LIMITS OF NATURAL SELECTION. 21 



because the first two nested on the ground, and the 

 last in a pine tree. It seems to be very seldom the 

 case that birds lose their young by mistakes in 

 location or manner of building, such as could have 

 been guarded against by the foresight of such intelli- 

 gence as a bird may be supposed to possess. Yet 

 it is exactly such odd and unintelligible mistakes 

 which are demanded by the theory of undirected 

 variation and selection, in order to produce the nest- 

 ing instincts. On the other hand, in the change of 

 the nesting habits of the chimney-martins and barn- 

 swallows, from hollow trees and rocks, to chim- 

 neys and barns, we find two cases of change of 

 instinct which there is no ground for assuming to 

 be the result of the general destruction of a hypo- 

 thetical number of birds which refused to make the 

 change. That the kea-bird of New Zealand has 

 learned to dig the kidney fat out of living sheep 

 since the introduction of sheep into that country, 

 is another wonderful instance of change of instinct 

 hardly to be accounted for by means of natural 

 selection, but rather as the result of intelligent 

 experience. 



If we could conceive of a species as having one 

 long uninterrupted mental existence, in which gen- 

 eration after generation of individuals replace each 

 other in some way without interrupting the mental 



