148 GRADUAL MODIFICATION OF ANIMAL FORMS xi 



considering how little evidence we have of the animal life of 

 the middle or older tertiaries of South America, this is not to 

 be wondered at. 



On the whole it will be seen that, taking actual anatomical 

 characters alone, palseontological research, even so far as it 

 has yet been carried, bridges over most of the gaps existing 

 between the modern forms of Perissodactyles, entirely abolish- 

 ing, for instance, the order Solipedia, as it is impossible to draw 

 a satisfactory line where the animal ceases to be equine and 

 becomes a palaeotheroid ; some drawing it between Anclii- 

 therium and Palceotherium, some between Anchitherium and 

 Hipparion. Moreover, and this is most important, the lines 

 from the modern more specialised forms converge towards the 

 ancient more generalised forms ; so that if we could get a 

 side-view of what is shown in the diagram, the earliest forms 

 at the bottom and the latest at the top, we should have lines 

 (broken, it is true, here and there) diverging from a common, 

 or near a common centre, towards a circumference above a 

 view, in fact, of the conventional genealogical tree. 



We turn now to the Artiodactyles, represented at present 

 by the scattered groups before spoken of, clustering round two 

 type-forms so widely sundered in their structure and habits 

 as the pig and the ox ; but the former history of this division 

 yields a totally different state of things. Of early eocene 

 Artiodaetyles we know very little at present ; but in the 

 later divisions of the same epoch forms appeared, such as 

 Anoplotherium, Dicliobunz, Cheer opotamus, and Hyopotamus, 

 which were certainly neither pigs nor ruminants, but which 

 partook remarkably of the characters of both. They had 

 the complete number of teeth, i.e. incisors and canines, like 

 modern pigs, but molars with indications of the crescentic 

 pattern so characteristic of ruminants. They had two or four 

 toes ; but the metacarpals and metatarsals were not united to 

 form a cannon bone as in ruminants, and they wanted the 

 horny appendages to the head, so usually met with in the 

 modern representatives of that group. From some of these 

 central forms, or more probably from a still earlier allied 

 group indicated by the genus Acotlierulum, or by some other 

 still undescribed remains from Mauremont, transitions can be 



