As an example of the kind of theorising to which this method leads, we may cite the often-quoted theory which ascribes the green colouring of some arboreal fruit-eating pigeons to adaptation to an existence among tropical foliage, and ignores the fact that in America tree-haunting pigeons are never of this colour, and that it is not by any means universal even among the old-world pigeons. White Down of Nestlings Similarly, a theory has been advanced (W. P. Pycraft, _Knowledge_, 1904, p. 275) that the white down of some nestling birds, is an adaptation to resisting the heat of the sun in open nests. This is at once negatived by the fact that young owls, usually hatched in shaded places, are also generally white, while young cormorants, living in open nests, are black; yet the allied darters, with the same breeding haunts in some cases, have white young. Lest it should be thought that black has some especial value in a nestling living exposed, we may mention that young petrels, which are born in holes, have black or dark down. As we have already pointed out, naturalists in too readily accepting the theory that variation is minute in degree and indefinite in direction, have raised quite unnecessary difficulties, even for the selection hypothesis. We have cited certain facts, which seem to show that variations, as a rule, are not indefinite in direction; of these the most striking is furnished by birds in which the tail feathers are greatly elongated. Were variations indeterminate, we might reasonably expect to find that the elongation occurred in one particular feather or pair of feathers in one species, in another pair in a second species, in a third pair in a third species, and so on. But this is not the case; no bird has one _single_ long feather in its tail, and when two are elongated, as is so commonly the case, these are almost invariably the middle or the outside pair; _e.g._, in the European bee-eater and pheasant it is the former, in the swallow and blackcock, the latter. Exceptions are so rare that they may almost be said to prove the rule; _e.g._, although most terns have the outer-tail feathers elongated, in some of the Noddy Terns (_Anous_, _Gygis_) the third pair, in others the fourth pair, of tail feathers are the longest. This must mean one of two things, either that the variation, as regards length in tail feathers, other than middle or outer, does not ordinarily occur, or that it occurs, but is, in some way, inimical to the welfare of the species. The latter hypothesis does not seem probable, as the Noddies are particularly abundant birds where they occur, that is to say, in the tropical seas; therefore, we can only conclude that that particular variation has not occurred in birds as a whole.