xix. 3 ANCESTRY OF MAMMALS 537 



times, as a special type of cotylosaurian reptile, with a tympanum 

 placed behind the jaw and later a lateral temporal aperture. This stock 

 quickly became very abundant and successful as the theromorphs 

 and therapsids of Permian and Triassic times, but then nearly died 

 out in the Jurassic, during which period we know of the mammals 

 only from fragmentary remains of a few small animals. Then, in the 

 Cretaceous, some of these small forms became more numerous and 

 from them arose, before the Eocene, a variety of different types of 

 mammal, from which the histories of the modern orders can be fol- 

 lowed in some detail. 



We have, therefore, abundant evidence of the earliest stages of 

 mammalian evolution, say, 20 million years, from fossils in the Per- 

 mian Red Beds of Texas. For the next following 40 million years or 

 so we have also a rich material, from the Upper Permian and Triassic 

 Karroo beds of South Africa. In these early times the mammal line 

 was a flourishing one, more so indeed than the diapsids or any other 

 of the descendants of the cotylosaurs. The animals were mostly 

 carnivorous, though there were also herbivorous types. Many of these 

 early mammal-like forms became quite large and numerous, in fact 

 this stock dominated the land scene in the Permian. In the later Trias, 

 however, if our fossil evidence is a safe guide, their numbers became 

 reduced, perhaps the carnivorous dinosaurs took their place. How- 

 ever, already at this time many of the essential features of mammalian 

 organization had been developed, so far as these can be judged from 

 the bones. Possibly the soft parts of these Permian forms were also 

 mammal-like, the animals may have been active and 'intelligent', have 

 possessed hair and warm blood. However, their brains were small and 

 reptilian in structure. 



Throughout the succeeding 100 million years of Jurassic and Creta- 

 ceous times our knowledge of the mammalian stock is dim. Mammals 

 of a somewhat rodent-like type, the multituberculates, became quite 

 numerous and produced some large forms, but then became extinct 

 in the Eocene. Unless we have been singularly unlucky in the pre- 

 servation of mammalian remains, no other mammal-like animals 

 larger than a polecat existed throughout this long period. Since we 

 possess detailed information, based on numerous fossils, about scores 

 of large and small diapsid reptilian types, it can hardly be only an 

 accident that mammal-like fossils are so rare. 



We must conclude that the mammalian organization, after an initial 

 success in the Permian and Triassic, was almost supplanted in the 

 Jurassic and Cretaceous by the various diapsid creatures. However, 



