ITS ANCESTORS AND RELATIONS 



20 



is still the largest, and forms the centre of support; the 

 second and third are of equal size and form a pair 

 arranged on each side of it. The fifth is an odd toe 

 with nothing to balance it on the inner side of the foot. 

 There is no trace of the symmetry around a line drawn 

 between the third and fourth toes, or of the equality of 

 these two which is seen in Artiodactyles. By referring 

 to the diagram at p. 14 (fig. 1), it will be easily under- 

 stood that A is not converted into B by merely taking 

 away the digit I. 



It can hardly be supposed that the change took place 

 suddenly from a five-toed to a four-toed fore-foot, and 

 indications have been discovered of intermediate forms 

 (Eohippus of Marsh) in which a rudimentary first toe, 

 represented only by the metacarpal bone, existed, but 

 these have not yet been fully described. 



The hind foot of Hyracotherium presents a still greater 

 modification, both the outer digits, first and fifth, having 

 disappeared. The three middle toes, symmetrically dis- 

 posed to the axis of the third, are alone present. This is 

 the condition of the hind foot of all known true Perisso- 

 dactyles of the Eocene and Miocene epoch, and of the 

 greater number of those still existing, the horse alone 

 having advanced to a still more specialised state. 



The teeth of Hyracotherium and its allies are the same 

 in number as in Phenacodus — the same, in fact, as in the 

 vast majority of Eocene mammals ; but they begin to 



