®l|p i. B. aHill ICtbrarg North (taroltna g>tatf llntoeraitg QH431 MS6 cop»2 «'^,,,^T*7, UNIVERSITY DH HILL L IBRARY S00376711 P THIS BOOK IS DUE ON THE DATE INDICATED BELOW AND IS SUB- JECT TO AN OVERDUE FINE AS POSTED AT THE CIRCULATION DESK. APR 1 7 1985 "^Cr 2JJ986 OEG-2 8 1,989 %• PEi ^/^% cnK4/o Q/l Monographs on Experimf.ntm biology f EDITED BY JACQUES LOEB, Rockefeller Institute T. H. MORGAN, Columbia University W. J. V. OSTERHOUT, Harvard University THE PHYSICAL BASIS OF HEREDITY BY THOMAS HUNT MORGAN MONOGRAPHS ON EXPERIMENTAL BIOLOGY PUBLISHED VOLUME I FORCED MOVEMENTS, TROPISMS, AND ANIMAL CONDUCT By JACQUES LOEB, Rockefeller Institute IN PREPARATION THE CHROMOSOME THEORY OF HEREDITY By T. H. MORGAN, Columbia University INBREEDING AND OUTBREEDING: THEIR GENETIC AND SOCIOLOGICAL SIGNIFICANCE By E. M. EAST and D. F. JONES, Bussey Institution, Harvard University PURE LINE INHERITANCE By H. S. JENNINGS, Johns Hopkins University THE EXPERIMENTAL MODIFICATION OF THE PROCESS OF INHERITANCE By R. PEARL, Johns Hopkins University LOCALIZATION OF MORPHOGENETIC SUBSTANCES IN THE EGG By E. G. CONKLIN, Princeton University TISSUE CULTURE By R. G. HARRISON, Yale University PERMEABILITY AND ELECTRICAL CONDUCTIVITY OF LIVING TISSUE By W. J. V. OSTERHOUT, Harvard University THE EQUILIBRIUM BETWEEN ACIDS AND BASES IN ORGANISM AND ENVIRONMENT By L. J. HENDERSON, Harvard University CHEMICAL BASIS OF GROWTH By T. B. ROBERTSON, University of Toronto PRIMITIVE NERVOUS SYSTEM By G. H. PARKER, Harvard University COORDINATION IN LOCOMOTION By A. R. MOORE, Rutgers College OTHERS WILL FOLLOW Monographs on Experimental Biology THE PHYSICAL BASIS OF HEREDITY BY THOMAS HUNT MORGAN PROFESSOR OF EXPERIMENTAL ZOOLOGY IN COLUMBIA UNIVERSITY JI7 ILLUSTRATIONS PHILADELPHIA AND LONDON J. B. LIPPINCOTT COMPANY COPYRIGHT, I919. BY J. B. LIPPINXOTT COMPANY Electrotyped and Printed by J. B. LiPPincott Company The Washington Square Press, Philadelphia, U. S. A. EDITORS' ANNOUNCEMENT The rapid increase of specialization makes it im- possible for one author to cover satisfactorily the whole field of modern Biology. This situation, which exists in all the sciences, has induced English authors to issue series of monographs in Biochemistry, Physiology and Physics. A number of American biologists have decided to provide the same opportunity for the study of Experimental Biology. Biology, which not long ago was purely descriptive and speculative, has begun to adopt the methods of the exact sciences recognizing that for permanent progress not only experiments are required but quantitative experi- ments. It will be the purpose of this series of monographs to emphasize and further as much as possible this develop- ment of Biology. Experimental Biology and General Physiology are one and the same science, in method as well as content, since both aim at explaining life from the physico-chemical constitution of living matter. The series of monographs on Experimental Biology will therefore include the field of traditional General Physiology. Jacques Loeb, ^' T. H. Morgan, r\V\ W^\ ^* *^* ^* OSTERHOUT. lA CONTENTS CHAPTER PAOB I. Introduction 15 11. Mendel's First Law — Segregation of the Genes 19 III. The Mechanism of Segregation 39 IV. Mendel's Second Law — The Independent Assortment of the Genes 59 V. The Mechanism of Assortment 73 VI. Linkage 80 VII . Crossing Over 87 VIII. Crossing Over and Chromosomes 96 IX. The Order of the Genes 118 X. Interference 126 XI. Limitation of the Linkage Groups 133 XII. Variation in Linkage 139 XIII. Variation in the Number of the Chromosomes and its Re- lation TO the Totality of the Genes 147 XIV. Sex-Chromosomes and Sex-linked Inheritance 165 XV. Parthenogenesis and Pure Lines 204 XVI. The Embryological and Cyto logical Evidence that the Chromosomes are the Bearers of the Hereditary Units. . 212 XVII. Cytoplasmic Inheritance 219 XVIII. Maternal Inheritance 227 XIX. The Particulate Theory of Heredity and the Nature of THE Gene 234 XX. Mutation 247 ILLUSTRATIONS Fig. page 1. Cross Between a Tall and a Short Race of Garden Peas 20 2. Cross Between White and Red Flowered Four-o'clocks 24 3. Cross Between Splashed-White and Black, in Andalusian 26 4. Male and Female Vinegar Fly 28 5. Normal and Abnormal Abdomen of D. melanogaster 29 6. Relation of Black Body Color to Wild Type as Shown by Classes of Fhes 30 7. Normal, Heterozygous, and Bar Eye of the Vinegar Fly 31 8. Relation of Bar Eye to Normal Eye 31 9. Relation of Andalusian to Splashed White and to Black as Shown by Classes of Birds 32 10. Relation of Tall to Short Peas 32 11. Relation of Normal to Abnormal Abdomen as Shown by Classes of Flies 32 12. Relation of Normal to Duplicate Legs of Fhes 33 13. Notch Wings in the Vinegar Fly 35 14. Oocyte of Ancyracanthus; Growth Period; Nucleus with Tetrads. . . 40 15. Egg of Ancyracanthus 40 16. Eggs of Ancyracanthus within Membrane 41 17. Spermatogenesis of Ancyracanthus 42 18. Last Spermatogonial Division of Tomopteris and Stages Before and During Synapsis 45 19. Thin-Thread Stage of Tomopteris Spermatocyte; Tetrads, and First and Second Spermatocyte Divisions 47 20. Synaptic Stages and Those Immediately Following in Batracoseps ... 48 21. Synaptic Stages and Those Immediately Following in the Egg of Pristiurvs 50 22. Sister Blastomeres of Ascaris Preparatory to Another Division .... 52 23. Normal and Reduced Chromosomes of Biston 53 24. Division Figures in Egg of Ctenolabrus FertiUzed by Fundulus 54 25. Female and Male Chromosome Groups of Protenor 55 26. Reduced Chromosome Group; and Extrusion of Polar Bodies in Protenor "^^ 9 10 ILLUSTRATIONS 27. Reduced Chromosome Group of Male; and Spermatogenesis in Protenor ; 56 28. Diploid and Haploid Chromosome Groups of Drosophila busckii and D. melanica (neglecta) 57 29. Cross Between Wingless and Ebony Vinegar Fly 65 30. Miniature Wing, Dumpy, and Miniature Dumpy 66 31. Combs of Fowls 69 32. Eight Chromosome Groups of Twelve Chromosomes Each of Trimeroiropis 77 33. Back-cross of Male (Out of Black Vestigial by Wild) to Black Vestigial 81 34. Back-cross of Male (Out of Gray Vestigial by Black) to Black Vestigial 83 35. Scheme Showing the Inheritance of the X-Chromosome in Drosophila 84 36. Back-cross of Female (Out of Black Vestigial by Wild) to Black Vestigial Male 89 37. Back-cross of Female (Out of Gray Vestigial by Black) to Black Vestigial Male 90 38. Scheme to Illustrate Double Crossing Over Between White and Forked 93 39. Curve Showing the Influence of Temperature on Crossing Over Control 98 40. Curve Showing the Influence of Temperature on Crossing Over .... 98 41 . Diagram Showing Crossing Over of Two Chromosomes at Four-strand Stage and the Subsequent Opening Out of the Tetrad 101 42. Scheme Showing the Opening Out of the Strands of the Tetrad 102 43. Scheme Showing Crossing Over Involving Both Strands of Each Chromosome 103 44. Spermatogonial Cells in the Last Phase of Division and the Following Resting Stages 105 45. Cells Emerging From the Resting Stages Preparatory for the Next Spermatogonial Division 106 46. Cells Emerging From Their Last Spermatogonial Division 106 47. Formation of a Thick Thread after Synapsis and the Following Condensation of a Tetrad 107 48. A Pair of Chromosomes in Conjugation 109 ILLUSTRATIONS 11 49. The Same Chromosome Pair in Conjugation from Thirteen Different Cells 110 50. Conjugation of an Unequal Pair of Chromosomes and Their Subse- quent Separation Ill 51. Two Schemes Illustrating the Idea of Reduplication by Bateson and Punnett 116 52. Scheme Illustrating How Double Crossing Over Between Two Distinct Genes takes Place 121 53. Chromosome Groups of Pea, Wheat, and Primula 135 54. Types of Chromosome Groups Found in Drosophila 136 55. Haploid Group of Chromosomes of the Silkworm Moth 137 56. Curve Showing Influence of Crossing Over at Different Temperatures 142 57. Diagram Illustrating the Effect on Crossing Over Due to the Presence of Crossover Grenes 143 58. Chromosome Group of (Enothera Lamarckiana and 0. gigas, and Triploid Group 149 59. Life Cycle of Moss 152 60. Diagram Illustrating the Formation of Individuals from the Regener- ation of the Sporophyte in a Dioecious Species 153 61. Diagram Illustrating the Formation of Individuals from the Regener- ation of the Sporophyte in a Hermaphroditic Species 153 62. Somatic Chromosomes Groups of (Enothera sciniillans 156 63. Scheme Showing the Probable Relation Between the Extra Chromo- some Pieces of Fig. 62, and the Normal Fifteen Chromosomes of This Mutant 158 64. An Egg of Ascaris hivalens Fertilized by Sperm of A. univalens . . . . 160 65. Diploid and Haploid Groups of the Sundew Drosera 160 66. A Scheme Illustrating the Fertilization of the Egg of One Species of Moth by the Sperm of Another 161 67. Scheme Illustrating the History of the Chromosomes, and the Back- cross Between a Hybrid Male and One or the Other Parent .... 162 68. Scheme Showing the Relation of the Sex- Chromosome to Sex-De- termination 166 69. Cross Between White-Eyed Male and a Red-Eyed Female of the Vinegar Fly 168 12 ILLUSTRATIONS 70. Cross Between White-Eyed Female and a Red-Eyed Male of the Vinegar Fly 169 71. Cross Between a Yellow White-Eyed Female and a Wild-Type ("Gray"), Red-Eyed Male 171 72. The Results from the Reciprocal Cross of That Shown in Fig. 71 . . . 173 73. Scheme Showing the Relation of the Sex-Chromosomes of the Moth in Sex Determination 174 74. Cross Between Abraxas lacticolor Female, and Grossulariata Male .... 175 75. Cross Between Abraxas grossulariata Female and Lacticolor Male. . . 176 76. Cross Between Barred Plymouth Rock Male and Black Langshan Female 178 77. Scheme Showing the Transmission of the Sex- Linked Characters .... 178 78. Cross Between Black Langshan Male and Barred Plymouth Rock Female 178 79. Scheme Showing the Transmission of the Sex-Linked Characters Shown in Fig. 78 - 179 80. First and Second Spermatocyte Divisions in the Bee 181 81. First and Second Spermatocyte Divisions in the Hornet 182 82. Life Cycle of Phylloxera caryoecaulis 182 83. Extrusion of the Polar Body from a Male-Producing Egg 183 84. First and Second Spermatocyte Divisions in the Bearberry Aphid .... 184 85. Hydatina senta: Adult Female, Young Female Soon After Hatching, Adult Male, Parthenogenetic Egg, Male-Producing Egg, Resting Egg 186 86. Diagram Showing How a Continuous Diet of Polytoma Through Twenty-Two Months Yielded Only Female-Producing Females. . . 187 87. A, Gynandromorph of Drosophila melanogaster , that was Female on Right Side and Male on the Left; B, Female on the Left Side and Male on the Right 190 88. Diagram Showing Ehmination of X' at an Early Cell Division 191 89. Caterpillars of the Silkworm Moth 192 90. Diagram Illustrating How a Heterozygous Egg With Two Nuclei Fertilized by Two Sperms Might Produce a Gynandromorph Uke that Shown in Fig. 89 193 91. Scheme Showing the Transmission of a Lethal Sex-Linked Factor in an X-Chromosome 199 92. Normal Female and Male Groups of Chromosome of the Vinegar Fly 200 ILLUSTRATIONS 13 93. Non-Disjunction. Egg Fertilized by X-Sperm 201 94. Non-Disjunction. Egg Fertilized by Y-Sperm 202 95. A Wingless Aphid and a Winged One 207 96. Curve Showing the Non-efifect of Selection for the First Twelve Generations for Increase in Body Length 208 97. Curve Showing the Effect of Selection for the Second Score of Generations 209 98. Scheme Showing Dispermic Fertilization of the Egg of the Sea Urchin 214 99. First Division of a Hybrid Egg 215 100. Fertilization of an Egg Starting to Develop Parthenogenetically . . . 216 101. Larval Sea Urchin Seen in Side View 217 102. Green Leaf and Checkered Leaf of Four-o'clock 220 103. Pelargonium that Gave Rise to a White Branch 221 104. Diagram to Show How a Sectorial Chimera May be Produced 221 105. Diagram to Illustrate Maternal Inheritance 228 106. Diagram to Show the Inheritance of Two Pairs of Mendelian Characters 238 107. A, Hen-Feathered Campine Male; B, Adult Castrated Campine Male ; C, Sebright Hen-Feathered Male, D, Adult Castrated Sebright Male 246 108. Diagram Illustrating Mutation in a Nest of Genes 252 109. Two Flies (Drosophila) with Beaded Wings 258 110. Diagram Showing the Relation of the Chromosomes 258 111. Diagram to Show how the Appearance of a Lethal Near Beaded Causes the Stock to Produce only Beaded 259 112. Diagram Showing the Results of Crossing Over in a Stock Contain- ing Both Beaded and Lethal 260 113. Diagram Illustrating How in the Presence of a Dominant Factor, Dichete, and a Lethal in Its Homologous Chromosome at About the Same Level, Together with Another Factor, Peach Colored Eyes, Gives the Result Shown in the Squares 261 114. Diagram Illustrating Crossing Over of Factors in Fig. 113 262 115. Rosettes of the Twin Hybrids of the Evening Primrose 263 116. Diagram Illustrating Balanced Lethals and Twin Hybrids 264 117. Diagram Illustrating Lethals and Four Types 265 "I THE PHYSICAL BASIS OF HEREDITY CHAPTER I INTRODUCTION That the fundamental aspects of heredity should have turned out to be so extraordinarily simple supports us in the hope that nature may, after all, be entirely approach- able. Her much-advertised inscrutability has once more been found to be an illusion due to our ignorance. This is encouraging, for, if the world in which we live were as complicated as some of our friends would have us believe we might well despair that biology could ever become an exact science. Personally I have no sympathy with the statement that *^the problem of the method of evolution is one which the biologist finds it impossible to leave alone, although the longer he works at it, the farther its solution fades into the distance.'' On the contrary, the evidence of recent years and the methods by means of which this evidence is obtained have already in a reasonably short time brought us nearer to a solution of some of the import- ant problems of evolution than seemed possible only a few years ago. That new problems and developments have arisen in the course of the work — as they are bound to do in any progressive science, as they do in chemistry and in physics for example — goes without saying, but only a spirit of obscurantism could pretend that progress of this kind means that we see the solution of our problem fading away into the distance. Mendel left his conclusions in the form of two general laws that may be called the law of segregation and the 16 fROPERTY IIBURY n. C. StaU College 16 PHYSICAL BASIS OF HEEEDITY law of independent assortment of the genes. They rest on numerical data, and are therefore quantitative and can be turned into mathematical form wherever it seems desir- able. But though the statements were exact, they were left without any suggestion as to how the processes involved take place in the living organism. Even a purely mathematical formulation of the principles of segregation and of free assortment would hardly satisfy the botanist and zoologist for long. Inevitably search would be made for the place, the time, and the means by which segre- gation and assortment take place, and attempts would sooner or later be made to correlate these processes with the remarkable and unique changes that take place in the germ-cells. Sutton, in 1902, was the first to point out clearly how the chromosomal mechanism, then known, supplied the necessary mechanism to account for Mendel's two laws. The knowledge to which Sutton appealed, had been accumulating between the years 1865, when MendePs work was published, and 1900, when its importance became generally known. An account of the chromosomal mechanism may be deferred, but I have spoken of it here in order to call attention to a point rarely appreciated, namely, that the acceptance of this mechanism at once leads to the logical conclusion that MendePs discovery of segregation applies not only to hybrids, but also to normal processes that are taking place at all times in all animals and plants, whether hybrids or not. In conse- quence we find that we are dealing with a principle that concerns the actual composition of the material that car- ries one generation over to the next. Segregation and independent assortment were the two fundamental principles of heredity discovered by Mendel. Since 1900, four other principles have been added. These are known as linkage, the linear order of the genes, inter- ference, and the limitation of the linkage groups. In the same sense in which in the physical sciences it is custo- INTRODUCTION 17 mary to call tlie fundamental generalizations of the science the ''laws" of that science, so we may call the foregoing generalizations, the six laws of heredity known to us at present. Despite the fact that the use of this word '4aw" has been much abused in popular biological writing we need not apologize for using it here, because the postu- lates in question have been established by the same scien- tific procedure that chemists and physicists make use of, viz., by deductions from quantitative data. Excepting for the sixth law they can be stated independently of the chro- mosomal mechanism, but on the other hand they are also the necessary outcome of that mechanism. The theory of the constitution of the germ-plasm, to which Mendel's discoveries led him, not only failed to receive any recognition for fifty years, but the principle of particulate inheritance to which it appeals has met with a curious reception even in our own time, leading a recent writer to state that particulate theories in general "do not help us in any way to solve any of the funda- mental problems of biology, ' ' and another writer to affirm that if the chromatin of the sperm is "pictured" as com- posed of individual units that represent "some specific unit-characters of the adult, ' ' then we should expect it to be extremely complex, "more complex indeed than any chromatin in the body, since it is supposed to represent them all," but "as a matter of fact chemical examination shows the chromatin in the fish sperm to be the simplest found anywhere. ' ' Were our knowledge of the chemistry of the "chromatin" as advanced as these very positive statements might lead one to suppose, the objection raised might appear to be serious, but there is no evidence in favor of the statement that the sperm-chromatin should be expected to be more complex than the same chromatin in the cells of the embryo or adult. And even were it different in the germ-tract and soma the criticism would miss its mark, because heredity deals with the constitution of the chromatin of the germ-tract and not with that of 2 18 PHYSICAL BASIS OF HEEEDITY the soma. Until physiological chemists are in position to furnish more complete information concerning the com- position of the chromosomes, or more illuminating criti- cism of the situation as it exists, we need not, I think, be over-much troubled by such views so long as we handle our own data in a manner consonant with the recognized methods of scientific procedure. Other critics object for one reason or another to all attempts to treat the problem of heredity from the stand- point of the factorial hypothesis. It has been said, for instance, that since the postulated genetic factors are not known chemical substances the assumption that they are such bodies is presumptuous, and gives a false analogy with chemical processes. Such critics claim that the pro- cedure is at best only a kind of symbolism. Again, it has been said, that the factorial hypothesis is not a real scientific hypothesis, for it merely restates its facts in terms of factors, and then by juggling with numbers pre- tends that something is being explained. It has been argued that Mendelian phenomena relate to unnatural conditions and that they have nothing to do with, the normal process of heredity in evolution that takes place in ''nature." It has been objected that such a hypoth- esis assumes that genetic factors are fixed and stable in the same sense that molecules are stable, and that no such hard lines are to be found in the organic world. And finally it has been urged that the hypothesis rests on dis- continuous variation which, it is said, does not exist. If the implications in any or in all of these objections were true, the attempt to explain the traditional prob- lem of heredity by the factorial hypothesis would appear fantastic in the extreme. An attempt will be made in the following chapters to present the evidence on which our present views concerning heredity rest, in the hope that an understanding of this evidence will go far towards removing these a priori objections, and will show that they have no real foundation in fact. CHAPTER II MENDEL'S FIRST LAW— SEGREGATION OF THE GENES Mendel succeeded in discovering the principle of segregation because he simplified the conditions of his experiments so that he had to deal with one process at a time. Others before him had failed because they worked with too complex a situation. In each case Mendel picked out for study a pair of contrasted characters of a kind that were sharply distinguishable from each other when- ever they appeared. He chose plants that normally self- fertilize and are little liable to accidental cross-fertiliza- tion, which made it possible easily to obtain in the second generation numbers large enough to give significant results. To MendePs foresight in arranging the condi- tions of his work, as much as to his astuteness in interpret- ing the data, is due his remarkable success. Mendel used varieties of the common edible garden pea (Pisum sativum). Many of these varieties (races) differ from each other in a particular character. Some races are tall, others short; some have green peas (seeds in the pods), others have yellow peas ; some of these seeds have a smooth surface, others are wrinkled; some of the pods are hard, others are soft. One of the crosses made by Mendel will serve as an illustration of his work (Fig. 1). Pollen from a race of tall peas was put artificially on the stigma of a plant of a short race, whose own stamens, and therewith the pollen, had been previously removed. The hybrid plants that came from the seed were tall. These hybrids were allowed to self-fertilize and their seeds collected. Some of the seeds produced tall jolants, 19 20 PHYSICAL BASIS OF HEREDITY Tall @@ Sfiott \s\ \s\ Fig. 1. — Cross between a tall and a short race of garden peas. The F\ generation is tall. In the second generation, Fi, there are three talis to one short. {P\, F\ and F2 were reared from peas supplied by Dr. O. E. White.) MENDEL'S FIRST LAW 21 others produced short plants; in the ratio of 3 tall to 1 short. In other words, the contrasted characters of the grandparents reappeared in the grandchildren in the ratio of 3 to 1. The experiment was carried through one more generation, which was necessary in order to get data for finding out what had been taking place. The short peas were allowed to fertilize themselves. They pro- duced only short peas. The tall peas were also allowed to fertilize themselves. One-third of the tall peas produced only tall offspring; two-thirds produced both tall and short offspring in the ratio of 3 : 1, as had the first genera- tion hybrids. Evidently then the grandchildren had been of three kinds, one kind was pure for shortness, others were hybrids, and the remaining kind was pure for tall- ness. These kinds appeared in the proportion of 1 : 2 : 1. Some factor or factors in the original tall peas must cause the peas of that race to be always tall, and some factor in the original short peas must cause them to be short. The short factor may be represented by s, and the long factor by S. When crossed, the fertilized egg should contain both factors (sS), and since the hybrids coming from this egg were tall, it is evident that tall must dominate over short. Now if the two factors (sS) present in the hybrid should separate {i.e., "segregate'') when its ovules and its pollen-grains are formed, half of the eggs would contain the factor that represents the short peas (5), and half of the eggs the factor that represents tall peas (S) ; also half of the pollen grains would contain the factor that represents the short peas (5), and half of them would contain the factor that represents the tall peas (S). Chance meeting between egg-cells and pollen-cells (one ovule being always fertilized by one pollen grain), would, on the average, give one fertilized egg containing two factors for short (55) ; to two fertilized eggs that contain one of each kind of factor (sS) ; to one that contains two 22 PHYSICAL BASIS OF HEREDITY factors for tall (SS). The chance combination just given may be represented graphically as follows: Tall ,^ ^ Short Tall / \. Short (Tall-Short. Tall-Tall. -f^ +Sliort-Short. ^ riall-Short. In the actual experiment that Mendel carried out, plants of the tall race measured from 6 to 7 feet, and those of the short plants three-quarters to one foot and a half. The F^ plants were as tall as, or even taller than the tall parent. When these 7^"'/s were self-fertilized, the seeds (either from the same plant or from a random collection of seeds from different F^ plants) produced 787 long plants and 277 short plants — a ratio of 2.84 to 1. As a fair sample of each plant, ten seeds were taken from each of 100 tall plants of tliis second (or F2) genera- tion. Out of the 100 plants so tested, 28 plants produced only tall plants, while 72 of them produced some tall and some short offspring. This means that 28 plants were pure (homozygous) tall, whilst 72 were hybrid like the F^ plants. Taking, then, all F2 plants together, the results show % were short, V4 were hybnd, and % were tall, i.e., they stand in a ratio of 1 : 2 : 1. This relation is illustrated in the scheme below, based on what 16 F2 plants might give. Twelve would be tall to 4 short. If the tall plants are tested, they are found to consist of 4 pure tails {.SS) and 8 hybrid tails {sS). Altogether, then, there are 4 tails to 8 hybrid tails to 4 short, i.e., there are three kinds of F2 peas in the ratio of 1:2:1. 12 tall -f 4 short 4SS 4- 8sS + 4ss 1 2 1 The process of disjunction, or separation of the mem- bers of a pair of factors, is known technically as segre- gation. While we sometimes also speak of the segrega- MENDEL ^S FIRST LAW 23 tion of the characters themselves, it seems better, I think, to avoid as far as possible this application of the word. The factor for tall and the factor for short are said to be allelomorphic to each other. The parents are generally designated by P^ ; the first hybrid generation is known as the first filial generation, or briefly F^. The next generation, derived from F^ is called Fg, etc. When one member of the pair of contrasted characters appears in F^ to the exclusion of the other it is said to be dominant, the eclipsed character is said to be recessive. The hybrid itself is said to be heterozygous, meaning that it contains one factor or gene of each kind, while an individual con- taining both genes of the same sort is said to be homo- zygous for the genes involved. Mendel did not emphasize the idea that even in pure races each character is also represented, as a rule, by a pair of factors or genes that segregate in the formation of the germ-cells in the same way as do the pair of contrasted genes in the hetero- zygotes, but at the present time this idea is accepted by all geneticists. It was at least implied on Mendel's view that the two pure classes in F^ {SS and ss), formed by the recombination of two like genes, are identical with the two grandparental races (Pi). A crucial test of the correctness of the assumption that segregation of the members of a pair of elements takes place in the germ-cells of the hybrid, consists in back- crossing the hybrid (F^) to one of the parent stock,' viz., to the not dominant stock, here the short pea. Since short is recessive to tall, it will not influence the height of the offspring when a tall and a short factor are brought together. Such a cross should show whether the germ- cells of the hybrid are, as postulated, of two sorts, and whether equal numbers of each sort are produced. Mendel made such tests, and obtained equal numbers of two kinds of offspring. Mendel obtained results like these with tall versus short peas for other pairs of characters, such as f asciated versus normal stems, hard versus soft pod, yellow versus 24 PHYSICAL BASIS OF HEREDITY green pods, gray versus white -skinned peas, yellow versus green cotyledons (seen through the skin of the seed), and round versus wrinkled seeds (determined by the nature of the cotyledons within the seed coat). The 3 : 1, Fo, ratio characteristic for a single pair of characters is the expectation based on the chance meeting of either one of two kinds of eggs with either one of two kinds of pollen grains. In actual numbers this ratio is, of course, not always exactly realized, but only approxi- mately. For the seven pairs of characters that Men- del examined, the F2 ratios were as follows : Dominants Receesivee No's, per 4 Form of seed 7,324 8,023 929 1,181 580 858 1,064 5,474 6,022 705 882 428- 651 787 1,850 2,001 224 299 152 207 277 2.99 3.00 3.04 2.99 2.95 : 3.03 2.92 : 1.01 Color of cotyledons Color of seed coats : 1.00 0.96 Form of pod 1.01 Color of p>od 1.05 Position of flowers 0.97 Lenjrth of stem 1 ns Totals 19,959 14,949 5,010 2.996 : 1 .004 The following collective data for the inheritance of color of the cotyledons of garden peas show that the approximation to a 3 to 1 for the recessive character is very close ; Yellow Green Total No's, per 4 Probable errors Mendel Correns 6,022 1,394 3,580 1,310 11,903 1,438 109,060 1,089 1,647 1,012 3,000 3,082 222 2,405 50 2,001 453 1,190 445 3,903 514 36,186 354 543 344 959 1,008 1,856 70 850 8,023 1,847 4,770 1,755 15,806 1,952 145,246 1,443 2,190 1,356 3,059 4,090 7,518 295 3,250 3.002 : 0.998 3.019 : 0.981 3.002 : 0.998 2.986 : 1.014 3.012 : 0.988 2.947 : 1.053 3.004 : 0.996 3.019 : 0.981 3.008 : 0.992 2.985 : 1.015 3.0;U : 0.969 3.014 : 0.986 3.013 : 0.987 3.051 : 0.949 2.954 : 1.046 ±0.0130 ±0.0272 Tschermak Hurst ±0.0169 ±0.0279 Bateson ±0.0093 Lock Darbishire Darbyshire White ±0.0264 ±0.0030 ±0.0308 ±0.0250 Correns ±0.0319 Tschermak Lock ±0.0186 ±0.0183 Darbishire Correns Lock ±0.0135 ±0.2151 ±0.0205 Totals 218,425 50,676 203,500 3.004 : 0.996 ±0.0026 M r State CoUeJF PARENTS Fig. 2. — Cross between white and red flowered four-o'clor-ks {Mirahilis jalapa). In the lower part of the diagram the large circles represent somatic conditions, the included small circles the genes that are involved. MENDEL'S FIRST LAW 25 That MendePs principles apply to animals was first made out by Bateson and by Cuenot in 1902. Since then many characters both in domesticated and in wild animals and plants have been studied, and there can be no question of the wide application of MendePs discovery. During the years immediately following the re-dis- coveiy of MendePs principles (1900) much attention was paid to the phenomena of dominance and recessiveness. This was due, no doubt, to the striking fact that the hybrid sometimes resembles only one parent in some particular trait, whereas the older observations, where many charac- ters were generally involved in the cross, seemed to have shown that hybrids are intermediate in regard to their parents. We now know, however, that although there are cases in which the dominance is as complete as in those described by Mendel, yet in a very large number of forms the hybrid is intermediate between the parents, even when only a single pair of characters is involved. A few examples will serve to illustrate these relations. The common garden four o 'clock, Mirahilis jalapa, has a white-flowered and a red-flowered variety (Fig. 2). When crossed, the hybrid has a pink flower, which may be said to be intermediate in color between white and red. Here neither color can strictly be said to dominate. When the hybrid (F^) is self-fertilized the offspring {F2) are in the proportion of one white, to two pink, to one red- flowered plant. The Fn reds and the F2 whites breed true ; the pinks when self-fertilized give white, pink and red in the proportion of 1 : 2 : 1. In a case of this kind the color of the F2 plants reveals the nature of the three classes present, so that it is not necessary to test them out, as was the case in the F2 generations of MendePs peas, where the Fo tails were found in this Avay to be of two sorts. The F2 results w^ith the four o'clock also show that the segregation of the genes is clean, for the F2 whites never produce in subsequent generations anything 26 PHYSICAL BASIS OF HEREDITY but white descendants, and the F2 reds never anything but red descendants. In this case the color of the F^ flowers is obviously somewhere between red and white. In so far as the F^ flower is colored, it may be said that red is dominant ; in which case the red and the pink Fn classes (1 + 2 = 3) are to be counted together as contrasted wath the white, giving a 3 : 1 ratio. On tlie other hand, if one chose to emphasize the fact that the F^ pink flower is not red, but affected by the white-producing element in its make-up, then not red, but white, might be said to be the dominating character; in which case the white and the pink F2 classes (1 + 2 = 3) would be counted together as contrasted with the red giving an inverse 3 : 1 ratio. It appears then largely a matter of choice as to what is to be called dominance (see below). The essential fact of segrega- tion is not affected by the decision, and it is this that is fundamentally important. Another example of failure of complete dominance is shown in the race of Andalusian fowls. In this race there are blue, splashed-white, and black birds ; the blue birds going under the name of Andalusians. When splashed- white is mated to black, all the offspring {F^) are blue (Fig. 3) ; when these blues are bred together they give 1 splashed-white : 2 blues : 1 black. Evidently the blue birds are the heterozygous type. Their feathers show under the microscope less black pigment, somewhat dif- ferently distributed from that in the black birds. The intermediate blue color is due in this case to the less dense distribution of the pigment in the heterozygote. Lippin- cott, who has recently examined this cross in greater detail than heretofore, states that the colored areas or splashes in the white males are either blue or blackish according to the part of the body on which they occur, and that this corresponds with the distribution of the color on the Anda- lusian, for while the latter is said to be blue, this applies i' ■ , ^ P, Fig. 3. — Cross between splashed-white and black, giving in Fi Andalusian, and in Fj one splashed-white, two Andalusian, and one black. MENDEL ^S FIRST LAW 27 strictly only to the hen and to the lower parts of the body in the cock whose upper surface is very dark blue or even black. In this case neither black nor white can be said to be dominant. The blue brought in as splashes by the splashed-white might indeed be regarded as dominant over the black of the other (black) parent, but if so, then the uniform distribution of the blue must be determined by dominance of the allelomorphic gene brought in by the black parent. Each parent then would contribute at the same time a dominant and a recessive effect, each the product of one member of the same pair of allelomorphs. There are other cases in which the hybrid is inter- mediate in color, and, in addition, its range of variation is so large that the extremes overlap one or even both of the two parental types. For example : In the vinegar fly, Drosophila melanogaster, there is a race with ebony wings and another race with sooty wings. Wlien such flies are crossed to each other, the wings of the F^ fly are intermediate in color, ranging from wings like those of sooty to wings as black as ebony. When the F^ flies are in- bred they give rise to a series that at one extreme has gray wings and at the other black wings. Separation into three classes is diflicult or impossible. Here it may appear that the two original characters have completely blended in F^ and in F^, but that there are in reality three classes of flies in Fo can be demonstrated by suitable tests. If, for instance, we pick out a sufficient number of i^\ males to give a fair sample of the population, and mate each male first to an ebony female of pure stock, and then to a female of sooty stock, we shall find that one-quarter of the males mated to ebqny give only ebony, one-quarter mated to sooty give only sooty, while the remaining two- quarters give, both in the back-cross to sooty, and in that to ebony, a wider ranging group, which is darker on the whole when mated to ebony, and lighter when mated to 28 PPIYSICAL BASIS OF HEREDITY sooty. These and other tests show that in the F^ hybrid segregation of tlie same kind as in the preceding cases has taken place, but the results are obscured by the wide variability of the hybrid flies. In other words, evidence can be obtained that the segregation of the genes has been clean cut, even although this is obscured by the character of the heterozygous flies. Fia. 4. — Male and female vinegar fly (Drosophila melanoyaster). In the iDreceding illustrations the character difference between the two races is supposed to show itself in the same environment. It has been found in a few other cases that the dominance of one character over the other may depend on the environment. For example, in the normal vinegar fly the black bands of the abdomen show great regularity (Fig. 4), but in a mutant race called "abnormal abdomen'^ (Fig. 5) the bands may be irregu- larly broken up, or even absent. In cultures with abund- ance of fresh food and moisture, all the individuals have very irregular bands, but as the culture gets old, and the MENDEL'S FIRST LAW 29 food and moisture become less and less, the bands become more and more regular until at last the flies are indistin- guishable from normal flies. If a cross is made between a female with abnormal bands and a wild male, the off- spring that first hatch under favorable conditions are all very abnormal. Here abnormal completely dominates normal bands. But as the culture dries up, the hybrid offspring become more and more normal, until finally they are all normal. At this time it might be said that normal dominates abnormal. Both statements are correct, if we add that in one environment abnormal banding dominates, Fig. 5. — Normal and abnormal abdomen of D. melanogaster. in another environment normal banding dominates. The genetic behavior of the pairs of genes is the same here as in all other cases of Mendelian behavior, but this is revealed onlv when the environment is one in which the abnormal gene produces one effect, the normal a different one. That the gene is not itself affected by the environ- ment can be shown very simply. If a female from the abnormal stock be picked out, at a time when the stock has onlv normal bands, and crossed to a wild male, the offspring will all be as ' ' abnormal ' ' as when the mother herself is abnormal, provided the food and moisture conditions are of the right kind. The late hatched normal flies of abnormal stock may be bred from for several 30 PHYSICAL BASIS OF HEEEDITY generations, but as soon as a "generation hatches under favorable conditions they are as abnormal as though all their ancestors had been of this sort. Thus it is evident that no fundamental importance is to be attached to domi- nance of characters. On the other hand, it is equally obvious that it would be entirely unwarranted to suppose that incompleteness of dominance is due to failure of segregation of the genes that stand for the characters. While the problem of segregation can be studied to greatest advantage where the characters of a pair are sharply separated, yet even where the pair does not possess this advantage, the cleanness of the segrega- tion process can be just as definitely, though more laboriously, demonstrated. In cases where there is an overlap between the hetero- zygous type and one of the parental types it may, simply as a matter of convenience, be advantageous to call that character that gives the more continuous Fo group the dominant, thus leaving the smaller more sharply defined group as the recessive. For example, the F2 group from black by wild-type Drosophila may be represented by such a scheme (Fig. 6) as the following: Fig. 6. — Relation of black body color to wild type as shown by the claBses of Fi flics. The heavy outline includes the mutant class, the lighter line the wild type, and the dotted line the heterozygous class. Here the heterozygous flies are typically intermediates, but their variability overlaps that of the wild type to such an extent that separation of the intermediate from the wild type is practically impossible. On the other hand, there is no difficulty in making a complete separation between the heterozygous class and the homozygous black. MENDEL'S FIRST LAW 31 Black is accordingly treated as a recessive in nearly all experiments. a at Fig. 7. — Normal eye, o, a', heterozj'gous eye h, V , and bar eye c, c', of the vinegar fly. A mutant eye shape of Drosophila, called ^^bar'' (Fig. 7, a) , has an intermediate hybrid type (Fig. 7,h), The F2 group may be represented (Fig. 8) in the following scheme: v._ Fig. S. — Relation of bar eye to normal eye, as shown by the Ft classes. In this case the hybrid, intermediate type, overlaps the bar type, so that in Fc these two latter types give a nearly continuous class. At the other end of the Fn series, the round eyed normal (or wild) type can be distinguished without difficulty from either of the other classes. Bar is therefore normally treated as a dominant. 32 PHYSICAL BASIS OF HEREDITY The case of Mirahilis, or of the Andalusian fowl, might be represented (Fig. 9) in the following scheme: \ \ FiQ. 9. — Relation of Andalusian to splashed white and to black as shown by classes of f 2 birds. Here all three types are fully separable, in which case either homozygote might be considered the dominant. Finally, to return to the case of the tall and short peas, the following scheme (Fig. 10) represents the Fg Fig, 10. — Relation of tall to short peas as shown by F2 classes. group: Here the tall and the heterozygous group are alike, and inseparable by ordinary inspection, even at the extreme end of their variation curves, and short is ' ^ completely * * recessive. In cases in which the environment enters more obviously into the result (as in ' ' abnormal abdomen, ' ' Fig. 5), the following scheme (Fig. 11) represents the relation : Dry IVef Fia. 11. — Relation of normal to abnormal abdomen as shown by classes of F2 flies. "Dry" signifies conditions that make for normal; wet for abnormal. In this case both the heterozygous and the parental ^^abnonnaP^ type may show *' normal' ' abdomen like the MENDEL'S FIEST LAW 33 wild type. The abnormal type is treated as the dominant although only when the conditions are favorable to its appearance is the hereditary phenomenon seen. In another case (duplicate legs) only the homozygous form may show the duplications (in a special environment). The following scheme (Fig. 12) represents this relation, reduplication of legs being treated as a recessive: Fio. 12. — Relation of normal to duplicate legs. There are still other relations that affect the dominance of characters. For example, there may be internal fac- tors, which when present, determine that a character shall be dominant over its allelomorph, or recessive to it. In this connection might be mentioned what has been called '^ reversal of dominance.'' An example from Davenport will illustrate what is meant. In a certain strain of fowls there is a tendency for the toes to be united by a web at the base. Crossed to birds with normal feet, no birds with united toes (syndactyls) appeared in F^. The F^ birds inbred gave in Fo only about 10 per cent, of s}Tidactyl birds. It would appear that the latter character is reces- sive, and that the recessive type overlaps largely the dominant heterozygous type. Davenport interpreted, however, the syndactyl as the dominant type, because ^Hwo syndactyls may give nor- mals, but no true normals give syndactyls." In other words, he defines the dominant type as the one that can carry the other type, because he says dominance is due to presence of factors, recessiveness to absence. *^Now dominance may fail to develop but recessiveness never can do so.*' For this reason two syndactyls may give 3 34 PHYSICAL BASIS OF HEREDITY normals, because a dominant cliaracter may fail to develop, even though its factors be present. Since normal feet never give syndactyls, the normal type must be recessive. But Davenport's definition of a recessive type as one that never shows in the heterozygous condition is in my opinion based on an arbitrary distinction of what is the cause of dominance and recessiveness. The evidence may, I think, be better interpreted as indicated in the same diagram as that for abnormal abdomen (Fig. 11) in that part marked "dry," in which the syndactyl condition would be represented as recessive (heavy line). In the hybrid the character is usually seen only in a few individ- uals, i.e., it is intermediate, overlapping both parent types. While this case shows that it is often only a convention as to which type is called the dominant and which the recessive, I can see no special reason why in these cases of syndactylism the usual convention may not be followed .which recognizes the small F2 class as the recessive. Mendelism rests on the theory of a clean separation of the members of each pair of factors (genes). In every heterozygote the factor for the dominant and that for the recessive are supposed to come into relation to each other and then to separate at the ripening of the germ-cells. If we think of the two genes coming together and afterwards separating, it w^ould seem that a favor- able situation might exist for the two to become mixed, and one "contaminate" the other. If any extensive process of this kind occurred the Mendelian phenomena would be so irregular and erratic that they would have little interest. But even those who are inclined to appeal to contamination as an exceptional phenomenon, grant that clean separation of the genes is the rule. The best critical evidence against contamination is in cases in which for many successive generations breeding has taken place from heterozygous forms only (which creates a favorable situation for contamination to take place w^ere it possible). No influence of contamination has been found in such cases. MENDEL'S FIRST LAW 35 Marshall and Muller kept flies heterozygous for three re- cessive mutant factor for about seventy-five generations, and at the end of that time found that these factors had not been weakened in any way as a result of juxtaposition Fig. 13. — Notch wings in the vinegar fly, extreme condition, a; average condition, 6; nearly normal condition, c. with their normal dominant allelomorphs. I have kept a stock of notch-winged flies under selection for twenty- five generations. Notch (Fig. 13) is a character varying in the direction of normal wings (Fig. 13, c) ; in eveiy generation of notch, many notch flies have normal wings. The character is dominant, and exists only in heterozy- 36 PHYSICAL BASIS OF HEREDITY gous condition, since a fly homozygous for notch dies. The race is therefore necessarily maintained in a hetero- zygous state. In each generation females that were genetically notch, but had normal wings, were selected and bred to normal males. Tlie selection was away from notch (i.e., toward normal). After a tune more than half of the notch flies had nonnal wings. The effect produced proved to be due not to a change in the notch gene through contamination, but to modifying genes; for at the end of the selection tlie original notch could be recovered at any time by removing the influence of the modifying factor. It has been sometimes stated, usually by the opponents of Mendel's theory, or by advocates of doctrines of evolu- tion that appeared to be compromised by the Mendelian conception of *^unit factors,'* that Mendelism deals only with such superficial characters as the color of flowers or the hair color of mammals. This statement contains an element of truth in so far as it covers most of the kinds of characters that students of heredity find most convenient to study; but it contains an entirely false inference as to the limitations of Mendelism. The issue involved is this : changes in superficial characters are not so likely to affect the ability of the organism to survive as are changes in essential organs ; hence they are the best kind of hereditary characters for study. But there is no evidence that such superficial characters are inherited in a different way from "fundamental" characters, and there is evidence to the contrary. A common class of characters showing perfect Mendelian behavior are so-called lethals that destroy the individual when in homo- zygous condition. There can be no question as to the fundamental importance of such factors. Between these extreme cases and the superficial shades of eye color, for example, all possible gradations of structure, physio- logical and pathological, are known. The only possible question that might be seriously raised is whether these characters are all losses or deficiencies, while progres- MENDEL'S FIRST LAW 37 sive advances may belong to a different category. This may be a serious question for the evolutionist, but has nothing to do with the problem that concerns us here. In recent years an entirely unexpected and important discovery in regard to segregating pairs of genes (allelo- morphs) has been made. In an ever-increasing number of cases it has been found that there may be more than two distinct characters that act as allelomorphs to each other. For example, in mice, yellow, sable, black, wliite- bellied gray, and gray-bellied gray (wild type) are allelo- morphs, i.e., any two may be present (as a pair) in an individual, but never more than two. In DrosopJiila the eye colors white, eosin, cherry, blood, tinged, buff, milk, ivory, coral and the normal allelomorph form a series of multiple allelomorphs. In the grouse locust, Paratettix, there are nine types that may be allelomorphic, all of which exist in the wild state (Nabours). In Drosophila, again, there are as many as twelve other series of allelo- morphs known at present ; in rats there is a small allelo- morphic series, also two in guinea pigs and two in rabbits. In plants there are a few cases known, especially in corn. In all these series it is the same organ that is mainly affected by the different allelomorphs, which seems ''natu- ral, ' ' but was not necessarily to have been expected. The chief interest of these series is that they appear to demon- strate that the normal (wild type) allelomorph, and its mutant mates need not be due to presence and absence, but rather represent modifications of the same unit in the hereditary material ; for, taken literally, only one absence is thinkable, and yet in Drosophila there are eight such "absences'' in one series. As has been stated, Mendel did not make it clear that there exists in the normal animal or plant the same dual- ity that comes to light when a hybrid is produced ; never- theless this condition is implied, at least, in his paper, and has been taken for granted in practically all of the modern w^ork on heredity. The demonstration that such 38 PHYSICAL BASIS OF HEEEDITY is the case is, however, not a simple matter. It could not have been made by Mendel or in the earlier days after the rediscovery of Mendelism (1900). An attempt to furnish this demonstration is given in Chapter XX. Assuming the demonstration to be satisfactory, we reach the highly important conclusion that segregation is not something m peculiar to hybrids, but something most readily demon- strated by means of hybrids, and that in all probability the germ-plasm is at first made up of pairs of elements, but at the ripening of the germ-cells these elements (genes) separate, one member of each pair going to one daughter cell, the other member to the other cell. The mechan- ism by means of which such a process might take place had been known for several years before its relation to Mendel's principles of segregation was realized. This mechanism is to be found in the conjugation and reduc- tion processes that take place in the maturation of egg- and sperm-cell. An account of this process is given in the next chapter. CHAPTER III THE MECHANISM OF SEGREGATION One of tlie most secure generalizations of modern work on the cell is that every cell of the individual contains a constant niimher of self-perpetuating bodies (called chro- mosomes), half of which are traceable to the father and half to the mother of the individual. No matter how specialized cells may be, they contain the same number of chromosomes. Equally important is the fact that after the eggs of the female and the sperm-cells of the male have passed through the ripening or maturation divisions the number of chromosomes is reduced to half.^ Lastly, there is convincing evidence that the reduced number of chromosomes is brought about as the result of a separa- tion of such a kind that each mature germ-cell gets only a paternal or a maternal member of each chromosome pair. The reduction takes place in the female at the time when the polar bodies are given off from the Qgg ; and in the male just prior to the formation of the spermatozoa. A characteristic process is seen in the oogenesis and sper- matogenesis of the nematode worm Ancyr acanthus cysti- dicola (a parasite in the swim-bladder of fresh-water fishes) described by Mulsow. The young eggs contain twelve chromosomes (Fig. 14, a). As the result of the later union of these twelve in pairs, six short threads appear in the nucleus of the Qg^ just before it extrudes its polar bodies. The threads contract to six short rods (split in two planes at right angles to each other), the tetrads (Fig. 14, c). With the dissolution of the nuclear wall these tetrads are set free in the protoplasm, and a spindle develops about them (Fig. 15, a). They pass to the equator of the spindle, and there dividing lengthwise, * Exceptions occur in certain cases of parthenogenesis. 39 40 PHYSICAL BASIS OF HEREDITY half of each goes to one pole, and half to the other pole of the spindle (Fig. 15, h). One end of the spindle pro- trudes from the egg, and around it the protoplasm con- a ■fr Fig. 14. — 06cyte of Ancyr acanthus, a; growth period, b; nucleus with tetrads, c. (After Mulsow.) stricts off (Fig. 15, c) to form the first polar body. About the six ovoidal chromosomes left in the egg a new spindle develops; and these chromosomes become drawn into its equator, where they divide again, half of each going a c e Fig. 15. — Egg of Ancyr acanthus with six tetrads, a; egg with first polar spindle, h; egg after extrusion of first polar body, c; egg with second polar spindle, d; egg after the extrusion of both polar bodies, e. to one pole and half to the other (Fig. 15, d). A sec- ond protrusion takes place from the surface of the egg which pinches off to form the second polar body (Fig. 15, e). Thus, after two mitotic divisions, the egg has lost three-quarters of its chromatin, but retains half the full MECHANISM IN SEGREaATION 41 nmnber of chromosomes, and as a result, tlie original twelve chromosomes have been reduced to six. Around the six chromosomes left in the egg, a nuclear wall forms, and the chromosomes become spun out into delicate fibres. Meanwhile a spermatozoon has entered the eggj and out of its head another nucleus develops. The two nuclei, the egg nucleus and the sperm nucleus, move toward the center of the egg (Fig. 16, a), where they come into contact with each other. After a time, the chromatin threads begin to condense again into rods. Fig 16. — Eggs of Ancyracanthus within membrane. Egg with two pronuclei, a; egg pronucleus with six chromosomes and sperm nucleus with six chromosomes, h; egg pro- nucleus with six chromosomes and sperm nucleus with five chromosomes, c; union of male and female pronuclei, d. (After Mulsow.) Six appear in the egg nucleus, and six in the male nucleus (Fig. 16, ty. A spindle develops in the protoplasm of the egg around the twelve chromosom.es of which six have come from the father (the paternal chromosomes) and six from the mother (the maternal chromosomes) (Fig. 16, d). Each chromosome now splits lengthwise into equivalent halves, and a half moves to each pole of the mitotic spindle. The spindle rotates in the cytoplasm of this egg until its long axis corresponds with that of the egg. As the daughter chromosomes move towards the poles of the mitotic spindle the egg protoplasm constricts * Assuming a female producing sperm to have entered. 42 PHYSICAL BASIS OF HEREDITY between them so that two cells are formed, each cell con- taining twelve chromosomes, six paternal and six mater- nal. Thus, through fertilization, the whole number of chromosomes is restored to the egg. This number remains through all subsequent divisions of the cells of the embryo. The male of Ancyracantlius has only eleven (Fig. 17, a) chromosomes ; because the male has only one sex-chromo- a Fig. 17. — Spermatogenesis of Ancyracanthus. Spermatogonia! cell, o; cell after growth period with tetrads, h; first spermatocyte division, c; two cells resulting from first division with six and with five chromosomes, respectively, d; four cells resulting from the next division, c; ditto, /; mature spermatozoa, one with six, the other with five, chromoBomes.g; ditto, living spermatozoa, h. (After Mulsow.) some, while the female has two sex-chromosomes. Both sexes have ten other chromosomes, sometimes called auto- somes. Just before the maturation divisions take place, there are six rods in each sperm-cell, five of which (the autosomes) condense into tetrads, the sixth (the sex- chromosome) into only a double body (Fig. 17, &). A spindle develops about these and each of the five auto- somes divides. The sex-chromosome does not divide, but passes to one pole of the spindle (Fig. 17, c). The result MECHANISM IN SEGREGATION 43 is that two cells are produced, one with six, the other with five chromosomes (Fig. 17, d). Without a resting stage a new spindle develops in each cell, and a new division takes place — each dumb-bell- shaped body dividing, as well as the sex chromosome in the cell that contains it. In all, four cells result (Fig. 17, e and /) — two with five chromosomes each, two with six each. Each becomes a spermatozoon, which in this worm is a round cell with the chromosomes at one pole (Fig. 17, g). Half of the spermatozoa contain six, half five chromosomes. They can be distinguished even in the living sperms (Fig. 17, h). If a six-chromosome sperm fertilizes an egg (Fig. 16, b), a female (with 12 chromosomes) is produced — if a five-chromosome sperm fertilizes an egg (Fig. 16, c), a male (with 11 chromosomes) is produced. The two chromosome divisions (or separations) that take place when the polar bodies are extruded from the egg are, for a number of reasons that need not be entered into here, generally regarded as equivalent to the two final divisions in the ripening of the sperm-cells. One of the two divisions is interpreted as an ordinary cell- division in which the chromosomes split lengthwise into equivalent halves — half going to each pole. The other division is interpreted as a separation of whole chromo- somes that have come together side by side at an earlier stage. The tetrad is, then, looked upon as a pair of chromosomes that have conjugated in the sense that they have come to lie side by side (with interchange of mate- rials at times in a way to be described later). One split is supposed to correspond to the line between the conju- gated pairs ; the other split represents a division in each chromosome of the pair. As a consequence when the chromosomes move apart (at the maturation division) one of the two divisions is said to be a *^ reducing division** because whole chromosomes are supposed to separate ; the other division is said to be an ^^ equation di\dsion,** each 44 PHYSICAL BASIS OF HEREDITY chromosome splitting lengthwise into equivalent halves as in ordinary cell-division. The interpretation of these two divisions that occur in the egg and in the sperm-cell has been the subject of much speculation. It is apparent that the process reduces the number of chromosomes by half, and that the whole number is regained by fertilization. It is sometimes said that the "purpose" of this division is to keep the number of chromosomes constant, for, if not reduced, they would increase in number with each fertilization. The '^reason" for the other, the second, division is acknowledged to be obscure. For present purposes it is futile to speculate concerning these two divisions, but it should be pointed out here that the genetic evidence is in full accord with the interpretation of these two divisions that is generally accepted to-day by cytologists, i.e., that one of the divisions separates the conjugating pair, and that the other represents a longitudinal division within a paternal and within a maternal chromosome of each pair. If w^e follow the history of the germ-cells further back before the maturation divisions, we find that between the stage when the half number of chromosomes reappears (tetrads) and the stage at which the full number was present, there is a very obscure period in the history of the germ-cells. This period has been studied chiefly in the male. Only a few types have been found favorable for the study of this period. One of the most favorable ones is a marine annelid, Tomopteris, studied by the Schreiners. The early division of the germ-cells (the spermatogonia) of Tomopteris, when the full number of chromosomes is present, is show^n in Fig. 18, a-g. The division is like that of all the other cells of the body. The chromosomes appear as thick bent threads that split lengthwise (Fig. 18,a, b). The nuclear wall disappears and a spindle appears near the group of split chromosomes (Fig. 18, c). As the poles of the spindle move apart the chromosomes become arranged at the equator of the spin- MECHANISM IN SEGEEGATION 45 die, each half of each chromosome becoming attached by a spindle fibre to one pole (Fig. IS^ d). The halves move Fig. 18. Last spermatogonia! division of Tomopteris, a-h; stages before and during synapsis, i-l. (After Schreiner.) apart towards their respective poles (Fig. 18, e) and as they become separated into two groups the cell protoplasm 46 PHYSICAL BASIS OF HEREDITY constricts between them to produce new cells (Fig. 18,/). When the chromosomes have reached the pole they shorten (Fig. 18, g) and appear to send out anastomosing threads. Around this group of threads a new nuclear wall is formed (Fig. 18, /i). All trace of the separate chromosomes is now lost, but between the last stage just described and the stage now to be described it is supposed that important changes in the chromosomes take place. This new phase is spoken of as the synapsis stage. At the beginning of this stage (Fig. 18, i and j) faint indications of the chromo- some appear, and soon they can be seen again (Fig. 18, A:) as long thin threads whose free ends place themselves in parallel pairs. The pairing of the threads continues to extend inwards from the ends (Fig. 18, 1) until they have united throughout the length of the loops (Fig. 19,(2-). There are exactly half as many of these loops as there were original chromosomes, which is expected if they have united in pairs. The conjugation has been accomplished. During the stages that follow, the double chromosomes shorten and become thicker (Fig. 19,h,c,d), and con- dense into the form of tetrads (Fig. 19, e). They begin to separate into halves, each half is also split lengthwise. A spindle appears, and the cells divide (Fig. 19, f,g,h). In each cell the chromosomes show indications of passing into a resting stage, as happens after all ordinary cell divisions, but before this change has gone very far a new spindle appears (Fig. 19, ^), and preparations for another division are rapidly made. The new division completes the maturation of the sperm-cells (Fig. 19, j, k, I). Each of the four cells resulting from the original sperm-mother- cell differentiates into a spermatozoon. In one of the salamanders, Batrocoseps, the matura- tion stages of the male are particularly well shown. The essential stages in synapsis are shown in Fig. 20, a-d as worked out by Janssens. These stages are essentially the same as those of Tom,opteris. During the early multi- plication stages the cells of the future testes divide by i MECHANISM IN SEOEEGATION 47 the ordinary mitotic process. The cells then pass into the synaptic stage (Fig. 20, a-d). As the chromosomes begin a g Fig. 19. — Thin-thread stage of Tomopferis spermatocyte, a-d; tetrads, e; first sperma- tocyte division, /-/; second spermatocyte division,./-/. (After Schreiner). to emerge as thin threads, it is found in Batracoseps that their ends are all pointed towards one pole (Fig. 20, d)- This is the same pole as that towards which the two ends 48 PHYSICAL BASIS OF HEREDITY of each V-shaped chromosome pointed as the cell went into the resting stage. It appears then that the chromo- somes not only retain tlieir original orientation, but that the ends of homologous chromosomes have already come a r:: <^ti *•■ - ."»>><■.. ^ ^^^V^ y .'1?, ^^ ^^i^ • • -'^-'•r #• .■•-•-K ■*^^ 5^ • -r^ :%■ y^ • • ^^^ / Fig. 21. — Synaptic stages, and those immediately following, in the egg of Pristiurus. (After Mar6chal.) MECHANISM IN SEGREGATION 51 quence for genetics in so far as the chromosomes represent the bearers of genes, for while side-to-side union offers an opportunity for interchange between the paternal and maternal members of a pair, no such interchange could be postulated if end-to-end conjugation took place. So far as segregation is concerned either method supplies all that is called for.^ A discussion of other matters will be left until later. Individuality of the Chromosomes During the period of cell-division there can scarcely be any question concerning the persistence of the individ- ual chromosomes, because they remain visibly distinct elements in the cell; but when the nucleus re-forms after each division the chromosomes spin out threads laterally, and these appear to fuse, making a continuous network throughout the nucleus. Whether there is actual fusion between these threads or whether they occupy delimited contact areas, and whether the branches represent the essential part of the chromosome concerned in heredity, are questions impossible to answer at present. The genetic evidence at least consistently shows that no real fusion of the hereditary material occurs even in cells that have passed through many such resting periods. From several other sources there are strong indica- tions that the chromosomes retain their individuality dur- ing the resting stage. In Ascaris, where the chromosomes are few and long, they are often drawn out in an irregular way in the cleavage cells as they pass to the poles of the spindle of the dividing cells. Daughter halves of the same chromosomes show the same identical irregularity. Boveri has shown by an examination of a large number of daughter cells (pairs) that are getting ready for the next division, that when the chromosomes of sister cells reap- ® If the pairs fused end to end and the tetrad arose by two loncritudinal divisions, the outcome would not be in harmony with the theory of segrega- tion based on separation of maternal and paternal chromosomes at reduction. 52 PHYSICAL BASIS OF HEEEDITY pear they show the identical irregularities (Fig. 22, a to d) . It is probable, therefore, that each chromosome has retained the particular form that it had when it passed into the resting stage ; or at least that the axial thread from which the network was spun out has remained in place. In a few cases the chromosomes appear more or less visible during the resting stages. This, however, is such a rare event that it is doubtful whether it can be appealed to in support of the view that in other cases the chromo- somes remain intact. ^& a b c FiQ. 22. — Sister blastomeres of Ascaris preparatory to another division, showing similar arrangements of chromosomes. (After Boveri.) The most convincing evidence comes from exceptional cases of accidental or irregular distribution of one or more chromosomes, so that an ^^^^ or a cell comes to have one more chromosome than is usually present. In the thread-worm Ascaris there are two varieties — one that has four chromosomes in the embryonic cells (with two as the reduced number) and another variety with two chromosomes (with one as the reduced number). A few females have been found in which the unfertilized eggs contain one of these numbers, and all of the spermatozoa that have been received from another individual the other number. In such cases the fertilized eggs, and MECHANISM IN SEGREaATION 53 all embryonic cells, have three chromosomes each (Fig. 64), showing that when an egg starts with three chromo- somes, this number is retained through all subsequent divisions, despite the fact that after each division a rest- ing stage intervenes. \ ^ 9^\ • - •'-•r« **• '* *• ;«• ••i^. ^, 'A am !»• •.'5^ •«• .T: f^® ft^*-' "*. ^J;. i •»■ c Fig. 23. — Normal a, b, and reduced a', b', chromosomes of two species of Biston; and of hybrid c, c'. The evening primrose, CEnothera Lamar ckiana, has 14 chromosomes (reduced number 7). Individuals are known in which there are 15 chromosomes. As a result of accidental displacement at a division in a germ-cell, possibly one cell came to contain an additional chromo- some. Such a cell combining with a normal one, at fertil- 54 PHYSICAL BASIS OF HEEEDITY ization, would produce a plant of the 15-cliromosonie type. Here again, the additional chromosome persists as an individual element of the cell throughout subsequent cell-srenerations. / w" Wni^" /l! U / f///iimi!liPHl\\ ^ fe^ /Wi^iiiiiii Fig. 24. — Lagging and elimination of chromosomes in hybrid fish embryos. (After Pinney) . In Drosophila a female occasionally appears with two X's and a Y-chromosome. There are several ways in which this may arise, but the most common way appar- ently is for an egg to retain both of its X elements. Such an egg fertilized by a Y-bearing sperm produces an XXY embryo. Such an embryo retains throughout the entire MECHANISM IN SEGREaATION 55 development (cell-divisions) its two X's and its Y. There is evidence for this, obtained by Bridges, both from observation of the cells themselves and from the genetic behavior of such an individual. In certain crosses between moths with different num- bers and sizes of chromosomes, Federley, and Harrison, and Doncaster have shown that the cell of the hybrid contains half the number of each species, even with their characteristic size differences (Fig. 23). In crosses between different species of fish, where the size differ- ences are quite conspicuous, it has been shown by Moenk- haus, Morris and Pinney (Fig. 24) that the embryonic cells may continue through their divisions to retain the characteristic chromosomes of both species. These hybrid cases are particularly significant ; for the chromo- somes derived from the father are in the foreign medium of the protoplasm of the other species. Never- theless, in some cases they retain their own peculiarities, through successive cell generations. Evidence That Homologous Chromosomes Mate with Each Other That the mating of the chromosomes in pairs is not a haphazard process, but that each paternal chromosome 'c^% Fig. 25. — Female and male chromosome groups of Protenor. (After Wilson.) mates with a definite maternal chromosome, has been established by evidence from several sources. In many species the chromosomes are of different sizes, and some- times certain ones are markedly different in size from the others. In the bug Protenor the two sex-chromosomes 56 PHYSICAL BASIS OF HEEEDITY of the female are conspicuously larger than the others (Fig. 25,(2). When reduction takes place the sizes of Maturation Divisions of Protlnor Q I ■Sip Oocyte r Polar Spmdlc 2"^ Polar Spindle Mature C^^ and Polav Bodies Fig. 26. — Reduced chromoBome group; and extrusion of polar bodies in Proienor. Maturation Divisions or Protenor (5 ^1 ' 1 1 jf. OQ 0 OO Ifim ^o o ^^ \id WQo Melaphase oj the ^ j/^ ^ O q 1 ^' Spermatocyte -^/i^ Amapliase ot the P^ Spermalocyle %C^9o Metapkseoj the ^ -^J/^" 2^ Spermatocyte Anaphase o| the Fig. 27. — Reduced chromosome group of male; and spermatogenesis in Protenor. the fused jDairs show that these two large chromosomes must always unite with each other (Fig. 26). In the male of certain species, as in Protenor (Fig. 27), the MECHANISM IN SEGREQATION 57 sex chromosome has no mate, and therefore nothing to fuse with. Its size, after the others have conjugated (Fig. 27) shows that it remains single; while its failure to divide twice, as do the other chromosomes, corrobo- rates the view that having no mate of its own it never combines with any other. At the other extreme, the two very minute chromosomes in several of the Drosophila species must have united to form the smallest chromo- some of the reduced series (Fig. 28, a-a\ h-b'). In a few cases the X and the Y are different in size. Wlien they fuse (in the male) the size of the fused mass is what ¥.?!^ 0 a (t Fia. 28. — Diploid and haploid chromosome groups of Drosophila lu^ckii, a, a', and D. melanica, (neglecta) b, b'. (After Metz.) is expected, viz., the sum of the masses of X and Y, and their subsequent separation into parts corresponding in size to the fused bodies supports the view that conjuga- tion amongst the chromosomes is a very definite process. In the very exceptional case of a bug, Metapodius, there is a pair of small chromosomes called m's. When the other pairs enter the spindle the two m's come together, touch, and then separate, to pass to opposite poles. Resume The evidence from studies of the maturation of eggs and sperm shows that the paternal and maternal chromo- 58 PHYSICAL BASIS OF HEREDITY somes come together at this time in pairs, and subse- quently separate, so that each egg comes to contain one or the other member of a pair. The same process takes place in the formation of the sperm-cells. It is obvious that if one member of any pair contains material that produces an effect on some character as one of the end results of its activity, and the other member of the pair contains a different material, the behavior of the chromosomes at the time of maturation supplies exactly the mechanism that MendePs law of segregation calls for. ii CHAPTER IV MENDEL'S SECOND LAW— THE INDEPENDENT ASSORTMENT OF THE GENES Mendel proved that when races differ from each other in two pairs of characters, each pair considered by itself alone gives the 3 : 1 ratio, and the inheritance of one pair is independent of that of the other. If a tall race of peas with colored flowers is crossed to a short race with white flow^ers the offspring show the two domi- nant characters, i.e., they are tall and have colored flowers. If these are inbred they produce tall and short offspring (Fo) in a ratio of 3: 1, and these same individ- uals, if reclassified for pigment, are colored or white in the ratio of 3 : 1. For example, the ideal for 12 tall peas would be 9 colored and 3 white; and for 4 short peas there would be 3 colored and 1 white. Expressed in a diagram we have: 12 tall 4 short 9 colored : 3 white 3 colored. : 1 white The preceding w^ay of stating the results deals directly with the facts. The explanation of these results, based on the segregation of the members of the two independent pairs of factors, is as follows: If we call the gene for tallness by the same name as the character itself, viz., tall, and the gene for shortness by the name of this character, viz., short, and similarly for the other pair of characters, viz., color versus white, then when crossed the hybrid has two pairs of allelomorphs, tall color short white 59 60 PHYSICAL BASIS OF HEREDITY If at the maturation (whether of egg or sperm) tall and color go to one cell, then short and white go to the other cell; but if one of the pairs is turned, so to speak, the other way, thus short color tall white so that short and color go to one cell, then tall and white go to the other. Four classes of germ-cells are expected in i^i, namely, tall coior, tall white, short color, short white. Chance meeting of any one of these four kinds of pollen grains with any one of the same four kinds of eggs will give the sixteen recombination classes shown in the fol- lowing table: Eggs Sperm tall color tall white short color short white tall color tall white short color short white tall color tall color tall white tall color short color tall color short white tall color tall color tall white tall white tall white short color tall white short white tall white tall color short color tall white short color short color short color short white short color tall color short white tall white short white short color short white short white short white The four kinds of eggs are written above and the four kinds of sperm are written to the left. There are 16 pos- sible combinations. Since tall and color are dominant the recombinations give : 9 tall color, 3 tall white, 3 short color, 1 short white. In this table the genes have the same name as the character for which they stand, and these names are written out in full, but it is generally more convenient to use symbols for the genes in order MENDEL'S SECOND LAW 61 to save space and time. It is customary to represent the members of a pair by the same letter, as Mendel himself did, and to represent the dominant member by the capital letter, the recessive member by a small letter. Thus if ^ = tall and a = short; and 5=^ color and & = white, the recombination square becomes: Eggs AB Sperm AB Ab aB ab Ab aB ab AB AB Ab AB aB AB ab AB AB Ab Ab Ab aB Ab ab Ab AB aB Ab aB aB aB ab aB AB ab Ab ab aB ab ab ab Instead of using arbitrary letters for the characters as above, it has been found more convenient to use a mnemonic system in which the first letter of one of the members of each pair becomes the symbol. The two members of such a pair are then distinguished from each other by using a capital letter for one and a correspond- ing small letter for the other. For example, we might let t = short, T = tall, c = white, C = color. In this case the capital letter represents the dominant character, and the small letter represents the loss of that character, as seen in the recessive type. But besides prejudging the question as to what kind of a change took place in the germ-plasm to change a dominant to a recessive by assuming that it is due to a loss, this system is unsatis- factory in cases where many modifications of the same organ exist (such as the 40 eye colors of the vinegar fly), 62 PHYSICAL BASIS OF HEREDITY and where new ones are being found. For example, if the symbol R (red) is used for the dominant wild eye color, small r would stand for any one of 40 mutant eye colors, and when several of these occur in the same experiment there would be no way of telling for which one the small letter stood. Some other system becomes imperative in such cases, ^ and the most consistent seems to be to use a small letter for the mutant gene in question (or when unknown for the recessive gene), and the corre- sponding capital letter for its allelomorph (usually the wild type). Thus, 5 = short, iS^^tall, 21;:= white, TF = color. The recombination square for the same characters treated above is then: Eggs Sperm SW Sw sW sw SW Sw sW sw sw sw Sw SW sW SW sw sw sw Sw Sw Sw sW Sw sw Sw sw sW Sw sW sW sW sw sW sw sw Sw sw sW sw sw sw Since the large letters simply represent the wild type of each particular character it may sometimes simplify the formulae to omit them, or since this may lead to confusion in making up pairs of genes, some convention for wild type, such as N (normal), T (type), or the + sign, or a dash, or a dot may be used. Such short-hand methods are followed by many workers, but it is not necessary to advance the claims of any one of them here. If, for * An even more serious objection to the system is explained in "The Mechanism of Mendelian Heredity, " pages 233-235. MENDEL ^S SECOND LAW 63 example, the normal, meaning the wild type in each factor pair, is represented by N, the foregoing table becomes : Eggs » Sperm NN Nw sN 8W NN Nw sN 8W NN NN Nw NN sN NN sw NN NN Nw Nw Nw sN Nw sw Nw NN sN Nw sN sN sN sw sN NN sw Nw 8W sN sw sw sw In the preceding illustration one grand-parent (Pj) was assumed to have had both dominant characters (tall and colored), while the other grand-parent had both reces- sives (short and white). Obviously the grand-parents might have happened to be made up differently — one might have been tall and white, the other short and col- ored. The Fi plants {Ss, Wiv) would have been the same in either case, and so would the F2 results. In other words, for the principle of assortment it should make no difference from which parent the characters have come. This is illustrated in the following cross (Fig. 29), in which a wingless vestigial (recessive) Drosophila male having the wild-type color (dominant) is bred to long- winged (dominant) female with ebony (recessive) body color. The F^ flies have long wings and wild type body color. Inbred, they give 9 long wild type color, 3 long ebony, 3 vestigial wild type color, and 1 vestigial ebony. In the diagram the gene for vestigial is repre- sented by V, and its allelomorph for long wings by V; the gene for ebony by e, its allelomorph for wild type color by E, The germ-cells of the two Pj flies are therefore vE and Ve. Each contains the wild-type allelomorph of the recessive mutant gene in the other parent. The 64 PHYSICAL BASIS OF HEREDITY F^ fly has the formula vVeE. Independent assortment of the two pairs of factors v_ E^ T T gives four kinds of germ-cells both in males and fe- males, thus : vE Ve VE ve Any one of the four kinds of egg may be fertilized by any one of the same four kinds of sperm giving the same result as in the case of the peas, viz., four kinds of F2 individuals in the ratios of 9 : 3 : 3 : 1. In practical tests the occur- rence of or the possession of a race with both recessives in it is highly desirable for use in making a back-cross to Pj (instead of inbreeding F^'s), because the numerical results obtained by back-crossing furnish, for a smaller number of individuals, more significant data. For exam- ple, if a tall pea with colored flowers is crossed to a short pea with white flowers, and the F^ individuals (SsWw) are back-crossed to short white peas (sw), the expected ratio will be 1:1:1:1, because the four kinds of gametes in Fj (SW, Sw, sW, sw) will then reveal themselves in the offspring, since the double recessive individual (sw) used for back-crossing (having only recessive gametes) will not ^^ cover up'' any of the factors coming from the Fj hybrid. For instance, as shown in our type example, the Fj gametes are SW, Sw, sW, sw. The only kind of gamete produced by the double recessive, short white, is sw. When this meets each of the above gametes only four kinds of combina- tions are possible, vis^., SWsiv, Swsw, sWsw, swsw; and these zygotes, containing only the same dominants as the Fj gametes, will reveal what the kinds of gametes were. In practice an approximation to a 1 : 1 : 1 : 1 ratio is much more likely to be evident than an approximation to a 9 : 3 : 3 : 1 in which only one double recessive individual out MENDEL'S SECOND LAW 65 [HD CUD or:) CUD EH) CUD 31 Canutits. Ft Zjcpte ), and then FiQ. 42. — Scheme showing the opening out of the strands of the tetrad, a, in two planes, h, according to Robertson and W enrich. later the four strands come to lie side by side, i.e., ^'fuse," the result will be that shown in Fig. 41, c. Two of the strands represent crossovers in the sense that an interchange has taken place between a maternal and a paternal strand ; and if at the first spermatocyte division, when the threads begin to pull apart, the maternal sepa- rate from the paternal threads, two threads may be seen actually crossing each other (Fig. 41, d). They are here the two non-crossover strands, but if the two strands thrown to the left had been thrown to the right the two crossover strands would cross over. The scheme is essen- tiallv the same as the chiasma of Janssens, but the strands that cross may or may not (as here) represent the cross- over strands. CROSSING OVER AND CHROMOSOMES 103 The next two figures (Pig. 42, a, h) show how Robert- son and Wenrieh interpret the crossed threads, that they have observed in the spermatogenesis of some of the grasshoppers. The four strands are represented as con- jugating side by side in Fig. 42, a. Wlien the strands begin to open out preparatory to the first spermatocyte division the two maternal separate from the paternal at the ends of the tetrad, while in the middle of the tetrad the opening up involves the separation of a maternal and a paternal strand from a maternal and a paternal. In other words, the tetrad opens up in two planes at right angles to each other. This scheme also gives an appa- ymmfm'm«>'^ffmfm?immmmmif!imt^m n The two ''like'' pieces in all cases contain identical series of loci. 126 INTERFERENCE 127 The data also show that the series may break at two points, and that when this happens the three blocks oi' one set always correspond to the three blocks of the other series of genes. Thus interchange at two levels gives : ahcdEFGHijkl ABCDefghlJKL The same relation holds in principle for three or more breaks in the series. If in such a system the blocks have no commonest length, the break in the series at one level should not bear any relation to the place at which another break takes place. For example, if it is true that when a break occurred between D and E it had no influence on a break at any other point of the series, the blocks resulting from two breaks would not tend to be more of one length than of any other length. But if the evidence shows that when a break occurs between D and E the chance of another break occurring in that vicinity is decreased, or increased, the results vvould be expected to follow some definite law or principle, rather than be simply the result of chance. This IS in fact the case. An illustration may make this clear. Suppose when crossing over takes place within the blocks A B C D, and E F G H, and I J K L it can be recorded. If we know how often, when the break occurs only once in the series, it takes place in the first, in the second, or in the third block, we can then determine in those cases where breaking occurs in the first block, whether it is as likely to take place in the second block as when no break occurs in the first, etc. Such tests have been made (Muller, Sturtevant, Bridges, Weinstein, Gowen) with Drosophila, and the same kind of results con- sistently obtained. It has been found, for example, that when a crossing over takes place between G and H, a sec- ond one is less likely to take place on either side, i.e., between F and G or between H and / than when no cross- 128 PHYSICAL BASIS OF HEREDITY ing over takes place between G and H. Stated in another way, crossing over in one region protects neighboring regions from crossing over. Moreover, this relation fol- lows a perfectly definite law according to the "distances," as determined by linkage relations of genes outside of the region of crossing over. If we take two pairs of factors closely linked together we find that the genes lying G H g h immediately to the right and left of ^ never cross over independently of - and ~ at the time that a crossover separates - and f . In other words, the genes imme- diately to the right of H always go over with Hy and those to the left of G always go over with (7, when G separates from H, If we consider genes that are less closely linked with G and with Hj we find that while their crossing over is interfered with by the crossing over between G-Hj it is affected to a limited extent. G-enes still less linked with G or with H are still less interfered with ; until finally there is no relation at all between crossing over between G-H, and other more loosely linked genes, i.e., crossing over between G-H is found to have no relation to crossing over between L and M. Put in another way, one may say that crossing over at L and M is no more likely to take place when none occurs between G-H, than when it does. For different pairs of chromosomes the regions that bear this relation to each other have been found to be different. Even within the same chromosome this rela- tion may be different at the ends and in the middle. There are also special factors that affect special chromosomes and special regions of chromosomes. An example will illustrate this relation that is called interference. If in a group of genes A B C D E F 3, break occurs somewhere between A and D in 6 per cent, of cases, and if between M and T in the same series {M N 0 P Q R S T), inlOipeT cent, of cases, a double break involving both regions simul- taneously should, if the breaks occurred independently of INTERFERENCE 129 each other, take place in 0.6 per cent, of the cases. But if the regions in question are close together, that is, if the intervening block (i.e., G F E J K L) oi genes is short, it is found that there are fewer double crossovers than the 0.6 per cent, expected on a purely random basis. This was shown by Sturtevant in his paper on chromosome maps. It means that a break in one region interferes with a break in the other region when the intervening block is short. The ratio of the number* of actual double breaks obtained to the number of double breaks that would occur if one of them did not interfere with the other is termed coincidence. If in the above example only 0.3 per cent, of the cases were double crossovers involving the regions ABCDEF SiTid MN0PQR8T the coincidence would be 0.3 per cent, divided by 0.6 per cent., or 0.5. It has been found that as the distance between two regions increases, crossing over in one of them interferes less and less with crossing over in the other; that is, the number of double crossovers obtained approaches the number expected on a random basis, and coincidence rises gradually to the value of 1. This phenomenon is shown in all the cases where more than one block of genes has been followed. It is especially clear in the work of Muller, who studied a large number of factors in the sex-chromo- some of Drosophila simultaneously. When the intervening block becomes sufficiently long so that the coincidence attains the value of 1, interference has entirely disappeared. When, however, the distance is increased still further interference reappears, i.e., coin- cidence decreases again. There was a suggestion of this in Muller 's work; and the work of Weinstein undertaken to get critical evidence on this point indicates clearly that such a decrease exists. For the second chromosome a similar rise and fall with increase of distance is indicated by Bridges ' data. 9 130 PHYSICAL BASIS OF HEREDITY The fact that interference reappears, i.e., that coinci- dence decreases after reaching a maximum, indicates that the segment of a chromosome between the breaking points tends to be of a particular (modal) length; and that breaks which are closer together or farther apart than this modal length are less frequent. That is, genes not only stick together in blocks, but the blocks tend to be of a definite size, and longer and shorter blocks are less frequent. In the sex-chromosome of Drosophila, which is 65 units long, Weinstein's data indicate that the most frequent length of block is about 46. In the second chromosome (which is 107 units long). Bridges^ data indicate a modal length of about 15 in the centre of the chromosome and of about 30 on either side of the middle point. The work on coincidence throws light on the behavior of the chromosomes during crossing over. The cytologi- cal evidence has not determined whether when crossing over takes place the chromosomes are twisted loosely or tightly. But Muller has shown that this question may be attacked by certain calculations based on the data of inter- ference. If, as a rule, chromosomes twist in long loops, crossing over at two points close together would be rare, for it would require a shorter twist than usually occurs. The occurrence of long loops would explain the interfer- ence of neighboring regions. Moreover the decrease of interference as distance increases would be accounted for, because short loops would be less frequent than longer ones. The reappearance of interference for widely separ- ated regions is explained by supposing that extremely long loops are infrequent as are very short ones. That is, on the supposition of long twists there would be a modal length of loop, and loops of greater or lesser length would be less frequent. If, however, the chromosomes are tightly twisted into short loops, the interference of neighboring regions might be explained on the supposition that a break at one point allows the chromosomes partly to unravel in the neighbor- INTERFERENCE 131 hood of the break, and that this loosens the twisting and prevents another break near by. In regions farther away from the break, the threads would not be so much unrav- elled, so that the greater the distance from the first point of breaking tlie more would a second break be likely to occur. That is, interference should grow less at greater distances. But the reappearance of interference at still greater distances seems incompatible with this scheme; thus the actual data favor the first view of crossing over, in which the break occurs during a stage of loose twisting. At any rate, as Weinstein has pointed out, the variation of coincidence with distance must be dependent on other conditions than the mere tension due to the twisting of the chromosomes, and any view which refers the breakage of the threads to the tension of tight twisting must be rejected or supplemented. Castle has recently suggested that the difference between the values for a long ''distance'' and summation of short ''distances'' is due to the loci not lying in a straight line but "out of line." He suggested that when short steps are taken as the basis for map distance they represent the "long way round," as, for instance, in passing from one end of a F to the other end, keeping on the line; while when a direct cross is made, giving a shorter "distance," this is a measure of the direct or air-line between the two ends of the F. Such a theory is not in harmony with the following facts. The best data (i.e., data sufficient in amount and free from crossover variations) show that Castle's three dimensional fi,2:ures reduce to a curved line in a plane. In such a curved line the most distant points are nearer to each other in an " air- line" than along the line. Such a graphic representation of the data is possible, but leads to certain inconsistencies. If Castle's procedure is followed it leads to the placini,' of the same locus in two or more different places on the basis of adequate and comparable data for both positions. The two cases that Castle says furnish the crucial evi- 132 PHYSICAL BASIS OF HEREDITY dence for his view demonstrate just the opposite, when complications due to crossover variations are excluded, by using only data in which three or more loci are recorded simultaneously. In his attempt to explain the all-import- ant fact of rarity of double crossovers, Castle is obliged to assume that there is a difference in frequency of cross- ing over in different planes (directions). This assump- tion can be shown to be inconsistent with the primary assumption that he accepts, viz.y that crossing over is proportional to the distance between genes. CHAPTER XI LIMITATION OF THE LINKAGE GROUPS It may be questioned whether we are at present justi- fied in speaking of the limitation of the linkage groups to the number of chromosome pairs as one of the funda- mental principles of heredity, since the only species in which a correspondence that is numerically significant between the two has been proved is Drosophila melano- gaster. But despite the absence of other positive evi- dence, the fact that in no other animal or plant does the number of linkage groups exceed the number of the chromosome pairs, may be, I think, legitimately inter- preted in favor of the view. It may also be argued, that if the phenomena of linkage are assumed to be due to the genes being carried by the chromosomes, it follows that there could be no more groups of linked genes than there are chromosome pairs; hence one relation is the direct outcome of the other. But the proof of the linear order that has been developed here rests directly on the linkage data, and is independent of any assumption concerning the chromosomes. It has been shown, secondarily so to speak, that the chromosomes ful- fill all the requirements of the abstract reasoning from the data, and therefore give a mechanism capable of per- forming all that the theory demands. The demonstration, then, that in Drosophila the linkage groups correspond in number to the chromosome pairs may be taken as a con- clusion or a discovery independent of the other relations furnished by linkage. If then, as I anticipate mil be the case, further work in other groups should show that the same relation holds everywhere, we should be fully war- ranted in stating the result as one of the general prin- ciples of heredity. 133 134 PHYSICAL BASIS OF HEEEDITY In Drosophila melanogaster the evidence is now very strong in favor of the identity in number of linkage groups and chromosome-pairs. As the new characters coming up, one after the other, have continued to fall into the four kno^\Ti groups, and as something like 200 characters have been so placed, and as none of them has failed to show link- age with one of the four established series, the probability is enormously in favor of a causal relation between the two events, especially in the light of the evidence from other sources that the chromosomes are the bearers of the hereditary factors — evidence from the sex-chromosomes, for example. The only other species in which the heredity of known mutant characters approaches that of the chromosome group is the garden pea in which about 35 mutant factors have been studied. From the summary of what has been so far recorded, as well as from the results of his own work. White has recently given an account of what is fairly well established. Of the 35 mutant factors in this pea, seven independently inherited groups have been recorded, i.e., each one of seven factors has been tested out and found to assort independently of the other six. There are seven pairs of chromosomes in the edible pea (Fig. 53, a). The agreement between the two is to date perfect. It is, of course, possible that the linkage between some of the factors tested was so loose that they a])peared to give free assortment, and that until more fac- tors have been studied the evidence is not above sus- picion. Nevertheless, it is important to find that the number of independent mutant factors in Pisum sativum does not exceed the number of chromosome pairs. White's study of the linkage of factors in the edible peas shows further that there are four linkage groups-— three of them include factors that are also included in those that freely assort. It is fair, perhaps, to conclude that four of the possible seven-linked groups have been found. There are no other forms known in which the LIMITATION OF THE LINKAGE GEOUPS 135 number of linkage groups approaches so near the number of the chromosome pairs. In the snapdragon, Baur has described two linked groups. He states that there are 16 pairs of chromosomes. In wheat one linked group has been described. There are 8 pairs of chromosomes (Fig. 53, h). In Indian corn there appear to be a few linkage groups, and probably 10 pairs of chromosomes. In oats, Surface finds two linked genes. In Primula there is one group composed of several linked genes, and 12 pairs of chromosomes (Fig. 53, c). In the silkworm moth one linked group of genes has been found by Tanaka, and Yatsu records (Fig. 55) 20 pairs of chromosomes. In Drosophila virilis three linked groups of genes have been found by Metz, who has also .f«r • Fig. 53. — Chromosome group of pea, a, wheat, b, and primula, c. described six pairs of chromosomes for this fly. In Droso- phila huscMi there is one group of linked genes and four pairs of chromosomes. In D. repleta one group, and six pairs of chromosomes. The groups of chromosomes in some of the different species of Drosophila, as described by Metz, are shown in Fig. 54. As indicated by the arrangement of the figures (that correspond fairly closely with the actual arrangement of the chromosome in the cells themselves) it appears that one pair of chromosomes in one species is at times represented by two pairs in related species, and this view is borne out by the attach- ment of the spindle fibre to the middle of the chromosomes in the bent pairs, but to the inner ends of the two that supposedly correspond to its halves in other species. In the mouse one group of linked genes has been reported. There are 20 pairs of chromosomes ( Fig. 55, b ) . In man no linked genes are known, if we do not count sex- 136 PHYSICAL BASIS OF HEREDITY lined genes, which must, however, if carried by the sex- chromosomes, be linked to each other. The number of chromosome pairs in man is, according to Grnyer, 12 (Fig. 55, c), but Winewarter describes 24 pairs (Fig. 55, d). The difference would seem to be due to technic, rather than to differences in different races of men. Ma IK ih IK IK II. IL «„ ^.^^ ^.^ M II ? I ^ ni i J T/m) ® x(nt Fig. 60. xy(4n) Fxa. 60. — Diagram illustrating the formation of £n individuals from the regeneration of the sporophyte in a dioecious species. (According to Marchal.) Fia. 61. — Diagram illustrating the formation of 2n individuals from the regeneration of the sporophyte in a hermaphroditic species. (According to Marchal.) sider the former first. If the sporophyte is removed and cut across, its cells regenerate a tangle of threads (pro- tonema), which is the beginning of a new moss plant (Fig. 60). Since the sporophyte had the double number (2iY) of chromosomes, it is to be expected that the young moss plant that regenerates from its tissue (sporophyte) will also have the double number, and such proves to be the 154 PHYSICAL BASIS OF HEREDITY case. The new moss-plant is therefore 2N (or diploid) instead of being liV, as in the normal mode of propaga- tion. Since no reduction has taken place into male- and female-producing individuals, it would seem possible that such a plant might produce either or both sexes. Such is the case, for Avhen the 2N moss plant produces its *' flowers^' some contain archegonia, others spermato- gonia (with their contained germ-cells) and other flowers contain both. The hermaphroditism here produced would seem to be the sum of both the contrasted elements. The expectation from such a 2N' plant would be that its germ- cells i2N) would produce a 4iV sporophyte — unfortunately the plants proved sterile. Imperfect germ-cells were present incapable of fertilizing or of being fertilized, so that it was not possible to perpetuate the 2A^ plant by sexual reproduction. The results mth the 2N plants derived from the regen- erating sporophyte of the hermaphroditic species (Fig. 61) is different in one important respect. When, as before, a diploid (2N) plant is obtained by regeneration from the sporophyte it produces hermaphroditic flowers, i.e., flowers containing both oogonia and spermatogonia, and these are fertile. The sporophyte that they produce is tetraploid (4iV), due to the union of a diploid anther- ozooid with diploid egg. Regeneration from the tetraploid sporophyte (4iV) should produce fertile gametes, which might give rise by their union to an octoploid sporophyte (SN). So far the Marechals have not been able to produce such plants, for although in a few cases the 4:N sporophyte regenerated it failed to produce flowers. The difference then between the results from mosses with separate sexes and mosses that are hermaphrodite is that the 2N plant of a race with separate sexes does not form normal gametes, while a 2N plant of hermaphroditic races forms fertile gametes. It may appear more or less plausible that the failure of the former is due to failure in the reduction of the spores into two alternative types, VARIATION OF CHROMOSOMES 155 while in the latter case, since there are presumably no such types found, there is no conflict. Some other dif- ference would have to be appealed to to explain why the octoploid forms fail to develop. A triploid condition {3N) has been found to occur in certain types of the evening primrose (Stomps, Lutz, Gates). De Vries has found in crosses in which Lamarck- iana was the mother and some other species (muricata, cruciata, etc.), the father, that triploid types appear three times in 1000 cases. He interprets the results to mean that three in 1000 times the egg-cell of Lamarchiana has the double number of chromosomes (14), which being fer- tilized by a normal pollen grain with seven chromosomes, gives the triploid number, viz., twenty-one chromosomes. The same result would be reached if a diploid pollen grain fertilized a normal Qgg, That such pollen grains appear is as probable a priori as that diploid eggs occur. It may be recalled that one explanation of the tetraploid evening primrose (gigas) is that it arises from a 22V pollen grain meeting a 2N egg-cell. How reduction takes place in the triploid Oenotheras is uncertain, since the accounts of the process are different. Geerts states that, as a rule, only seven chromosomes conjugate (7 + 7), while the remaining seven chromosomes are irregularly distrib- uted in the dividing germ-cells. On the other hand, Gates finds in a 21-chromosome type that the chromosomes separate into groups of 10 and 11, or occasionally into 9 and 12. The former account fits in better with results of the same kind obtained by others, and is more easily understood from a general point of view, because seven homologous pairs would correspond to the normal conju- gation, while the seven chromosomes left over would have no mates and fail to divide at the reduction division, hence their erratic distribution. It has also been shown in (Enothera that there are three 15-chromosome types. If the 15th chromosome is 156 PHYSICAL BASIS OF HEREDITY sometimes one, sometimes another chromosome, there may he genetically several types, but as yet evidence on this point is lacking. Irregularities in the germ-cells of (Enothera have been observed by Gates of such a kind that one cell gets 6, the J ^ r Fig. 62. — Somatic chromosomes groups of (Enothera scintiUans, showing variable numbers of chromosomes. (After Hance.) other 8 chromosomes. A pollen grain with 8 chromo- somes fertilizing an egg with 7 would give a 15-chromo- some type. When such a 15-chromosome plant forms its egg-cells the supernumerary chromosome having no mate may go to either pole of the spindle, hence eggs of two VARIATION OF CHROMOSOMES 157 sorts would result, vi^., 7- and 8-chromosome cells.« Such a plant if crossed to a normal plant should give half nor- mal (14), half 15-chromosome types. Such plants have been shown, in fact, to be produced (Lutz). Other com- binations that would give 22, 23, 27, 29 chromosomes have been reported. A variation in the number of the chromosomes of a somewhat different kind has been described by Hance for CEnothera scmtillans, one of the 15-chromosome types of 0, Lamarchiana. No variation in number was found in the germ-tract of the same individuals that consistently gave two types of pollen grains, one with 7 and the other with 8 chromosomes. The number of chromosomes in the somatic cells was found to vary from 15 to 21. Some of the groups are shown in Fig. 62. When the 15 chromo- somes of the type-group are measured, it is found that they can be arranged in respect to length in 7 pairs, with one odd one (marked a in the figures). There is also found a constant length difference between the pairs. In those cases where there are more than 15 chromosomes in a cell, measurements show that the pieces can be assigned to particular chromosomes. When this is done. Fig. 63, the lengths of the chromosomes come out as in the typical cells. There can be no doubt that the extra chromosomes in these cases represent pieces that have broken off from typical chromosomes. This process of fragmentation does not destroy the ^'individuality of the chromosomes'' since the increase in this way of the number of chromo- somes would not lead to any immediate change in the number of the genes. The peculiarity of the mutant 0. scintilluns is not connected with the increase in the number of its chromosome bodies, but rather to the presence of a 15th chromosome. Bridges has called attention to a peculiar case in Drosophila (1917) in which an individual behaves as ' No pollen is produced by most of the lata plants. 158 PHYSICAL BASIS OF HEEEDITY though a piece of the X-chromosome (recognizable from its genes that normally lie in the middle of the chromo- some) had become attached to one end of the other X-chro- mosome. Owing to this piece (including the region that contains the normal allelomorphs of vermilion and sable) the individuals give unexpected results in relation to domi- nance or recessiveness of certain factors. For example, ABCDEFGHIJKLMNO A^BCDEFQHI JKLMNO m 1 h 0 ABCOErdHijKlMNO a P AeCDE^eVil JKtMfiO b c e f S K k 4 < p g 1 g 1 m •• "^ :V| BCDEFGHt JKLMNO ■1 * ^ 9 k I ^ 6 Fig. 63. — Scheme showing the probable relation between the extra chromosome pieces of Fig. 62, and the normal 15 chromosomes of this mutant. (After Hanse.) a male that contains the recessive genes for vermilion and for sable, normally located, and having attached to this chromosome the duplicated piece (containing the normal allelomorphs of vermilion and sable) is in appearance a wild-type fly, instead of being vermilion sable as it would otherwise be without the piece. On the other hand, a female having one such chromosome and a normal ver- milion sable chromosome is in appearance not wild type VARIATION OF CHROMOSOMES 159 (as might have been expected), but shows vermilion and sable, because in this case the two recessive genes for vermilion and for sable dominate the single normal allelo- morphs. But a female having two such duplicated chro- mosomes {i.e., tetraploid for the genes of certain regions of the sex-chromosome) is now wild type in appearance, because the two dominants dominate the two recessives. Such a female crossed to a vermilion sable male gives wild- type sons and vermilion sable daughters, which is criss- cross inheritance in an opposite sense from that ordinarily met with in Drosophila. A second instance discovered by Bridges, but not yet reported, seems best explained on the assumption that a piece taken from the second chromosome has become attached to the middle of the third chromosome. This condition makes possible the linkage of mutant characters to genes in both the second and the third chromosome at the same time. The second chromosome that lost a piece, and the third chromosome that gained the piece (both were of course in the same cell), have been easily kept together in the same stock ever since, because in those cases where they become separated through assortment every zygote that receives the deficient (2nd) chromosome dies unless the same zygote has received the third chromosome with the duplicate piece. The preceding results show that chromosomes may not only gain genes by the attachment of pieces (duplication), but also that chronaosomes may lose pieces ( deficiency) . Other instances of deficiency have been reported by Bridges which can be explained either as total losses of certain regions, or due to their inactivation. Unless the lost pieces happen to have been retained as in the^ last case, the distinction between these possibilities is difficult. A study of one case has shown that no crossing over takes place in the region of deficiency, although the rest of the chromosome was little or not at all affected. As a result 160 PHYSICAL BASIS OF HEREDITY the chromosome is ' * shortened ' ^ by an amount correspond ing to the "length" of the deficient region. It is not Avithout interest to notice that in the first case the duplicating piece is attached to that end of the first chromosome where the spindle fibre is attached. In the other case the duplicating piece is attached to the a b Fio. 64.— An egg of Ascariabivalena fertilized by sperm of A. univalens, a; later stage of same, b. middle of the third chromosome, and in this chromosome the spindle fibre is attached to the middle. An interesting case of triploidy has been reported in the threadworm Ascarw (Boveri). Two varieties occur, one with four chromosomes (haploid two), and one with two (haploid one). Rarely a female of one variety is a *':riW^^ ^--^^s V^V-' Fig. 70.— Cross between white-eyed female and a red-eyed male of the vinegar fly. 170 PHYSICAL BASIS OF HEREDITY are "bred together, the results are as follows : Half the daii.srhters have white eyes (wiv), half red eyes (wW) ; half the sons have Avhite eyes {w)j half red eyes (TF). It will he seen tliat the red-eyed Fg daughters are all hetero- zygous, and should give 50 per cent, white and 50 per cent, red offspring if mated to white-eyed males. This occurs. Similar illustrations might be given for any of the 50 sex-linked characters of Drosophila. Of these the sex- linked lethals form the most interesting cases and will be spoken of in another connection. Despite the fact that the results in one of the two fore- going crosses gave a 3 : 1 ratio, and in its reciprocal a 1 : 1 ratio, the results in both cases conform to Mendel's first law of segregation. The peculiarity of the 1 : 1 ratio is due to the fact that the P^ red-eyed male is in a sense heterozygous for the wild-type eye color (since he has but one X-chromosome that carries the factor for red eyes). Since in the second cross the F^ male gets no red- producing X from either parent, he is pure for white eyes in the sense that he has an X bearing the factor for white eyes and a Y that bears no factor making red. Plence this F^ cross is exactly like a back-cross of a hetero- zygous female to a recessive male, and gives the same numerical result, viz., 1:1. Cases of sex-linked inheritance of this kind are also known in man. Color blindness in man appears to follow exactly the same procedure as sex-linked inheritance in the vinegar fly — at least certain kinds of color blindness have been sho^vn to do so. Haemophilia also is sex-linked, and there are four or five other defects in man that appear to come under this head. According to several accounts there is an unpaired sex-chromosome (or two of them) in man, which is also called for by the genetic evidence relating to sex -linkage in man, but since the female number of chromosomes in man is stated by Guyer to be 24, and by von Winiwarter to be 48, it is unsafe as yet to appeal SEX-CHROMOSOMES AND INHERITANCE 171 FiQ. 71. — Cross between a yellow, white-eyed female and a wild-typeC'gray"), red-eyed male. 172 PHYSICAL BASIS OF HEREDITY to this evidence as showing the identity of the sex-deter- mining mechanism of man and the vinegar fly. When two or more sex-linked characters are involved at the same time, the situation is different only in so far as crossing over may take place in the female. It will be simpler to consider such a cross and its reciprocal in the reverse order from that just given. If a female with yellow wings {yy) and white eyes {wiv) is crossed to a wild-type male, ''gray^' wings (Z) and red eyes (TF), the sons are yellow white and the daughters are gray red (Fig. 71). Wlien these are inbred there are four types in F2 (ignoring sex), viz., the two original combinations yellow white and gray red, and the two crossover com- binations yellow red and gray white. They occur in the following ratios : Yellow white Gray red Yellow red Gray white 99 per cent. 1 per cent. In this case the F^ male acts as a double recessive, reveal- ing the amount of crossing over in the F^ female. Since neither his female-producing nor his male-producing sperms carry factors that cover up the characters carried by the four classes of gametes in the F^ female, aU four classes of her gametes are revealed in their numerical proportions. Eeciprocally, when a male with yellow wings {y) and white eyes {w) is crossed to a wild-type female (gray {YY) red {WW), both sons and daugh- ters are gray red, because both get the dominating genes for these characters carried by the X-chromosome received from the mother. If these F^^^s are inbred (Fig. 72), the F2 females are gray red, since each contains an X with the two dominant genes derived from the father whose genes have remained completely linked, as there is no crossing over in the male. On the other hand, there are four kinds of JPg males : yellow white ; gray red; yellow red; gray white; because each male shows the SEX-CHEOMOSOMBS AND INHERITANCE 173 character of his single X-chromosome, and there are four kinds of these chromosomes in his mother on account of crossing over in the female. The other sex-chromosome, the Y, has no dominating influence. Fia. 72. — The F2 results from the reciprocal cross of that shown in Fig. 71. Sex-linked Inheritance of the Abraxas Type In certain moths and birds it has been shown by the genetic evidence that the female is heterozygous for sex- linked factors. The cytological evidence, as far as it goes, supports this evidence, but for birds the material is so 174 PHYSICAL BASIS OF HEREDITY difficult to interpret that Guyer's conclusions do not seem to me as yet to be on as secure grounds as those of Seller 's for moths. Both descriptions give, however, the bases for a consistent explanation of sex-linked inheritance in this type (WZ-ZZ). Since we do not know as yet whether the same or dif- ferent sex factors are involved in the Drosophila and in the Abraxas types, it seems best not to use the same sym- fVZ zz IV z zz Fig. 73. — Scheme showing the relation of the eex-chromoeomeB of the moth (and of the bird) in Bex-determination. WZ-ZZ type. bols in both for the sex-factors. If in both types a single sex-factor is concerned, and if it is the same in both, the conditions that make for a female in one case and for a male in the other must be due to a difference in the rest of the hereditary complex that reverses the reaction. It would appear simpler to assume that the sex-factor itself is different in the two cases. If there is more than one factor for sex, the two types may have some in com- mon, but the theoretical situation would remain the same. For our present purpose these possible distinctions are of no importance. SEX-CHEOMOSOMES AND INHERITANCE 175 If the sex-chromosome that carries the sex-linked genes in birds and moths be symbolized by Z, and its homologiie that occurs in the female by W, the scheme for sex-deter- mination is that shown in Fig. 73 : The eggs of the female extrude either one or the other sex-chromosome. If Z stays in, and this egg is fertilized by a sperm (Z-bearing LACTICOLOR 9 OL CroSULARlATA c5 CG O (T) GERM CEaS (g) GR0S5ULAR1ATA 9 OG , CRQSaJLARlATA 6 LG O ® GERM CELLS , ® ® o ©■• LACTICOLOR 9 OL o ©• GR0SSULARIATA90G ® ® ^® ® CROSSULARlATAd GL GR055UU\R1ATA d CO Fig. 74. — Cross between Abraxas lacticolor female and grosaulariata male. also) a male (ZZ) is produced; if W stays in, and the egg is fertilized by a Z-bearing sperm, a female (WZ) is produced. The way in which sex-linked characters are transmitted may be illustrated by the inheritance of a color difference in the currant moth Abraxas. The wild species (grossulariata) has a mutational variety called lacticolor, that differs from the former by having less 176 PHYSICAL BASIS OF HEKEDITY black pigment in the wings. When a dark (grossulariata) male is mated to a light (lacticolor) female, both sons and daughters are dark (Fig. 74). If these are inbred all the F2 sons are dark, half the daughters are dark, half light. As the diagram shows, the distribution of the Z-ohromo- some furnishes the mechanism by means of which we can C3^05SULAR1ATA 9 OG LACnCOLOR d LL O ® GERM CELLS ® / \ \ LACTICOLOR 9 OL \ GR05SUL^R1ATA c5 GL O ©^GERMCEaS^^® ® O ®'' GROSSULARIATA 9 OG o ® ® ® LACTICOLOR 9 OL GROSSULARIATA d LG ® ® LACTICOLOR 6 LL Fia. 75. — CroBB between Abraxas grossulariata female and lacticolor male. explain, as in Brosophila, the process of sex-linked inheri- tance in this moth. The reciprocal cross is shown in the next diagram (Fig. 75) in which a dark (gross.) female is mated to a light (lact.) male. The daughters are light like the father, the sons dark like the mother — criss-cross inheritance. The daughters get their one Z-chromosome carrying the light SEX-CHROMOSOMES AND INHERITANCE 177 factor from their f athe •, the sons get in addition to a light Z from their father a dark dominating Z from their mother. When the F^'s are bred together four classes result in the proportion of 1:1:1:1, when sex is taken into consideration, or in the ratio 1 : 1 for the color differ- ences alone. According to Doncaster, the male and the female Abraxas have each 56 chromosomes, i.e., the female is ZW rather than ZO; but as yet the sex-chromosomes as such have not been identified. That sex is connected with such chromosomes is not only established by sex-linked in- heritance, but is also indicated by an aberrant race of Ahraaxis found by Doncaster. The males of the race had the normal number of chromosomes (56), but the females had only 55 chromosomes. Doncaster found that in these females an unpaired chromosome, presumably the Z-chro- mosome, was more often thrown out into the polar body than left in the Qg^, so that most of the resulting eggs had only 27 chromosomes. Any Qgg of this kind fertilized by a spermatozoon should give a 55-chromosome indivi- dual, i.e., a female. The few eggs that retained the unpaired Z-chromosome, fertilized by a Z-spermatozoon, would be expected to give rise to the rare males, which like normal males have 56 chromosomes. The excess of females is thus accounted for, and incidentally the results show that the TF-chromosome carries no essential factors for the life of the individual, since females without it develop and look like normal females. Probably it is empty as is the Y of Drosophila. In poultry there are several cases of sex-linked inheri- tance that follow the Abraxas type. One of the most striking cases is the cross between Barred Pl5anouth Rock and Langshan. When a barred male is crossed to a black female, the sons and daughters are barred (Fig. 76). Barring is dominant to black. Two such F\ 's, inbred, give all barred males ; half the hens are barred, half are black. It may be said here that the black grandmother transmits 12 178 PHYSICAL BASIS OF HEKEDITY lier black color to only half of her grandsons. The chro- mosomal explanation can obviously be worked out on the same scheme as in Abraxas (Fig. 77). But if Guyer's recent account of spermatogenesis in birds is correct, the situation is different. Guyer describes the ripening of the sperm as follows : There are 18 chromosomes in the male, including two large Z^s (16 + 2). After synapsis there are 9 double chromosomes in the first spermatocyte, all of which, except ZZ separate at the first maturation division, 8 going to one pole and 8 to the other. One Px Darned (f Blacky zV F, z^z"^ zV Darredc? Barred o z°z" z^' zV zV Darredc? Darredc? Darredo Black: 9 FiQ. 77. — Scheme showing the transmission of the sex-linked characters B = barred, and b =black in the cross shown in Fig. 76. daughter cell gets both Z's ( 8 + 2 ) . This cell then divides again, the Z's presumably separating so that two second spermatocytes are produced, each with 9 chromosomes (8+1), including the Z. These become the functional sperm. The other spermatocyte, the one without a Z, may divide again, but it, or its products, degenerate, and never produce sperm. According to Guyer, there are 17 chromosomes in the female, including one Z. Presum- ably, then, after reduction half of the eggs* will contain a Z (8 + 1), the other half will be without it (8). The egg that carries a Z (8 + 1), fertilized by a sperm (each sperm carries a Z (8 + 1)), wiU make a male with 18 chromo- Fig. 76. — Cross between Barred Plymouth Rock male and Black Langshan female. I\ /•; mm€k Fig. 78. — Cross between Black Langshan male and Barred Plymouth Rock female. SEX-CHROMOSOMES AND INHERITANCE 179 somes, including two Z's, The egg that lacks a Z (8), fer- tilized by a sperm (8 + 1), makes a female with 17 chro- mosomes, including one Z. This scheme gives consistent results for sex-linked inheritance in birds. Since the daughter gets her single Z-chromosome from her father, she will show any sex- linked characters carried by his Z-chromosome. If the father carries a sex-linked dominant gene his sons and his daughters will be alike. It should be noticed that while Guyer's scheme gives the same results so far as sex-link- n BlackcJ* Barred Barredc? Black 15 zV tz\ ^ z V zV Barredd Black d DarredQ Black o Fig. 79. — Scheme showing the transmission of the sex-linked characters B = barred, and b =black in the cross shown in Fig. 78. age is concerned, as the one described by Seller for some moths, the machinery in the male is different in the two cases, while that in the female is presumably the same. In both the female is heterozygous for Z ; in the moth the male is homozygous (ZZ), but in the bird the two Z'5 described by Guyer both go to one pole at one of the maturation divisions, and reduce at the other — a proce- dure not known in any other animal. In the reciprocal cross (Fig. 78) a black cock is bred to a barred hen. The sons are barred — like their mother — the daughters are black — like their father, criss-cross inheritance. When the barred Fj cock and the black hen 180 PHYSICAL BASIS OF HEREDITY are inbred, there are four Fo classes with sex taken into account in the proportion of 1:1:1:1; or ignoring sex, 1 barred to 1 black. The barred and the black races differ by one factor difference (Fig. 79), viz., barred Z^ and its nonnal recessive allelomorph Z^. This seems to mean that the Barred Plymouth Rocks is a black race with an additional dominant factor for barring. The Black Langshan is the same black race but without the barring factor. Until quite recently no cases of crossing over had been observed in forms having the Abraxas type of sex-linked inheritance, for, except in one or two cases in poultry, only a single pair of sex-linked genes were known, and two at least must be studied together in order to demonstrate linkage. Goodale has recently studied two sex-linked characters in poultry, and states that crossing over occurs in the male, but whether or not in the female is not stated. Sex-detekmination and Natural. Paethenogenesis Variations in the ordinary sex-determining mechanism account in some cases for the normal output of males and females produced by parthenogenesis, and determine the exceptional sex-ratios of such species. The honey bee furnishes the best known example. The queen comes from a fertilized Qgg, and has therefore the double (22V) number of chromosomes. Her eggs give off two polar bodies, hence have the reduced, or single number of chromosomes. Any Qgg that is not fertilized develops parthenogenetically into a male. If there are two X-chro- mosomes in the bee, as in some of the otier insects, the Qgg is expected to contain only one of them after the extrusion of the polar bodies. Hence, if it develops with- out doubling its chromosomes, it should give rise to a male. That the male has the smgle number of chromo- somes is also borne out by the evidence from a peculiarity of the first spermatocyte division in which the cytoplasm divides, but the chromosomes do not separate into two SEX-CHROMOSOMES AND INHERITANCE 181 groups. Several stages in the maturation of the sperma- tozoon of the bee are shown in Fig. 80. In a, the spindle for the first spermatocyte division has appeared. A small piece of the cytoplasm cuts off, but the chromosomes do not separate, and they return again {h and c) to a resting a d e FiQ. 80. — First spermatocyte divisions a-c, and the second sperniatocyte division d-\i in the bee. (After Meves.) stage. Another spindle forms (eZ), and the chromosomes separate into two groups, one of which is pinched off as a rudimentary cell that never becomes a spermatozoon. Hence only one, and not four spermatozooa as in ordi- nary cases, is formed from each spermatocyte. In the hornet (Fig. 81), the spermatogenesis is similar to that of the bee in that the first division is abortive. It is different 182 PHYSICAL BASIS OF HEEEDITY d f Fig. 81. — First spermatocyte division a-c, and the second spermatocyte division d-f in the hornet. (After Meves.) in that the second division produces two functional sperms, both female producing. Since the male comes from an unfertilized egg^ the '/^//j///{ : rr 'rr/ rrr 'f//auY/////l'i i*=^5>».^«i« St^m ^fff'f/it I 41 r-r iliy^aait. ijo^yrrfuef.x, ? Fig. S2. — Life cycle of Phylloxera carycBcaulis. SEX-CHROMOSOMES AND INHERITANCE 183 queen must transmit to him all her characters, thus giving rise to a form of inheritance that has a superficial resem- blance to sex-linked inheritance. A queen of a pure race, bred to a male of another race with a dominant factor, produces daughters all showing the dominant character of the father, and sons all showing the recessive character of the mother. Since the son gets his entire chromosome- complex from his mother, he must necessarily be like her, whether the character in question is in the sex-chromo- some, or in some other one. CO (D I— ,• ^ ~§ 2.P o t; o 3 o-r^ (H re i-i 3 O p 3 re n 3 1 o D-3- ■^ »> " R's ?-3':5 >< 3-^ CTre re H p i-» 1 3 O 3 re C re >-' re Q.O CO O ?3 OG- B P p 3 3 CL 3 P 3X^! p ■-! Q.re c*- o 3 C 3 3 g re o " • CD o B re O p B ^ n Pre 2.2 C- b'I"^ p ^. re ^^ p re IR '^ w 5 o S «> 3 '-'B re fi> 3 (R fi- re rs.; re p 5' (R Br- as '^2 "^ 2 B S'p <: re o oi B"^ "^ x-^ t^p p 3 o B O (R re 3 "^ = re £T^.re 3 re re (R P cr a- o 3 re 3 O 3 3 re P B P 3 S 3 a o o o — ^o cc re 1-j re re >^ re 2 P B-a'' -J- ►— rt- ^p O =3 2 p re c CK p ^^ C-re O B-^ 73 re P O 00 3 c'vi p ►-.(R . CO CO re cB re _■.- '■" at K— B^p b2£:p 0) •o O B p re re o a c o re ►* re 3 B^ >-•>->,< -■3 re „. re B B re o B tTc ^a-r- B re 2- 3"ffi re SL « 3 ^S B P 3t3 ^B £.re ^ - -'^ ^2 CB -^ 3 2.P 3 • B.p re ~<; re - I • ^.v I SEX-CHEOMOSOMBS AND INHERITANCE 191 character. It, or its normal allelomorph, should therefore be present in both nuclei if all the chromosomes of the fertilized egg have divided normally except the X-chromo- somes. This, in fact, has been found to be the case (Mor- gan, Bridges, Sturtevant). Nearly all of the many hybrid gynandromorjjhs of Drosophila can be explained as above. In a few cases, when the abdomen of the fly was sufficiently female to make mating possible, it has been found that the eggs give the results expected for a female having the sex linked factors that entered the cross. FiQ. 88. — Diagram showing elimination of X' at an early cell-division, so that the nucleus to the right gets X and X' and that to the left only X. In a few cases in Drosophila the explanation of chro- mosomal dislocation will not cover the results. Some of these cases can, however, be accounted for by another hypothesis. Should an egg arise with two nuclei (there are several possible ways for this to occur), one nucleus having one set of factors, the other the other set (the parent being heterozygous), then if each nucleus is sepa- rately fertilized a different combination of factors is pos- sible from that possible on the elimination theory. A gynandromorph, described by Toyama, appears to belong to this category. Toyama found two gynandromorphs of the silkworm (Fig. 89) whose mother belonged to a race with banded caterpillars, and whose father belonged to a 192 PHYSICAL BASIS OF HEREDITY race with pale caterpillars. One of these was banded on the left side (which side was also female) and pale on the right side (which was also male). The sex of the two sides was only apparent after the moth had appeared. The banded character of the worm is kno^^^n to be dominant to the pale character, but neither is sex-linked. The case can be explained, if as the evidence indicates, the mother was ^V^ rc t) -M '%pri striped gynandromorph plain Fig. 89. — Caterpillars of the silkworm moth. A striped one to the left, a plain one to the right, a hybrid gynandromorph in the middle. heterozygous for a not sex-linked character, banded, and if she produced an egg with two nuclei (Fig. 90). Don- caster has found such eggs in Abraxas, and has shown that each nucleus extrudes separately polar bodies, and that each reduced egg nucleus is fertilized by a separate spermatozoon. If as shown in the next diagram one reduced nucleus has a TF-chromosome, and a factor for banded carried in one of the autosomes, and the other reduced nucleus has a Z-chromosome, and in one of the SEX-CHROMOSOMBS AND INHERITANCE 193 autosomes a factor for pale, and if a spermatozoon, carry- ing the factor for pale, fertilizes each nucleus, the two zygotic nuclei will be ZW female and banded, and ZZ male and pale. This gives at least a formal explanation of the results, and helps us to see how such a rare event, the appearance of two gynandromorphs in the same brood, happened to occur at the same time ; because, as Doncas- ter's evidence shows, a double nuclear condition may be characteristic of the eggs of certain females. Fia. 90. — Diagram illustrating how a heterozygous egg with two nuclei fertilized by two sperms might produce a gynandromorph like that shown in Fig. 89. * ' Intebsexes ' ' AND Sex Genes The quantitative relation of one X for male and two X's for female that has been found to hold in many of the groups of animals might seem from a purely a priori point of view capable of being modified in such a way that an intermediate condition might be realized, but whether such conditions should be expected to give rise to her- maphrodites or to non-sex- somethings (intermediates) — or to a mosaic of both sexes, or should rather be expected to die could scarcely be foretold. There are three cases in which individuals called ^ ' intersexes ' ' have been found, or produced; and since their interpretation has led to a view that has appeared to contradict the ordinaiy sex- determination scheme, these cases must be briefly referred to here. Gold«chmidt has studied very thoroughly ' ' inter- 13 194 PHYSICAL BASIS OF HEEEDITY sexes ^' that arise when the European and Japanese race of gypsy moths, Lymantria dispar and L. japonica, are crossed. Eiddle has described doves obtained by crossing the white ring dove (Streptopelia alba) and the Japanese turtle dove (Turtur orieiUalis) that are intersexual in their mating habits. Olga Kuttner and Banta have found that certain lines of Cladocerans (Simocephalus) may produce (parthenogenetically) '^intersexual individuals'^ in the sense that an individual may possess some of the secondary sexual differences of one sex and some of the other. Some of Goldschmidt's combinations between different races of gypsy moth produce only intersexual females, i.e., individuals that are mostly female, but have also, in spots, male characters. In the most extreme cases they are almost like males, not only in color, but even in the partial production of testes. Other racial combinations give male intersexes, i.e., individuals that are for the most part males, but show, in spots, some of the characteristics of the female. Goldschmidt explains these results by the assumption that the sex factors have different quantita- tive values in the different races. He represents the female by FFMm, and the male by FFMM. If the FF ' ' fac- torial set'' is represented by 80 units, and the '' present" male factor, M, by 60 units, then the above formula for the female becomes 80-60 = + 20, and the male formula becomes 80- (60 + 60)= -40. In the former, female units '' dominate," in the latter, the male. Values like these can be arbitrarily set for all the different races. For instance, to the 'Sveak" European race and the "strong" Japanese the following values are assigned : Weak European Race Strong Japanese Race 9 FF Mm FF Mm 80, 60 100, 80 (J^ FF MM FF MM 80, 60,60 100, 80,80 SEX-CHEOMOSOMBS AND INHERITANCE 195 If a Japanese female is crossed to a European male, the Fi female and male may be represented in the fol- lowing* formula: Fi 9 FF Mm Fi c? FF MM 100, 60 100, 80, 60 Both ^^normaP' female and male offspring are expected in equal numbers. The reciprocal cross gives a different result, vi^.: Fi cf FF MM Fi c? FF MM 80 -(80+60) 80, 80, 60 The Fi female is FF-M==0; and is therefore repre- sented as intersexual. It will be observed that the so- called ^^female factors'^ in these formulas are supposed to be inherited entirely through the mother. By assigning different values to FF and M in the dif- ferent races it is possible to express the results in such a way that the sexes obtained by various crosses have different minimal values — those less or more than any assigned value for a given sex are interpreted as inter- sexes. In the example cited, an exact balance (=0) between the conflicting factors produces an individual that is represented as neither male nor female. It is not obvious, however, why it should be made up of parts each of which is strictly comparable to the same part in a male or a female. While the assignment of arbitrary values to sex fac- tors is a legitimate procedure, it is not a quantitative analysis in the ordinary sense, since the quantities are not referred to some external measure, but are purely arbitrary. How far an erratic elimination of sex-chromosomes in later stages of cell-division might account for the result cannot be stated, since there are at present no facts to go upon — the chromosome count in somatic cells of the hybrid has not yet been reported, but Goldschmidt thinks 196 PHYSICAL BASIS OF HEREDITY that the mode of development of the embryo precludes this interpretation. Riddle obtained his intersexual hybrids by causing their mother to produce many more eggs than she would ordinarily produce. This was done by removing the eggs from the nest as soon as they were laid. Towards the end of a series obtained in this way an overworked female produced an excess of males. Some of these males Riddle regards as females that have been changed into males — the completeness of the change being shown in their sexual behavior towards other males, etc. But there is involved in the cross a sex-linked factor that behaves, as R. M. Strong had already shown several years ago, as do sex- linked factors in other birds. It is thus possible to identify the chromosomal make-up of Riddle 's intersexual hybrids. His own results show that the hybrids have the expected combination of chromosomes for males. It appears, there- fore, that whatever it may be that affects their behavior their sex is determined by their possessing the ordinary chromosome constitution for males. Hermaphroditism and Sex As has been shown, the sex-mechanism, whether XX- XY or WZ-ZZ, gives rise to two kinds of individuals — males and females. There are, however, many groups and species of animals where both eggs and sperm are found within the same individuals, and in typical cases there are in such individuals special ducts that are outlets for the male germ-cells and others for the female germ-cells. In these hermaphrodites '^sex-chromosomes^' are not known to be present, or if present as in Ascaris nigrovenosa, they act as sex determinants only in alternate generations. The usual interpretation of the determination of the sex-ceUs of hermaphrodites is that their differentiation is determined by the same kind of specific influences that determine, for example, that certain cells of the primitive gut develop into liver cells, others into lung cells, still SEX-CHROMOSOMES AND INHERITANCE 197 others into pancreas cells, etc. There is nothing inconsist- ent in such a view with the theory that in other cases a different mechanism produces different kinds of germ- cells. Logically, this viewpoint is consistent, but I can sympathize with efforts that are continually being made to find an explanation that makes use of the same kind of process in genetic segregation and in embryonic differen- tiation. In fact, in 1902, while still under the influence of the then recent advances in the field of experimental em- bryology (developmental mechanics), I suggested that one might attempt to treat the phenomenon of segregation from the same theoretical standpoint {viz., the realization of alternative states) as was then appealed to for embry- onic differentiation. It soon became apparent to me, how- ever, that (1) the two kinds of results depended upon entirely different situations, and therefore need not have a common explanation ; (2) that the genetic evidence showed the improbability of explaining segregation and differ- entiation in the same way; (3) that special tests that I carried out failed to support the supposition of a common explanation; (4) that while no detailed explanation is possible at present for the general phenomena of specific differentiation, yet for Mendelian segregation the reduc- tion division supplies all that the results call for. Sex Ratios The theory of sex-determination has been deduced from the evidence of equality of males and females as well as from the cytological evidence. It remains to explain why in some cases the machine fails to give equality of the two sexes ; why, for example, all fertilized eggs of phylloxerans and aphids, or daphnians, or roti- fers, or bees, are female ; why certain mutant races of flies give twice as many daughters as sons; why other races of flies produce nearly all sons ; why the sex ratio in man is about 106 males to 100 females. It is perhaps needless to point out that if, in a species 198 PHYSICAL BASIS OF HEREDITY in which sex is determined by a chromosome mechanism, it were possible to change the sex by other agencies in spite of the chromosome arrangement, the latter relation would be entirely thrown out of gear and males would transmit sex-linked characters and sex itself like females, and females like males. As no such cases have been found, it is futile to discuss such a possibility. It has been shown that only the female-producing sperm in phylloxerans and aphids becomes functional, hence it is obvious why all the fertilized eggs develop into females. In daphnians and other Crustacea it is not known whether one class of spermatozoa degenerates, but the results are explicable on such a view. In rotifers the production of males only by certain females is due to the eggs developing by parthenogenesis with the haploid number of chromosomes and this explains also the case of the bees, wasps and other hymenoptera. If a queen bee is unfertilized or if her supply of sperm gives out she produces only males. If she contains sperms, then any egg that is fertilized produces a female, and as Petrunke- witch showed several years ago, spermatozoa are to be found in eggs laid in worker cells — such eggs being known to produce workers (9 9). In rotifers, too, the presence of a large and a small class of sperm suggests that only the former is functional. Certain females of Drosophila give a sex ratio of two females to one male. By making such a female hetero- zygous as to her X-chromosomes (each carrying different factors) it can be determined that the half of the expected sons that die are the ones containing one of these two chromosomes. It is easily possible by means of linked genes to locate a factor in the sex-chromosome (Fig. 91) and to show that whenever it goes to a male the fly dies. All the daughters survive because the lethal factor being recessive does not harm a female whose other_ chromo- some comes from a normal father. The scheme is shown on the next page. SEX-CHROMOSOMES AND INHERITANOE 199 As many as 20 different lethals have been found in the Z-chromosomes of Drosophila. Their occurrence in these chromosomes is first noticed by the appearance of such exceptional sex ratios. Lethal factors like these need not be thought of as different in kind from any other mutant factors. They may mean only that the changes that they cause are of such a kind, structural or physiological, that the affected individual cannot develop normally. Some of the lethals may be fatal in Fig. 91. — Scheme showing the transmission of a lethal sex-linked factor in an X-chromoiome the black one in the diagram. the egg stages, others are known to cause the death of the larvae, others probably act on the pupae, and a few- even allow an affected male to occasionally come through. In man and in several other mammals there is at birth a slight excess of males over females. Since male babies die oftener than females, the difference has been said to be an '^adaptation,'' with the implication that it calls for no further explanation. Several possible solutions suggest themselves. The male-producing sperm bearing the sex- chromosome may more frequently develop abnormally than the female-producing sperm. Again, since the sper- matozoa must, by their own activity, travel the entire 200 PPIYSICAL BASIS OF HEREDITY length of the oviduct to reach the egg as it enters the tube, the greater size or weight of the female-producing spemi may give a slight advantage to the male-produc- ing sperm in the long trip up the tube. This would lead to an excess of males. There are still other possibilities, which if realized, would suffice to slightly change the equal- ity of the output of the machine. ' NoN-DISJUNCTION Females of Drosophila are occasionally found that give exceptional breeding results which have been explained by Bridges on the view that these females are FEMALE MALE IXT FEMALE FiQ. 92. — Normal female and male groups of chromosomes of the vinegar fly, with the XXY female group below. XXY individuals (Fig. 92). It has been shown by cyto- logical examination that such females do actually contain an additional T-ehromosome. The four possible ways in which these three chromosomes might be expected to behave at the reduction division when the polar bodies il[ SEX-CHROMOSOMES AND INHERITANCE 201 are given off by the egg are shown in the next diagram (Fig. 93). One X may go out of the egg, and the other X and the Y stay in ; or one X may stay in the egg and the other X and the Y go out. In these hvo cases, X and X may be thought of as members of a pair that conjugate, as in the normal female, and then separate, and chance alone determines whether the Y stays in or passes out. Again Y may go out of the egg and X and X stay in ; or X and X POLAR BODY EGGS SPERM 4 Fig. 93. — Non-disjunction. In the upper part of the figure the four possible modes of elimination of the sex-chromosome from XXY eggs are shown; the results of their fertili- zation by an X-bearing sperm of the male is shown below. go out and Y stay in. Here X and Y may be supposed to be members of the conjugating pair, and the free X goes to the same pole as the X that conjugated. In the diagram, each of these four types of eggs is represented as fertilized by an X-bearing sperm. In order to make the outcome more apparent the original XXF female may be supposed to have had white eyes (clear X^s) and the male that fertilized her red eyes (here repre- sented by the black X carrying the gene for red eyes). Four classes of individuals are expected : (1) Rod-eyed females (XXZ) ; (2) red-eyed females (XX) ; (3) red-eyed 202 PHYSICAL BASIS OF HEREDITY females (XXX) that die, and (4) red-eyed males (XY). The last are exceptional, since white-eyed females nor- mally never produce anything but white-eyed sons. Here the exceptional male is due to an egg without an X, being fertilized by a '^female-producing" (or X-bearing) sperm. The three X individuals have never been found, and undoubtedly die, presumably from too many X's. The remaining red females are of two kinds, one normal XX FOLAP BODY ECCS SPElBIi WHITE d \X/HlTEd WHITE 9 (EXCEPTION) DIl 5 6 7 8 Fig. 94. — Non-disjunction. In the upper part of the figure the four possible modes of elimination of the sex-chromosome from the XXY eggs are shown, ancl the results of their fertilization by a Y-bearing sperm of the male is shown below. and the other (XXF), which is expected to repeat the exceptional behavior of her mother. In fact, this is what she does. In the next diagram (Fig. 94) the fate of the same four kinds of eggs is shown if they are fertilized by a F-bearing sperm. Four classes of individuals are expected (5) white males (XYY) ; (6) white males (XY) ; (7) white females (XXZ) ; and (8) YY individuals. No individuals having the last make-up have ever been found, and there can be no doubt that an individual without at least one X dies. The white-eyed females are exceptional, since white-eyed SEX-CHROMOSOMES AND INHERITANCE 203 mothers by red-eyed fathers have normally only red-eyed daughters. These exceptional white-eyed females (XXY) must repeat the phenomena of non-disjunction, and it has been found that they do so invariably. The white-eyed male XY is normal ; the other male should produce some XY sperm and thus transmit both X and Y to some of his daughters. Such daughters as get both X and Y from the entering sperm should show non-disjunction. This has been proven to occur. An analysis of the data has shown that two of the four types of eggs are more common than the other two. As indicated in both diagrams the types of eggs that result after X and X have united occurs in 92 per cent, of the cases, and since in this type the unmated Y has a random distribution, the XY egg is found in 46 per cent, of cases and the X egg in 46 per cent. The more uncommon type of egg would be expected to result if X and Y united and then separated while the other X had a random distribu- tion.^ Eight per cent, of such cases occur, giving XX eggs in 4 per cent., and Y eggs in the other 4 per cent, of cases. These results not only furnish very strong proof of the chromosome theory of sex, but serve also to show how a knowledge of the actual mechanism involved leads to the discovery of how a change in the mechanism gives a new output. The conclusion that females behaving in this way must contain a Z-chromosome was confirmed by the cytological demonstration that showed in them two X^s and a Y. ^ Since this was written it has been found that after X Y ajniapsis the free X always goes to the same pole as the synapsed X. CHAPTER XV PARTHENOGENESIS AND PURE LINES In so far as partheiiogenetic reproduction takes place without reduction in number of the chromosomes, the expectation for any character is that it will have the same frequency distribution in successive generations, because the chromosome group is identical in each generation. There are a few cases where parthenogenetic inheritance has been studied. The results conform to expectation. The only difference between a species reproducing by diploid parthenogenesis and one propagating vegetatively is that in the latter a group of cells starts the new genera- tion and in the former only one cell, viz., an egg, that no longer undergoes reduction, or needs to be fertilized. In both, the chromosome complex remains the same as in the parent. Strictly analogous to the two foregoing methods of propagation are the cases of sexual reproduction in a homozygous group of individuals, composed of males and females or in a group of hermaphroditic forms that are homozygous. Successive generations are here also expected to have the same frequency distribution, whether selected or not, because they have the same germ-plasm. Johannsen's pure lines furnish an example of the last case, for, in principle, pure lines, parthenogenetic repro- duction, and vegetative propagation, are concerned with nearly the same situation. Johannsen worked with one of the garden beans {Phaseolus vulgaris) taking the weight of the seeds, in some cases, and measuring their sizes in other cases. It is known that this bean regularly fertilizes itself. As a consequence of self-fertilization there is a tendency for the descendants of any form to become in time homozy- gous, even when heterozygous forms were present at first. 204 PARTHENOGENESIS AND PURE LINES 205 In fact, in a few generations perpetuated by self-fertiliza- tion with chance elimination of individuals, a homozygous race will result. This comes about as follows: Starting with a heterozygous hermaphroditic individual, some of its offspring will, through recombination of factors, become homozygous, and if self-fertilization prevails they will continue homozygous ; other offspring will be hetero- zygous. From the latter both homo- and hetero-zygous offspring will again be produced, the former remaining such in later generations, the latter continuing the process of splitting. Since only a part of each generation sur- vives, there is in the long run a better chance that the homozygous individuals will be the survivors, because those that have become such in each generation are fixed, and those that are not will continue to produce some homozygotes. There will be in consequence a steady proc- ess of recurrence of homozygotes which, on chance alone, will sooner or later win out. The beans that Johannsen worked with had apparently reached a homozygous condition, and at the start there must have been several such lines. He studied nineteen of them. The offspring of any one plant produced beans that gave the same frequency distribution as the beans of the last generation. This condition continued through all successive generations. It is to be noted that the beans on any one plant differ in size, but any one will give the same frequency distribution as the beans of the preceding generation. It made no difference whether the larger or the smaller beans were chosen for planting — they gave the same group in the next generation. It is interesting to compare this result with what would have happened had the beans been propagating by cross- fertilization at the time when Johannsen began his work with them. If this had been their nonnal method of reproduction they would probably have been heterozygous at the start, and would have given different genetic types for several generations, even if self -fertilized. Pure linos 206 PHYSICAL BASIS OF HEREDITY would have appeared only after the beans had become homozygous through repeated inbreeding. But Johann- sen, starting with homozygous beans, was able to obtain his extremely important results, because if selection could bring about any change it would have to be due to a change in the genes themselves. Here, by means of a crucial experiment, he exposed an error that had been accepted by selectionists from 1859 to 1903. It would have been difficult, almost impossible, to give this demonstra- tion on an}" plant or animal in which self-fertilization or asexual reproduction was not the rule ; for, if the material had been heterozygous either for the main factors for a character, or for modifying factors for that character, selection in one or another direction would be expected through recombination of factors to change the original frequency distribution. It is true that any stock, even such as reproduces by males and females, may be made homozygous by inbreeding brother and sister for ten or more generations, but even such stock would have to be constantly watched for mutation. Johannsen defined a pure line as a race or family of individuals descended through an unbroken series of self- fertilizations from an ancestor homozygous in all its genes. By making this definition precise he made clear the essential point of his demonstration. Now" that his point is made, it seems no longer necessary or even desir- able, I think, to narrow the definition of a pure line to races that self-fertilize, since this is only one form of inbreeding, resulting in the production of homozygous individuals. By extending the definition of a pure line to all forms whose genes are the same in all individuals (whether the pairs are homozygous or not), the definition covers all cases of parthenogenesis that do not undergo reduction, and all cases propagating by non-sexual means, for, in these cases the same complex of genes is present in successive generations. Many plants are propagated by offshoots, stolons, PARTHENOGENESIS AND PURE LINES 207 tubers, cuttings, etc. East has studied the effect of selec- tion of tubers of certain races of the common potato. A race was first grown from a single tuber. By boring holes into the tubers enough material could be obtained for a chemical test of the amount of nitrogen in them. The rest of each tuber could, if desired, be cut into pieces of standard size and planted. Ten tubers, high in nitrogen, and ten, low in nitrogen, were selected. The tubers of the next generation showed that there was no relation found between the amount of nitrogen in the original tuber and in those that came from it. A repetition of the experi- Fia. 95. — A wingless aphid to the left and a winged to the right, both belonging to the same species. (After Webster and Phillips.) ment in another generation gave only meagre results owing to drought. As far as the facts went, this genera- tion, too, showed no effect of selection. Most of the protozoa propagate by dividing into equal or nearly equal parts — i.e., by a process of cell-division. Jennings has studied the effect of selection in a culture of Paramecium, all members of which had descended from a single individual. No change was induced. Later, how- ever, working on another protozoon, Difflugia corona, Jennings found that selection brought about changes in the direction of selection. In this case, the method of division may possibly include irregular distribution of the chromatin material, and the recent work of Hegner indi- cates that such an interpretation is not improbable. Pos- 208 PHYSICAL BASIS OF HEREDITY sibly, too, the irregular distribution of ehromatiii par- ticles (chromidia) in the cytoplasm — aside from the nuclear phenomena, or in connection with them — may make the results similar in certain aspects to the distri- bution of plastids in certain plant cells. Many species of plant lice — aphids — (Fig. 95, a) propagate throughout the summer by parthenogenesis. There is no chromosomal reduction during the develop- ment of the egg. Each egg gives off only one polar body. ^3 hi h5 ^^ ^. / N> _4 ^} \. \^ ^ ^ V / V .-^ r> \ /' \ H A \^ ^ N, V ^.^ / > Ln ^j^ /" \ / ( V X ^ J\ y ^ ■-^^ > / V^ / V UO I.5i 1 HO \ ^^' r^ - \X / 1 .—— • t 1 \ s r^ / \ ^ V ^ \ # \ * 1 *^ " \ \\ \ /^ ^' '^ 1 ^ ' \ 1 — -. ^1> / 1.10 J.JO I 0( , '' y ^-^ f '' / ' % \ ■ 1 ■ ■ 0 90 o.ito ? 9 6 \H 13 20 22 ii 5 II (0 (8 5 16 IN 10 13 Fig. 96. — Curve showing the non-efFect of selection for the first twelve generations for increase in body length, the heavy solid lines represent the fluctuations of the fraternal means; the light solid line the fluctuations of the longest variant; the broken line the fluctuation of the shortest variant. (After Ewing.) each chromosome splitting into two daughter chromo- somes, so that the egg retains the whole number of chromo- somes. Ewing has carried out an extensive experiment with Aphis avence, selecting individuals through a num- ber of generations for the length of the cornicles (honey- dew tubes), for the length of the antennas, and for body length. Considering here only the last, individuals were selected for forty-four generations in a plus and in a minus direction. The graph for the fourty-fourth to the sixty-third generation is shown in Fig. 96. The heavy solid line represents the fluctuations of the longest vari- PARTHENOGENESIS AND PURE LINES 209 ants, the broken line the fluctuations of the shortest variants. It was found that much of the fluctuation observed was connected with temperature. The tempera- ture was therefore kept constant at about 65° F. for the next twenty generations, and as shown in Fig. 97, the fluctuation in the fraternal line was cut down. No in- fluence of the selection is observable in the chart. This evidence, in conjunction with that for other characters, shows that no change takes place in the characters of 63 C* 65 66 67 6i 69 70 7J 71 73 7'/ 75 76 77 78 79 20 ?I n IM —Curve showing the effect of selection for the second score of generations. (See Fig. 96.) the insect so long as the same group of chromosomes remains. It would be difficult to find a better example than these parthenogenetic insects to test the claim that selection can change the germ-plasm, for here the con- ditions are even simpler than in unisexual forms unless they have first been made homozygous. The aphids also furnish favorable material to illus- trate how the environment may cause very great changes, even when the genetic complex remains the same. The parthenogenetic aphids appear often as winged individ- uals (Fig. 95, 6). There is an entire change in structure involving practically every part of the body. The winged 14 210 PHYSICAL BASIS OF HEREDITY and wingless individuals may differ more strikingly than do species of the same genus. The winged forms arising from the wingless produce wingless forms again in the next generation that may be identical with those from which they came. It has long been believed that environ- mental influences bring about these transitions in aphids, but only recently has critical evidence been obtained. The clearest evidence is that of Shinji, with the rose aphid. By sticking twigs of the rose in sand and flooding the sand with water containing substances in solution — a method first suggested by W. T. Clarke — the fluid being drawn up into the leaves is sucked out by the aphids on the leaves. As the following table shows, young aphids reared on the Winged Apterous IndividualB. Individuals AgNO, 51 0 CuSo, 34 1 HgCl, 31 6 NiSO* 955 5 SbCla 41 5 PbCl, 12 2 SnCl. 579 8 ZnClj 49 2 Mg salts 840 9 Sugar 365 160 Alcohol 2 288 Alum 3 34 Acetic acid 0 67 Na salts 2 1029 Ca salts 1 433 K salts 3 324 Sr Saltsi 1 220 Tannin 1 14 Urea 5 153 Water, distilled 0 394 Water, tap and creek 17 461 Peptone 15 salts of the heavy metals as well as on magnesium salts and sugar became winged, while those reared on the other substances in this list remain apterous. Here we have an excellent example of how in one environment a given germ-plasm produces one result, and in another environ- ment a different result without any intermediate forms. PARTHENOGENESIS AND PURE LINES 211 The change from wingless to winged aphids is far greater than most mutational changes that we know, yet must involve a different kind of change because the result is reversible, while a mutation, having once taken place, is relatively irreversible. Summing up, it may be said that the evidence shows that whenever the same chromosomal complex containing the same genes is found, the measurements of any charac- ter in successive generations show the same frequency distributions of the measurements, and the form may be said in a general sense to belong to a pure line. The evidence shows that whether the chromosomal complex is heterozygous or homozygous, the results are the same, so far as the pure line is concerned ; but it is also obvious that in most animals and plants, where redistribution (reduction) of the chromosomes takes place in each gen- eration, only forms already homozygous will give pure lines. This was the special feature of the material that Johannsen worked with, but aside from its practical value in studying the selection problem, the limitation of the definition of pure lines to such an exceptional situation leaves out of sight the wider bearing of the evidence. CHAPTER XVI THE EMBKYOLOGICAL AND CYTOLOGICAL EVI- DENCE THAT THE CHROMOSOMES ARE THE BEARERS OF THE HEREDITARY UNITS Long before the genetic evidence brought forward its abundant data that are explicable on the theory that the chromosomes carry the genes, embryologists had already found other evidence that led them to regard the chromo- somes as the bearers of the hereditary factors. Taken as a whole, this evidence makes out a very strong case for the chromosomes, but since it did not establish the relation beyond question, the genetic evidence was all the more welcome. The earliest evidence, sometimes cited in favor of chromosomal inheritance, was based on the statements that in some cases at least, only the head of the spermato- zoon enters the egg. Since it was then thought that the head is composed almost entirely of the nucleus, and since the child inherits equally (in the older parlance) from its father and from its mother, it followed that the nucleus carries the hereditary elements. When later it became known that the head of the sperm represents almost exclusively the mass of condensed chromatin, it was sup- posed that the chromosomes, in particular, must be that part of the nucleus that is the bearer of hereditary charac- ters. Such a conclusion received indirect support from the facts, then becoming kno^\Ti, that the chromosomes remain constant through successive generations of cells, whereas the nuclear sap becomes lost in the gen- eral cytoplasm each time that the nuclear wall is dis- solved. It was also found that the spindle fibres disappear in the resting: stages, while the nuclear reticulum (chro- matin) remains. 212 BEAEERS OF HEREDITARY UNITS 213 This evidence failed, however, in so far as there might be present a certain amount of nuclear plasm in the sperm- head that is carried in with the head, and if so, would be later mixed with the egg cytoplasm. The discovery that at the base of the sperm-head there is present in some eggs a centrosome that becomes, through division, the dynamic centre of the next division, opened the door to suspicion that the sperm might bring in other things than the chro- mosomes to influence development, and hence heredity. In conclusion then, while it may be said that the evi- dence that the sperm-head alone enters the egg may be claimed as favorable for the chromosome view^, it cannot be accepted as critical proof, because it is uncertain wdiether other things also may not be brought in besides the chromatin of the sperm. Boveri^s evidence for chromosomal heredity from di- spermic sea urchin eggs v/as ojjen to less objection. It w^as known that when two sperms enter the sea urchin's egg- simultaneously, the first division of the egg is into three or into four parts, because four (instead of two) division- centres appear in these dispermic eggs. It was also known that these eggs rarely produce normal embryos or larvae. Boveri, studying the mode of division of the dispermic eggs, found that there was an irregular distribution of the chromosomes to the three or four poles that appear, and consequently to the three or four resulting cells (i^'ig. 98). The abnormal development of the whole egg that generally follow^s might be ascribed to the irregular dis- tribution of chromosomes to different regions; for, quite apart from the specific nature of each chromosome or group of chromosomes, the activity of one region being quantitatively different from that of a corresponding region in another part of the egg might be responsible for the failure to develop normally. But Boveri went further in his analysis. He shook apart the three or four blasto- meres coming from dispermic eggs (by using Herbst's calcium-free sea-water method), and compared the num- 214 PHYSICAL BASIS OF HEEEDITY ber that developed into normal plutei with the number of plutei from one-fourth normally fertilized blastomeres. From the latter a large proportion give rise to normal embryos, from tlie former normal embryos are rarer. Their greater rarity, Boveri thought safe to attribute to the chromosomal deficiencies present in most of such iso- FiQ. 98. — Scheme showing dispermic fertilization of the egg of the sea urchin with the subsequent irregular distribution of the chromosomes. (After Boveri.) lated blastomeres. He suggested that the chance of a blastomere developing normally depends on its having at least one full set of chromosomes. For these triploid sea urchin eggs with three times 18 chromosomes, the chance of one full set of chromosomes getting into each blastomere is, according to Boveri ^s calculation, only one to 10,000. The chance of getting at least one chromosome of each kind in one cell is greater. He concluded that the few embryos he obtained came from quadrants that had at least one haploid set of chromosomes. There is, however, BEARERS OF HEREDITARY UNITS 215 to-day some uncertainty concerning the assumption that normal development is to be expected if in adflition to one haploid set of chromosomes other chromosomes arc also present, because while one set alone might permit normal development, it is by no means certain that if there were one, two, or more additional chromosomes, the balance might not be upset and abnormal development fol- low. On chance distribution alone the isohition of just one set and no more would seem a very remote possibility, I Vi/ii ,t^^^ ^n^i X 1^' a Fig. 99. — First division of a hybrid egg showing the elimination of chromosomes at the equation of the spindle, o. The reciprocal cross, 6, shows no such elimination. (After Baltzer). but if there is to some degree a tendency for a group of daughter chromosomes to move off together as a result of their method of division, there might be a better chance of such a group getting into one of the three or four blastomeres than by chance distribution alone. At pres- ent it is not possible to make any calculation based on such an assumption. While, therefore, Boveri^s argument can- not be accepted as demonstrative, yet it has probability in its favor. Baltzer has found a different kind of evidence of chromosomal influence. When the eggs of one sea urchin. 216 PHYSICAL BASIS OF HEEEDITY Strongylocentrohis, are fertilized by the sperm of another sea urchin, SpliaerecJdnus, the segmentation nucleus, formed by the union of the a^^- and sperm-nucleus shows irregularities in the movements of the daughter chromo- somes to the poles of the spindle. While some of the chromosomes after dividing pass normally to the poles, others become scattered irregularly between the two poles and fail to become incorporated in the two-daughter nuclei (Fig. 99,(2). They appear to become lost and take no ^wmmr-' ' -mm^ ^^ ///iUV ^ ' / ■■ 1 \ \ -^ a a Fig. 100. — Fertilization of an egg that had started to develop parthenogenetically. The belated sperm unites with one of the daughter chromosomes groups only, a; an earlier condition of the same proeedure. (After Herbst.) part in the further development. Counts of the chromo- some plates in the later divisions of the e^g give about 21 chromosomes, whereas 36 are expected as the whole number. It appears that 15 chromosomes are lost, and presumably they belong to the foreign sperm. Many of these eggs develop abnormally, but those that reach the pluteus stage show a maternal skeleton only. This seems to mean that the sperm has done no more than start the development. It has contributed nothing, or little, to the embrj'-o, and it seems reasonable to attribute this to the BiEARERS OF HEREDITARY UNITS 217 loss of the paternal chromosomes, especially in the light of the reciprocal cross. In this reciprocal cross, the egg oif Splicer ecUnm is fertilized by the sperm of Strongylocentrotus. All the chromosomes of the segnientation nucleus divide and pass regularly to the two poles (Fig. 99, &). The hybrid embryo shows characters of both parental species. "*■-"•-- -. " "^ /'o "oON 0 0 oo VOn O 0 0 O,^' \ O; i o O; Fig. 101. — Larval sea urchin seen in side view. On one side it shows hybrid characters, on the other side it is maternal. The sizes of the nuclei on these two sides, as seen in the figure, coincide with the view that the hybrid side is diploid and the maternal side haploid. After Herbst.) The difference in the two cases can be safely attributed to the observed differences in the fate of the chromosomes, rather than to unrecognized ditferences in other elements brought in by the sperms. Herbst 's experiments contribute further evidence in favor of the chromosome interpretation. He caused the unfertilized eggs of a sea urchin to begin to develop parthenogenetically by adding a little acid to the sea water. After five minutes the eggs w^ere removed to pure sea water, and sperm of another species, Stronciylocen- 218 PHYSICAL BASIS OF HEREDITY trotiiSj was added. The sperm entering the egg after its nucleus had started to divide, failed to reach the egg nucleus until the latter had divided (Fig. 100). The sperm nucleus then formed a nucleus of its own, that passed into one only of the daughter cells. This cell got two nuclei. The other cell had only one of the daughter nuclei. Such half-fertilized eggs give rise to larvae that are maternal on one side, and hybrid on the other — or at least larvae of this kind are sometimes found in such cultures (Fig. 101), and Herbst believes it is safe to refer them to the half -fertilized eggs. If so, there can be little doubt that the hybrid half owes its peculiarities to the presence of both sets of chromosomes in its cells, while the maternal half owes its peculiarities to its single set of maternal chromosomes. This in itself, however, shows little more than do whole hybrids and whole parthenogenetic eggs themselves, for the demonstration that it is the chromo- somes and not other constituents of the sperm-nucleus that make the difference in the two sides rests on the unproven inference that if other things than the nucleus are involved they would be distributed equally throughout the cytoplasm, but produce no effects. There is no reason to suppose that they would be so distributed, and no evi- dence that they are. Hence the proof is not cogent, how- ever probable it may seem that only the sperm-nucleus is responsible for those cases where there is a difference in the two sides. On the whole, then, while I am inclined to give much weight to this evidence from experimental embryology as very favorable to the hypothesis that the chromosomes carry the hereditary characters, it is the genetic evidence that furnishes convincing evidence in favor of this view. CHAPTER XVII CYTOPLASMIC INHERITANCE In the preceding pages so much emphasis has heen laid on the chromosomes as bearers of the hereditary material that it may apj)ear that no very important role is left to the rest of the cell. Such an impression would be quite misleading; for the evidence from embryology appears to show that the reactions by means of which the embryo develops, and many physiological processes themselves, reside at the time in the cytoplasm. Further- more, there is also genetic evidence to show that certain forms of inheritance are the outcome of self -perpetuating bodies in the cytoplasm, most of which go under the name of plastids. Recognition of plastid inheritance carries with it the idea that if there are such materials in the cytoplasm that are self-perpetuating they will have to be taken into account in any complete theory of heredity. In the case of certain chlorophyll characters there is excellent genetic evidence to show that a peculiar kind of inheritance is due to the mode of transmission of plastids in the cytoplasm. There is a race of f our-o ^clocks known as Mirabilis Jalapa albomaculata, whose leaves are made up of patches of green and white. Such leaves are said to be checkered (Fig. 102, h). The amount of green, or of white, varies on different leaves, and on such plants there frequently appear leaves and entire branches that are green and others that are white. The white is due to the absence of green in the chlorophyll grains. Some cells have only green chlorophyll bodies, and others only white, still others may have the two mixed in various amounts. Correns has shown that if the flowers on a green branch are self -fertilized they produce only green plants, and these again only green plants. Flowers on wliite 1219 220 PHYSICAL BASIS OF HEREDITY brandies give only white offspring. Flowers on the check- ered branches give some checkered plants, some white plants and some green plants. The proportions in which these different types arise varies according to the amount of green in the branch from which the self-fed seed came. Wlien the ovary of a flower on a green branch is fertil- ized by pollen from a white branch, the plant produced is green like the maternal branch. If the ovary of a flower on a white branch is fertilized by pollen from a Fig. 102. — Gipen leaf and checkered leaf of four-o't-lock. (After Baur.) green branch the offspring is white like the maternal branch. These and other combinations show that this color inheritance is only through the mother. The results are explicable on the assumption that there are normal (green) chlorophyll bodies and abnormal chlorophylless bodies, both kinds propagating in the cytoplasm by divi- sion, and that these two kinds are transmitted only through the egg-cell. The green or white color of the leaves of a given branch is an index of the kind of chlorophyll body that the ovaries will probably contain. At each division of the body-cells the chlorophyll grains present in it are sorted out more or less at random — hence from a cell that CYTOPLASMIC INHERITANCE 221 contains both kinds, more white granules than green ones may at times get into a cell, and at other times only white granules will get into one daughter cell, so that a wliite branch arises. In other species of plants that have white leaves and branches and green leaves and branches, the cross may give a different result. Thus in Melandrium and Antirrhi- num, green by white gives green F, (whicliever way the cross is made), in F. there are 3 green to 1 white plant. In this case the results can be explained as due to the action of genes in the chromosome on the production of chlorophyll in the cytoplasm— an action of such a kind that Fig. 103. — Pelargouium that gave rise to a white branch. (After Baur.) the granules do not develop green color unless the (nor- mal) gene is present, in single dose at least. In this case, even if the eggs only transmit plastids, the F^ individual from a white-leaved mother by a green-leaved father is green, because the paternal nucleus introduces a gene that causes the green color to develop in the plastids. It is the segregation of the genes in the germ-cells of the F^ individual that leads to the 3 : 1 ratio in F., and not the distribution of the plastids as in the preceding case. The most peculiar case is that of Pelargonium de- scribed by Baur. White leaves and branches, and green leaves and branches occur on the same plant (Fig. 103). Self-fertilized seeds from each breed true to color of branch. White to green gives a different result, viz., 222 PHYSICAL BASIS OF HEREDITY mosaic seedlings with patches of green and white on stems and leaves (Fig. 104). When these seedlings grow into plants, the color of the leaves will depend on the color of that part of the stem from which the terminal bud, and lateral buds grow out. If a bud lies in a green part of the stem the new part will be green (Fig. 104, a) : if the new bud lies in a wliite part of the stem the new part will be white (Fig. 104, c) : and if it lies in a partly green, partly white region the new part will have some white, some Fig. 104. — Diagram to show how a sectorial chimera may be produced. If the ter- minal bud has come from a region of the seedling entirely green, all of the future leaves will be green, o; if from a region without chlorophyll, all the future leaves will be white, c; but if the terminal bud lies partly in one, partly in the other region, some white and some green leaves will arise, b. (After Baur.) green parts (Fig. 104, h). The only explanation that is suggested by Baur is that in this plant the plastids are transmitted both by the egg and by the pollen. The white plant with defective plastids contributes part of the plas- tids in the fertilized egg, the green plant with normal plastids the other part. The fertilized egg contains there- fore both kinds of plastids. During division of the egg and embryo, the granules become irregularly distributed in the cells. Whenever a cell gets only defective granules, it and its descendants are white, producing white parts : when a cell gets mostly or only green granules, it and its descendants are green, producing green parts. Hence CYTOPLASMIC INHERITANCE 223 arise the checkered seedlings from which white or green branches grow out. The preceding facts and theories relating to plastid inheritance show that if any element outside the nucleus has the power to propagate itself it may be transmitted through the egg, and even possibly through the sperm (pollen) also. There is no contradiction here iii any sense to Mendelian inheritance but only an additional type of inheritance that can be studied by as exact methods as those used in Mendelian work. The chief difference between chromosomal and plastid inheritance lies in the orderly sequence of the distribution of the genes in all divisions by means of the mitotic figure, whereas the plas- tids are supposed to be shuffled about at random to the daughter cells (partly because their division period does not correspond with that of the cell). This haphazard distribution of the plastids at any and all divisions is in striking contrast to the sorting out of the genes that occurs only at one specific cell-division when the germ-cells pass through the maturation stage. Hence the orderliness of Mendelian inheritance as contrasted with the more irregu- lar procedure in plastid inheritance. To embryologists familiar with the fact that differen- tiation of the egg is closely associated with the cleavage pattern, it was a natural inference that in the cytoplasm lay the inherited characteristics that gave form to the embryo, and even to all of its essential features. Little room would seem to be left for the action of the chromo- somes except to fill in the details of the characters already outlined by cytoplasmic activity. This view might l)e la- conically referred to as the theory of the ^'Embr^^o in the Rough, '^ or more generally as the *' Theory of the Organ- ism as a Whole.'' Boveri discussed some such view (1903), and at first considered it favorably. It has since been seriously discussed by others. Boveri pointed out that when a horse is crossed to an ass it makes no differ- ence which way the cross is made, for both egg and sperm 224 PHYSICAL BASIS OF HEREDITY bring' in the characteristics that make the organism first a bilateral one, then a vertebrate, then a mammal, and, lastly, a perissodactyl. In all these aspects, both parents agree, and beyond these limits hybridizing is impossible. Whatever the germ develops into must contain these com- mon characters. The important point to determine, Boveri thought, is whether the species characteristics are or are not in the nucleus. He concluded, after discussing the pros and cons, that it is doubtful if these preformed qualities of the egg-protoplasm extend beyond the larval periods, but that in general all characteristics that distin- guish the individual from all others of its species and from the characteristics of related species are inherited through the chromosomes. Later he restated his con- clusion as follows: ^^All essential characteristics of the individual and of the species are epigenetic, and the deter- mination is brought about through the nucleus. ' ' Conklin at one time expressed even more sharply the idea that group characteristics may be inherited in a different way from specific characters in the following paragraph: We are vertebrates because our motbei's were vertebrates ajid pro- duced eggs of the vertebrate pattern; but the color of our skin and hair and eyes, our sex, stature, and mental peculiarities were determined by the sperm as well as by the egg from which we came. There is evidence that the chromosomes of the egg and sperm are the seat of the differential factors or determinei's for Mendelian characters, while the general polarity, sjnnmetry and pattern of the embryo are deter- mined by the cytoplasm of the egg. In another statement, however, Conklin takes what seems to me to be more nearly a correct view in regard to the question, viz., that ' ^ There is no doubt that most of the differentiations of the egg cytoplasm have arisen during the ovarian history of the egg, and as a result of the interaction of nucleus and cytoplasm ; but the fact remains that at the time of fertilization the hereditary potencies of the two germ-cells are not equal, all the early stages of development, including the polarity, symmetry, type of CYTOPLASMIC INHERITANCE 225 cleavage, and the pattern, or relative positions and pro- portions of future organs, l)cing- foreshadowed in the cyto- plasm of the egg-cell, while only the differentiations of later development are influenced by the sperm. In short, the egg cytoplasm fixes the general type of development, and the sperm and egg nuclei supply only the details." If, as implied, the egg nucleus at first has already pro- duced its effect on the cytoplasm, it has done something more than supply the details ; and as to the sperm nucleus I should substitute nearly all the stages of development later than the gastrula. Moreover, sex is certainly one of the fundamental characters of the organism, yet it appears to be determined at fertilization by the chromo- somal combination formed at that time. Conklin later abandoned his earlier interpretation. Quite recently, in his book on ^'The Organism as a Whole," Loeb has discussed the question as to whether the protoplasm of the egg is ^Hhe future embryo in the rough, ^ ' the sperm furnishing only the ^ ' individual charac- ters. ' ' Loeb suggests that the ' ' specificity of the species ' ' must be due to their proteins, and that the ' 'heredity of the genus is determined by proteins of a definite constitution ditfering from the proteins of other genera. This consti- tution of the proteins would therefore be responsible for the genus heredity. The different species of a genus have all the same genus proteins, but the proteins of the species of the same genus are apparently different again in chemi- cal constitution and hence may give rise to the specific bio- logical or immunity reactions. ' ' The possible relations of these considerations to heredity are summed up in the following paragraph : It is thu3 doubtful whether or not any of the constituents of the nucleus contribute to the detennination of the species. This in it^ ultimate consequences might lead to the idea that the Mendelian chara-o is like the maternal race. Here we are dealing not so much with maternal inheritance, but rather with a special kind of parthenogenesis. Such eggs, however, rarely go beyond the cleavage stages. The rate of cleavage of an egg fertilized by foreign sperm usually coincides with that of the species to which the egg belongs. Since the cytoplasm of the egg has, prior to fertilization, always been under the influence of its own nucleus, this relation is what might be expected. It is necessary to study eggs from an F^ generation in such cases in order to judge how far paternal chromosomes may influence the cleavage. It is thinkable, for example, that a spermatozoon might bring in a factor dominant for rate of cleavage, but because this factor had not had time to influence the cytoplasm its effect would not show in the Pi cross. In the F^, on the other hand, the paternal char- acter might prove dominant. Both Driesch and Boveri have shown in the sea urchin that the rate of cleavage, the pigmentation, and the kind of gastrulation are entirely or largely determined by the egg — they differ in opinion only as to how soon the influence of the sperm can be seen. At the time when the larval skeleton is formed most observers agree that the influence of the foreign sperm makes itself felt. Most of the accounts of the skeleton of hybrid sea urchins describe it as intermediate in struc- ture, but one that varies widely under different external conditions. Tennent has shown, in fact, that the character of the hybrid larval skeleton is so greatly influenced by the alkalinity or acidity of the sea water that it can be artificially thrown towards one or the other side — mater- nal or paternal. Loeb, King and Moore have attempted to determine whether the larval skeleton has dominant char- acters in certain parts and recessive ones in other parts. They crossed the sea urchins, Stronqyloceyitrotus Fraucis- canus and S. purpureus. Both the straight cross and its 232 PHYSICAL BASIS OF HEREDITY reciprocal showed neither a great predominance of the characters of the paternal race, nor of the maternal race, but rather certain characteristic features of purpuratus and others of Franciscanns. The larval characters appeared to be dominant or recessive taken singly. Until an F2 generation can be raised it is obviously hazardous to speak here of Mendelian dominance and recessiveness of characters that are based on F^ observations alone, especially since it is becoming more and more apparent that many F, characters are more or less intermediate, and there are no general grounds for expecting pure domi- nance or recessiveness. Many crosses have been made between different species of fish, and in some of these the young, at the time of hatch- ing, are maternal. It has generally been supposed that such cases are due to the absorption of the paternal chro- mosomes at the first or at later cleavage stages. Loss of chromosomes has in fact been recorded in several of these cases of maternal inheritance. On the other hand, Miss Pinney's observations, summarized in the following table. Cross Development Results Chromosomal Behavior Ctenolabrus 9 X Fundulus cf" Development cases Early mitotic behavior during gastrulation. is prevailingly nor- mal. Ctenolabrus 9 X Stenoto- mus o^ Many hatching em- Early mitoses are nor- bryos of the mater- mal. nal type. Ctenolabrus 9 X Menidia cf" Advanced development. Early mitoses are nor- mal. Ctenolabrus cf X Fundulus 9 One hatching embryo reported. Many ad- vanced embryos — maternal type. Ctenolabrus cf X Stenoto- mus ^ Development ceases during gastrulation. Ctenolabrus o' X Menidia 9 Two hatching embryos reported. Maternal type. show that the maternal type may appear not only when the early mitoses are abnormal, but in one case at least when they are normal. It is quite possible, therefore, that while Abnormal nuclear be- havior occurs. Abnormal mitosis pre- dominant. Abnormal mitosis is of frequent occurrence. MATERNAL INHERITANCE 233 early loss of the paternal chromosomes may account for some of the cases of maternal embryos, there may also be cases where the chromatin may divide normally but fail to produce any conspicuous effects on the cytoy)lasm sufficiently soon to become apparent in the young fish. In this connection the tobacco crosses described by Groodspeed and Clausen may be recalled. In these cases it was a particular group of chromosomes, regardless of whether it was of paternal or of maternal origin, whose ^^ reaction system'^ dominated in the F, hybrid. CHAPTER XIX THE PARTICULATE THEORY OF HEREDITY AND THE NATURE OF THE GENE The attempt to explain biological phenomena by means of representative particles has often been made in the past. The superficial resemblance of the theory of the gene to some of the older theories, long since abandoned, has furnished the opponents of the Mendelian theory of heredity an opportunity to injure the latter by pretending that the modern idea of the gene is the same as the older ideas of Herbert Spencer concerning physiological units, of Darwin relating to pangenes, and especially of Weis- mann about biophors. There is no need for such con- fusion, for even a little knowledge of the evidence on which the old and the new views rest ought to have sufficed to make evident some important and essential differences. It need not be denied, however, that there is an historical connection between the mediaeval theory of preformation and the particulate theory of heredity. Bonnet, one of the best known adherents of preformation, believed at first in ^' whole '^ germs, but later admitted that pieces of germs might be stowed away in regions of the body likely to be injured. Weismann, also, the most prominent modem adherent of preformation, held that whole germs, ids, are present in the germ-plasm, each standing for a whole organism — each (or most or one ?) becoming unravelled as the embryonic development proceeded. In fact, Weis- mann's entire theory was invented primarily to explain embryonic development rather than genetics. Its connec- tion with the modern idea of the germ-plasm is little more than an analogy — for reduction in Weismann 's original 234 PARTICULATE THEORY OF HEREDITY 235 sense meant the sorting out of the wholes of ancestral germ-plasms with which he peopled the chromosomes.^ The danger of any appeal to a theory of representative particles obviously lies in the ease with which by its means any phenomenon might be accounted for, if the theorizer is allowed to endow the particles with any and all tlie attributes that he wishes to use in his explanation. It was because Bonnet, Spencer, and AYeismann assigned arbitrarily attributes to the ultimate particles of living matter, that these views appear to-day highly speculative. The different kind of evidence to which the modern theory of the gene appeals is what I wish to emphasize here. The Evidence for the Gene The evidence that Mendelian inheritance rests on the distribution of separate elements has already been given. The numerical results obtained in the second generation from any Mendelian cross involving a pair of contrasted characters, find their explanation on the assumption that the two original germ-plasms (or some element in them) separate cleanly in the germ-cells of the F^ hybrid. Tested by back-crossing the assumption is verified. Recombining the Pj, F^, F2 individuals in all possible ways also gives results consistent with the very simple assumption that whatever it is that causes one race to produce one charac- ter, and another race another character, the two separate in the hybrid in such a way that equal numbers of germ- cells of each kind are produced. Up to this point the results do not tell us whether the two germ-plasms separ- ate as wholes — one from the other — or whether only some part or parts behave in this way. But when two or more ^The nominal adoption (1904) toward the end of his career of heredi- tary units in the Mendelian sense did not go deep. Weismann still adhered to his view of dissociation of the ids as their most characteristic feature— the only one in fact for which they were originally invented. The e>'idence on which Mendelian units rest has nothing whatever to do with this cardinal doctrine of Weisraann's teaching. 236 PHYSICAL BASIS OF HEREDITY pairs of contrasted characters are involved in the same cross, we get further information as to the situation. For example, Mendel showed that when peas that are both yellow and round are crossed to peas that are both green and wrinkled, there appear in the F^ generation not only the original combinations, but also recombinations of these, viz., yellow and wrinkled; and green and round (Fig. 10()). Here also the numerical results 9:3:3:1 can be explained on the theory that the representatives of each pair of characters separate in the germ-plasm, and that the separation of each pair is independent of what takes place in the other pair. Obviously it can no longer be whole germ-plasms that separate, but there must be different pairs of elements in the germ-plasm that assort independently of each other. It has been found that this principle of independent assortment may apply to a considerable number of pairs of characters segregating at the same time. The only restriction that is found is in the case of linked pairs of characters. This relation will be considered later. The independent assortment of the pairs of characters proves that the elements that stand for the characters in the two original germ-plasms may separate from each other. If each such pair of characters represented one of the pairs of homologous chromosomes, the evidence, so far considered, would be in accord with the view that the chromosomes were the ultimate units involved in the proc- esses of segregation and assortment. The chromosomes are, as has been shown, independent units in the germ- plasm. But as Drosophila shows, there are many more pairs of characters than there are pairs of chromosomes. It is obvious that if the chromosomes are the ulti- mate units involved, and remain intact, there could be no more independent pairs of characters than there are pairs of chromosomes. In animals and in plants there are no cases known where there are more independent pairs than there are chromosomes, so that, as has been pointed PAETICULATE THEORY OF HEREDITY 237 out in another connection, this evidence may also be appealed to as favorable. The behavior of linked pairs shows, however, that the analysis must be carried further, because, despite linkas^e, the elements that went in together may be separated. The evidence shows that while some linked genes separate almost as freely as do independent genes, so that their linkage to each other can only be safely determined by their relation to certain other genes, other linked genes may separate not once in a hundred times, or even less often. Between these extremes all intermediate linkage values are found. These results indicate that the chromo- somes do not represent the ultimate elements that may be separated out of the original complex (germ-plasm). We are led, then, to the conclusion that there are ele- ments in the germ-plasm that are sorted out independently of one another. The Brosophila evidence shows at least several hundred independent elements, and as new ones still appear as frequently as at first, the indications are that there are many more such elements than those as yet identified. These elements we call genes, and what I wish to insist on is that their presence is directly deducible from the genetic results, quite independently of any further attributes or localizations that we may assign to them. It is this evidence that justifies the theory of partic- ulate inheritance. So far as representative elements in the germ-plasm are concerned, we might be content to rest the case on the preceding analysis of the results ; but recent work has now advanced far enough to tempt us to assign further attri- butes to the genes than those deducible from the preceding analysis alone. Some of these attributes may appear better established than others, but, all together, they give a consistent body of data, and have therefore a certain value and use. 238 PHYSICAL BASIS OF HEEEDITY It has been pointed out that the evidence shows not only that the genes are carried by the chromosomes, but that there may be interchanges between paternally- derived and maternally-derived chromosome pairs. The evidence shows that this interchange is a normal feature of the germ-cell, and not peculiar to hybrids, or to a heterozygous condition of the pairs. This analysis leads then to the view that the gene is a certain amount of material in the chromosome that may separate from the chromosome in which it lies, and be replaced by a corresponding part (and by none other) of the homologous chromosome. It is of fundamental sig- nificance in this connection to recognize that the genes of the pair that interchange do not jump out of one chro- mosome into the other, so to speak, but are changed by the thread breaking as a piece in front of or else behind them, but not in both places at once, as would be the case if only a single pair of allelomorphs were involved each time. That the gene does not stand for the whole length of the chromosome between two other known genes is shown by the fact that new genes arising by mutation in the inter- mediate region do not affect the character of the gene already known. This fact recurring continually in Droso- phila, where new mutations frequently appear, reassures us that the idea of the gene as a very small part of the thread is a legitimate conclusion, even if we can not tell how large or how small that region is. 1. The Manifold Effects of Each Gene If we examine almost any mutant race, such as the race of white -eyed Drosophila, we find that the white eye is only one of the characteristics that such a mutant race shows. The productivity of the individual is also much affected, and the viability is lower than in the wild fly. AU of these peculiarities are found whenever the white eye emerges from a cross, and are not separable from the PARENTS A ^ Fig. 106. — Diagram to show the inheritance of two pairs of Mendelian characters, viz. yellow versus green peas, and round versus wrinkled skin in garden peas. PAETICULATE THEOEY OF HEREDITY 239 white-eyed condition. It follows that whatever it is in the germ-plasm that produces white eyes, also produces other modifications as well, and modifies not only such "superficial" things as color, but also such '* fundamen- tal ^ ' things as productivity and viability. Many examples of this manifold effect are known to students of heredity. It is perhaps not going too far to say that any change in the germ-plasm may produce many kinds of effects on the body. Clearly then the character that we choose to follow in any case is only the most conspicuous or (for purposes of identification) the most striking or convenient modification that is produced. Since, however, these effects always go together, and can be explained by the assumption of a single unit difference in the germ-plasm, the particular difference in the germ-plasm is more sig- nificant than the character chosen as its index. 2. The Variability of the Character is Not Due to the Corresponding Variability of the Gene All characters are variable, but there is at present abundant evidence to show that much of this variability is due to external conditions that the embryo encounters during its development. Such differences as these are not transmitted in kind — they remain only so long as the environment that produces them remains. By inference the gene itself is stable, although the character varies ; yet this point is very difficult to establish. The evidence is becoming stronger nevertheless that the germ-plasm is relatively constant, while the character is variable. 3. Characters That are Indistinguishable May be the Product of Different Genes We find, in experience, that we cannot safely infer from the appearance of the character what gene is pro- ducing it. There are at least three white races of fowls, produced by different genes. We can synthesize white- 240 PHYSICAL BASIS OF HEREDITY eyed flies that are somatically indisting-uisiiable from the ordinary white-eyed race, yet they are the combined prod- uct of several known color-producing genes. The purple eye color of Drosophila is practically indistinguishable from the eye colors maroon and garnet. In a word, we are led again to units in the germ-plasm in our final analysis rather than to the appearance of a character. 4. Inference That Each Character is the Product of Many Genes We find that any one organ of the body (such as an eye, leg, wing) may appear under many forms in different mutant races as a result of changes of genes in the germ- plasm. It is a fair inference, I think, that the normal units — ^the allelomorphs of the mutant genes — also often affect the same part. We have found about 50 different factors that affect eye-color, 15 that affect body-color, and at least 10 factors for length of wing in Drosophila. If, then, it is a fair inference that the units in the wild fly, that behave as Mendelian mates to the mutant genes, also affect the same organ that the mutant gene affects, it follows that many genes, and perhaps a very large num- ber, are involved in the production of each organ of the body. It might perhaps not be a very great exaggeration to say that every gene in the germ-plasm affects several or many parts of the body ; in other words, that the whole germ-plasm is instrumental in producing each and every part of the body. Such a statement may seem at first hearing to amount almost to an abandonment of the particulate conception of heredity, but on the contrary, the statement conveys a very important idea in the modern conception of the nature of the genes and the way they act. The essential point here is that even although each of the organs of the body may he largely a product of the entire germ-plasm, yet this germ-plasm is made up of units that are independent of each other in at least two PAKTICULATE THEORY OF HEREDITY 241 respects, viz., in that each one may change {mutate) with- out the others changing, and in segregation and in crossing over each pcdr is separable from the others. 5. "The Organism as a Whole," or The Collective Action of the Gtenes Several writers have stated their objections to the particulate theory of heredity on the grounds of their belief that the organism is a "whole." If this phrase is intended to mean that there is some sort of an entity or entelechy that directs all processes that go on in each living thing, there is little to be said here, except that this very old idea has not been found profitable as a working hypothesis. It is improbable, however, that many biologists mean to appeal to any such vitalistic agency when they speak of the "organism as a whole," but have rather some other idea in mind. I am inclined to think that certain phenomena of embryonic development are responsible for the slogan of the "organism as a whole. ' ' In the segmentation of the egg the entire chromo- somal complex is distributed to every cell in the body. Each cell inherits the whole germ-plasm. How then it may be asked can the result depend on the particular make-up of its chromosomes rather than on the action of the whole material! G-ranted that we know very little about the interactions between the cells that cause some of them to differentiate in one direction, others in other directions, yet if one fer- tilized egg should begin its development with one kind of material, and another egg with a different material, should we not expect the end products to be different, irrespective of the way in which the materials were present in the original egg^. No matter where the differences may lie, i.e., whether in the nucleus or in the cytoplasm, there is nothing here in any way inconsistent with this particulate theory of the composition of the germ-plasm. On the contrary, the only conclusion that seems at all reasonable 16 242 PHYSICAL BASIS OF HEEEDITY is that if differences are present at the beginning, the end product is expected to be correspondingly different. So much is clear. But why, it may still be asked, are not two organisms that are diff'erent at the start, if only in some one difference, different later in every part, rather than in only some one small part such as in a red or in a white eye. The answer is, of course, that the first difference was such that it affected principally a particular process, viz., the formation of the red pigment of the eye, and to a less degree, or not at all, other chemical processes. This seems to me an entirely consistent view. Perhaps the difficulty in accepting the particulate theory lies in the erroneous idea that the specific effect comes into action only at the moment when the red pig- ment is about to form. But no one has, so^ far as I know, made such a claim. It may be true, but it has not been proven, and is moreover not in any way essential to the assumption of the particulate theory. On the contrary, as our knowledge of Mendelian heredity has increased many cases have been found where a special factor-differ- ence affects not only one part of the body but many parts. It is true that the particulate theory as held at one time by Eoux and for a long time by Weismann was used to explain the differentiating changes in the segmenting egg and embryo in the sense that development was looked upon as a process that resulted immediately in the sorting out of the inherited chromosomal particles to the differ- ent parts of the organism. Differentiation resulted in the sorting out of particular genes to particular groups of cells whose development they controlled. But the cyto- logical evidence in regard to the chromosomes gave no evidence in support of the view, and the evidence from the experimental study of embryology seemed to entirely disprove any such basis for the developmental phe- nomena. In fact, Eoux himself abandoned this view in the light of the brilliant experiments of Driesch and of other embryologists. PAETICULATE THEORY OF HEREDITY 243 Our present conception of the relation of the germ- plasm to developmental phenomena has then only a most superficial resemblance to the older theories. The newer point of view may be summed up in a few words, and has in fact been stated already. First, that each gene may have manifold effects on the organism, and second, that every part of the body, and even each particular character, is the product of many genes. The evidence for these two conclusions has been so repeatedly referred to in the pre- ceding pages that it is not necessary to go over it again, but it may be worth while to emphasize that these two conclusions are not pure speculations, but derived from the evidence itself. It may also be well to point out that even if the whole germ-plasm — the sum of all the genes — acts in the formation of every detail of the body, still the evidence from heredity shows that this same material becomes segregated into tw^o parts during the maturation of the egg and sperm, and that at this time individual elements separate from each other largely independently of the separation of other pairs of elements. It is in this sense, and in this sense only, that we are justified in speak- ing of the particulate composition of the germ-plasm and of particulate inheritance. There is a further idea deducible from well-kno\\Ti facts of physiology that may at first sight seem to give an impression that the organism is a *' whole.'' This is the action of one part of the body on other parts by means of substances set free in the blood, called hor- mones. Many of them arise through the action of certain so-called endocrine glands. But the relation here is so obviously different from the problem dealt with as par- ticulate inheritance that it calls for little more than passing notice. It may, however, not be without interest to refer to one case of the kind in which an endocrine secretion depends on a genetic factor inherited in the same way as are other genetic factors. There is a race of poultry known as Sebrights (Fig. 107, a) in which the 244 PHYSICAL BASIS OF HEREDITY males are always hen-feathered. This means that the feathers of the neck and back and the tail coverts of the Sebright cock are nearly like those of the hen of this breed, and not long and pointed as in the ordinary cock. AYhen Sebrights are crossed to game bantams (which have ordinary males), the F^ males are hen-feathered. When these are inbred the two types reappear in the Fo males. One, or probably two, Mendelian factor differ- ences account for the results. It has been sho^^m that when the testes are removed from the Sebright male, he then develops at the next moult (or at once if some feathers are plucked out) the long and highly colored feathers of the ordinary male (Fig. 107, h). It is probable, therefore, that the testes of the Sebright produce an internal secretion that inhibits in the male the full development of certain feathers. This makes him like the hen, and in this connection it is interesting to note that when the ovary of a hen of an ordinary breed is removed she also develops the full plumage of the cock, as Goodale has clearly demonstrated. Whether the testes of a male are of the sort to develop this inhibiting substance, depends on the presence in the cells of the testes of certain genetic factors. These factors are present, presumably, in all the cells of the body, but if they are, their activity is ineffective in the absence of secretions produced by the testes, as is shown by the castrated Sebright becoming cock-feathered. Whether this substance belongs in the heterogeneous group of substances called hormones — defined by the kind of action they produce rather than by any chemical peculiarity — or to the groups of enzymes that have a more or less specific action, caimot be stated. The fore^^oing discussion touches upon the question as to whether there is any evidence that the genes themselves are to be regarded as enzymes.* In almost all of the * Inadequate as is our knowledge of the physico-chemical processes that go on in development, it is enough to indicate that many processes are at work. PARTICULATE THEORY OF HEREDITY 245 recent papers (Beijerinck, Riddle, Goldschmldt) that touch on this question it is argued, from the evidence of the specific enzymes supposed or demonstrably involved in the production of some final stage in the chemical reaction that leads to the character in question, that the gene itself is the same specific enzyme. The argument shifts back and forth from unit-character to unit-factor. The reason- able position to take in this matter is, in my opinion, that stated by Loeb and Chamberlain (1915), ''The hereditary factor in this case must consist of material which deter- mines the formation of a given mass of these enzymes, since the factors in the chromosomes are too small to carry the whole mass of the enzymes existing in the embryo or adulf It should not be forgotten, however, that the evidence in favor of enzyme action as the most important developmental process is by no means established, and even were the evidence for this view adequate, the stages between such action and the ultimate chemical nature of the gene may be too great to be cleared at a single bound. Some of the modern work on the chemical composition of the nucleus indicates that extremely complex protein com- pounds may be present in it^ — even though some of the split products obtainable from it may be relatively simple. It seems to me therefore that it is both premature and highly speculative at present to tie up the genetic evi- dence concerning the genes with hypotheses concerning their chemical composition. I urge this, but at the same time I realize of course that we should endeavor to obtain as soon as possible better knowledge as to the chemical nature of the .chromatin. Another question concerning the gene, that has been raised, is whether it is to be regarded as something having a definite molecular constitution, or whether the gene is to be regarded as a quantity of material fluctuating about a mode— its definiteness representing only a general ten- dency for the same frequency distribution to recur in each species. From the nature of the case such a question 246 PHYSICAL BASIS OF HEREDITY is speculative, and would have little importance were it not that, by imputing to the advocates of Mendelian heredity the assumption of absolute fixity to the gene, attempts have been made to throw the burden of proof that the genes are ^'constant'' on the advocates of Mendelism. So far as the genetic evidence is involved, I see at pres- ent no way of deciding whether the gene has a definite molecular constitution, or is only something that fluctuates under the condition of its occurrence about a mode. Inter- esting as it might be to speculate about these alternatives, it seems futile to do so at present, but there is one impli- cation that I should like to examine. If the gene is a chemical molecule it is not evident how it could change except by altering its chemical constitution. Its influence, i,e., the chemical effects it produces, might, however, be altered by changing other substances with which the mate- rial it produces reacts. This is the idea involved in the theory of ** modifying genes.'' But if the gene is a fluctuating amount of something it might seem that any *' fluctuation'' that is present at one time might be perpetuated by selection, and that a further fluctuation in the same direction might be utilized for a further advance, etc. It may be pointed out that this picture of the process is quite fanciful, and its success would depend largely on a denial of the premise as to the nature of the gene, viz., that it is of a fluctuating amount. Johannsen's facts contradict an interpretation of the fluctuations of the character being due to a new modal position of the gene standing for that character. And his facts furnish the only crucial evidence we have at present. B D Fig. 107. — A. Adult hen-feathered Campine male. B. Adult male of same race that had been castrated while still a young bird. When it became older it developed cock-feathering. It resembles the male of another race of Campines in which the male is normally cock-feathered. C. Adult hen- feathered Sebright male. D. Adult male Sebright, that had been castrated while still a younj; bird. It developed cock-feathering when it became older. CHAPTER XX MUTATION Concerning the origin of the germinal differences that give rise to mutant characters very little is known at pres- ent except, (1) that they appear infrequently, (2) that the change is definite from the beginning, (3) that some of the changes at least are recurrent, and (4) that the differ- ence between the old character and the new one is small in some cases and greater in others. I do not think that any of the work purporting to produce specific mutational changes has succeeded in establishing its claims, at least in the sense that we can pretend at present to control the appearance of specific mutant changes, and until this is done we can not hope to find out very much as to the nature of these changes. Our study of the germ-plasm is largely confined, therefore, for the present, to a study of transmission of the genes, to the kinds of effects they produce on the organism, and to the special relations of the genes in the chromosomes where they are located. Concerning the frequency of mutation there is a slowly increasing body of evidence showing in some animals and plants how often or how rarely changes of this kind take place. The impression prevails that mutation is less rare in some species than in others, and while I am inclined to think that this may be true, not much value can be ascribed to such impressions ; for it is not improbable that the frequency with which mutations are found is often directly in proportion to the number of individuals exam- ined and to familiarity with the type in question, so that the smaller changes are not overlooked. The dis- covery of new mutant types in almost every plant and animal that has been carefully examined indicates at least the very general occurrence of definite mutations, and the 247 248 PHYSICAL BASIS OF HEREDITY great variety of types shown by nearly all of our domesti- cated animals and plants — varieties that follow MendePs law — appears to give further support to the view that the process of mutation is widespread. One of the most interesting phenomena connected with mutation is the recurrence of the same change. It has long been recognized that certain *^ sports'' such as albi- nos and melanic forms are found again and again in nature. In insects there are many records of the sporadic appearance of the same type, such as the light form (lacti- color) of the moth Abraxas. It is true that not all such appearances are to be accepted offhand as the first appear- ance of the mutative change, since when these are reces- sive it is probable in most cases * that the actual mutation occurred several generations before the mutated genes came together to produce the mutant character. But granting this, it is at least probable that the same type has appeared in many cases independently. The only evidence that can be relied upon in such cases is from pedigreed cultures, followed up by evidence that the mutants that look alike are really due to mutations in the same locus. Fortunately there is actual evidence, both for plants and for animals, that can be appealed to to show that the same mutations recur. The most extensive evidence is from Drosophila melanogaster. One of the first mutants that appeared, viz., white eyes, has appeared anew in our cultures about three times, in cultures known to be free from it before and not contaminated. The same mutant has been found by several other observers. The eye-color vermilion has appeared at least six times; the wing character called rudimentary, five times; cut wing has been found four times; truncate wing has frequently appeared, but has not necessarily been always produced by the same change. Certain characters such as notch wings, that have * Recessive mutations in the X-chromosomes of the XX-XY type may appear in the male in the next generation. MUTATION 249 appeared quite often, represent, it seems, a pecuHar change whose relation to the changes that stand behind other mutant characters is not yet worked out. In plants the best evidence is that reported by Emerson for Indian corn. Emerson has shown that when a race of corn {Zea mais) having red cobs and red seeds is crossed to a race having white cobs and white seeds only, the two original combinations appear in the second (Fo) generation giving plants ^vith red cobs and red seeds and plants with white cobs and white seeds. Either a single factor determines that both cob and seed are red in one case and white in the other, or if the color of each part is due to a separate factor these factors are completely linked. Now striped seeds with white cobs sometimes mutate to red seeds and red cobs. The new combination (red and red) acts as a unit toward the other kno^vn com- binations. Therefore a single factor must have changed, for, if not, mutation must occur in two (or more) closely linked factors, i.e., for seed and cob color at the same time, which is highly improbable. In forms propagating by sexual methods it cannot be told whether mutation has occurred in one locus or in both homologous loci at the same time, because in the Qgg one of each pair of genes is lost in the polar body, and irrespective of whether one or' two mutated genes were present only one member of the pair is left in the ripe Qgg ; and in the sperm the chance of any one sperm reaching the egg is so small that it is unlikely that the difference between one sperm or two sperms having the mutated locus could be detected. It is true that of the twelve domi- nant mutants that have appeared in Drosophila each appeared at first in a single individual — never two — which might appear to favor the single locus view, but this evi- dence is too meagre to be significant. Mutants from reces- sive genes usually come to light in about a quarter of the offspring of a given pair. This means that both parents were heterozygous for the mutant gene, but this gene 250 PHYSICAL BASIS OF HEREDITY must have arisen at least one generation earlier, and have been carried over into the two heterozygous indi- \dduals in question. It would be a point of capital importance if it could be determined beyond doubt that at times recessive mutant genes change back to the original (wild type) gene, or even if a recessive gene could mutate to a dominant one. The appearance of the wild type in a pure culture of a mutant race can be accepted as good evidence of such a change only when every possibility of contamination by the wild type is excluded, and this is difficult to regulate. In our cultures we have come across such cases, but have not ventured to exploit them, since wild-type flies are always present in the laboratory, and hence the discovered form may have arisen through accidental contamination. Thus even when a red-eyed yellow fly appeared in the white-eyed yellow stock there is the barest chance that a yellow red-eyed fly, or an egg of such a fly, had somehow gotten into the stock. Certainty can be attained only when a stock, pure for several mutant characters, reverts to the normal in one of these characters, and not in the others. Only one case of this kind that is above suspicion has been as yet recorded. This is a mutant stock in which, as May has recorded, reversion to the wild type occurs with such frequency that there can be no chance of error. The stock in question, bar eye, is a dominant mutant and the rever- sion therefore is to the recessive wild type of eye (round eye). The change back to normal is complete, since such individuals give only normal offspring. When such a mutant chromosome comes from the mother and goes into a son he has normal (wild type) eyes ; when it comes from the father, and goes to a daughter, she is heterozygous for bar eye. Baur has recently recorded the appearance of recessive ( ?) mutants from self -fertilized plants (snap- dragon) that bred true at once. Punnett has described n similar case (1919). The result can be accounted for, if a mutation occurred in only a single chromosome far enough MUTATION 251 back in the germ-tract to give rise, after reduction, botli to pollen and to ovules, each one carrying the mutated genes. Such an interpretation is supported by the evi- dence from Drosophila, Avhere, although mutations are much more numerous, no such cases have been observed, and none such would be expected if mutation occurs in a single chromosome at a time, since here the germ-cells come from separate individuals. Probably the most important evidence bearing on the nature of the genes is that derived from multiple allelo- morphs. Now that the proof has been furnished that the phenomena connected with these cases are not due to nests of closely linked genes, we can properly appeal to these as crucial cases. As already explained, in ever-increasing numbers of animals and plants, series of genes have been found in each of which mutant characters with the same normal allelomorph have been found. These mutant char- acters of each series are also allelomorphs of one another — only two ever existing in the same individual. Ob- viously, not all such mutants can be due to the absence of a factor present in the germ-plasm of the wild t>ije, since only one kind of absence is thinkable. If to save the situation for the theory of presence and absence it be assumed that only a part of the original gene is absent, and a different part in each case, then nothing is gained by the admission; and while this may be true it is equally possible that the genes change in other ways. It is not essential that we should specify the nature of the change, but simpler to look upon the mutant gene as due to some kind of change or changes that have taken place in the original germ-plasm at a specific locus — there is nothing known at present to furnish even a clue as to the nature of this change. The demonstration that multiple allelomorphs are modifications of the same locus in the chromosome, rather than cases of closely linked genes, can come only where their origin is known, and at present this holds only in 252 PHYSICAL BASIS OF HEREDITY the case (just stated) for Indian corn and for the fruit fly. If each member of such a series of allelomorphs has arisen historically from the preceding one in the series, by a mutation in a locus closely associated with the locus responsible for the first, they would be expected to give the wild type when crossed; and as the proof of their allelomorphism turns on the failure of members of the B 1 • 2 • A 3 • 4 t 5 t 1 0 1 • 1 • 1 • 1 • 2 • 2 0 2 • 2 • 2 • 3 • 3 • 3 0 3 • 3 • 4 • 4 t 4 • 4 0 4 • 5 • 5 • 5 • 5 • 5 0 a b c d e 1 0 a 0 1 0 1 0 1 0 2 t 2 0 2 0 2 0 2 0 3 • 3 • 3 0 3 0 3 0 4 • 4 • 4 • 4 0 4 0 5 t 5 • 5 # 5 • 6 0 Fig. 108. — Diagram illustrating mutation in a nest of genes so closely linked that no crossing over' takes place. series to show the atavistic behavior on crossing, it is necessary, as stated, to know how they arose. This may be made clear by the following illustration : Let the five circles of Fig. 108, A represent a nest of closely linked genes. If a recessive mutation occurs in the first one (line B, a) and another in the second gene (line B, h), the two mutants a and h if crossed should give the atavistic type, since a brings in the normal allelo- morph (B) of b, and h that (A) of a. If a third mutation should occur in the third gene it, too, will give the atavistic MUTATION 253 type if crossed to a or to h. Similarly for a mutation in the fourth and in the fifth normal gene. Now this is exactly what does not take place when memhers of an allelomorphic series are crossed — they do not give the wild type, but one of the other mutant types or an inter- mediate character. Evidently independent mutation in a nest of linked normal genes will not explain the results if the new genes arise directly each from a different nor- mal allelomorph. But suppose, as shown in Fig. 4 (line C) after a muta- tion had occurred in the first gene a new mutant, h, arose from a new gene, and from h a mutation arose in a third gene c, and o similarly gave rise to d; then a crossed to h will give a (or something intermediate if the heterozygote is an intermediate type). Likewise c crossed to h will give h, or c crossed to a will give a, etc. If mutant allelo- morphic genes in a series such as 6\ a, h, c, d, e, arise as successive steps, i.e., Ca to Cb and Cb to Cc, etc., then the hypothesis of closely linked genes would seem to be a possible interpretation of the data, but if they do not arise in this way, but by independent mutations from the wild type (or even from each other, but not seriatim), then they must be due to mutations in the same gene : for, to assume that they are not, requires that, when the second mutation took place both gene a and gene b mutated at the same time, and that when c appeared three genes mutated, when gene d appeared four; when gene e five genes mutated at once, four of them being mutant genes that have already arisen independently. Such an inter- pretation is excluded, since it is inconceivable, even in a readily mutating form like Brosophila, that five muta- tions could have occurred at the same time in distinct but neighboring loci. As has been stated, the evidence from Brosophila shows positively that multiple allelomorphs arise at random. Only two members of a series of multiple allelomorphs can be present in any one individual, and in the case of 254 PHYSICAL BASIS OF HEREDITY genes carried by the sex-chromosome only one can exist at a time in the sex that has only one of these chromosomes. In the individual with two mutant allelomorphs one of them replaces the normal allelomorph of the ordinary Mendelian pair. The two mutant allelomorphs behave towards each other in the same way as does the normal towards its mutant allelomorphs. It is doubtful whether we can conclude anything more from this relation of Men- delian pairs than we knew before/ although there is at least a sentimental satisfaction in knowing that both nor- mal allelomorphs can be replaced by mutant ones without altering the working of the machinery. The linkage relation of each member of a series of multiple allelomorphs to all other genes of its chromo- some is, of course, the same. While the theory of identical loci requires this as a primary condition it is not legiti- mate to use this evidence as a proof of the identity of the loci, because it is not possible to work with sufficient pre- cision in locating genes by their relation to other linked genes to distinguish between identical loci and close- linked genes. The question of lethal genes has attracted in recent years increasing attention, both on account of their fre- quency and because of a curious complication they may produce in hiding the effects of other genes also present. In Drosophila we have records of more than 20 sex-linked lethals, and about 15 not sex-linked, and scattering records of many others. Gametic lethal genes are those that destroy eggs or pollen cells that contain such genes. Zygotic lethal genes affect the embryo, the larva, or the adult, so that it dies. In the case of the garden plant known as double ^^ stocks," the genetic evidence obtained by Miss Saunders indicates that certain kinds of pollen are not produced, and presumably die because of a con- tained factor. The same factor does not kill the ovules, ' The substitution by crossing over really furnishes as good a demon- -^tration of this point. MUTATION 255 which may therefore transmit the recessive lethal gene to half the progeny. How far the frequent occurrence of im- perfect pollen grains in many species of plants is due to such factors is still uncertain. Belling found that while the Florida velvet bean produces normal pollen grains and ovules, and the Lyon bean, another bean of the same genus, also produces nor- mal gametes, the F^ hybrid contains 50 per cent, abortive pollen grains, and possibly about 50 per cent, of the ovules are abortive. In the second generation (Fo) half of the pollen grains of half of the plants are abortive. The other half of the plants have normal pollen grains. This is the result expected if there are present in one of the species the factors AAbh, and in the other species the factors aaBB, the viable gametes in the F^ generation being those containing Ab, Ba, and the two gametes that die being AB, db. Other observers have made records of abortive pollen in hybrids, but without knowing the condition of the pollen in the parents the interpretation of the results is doubtful, for, as Jeffrey has emphasized, abortive pollen is a characteristic of many wild species. There is one fact of capital importance recorded by several botanists, viz., that the degeneration of the germ-ceUs only takes place after the tetrad has been produced, and only in some of the cells of each tetrad. In other words, the lethal effect is not observed until the chromosomes have under- gone reduction. It is obvious that if there is present a recessive lethal for the germ-cells (or for any cells, in fact), it causes no injury in the presence of its normal allelomorph, but kills when the counter-effect of its part- ner is removed. Tischler found in a hybrid currant that tetrad fonna- tion was normal, and that the shrinking of the pollen grains occurred afterwards. Geerts found that one-half of the pollen grains of CEnothera Lamarckiana degen- erate, and that half of the embryo sacs abort in the tetrad 256 PHYSICAL BASIS OF HEREDITY stage. Other related (wild) species and genera of the evening primrose have also been found to have some abortive pollen and ovules. Complete or nearly complete abortion has been seen in other hybrids; viz., by Rosenberg in the sundew, by Osawa in the Satsuma orange, by Goodspeed and others in the hybrid tobacco {N. tabacum by N. sylvestris), by Jesenko in the wheat-rye hybrid, and by Sutton in the hybrid between the Palestine pea {Pisum humile) and the edible pea. These cases may be in part the same phe- nomenon and in part a different one comiected with fail- ure of the chromosome to conjugate or to be properly distributed during the maturation divisions. The ^* yellow mouse case'' is an example of a zygotic lethal effect. The gene that produces the dominant yellow color is lethal in double dose, so that all homozygous yel- low mice die, as Cuenot first discovered, and as has been more positively demonstrated by the work of Castle and Little. There is some evidence indicating that these homo- zygotes die as young embryos. Little has also shown that black-eyed white mice carry a lethal, that acts in the same way. In Drosophila there is a sex-linked recessive lethal factor that causes the development of tumors in the larvae, destroying every male larva that contains the sex-chromo- some carrying this gene. This effect, discovered by Bridges, has been the basis for an extensive series of experiments by Miss Stark. The gene is present in the X-chromosomes ; it follows the rules for all sex-linked genes in its inheritance. The females of the stock are of two kinds : One has the lethal in one sex-chromosome, and its normal, dominant allelomorph in the other. Such a female has survived because the effect of the lethal gene is counteracted by the effect of its normal allelomorph. Half of her sons get the affected chromosome. All such sons develop the tumor — one or more melanitic growths that appear in the imaginal discs or in other parts of the larva. The other sons get the other chromosome with the MUTATION 257 normal allelomorph. They never produce a tumor and never transmit the disease. The same mother that gave these two kinds of sons — having been fertilized by a nor- mal male, since no affected males exist— produces also two kinds of daughters, one containing the gene for the tumor (and its normal allelomorph), the other having two nor- mal genes. The former transmit the disease as just explained, the latter daughters are perfectly normal and do not transmit the disease. Other lethal genes kill the pupae, a few of them even allow the fly occasionally to come through, but such flies rarely propagate. Certain races of Drosophila have ster- ile or nearly sterile females, other races sterile males. The sterility is here lethal in so far as it affects the germ- cells. Some effects on other characters are also generally to be seen. The presence of a lethal gene near to, i.e., linked to, another mutant gene may affect the kinds of individ- uals that appear because owing to the linkage the other mutant character fails to appear, except when crossing over takes place. Some examples of this relation may be given. There is a mutant race called beaded (Fig. 109) in which the margin of the wing is irregularly broken, giving the appearance of a beaded edge. The gene for beaded is dominant, and lethal when homozygous. As in the case of the yellow mouse, only the hybrid (heterozygous) combination exists, and consequently when two beaded flies mate they produce two beaded to on^ normal fly, as shown in Fig. 110. Here the first pair of vertical lines stand for the pair of third chromosomes present in the egg before its reduction. The two genes here involved, that for beaded and its allelomorph for normal, are indicated at the lower end of the vertical lines. The two corresponding chromosomes in the male are represented to the right of the last. After the ripening of the germ-cells each egg and each sperm carries one or the other of these chromosomes. Chance meetings of egg 17 258 PHYSICAL BASIS OF HEREDITY and sperm are indicated in the figure by the arrow-scheme below, which gives the combinations (classes) included in the four squares. The double dominant BB is the class that does not come through. The result is two beaded (heterozygous) to one normal fly. The beaded stock remained in this condition for a long time ; although selected in every generation for beaded, it did not improve, but continued to throw 33 per cent, of normal flies. Then it changed and bred nearly true. Eggs Sperm B-- t N B"- --N 5 ^^ ixt B N. Beaded : 1 Normal, Fig. 110. — Diagram showing the relation of the chromosomes (represented by the vertical rods) in a cross of "beaded" by "beaded." Flies homozygous for beaded die as indicated by the cross-hatched square. The change must have been due to the appearance of another lethal factor (now called lethal three, here Zi) in Fig. 111). Such a gene was found in the race when studied later by Muller. The lethal gene that appeared in the beaded stock was also in the third chromosome, and in the chromosome that is the mate of the one carrying the gene for beaded, i.e., in the normal third chromosome of the beaded stock. The lethal gene lies so near to the level of the beaded-normal t-H Q O CO o a o Co ^ a 0 1 Acad. Sci., 1908, cxlvii. Payne, F. : An Experiment to Test the Nature of the Variations on which Selection Acts. Indiana Univ. Studies, 1918, v. PaynE; F. : The Effect of Artificial Selection on Bristle Number in Drosophila ampelophila and Its Interpretation. Proc. Nat. Acad. Sci., 1918, iv. Pearl_, R. : The Mode of Inheritance of Fecundity in the Domestic FowL Jour. Exp. Zool., 1912, xiii. Rawls, E.: Sex Ratios in Drosophila ampelophila. Biol. Bull., 1913, xxiv. Riddle^ 0. : Our Knowledge of Melanin Color Formation and Its Bear- ing on the Mendelian Description of Heredity. Biol. Bull., 1909, xvi. Riddle^ 0. : Preliminary Chemical Studies on Male- and Female-Produc- ing Eggs of Pigeons. Science, 1912, xxxv. Riddle, Oscar: Sex Control and Known Correlations in Pigeons. Am. Nat., 1916, 1, pp. 385-410. Riddle^ Oscar : Success in Controlling Sex. Jour. Hered., 1916, vii. Riddle, Oscar: The Theory of Sex as Stated in Terms of Results of Studies on Pigeons. Science, 1917, n.s., xlvi. Riddle, Oscar: The Control of the Sex Ratio. Journ. Wash. Acad. Sc, 1917, vii. Riddle, Oscar: The Theory of Sex as Stated in Terms of Residts of Studies on Pigeons. Science, 1917, n.s., xlvi. Robertson, W. R. B. : Chromosome Studies. III. Inequalities and Deficiencies in Homoiogous Chromosomes: Their Bearing upon Synapsis and the Loss of Unit Characters. Jour. Morph., 1915, xxvi. Robertson, W. R. B. : Chromosome Studies. I. Taxonomic Relation- ships .Shown in the Chromosomes of Tettigidce and Acrididce: V- shaped Chromosomes and Their Significance in Acrididce, Locustid