Syndactylism illustrates another step in the series of decreasing potency of the dominant. On not one of the F1 generation was the dominant (syndactyl) condition observed; and when these hybrids were mated together the dominant character appeared not in 75 per cent but in from 10 per cent to 0 per cent of the offspring. The question may well be asked: What is then the criterion of dominance? The reply is elaborated to the effect that, since dominance is due to the presence of a character and recessiveness to its absence, dominance may fail to develop, but recessiveness never can do so. Consequently two extracted recessives mated _inter se_ can not throw the dominant condition; but two imperfect dominants, even though indistinguishable from recessives, will throw dominants. On the other hand, owing to the very fact that the dominant condition often fails of development, two extracted "pure" dominants will, probably always, throw some apparent "recessives." Now, two syndactyls have not been found that fail (in large families) to throw normals, but extracted normals have been found which, bred _inter se_, throw only normals; hence, "normal-toe" is recessive. In this character, then, dominance almost always fails to show itself in the heterozygote and often fails in pure dominants. The series of diminishing potency has now brought us to a point where we can interpret a case of great difficulty, namely, a case of rumplessness. Here a dominant condition was originally mistaken for a recessive condition, because it never fully showed itself in F1 and F2. Nevertheless, in related individuals, the condition is fully dominant. We thus get the notion that a factor that normally tends to the development of a character may, although present, fail to develop the character. Dominance is lacking through _impotence_. The last term of the series is seen in the wingless cock which left no wingless offspring in the F1 and F2 generations. In comparison with the results gained with the rumpless cock, winglessness in this strain is probably dominant but impotent. When a character, instead of being simply present or absent, is capable of infinite gradations, inheritance seems often to be blending and without segregation. Two cases of this sort--booting and nostril-height--are examined, and by the aid of the principle of imperfect dominance the apparent blending is shown to follow the principle of segregation. Booting is controlled by a dominant inhibiting factor that varies greatly in potency, and nostril-height is controlled by an inhibiting factor that stops the over-growth of the nasal flap which produces the narrow nostril.