228 THE HORSE AND ITS RELATIVES 



cannot establish a mixed race, and are therefore 

 distinct species judged by that test as well as by 

 their form and marking. It is not known whether 

 the so-called species of wild ass — the Asiatic and 

 the African — would prove to produce fertile or in- 

 fertile mules if inter-crossed, nor has the test been 

 applied to the very differently marked local races 

 of the African zebras — Gravy's zebra, Burchell's 

 zebra, and the mountain zebra. It is likely enough 

 that the three or more species distinguished among 

 zebras on account of their being differently striped, 

 and existing in different localities, would be found 

 to breed freely together, and prove themselves thus 

 to be entitled to be regarded as local ' varieties ' 

 or ' races,' but not as fully-separated, true species." 



In this passage the writer definitely commits 

 himself to the opinion that the fertility or sterility 

 of the hybrids produced by crossing two distinct 

 members of the horse family (or, for that matter, of 

 any other family) affords a definite and decisive 

 test whether such members should be regarded as 

 races or breeds of one and the same species, or as 

 distinct species. Such information as we possess 

 on the subject,^ comparatively meagre as it is, does 

 not, however, justify a sweeping generalisation like 

 the above, for there is really no hard-and-fast line 



1 Much interesting information on horse and zebra hybrids will 

 be found in Professor J. C. Ewart's The Penicuik Experiments, 

 London, 1899. 



