American Breeders Association, 6: 29-32 (1911)
C. B. Davenport, Cold Spring Harbor, N. Y.

The opposition to mendelian principles of heredity is gradually vanishing in the face of facts. The principle of segregation of characters, according to which dissimilar parental characters do not blend in the progeny but come out intact as in the grandparents, is recognized as widespread if not universal. But concerning dominance there still remains much skepticism without and much confusion within the fold.

What is dominance? Bateson, in his first statements of Mendel's doctrine (in 1902), says (page 9): "In the case of each pair of characters there is thus one which in the first cross prevails to the exclusion of the other. This prevailing character Mendel calls the dominant character." We have progressed far in the last seven years and now we think of dominance as occurring when one parent has a characteristic that the other lacks. Under these circumstances the offspring possess the character—that is the whole story and it seems simple enough.

Two complications must, however, be considered. It appears that in certain characters which show a great variety of grades, as pigment does in human hair, the more advanced grade, e. g. the heavier pigmentation, dominates over the lesser grade. This would seem to indicate either that a low grade of a character may act toward a high grade as absence towards presence, or else that hair, despite its apparent continuity, really consists of a multitude of discontinuous units; and that a lower grade means absence of a unit present in a higher grade.

The second complication is the fact of imperfect dominance. Even Mendel recognized that dominance is not always perfect in sweetpeas, for the hybrids between white-flowered and purple-red-flowered peas have flowers less intensely colored than the darker parent. Correns, in 1900, showed that in a certain set of crosses between good species the hybrids had the characters as in both parents, only reduced in varying degrees. Bateson and Saunders, in 1902, found that the "intensity of the dominant character (comb and extra toe) is often considerably reduced." Correns, in 1905, stated that there was known a complete series of cases at one extreme of which one "allelomorph" completely hindered the appearance of the other, while at the opposite end of the series the hybrid character showed an intermediate condition, both "allelomorphs" appearing with equal strength. In my study of poultry such a series is striking. When the median comb is mated with no-median the median comb appears in the offspring but reduced in length between 30 and 100 per cent. Extra toe mated with normal gives extra toe in only 73 per cent. of the offspring. Syndactylism is dominant over its absence, but no syndactyl offspring were observed in the first hybrid generation; nevertheless two syndactyl parents yield about 56 per cent syndactyl offspring. Rumplessness is dominant. But a tailless cock mated to tailed hens produced no rumpless offspring; neither when mated to his daughters did any tailless appear, but the reduced "pope's nose" and shortened back showed that there was an imperfect modification. However, another strain has given results more in accord with expectation. Finally, winglessness appears not to be inherited at all, but the hypothesis is tenable that winglessness is an imperfectly dominant character. Thus the series of potency in dominance is complete.

The failure fully to recognize and accept the consequences of imperfection of dominance has led to misunderstandings and to unnecessary subsidiary hypotheses. The Mendelians, knowing the fact of imperfect dominance, have placed too much stress on dominance and recessiveness, and so laid themselves open to just criticism by their opponents. Dominance is a corollary of segregation and its discovery marks a great step in advance, but its variation in degree must be constantly recognized and insisted upon. The consequences are most important. First, because of imperfect dominance, and the accompanying diluted nature of the determiner in the heterozygotes, the character may appear in reduced amount and permit a hypostatic quality to appear. Thus poultry with deep black plumage mated with albinos showi in part the Jungle fowl plumage in the offspring that is covered over by the black pigment in the dark parent. Thus a new character is revealed by the heterozygote;

Secondly, the weakened character may be retarded in development so that it fails to appear at the normal period but develops later. Thus Lang found hybrids between red and non-red snails to be at first non-red but finally red. Some authors have spoken of this as reversal of dominance, and even Bateson uses this terminology in his latest book. But this is obviously an unfortunate term if dominance means the presence of a'quality. For, a given quality, that is due to the absence of a factor, like blue iris color, cannot be at one time recessive and at another dominant. If a blue iris appears where brown is expected, the clear reason is that brown pigment has merely failed to develop and is potentially present. A similar case occurs in hybrids between albino and some buff birds; the chicks have a pure white down, only later acquiring the black and buff of the adult plumage. Albinism is here not momentarily dominant, but merely pigmentation, owing to the weakened stimulus to its production in the heterozygote, is delayed in development.

Not in heterozygotes only but even in homozygotes the character may fail to develop. Extra toe is dominant in certain fowls. Even in pure-bred fowls that are certainly homozygous dominants, two or three per cent of the offspring may fail to develop the extra toe. Evidently, for some reason, the proper internal stimulus is lacking to complete the development of the toe in these few individuals.

In many cases, Shull has stated, imperfect dominance may make it impossible to tell which character is dominant, so that an apparent recessive "presence" of a character may be an imperfectly dominant "presence." But I cannot follow him entirely. Between imperfect dominance and recessiveness there is clear distinction—two perfect and imperfect dominants alike may throw some recessives; but two homozygous recessive parents can throw no dominants; where neither parent possesses a character in its germ plasm their progeny will not have the character. Some families of two recessive parents will, then, throw only offspring of one kind; such are homozygous parents and their offspring are pure recessives.

Many cases of apparent blending are due to imperfect dominance. Thus I recognize 10 grades of booting or foot-feathering of poultry and one of no boot. If a booted be mated to a non-booted the grades in the offspring run from 0 to 8; it is impossible to say whether booting or clear shanks is dominant. In F-2, booted and clean shanks appear again, with boot of some grade in a large majority. All of the extracted clear-shanked mated together produce some booted birds and some of the booted, mated inter se, may produce clear-shanked offspring; such parents are clearly heterozygotes with arrested dominance. But, on the other hand, some of the booted, bred inter se, produce only booted offspring; such are clearly extracted recessives. I have a set of such matings yielding 287 offspring, 100 per cent booted. My best extracted homozygous dominant clean-shanks produce only 90 per cent of offspring with clean shanks. Presence X presence of a character may give absence through failure of the character to develop, but absence X absence of a character can give no trace of the character.

Finally, dominance may be so imperfect that the dominant character fails completely to develop in the heterozygote and develops only imperfectly or sporadically in the homozygote. This is apparently the explanation of many cases where a character seems not to be inherited. The character has perhaps arisen once as a sport but cannot be propagated. It is probable in such cases that the internal stimulus to development of the organ is too weak. This case is illustrated by a rumpless cock which, crossed with tailed birds, produced only tailed offspring. These offspring mated among themselves and with their father produced no tailless progeny. Nevertheless a disturbance in the germ plasm was indicated by the small "pope's nose" that carries the tail feathers and by shortened, bent backs. These tailed heterozygotes mated to a tailless son of the first cock produced a large proportion of tailless birds. Indeed, taillessness proves to be dominant, but its determiner in the germ plasm is very feeble.

This study, then, leads to the conclusion that alongside of dominance we must place an important modifying factor—the factor of the strength or potency of the representative of the given character in the germ plasm. This is clearly a very variable quality. If it is very potent we get a typically mendelian result; but if it is weak, we will have imperfect dominance or failure to develop altogether.

(Presented by the Committee on Theoretical Research.]