MATTHEW, CLIMATE AND EVOLUTION 265 



Eocene of Wyoming, in the Oligocene of Colorado and in the Upper 

 Eocene to lower Miocene of France and Germany. They are not known 

 from any later formation in any of the northern continents. 



In the Southern continents, they assumed a much more important 

 position. In South x\merica, in the absence of placental carnivora, the 

 polyprotodont marsupials developed into a number of large and small 

 predaceous mammals (Borhyaenidse), so closely paralleling some of the 

 predaceous marsupials of Australia that they have been referred to the 

 same family (Thylacinidre). Pseudo-dip rotodont marsupials were also 

 fairly common, taking the place in the fauna held by Insectivora in the 

 North, this group of placentals (except for a single type) not having 

 reached South America. The marsupials of South America did not de- 

 velop into groups taking the place of northern ungulates, rodents or 

 primates, since primitive placentals of these groups (Condylarthra, 

 ? Hystricomorpha, ? Lemuroidea) had penetrated into South America 

 before it was separated from the Northern world, and there developed 

 along lines sub-parallel to the development of the higher placental groups 

 in the North, but distinct and less progressive. 



In Australia, the marsupials assumed a still more important position, 

 as the only mammals of that continent. The placental mammals of the 

 northern Tertiary did not reach Australia, except for a few strays — bats 

 and mice and the dingo — which were too few in numbers and of too re- 

 cent introduction to affect seriously the course of mammalian evolution 

 on that continent. In the absence of placentals, the marsupials developed 

 into a wide variety in size, form and habits of life, partially paralleling 

 the higher mammals. 



The near resemblance between the modern Australian Thylacinus and 

 the Borhygenidaj of Tertiary South America has been used as an argu- 

 ment for an Antarctic connection between the two. Such a hypothesis 

 will not bear close examination. The resemblance is not closer than 

 between parallel adaptations in distinct families of true Carnivora, whose 

 genealogy has been more or less completely traced back through inde- 

 pendent lines of descent from unspecialized common ancestors. It is 

 not closer, for instance, than that between the Oligocene Felidse and 

 the modem Cryptoprocta of Madagascar, whose common descent from 

 an unspecialized placental carnivore (Viverrid or Miacid), analogous to 

 the marsupial didelphyids, is generally admitted. The common char- 

 acters distinguishing thylacinids and borhyaenids from the didelphyids 

 are, without exception, such as would naturally be assumed independently 

 in adaptation to predaceous terrestrial life and have been so assumed in 

 numerous independent parallel adaptations of the same sort among 



