E-text prepared by Paul Hollander, Malcolm Farmer, Keith Edkins,
and the Project Gutenberg Online Distributed Proofreading Team
(http://www.pgdp.net)

MENDELISM

BY

R. C. PUNNETT

FELLOW OF GONVILLE AND CAIUS COLLEGE

PROFESSOR OF BIOLOGY IN THE UNIVERSITY OF CAMBRIDGE

THIRD EDITION
ENTIRELY REWRITTEN AND MUCH ENLARGED

New York

THE MACMILLAN COMPANY

1911

All rights reserved

Copyright, 1911,

By THE MACMILLAN COMPANY.

Set up and electrotyped. Published May, 1911.

Norwood Press
J. S. Cushing Co.—Berwick & Smith Co.
Norwood, Mass., U.S.A.

PREFACE

A few years ago I published a short sketch of Mendel's discovery in heredity, and of some of the recent experiments which had arisen from it. Since then progress in these studies has been rapid, and the present account, though bearing the same title, has been completely rewritten. A number of illustrations have been added, and here I may acknowledge my indebtedness to Miss Wheldale for the two coloured plates of sweet peas, to the Hon. Walter Rothschild for the butterflies figured on Plate VI., to Professor Wood for photographs of sheep, and to Dr. Drinkwater for the figures of human hands. To my former publishers also, Messrs. Bowes and Bowes, I wish to express my thanks for the courtesy with which they acquiesced in my desire that the present edition should be published elsewhere.

As the book is intended to appeal to a wide audience, I have not attempted to give more experimental instances than were necessary to illustrate the story, nor have I burdened it with bibliographical reference. The reader who desires further information may be referred to Mr. Bateson's indispensable Volume on Mendel's

Principles of Heredity (Cambridge, 1909), where a full account of these matters is readily accessible. Neither have I alluded to recent cytological work in so far as it may bear upon our problems. Many of the facts connected with the division of the chromosomes are striking and suggestive, but while so much difference of opinion exists as to their interpretation they are hardly suited for popular treatment.

In choosing typical examples to illustrate the growth of our ideas it was natural that I should give the preference to those with which I was most familiar. For this reason the book is in some measure a record of the work accomplished by the Cambridge School of Genetics, and it is not unfair to say that under the leadership of William Bateson the contributions of this school have been second to none. But it should not be forgotten that workers in other European countries, and especially in America, have amassed a large and valuable body of evidence with which it is impossible to deal in a small volume of this scope.

It is not long since the English language was enriched by two new words—Eugenics and Genetics—and their similarity of origin has sometimes led to confusion between them on the part of those who are innocent of Greek. Genetics is the term applied to the experimental study of heredity and variation in animals and plants, and the main concern of its students is the establishing of law and order among the phenomena

there encountered. Eugenics, on the other hand, deals with the improvement of the human race under existing conditions of law and sentiment. The Eugenist has to take into account the religious and social beliefs and prejudices of mankind. Other issues are involved besides the purely biological one, though as time goes on it is coming to be more clearly recognised that the Eugenic ideal is sharply circumscribed by the facts of heredity and variation, and by the laws which govern the transmission of qualities in living things. What these facts, what these laws are, in so far as we at present know them, I have endeavoured to indicate in the following pages; for I feel convinced that if the Eugenist is to achieve anything solid it is upon them that he must primarily build. Little enough material, it is true, exists at present, but that we now see to be largely a question of time and means. Whatever be the outcome, whatever the form of the structure which is eventually to emerge, we owe it first of all to Mendel that the foundations can be well and truly laid.

R. C. P.

Cambridge, March, 1911.

CONTENTS

ILLUSTRATIONS

For although it be a more new and difficult way, to find out the nature of things, by the things themselves; then by reading of Books, to take our knowledge upon trust from the opinions of Philosophers: yet must it needs be confessed, that the former is much more open, and lesse fraudulent, especially in the Secrets relating to Natural Philosophy.

William Harvey,

Anatomical Exercitations, 1653.

CHAPTER I

THE PROBLEM

A curious thing in the history of human thought so far as literature reveals it to us is the strange lack of interest shown in one of the most interesting of all human relationships. Few if any of the more primitive peoples seem to have attempted to define the part played by either parent in the formation of the offspring, or to have assigned peculiar powers of transmission to them, even in the vaguest way. For ages man must have been more or less consciously improving his domesticated races of animals and plants, yet it is not until the time of Aristotle that we have clear evidence of any hypothesis to account for these phenomena of heredity. The production of offspring by man was then held to be similar to the production of a crop from seed. The seed came from the man, the woman provided the soil. This remained the generally accepted view for many centuries, and it was not until the recognition of woman as more than a passive agent that the physical basis of heredity became established. That recognition was effected by the microscope, for only with its advent was actual

observation of the minute sexual cells made possible. After more than a hundred years of conflict lasting until the end of the eighteenth century, scientific men settled down to the view that each of the sexes makes a definite material contribution to the offspring produced by their joint efforts. Among animals the female contributes the ovum and the male the spermatozoon; among plants the corresponding cells are the ovules and pollen grains.

As a general rule it may be stated that the reproductive cells produced by the female are relatively large and without the power of independent movement. In addition to the actual living substance which is to take part in the formation of a new individual, the ova are more or less heavily loaded with the yolk substance that is to provide for the nutrition of the developing embryo during the early stages of its existence. The size of the ova varies enormously in different animals. In birds and reptiles where the contents of the egg form the sole resources of the developing young they are very large in comparison with the size of the animal which lays them. In mammals, on the other hand, where the young are parasitic upon the mother during the earlier stages of their growth, the eggs are minute and only contain the small amount of yolk that enables them to reach the stage at which they develop the processes for attaching themselves to the wall of the maternal uterus. But whatever the differences in the size and appearance of the ova produced by different

animals, they are all comparable in that each is a distinct and separate sexual cell which, as a rule, is unable to develop into a new individual of its species unless it is fertilised by union with a sexual cell produced by the male.

The male sexual cells are always of microscopic size and are produced in the generative gland or testis in exceedingly large numbers. In addition to their minuter size they differ from the ova in their power of active movement. Animals present various mechanisms by which the sexual elements may be brought into juxtaposition, but in all cases some distance must be traversed in a fluid or semifluid medium (frequently within the body of the female parent) before the necessary fusion can occur. To accomplish this latter end of its journey the spermatozoon is endowed with some form of motile apparatus, and this frequently takes the form of a long flagellum, or whip-like process, by the lashing of which the little creature propels itself much as a tadpole with its tail.

In plants as in animals the female cells or ovules are larger than the pollen grains, though the disparity in size is not nearly so marked. Still they are always relatively minute cells since the circumstances of their development as parasites upon the mother plant render it unnecessary for them to possess any great supply of food yolk. The ovules are found surrounded by maternal tissue in the ovary, but through the stigma and down the pistil a

potential passage is left for the male cell. The majority of flowers are hermaphrodite, and in many cases they are also self-fertilising. The anthers burst and the contained pollen grains are then shed upon the stigma. When this happens, the pollen cell slips through a little hole in its coat and bores its way down the pistil to reach an ovule in the ovary. Complete fusion occurs, and the minute embryo of a new plant immediately results. But for some time it is incapable of leading a separate existence, and, like the embryo mammal, it lives as a parasite upon its parent. By the parent it is provided with a protective wrapping, the seed coat, and beneath this the little embryo swells until it reaches a certain size, when as a ripe seed it severs its connection with the maternal organism. It is important to realise that the seed of a plant is not a sexual cell but a young individual which, except for the coat that it wears, belongs entirely to the next generation. It is with annual plants in some respects as with many butterflies. During one summer they are initiated by the union of two sexual cells and pass through certain stages of larval development—the butterfly as a caterpillar, the plant as a parasite upon its mother. As the summer draws to a close each passes into a resting-stage against the winter cold—the butterfly as a pupa and the plant as a seed, with the difference that while the caterpillar provides its own coat, that of the plant is provided by its mother. With the advent of spring both butterfly and

plant emerge, become mature, and themselves ripen germ cells which give rise to a new generation.

Whatever the details of development, one cardinal fact is clear. Except for the relatively rare instances of parthenogenesis a new individual, whether plant or animal, arises as the joint product of two sexual cells derived from individuals of different sexes. Such sexual cells, whether ovules or ova, spermatozoa or pollen grains, are known by the general term of gametes, or marrying cells, and the individual formed by the fusion or yoking together of two gametes is spoken of as a zygote. Since a zygote arises from the yoking together of two separate gametes, the individual so formed must be regarded throughout its life as a double structure in which the components brought in by each of the gametes remain intimately fused in a form of partnership. But when the zygote in its turn comes to form gametes, the partnership is broken and the process is reversed. The component parts of the dual structure are resolved, with the formation of a set of single structures, the gametes.

The life cycle of a species from among the higher plants or animals may be regarded as falling into three periods: (1) a period of isolation in the form of gametes, each a living unit incapable of further development without intimate association with another produced by the opposite sex; (2) a period of association in which two gametes become yoked together into a zygote and react upon one

another to give rise by a process of cell division to what we ordinarily term an individual with all its various attributes and properties; and (3) a period of dissociation when the single structured gametes separate out from that portion of the double structured zygote which constitutes its generative gland. What is the relation between gamete and zygote, between zygote and gamete? how are the properties of the zygote represented in the gamete, and in what manner are they distributed from the one to the other?—these are questions which serve to indicate the nature of the problem underlying the process of heredity.

Owing to their peculiar power of growth and the relatively large size to which they attain, many of the properties of zygotes are appreciable by observation. The colour of an animal or of a flower, the shape of a seed, or the pattern on the wings of a moth are all zygotic properties, and all capable of direct estimation. It is otherwise with the properties of gametes. While the difference between a black and a white fowl is sufficiently obvious, no one by inspection can tell the difference between the egg that will hatch into a black and that which will hatch into a white. Nor from a mass of pollen grains can any one to-day pick out those that will produce white from those that will produce coloured flowers. Nevertheless, we know that in spite of apparent similarity there must exist fundamental differences among the gametes, even

among those that spring from the same individual. At present our only way of appreciating those differences is to observe the properties of the zygotes which they form. And as it takes two gametes to form a zygote, we are in the position of attempting to decide the properties of two unknowns from one known. Fortunately the problem is not entirely one of simple mathematics. It can be attacked by the experimental method, and with what measure of success will appear in the following pages.

CHAPTER II

HISTORICAL

To Gregor Mendel, monk and abbot, belongs the credit of founding the modern science of heredity. Through him there was brought into these problems an entirely new idea, an entirely fresh conception of the nature of living things. Born in 1822 of Austro-Silesian parentage, he early entered the monastery of Brünn, and there in the seclusion of the cloister garden he carried out with the common pea the series of experiments which has since become so famous. In 1865 after eight years' work he published the results of his experiments in the Proceedings of the Natural History Society of Brünn, in a brief paper of some forty pages. But brief as it is the importance of the results and the lucidity of the exposition will always give it high rank among the classics of biological literature. For thirty-five years Mendel's paper remained unknown, and it was not until 1900 that it was simultaneously discovered by several distinguished botanists. The causes of this curious neglect are not altogether without interest. Hybridisation experiments before Mendel there had been in plenty. The classificatory work of

Linnaeus in the latter half of the eighteenth century had given a definite significance to the word species, and scientific men began to turn their attention to attempting to discover how species were related to one another. And one obvious way of attacking the problem was to cross different species together and see what happened. This was largely done during the earlier half of the nineteenth century, though such work was almost entirely confined to the botanists. Apart from the fact that plants lend themselves to hybridisation work more readily than animals, there was probably another reason why zoologists neglected this form of investigation. The field of zoology is a wider one than that of botany, presenting a far greater variety of type and structure. Partly owing to their importance in the study of medicine, and partly owing to their smaller numbers, the anatomy of the vegetable was far better known than that of the animal kingdom. It is, therefore, not surprising that the earlier part of the nineteenth century found the zoologists, under the influence of Cuvier and his pupils, devoting their entire energies to describing the anatomy of the new forms of animal life which careful search at home and fresh voyages of discovery abroad were continually bringing to light. During this period the zoologist had little inclination or inducement to carry on those investigations in hybridisation which were occupying the attention of some botanists. Nor did the efforts of the botanists afford much

encouragement to such work, for in spite of the labour devoted to these experiments, the results offered but a confused tangle of facts, contributing in no apparent way to the solution of the problem for which they had been undertaken. After half a century of experimental hybridisation the determination of the relation of species and varieties to one another seemed as remote as ever. Then in 1859 came the Origin of Species, in which Darwin presented to the world a consistent theory to account for the manner in which one species might have arisen from another by a process of gradual evolution. Briefly put, that theory was as follows: In any species of plant or animal the reproductive capacity tends to outrun the available food supply, and the resulting competition leads to an inevitable struggle for existence. Of all the individuals born, only a portion, and that often a very small one, can survive to produce offspring. According to Darwin's theory, the nature of the surviving portion is not determined by chance alone. No two individuals of a species are precisely alike, and among the variations that occur some enable their possessors to cope more successfully with the competitive conditions under which they exist. In comparison with their less favoured brethren they have a better chance of surviving in the struggle for existence and consequently of leaving offspring. The argument is completed by the further assumption of a principle of heredity, in virtue of which offspring tend to

resemble their parents more than other members of the species. Parents possessing a favourable variation tend to transmit that variation to their offspring, to some in greater, to others in less degree. Those possessing it in greater degree will again have a better chance of survival, and will transmit the favourable variation in even greater degree to some of their offspring. A competitive struggle for existence working in combination with certain principles of variation and heredity results in a slow and continuous transformation of species through the operation of a process which Darwin termed natural selection.

The coherence and simplicity of the theory, supported as it was by the great array of facts which Darwin had patiently marshalled together, rapidly gained the enthusiastic support of the great majority of biologists. The problem of the relation of species at last appeared to be solved, and for the next forty years zoologists and botanists were busily engaged in classifying by the light of Darwin's theory the great masses of anatomical facts which had already accumulated and in adding and classifying fresh ones. The study of comparative anatomy and embryology received a new stimulus, for with the acceptance of the theory of descent with modification it became incumbent upon the biologist to demonstrate the manner in which animals and plants differing widely in structure and appearance could be conceivably related to one another. Thenceforward the energies of both

botanists and zoologists have been devoted to the construction of hypothetical pedigrees suggesting the various tracks of evolution by which one group of animals or plants may have arisen from another through a long continued process of natural selection. The result of such work on the whole may be said to have shown that the diverse forms under which living things exist to-day, and have existed in the past so far as palaeontology can tell us, are consistent with the view that they are all related by the community of descent which the accepted theory of evolution demands, though as to the exact course of descent for any particular group of animals there is often considerable diversity of opinion. It is obvious that all this work has little or nothing to do with the manner in which species are formed. Indeed, the effect of Darwin's Origin of Species was to divert attention from the way in which species originate. At the time that it was put forward his explanation appeared so satisfying that biologists accepted the notions of variation and heredity there set forth and ceased to take any further interest in the work of the hybridisers. Had Mendel's paper appeared a dozen years earlier it is difficult to believe that it could have failed to attract the attention it deserved. Coming as it did a few years after the publication of Darwin's great work, it found men's minds set at rest on the problems that he raised and their thoughts and energies directed to other matters.

Nevertheless one interesting and noteworthy attempt to give greater precision to the term heredity was made about this time. Francis Galton, a cousin of Darwin, working upon data relating to the breeding of Basset hounds, found that he could express on a definite statistical scheme the proportion in which the different colours appeared in successive generations. Every individual was conceived of as possessing a definite heritage which might be expressed as unity. Of this, ½ was on the average derived from the two parents (i.e. ¼ from each parent), ¼ from the four grandparents, ⅛ from the eight great-grandparents, and so on. The Law of Ancestral Heredity, as it was termed, expresses with fair accuracy some of the statistical phenomena relating to the transmission of characters in a mixed population. But the problem of the way in which characters are distributed from gamete to zygote and from zygote to gamete remained as before. Heredity is essentially a physiological problem, and though statistics may be suggestive in the initiation of experiment, it is upon the basis of experimental fact that progress must ultimately rest. For this reason, in spite of its ingenuity and originality, Galton's theory and the subsequent statistical work that has been founded upon it failed to give us any deeper insight into the nature of the hereditary process.

While Galton was working in England the German zoologist August Weismann was elaborating the complicated

theory of heredity which eventually appeared in his work on The Germplasm (1885), a book which will be remembered for one notable contribution to the subject. Until the publication of Weismann's work it had been generally accepted that the modifications brought about in the individual during its lifetime, through the varying conditions of nutrition and environment, could be transmitted to the offspring. In this biologists were but following Darwin, who held that the changes in the parent resulting from increased use or disuse of any part or organ were passed on to the children. Weismann's theory involved the conception of a sharp cleavage between the general body tissues or somatoplasm and the reproductive glands or germplasm. The individual was merely a carrier for the essential germplasm whose properties had been determined long before he was capable of leading a separate existence. As this conception ran counter to the possibility of the inheritance of "acquired characters," Weismann challenged the evidence upon which it rested and showed that it broke down wherever it was critically examined. By thus compelling biologists to revise their ideas as to the inherited effects of use and disuse, Weismann rendered a valuable service to the study of genetics and did much to clear the way for subsequent research.

A further important step was taken in 1895, when Bateson once more drew attention to the problem of the origin

of species, and questioned whether the accepted ideas of variation and heredity were after all in consonance with the facts. Speaking generally, species do not grade gradually from one to the other, but the differences between them are sharp and specific. Whence comes this prevalence of discontinuity if the process by which they have arisen is one of accumulation of minute and almost imperceptible differences? Why are not intermediates of all sorts more abundantly produced in nature than is actually known to be the case? Bateson saw that if we are ever to answer this question we must have more definite knowledge of the nature of variation and of the nature of the hereditary process by which these variations are transmitted. And the best way to obtain that knowledge was to let the dead alone and to return to the study of the living. It was true that the past record of experimental breeding had been mainly one of disappointment. It was true also that there was no tangible clue by which experiments might be directed in the present. Nevertheless in this kind of work alone there seemed any promise of ultimate success.

A few years later appeared the first volume of de Vries' remarkable book on The Mutation Theory. From a prolonged study of the evening primrose (Oenothera) de Vries concluded that new varieties suddenly arose from older ones by sudden sharp steps or mutations, and not by any process involving the gradual accumulation of minute

differences. The number of striking cases from among widely different plants which he was able to bring forward went far to convincing biologists that discontinuity in variation was a more widespread phenomenon than had hitherto been suspected, and not a few began to question whether the account of the mode of evolution so generally accepted for forty years was after all the true account. Such in brief was the outlook in the central problem of biology at the time of the rediscovery of Mendel's work.

CHAPTER III

MENDEL'S WORK

The task that Mendel set before himself was to gain some clear conception of the manner in which the definite and fixed varieties found within a species are related to one another, and he realised at the outset that the best chance of success lay in working with material of such a nature as to reduce the problem to its simplest terms. He decided that the plant with which he was to work must be normally self-fertilising and unlikely to be crossed through the interference of insects, while at the same time it must possess definite fixed varieties which bred true to type. In the common pea (Pisum sativum) he found the plant he sought. A hardy annual, prolific, easily worked, Pisum has a further advantage in that the insects which normally visit flowers are unable to gather pollen from it and so to bring about cross fertilisation. At the same time it exists in a number of strains presenting well-marked and fixed differences. The flowers may be purple, or red, or white; the plants may be tall or dwarf; the ripe seeds may be yellow or green, round or wrinkled—such are a few of the characters in which the various races of peas differ from one another.

In planning his crossing experiments Mendel adopted an attitude which marked him off sharply from the earlier hybridisers. He realised that their failure to elucidate any general principle of heredity from the results of cross fertilisation was due to their not having concentrated upon particular characters or traced them carefully through a sequence of generations. That source of failure he was careful to avoid, and throughout his experiments he crossed plants presenting sharply contrasted characters, and devoted his efforts to observing the behaviour of these characters in successive generations. Thus in one series of experiments he concentrated his attention on the transmission of the characters tallness and dwarfness, neglecting in so far as these experiments were concerned any other characters in which the parent plants might differ from one another. For this purpose he chose two strains of peas, one of about 6 feet in height, and another of about 1½ feet. Previous testing had shown that each strain bred true to its peculiar height. These two strains were artificially crossed[^[1]] with one another, and it was found to make no difference which was used as the pollen parent and which was used as the ovule parent. In either case the result was the same. The result of crossing tall with dwarf was in every case nothing but talls, as tall or even a little taller than the tall parent. For this reason Mendel termed tallness the dominant and

dwarfness the recessive character. The next stage was to collect and sow the seeds of these tall hybrids. Such seeds in the following year gave rise to a mixed generation consisting of talls and dwarfs but no intermediates. By raising a considerable number of such plants Mendel was able to establish the fact that the number of talls which occurred in this generation was almost exactly three times as great as the number of the dwarfs. As in the previous year, seed were carefully collected from this, the second hybrid generation, and in every case the seeds from each individual plant were harvested separately and separately sown in the following year. By this respect for the individuality of the different plants, however closely they resembled one another, Mendel found the clue that had eluded the efforts of all his predecessors. The seeds collected from the dwarf recessives bred true, giving nothing but dwarfs. And this was true for every dwarf tested. But with the talls it was quite otherwise. Although indistinguishable in appearance, some of them bred true, while others behaved like the original tall hybrids, giving a generation consisting of talls and dwarfs in the proportion of three of

the former to one of the latter. Counting showed that the number of the talls which gave dwarfs was double that of the talls which bred true.

If we denote a dwarf plant as D, a true breeding tall plant as T, and a tall which gives both talls and dwarfs in the ratio 3 : 1 as T(D), the result of these experiments may be briefly summarised in the foregoing scheme.[^[2]]

Mendel experimented with other pairs of contrasted characters and found that in every instance they followed the same scheme of inheritance. Thus coloured flowers were dominant to white, in the ripe seeds yellow was dominant to green, and round shape was dominant to wrinkled, and so on. In every case where the inheritance of an alternative pair of characters was concerned the effect of the cross in successive generations was to produce three and only three different sorts of individuals, viz. dominants which bred true, dominants which gave both dominant and recessive offspring in the ratio 3 : 1, and recessives which always bred true. Having determined a general scheme of inheritance which experiment showed to hold good for each of the seven pairs of alternative characters with which he worked, Mendel set himself to providing a theoretical interpretation of this scheme which, as he clearly realised, must be in terms of germ cells. He

conceived of the gametes as bearers of something capable of giving rise to the characters of the plant, but he regarded any individual gamete as being able to carry one and one only of any alternative pair of characters. A given gamete could carry tallness or dwarfness, but not both. The two were mutually exclusive so far as the gamete was concerned. It must be pure for one or the other of such a pair, and this conception of the purity of the gametes is the most essential part of Mendel's theory.

Scheme of inheritance in the cross of tall with dwarf pea. Gametes represented by small and zygotes by larger circles.

We may now proceed with the help of the accompanying scheme (Fig. 1) to deduce the results that should flow from Mendel's conception of the nature of the gametes, and to see how far they are in accordance with the facts. Since the original tall plant belonged to a strain which bred true, all the gametes produced by it must bear the tall character. Similarly all the gametes of the original dwarf plant must bear the dwarf character. A cross between these two means the union of

a gamete containing tallness with one bearing dwarfness. Owing to the completely dominant nature of the tall character, such a plant is in appearance indistinguishable from the pure tall, but it differs markedly from it in the nature of the gametes to which it gives rise. When the formation of the gametes occurs, the elements representing dwarfness and tallness segregate from one another, so that half of the gametes produced contain the one, and half contain the other of these two elements. For on hypothesis every gamete must be pure for one or other of these two characters. And this is true for the ovules as well as for the pollen grains. Such hybrid F₁ plants, therefore, must produce a series of ovules consisting of those bearing tallness and those bearing dwarfness, and must produce them in equal numbers. And similarly for the pollen grains. We may now calculate what should happen when such a series of pollen grains meets such a series of ovules, i.e. the nature of the generation that should be produced when the hybrid is allowed to fertilise itself. Let us suppose that there are 4x ovules so that 2x are "tall" and 2x are "dwarf." These are brought in contact with a mass of pollen grains of which half are "tall" and half are "dwarf." It is obvious that a "tall" ovule has an equal chance of being fertilised by a "tall" or a "dwarf" pollen grain. Hence of our 2x "tall" ovules, x will be fertilised by "tall" pollen grains and x will be fertilised by "dwarf" pollen grains. The former must give rise to tall

plants, and since the dwarf character has been entirely eliminated from them they must in the future breed true. The latter must also give rise to tall plants, but since they carry also the recessive dwarf character they must when bred from produce both tails and dwarfs. Each of the 2x dwarf ovules, again, has an equal chance of being fertilised by a "tall" or by a "dwarf" pollen grain. Hence x will give rise to tall plants carrying the recessive dwarf character, while x will produce plants from which the tall character has been eliminated, i.e. to pure recessive dwarfs. Consequently from the 4x ovules of the self-fertilised hybrid we ought to obtain 3x tall and x dwarf plants. And of the 3x talls x should breed true to tallness, while the remaining 2x, having been formed like the original hybrid by the union of a "tall" and a "dwarf" gamete, ought to behave like it when bred from and give talls and dwarfs in the ratio 3 : 1. Now this is precisely the result actually obtained by experiment (cf. p. [17]), and the close accord of the experimental results with those deduced on the assumption of the purity of the gametes as enunciated by Mendel affords the strongest of arguments for regarding the nature of the gametes and their relation to the characters of the zygotes in the way that he has done.

It is possible to put the theory to a further test. The explanation of the 3 : 1 ratio of dominants and recessives in the F₂ generation is regarded as due to the F₁ individuals producing equal numbers of gametes bearing the

dominant and recessive elements respectively. If now the F₁ plant be crossed with the pure recessive, we are bringing together a series of gametes consisting of equal numbers of dominants and recessives with a series consisting solely of recessives. We ought from such a cross to obtain equal numbers of dominant and recessive individuals, and further, the dominants so produced ought all to give both dominants and recessives in the ratio 3 : 1 when they themselves are bred from. Both of these expectations were amply confirmed by experiment, and crossing with the recessive is now a recognised way of testing whether a plant or animal bearing a dominant character is a pure dominant, or an impure dominant which is carrying the recessive character. In the former case the offspring will be all of the dominant form, while in the latter they will consist on the average of equal numbers of dominants and recessives.

So far we have been concerned with the results obtained when two individuals differing in a single pair of characters are crossed together and with the interpretation of those results. But Mendel also used plants which differed in more than a single pair of differentiating characters. In such cases he found that each pair of characters followed the same definite rule, but that the inheritance of each pair was absolutely independent of the other. Thus, for example, when a tall plant bearing coloured flowers was crossed with a dwarf plant

bearing white flowers the resulting hybrid was a tall plant with coloured flowers. For coloured flowers are dominant to white, and tallness is dominant to dwarfness. In the succeeding generation there are plants with coloured flowers and plants with white flowers in the proportion of 3 : 1, and at the same time tall plants and dwarf plants in the same proportion. Hence the chances that a tall plant will have coloured flowers are three times as great as its chance of having white flowers. And this is also true for the dwarf plants. As the result of this cross, therefore, we should expect an F₂ generation consisting of four classes, viz. coloured talls, white talls, coloured dwarfs, and white dwarfs, and we should further expect these four forms to appear in the ratio of 9 coloured talls, 3 white talls, 3 coloured dwarfs, and 1 white dwarf. For this is the only ratio which satisfies the conditions that the talls should be to the dwarfs as 3 : 1, and at the same time the coloured should be to the whites as 3 : 1. And these are the proportions that Mendel found to obtain actually in his experiments. Put in a more general form, it may be stated that when two individuals are crossed which differ in two pairs of differentiating characters the hybrids (F₁) are all of the same form, exhibiting the dominant character of each of the two pairs, while the F₂ generation produced by such hybrids consists on the average of 9 showing both dominants, 3 showing one dominant and one recessive,

3 showing the other dominant and the other recessive, and 1 showing both recessive characters. And, as Mendel pointed out, the principle may be extended indefinitely. If, for example, the parents differ in three pair of characters A, B, and C, respectively dominant to a, b, and c, the F₁ individuals will be all of the form ABC, while the F₂ generation will consists of 27 ABC, 9 ABc, 9 AbC, 9 aBC, 3 Abc, 3 aBc, 3 abC, and 1 abc. When individuals differing in a number of alternative characters are crossed together, the hybrid generation, provided that the original parents were of pure strains, consists of plants of the same form; but when these are bred from a redistribution of the various characters occurs. That redistribution follows the same definite rule for each character, and if the constitution of the original parents be known, the nature of the F₂ generation, i.e. the number of possible forms and the proportions in which they occur, can be readily calculated. Moreover, as Mendel showed, we can calculate also the chances of any given form breeding true. To this point, however, we shall return later.

Of Mendel's experiments with beans it is sufficient to say here that they corroborated his more ample work with peas. He is also known to have made experiments with many other plants, and a few of his results are incidentally given in his series of letters to Nägeli the botanist. To the breeding and crossing of bees he also devoted much

time and attention, but unhappily the record of these experiments appears to have been lost. The only other published work that we possess dealing with heredity is a brief paper on some crossing experiments with the hawkweeds (Hieracium), a genus that he chose for working with because of the enormous number of forms under which it naturally exists. By crossing together the more distinct varieties, he evidently hoped to produce some of these numerous wild forms, and so throw light upon their origin and nature. In this hope he was disappointed. Owing in part to the great technical difficulties attending the cross fertilisation of these flowers he succeeded in obtaining very few hybrids. Moreover, the behaviour of those which he did obtain was quite contrary to what he had found in the peas. Instead of giving a variety of forms in the F₂ generation, they bred true and continued to do so as long as they were kept under observation. More recent research has shown that this is due to a peculiar form of parthenogenesis (cf. p. [135]), and not to any failure of the characters to separate clearly from one another in the gametes. Mendel, however, could not have known of this, and his inability to discover in Hieracium any indication of the rule which he had found to hold good for both peas and beans must have been a source of considerable disappointment. Whether for this reason, or owing to the utter neglect of his work by the scientific world, Mendel gave up his experimental

researches during the latter part of his life. His closing years were shadowed with ill-health and embittered by a controversy with the Government on a question of the rights of his monastery. He died of Bright's disease in 1884.

Note.—Shortly after the discovery of Mendel's paper a need was felt for terms of a general nature to express the constitution of individuals in respect of inherited characters, and Bateson accordingly proposed the words homozygote and heterozygote. An individual is said to be homozygous for a given character when it has been formed by two gametes each bearing the character, and all the gametes of a homozygote bear the character in respect of which it is homozygous. When, however, the zygote is formed by two gametes of which one bears the given character while the other does not, it is said to be heterozygous for the character in question, and only half the gametes produced by such a heterozygote bear the character. An individual may be homozygous for one or more characters, and at the same time may be heterozygous for others.

CHAPTER IV

THE PRESENCE AND ABSENCE THEORY

It was fortunate for the development of biological science that the rediscovery of Mendel's work found a small group of biologists deeply interested in the problems of heredity, and themselves engaged in experimental breeding. To these men the extraordinary significance of the discovery was at once apparent. From their experiments, undertaken in ignorance of Mendel's paper, de Vries, Correns, and Tschermak were able to confirm his results in peas and other plants, while Bateson was the first to demonstrate their application to animals. Thenceforward the record has been one of steady progress, and the result of ten years' work has been to establish more and more firmly the fundamental nature of Mendel's discovery. The scheme of inheritance, which he was the first to enunciate, has been found to hold good for such diverse things as height, hairiness, and flower colour and flower form in plants, the shape of pollen grains, and the structure of fruits; while among animals the coat colour of mammals, the form of the feathers and of the comb in poultry, the waltzing habit of Japanese mice, and eye

colour in man are but a few examples of the diversity of characters which all follow the same law of transmission. And as time went on many cases which at first seemed to fall without the scheme have been gradually brought into line in the light of fuller knowledge. Some of these will be dealt with in the succeeding chapters of this book. Meanwhile we may concern ourselves with the single modification of Mendel's original views which has arisen out of more ample knowledge.

A wing feather and a contour feather of an ordinary and a silky fowl. The peculiar ragged appearance of the silky feathers is due to the absence of the little hooks or barbules which hold the barbs together. The silky condition is recessive.

As we have already seen, Mendel considered that in the gamete there was either a definite something

corresponding to the dominant character or a definite something corresponding to the recessive character, and that these somethings whatever they were could not coexist in any single gamete. For these somethings we shall in future use the term factor. The factor, then, is what corresponds in the gamete to the unit-character that appears in some shape or other in the development of the zygote. Tallness in the pea is a unit-character, and the gametes in which it is represented are said to contain the factor for tallness. Beyond their existence in the gamete and their mode of transmission we make no suggestion as to the nature of these factors.

Two double and an ordinary single primula flower. This form of double is recessive to the single.

Fowls' combs. A, pea; B, rose; C, single; D, walnut.

On Mendel's view there was a factor corresponding to the dominant character and another factor corresponding to the recessive character of each alternative pair of unit-characters, and the characters were alternative because no gamete could carry more than one of the two factors belonging to the alternative pair. On the other hand, Mendel supposed that it always carried either one or the other of such a pair. As experimental work proceeded,

it soon became clear that there were cases which could not be expressed in terms of this conception. The nature of the difficulty and the way in which it was met will perhaps be best understood by considering a set of experiments in which it occurred. Many of the different breeds of poultry are characterised by a particular form of comb, and in certain cases the inheritance of these has been carefully worked out. It was shown that the rose comb (Fig. 4, B) with its flattened papillated upper surface and backwardly projecting pike was dominant in the ordinary way to the deeply serrated high single comb (Fig. 4, C) which is characteristic of the Mediterranean races. Experiment also showed that the pea comb (Fig. 4, A), a form with a low central and two well-developed lateral ridges, such as is found in Indian game, behaves as a simple dominant to the single comb. The interesting question arose as to what would happen when the rose and the pea, two forms each dominant to the same third form, were mated together. It seemed reasonable to suppose that things which were alternative to the same thing would be alternative to one another—that either rose or pea would dominate in the hybrids, and that the F₂ generation would consist of dominants and recessives in the ratio 3 : 1. The result of the experiment was, however, very different. The cross rose × pea led to the production of a comb quite unlike either of them. This, the so-called walnut comb (Fig. 4, D),

from its resemblance to the half of a walnut, is a type of comb which is normally characteristic of the Malay fowl. Moreover, when these F₁ birds were bred together, a further unlooked-for result was obtained. As was expected, there appeared in the F₂ generation the three forms walnut, rose, and pea. But there also appeared a definite proportion of single-combed birds, and among many hundreds of chickens bred in this way the proportions in which the four forms walnut, rose, pea, and single appeared was 9 : 3 : 3 : 1.

The Presence and Absence theory is to-day generally accepted by students of these matters. Not only does it afford a simple explanation of the remarkable fact that in all cases of Mendelian inheritance we should be able to express our unit-characters in terms of alternative pairs, but, as we shall have occasion to refer to later, it suggests a clue as to the course by which the various domesticated varieties of plants and animals have arisen from their wild prototypes.

Fowls' combs. A and B, F₁ hen from rose × Breda; C, an F₁ cock from the cross of single × Breda; D, head of Breda cock.

Before leaving this topic we may draw attention to some experiments which offer a pretty confirmation of the view that the rose comb is a single to which a modifying factor for roseness has been added. It was argued that if we could find a type of comb in which the factor for singleness was absent, then on crossing such a comb with a rose we ought, if singleness really underlies rose, to obtain some single combs in F₂ from such a cross. Such a comb we had the good fortune to find in the Breda fowl, a breed largely used in Holland. This fowl is usually spoken of as combless, for the place of the comb is taken by a covering of short bristlelike feathers (Fig. 6, D). In reality it possesses the vestige of a comb in the form of two minute lateral knobs of comb tissue. Characteristic also of this breed is the high development of the horny nostrils, a feature probably correlated with the almost complete absence of comb. The first step in the experiment was to prove the absence of the factor for singleness in the Breda.

On crossing Breda with single the F₁ birds exhibit a large comb of the form of a double single comb in which the two portions are united anteriorly, but diverge from one another towards the back of the head (Fig. 6, C). The Breda contains an element of duplicity which is dominant to the simplicity of the ordinary single comb. But it cannot contain the factor for the single comb, because as soon as that is put into it by crossing with a single the comb

CHAPTER V

INTERACTION OF FACTORS

We have now reached a point at which it is possible to formulate a definite conception of the living organism. A plant or animal is a living entity whose properties may in large measure be expressed in terms of unit-characters, and it is the possession of a greater or lesser number of such unit-characters renders it possible for us to draw sharp distinctions between one individual and another. These unit-characters are represented by definite factors in the gamete which in the process of heredity behave as indivisible entities, and are distributed according to a definite scheme. The factor for this or that unit-character is either present in the gamete or it is not present. It must be there in its entirety or completely absent. Such at any rate is the view to which recent experiment has led us. But as to the nature of these factors, the conditions under which they exist in the gamete, and the manner in which they produce their specific effects in the zygote, we are at present almost completely in the dark.

The case of the fowls' combs opens up the important question of the extent to which the various factors can

influence one another in the zygote. The rose and the pea factors are separate entities, and each when present alone produces a perfectly distinct and characteristic effect upon the single comb, turning it into a rose or a pea as the case may be. But when both are present in the same zygote their combined effect is to produce the walnut comb, a comb which is quite distinct from either and in no sense intermediate between them. The question of the influence of factors upon one another did not present itself to Mendel because he worked with characters which affected different parts of the plant. It was unlikely that the factor which led to the production of colour in the flower would affect the shape of the pod, or that the height of the plant would be influenced by the presence or absence of the factor that determined the shape of the ripe seed. But when several factors can modify the same structure it is reasonable to suppose that they will influence one another in the effects which their simultaneous presence has upon the zygote. By themselves the pea and the rose factors each produce a definite modification of the single comb, but when both are present in the zygote, whether as a single or double dose, the modification that results is quite different to that produced by either when present alone. Thus we are led to the conception of characters which depend for their manifestation on more than one factor in the zygote, and in the present chapter we may consider a few of the

phenomena which result from such interaction between separate and distinct factors.

One of the most interesting and instructive cases in which the interaction between separate factors has been demonstrated is a case in the sweet pea. All white sweet peas breed true to whiteness. And generally speaking the result of crossing different whites is to produce nothing but whites, whether in F₁ or in succeeding generations. But there are certain strains of white sweet peas which when crossed together produce only coloured flowers. The colour may be different in different cases, though for our present purpose we may take a case in which the colour is red. When such reds are allowed to self-fertilise themselves in the normal way and the seeds sown, the resulting F₂ generation consists of reds and whites, the former being rather more numerous than the latter in the proportion of 9 : 7. The raising of a further generation from the seeds of these F₂ plants shows that the whites always breed true to whiteness, but that different reds may behave differently. Some breed true, others give reds and whites in the ratio 3 : 1, while others, again, give reds and whites in the ratio 9 : 7. As in the case of the fowls' combs, this case may be interpreted in terms of the presence and absence of two factors.

The theory was further tested by an examination into the properties of the various F₂ whites which come from a coloured plant that has itself been produced by the mating of two whites. As Fig. 7 shows, these are, in respect of their constitution, of five different kinds, viz. AAbb, Aabb, aaBB, aaBb, and aabb. Since none of them produce anything but whites on self-fertilisation it was found necessary to test their properties in another way, and the method adopted was that of crossing them together. It is obvious that when this is done we should expect different results in different cases. Thus the cross between two whites of the constitution AAbb and aaBB should give nothing but coloured plants; for these two whites are of

the same constitution as the original two whites from which the experiment started. On the other hand, the cross between a white of the constitution aabb and any other white can never give anything but whites. For no white contains both A and B, or it would not be white, and a plant of the constitution aabb cannot supply the complementary factor necessary for the production of colour. Again, two whites of the constitution Aabb and aaBb when crossed should give both coloured and white flowers, the latter being three times as numerous as the former. Without going into further detail it may be stated that the results of a long series of crosses between the various F₂ whites accorded closely with the theoretical explanation.

From the evidence afforded by this exhaustive set of experiments it is impossible to resist the deduction that the appearance of colour in the sweet pea depends upon the interaction of two factors which are independently transmitted according to the ordinary scheme of Mendelian inheritance. What these factors are is still an open question. Recent evidence of a chemical nature indicates that colour in a flower is due to the interaction of two definitive substances: (1) a colourless "chromogen," or colour basis; and (2) a ferment which behaves as an activator of the chromogen, and by inducing some process of oxidation, leads to the formation of a coloured substance. But whether these two bodies exist as such

in the gametes or whether in some other form we have as yet no means of deciding.

Since the elucidation of the nature of colour in the sweet pea phenomena of a similar kind have been witnessed in other plants, notably in stocks, snapdragons, and orchids. Nor is this class of phenomena confined to plants. In the course of a series of experiments upon the plumage colour of poultry, indications were obtained that different white breeds did not always owe their whiteness to the same cause. Crosses were accordingly made between the white Silky fowl and a pure white strain derived from the white Dorking. Each of these had been previously shown to behave as a simple recessive to colour. When the two were crossed only fully coloured birds resulted. From analogy with the case of the sweet pea it was anticipated that such F₁ coloured birds when bred together would produce an F₂ generation consisting of coloured and white birds in the ratio 9 : 7, and when the experiment was made this was actually shown to be the case. With the growth of knowledge it is probable that further striking parallels of this nature between the plant and animal worlds will be met with.

Before quitting the subject of these experiments attention may be drawn to the fact that the 9 : 7 ratio is in reality a 9 : 3 : 3 : 1 ratio in which the last three terms are indistinguishable owing to the special circumstances that neither factor can produce a visible effect without

the co-operation of the other. And we may further emphasise the fact that although the two factors thus interact upon one another they are nevertheless transmitted quite independently and in accordance with the ordinary Mendelian scheme.

One of the earliest sets of experiments demonstrating the interaction of separate factors was that made by the French zoologist Cuénot on the coat colours of mice. It was shown that in certain cases agouti, which is the colour of the ordinary wild grey mouse, behaves as a dominant to the albino variety, i.e. the F₂ generation from such a cross consists of agoutis and albinos in the ratio 3 : 1. But in other cases the cross between albino and agouti gave a different result. In the F₁ generation appeared only agoutis as before, but the F₂ generation consisted of three distinct types, viz. agoutis, albinos, and blacks. Whence the sudden appearance of the new type? The answer is a simple one. The albino parent was really a black. But it lacked the factor without which the colour is unable to develop, and consequently it remained an albino. If we denote this factor by C, then the constitution of an albino must be cc, while that of a coloured animal may be CC or Cc, according as to whether it breeds true to colour or can

throw albinos. Agouti was previously known to be a simple dominant to black, i.e. an agouti is a black rabbit plus an additional greying factor which modifies the black into agouti. This factor we will denote by G, and we will use B for the black factor. Our original agouti and albino parents we may therefore regard as in constitution GGCCBB and ggccBB respectively. Both of the parents are homozygous for black. The gametes produced by the two parents are GCB, and gcB, and the constitution of the F₁ animals must be GgCcBB. Being heterozygous for two factors they will produce four kinds of gametes in equal numbers, viz. GCB, GcB, gCB, and gcB. The results of the mating of two such similar series of gametes when the F₁ animals are bred together we may determine by the usual "chessboard" method (Fig. 8). Out of the 16 squares 9 contain both C and G in addition to B. Such animals must be agoutis. Three squares contain C but not G. Such animals must be coloured, but as they do not contain the modifying agouti factor their colour will be black. The remaining four squares do not contain C, and in the absence of this colour-developing factor they must all be albinos. Theory demands that the three classes agouti, black, and albino should appear in F₂ in the ratio 9 : 3 : 4; experiment has shown that these are the only classes that appear, and that the proportions in which they are produced accord closely with the theoretical expectation. Put briefly, then, the explanation

Though albinos, whether mice, rabbits, rats, or other animals, breed true to albinism, and though albinism behaves as a simple recessive to colour, yet albinos may be of many different sorts. There are in fact just as many kinds of albinos as there are coloured forms—neither more nor less. And all these different kinds of albinos may breed together, transmitting the various colour factors according to the Mendelian scheme of inheritance,

and yet the visible result will be nothing but albinos. Under the mask of albinism is all the while occurring that segregation of the different colour factors which would result in all the varieties of coloured forms, if only the essential factor for colour development were present. But put in the developer by crossing with a pure coloured form and their variety of constitution can then at last become manifest.

So far we have dealt with cases in which the production of a character is dependent upon the interaction of two factors. But it may be that some characters require the simultaneous presence of a greater number of factors for their manifestation, and the experiments of Miss Saunders have shown that there is a character in stocks which is unable to appear except through the interaction of three distinct factors. Coloured stocks may be either hoary, with the leaves and stem covered by small hairs, or they may lack the hairy covering, in which case they are termed glabrous. Hoariness is dominant to glabrousness; that is to say, there is a definite factor which can turn the glabrous into a hoary plant when it is present. But in families where coloured and white stocks occur the white are always glabrous, while the coloured plants may or may not be hoary. Now colour in the stock as in the sweet pea has been proved to be dependent upon the interaction of two separate factors. Hence hoariness depends upon three separate factors, and a stock cannot be hoary unless

it contains the hoary factor in addition to the two colour factors. It requires the presence of all these three factors to produce the hoary character, though how this comes about we have not at present the least idea.

Sections of primula flowers. The anthers are shown as black. A, "pin" form with long style and anthers set low down; B, "thrum" form with short style and anthers set higher up; C, homostyle form with anthers set low down as in "pin," but with short style. This form only occurs with the large eye.

Two primula flowers showing the extent of the small and of the large eye.

A somewhat different and less usual form of interaction between factors may be illustrated by a case in primulas recently worked out by Bateson and Gregory. Like the common primrose, the primula exhibits both pin-eyed and thrum-eyed varieties. In the former the style is long, and the centre of the eye is formed by the end of the stigma which more or less plugs up the opening of the corolla (cf. Fig. 9, A); in the latter the style is short and hidden by the four anthers which spring from higher up in the corolla and form the centre of the eye (cf. Fig. 9, B). The greater part of the "eye" is formed by the greenish-yellow patches on each petal just at the opening

of the corolla. In most primulas the eye is small, but there are some in which it is large and extends as a flush over a considerable part of the petals (Fig. 10). Experiments showed that these two pairs of characters behave in simple Mendelian fashion, short style ( = "thrum") being dominant to long style (= "pin") and small eye dominant to large. Besides the normal long and short styled forms, there occurs a third form, which has been termed homostyle. In this form the anthers are placed low down in the corolla tube as they are in the long-styled form, but the style remains short instead of reaching up to the corolla opening (Fig. 9, C). In the course of their experiments Bateson and Gregory crossed a large-eyed homostyle plant with a small-eyed thrum ( = short style). The F₁ plants were all short styled with small eyes.

CHAPTER VI

REVERSION

As soon as the idea was grasped that characters in plants and animals might be due to the interaction of complementary factors, it became evident that this threw clear light upon the hitherto puzzling phenomenon of reversion. We have already seen that in certain cases the cross between a black mouse or rabbit and an albino, each belonging to true breeding strains, might produce nothing but agoutis. In other words, the cross between the black and the white in certain instances results in a complete reversion to the wild grey form. Expressed in Mendelian terms, the production of the agouti was the necessary consequence of the meeting of the factors C and G in the same zygote. As soon as they are brought together, no matter in what way, the reversion is bound to occur. Reversion, therefore, in such cases we may regard as the bringing together of complementary factors which had somehow in the course of evolution become separated from one another. In the simplest cases, such as that of the black and the white rabbit, only two factors are concerned, and one of them is brought in from each of the

two parents. But in other cases the nature of the reversion may be more complicated owing to a larger number of factors being concerned, though the general principle remains the same. Careful breeding from the reversions will enable us in each case to determine the number and nature of the factors concerned, and in illustration of this we may take another example from rabbits. The Himalayan rabbit is a well-known breed. In appearance it is a white rabbit with pink eyes, but the ears, paws, and nose are black (Pl. I., 2). The Dutch rabbit is another well-known breed. Generally speaking, the anterior portion of the body is white, and the posterior part coloured. Anteriorly, however, the eyes are surrounded by coloured patches extending up to the ears, which are entirely coloured. At the same time the hind paws are white (cf. Pl. I., 1). Dutch rabbits exist in many varieties of colour, though in each one of these the distribution of colour and white shows the same relations. In the experiments about to be described a yellow Dutch rabbit was crossed with a Himalaya. The result was a reversion to the wild agouti colour (Pl. I., 3). Some of the F₁ individuals showed white patches, while others were self-coloured. On breeding from the F₁ animals a series of coloured forms resulted in F₂. These were agoutis, blacks, yellows, and sooty yellows, the so-called tortoise shells of the fancy (Pl. I., 4-7).

1, Yellow Dutch Rabbit; 2, Himalayan; 3, Agouti ( = grey) F₁ reversion; 4-8, F₂ types, viz.: 4, Agouti; 5, Yellow; 6, Black; 7, Tortoiseshell; 8, Himalayan.

In addition to these appeared Himalayans with either black points or with lighter brownish ones, and the proportions in which they came showed the Himalayan character to be a simple recessive. A certain number of the coloured forms exhibited the Dutch marking to a greater or less extent, but as its inheritance in this set of experiments is complicated and has not yet been worked out, we may for the present neglect it and confine our attention to the coloured types and to the Himalayans. The proportion in which the four coloured types appeared in F₂ was very nearly 9 agoutis, 3 blacks, 3 yellows, and 1 tortoiseshell. Evidently we are here dealing with two factors: (1) the grey factor (G), which modifies black into agouti, or tortoiseshell into yellow; and (2) an intensifying factor (I), which intensifies yellow into agouti and tortoiseshell into black. It may be mentioned here that other experiments confirmed the view that the yellow rabbit is a dilute agouti, and the tortoiseshell a dilute black. The Himalayan pattern behaves as a recessive to self-colour. It is a self-coloured black rabbit lacking a factor that allows the colour to develop except in the points. That factor we may denote

by X, and as far as it is concerned the Himalayan is constitutionally xx. The Himalayan contains the intensifying factor, for such pigment as it possesses in the points is full coloured. At the same time it is black, i.e. lacking in the factor G. With regard to these three factors, therefore, the constitution of the Himalayan is ggIIxx. The last character which we have to consider in this cross is the Dutch character. This was found by Hurst to behave as a recessive to self-colour (S), and for our present purpose we will regard it as differing from a self-coloured rabbit in the lack of this factor.[^[3]] The Himalayan is really a self-coloured animal, which, however, is unable to show itself as a full black owing to its not possessing the factor X. The results of breeding experiments then suggest that we may denote the Himalayan by the formula ggIIxxSS and the yellow Dutch by GGiiXXss. Each lacks two of the factors upon the full complement of which the agouti colour depends. By crossing them the complete series GIXS is brought into the same zygote, and the result is a reversion to the colour of the wild rabbit.

Most of the instances of reversion yet worked out are those in which colour characters are concerned. The sweet pea, however, supplies us with a good example of reversion in structural characters. A dwarf variety known as the "Cupid" has been extensively grown for

some years. In these little plants the internodes are very short and the stems are few in number, and attain to a length of only 9-10 inches. In course of growth they diverge from one another, and come to lie prostrate on the ground (Pl. II., 2). Curiously enough, although the whole plant is dwarfed in other respects, this does not seem to affect the size of the flower, which is that of a normal sweet pea. Another though less-known variety is the "Bush" sweet pea. Its name is derived from its habit of growth. The numerous stems do not diverge from one another, but all grow up side by side, giving the plant the appearance of a compact bush (Pl. II., 1). Under ordinary conditions it attains a height of 3½-4 feet. A number of crosses were made between the Bush and Cupid varieties, with the somewhat unexpected result that in every instance the F₁ plants showed complete reversion to the size and habit of the ordinary tall sweet pea (Pl. II., 3), which is the form of the wild plant as it occurs in Sicily to-day. The F₂ generation from these reversionary talls consisted of four different types, viz.

talls, bushes, Cupids of the procumbent type like the original Cupid parent, and Cupids with the compact upright Bush habit (Pl. II., 4). These four types appeared in the ratio 9 : 3 : 3 : 1, and this, of course, provided the clue to the nature of the case. The characters concerned are (1) long internode of stem between the leaves which is dominant to short internode, and (2) the creeping procumbent habit which is dominant to the erect bush-like habit. Of these characters length of internode was carried by the Bush, and the procumbent habit by the original Cupid parent. The bringing of them together by the cross resulted in a procumbent plant with long internodes. This is the ordinary tall sweet pea of the wild Sicilian type, reversion here, again, being due to the bringing together of two complementary factors which had somehow become separated in the course of evolution.

To this interpretation it may be objected that the ordinary sweet pea is a plant of upright habit. This, however, is not true. It only appears so because the conventional way of growing it is to train it up sticks. In reality it is of procumbent habit, with divergent stems like the ordinary Cupid, a fact which can easily be observed by anyone who will watch them grow without the artificial aid of prepared supports.

1, Bush Sweet Pea; 2, Cupid Sweet Pea; 3, F₁ reversionary Tall; 4, Erect Cupid Sweet Pea; 5, Purple Invincible; 6, Painted Lady; 7, Duke of Westminster (hooded standard).

The cases of reversion with which we have so far dealt have been cases in which the reversion occurs as an immediate result of a cross, i.e. in the F₁ generation. This is perhaps the commonest mode of reversion, but instances are known in which the reversion that occurs when two pure types are crossed does not appear until the F₂ generation. Such a case we have already met with in the fowls' combs. It will be remembered that the cross between pure pea and pure rose gave walnut combs in F₁, while in the F₂ generation a definite proportion, 1 in 16, of single combs appeared (cf. p. [32]). Now the single comb is the form that is found in the wild jungle fowl, which is generally regarded as the ancestor of the domestic breeds. If this is so, we have a case of reversion in F₂; and this in the absence of the two factors brought together by the rose-comb and pea-comb parents. Instead of the reversion being due to the bringing together of two complementary factors, we must regard it here as due to the association of two complementary absences. To this question, however, we shall revert later in discussing the origin of domesticated varieties.

There is one other instance of reversion to which we must allude. This is Darwin's famous case of the occasional appearance of pigeons reverting to the wild blue rock (Columba livia), when certain domesticated races are crossed together.[^[4]] As is well known, Darwin made use of this as an argument for regarding all the domesticated varieties as having arisen from the same wild species. The original experiment is somewhat complicated, and is shown in the accompanying scheme. Essentially it lay in

following the results flowing from crosses between blacks and whites. Experiments recently made by Staples-Browne have shown that this case of reversion also can be readily interpreted in Mendelian terms. In these experiments the cross was made between black barbs and white fantails.

Diagram to illustrate the appearance of the reversionary blue pigeon in F₂ from the cross of black with white.

CHAPTER VII

DOMINANCE

Primula flowers to illustrate the intermediate nature of the F₁ flower when sinensis is crossed with stellata.

In the cases which we have hitherto considered the presence of a factor produces its full effect whether it is introduced by both of the gametes which go to form the zygote, or by one of them alone. The heterozygous tall pea or the heterozygous rose-combed fowl cannot be distinguished from the homozygous form by mere inspection, however close. Breeding tests alone can decide which is the heterozygous and which the homozygous form. Though this is true for the majority of characters yet investigated, there are cases known in which the heterozygous form differs in appearance from either parent. Among plants such a case has been met with in the primula. The ordinary Chinese primula (P. sinensis) (Fig. 12) has large rather wavy petals much crenated at the edges. In the Star Primula (P. stellata) the flowers are much smaller, while the petals are flat and present only a terminal notch instead of the numerous crenations of P. sinensis. The heterozygote produced by crossing these forms is intermediate in size and appearance. When self-fertilised such plants behave in simple Mendelian fashion,

giving a generation consisting of sinensis, intermediates, and stellata in the ratio 1 : 2 : 1. Subsequent breeding from these plants showed that both the sinensis and stellata which appeared in the F₂ generation bred true, while the intermediates always gave all three forms again in the same proportion. But though there is no dominance of the character of either parent in such a case as this, the Mendelian principle of segregation could hardly have a better illustration.

Among birds a case of similar nature is that of the Blue Andalusian fowl. Fanciers have long recognised the difficulty of getting this variety to breed true. Of a slaty blue colour itself with darker hackles and with black lacing on the feathers of the breast, it always throws "wasters" of two kinds, viz. blacks, and whites splashed with black. Careful breeding from the blues shows that the three sorts are always produced in the same definite

proportions, viz., one black, two blues, one splashed white. This at once suggests that the black and the splashed white are the two homozygous forms, and that the blues are heterozygous, i.e., producing equal numbers of "black" and "white splashed" gametes. The view was tested by breeding the "wasters" together—black with black, and splashed white with splashed white—and it was found that each bred true to its respective type. But when the black and the splashed white were crossed they gave, as was expected, nothing but blues. In other words, we have the seeming paradox of the black and the splashed white producing twice as many blues as do the blues when bred together. The black and the splashed white "wasters" are in reality the pure breeds, while the "pure" Blue Andalusian is a mongrel which no amount of selection will ever be able to fix.

In such cases as this it is obvious that we cannot speak of dominance. And with the disappearance of this phenomenon we lose one criterion for determining which of the two parent forms possesses the additional factor. Are we, for example, to regard the black Andalusian as a splashed white to which has been added a double dose of a colour-intensifying factor, or are we to consider the white splashed bird as a black which is unable to show its true pigmentation owing to the possession of some inhibiting factor which prevents the manifestation of the black. Either interpretation fits the facts equally well,

and until further experiments have been devised and carried out it is not possible to decide which is the correct view.

Besides these comparatively rare cases where the heterozygote cannot be said to bear a closer resemblance to one parent more than to the other, there are cases in which it is often possible to draw a visible distinction between the heterozygote and the pure dominant. There are certain white breeds of poultry, notably the White Leghorn, in which the white behaves as a dominant to colour. But the heterozygous whites made by crossing the dominant white birds with a pure coloured form (such as the Brown Leghorn) almost invariably show a few coloured feathers or "ticks" in their plumage. The dominance of white is not quite complete, and renders it possible to distinguish the pure from the impure dominant without recourse to breeding experiments.

Diagram to illustrate the nature of the F₂ generation from the cross between dominant white and recessive white fowls.

This case of the dominant white fowl opens up another interesting problem in connection with dominance. By accepting the "Presence and Absence" hypothesis we are committed to the view that the dominant form possesses an extra factor as compared with the recessive. The natural way of looking at this case of the fowl is to regard white as the absence of colour. But were this so, colour should be dominant to white, which is not the case. We are therefore forced to suppose that the absence of colour in this instance is due to the presence of a factor whose

property is to inhibit the production of colour in what would otherwise be a pure coloured bird. On this view the dominant white fowl is a coloured bird plus a factor which inhibits the development of the colour. The view can be put to the test of experiment. We have already seen that there are other white fowls in which white is recessive to colour, and that the whiteness of such birds is due to the fact that they lack a factor for the development of colour. If we denote this factor by C and our postulated inhibitor factor in the dominant white bird by I, then we must write the constitution of the recessive white as ccii, and the dominant white as CCII. We may now work out the results we ought to obtain when a cross is made between these two pure white breeds. The constitution of the F₁ bird must be CcIi. Such birds being heterozygous for the inhibitor factor, should be whites showing some coloured "ticks." Being heterozygous for both of the two factors C and I, they will produce in equal numbers the four different sorts of gametes CI, Ci, cI, ci. The result of bringing two such similar series of gametes together is shown in Fig. 13. Out of the sixteen squares, twelve contain I; these will be white birds either with or without a few coloured ticks. Three contain C but not I: these must be coloured birds. One contains neither C nor I; this must be a white. From such a mating we ought, therefore, to obtain both white and coloured birds in the ratio 13 : 3. The results thus theoretically

deduced were found to accord with the actual facts of experiment. The F₁ birds were all "ticked" whites, and in the F₂ generation came white and coloured birds in the expected ratio. There seems, therefore, little reason to doubt that the dominant white is a coloured bird in which the absence of colour is due to the action of a colour-inhibiting factor, though as to the nature of that factor we can at present make no surmise. It is probable that other facts, which at first sight do not appear to be in agreement with the "Presence and Absence" hypothesis, will eventually be brought into line through the action of inhibitor factors. Such a case, for instance, is that of bearded and beardless wheats. Though the beard is obviously the additional character, the bearded condition is recessive to the beardless. Probably we ought to regard the beardless as a bearded wheat in which there is an inhibitor that stops the beard from growing. It is not unlikely that as time goes on we shall

find many more such cases of the action of inhibitor factors, and we must be prepared to find that the same visible effect may be produced either by the addition or by the omission of a factor. The dominant and recessive white poultry are indistinguishable in appearance. Yet the one contains a factor more and the other a factor less than the coloured bird.

Ears of beardless and bearded wheat. The beardless condition is dominant to the bearded.

A phenomenon sometimes termed irregularity of dominance has been investigated in a few cases. In certain breeds of poultry such as Dorkings there occurs an extra toe directed backwards like the hallux (cf. Fig. 15). In some families this character behaves as an ordinary dominant to the normal, giving the expected 3 : 1 ratio in F₂. But in other families similarly bred the proportions of birds with and without the extra toe appear to be unusual. It has been shown that in such a family some of the birds without the extra toe may nevertheless transmit the peculiarity when mated with birds belonging to strains in which the extra toe never occurs. Though the external appearance of the bird generally affords some indication of the nature of the gametes which it is carrying, this is not always the case. Nevertheless we have reason to suppose that the character segregates in the gametes, though the nature of these cannot always be decided from the appearance of the bird which bears them.

Fowls' feet. On the right a normal and on the left one with an extra toe.

Scheme to illustrate the inheritance of horns in sheep. Heterozygous males shown dark with a white spot, heterozygous females light with a dark spot in the centre.

There are cases in which an apparent irregularity of dominance has been shown to depend upon another character, as in the experiments with sheep carried out by Professor Wood. In these experiments two breeds were crossed, of which one, the Dorset, is horned in both sexes, while the other, the Suffolk, is without horns in either sex. Whichever way the cross was made the resulting F₁ generation was similar; the rams were horned, and

the ewes were hornless. In the F₂ generation raised from these F₁ animals both horned and hornless types appeared in both sexes but in very different proportions. While the horned rams were about three times as numerous as the hornless, this relation was reversed among the females, in which the horned formed only about one-quarter of the total. The simplest explanation of this interesting case is to suppose that the dominance of the horned character depends upon the sex of the animal—that it is dominant in the male but recessive in the female. A pretty experiment was devised for putting this view to the test. If it is true, equal numbers of gametes with and without the horned factor must be produced by the F₁ ewes, while the factor should be lacking in all the gametes of the hornless F₂ rams. A

hornless ram, therefore, put to a flock of F₁ ewes should give rise to equal numbers of zygotes which are heterozygous for the horned character, and of zygotes in which it is completely absent. And since the heterozygous males are horned, while the heterozgyous females are hornless, we should expect from this mating equal numbers of horned and hornless rams, but only hornless ewes. The result of the experiment confirmed this expectation. Of the ram lambs 9 were horned and 8 were hornless, while all the 11 ewe lambs were completely destitute of horns.

Sheep

CHAPTER VIII

WILD FORMS AND DOMESTIC VARIETIES

In discussing the phenomena of reversion we have seen that in most cases such reversion occurs when the two varieties which are crossed each contain certain factors lacking in the other, of which the full complement is necessary for the production of the reversionary wild form. This at once suggests the idea that the various domestic forms of animals and plants have arisen by the omission from time to time of this factor or of that. In some cases we have clear evidence that this is the most natural interpretation of the relation between the cultivated and the wild forms. Probably the species in which it is most evident is the sweet pea (Lathyrus odoratus). We have already seen reason to suppose that as regards certain structural features the Bush variety is a wild lacking the factor for the procumbent habit, that the Cupid is a wild without the factor for the long inter-node, and that the Bush Cupid is a wild minus both these factors. Nor is the evidence less clear for the many colour varieties. In illustration we may consider in more detail a case in which the cross between two whites resulted

in a complete reversion to the purple colour characteristic of the wild Sicilian form (Pl. IV.). In this particular instance subsequent breeding from the purples resulted in the production of six different colour forms in addition to whites. The proportion of the coloured forms to the whites was 9 : 7 (cf. p. [44]), but it is with the relation of the six coloured forms that we are concerned here. Of these six forms three were purples and three were reds. The three purple forms were (1) the wild bicolor purple with blue wings known in cultivation as the Purple Invincible (Pl. IV., 4); (2) a deep purple with purple wings (Pl. IV., 5); and (3) a very dilute purple known as the Picotee (Pl. IV., 6). Corresponding to these three purple forms were three reds: (1) a bicolor red known as Painted Lady (Pl. IV., 7); (2) a deep red with red wings known as Miss Hunt (Pl. IV., 8); and (3) a very pale red which we have termed Tinged White[^[5]] (Pl. IV., 9). In the F₂ generation the total number of purples bore to the total number of reds the ratio 3 : 1, and this ratio was maintained for each of the corresponding classes. Purple, therefore, is dominant to red, and each of the three classes of red differs from its corresponding purple in not possessing the blue factor (B) which turns it into purple.

1, 2, Emily Henderson; 3, F₁ reversionary Purple; 4-10, Various F₂ forms: 4, Purple; 5, Deep Purple; 6, Picotee; 7, Painted Lady; 8, Miss Hunt; 9, Tinged White; 10, White.

Again, the proportion in which the three classes of purples appeared was 9 bicolors, 3 deep purples, 4 picotees. We are, therefore, concerned here with the operation of two factors: (1) a light wing factor, which renders the bicolor dominant to the dark winged form; and (2) a factor for intense colour, which occurs in the bicolor and in the deep purple, but is lacking in the dilute picotee. And here it should be mentioned that these conclusions rest upon an exhaustive set of experiments involving the breeding of many thousands of plants. In this cross, therefore, we are concerned with the presence or absence of five factors, which we may denote as follows:—

A colour base, R.

A colour developer, C.

A purple factor, B.

A light wing factor, L.

A factor for intense colour, I.

On this notation our six coloured forms are:—

It will be noticed in this series that the various coloured

forms can be expressed by the omission of one or more factors from the purple bicolor of the wild type. With the complete omission of each factor a new colour type results, and it is difficult to resist the inference that the various cultivated forms of the sweet pea have arisen from the wild by some process of this kind. Such a view tallies with what we know of the behaviour of the wild form when crossed by any of the garden varieties. Wherever such crossing has been made the form of the hybrid has been that of the wild, thus supporting the view that the wild contains a complete set of all the differentiating factors which are to be found in the sweet pea.

Moreover, this view is in harmony with such historical evidence as is to be gleaned from botanical literature, and from old seedsmen's catalogues. The wild sweet pea first reached England in 1699, having been sent from Sicily by the monk Franciscus Cupani as a present to a certain Dr. Uvedale in the county of Middlesex. Somewhat later we hear of two new varieties, the red bicolor, or Painted Lady, and the white, each of which may be regarded as having "sported" from the wild purple by the omission of the purple factor, or of one of the two colour factors. In 1793 we find a seedsman offering also what he called black and scarlet varieties. It is probable that these were our deep purple and Miss Hunt varieties, and that somewhere about this time the factor for the

light wing (L) was dropped out in certain plants. In 1860 we have evidence that the pale purple or Picotee, and with it doubtless the Tinged White, had come into existence. This time it was the factor for intense colour which had dropped out. And so the story goes on until the present day, and it is now possible to express by the same simple method the relation of the modern shades, of purple and reds, of blues and pinks, of hooded and wavy standards, to one another and to the original wild form. The constitution of many of these has now been worked out, and to-day it would be a simple though perhaps tedious task to denote all the different varieties by a series of letters indicating the factors which they contain, instead of by the present system of calling them after kings and queens, and famous generals, and ladies more or less well known.

From what we know of the history of the various strains of sweet peas one thing stands out clearly. The new character does not arise from a pre-existing variety by any process of gradual selection, conscious or otherwise. It turns up suddenly complete in itself, and thereafter it can be associated by crossing with other existing characters to produce a gamut of new varieties. If, for example, the character of hooding in the standard (cf. Pl. II., 7) suddenly turned up in such a family as that shown on Plate IV. we should be able to get a hooded form corresponding to each of the forms with the erect

standard; in other words, the arrival of the new form would give us the possibility of fourteen varieties instead of seven. As we know, the hooded character already exists. It is recessive to the erect standard, and we have reason to suppose that it arose as a sudden sport by the omission of the factor in whose presence the standard assumes the erect shape characteristic of the wild flower. It is largely by keeping his eyes open and seizing upon such sports for crossing purposes that the horticulturist "improves" the plants with which he deals. How these sports or mutations come about we can now surmise. They must owe their origin to a disturbance in the processes of cell division through which the gametes originate. At some stage or other the normal equal distribution of the various factors is upset, and some of the gametes receive a factor less than others. From the union of two such gametes, provided that they are still capable of fertilisation, comes the zygote which in course of growth develops the new character.

Why these mutations arise: what leads to the surmised unequal division of the gametes: of this we know practically nothing. Nor until we can induce the production of mutations at will are we likely to understand the conditions which govern their formation. Nevertheless there are already hints scattered about the recent literature of experimental biology which lead us to hope that we may know more of these matters in the future.

In respect of the evolution of its now multitudinous varieties, the story of the sweet pea is clear and straightforward. These have all arisen from the wild by a process of continuous loss. Everything was there in the beginning, and as the wild plant parted with factor after factor there came into being the long series of derived forms. Exquisite as are the results of civilization, it is by the degradation of the wild that they have been brought about. How far are we justified in regarding this as a picture of the manner in which evolution works?

There are certainly other species in which we must suppose that this is the way that the various domesticated forms have arisen. Such, for example, is the case in the rabbit, where most of the colour varieties are recessive to the wild agouti form. Such also is the case in the rat, where the black and albino varieties and the various pattern forms are also recessive to the wild agouti type. And with the exception of a certain yellow variety to which we shall refer later, such is also the case with the many fancy varieties of mice.

Nevertheless there are other cases in which we must suppose evolution to have proceeded by the interpolation of characters. In discussing reversion on crossing, we have already seen that this may not occur until the F₂ generation, as, for example, in the instance of the fowls' combs (cf. p. [65]). The reversion to the single comb occurred as the result of the removal of the two factors

for rose and pea. These two domesticated varieties must be regarded as each possessing an additional factor in comparison with the wild single-combed bird. During the evolution of the fowl, these two factors must be conceived of as having been interpolated in some way. And the same holds good for the inhibitory factor on which, as we have seen, the dominant white character of certain poultry depends. In pigeons, too, if we regard the blue rock as the ancestor of the domesticated breeds, we must suppose that an additional melanic factor has arisen at some stage. For we have already seen that black is dominant to blue, and the characters of F₁, together with the greater number of blacks than blues in F₂, negatives the possibility that we are here dealing with an inhibitory factor. The hornless or polled condition of cattle, again, is dominant to the horned condition, and if, as seems reasonable, we regard the original ancestors of domestic cattle as having been horned, we have here again the interpolation of an inhibitory factor somewhere in the course of evolution.

On the whole, therefore, we must be prepared to admit that the evolution of domestic varieties may come about by a process of addition of factors in some cases and of subtraction in others. It may be that what we term additional factors fall into distinct categories from the rest. So far, experiment seems to show that they are either of the nature of melanic factors, or of inhibitory

factors, or of reduplication factors as in the case of the fowls' combs. But while the data remain so scanty, speculation in these matters is too hazardous to be profitable.

CHAPTER IX

REPULSION AND COUPLING OF FACTORS

Although different factors may act together to produce specific results in the zygote through their interaction, yet in all the cases we have hitherto considered the heredity of each of the different factors is entirely independent. The interaction of the factors affects the characters of the zygote, but makes no difference to the distribution of the separate factors, which is always in strict accordance with the ordinary Mendelian scheme. Each factor in this respect behaves as though the other were not present.

A few cases have been worked out in which the distribution of the different factors to the gametes is affected by their simultaneous presence in the zygote. And the influence which they are able to exert upon one another in such cases is of two kinds. They may repel one another, refusing, as it were, to enter into the same zygote, or they may attract one another, and, becoming linked together, pass into the same gamete, as it were by preference. For the moment we may consider these two sets of phenomena apart.

One of the best illustrations of repulsion between factors occurs in the sweet pea. We have already seen that the loss of the blue or purple factor (B) from the wild bicolor results in the formation of the red bicolor known as Painted Lady (Pl. IV., 7). Further, we have seen that the hooded standard is recessive to the ordinary erect standard. The omission of the factor for the erect standard (E) from the purple bicolor (Pl. II., 5) results in a hooded purple known as Duke of Westminster (Pl. II., 7). And here it should be mentioned that in the corresponding hooded forms the difference in colour between the wings and standard is not nearly so marked as in the forms with the erect standard, but the difference in structure appears to affect the colour, which becomes nearly uniform. This may be readily seen by comparing the picture of the purple bicolor on Plate II. with that of the Duke of Westminster flower.

Now when a Duke of Westminster is mated with a Painted Lady the factor for erect standard (E) is brought in by the red, and that for blue (B) by the Duke, and the offspring are consequently all purple bicolors. Purples so formed are all heterozygous for these two factors, and were the case a simple one, such as those which have already been discussed, we should expect the F₂ generation to consist of the four forms: erect purple, hooded purple, erect red, and hooded red in the ratio 9 : 3 : 3 : 1. Such, however, is not the case. The F₂ generation

actually consists of only three forms, viz. erect red, erect purple, and hooded purple, and the ratio in which these three forms occur is 1 : 2 : 1. No hooded red has been known to occur in such a family. Moreover further breeding shows that while the erect reds and the hooded purples always breed true, the erect purples in such families never breed true, but always behave like the original F₁ plant, giving the three forms again in the ratio 1 : 2 : 1. Yet we know that there is no difficulty in getting purple bicolors to breed true from other families; and we know also that hooded red sweet peas exist in other strains.

On the assumption that there exists a repulsion between the factors for erect standard and blue in a plant which is heterozygous for both, this peculiar case receives a simple explanation. The constitutions of the erect red and the hooded purple are EEbb and eeBB respectively and that of the F₁ erect purple is EeBb. Now let us suppose that in such a zygote there exists a repulsion

between E and B, such that when the plant forms gametes these two factors will not go into the same gamete. On this view it can only form two kinds of gametes, viz. Eb and eB, and these, of course, will be formed in equal numbers. Such a plant on self-fertilisation must give the zygotic series EEbb + 2 EeBb + eeBB, i.e. 1 erect red, 2 erect purples, and 1 hooded purple. And because the erect reds and the hooded purples are respectively homozygous for E and B, they must thenceforward breed true. The erect purples, on the other hand, being always formed by the union of a gamete Eb with a gamete eB, are always heterozygous for both of these factors. They can, consequently, never breed true, but must always give erect reds, erect purples, and hooded purples in the ratio 1 : 2 : 1. The experimental facts are readily explained on the assumption of repulsion between the two

factors B and E during the formation of the gametes in a plant which is heterozygous for both.

Other similar cases of factorial repulsion have been demonstrated in the sweet pea, and two of these are also concerned with the two factors with which we have just been dealing. Two distinct varieties of pollen grains occur in this species, viz. the ordinary oblong form and a rather smaller rounded grain. The former is dominant to the latter.[^[7]] When a cross is made between a purple with round pollen and a red with long pollen the F₁ plant is a long pollened purple. But the F₂ generation consists of purples with round pollen, purples with long pollen, and reds with long pollen in the ratio 1 : 2 : 1. No red with round pollen appears in F₂ owing to repulsion between the factors for purple (B) and for long pollen (L). Similarly plants produced by crossing a red hooded long with a red round having an erect standard give in F₁ long pollened reds with an erect standard, and these in F₂ produce the three types, round pollened erect, long pollened erect, and long pollened hooded, in the ratio 1 : 2 : 1. The repulsion here is between the long pollen factor (L) and the factor for the erect standard (E).

Yet another similar case is known in which we are concerned with quite different factors. In some sweet peas the axils whence the leaves and flower stalks spring from the main stem are of a deep red colour. In others they are green. The dark pigmented axil is dominant to the light one. Again, in some sweet peas the anthers are sterile, setting no pollen, and this condition is recessive to the ordinary fertile condition. When a sterile plant with a dark axil is crossed by a fertile plant with a light axil, the F₁ plants are all fertile with dark axils. But such plants in F₂ give fertiles with light axils, fertiles with dark axils, and steriles with dark axils in the ratio 1 : 2 : 1. No light axilled steriles appear from such a cross owing to the repulsion between the factor for dark axil (D) and that for the fertile anther (F).

These four cases have already been found in the sweet pea, and similar phenomena have been met with by Gregory in primulas. To certain seemingly analogous cases in animals where sex is concerned we shall refer later.

Now all of these four cases present a common feature which probably has not escaped the attention of the reader. In all of them the original cross was such as to introduce one of the repelling factors with each of the two parents. If we denote our two factors by A and B, the crosses have always been of the nature AAbb × aaBB. Let us now consider what happens when both of the

factors, which in these cases repel one another, are introduced by one of the parents, and neither by the other parent. And in particular we will take the case in which we are concerned with purple and red flower colour, and with long and round pollen, i.e. with the factors B and L. When a purple long (BBLL) is crossed with a red round (bbll) the F₁ (BbLl) is a purple with long pollen, identical in appearance with that produced by crossing the long pollened red with the round pollened purple. But the nature of the F₂ generation is in some respects very different. The ratio of purples to reds and of longs to rounds is in each case 3 : 1, as before. But instead of an association between the red and the long pollen characters the reverse is the case. The long pollen character is now associated with purple and the round pollen with red. The association, however, is not quite complete, and the examination of a large quantity of similarly bred material shows that the purple longs are about twelve times as numerous as the purple rounds, while the red rounds are rather more than three times as many as the red longs. Now this peculiar result could be brought about if the gametic series produced by the F₁ plant consisted of 7 BL + 1 Bl + 1 bL + 7 bl out of every 16 gametes. Fertilization between two such similar series of 16 gametes would result in 256 plants, of which 177 would be purple longs, 15 purple rounds, 15 red longs, and 49 red rounds—a proportion of the four different kinds very close to

that actually found by experiment. It will be noticed that in the whole family the purples are to the reds as 3 : 1, and the longs are also three times as numerous as the rounds. The peculiarity of the case lies in the distribution of these two characters with regard to one another. In some way or other the factors for blue and for long pollen become linked together in the cell divisions that give rise to the gametes, but the linking is not complete. This holds good for all the four cases in which repulsion between the factors occurs when one of the two factors is introduced by each of the parents. When both of the factors are brought into the cross by the same parent we get coupling between them instead of repulsion. The phenomena of repulsion and coupling between separate factors are intimately related, though hitherto we have not been able to suggest why this should be so.

Nor for the present can we suggest why certain factors should be linked together in the peculiar way that we have reason to suppose that they are during the process of the formation of the gametes. Nevertheless the phenomena are very definite, and it is not unlikely that a further study of them may throw important light on the architecture of the living cell.

APPENDIX TO CHAPTER IX

As it is possible that some readers may care, in spite of its complexity, to enter rather more fully into the peculiar phenomenon

of the coupling of characters, I have brought together some further data in this Appendix. In the case we have already considered, where the factors for blue colour and long pollen are concerned, we have been led to suppose that the gametes produced by the heterozygous plant are of the nature 7 BL : 1 Bl : 1 bL : 7 bl. Such a series of ovules fertilised by a similar series of pollen grains will give a generation of the following composition:—

and as this theoretical result fits closely with the actual figures obtained by experiment we have reason for supposing that the heterozygous plant produces a series of gametes in which the factors are coupled in this way. The intensity of the coupling, however, varies in different cases. Where we are dealing with another, viz. fertility (F) and the dark axil (D), the experimental numbers accord with the view that the gametic series is here 15 FD : 1 Fd : 1 fD : 15 fd. The coupling is in this instance more intense. In the case of the erect standard (E) and blueness (B) the coupling is even more intense, and the experimental evidence available at present points to the gametic series here being 63 Eb : 1 EB : 1 eB : 63 eb. There is evidence also for supposing that the intensity of the coupling may vary in different families for the same pair of factors. The coupling between blue and long pollen is generally on the 7 : 1 : 1 : 7

basis, but in some cases it may be on the 15 : 1 : 1 : 15 basis. But though the intensity of the coupling may vary it varies in an orderly way. If A and B are the two factors concerned, the results obtained in F₂ are explicable on the assumption that the ratio of the four sorts of gametes produced is a term of the series—

In such a series the number of gametes containing A is equal to the number lacking A, and the same is true for B. Consequently the number of zygotes formed containing A is three times as great as the number of zygotes which do not contain A; and similarly for B. The proportion of dominants to recessives in each case is 3 : 1. It is only in the distribution of the characters with relation to one another that these cases differ from a simple Mendelian case.

As the study of these series presents another feature of some interest, we may consider it in a little more detail. In the accompanying table are set out the results produced by these different series of gametes. The series marked by an asterisk have already been demonstrated experimentally. The first term in the series,

in which all the four kinds of gametes are produced in equal numbers is, of course, that of a simple Mendelian case where no coupling occurs.

Now, as the table shows, it is possible to express the gametic series by a general formula (n + 1) AB + Ab + aB + (n - 1) ab, where 2n is the total number of the gametes in the series. A plant producing such a series of gametes gives rise to a family of zygotes in which 3n² - (2n - 1) show both of the dominant characters and n² - (2n - 1) show both of the recessive characters, while the number of the two classes which each show one of the two dominants is (2n - 1). When in such a series the coupling becomes closer the value of n increases, but in comparison with n² its value becomes less and less. The larger n becomes the more negligible is its value relatively to n². If, therefore, the coupling were very close, the series 3n² - (2n - 1) : (2n - 1) : (2n - 1) : n² - (2n - 1) would approximate more and more to the series 3n² : n², i.e. to a simple 3 : 1 ratio. Though the point is probably of more theoretical than practical interest, it is not impossible that some of the cases which have hitherto been regarded as following a simple 3 : 1 ratio will turn out on further analysis to belong to this more complicated scheme.

CHAPTER X

SEX