306 THE TRANSFORMATIONS OF THE ANIMAL WORLD 



take refuge in the vague hypothesis of common 

 ancestors yet unknown, from which these two 

 groups must have issued. 



If we adopt this entirely provisional hypothesis, 

 we should have to suppose at the same time a 

 South African migration, which, at the end of 

 the Trias, would have brought the Multituberculate 

 Mammals across the Mesogean arm of the sea into 

 Europe. As regards the Marsupials, hypotheses 

 are still more misty ; the most ancient representa- 

 tives of the group, the Dromatherium and the 

 Microconodon, appear suddenly in the Trias of 

 North Carolina without any known ancestor. 

 From the United States, thus considered as the 

 probable centre of dispersion of the group, the 

 flesh-eating Marsupials must have emigrated; on 

 the one hand, to Europe, where we discover them 

 in the lower Jurassic of Stonesfield and in the last 

 layers of the Jurassic of the Isle of Purbeck, and 

 still higher in the form of the Didelphids in the 

 upper Eocene and in the Oligocene of France and 

 Germany ; on the other hand, towards South 

 America, the present home of the Sarigues, where 

 Ameghino has discovered for us the Proteodidelphys 

 of the lower Cretacean of Patagonia. It is, no 

 doubt, from the South American Continent that 

 the Marsupials have been able, at a recent epoch 

 (the upper Tertiary or Quaternary) to emigrate 

 to Australia and Tasmania, the present geographical 

 centre of these non-placental Mammals. It ap- 

 pears probable, conformably with the hypothesis 

 of Osborn, of Matthew, and Ameghino, that this 

 American migration towards Australia can only 



