, that of Cni (see next section, and Muller, 1916). Unlike C///, it "allows" a few cross-overs between sooty and rough; but it causes a reduction of crossing-over farther to the left than does Cni- (6) Females with C///, // in one chromosome, and Cin in its mate, give nearly the same amount of crossing-over in the third chromo- some as do females heterozygous only for CHI, or perhaps less in the left-hand regions. A detailed comparison of the effects of these two genes, a study of their interaction, and also an investigation of the locus of CIH, // are now under way. COMPARISON WITH RESULTS OBTAINED FROM Cm. I have shown (Sturtevant, 1913a, 1915) that great linkage variations occur in the third chromosome. My own unpublished data and those presented by Muller (1916) show that the case is very similar to that of C//r. The factor Cin, present in the beaded stock and in several TABLE 23. Percentage of Father of tested 9 • Culture. No. of offspring. crossing over. «« e* aS r e r0 2,568o 2,608 291 0.0 0.0 2,568a 2,610 284 10.9 14.1 2, 568o 2,613 197 0.0 ifl.fi 2, 568o 2,614 252 0.0 0.0 2,568o 2,615 83 7.2 20.5 2, 568o 2,617 216 0.0 0.0 2,5686 2,618 201 0.5 0.0 2,5686 2,619 193 0.0 0.0 2,5686 2,620 156 11.6 17.3 2,5686 2,621 110 13.6 20.9 2,5686 2,622 187 0.0 0.0 2,5686 2,623 143 0.0 0.0 2,5686 2,624 224 0.0 0.0 Total — 4 high, 9 low. 1 Probably an error in classification. Such cross-overs are exceedingly rare. This individual was not tested. 326 INHERITED LINKAGE VARIATIONS stocks derived from it (ebony, spread, eosin), greatly decreases cross- ing-over in the right-hand end of the third chromosome when it is present in heterozygous form; but this result disappears in flies homo- zygous for CUr. Moreover, the gene is itself located in the region in which it produces its greatest effect. The following sample experi- ment will illustrate its action. Certain experiments carried out by Dr. C. B. Bridges, in investi- gating cream III, led to the hypothesis that the eosin stock was im- pure for Cm. Accordingly two males from this eosin stock were mated individually to sepia spineless sooty rough females, and daugh- ters were back-crossed to sepia spineless sooty rough males, with the results shown in table 23. The values for sepia spineless are not given, because sepia was not easily classifiable in the eosin males produced. There are clearly two quite distinct types of results here. In 9 of the cultures there is less than 1 per cent crossing-over between spine- less and rough; in the other 4 there is about 25 to 30 per cent crossing- over between these loci.1 The results are due to the presence of CHI in those females that gave the low result, and its absence in those that gave the high one. That the difference was due to the nature of the third chromosomes derived from the fathers was shown by testing the crossing-over in wild-type daughters of these females. In every case such daughters gave approximately the same results as their respective mothers. Daughters of all but 2615 and 2621 were so tested. In females homozygous for Cm the crossing-over between ss and e rises to about 40 per cent (j-fTr = 41.6 per cent, in one experiment selected at random), as against about 12 per cent in the absence of Cm, and less than 1 per cent when it is present in heterozygous form. This result is in agreement with Muller's (1916) conclusion that homozygous CHI results in the production of more crossing-over than occurs in "normal" females. The Cm experiments are still in pro- TABLE 24. gress, and will be reported in detail in connection with the other third-chromo- some data accumulated in this labora- tory. From the above account, how- ever, it may be seen that the parallel between Cn r and Cm is very close. The effect of each upon the region in which it lies is shown in table 24. The ss e values are only approximately correct. 1 It will be observed that both males from eosin stock were heterozygous for Cm There was later found to be a lethal near the Cm. This, in connection with other results obtained with the eosin stock, suggests that it was a "balanced lethal" stock for the third chromosome (see Muller, 1917). This stock has now died out, so that it is no longer possible to test such a hypothesis. Usual result Heterozygous C . Homozygous C . . ii-pT 8P 46.5 2.9 41.5 in-s, e 12.0 0.5 40.0 IN THE SECOND CHROMOSOME. 327 OTHER CASES OF LINKAGE VARIATIONS. The cases reported in this paper are not the only ones in which linkage variations are known. As has been pointed out above, there is a gene in the third chromosome that affects the percentage of cross- ing-over in that chromosome. It has been shown (Morgan, 1912; Sturtevant, 1913a; Morgan, Sturtevant, Muller, and Bridges, 1915; etc.) that there is no crossing-over in the male of Drosophila, even between loci that give almost 50 per cent of crossing-over in females. The reverse relation — crossing-over in males but not in females — has been shown by Tanaka (1914) to hold for at least two loci in the silkworm moth. Bridges (1915) has shown that the percentage may change with age, and Plough (1917) has shown that it may be changed by temperature. Genetic factors (other than sex) influencing the process are suggested by the results of Baur (1912) with Antirrhinum, of Punnett (1913, 1917) with sweet peas, of Tanaka (1913, 1914) with silkworm moths, and of Chambers (1914) with Drosophila. In none of these cases is the evidence yet clear enough to warrant detailed discussion. BEARING OF METHOD ON CHROMOSOME VIEW. The work reported in this paper deals with the effects on crossing - over produced by certain definite genes. These genes do not, so far as I have been able to discover, produce any visible somatic effects; and their presence can not be detected, except in females, and in females that are heterozygous for other genes in definite regions of the chromosomes, i. e., that are capable of being tested for linkage in those regions. In the case of other females, or of any males, such tests can not be made directly, but only by producing female de- scendants heterozygous for the necessary genes. The fact that it has been possible to work out in great detail the inheritance of these "invisible' ' genes and the effects produced by them is a striking illus- tration of the possibilities of the chromosome view of inheritance and of the advantages of using a rapidly breeding form like Drosophila. The chromosome view itself is perhaps not necessary for the handling of such a case; but the conception of genes that form independent groups that behave as units, the members of which are only separable according to definite rules, is necessary. And such a conception, I think, presupposes some material basis for the independent groups. The great body of evidence that points to the chromosomes as forming such a material basis is too familiar to need discussion here. 328 INHERITED LINKAGE VARIATIONS SIGNIFICANCE OF MAP DISTANCE. It has often been pointed out (e. g., Sturtevant, 1913, p. 49; Morgan, Sturtevant, Muller, and Bridges, 1915, pp. 67-68) that 1 per cent of crossing-over must not be supposed to represent the same actual morphological distance in different chromosomes or in different regions of the same chromosome. Actual distance is evidently an important factor in the result. Other things being equal, chromosome sections of equal length will give equal percentages of crossing-over; but in no case can we be certain that "other things" are equal. The terms ''distance" and "percentage of crossing-over" have unfortunately been sometimes used almost as though synonymous, and confusion has perhaps resulted. But it has been recognized from the beginning that different regions might show different frequencies of crossing-over for the same actual length of chromosome. The results presented in this paper show conclusively that this is the case, as has already been stated (Morgan, Sturtevant, Muller, and Bridges, 1915; Muller, 1916; Sturtevant, 1917). They show that even in the same chromosome pair the percentage of crossing-over shown by different regions is not only not always the same, but is not necessarily even proportional. For example, while S' b remains approximately 40.0, b c may be either 23.0 (neither Cm nor CIIr present), or 7.5 (heterozygous CIIr). LINEAR ARRANGEMENT OF GENES. The strongest evidence for the linear arrangement of genes is that derived from crosses in which more than two loci in the same chromo- some can be followed. The method of seriating the loci on the basis of such information has been described in detail elsewhere (Morgan, Sturtevant, Muller, and Bridges, 1915; Sturtevant, 1915; Morgan and Bridges, 1916), so need not be discussed here. When the linkage values are changed the question arises: Is the sequence of genes affected? It has already been shown (Bridges, 1915; Plough, 1917) that this sequence is not altered when the amount of crossing-over is changed by age or by temperature. In the case of the genetic changes reported here, the evidence presented in tables, 1, 3, 10, 12, 14, 17, and 19 shows that the sequence found in "normal" females is maintained. There are just three cases in which the data, uncor- rected by other data, might lead us to assign a different sequence. These three cases may now be taken up in turn. (1) In the case of 7 ~ — only one cross-over between 6 and 6 pr vg ar sp pr was obtained; and that was also a cross-over between pr and vg. This would lead us to suppose the sequence to be pr b vg, were no IN THE SECOND CHROMOSOME. 329 other data known. But the other data for Cm CIIr show con- clusively that b and pr give very little crossing-over (0.2 per cent), while either with vg, c, or sp, gives about 1.1 per cent; and vg and c give only 0.1 or 0.2 per cent with sp. That is, vg and c are on the same side of b and pr. And the extensive data for b pr c show that the sequence is b pr c. Therefore the one individual that suggested the sequence pr b vg must have been a double cross-over. (2) In the case of CIIr only three cross-overs between c and sp were obtained. Of these, two were also cross-overs between b and c, while one was not. These data alone would indicate the sequence as 6 Sp c, instead of the usual 6 c sp. No great significance can be attached to the difference between 2 flies and 1 fly among a total of 1,615. In any case, the data suggest a very high coincidence. More data of the same sort will be necessary before this exceptional case can appear significant.1 (3) In the case of -79 — -, only 3 cross-overs were observed ^11 r between b and pr. All of these were also cross-overs between pr and sp.2 If the coincidence in this case is 100, approximately the value usual for b pr c, then nearly half of the b pr cross-overs should be also pr sp cross-overs. Therefore the fact that all 3 were such doubles need not cause surprise; even though, taken alone, it would indicate the sequence as pr b sp. The three exceptional cases are, then, of no great significance, ex- cept as indicating rather high coincidence. There are a large number of cases in which the evidence is much clearer and in which the sequence is certainly the same as that usually found. HOW DO Cm AND CIIr PRODUCE THEIR EFFECTS? The question of the mechanism whereby the cross-over genes pro- duce their effects is not yet satisfactorily answered. Cytological examination might conceivably furnish the solution, but has not yet been seriously attempted. A study of coincidence might give a clue, but is difficult to make, because of the very small percentages that are concerned. In the case of Cn r and Cm it is to be noted that two like chromo- somes cross over freely, while two unlike ones do not.3 While this is only a restatement of the facts, it at least offers an attractive opening for speculation as to the nature of the case. JIn a culture derived from Cm, n experiments discussed above, a female that was apparently C n f of the constitution r (without Cm, it) has recently been tested. One av daughter was 0 J)f C 8p produced. If this record represents what it appears to, the count becomes 2 double cross-overs against 2 single cross-overs. 2Two of them were also recorded as cross-overs between S' and 6 ; but this is probably incorrect, as was pointed out above (p. — ). Cm 3So far as the evidence goes, this is also true for Cn j, but 7; — is unknown. V//I ( 330 INHERITED LINKAGE VARIATIONS TABLE 25. Cjj i Cjj r Cm Cu T Cm Cu i Cu T Cu T CUT CJIT Cut S' b... 37 9 0 0 0.0 42.4 0 0 0 3 38 2 S'pr.. .. 43 7 0 0 46.5 0 1 0 3 40 6 S' c 45.9 0.4 25.9 47.3 3 6 S' Sp . . . 48.3 0 4 49.0 47.2 3 7 47 5 49 0 b pT . . . . 6 2 0 2 0 5 6 2 0 2 0 1 3 5 b Vn 17 8 1 2 13.4 8.9 be .. . 22 7 1 4 25 4 7.4 3 1 b mr . . . 46 6 2 0 7 3 1 9 b Sn. . . 47.6 1.6 49.7 9.2 3.7 47.0 45.0 bba 48.1 6.1 11.8 1.2 0.6 pr c 19.9 1.1 20.5 1.6 2.9 2.0 3.4 1.9 46.5 1.1 47.4 2.9 3.6 47.4 41 5 35.9 0.2 44.6 0.0 c mr .... 0.0 0.0 0.0 C Sn. . . 30.2 0.1 35 3 0.1 0.1 Total1. . 94.2 1.4 56.3 50.3 3.2 47.8 83.2 1 S' b + b pr + pr c + c Sp, except in the last two columns, where pT sp is used. Probably all but the second and fifth columns are too low, since no correction has been made for unob- servable double cross-overs. SUMMARY. Two genes that affect the amount of crossing-over in the second chromosome are discussed. Females of various constitutions with respect to these genes give the results shown in table 25 and figure 1. C//j, located somewhere to the left of purple, decreases the amount of crossing-over between star and purple in females heterozygous for it. Cjjr, located between purple and speck, reduces the amount of crossing-over between purple and speck in females heterozygous for it; but females homozygous for CIIr show the usual amount of crossing-over. Neither of these genes causes any change in the usual condition of no crossing-over in males. An incompletely investigated case of increased crossing-over be- tween purple and curved is apparently due, in part at least, to a dominant third-chromosome gene. A cross-over gene, located in the third chromosome, affects that chromosome in much the same way that C//r affects the region in which it lies. In all these cases the amount of crossing-over is changed, often markedly so. But the sequence of the genes is unchanged ; and females of any one constitution give as consistent results as do " normal" females. APPENDIX. DETAILED DATA. In the following tables it is to be understood that when a theoretically possible cross-over class is not set down no flies representing such a cross- over appeared in the series involved. TABLE 26. ONE ORIGINAL NOVA SCOTIA CHROMOSOME. Culture. Non-cross-overs. Cross-overs. Total. + b mr b mr 645 98 96 38 40 0 2 1 2 137 140 646 Total 194 78 2 3 277 Culture. Non-cross-overs. Cross-overs. Total. + Vg Sp Vg SP 7 55 75 72 61 32 156 161 44 59 46 52 23 169 178 0 2 2 0 0 0 0 0 0 0 0 0 0 0 99 136 120 113 55 325 339 68 69 109 170 199 200 Total 612 571 4 0 1,187 Culture. Non-cross-overs. Cross-overs. Total. + C Sp c Sp 193 111 89 143 66 87 79 99 86 61 122 56 63 23 86 0 0 0 0 0 0 0 1 0 0 0 0 0 0 198 150 265 122 150 102 185 281 ; . . 282 308 310 527 541 Total 674 109 497 99 0 0 1 0 1,172 208 52701 1 Second brood of same. 331 332 APPENDIX. TABLE 26 — continued. Culture. Non-cross-overs. Cross-overs. Total. + b prc mr 1 2 1, 2 b Pr c "*r b pr c mr Pr b c mr 675 67 132 84 101 123 105 104 105 54 97 61 96 73 49 125 38 66 106 79 83 43 62 89 45 69 52 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 3 1 1 1 0 0 4 1 0 2 2 1 0 2 0 4 2 1 0 7 1 3 1 3 2 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 122 258 127 170 230 184 199 150 119 189 111 168 125 683 691 691o.... 692 695 726 727 758 759 760 761 762 Total . . 67501... 684a» . . . 69161. .. 69201 . . . Total . . 1,202 906 0 1 16 26 1 0 2,152 163 145 164 52 112 112 124 62 0 0 0 0 3 0 0 0 0 0 2 1 2 2 2 0 0 1 0 1 0 0 0 0 280 260 292 116 524 410 0 3 3 6 2 0 948 1 Second broods from 675, 684, 691, and 692 above. Non-cross-overs. Cross-overs. Culture. 2 Total. bpx + S'b Px 733 91 103 3 3 200 735 88 102 3 3 196 Total 179 205 6 6 396 Non-cross-overs. Cross-overs. Culture. 1 2 1, 2 Total. + pr c b prc b pr c Pr b c 368 90 42 0 0 0 0 0 0 132 369 124 70 0 1 0 1 0 0 196 370 82 57 0 0 3 0 0 0 142 410 67 33 0 0 0 0 0 0 100 411 100 68 1 1 3 2 0 0 175 412 94 80 0 0 1 0 0 0 175 434 236 160 0 1 1 0 0 0 398 436 88 87 0 0 0 0 0 0 175 437 82 43 1 0 0 0 0 0 126 441 151 126 0 0 2 2 0 0 281 442 112 76 1 0 1 2 0 0 192 443 156 125 1 0 1 2 0 0 285 444 124 101 0 0 0 0 0 0 225 445 247 227 1 1 0 0 0 0 476 APPENDIX. 333 TABLE 26 — continued. Culture. Non-cross-overs. Cross-overs. Total. + b pT c 1 2 1, 2 . b prc bp, c PT be 460 211 201 200 165 133 166 58 140 107 82 94 112 122 84 64 87 113 74 115 97 106 116 134 122 123 132 92 104 137 116 91 98 170 132 136 103 81 95 41 136 58 66 100 77 122 56 34 70 77 54 74 87 72 43 66 73 76 107 73 54 64 53 65 63 3 3 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 3 0 0 1 0 1 0 0 0 0 0 0 0 1 1 1 1 2 0 0 3 1 2 2 1 1 2 0 0 2 1 2 1 2 1 1 0 0 1 1 0 2 0 0 0 2 0 0 1 1 0 1 5 2 0 1 0 1 0 0 1 1 0 1 2 0 3 1 0 3 0 1 2 0 0 0 0 0 1 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 385 339 339 269 217 265 99 278 173 151 196 192 245 143 100 157 191 131 190 189 184 161 204 198 199 243 166 160 203 171 156 161 461 462 463 464 474 475 476 477 494 501 502 503 505 507 553 554... 581 601..... 639 676 677 678 679 680 681 ... 681a.... 682 684 685 693 694 Total.. 4781.... 502a!. .. 5101 60101 . . . 694C1. .. Total. . 5,549 3,873 82 75 81 52 93 14 12 44 38 3 0 9,533 109 113 142 87 94 0 0 0 0 0 2 0 0 0 0 4 0 1 1 0 5 0 3 0 1 0 0 1 0 0 0 0 0 0 0 202 188 228 140 188 545 383 0 2 6 9 1 0 946 1 Second broods from above. Culture. Non-cross-overs. Single cross-overs. Total. + bprvg(aT')8p 2 3 4 bpr vg (aT) Sp b PT vg (or) Sp 258 121 127 77 129 146 64 61 74 41 93 99 51 1 I 0 2 2 0 0 1 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 183 203 119 225 248 116 259 262 304 656 659 Total .... 664 419 6 4 0 0 1 1,094 334 APPENDIX. TABLE 26 — continued. Non-cross-overs. Cross-overs. Total. + S' b PT C 8p 2 3 S'b prc Sp S'bpr C Sp 819 112 108 110 111 0 1 0 0 1 0 0 2 223 222 SlQo1 1 Second brood from above. SPECK END REPLACED. Culture. Non-cross-overs. Cross-overs. Total. + b pr vj (ar) sp 2 1, 2 b pr vg (ar) sp Pr b vg (ar) Sp 422 133 144 3 2 1 0 283 Culture. Non-cross- overs. Cross-overs. Total. Culture. Non-cross- overs. Cross- overs. Total. 1 2 + b prc b Sp + b sp b prc bpr c 477 492 Total. 175 116 97 90 I 0 0 0 4 1 I 0 1 278 207 337 341 Total. 125 118 188 134 2 3 0 2 315 257 291 187 1 0 5 485 243 322 5 2 572 ONE RECONSTITUTED NOVA SCOTIA CHROMOSOME. Culture. Non-cross-overs. Cross-overs. Total. + b pT c Sp 2 bpr CSp 786 787 Total . 138 137 141 133 4 1 0 1 283 272 275 274 5 1 555 HETEROZYGOUS Cnr. Culture. Non-cross- overs. Cross- overs. Total. Culture. Non-cross- overs. Cross- overs. Total. + b pT b Pr + be b c 352 142 151 143 142 8 4 3 10 296 307 468 72 65 10 7 154 353 Total . . 293 285 12 13 603 APPENDIX. 335 TABLE 26 — continued. Culture. Non-cross- overs. Cross- overs. Total. Culture. Non-cross- overs. Cross- overs. Total. 6 c + be + 6 mr 6 mr 277 64 104 114 55 38 133 80 65 3 9 10 3 0 9 9 5 105 255 213 128 653 183 146 80 97 89 95 12 11 10 1 2 8 293 248 193 377 654 655 Total . 379 380 Total . . 409 281 33 11 734 337 316 25 23 701 Culture. Non-cross- overs. Cross- overs. Total. Culture. Non-cross- overs. Cross- overs. Total. + pcr Pr c + C Sp c 8p 351 64 173 70 69 156 77 4 0 0 2 0 0 139 329 147 226 102 44 0 0 146 495 517 Total . . 307 302 4 2 615 HETEROZYGOUS C//r. Culture. Non- crossovers. Crossovers. Total. + b pr c 1 2 1, 2 b PfC bpT c Pr be 318 136 106 103 72 109 96 136 143 80 177 58 142 156 196 205 184 178 178 145 177 132 137 51 91 105 81 42 49 77 114 111 53 179 39 124 151 167 191 124 136 147 95 104 89 122 49 7 7 7 3 8 9 4 7 2 9 1 2 6 14 11 21 10 7 2 7 7 7 2 6 4 4 6 4 4 10 8 3 6 1 2 8 11 15 18 16 10 5 2 4 8 4 2 4 1 0 3 1 6 3 0 0 1 0 1 5 4 3 2 3 6 4 1 1 2 1 0 1 0 2 4 0 2 0 1 0 0 1 1 2 3 4 0 1 3 3 1 1 0 0 0 0 0 0 0 0 0 1 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 243 226 197 124 175 191 270 274 138 373 100 271 323 395 430 353 346 345 254 297 237 277 109 354 355 416 417 430 431 432 433 446 448 449 450 467 469 470 471 472 473 480 496 500 511 Total. . 450a1.... 472a» Total . . 3,096 2,440 160 159 53 31 5 3 5,947 95 114 100 102 7 23 7 14 4 6 4 3 1 1 0 0 218 263 209 202 30 21 10 7 2 0 481 1 Second broods from 450 and 472 above. 336 APPENDIX. TABLE 26 — continued. Culture. Non- crossovers. Crossovers. Total. bpr c 1 2 1, 2 Pr be + bpfC b prc 686 114 52 48 44 53 94 108 89 49 55 62 44 67 94 6 5 3 11 1 1 8 4 3 6 2 1 7 7 4 1 0 0 1 0 0 3 2 0 2 0 0 3 0 0 1 0 0 0 0 1 0 0 0 0 0 0 221 112 113 121 100 169 220 687 707 708 709 719 721 Total. 70701... 708a»... Total. 513 460 35 30 6 10 1 1 1,056 107 108 82 122 10 6 1 2 1 0 1 1 0 0 0 0 1 202 240 215 204 16 3 1 2 1 442 Culture. Non- crossovers. Crossovers. Total. b prc 1 2 + bprc Pr be 512 136 104 142 118 11 23 6 3 0 1 2 0 259 287 513 Total 278 222 34 9 1 2 546 Culture. Non- crossovers. Crossovers. Total. -f- bprSj, 1 2 b PrSp bp, 8p 569 101 92 74 80 7 2 0 5 3 1 3 1 206 163 702 Total 175 172 9 5 4 4 369 Culture. Non- crossovers. Crossovers. Total. bpr 8P 1 2 1, 2 b sp Pr + bprSp 6 PfSp 545 546 547 Total. 546a»... 97 99 73 91 91 90 6 15 4 22 22 3 11 13 3 12 14 0 0 1 5 2 3 0 241 258 178 269 272 25 47 27 26 6 5 677 106 128 26 24 17 18 3 1 323 1 Second broods of 707 and 708 above. s Second broods of 546 above. APPENDIX. TABLE 26 — continued. 337 Culture. Crossovers, crossovers. Total. Culture. Non- crossovers. Cross- overs. Total. 117 + btgs 1 bpfSp vff 1 b Vg Sp Pr8p b vg 648 85 122 56 11 4 95 3 2 156 222 622 39 62 9 7 64801 .... Culture. Non- crossovers. Crossovers. Total. S' bpT8p 123 1, 2 1, 3 2, 3 *" ft e" * 1 4 £ SQ 03 b S'*j , bpr S'b £ PT Sp -o OQ Sp S'P, b Sp 696 45 20 61 37 39 47 27 60 42 42 26 48 5 20 15 2 38 45 1 20 24 3 29 32 3 2 4 1 1 6 0 2 0 1 2 0 0 2 0 1 1 0 2 0 0 2 0 1 1 2 0 0 0 1 0 1 1 0 0 0 0 0 0 0 1 0 1 0 0 0 181 90 217 128 151 699 715 716 717 . Total. 696a2 . . . 699a2 . . . Total. 202 218 133 164 14 12 7 3 3 6 1 2 1 1 767 63 17 55 28 50 46 1 23 20 2 2 1 6 0 0 1 0 0 2 0 1 1 0 1 0 0 0 0 220 100 80 83 73 66 3 8 1 1 0 3 1 1 0 0 320 Culture. Non-crossovers. Crossovers. Total. + b pr c n 1 1r 2 6 Pr c mr b pr c mr 723 93 82 13 115 94 7 7 101 79 4 5 64 72 9 3 2 4 5 3 4 5 1 2 185 232 195 153 724 725 763 Total 373 327 21 18 14 12 765 Culture. Non-crossovers. Crossovers. Total. Sp b pT c 1 2 1, 2 bsp prc b pr sp c Pr sp be 742 142 150 143 56 133 8 4 8 10 8 6 6 9 1 2 1 1 0 1 3 3 0 0 0 1 0 0 0 0 311 291 142 271 743 136 747 71 749 112 Total 4fi1 482 30 29 5 7 1 0 1,015 1 Second brood of above. 2 Second broods of 696 and 699, above. 338 APPENDIX. TABLE 26 — continued. Culture. Non- crossovers. Crossovers. 1 2 3 4 S'bsp Pr C S'pTc b sp S'bprc 8p S'bc Pr Sp S'b PrCSp 796 82 68 82 71 73 56 69 52 56 54 70 72 71 42 49 36 60 53 73 42 54 59 66 52 5 4 13 7 6 6 11 5 8 7 9 5 1 0 3 1 2 0 0 2 1 3 1 3 0 0 0 0 0 0 0 0 1 0 0 0 797 798 799 800 801 Total 432 373 311 346 41 45 7 10 0 1 79601 90 68 80 68 65 64 62 45 30 64 53 38 4 7 0 6 8 7 6 3 1 2 2 2 0 0 0 0 0 0 79701 798a* Total 238 197 137 155 11 21 10 6 0 0 Culture. Crossovers — Continued. 1, 2 1, 3 2, 3 2, 4 3, 4 1, 2, 3 Total. S'sp bprc Sr pr SP be S' b pT sp c S'bpTc sp + S' b c Sp Pr 1 0 0 0 0 1 S'c b PT sP 796 5 3 0 1 3 0 5 2 2 6 4 1 1 1 0 0 0 1 0 0 0 1 2 5 1 0 0 0 0 0 I 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 325 221 269 248 296 256 797 798 799 800 801 Total. . 79601 .... 797a» .... 798a*. ... Total. . 12 20 3 8 1 1 1 0 0 1 2 0 1,615 1 1 4 0 1 2 0 4 0 1 0 0 2 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 307 259 228 6 3 4 1 2 0 0 0 0 1 1 794 Second broods from 796, 797, and 798 above. Culture. Non- crossovers. Crossovers. Total. S'b c 2 S'bc + 776 795 Total . 795a1.... 52 51 54 43 22 8 16 8 144 110 103 97 30 24 254 51 45 6 5 107 1 Second brood of 795, above. APPENDIX. TABLE 26 — continued. 339 Culture. Non- cross- overs. Crossovers. Total. Culture. Non- cross- overs. Crossovers. Total. + ev 05 ^ H If £ .0 1 2 + A 00 A JS 1 2 1, 2 b ft, 9> £ ? & bpr A 00 ^ t 6 A 00 £ bpr Sp Pr bsp 404 . 114 84 7 3 2 0 210 785 64 74 0 1 54 55 1 0 249 Non- Crossovers. crossovers. Culture. 2 3 2, 3 Total. bprc 8p bpr C 8p + b Pr c sp 794 41 50 17 18 46 22 1 5 200 79401... 76 61 19 17 47 35 6 0 261 1 Second brood of 794, above. Non- Crossovers. crossovers. Culture. 2 3 4 Total. S'bc prsp S'bpr C 8p S'bpT C Sp 789 112 116 0 0 5 5 0 0 238 790 95 124 0 0 1 1 1 0 222 791 107 105 1 0 6 6 0 0 225 Total. 314 345 1 0 12 12 1 0 685 789ol... 141 117 0 0 3 4 0 1 266 791o l... 115 124 0 0 1 3 0 0 243 Total . 256 241 0 0 4 7 0 1 509 1 Second broods of 789 and 791 above. Non- Crossovers. Non- Crossovers. crossovers. crossovers. Culture. 2 Total. Culture. 1 2 Total. + b pr c mr Pr be + bprc 713 95 103 6 3 207 754 48 58 1 1 2 o 110 752 126 131 1 1 259 778 61 59 0 n 6 2 128 753 94 82 0 j Tr»tal 1OQ 117 1 i OQS Total. . 315 316 7 5 643 340 APPENDIX. TABLE 26 — continued. C//2 C/jf CUT Culture. Non- crossovers. Crossovers. Total. S' b pT Sp + 1 3 1, 3 1, 2, 3 S' 6 pT Sp S'bpr Sp O Sp bpr S'PT b Sp 7, 7, 7, 7' r 71 7! 7i 7S J6 54 78 37 39 73 31 45 59 84 47 71 35 50 73 49 73 60 89 I 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 39 61 41 32 68 29 40 56 91 44 66 36 27 72 30 52 64 94 0 0 0 0 0 0 0 i 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 186 277 151 148 286 139 210 240 358 J8 }9 . . . . to. . rg. 50 52 53 54 Total 500 547 1 2 457 485 1 0 2 0 1,995 Ci crossovers. Crossovers. otal. Culture. Non- crossovers. Cross- overs. Total. ilture. 2 1 1, 2 6 Sp + b Sp bpr Sp PT bsp 6f 67 0 27 31 18 24 41 49 47 0 0 1 0 101 192 34C 72 72 70 49 263 0 55 Total . 82 72 67 71 0 1 293 CUT Culture Non- crossovers. Crossovers. t 1 2 3 1, 2 1,3 2,3 1, 2 , 3 " Total. S' PT Sp b S' bpTsp + S'bsp PT S' bprsp S'pr b Sp S'bp, Sp S'sp bpr 885.... 886.... 887.... 888.... Total 40 43 26 30 44 47 30 31 17 27 14 14 54 31 24 23 3 1 1 3 2 0 0 6 30 37 15 45 27 39 29 20 0 0 0 0 0 1 1 0 23 22 7 32 16 18 17 21 1 4 1 2 2 1 0 1 0 1 0 0 0 2 1 0 243 278 144 262 155 148 94 98 8 8 127 115 0 2 84 72 8 4 1 3 927 Culture. Non- crossovers. Crossovers. Total. Cult Non- crossovers. Crossovers. Total. b pT Sp 1 2 1, 2 ure. b *, + 6 Sp bsp pr + bprsp b Pf Sp 745 75 90 79 51 2 1 4 2 41 43 30 47 1 3 0 2 243 228 623 41 36 39 35 151 748 Total.. 154 141 6 3 84 77 1 5 471 LITERATURE CITED. BAUR, E. 1912. Vererbungs-und Bastardierungsversuche mit Antirrhinum. II. Fak- torenkoppelung. Zts. ind. Abst. Vererb. 6; 201. BRIDGES, C. B. 1915. A linkage variation in Drosophila. Journ. Exper. Zool., 19; 1. , and T. H. MORGAN. 1919. The second chromosome group of mutant characters. This publication. CHAMBERS, R. 1914. Linkage of the factor for bifid wing. Biol. Bull. 27; 151. MORGAN, T. H. 1912. Complete linkage in the second chromosome of the male. Science, n.s., 36: 719. . and C. B. BRIDGES. 1916. Sex-linked inheritance in Drosophila. Carnegie Inst. Wash. Pub. 237. , A. H. STURTEVANT, H. J. MULLER, and C. B. BRIDGES. 1915. The mechanism of Mendelian heredity. New York. MULLER, H. J. 1916. The mechanism of crossing-over. Amer. Nat. 50; 193, 284, 350, 421. . 1917. An Oenothera-like case in Drosophila. Proc. Nat. Acad. Science, 3; 619. PLOUGH, H. H. 1917. The effect of temperature on crossing-over in Drosophila. Journ. Exper. Zool. 24; 147. PUNNETT, R. C. 1913. Reduplication series in sweet peas. Journ. Genet., 3; 77. . 1917. Reduplication series in sweet peas. II. Journ. Genet., 6; 185. STURTEVANT, A. H. 1913. The linear arrangement of six sex-linked factors in Droso- phila, as shown by their mode of association. Journ. Exper. Zool., 14; 43. . 1913a. A third group of linked genes in Drosophila ampelophila. Science, n.s., 37, 990. . 1915. The behavior of the chromosomes as studied through linkage. Zts. ind. Abst. Vererb., 13; 234. . 1917. Genetic factors affecting the strength of linkage in Drosophila. Proc. Nat. Acad. Science, 3; 555. TANAKA, Y. 1913. Gametic coupling and repulsion in silkworms. Journ. Coll. Agr. Tohoku Imperial Univ., Sapporo, Japan, 5; 115. . 1914. Further data on the reduplication in silkworms. Journ. Coll. Agr., Tohoku Imper. Univ., Sapporo, Japan, 6; 1. 341 IV. A DEMONSTRATION OF GENES MODIFYING THE CHARACTER "NOTCH," BY T. H. MORGAN. 343 A DEMONSTRATION OF GENES MODIFYING THE CHARACTER "NOTCH." BY T. H. MORGAN. Two main topics are dealt with in the following pages from the standpoint of the experimental results obtained. One of them con- cerns the demonstration of modifying genes that were involved in the results of a selection experiment. The other topic is a discussion of the possibility of contamination of genes as a method that has been appealed to as an influence vitiating the regularity of Mendelian phenomena. The claim of the Mendelians that genes have been found to be sta- ble in successive generations wherever a critical test of them was made has been challenged both on the grounds of empiric observation and on the more sentimental grounds that such hard and fast rules do not apply to living things which are rather to be thought of as variable quantities. In the following pages an account is given of a character that changed in the course of selection and a demonstration that the result was due to a modifying gene and not to contamination between the notch gene and its normal allelomorph, despite the fact that an exceptional opportunity was given to contaminate the gene, if contami- nation is a possible process. In 1915, Dexter described a mutant type of Drosophila called Notch or "perfect Notch," and made out the main points in the heredity of the character. The gene is sex-linked, and dominant for the serra- tion that it produces in the wings, but recessive in its lethal effect. Since the gene is carried by the X chromosome, any male that gets a chromosome with this gene dies, while the female that has another X carrying the normal allelomorph lives and shows the notch at the end of her wings (fig. 91). Since no male that has the notch gene can live, it is not feasible to determine whether a female containing two lethal bearing X's would also die.1 Every heterozygous Notch female gives twice as many daughters as sons, because, as stated, half of the sons die, namely, those that get the lethal-bearing Xn. The scheme is as follows: X — JTeggs X — Y sperm XX — XX* — XY — XT (dies) 9 9 tf rf 1 Unless an XX egg, arising through "equational non-disjunction," were fertilized by a Y sperm and lived. No such females have appeared. They would have no regular sons. 345 346 GENES MODIFYING NOTCH. Half of the daughters are normal, half are heterozygous Notch. The normal daughters and normal sons never transmit the Notch gene, which, therefore, never gets into the male line or into the line of normal daughters. FIG. 91. Dexter obtained his mutant in a cross between beaded and wild. The Notch that I used arose independently in descendants of ves- tigial flies, in which stock the factor may have already existed. This mutant has, however, originated several times in other cultures in the laboratory. It is by no means one of the rarer mutations. VARIATION OF NOTCH. The most conspicuous character of the female heterozygous for Notch is the serration at the end of the wings (fig. 91) caused by the absence of the marginal bristles and generally accompanied by a slight GENES MODIFYING NOTCH. 347 concavity of the edge. The range of variation of the notching is very wide. That the limit of variability overlaps in one direction the nor- mal wing is certain, for amongst the daughters without notching occa- sionally one is found half of whose daughters are notched. The not unusual occurrence of a fly with one entire wing and one with notching (fig. 91, c) indicates that the range of variation includes normal wings. The low productivity of the Notch female appears to be an incidental effect of the Notch factor, because the normal sisters of the same stock are, whenever tested, much better producers. The viability of the Notch females is fairly good, but they appear to run behind their nor- mal sisters in nearly all cultures. Change in the viability will be dis- cussed later. THE PROBLEM. Throughout the older literature dealing with selection, the idea that the grade of any character shown by animals or plants is a criterion of the condition of the genetic factor or gene responsible for the character continually recurs, and the same idea appears occasionally in more recent tunes, despite Johanssen's analysis showing the inconsequence of such an argument, and despite the accumulated demonstrations that the production of a given character depends on the environment and on internal modifying genes, as well as on the principal gene itself. The wide range of variability of the notching, the fact that the females genetically Notch may be identified by the 2 to 1 sex-ratio in the off- spring, even when the wings themselves have somatically the normal margin, as well as the fact that it is a dominant, and therefore any alteration in the gene may be tested directly by outbreeding; the fact that linkage relations made it possible to identify any changes that might follow selection in the individuals that were Notch, although with normal wings; all these made Notch excellent material on which to put to actual test some of the older as well as current views con- cerning the nature of Mendelian factors and the influence of selection. In each generation several (usually 2 to 10) virgin, notched- winged females (of the derived type) were picked out and put into a new bottle with one to 10 males. Occasionally pairs were used and then mass selection followed in the next generations. This prodecure is not un- like the rough procedure formerly practised by the breeder, but is not, of course, to be recommended for a thorough understanding of the changes that are taking place during the selection period. Moreover, by such a method the end result is attained only after a long tune, whereas the results here described could probably have been reached in two or three generations; for, as the duplicate experiments show, the modifying gene for "slight notch" did not arise in the course of the experiment, but was present in some flies of the stock at the beginning. 348 GENES MODIFYING NOTCH. On the other hand, had the sequel shown that the results were due to a number of factors present at the beginning, the mass-culture method would have offered a better chance of collecting the different modi- fiers in the same strain. The object of the selection process here prac- tised was, however, to produce by a rough method results of the kind familiar to the breeder, and then to show, by the refined tests that the Drosophila work has made possible, what had been done to the original stock. CONDITION OF STOCK BEFORE SELECTION. In table 1 there are records of the offspring of 11 pairs of Notch females by normal males. The totals give 577 Notch females, 608 normal females, 613 normal males. It is clear that the viability of the Notch females compares favorably with that of the normal females. Very few of the Notch flies could have had normal wings when this class comes so near to the realization of their expected numbers. How- ever, there were other females that had the same origin in which the ratios amongst the offspring were strikingly different. These are given here in table 2. TABLE 1. Ref. Notch. Normal 9 • Normal cf. Ref. Notch. Normal 9 - Normal c?1- FT 40 49 35 50-1 45 51 60 PN .. . 9 8 9 PN .... 42 52 55 Pn 29 21 32 SS 37 46 41 50-1 40 40 42 50-1 205 219 212 T>>T 42 46 11 PN 36 35 27 Total . . . 577 608 613 PN 52 40 46 TABLE 2. Ref. Notch. Normal 9 • Normal cf. Ref. Notch. Normal 9 • Normal d"1 u 25 91 36 SSG 89 137 114 Pu 47 77 65 SSO 49 115 99 SS 35 126 147 DO 52 170 106 SS 39 122 97 EVN 75 182 172 Total. . 409 1,020 836 In these 8 sets the Notch females are not half as numerous as the males and less than half as numerous as the normal females. The normal females are greatly in excess of the males. If we suppose that here a considerable number of the Notch females have normal wings — as was actually shown to be the case later in the offspring of some of these sets — the discrepancies between tables 1 and 2 may be accounted for. Thus, if we add the two classes of females (1429) and divide by GENES MODIFYING NOTCH. 349 2 to give the expected number of Notch females (viz, 714), the results would mean that about 300 of the Notch females had varied into the normal class of females. We may make the comparison in another way. If the number of the males be taken as the measure of each class of females, there will be over 400 too few Notch females, and about 200 too many normal females. It was the offspring of some of these lots, viz, the SS lots, that later furnished the materials for selection (SSO, SSO 1, etc.). If the above interpretation be accepted as plausible, then at the beginning of the experiment either different genes for Notch were present or modifying genes were there. The later tests proved the presence of a modifying gene, but since this is not sex-linked, it may have been present in certain of the females or males either in heterozygous or homozygous condition, hence, until the stock could be made homozygous for this gene, random selection would be expected to give for some time variable results. SELECTION OF FEMALES HAVING NOTCH IN ONE WING ONLY. If the somatic characters were an index of the condition of the differ- entiating factor for a character, it would appear that those flies in which the character appeared in only one wing should indicate a change towards the phenotypic normal end of the variation curve. Hence by selecting in successive generations as parents those flies that had the character only in one wing, and amongst these only those in which it was developed to the slightest visible extent, then one might expect to bring about a change, but of course this would be equally true whether the selection was based on a changing factor or on the more frequent presence of one or more modifying factors. An experiment of this sort was begun in the third generation after SSO (viz, in SSO 112)and continued through 11 generations, with the result shown in the table 3. In the first column are given the flies in which both wings are notched, in the second the flies with a notch in only one wing, in the third the females with normal wings, and the fourth the males. I have indicated by the star (*) those records in which it appears that a considerable portion of the potential Notch females fall into the phenotypic normal class as shown by the excess of normal females and the deficiency of Notched females over the number of the males. This change is notice- able in the sixth to the eleventh generation. In the last 4 generations this relation holds for all the cultures, with two exceptions only in the eighth generation. It is probable, therefore, that at this time the full force of selection has been accomplished and there is nothing to indicate that unless some new sort of change were to occur, selection would accomplish anything further after the ninth generation. 350 GENES MODIFYING NOTCH. SELECTION OF SOMATICALLY NORMAL WINGED FEMALES THAT ARE GENETICALLY NOTCHED FEMALES. At the beginning of the work a few lines were run with eosin ruby males which were bred to the Notch females, but the history of these TABLE 3.— SS Set. Gen. Both Notch 9. One Notch 9. Normal 9. Normal tf. Gen. Both Notch 9. One Notch 9. Normal 9- Normal cf. 1 1 1 1 Total 2 2 2 Total 3 3 3 3 Total 4 4 4 4 Total 5 5 5 5 5 5 5 5 5 Total 6 6 6 6 6 6 6 Total 17 15 8 23 14 2 7 19 88 18 27 136 *63 18 29 *109 7 7 7 7 7 7 7 7 7 7 7 7 Total 8 8 8 8 8 8 Total 9 9 9 9 9 Total 10 10 10 10 10 10 10 10 10 10 Total 11 11 11 11 Total 0 3 0 0 4 1 7 0 6 18 3 1 7 8 4 7 5 15 7 7 11 12 7 1 35 19 19 14 15 76 12 14 31 228 28 29 *9 15 21 14 12 *29 9 14 29 *167 *19 *11 63 42 269 219 9 9 38 4 4 14 19 19 55 4 4 64 56 22 93 72 5 12 11 16 1 9 4 12 17 28 38 49 7 27 *40 50 43 91 520 349 2 11 7 5 11 1 15 10 12 13 5 19 21 29 72 104 24 130 22 20 *37 *72 *14 *68 44 26 132 124 6 7 15 14 4 4 2 17 42 13 29 29 37 14 25 25 37 74 380 233 1 2 8 2 4 5 9 23 8 23 45 56 161 38 105 *30 *33 *75 *18 *73 42 17 113 101 13 3 16 4 19 4 11 3 6 7 10 4 1 3 5 13 2 4 24 35 38 5 45 17 52 6 22 14 26 *22 5 *33 17 *24 2 18 17 68 405 229 10 16 15 0 1 5 3 3 0 3 6 13 27 4 11 2 12 8 1 12 45 55 149 24 55 118 65 96 21 85 *19 *43 *88 *24 *26 *64 *42 *42 * 9 *42 79 49 254 161 9 3 6 3 6 5 1 4 8 20 3 14 7 10 11 32 127 14 102 40 39 8 18 *41 13 *61 *20 *19 56 96 713 393 6 10 8 1 19 10 8 6 107 39 63 50 *42 *19 *49 *27 33 66 365 180 25 43 259 137 GENES MODIFYING NOTCH. 351 lines is not clearly separable now from those recorded in the last sec- tion. There is, moreover, the possibility that during these early expe- riments, stock males of eosin ruby may have been introduced at one stage. That these conditions have not affected seriously the condition of the selected stock as a whole is shown by table 4, where the number of normal females belonging to the potential Notch class is as high in most cases as in the middle and latter parts of the preceding table. By introducing into the experiments the two genes eosin and ruby, it is a very simple matter to identify potentially Notch females from the other females with normal wings. Selecting the former makes it possible to carry on the experiment by breeding in every generation from those females that carry the factor for Notch, but do not show a notch in the wing. In other words, if the expression of a character (its phenotype) is a measure of the major factor that produces it, we should expect that an extreme selection of this kind would be an excellent way of fixing the factor altered by selection. The location of the Notch factor had shown that it lies in a region of the X chromosome (fig. 92), 2.8 units from the arbitrary zero-point yellow. Eosin lies 1.5 units and ruby lies 7.3 units from yellow. The distance between eosin and ruby is therefore a distance so short that double crossing-over never takes place within it. If, then, we use a male whose sex-chromosome contains the factors for eosin and for ruby, and a Notch female having red eyes (i. e., the normal allelomorphs of eosin and of ruby) the gene for Notch in one X of the daughters will be located in a position between the eosin ruby genes present in the opposite chromosome of the same daughter, as seen in figure 92. Now, as said above, it would necessitate double crossing-over to get the Notch gene in between the eosin and ruby genes, or, in other words, double crossing-over must take place within these limits to produce a Notch female with eosin ruby eyes. Of the many thousands of females obtained in the course of the experiment, not a single double cross- over of this kind has been observed. Single cross-overs have, however, been recorded in the expected numbers. Thus eosin and Notch females, and eosin as well as ruby males have appeared. It would of course be possible to obtain a Notch fly with eosin ruby eyes by first getting a single cross-over of eosin Notch and then after mating such a female to an eosin ruby male some daughters, in which a cross-over between Notch and ruby would result, having eosin Notch ruby in the chromosome. Such a female bred to an eosin ruby male would give some daughters of the desired class. As there was no need in my work for such females, I have not taken the trouble to make them. Turning to table 4, we see that nothing further resulted from select- ing the potentially normal females through about 5 more generations. By potentially normal females I mean that females with red eyes and 352 GENES MODIFYING NOTCH. without Notch in the wings were selected. All red-eyed females must cany the Notch, whether they show the character or not, as has been explained. At the end of the experiment the relation between the Notch and normal (of two kinds) females was about the same as that after a few generations. TABLE l—SS-162 Set. Gen. Both wings Notch 9 . One wing Notch 9 . Normal 9. Eosin ruby 9. Eosin ruby (7. Eosin Notch 9. Ruby 9. Eosin 9. Notch ruby 9. Ruby d1. Eosin rf. Normal ptarDiehaete S' ' tx 1 X pn \ tx \ 2 V X iv V / tx / 3 / X &d tx 4 s' s' FIG. 93. If this male is now back-crossed to a selected Notch female (see figure 93) any red-eyed Notch daughter that is also Star-Dichsete (upper line to right; No. 1) must have gotten the Star (II) and the Dichaete (III) chromosomes from her father (neither of which bears the modifiers sought for) and an X chromosome also from the father with genes for eosin ruby eyes and normal wings. She must also have gotten the second and third chromosomes that may carry in one or in both the modifiers sought for (which are recessive) from her mother, as well as an X chromosome bearing the genes for red-eye and Notch wing. Hence such a female should be atavistic Notch, because either the S' or D' genes will bring in the normal allelomorph of the postulated modifiers in II and III. Conversely, females that are not Star and not Dichaete (No. 2) should be of the selected type, since their second and third chromosomes, one or both, contain the modifiers. Continuing the analysis, it is evident that if the modifier (one or more) is in the second chromosome, then all Star Notch daughters (No. 3) should be atavistic, and all not-Stars (No. 4) the selected type of Notch; and if the modifier is in the third chromosome, then all Dichsete Notch daughters should be atavistic (No. 4) and all not-Dichsete (No. 3) selected type of Notch. The ability to pick out atavistic flies from selected-type flies is essen- tial to this test. In general, this can be done successfully, with, how- ever, a margin of error, but the error is expected from the information at hand to be so small as not to effect the main result. Moreover, the occurrence of red-eyed females with normal wings (flies that are known from the linkage relations of the experiment to have the Notch gene) in any of the classes named above is an almost certain index of the occurrance of the modifier. GENES MODIFYING NOTCH. 363 The results of such a test are given in table 7. The table includes only females and only the red-eyed females (the flies that are genet- ically Notch), while the eosin ruby females and all of the males were thrown away. Examination of the table shows that practically all of the not-Star, not-Dichsete females have normal wings (potentially TABLE 7. Not star 9 . Star 9- Not-Dichaete. Dichsete. Not-Dichsete. Dichsete. Notch Sel'od. Norm. Sel'ed. Ata- vistic. Notch Sel'ed. Norm. Sel'ed. Ata- vistic. Notch Sel'ed. Norm. Sel'ed. Ata- vistic. Notch Sel'ed. Norm. Sel'ed. Ata- vistic. A.. A.. A". A». A3.. A3.. A". A'. A°. B . 2+1 1+2 1+4 1+1 +2 3+1 1+1 2+4 5+7 6 5 7 6 3 5 9 1 8 10 5 5 4 18 7 1 13 +2 +2 + 1 1 7 8 6 3 7 3 27 5 1 6 2 9 4 8 17 6 2 7 6 13 12 3 10 2 18 9 6 16 3 14 4 27 18 5 3 +1 2 +1 + 1 +1 +1 +1 2+2 + 1 +5 1 +2 +2 2+2 1+5 +1 22 2 3 4 +3 1 C.. D . 2+6 2 E.. F.. H.. K.. P.. P°.. 1+6 +4 3+7 4+8 +3 +5 4 5 11 19 3 7 9 6 27 18 +2 1 1 11 26+62 112 1 + 1 2 152 6+29 122 4+6 176 Notch).1 This is the class that contains the original second and third chromosomes and their modifying genes if such were present. Con- versely, practically all of the Star-Dichsete females are atavistic, and this class contains the Notch females that have received the second and the third chromosomes from the Star-Dichsete males. Thus far the evidence shows that the change that took place during selection is caused by something in one or the other or both of these two chromo- somes. Whether both or only one is shown by further analysis of the results. For instance, the fact that all the Dichsete flies are atavistic, and the fact that all not Dichsete are selected type, shows that the modifier is in the third chromosome. Had the modifying gene or genes been in the second chromosome, then all Star-eyed females should be atavistic, which they are not, and, conversely, all not- Star- eyed females should be selected type, which they are not. Hence the modifier in question is not in the second chromosome. Finally, the same evidence proves that the modifiers that caused the change are not in the sex chromosome as recessive modifiers be- 1 In this table (also in tables 4 and 5) the + sign indicates that the number of flies that follow were notched in only one wing. 364 GENES MODIFYING NOTCH. cause the not-Star not-Dichsete females are practically all the selected type, and the Star-Dichsete are practically all classified as atavistic; yet the females of both classes contain the same Notch-bearing chromo- some that must be identical, since in both it is the X chromosome of the selected stock. In the FI generation (table 8), the parents of the flies in table 7, it was found that all of the Notch females were atavistic as expected. In some sets the extent to which notching was developed was greater than in others. It is important for present purposes to note that there is no difference in the extent of development of the character Notch in the Dichaete and in the not-Dichaste (straight-winged) females. This means that the wing-character Dichsete does not modify the Notch character when present with it. Consequently we should not expect in FI any difference between Dichsete and not- Dichsete Notch females, due to the Dichsete gene. TABLE 8. Not-Dichsete 9 . Dichaete 9 • Not-Dichsete d". Dichsete cf. Notch 9. Not- Notch 9. Notch 9. Not- Notch 9. Eosin ruby cf . Eosin d». Ruby 1__1_ Total Short. Intermediate. Atavistic. cf 9. 1 Short 6 4 x x 0 33 2 Do 5 2 0 x 0 20 3 Do 2 1 x 0 x 33 4 Do 3 1 0 0 x 12 5 Do 15 0 x 0 o 76 6 Do 3 4 X 0 x 15 7 .. .. Do 14 0 X x 0 47 8 . Do 10 1 X x x 20 9 .. . Do 5 1 0 0 x 17 10 Do 12 1 0 0 x 50 11 Do 9 5 0 x x 33 12 Intermediate. . 2 3 0 0 x 3 13 Do 4 19 x 0 x 75 14 Do 11 2 0 0 x 49 15 Do ... 11 6 0 0 x 53 16 Do 18 2 0 x x 40 17 Do 7 6 x 0 x 59 18 Atavistic . . 23 13 0 x x 97 19 Do. (almost). 2 10 0 x 0 21 RECOMBINATION OF BENT AND SHORT NOTCH. As a matter of curiosity, largely, the possible interaction of "bent" in double dose on Notch was determined. Flies of the desired recom- bination were made by crossing bent males to short-Notch females and then by breeding the FI Notch females to their brothers. Amongst the F2 flies were pure bent females that were also Notch. These showed the widest possible range of modification as does the bent character itself. In extreme cases, as shown in figure 95, the wing is as narrow as stumpy wings (see Critique of Theory of Evolution, p. 11, figure 5, c, d). It is noticeable, too, when looking through such a series, that the extent to which the bent factor manifests itself in other ways, such as in the shortening of the legs, is a sort of index of the extent to which the wing is affected. This might, on first thought, be interpreted as furnishing evidence in favor of the view that the extent to which a character develops may be an index of the quantity of the gene present GENES MODIFYING NOTCH. 371 in the egg. But since bent flies with the character well developed may not produce, when bred to each other, any more flies of their own kind than do their normal appearing brothers and sisters, there is nothing gained by making such an assumption. On the contrary, it seems more reasonable, I think, to suppose that the same environment (in the widest sense) that is favorable to the full development of the bent characters make that character the more effective in its influence on short Notch. It seems to me hazardous to base any view concerning the nature of the gene itself on evidence of this kind, as has been done by several recent writers. FIG. 95. CROSSES BETWEEN SHORT NOTCH AND OTHER STOCKS. Unless the modifying factor for short-Notch is partially dominant or unless other stocks carry the modifier, or some other dominant modi- fier, the expectation on crossing short-Notch females to males of other stocks is atavistic types of Notch females. The males of short-Notch stock carry in their sex chromosome, as has been shown, a modifier for short-Notch; hence crossing of such males to selected or to atavistic Notch are expected to show some influence on the character of Notch in their daughters if the factor is dominant or partly so. In the light of these expectations the following crosses are not without interest. 372 GENES MODIFYING NOTCH. SHORT NOTCH BY STAR DICH&TE. There were four crosses of this combination that gave in FI Notch females with intermediate wings shorter, on the average, than the atavistic type, and therefore more on the order of the short type. The FI records are as follows : Notch- Dichaete 9 • Notch 9. Normal 9. Normal ff. Dichaete