in holes, usually in banks, but sometimes in trees. The motmots and the puff-birds (Bucconidae) build in similar places; while the toucans, barbets, trogons, woodpeckers, and parrots all make their nests in hollow trees. This habit, pervading all the members of extensive families, must therefore be extremely ancient, more especially as it evidently depends in some degree on the structure of the birds, the bills, and especially the feet, of all these groups being unfitted for the construction of woven arboreal nests.[1] But in all these families the colour varies greatly from species to species, being constant only in the one character of the similarity of the sexes, or, at all events, in their being equally conspicuous even though differently coloured.
When I first put forward this view of the connection between the mode of nesting and the coloration of female birds, I expressed the law in somewhat different terms, which gave rise to some misunderstanding, and led to numerous criticisms and objections. Several cases were brought forward in which the females were far less brilliant than the males, although the nest was covered. This is the case with the Maluridae, or superb warblers of Australia, in which the males are very brilliant during the pairing season and the females quite plain, yet they build domed nests. Here, there can be little doubt, the covered nest is a protection from rain or from some special enemies to the eggs; while the birds themselves are protectively coloured in both sexes, except for a short time during the breeding season when the male acquires brilliant colours; and this is probably connected with the fact of their inhabiting the open plains and thin scrub of Australia, where protective colours are as generally advantageous as they are in our north-temperate zones.
As I have now stated the law, I do not think there are any exceptions to it, while there are an overwhelming number of cases which give it a strong support. It has been objected that the domed nests of many birds are as conspicuous as the birds themselves would be, and would, therefore, be of no use as a protection to the birds and young. But, as a matter of fact, they do protect from attack, for hawks or crows do not pluck such nests to pieces, as in doing so they would be
- ↑ On this point see the author's Contributions to Natural Selection, chap. v. i.
