ITS NEAREST EXISTING RELATIONS 



73 



with dark stripes on the shoulder and legs. Darwin 1 

 collected a number of cases of horses of various breeds 

 and countries so marked, and from them came to the con- 

 clusion of the c probability of the descent of all the exist- 

 ing races from a single dun-coloured, more or less striped 

 primitive stock, to which our horses occasionally revert/ 

 Fossil remains of true horses, differing but very 

 slightly from those now existing, are found abundantly 

 in the most recent geological ages in almost every part 

 of America, from Escholtz Bay in the north to Pata- 

 gonia in the south. Whether any of these remains should 

 really be referred to E. caballus or not — that is, whether 

 they belonged to animals which possessed all the ex- 

 ternal characters attributed above to that species — is, 

 of course, doubtful. Our knowledge of existing forms 

 teaches us that closely similar and perhaps identical 

 skeletal and dental characters may be associated with 

 considerable external differences, especially in the 

 character, distribution, and colour of the hair. If zebras 

 were only known from such portions of their structure 

 as could be preserved in a fossil state, we should never 

 have guessed how greatly they differed in outward aspect 

 from horses and asses. All that we can do with a fossil 

 bone or tooth is to assign it to any known species which 

 it resembles so closely that no actual, definable difference 



1 The Var iation of Animals and Plants under Domestication, vol. i. 

 chap. ii. (1868). 



