xxviii. 7 THOATHERIUM 703 



rabbit-like creatures with gnawing incisors. The notoungulates thus 

 radiated to form various types and for many millions of years they 

 were the dominant herbivores of the South American forests. 



6. *Order Litopterna 



Some descendants of the condylarths in South America developed 

 along lines astonishingly similar to the horses. The didolodonts 

 already show tendencies in this direction and are, indeed, sometimes 

 removed from the condylarths and placed with the South American 

 horse-like forms in the *order Litopterna. A series of fossils shows 

 that members of this order became first digitigrade and then unguli- 

 grade, the central metapodials elongating and the lateral ones reduc- 

 ing, first to three and then, in *Thoatheriwn (Fig. 469), to a single one, 

 with splint bones even smaller than remain in our horses. The general 

 appearance of the limbs was very horse-like, for instance in the 

 grooved talus, but the carpus never became interlocking. Other 

 respects in which these litopterns developed less far than the horses 

 were that the tooth row remained nearly complete and the molars 

 low-crowned, though provided with ridges. A post-orbital bar was 

 developed. These differences from our horses are as interesting as the 

 similarities, and they show that the features of the ungulate facies do 

 not necessarily all evolve together. It is impossible to say what 

 difference in conditions was responsible for forming horse feet on an 

 animal whose head was only partly horse-like. 



These ungulates were presumably evolved to meet conditions on 

 the South American plains in the Miocene. It is interesting that the 

 three-toed types outlived the one-toed *Thoatherium, lasting into the 

 Pliocene. *Macranchenia (Fig. 469), a creature looking like a camel 

 but perhaps living in swamps, was the only Pleistocene survivor. 



7. *Order Astrapotheria 



This order includes some Oligocene and Miocene South American 

 ungulates, with a short skull but long lower jaw, long canines, and 

 massive molar teeth (Fig. 469). There was probably a proboscis. The 

 feet were small and perhaps rested on pads. The weak vertebral spines 

 and transverse processes suggest that the animals may have been 

 aquatic and the large lower canines, diverging in older animals, 

 resemble those of a hippopotamus. 



