TERTIARY MAMMAL HORIZONS. 53 



Among other problems, a land connection between Africa and 

 South America across the South Atlantic enables us to explain 

 the remarkable distribution of the Sirenia, sea-cows, dugongs 

 and manatees, now found exclusively in the tropical belt of 

 Africa and the Americas. (See Sirenia, Fig. III.) These ani- 

 mals first appear in the Oligocene of Germany. It is also, of 

 course, possible that they may have taken a northern route, as 

 indicated by the remains of RJiytina in the North Pacific. 



Before confining our attention to Arctog.ea, let us further con- 

 sider the mesozoic relations of the three realms. (Fig. I and 

 Fig. III.) 



In the Jurassic period stem forms of insectivores, marsupials 

 and possibly of monotremes ^ are found in Arctogaea and seem 

 to establish the theoiy of the northward origin of the mammalia 

 as a class. 



DoLLO ('99), has recently endeavored to demonstrate that all 

 Marsupials have been evolved from arboreal forms like the 

 Opossum. If we can draw a parallel with the adaptive radia- 

 tion of the placentals during the 3,000,000 years, more or less, 

 of the Tertiary, we may safely conclude that such a primitive 

 family, entering the Australian region during the Cretaceous 

 period either by way of Antarctica (Spencer) or by way of the 

 Oriental region (Wallace and Lydekker), might have peo- 

 pled Australia with all its wonderfully diversified forms of 

 Marsupial life. The Didelphyida^ are to the Marsupials what the 

 Creodonta are to the Placentals in point of potential evolution. 

 The JMojiotrcnics also may have entered Notog.ea by either of 

 these routes. 



North America is the only part of the globe where Cretaceous 

 mammals are known at present. In the late Cretaceous we ap- 

 pear to discover evidence of the existence of the following 

 orders : Insectivora, Creodonta or ancestral carnivores, hoofed 

 animals or Amblypoda and perhaps the earliest monkeys or 

 Mesodonta. In the basal Eocene we certainly find primitive 



iThe writer's view (Osborn, '88) that the Jurassic mammals of England and 

 Wyoming embrace primitive placentals or insectivores as well as marsupials and 

 multituberculates (? monotremes) is now generally accepted. 



