262 THE HORSE AND ITS RELATIVES 



horse-line that constitutes the main difficulty ; for 

 if the gaps were wider than they are, the division 

 into family groups would be easier. Professor 

 Osborn/ like several other American naturalists, 

 cuts the knot by including all the forerunners of 

 the modern horse in the same family as the latter. 

 According to this arrangement, the Equidcc is 

 divided into the following four subfamily groups : — 



1. EQUiNit, including the singlc-tocd Equiis^ Hippidium, 



and Onohippidium. 



2. Protohippin/E, represented by the mostly three-toed 



Pliohippus, Protohippus, Merychippus, Hipparion, and 

 Hippodach Ins. 



3. ANCHiTHERiiN.f;, with the fully three-toed Hypohippus, 



Anchitherium, Mesohippus, Anchilophus, &c. 



4. HvRACOTHERiiN.'F., including the four-toed Lophiotherium, 



Orohippus^ Hyracotherium, tSrc. 



All these very different types are included by 

 the American naturalists under the general name of 

 " horses," which is, of course, distinctly straining 

 the use of that term to an unjustifiable extent. 

 Moreover, the attempt to include all the members 

 of one line of ancestry — or phylum, as it is called 

 by American naturalists — must break down some- 

 where, or otherwise we should have to include in 

 the EquidcB those members of the mammal-like 

 reptiles from which that group is ultimately 

 derived. 



Admitting, then, that arbitrary breaks must be 



^ 77^1? Age of Mammal Sy pp. 555, 556. 



