842 GEOLOGY 



type had evolved probably from a North American Eocene lemuroid, 

 while rodents of the porcupine type had been derived from some 

 unknown immigrant. That the connection of the continents in 

 the Eocene was only partial or temporary, seems to be implied by 

 the absence in South America of most of the great North American 

 groups (creodonts, carnivores, condylarths, artiodactyls, peri 

 dactyls, and insectivores) . The absence of proboscidians implies 

 a lack of connection between South America and Africa, where these 

 forms developed during the Eocene and Miocene. Many carniv- 

 orous and herbivorous marsupials, similar to those of Australia, 



Fig. 557. Teeth of mastodon (Mastodon longirostris] , showing slightly 

 worn tubercles at the right and much worn ones at the left. (From 

 Gaudry, after Kaup.) 



lived at this time in South America, implying either connection 

 in that direction, or striking parallel evolution. The remarkable 

 South American mammalian fauna is a striking instance of evolu- 

 tion on a large scale in comparative isolation, and in relative free- 

 dom from the severe stimulus of effective competition, powerful 

 carnivores, and shifting geographic relations. 1 



When connection between the two Americas was made in 

 Pliocene time, the fauna of each continent invaded the other. 

 Horses, mastodons, deer, carnivores, llamas, and tapirs, from the 

 north went to South America, while gigantic sloths from the south 

 came to our continent, but did not maintain themselves loiui. 



The herbivores continued to occupy the foremost place amonir 

 mammals. Both the odd- and even-toed ungulates completed 



1 Reports of the Princeton University expedition to Patagonia 1896-99. 



