is 1 long to about 3·5 rounds. There is partial coupling of long pollen with the purple colour and a complementary coupling of the red colour with round pollen. This result would be brought about if it were supposed that seven out of every eight purple gametes produced by the F1 plant carried the long pollen character, and that seven out of every eight red gametes carried the round pollen character. The facts observed fit in with the supposition that the gametes are produced in series of sixteen, but how such result could be brought about is a question which for the present must remain open.
Spurious Allelomorphism.—Instances of association between characters are known in which the connection is between the dominant member of one pair and the recessive of another. In many sweet peas the standard folds over towards the wings, and the flower is said to be hooded. This “hooding” behaves as a recessive towards the erect standard. Red sap colour is also recessive to purple. In families where purples and reds as well as erect and hooded standards occur it has been found, as might be expected, that erect standards are to hooded ones, and that purples are to reds as 3:1. Were the case one of simple dihybridism the F2 generation should be composed of 9 erect purples, 3 hooded purples, 3 erect reds and 1 hooded red in every 16. Actually it is composed of 8 erect purples, 4 hooded purples and 4 erect reds. The hood will not associate with the red, but occurs only on the purples. Cases like this are best interpreted on the assumption that during gametogenesis there is some form of repulsion between the members of the different pairs—in the present instance between the factor for purple and that for the erect standard—so that all the gametes which contain the purple factor are free from the factor for the erect standard. To the process involved in this assumption the term spurious allelomorphism has been applied.
Sex.—On the existing evidence it is probable that the inheritance of sex runs upon the same determinate lines as that of other characters. Indeed, there occurs in the sweet pea what may be regarded as an instance of sex inheritance of the simplest kind. Most sweet peas are hermaphrodite, but some are found in which the anthers are sterile and the plants function only as females. This latter condition is recessive to the hermaphrodite one and segregates from it in the ordinary way. Most cases of sex inheritance, however, are complicated, and it is further possible that the phenomena may be of a different order in plants and animals. Instructive in this connexion are certain cases in which one of the characters of an allelomorphic pair may be coupled with a particular sex. The pale lacticolor variety of the currant moth (Abraxas grossulariata) is recessive to the normal form, and in families produced by heterozygous parents one quarter of the offspring are of the variety. Though the sexes occur in approximately equal numbers, all the lacticolor in such families are females; and the association of sex with character exhibiting normal segregation is strongly suggestive of a similar process obtaining for sex also. Castle has worked out similar cases in other Lepidoptera and has put forward an hypothesis of sex inheritance on the basis of the Mendelian segregation of sex determinants. An ovum or spermatozoon can carry either the male or the female character, but it is essential to Castle's hypothesis that a male spermatozoon should fertilize only a female ovum and vice versa, and consequently on his view all zygotes are heterozygous in respect of sex. Whether any such gametic selection as that postulated by Castle occurs here or elsewhere must for the present remain unanswered. Little evidence exists for it at present, but the possibility of its occurrence should not be ignored.
More recently evidence has been brought forward by Bateson and others (3) which supports the view that the inheritance of sex is on Mendelian lines. The analysis of cases where there is a closer association between a Mendelian character and a particular sex has suggested that femaleness is here dominant to maleness, and that the latter sex is homozygous while the former is heterozygous.
Chromosomes and Unit-Characters.—Breeding experiments have established the conception of definite unit-characters existing in the cells of an organism: in the cell histology has demonstrated the existence of a small definite bodies—the chromosomes. During gametogenesis there takes place what many histologists regard as a differentiating division of the chromosomes: at the same period occurs the segregation of the unit-characters. Is there a relation between the postulated unit-character and the visible chromosome, and if so what is this relation? The researches of E. B. Wilson and others have shown that in certain Hemiptera the character of sex is definitely associated with a particular chromosome. The males of Protenor possess thirteen chromosomes, and the qualitative division on gametogenesis results in the production of equal numbers of spermatozoa having six and seven chromosomes. The somatic number of chromosomes in the female is fourteen, and consequently all the mature ova have seven chromosomes. When a spermatozoon with seven chromosomes meets an ovum the resulting zygote has fourteen chromosomes and is a female; when a spermatozoon with six chromosomes meets an ovum the resulting zygote has thirteen chromosomes and is a male. In no other instance has any such definite relation been established, and in many cases at any rate it is certain that it could not be a simple one. The gametic number of chromosomes in wheat is eight, whereas the work of R. H. Biffen and others has shown that the number of unit-characters in this species is considerably greater. If therefore there exists a definite relation between the two it must be supposed that a chromosome can carry more than a single unit-character. It is not impossible that future work on gametic coupling may throw light upon the matter.
Heredity and Variation.—It has long been realized that the problems of heredity and variation are closely interwoven, and that whatever throws light upon the one may be expected to illuminate the other. Recent as has been the rise of the study of genetics, it has, nevertheless, profoundly influenced our views as to the nature of these phenomena. Heredity we now perceive to be a method of analysis, and the facts of heredity constitute a series of reactions which enable us to argue towards the constitution of living matter. And essential to any method of analysis is the recognition of the individuality of the individual. Constitutional differences of a radical nature may be concealed beneath apparent identity of external form. Purple sweet peas from the same pod, indistinguishable in appearance and of identical ancestry, may yet be fundamentally different in their constitution. From one may come purples, reds and whites, from another only purples and reds, from another purples and whites alone, whilst a fourth will breed true to purple. Any method of investigation which fails to take account of the radical differences in constitution which may underlie external similarity must necessarily be doomed to failure. Conversely, we realize to-day that individuals identical in constitution may yet have an entirely different ancestral history. From the cross between two fowls with rose and pea combs, each of irreproachable pedigree for generations, come single combs in the second generation, and these singles are precisely similar in their behaviour to singles bred from strains of unblemished ancestry. In the ancestry of the one is to be found no single over a long series of years, in the ancestry of the other nothing but singles occurred. The creature of given constitution may often be built up in many ways, but once formed it will behave like others of the same constitution. The one sure test of the constitution of a living thing lies in the nature of the gametes which it carries, and it is the analysis of these gametes which forms the province of heredity.
The clear cut and definite mode of transmission of characters first revealed by Mendel leads inevitably to the conception of a definite and clear-cut basis for those characters. Upon this structural basis, the unit-character, are grounded certain of the phenomena now termed variation. Varieties exist as such in virtue of differing in one or more unit-characters from
