448 The American Naturalist. [June, 



th rdopntent at each stage of a species', development must be in the 

 directions thus ratified by intelligence. So also with imitation. 

 Only those imitative actions of a creature which are useful to 

 him will survive in the species , for in so far as he imitates 

 actions which are injurious he will aid natural selection in 

 killing himself off. So intelligence, and the imitation which 

 copies it, will set the direction of the development of the com- 

 plex instincts even on the Neo-Darwinian theory ; and in this 

 sense we may say that consciousness is a ' factor ' " (ref. 4). 



2. The mean of phijlogenetic variation bring thus made more de- 

 terminate, further phylogaafiv variations follow about this mean, 

 and these variations an again utilized by Organic Selection for on- 

 togenetic adaptation. So there is continual phylogenetic prog- 

 ress in the directions set by ontogenetic adaptation (ref. 3, 4, 

 5). " The intelligence supplements slight co-adaptations and 

 so gives them selective value ; but it does not keep them from 

 getting farther selective value as instincts, reflexes, etc., by far- 

 ther variation " (ref. 5). " The imitative function, by using 

 muscular co-ordinations, supplements them, secures adapta- 

 tions, keeps the creature alive, prevents the 'incidence of 

 natural selection,' and so gives the species all the time necessary 

 to get the variations required for the full instinctive perform- 

 ance of the function " (ref. 4). But, " Conscious imitation, while 

 it prevents the incidence of natural selection, as has been seen, 

 and so keeps alive the creatures which have no instincts for the 

 performance of the actions required, nevertheless does not sub- 

 serve the utilities which the special instincts do, nor prevent 

 them from having the selective value of which Romanes 

 speaks. Accordingly, on the more general definition of intel- 

 ligence, which includes in it all conscious imitation, use of 

 maternal instruction, and that sort of thing — no less than on 

 the more special definition — we still find the principal of nat- 

 ural selection operative " (ref. 5). 



3. Tltis completely disposes of the Lamarkian factor as far as 

 two lines of evidence for it are concerned. First, the evidence 

 drawn from function, " use and disuse," is discredited ; since 

 by " organic selection," the reappearance, in subsequent 

 generations, of the variations first secured in ontogenesis is ac- 



