222 CERTAIN LAWS OP VARIATION. 



the species of Primates examined, the muscles were 

 quite as variable as in man. 



Though Schwalbe's views may be correct as regards 

 some domesticated races, they obviously cannot apply 

 to all. For instance, domesticated pigeons, dogs, and 

 horses vary between very much wider limits than any 

 naturally occurring species. But it is obvious that 

 such extreme variations as these owe their origin chiefly 

 to careful artificial selection, extending over very large 

 numbers of generations. When no such selection is 

 practised, as in some of the domesticated animals kept 

 by savages, then it is highly probable that the varia- 

 bility is no greater than in allied feral species. Any 

 slight increase of variability actually present may 

 doubtless be attributed to the direct action of the 

 changed conditions of life. 



Darwin's conclusion that increased variability only 

 shows itself after several generations of domestication, 

 if true as it stands, is difficult to account for in any 

 other way than by supposing that the reproductive sys- 

 tem is affected. But it seems to me distinctly probable 

 that much, if not all, of the increased variability may 

 be ascribed merely to the careful selection of the most 

 aberrant individuals in each generation, and their 

 assortive mating. In such a case the variability of the 

 whole group would obviously increase from generation 

 to generation. 



