Inheritance of Acquired Characters 249 



higher branches could not be the cause of the lengthening of 

 the legs. But "the coadaptation of the parts, required to make 

 the giraffe's structure useful, is much greater than at first 

 appears." For example, the bones and the muscles of the 

 hind-legs have been also altered, and Spencer argues that it 

 is "impossible to believe" that all parts of the hind-quarters 

 could have been coadapted to one another, and to all parts 

 of the fore-quarters. A lack of coadaptation of a single muscle 

 " would cause fatal results when high speed had to be main- 

 tained while escaping from an enemy." 



Spencer claims that, since 1886, when he first published 

 this argument, nothing like an adequate response has been 

 made ; and I think he might have added that an adequate 

 answer is not likely to be forthcoming, since nothing short 

 of a demonstration of how the giraffe really evolved is 

 likely to be considered as sufficient. Wallace's reply, that 

 the changes in question could have been brought about 

 by natural selection, since similar changes have been brought 

 about by artificial selection, is regarded as inadequate by 

 Spencer, since it assumes a parallel which does not exist. 

 Nevertheless, Wallace's reply contains, in my opinion, the 

 kernel of the explanation, in so far as it assumes that con- 

 genital variation x may suffice to account for the origin of 

 a form even as bizarre as that of the giraffe. The ancon 

 ram and the turnspit dog were marked departures from the 

 normal types, and yet their parts were sufficiently coordi- 

 nated for them to carry out the usual modes of progression. 

 It would not have been difficult, if we adopted Spencer's 

 mode of arguing, to show that these new forms could not 

 possibly have arisen as the result of congenital variations. 



Again, it might be argued that the large, powerful dray- 

 horse could not have arisen through a series of variations 

 from the ordinary horse, because, even if variations in the 



1 Wallace assumes fluctuating variation to suffice, but in this I cannot agree 

 with him. 



