98 How Theories are Manufactured 



and the Doves, and still more markedly of the Herons 

 and Cranes. 



In face of all these examples, it seems hard to con- 

 ceive that there is not an internal directive force guiding 

 development, if development there be, along predeter- 

 mined lines, and not leaving it to find its way, fortui- 

 tously, like a butterfly in a hailstorm, between the blows 

 of destructive forces. 



It is likewise perhaps worthy of remark that although 

 brilliancy of colour is as a rule most conspicuous in the 

 breeding plumage, it sometimes manifests itself at a 

 season when the selective preference of a mate cannot 

 account for its genesis. The Redpoll, for instance, in 

 spring, has but a faint tinge of crimson on its forehead, 

 which develops in richer tints as the season advances 

 to the time of the great moult which follows ; being thus 

 at its best when the breeding season is done. A pheno- 

 menon of similar import is presented by the autumn 

 song of the Robin, which cannot be accounted for, like 

 spring melodies, by the advantage which it gives the 

 singer in securing a partner. 



But this threatens to lead me to the question of 

 manners and customs, another branch of my subject, 

 of even wider and deeper interest, which I had intended 

 to treat in this paper, but which, spatiis exclusus iniquis, 

 I must leave for another. 



Before passing from the question of colour, however, 

 some mention should be made of the solution recently 

 proposed by Mr. Wallace. In his opinion the Darwinian 

 explanation of the origin of decorative coloration, 

 through the selective agency of mates, is not the true 

 one ; 1 and in its place he suggests another, which, truth 

 to tell, is still less comprehensible ; here, in fact, as 

 elsewhere, his argument is not merely hard to follow, 

 but absolutely eludes perception. He calls attention' 2 

 to the fact that diversified coloration follows the chief 

 lines of structure, and changes at points, such as the 

 joints, where function changes. Thus among mammals 

 1 Danoinisnt, pp. 274-288. 2 P. 288. 



