Throughout the greater part of the year the Rouen drake has the nuptial plumage. The head is green and the breast is claret. Two median tail feathers are strongly curved; the next two are also often curved. These four are called the sex feathers. At the close of the breeding-season (July) both sexes molt. The male now has the same coat as the female, or nearly so. The green head becomes brown to buff; the sex feathers are straight. The change back again to the nuptial plumage begins at the end of summer and is completed early in October. Thus in the race of Rouens the eclipse plumage lasts only a very short time. In the mallard it lasts longer. The eclipse plumage develops, therefore, only when the testes are active, or, as Goodale puts it, “the presence of the active testis is necessary for the drake to assume this plumage.” Conversely, the nuptial plumage comes on in the late summer, when mating is over, and when the testes have shrunken and are not active, at least as far as the sex-cells are concerned. In some respects the situation is like that in the fowls, for in both the testes are not necessary for the development of the full plumage, but in other respects the situation is different, because at the time in the ducks when the testes are active the eclipse plumage develops. Are we to suppose that at the time of sexual activity a substance is produced analogous to that produced by the ovary of the female? This seems the most plausible assumption, for we know that if the testis is removed the eclipse plumage does not appear. Such a situation suggests a comparison with the Sebright, where it has been shown that the testis must actively produce some substance which, like that in the ovary, keeps down cock-feathering. It is plausible, even if it can not be established, that the substance in the duck and the inhibitory substance in the male Sebright are the same as that produced in the female.