4 THE ANCESTORS [CH. 



breed, was the descendant of a distinct and larger species : it 

 is far more probable that the domestic races of our various 

 animals were gradually improved in different parts of the 

 great European- Asiatic continent, arid thence spread to other 

 countries." He thus left Africa out of account as a possible 

 source for a race of horses. It will be observed that those who 

 hold a single origin for all domestic horses base their belief on 

 " the fertility of the most distinct breeds when crossed." Yet 

 this cannot be regarded as a true criterion, for animals which 

 are admittedly distinct species, such as the dog, the wolf, and 

 the jackal among carnivores, and the common ox, the zebu (Bos 

 gaurus), and the yak (Bos grunniens) among herbivores, freely 

 interbreed and produce fertile offspring. But though Darwin 

 leaned to the belief that all our horses come from a single stock, 

 he carefully pointed out that, " as several species and varieties 

 of the horse existed during the later Tertiary period, and as 

 Rutimeyer found differences in the size and form of the skull in 

 the earliest known domesticated horses, we ought not to feel 

 sure that all our breeds are descended from a single species 1 ." 

 He elsewhere 2 points out that "as the savages of North and 

 South America easily reclaimed the feral horses, so there is 

 no improbability in savages in various quarters of the world 

 having domesticated more than one native species or natural 

 race." 



Since Darwin wrote it has been generally held that all our 

 domestic horses have had but a single source, whether they be 

 the fine horses of slender build and great speed, of which the 

 Arab is the type, or the heavy cart-horses, whose origin is 

 commonly found in the coarse, thickset horses of Europe and 

 upper Asia, of which the unimproved Mongolian pony is the repre- 

 sentative, or hunters, roadsters, carriage-horses and trappers, 

 which are as everyone knows, the result of a judicious blending 

 of the two first-mentioned classes. Thus M. Sanson 3 now holds 

 that all our domestic breeds had a single origin, and divides 

 recent horses into two groups long-headed and short-headed 



g 



1 Variation of Animals and Plants under Domestication, Vol. i. p. 53. 



2 Ibid. Vol. i. p. 54. 



3 Traite de Zootechnie (ed. 4, 1901), pp. 2, 3.' 



