ROOKS AND BOWER-BIRDS 65 



especially the so-called garden of amblyornis, their 

 gradual elaboration from a much simpler structure 

 presents no more difficulty than does that of a com- 

 plicated nest from a quite ordinary one. All 

 that we want is the initial directing impulse, and 

 this we have when once a bird uses its nest, not only 

 as a cradle for its young, but, also, as a nuptial bed 

 or sporting-place. In a passage of this nature, the 

 nest, indeed, must remain, but why should it not ? 

 Let us suppose that, like the rooks, the bower-birds 

 or, rather, their ancestors used, at one time, to 

 use their old nests of the spring, as play-houses 

 during the winter. If, then, they had built fresh 

 nests as spring again came round, might they not 

 gradually have begun to build fresh play-houses too ? 

 The keeping up of the old nest but for a secondary 

 purpose would naturally have passed into this, and 

 the playing about it would, as naturally, have led 

 to the keeping of it up. That duality of use should 

 gradually have led to duality of structure that from 

 one thing used in two different ways there should 

 have come to be two things, each used in one of 

 these ways does not seem to me extraordinary, but, 

 rather, what we might have expected, in accordance 

 with the principle of differentiation and specialisa- 

 tion, which has played so great a part in organic 

 evolution. It is by virtue of this principle that 

 limbs have been developed out of the vertebral 

 column, and the kind of advantage which all verte- 

 brate animals have gained by this multiplication and 

 differentiation of parts, in their own bodily structure, 

 is precisely that which a bird of certain habits would 

 have gained, by a similar increase in the number and 



