MATTHEW, CLIMATE AND EVOLUTION 27a 



lelism from immediate affinity,^' by the relative scarcity of fossils as com- 

 pared with living species (among land animals), and by our less certain, 

 knowledge of the causes which may control their evolution, their means 

 of migration, and their true evolutionary history and affinities. 



INTERPRETATION" OF NEGATIVE EVIDENCE IN FOSSIL MAMMAL FAUN.5; 



In considering a Tertiary mammal fauna, we must keep in mind the 

 facts that there may be large facies of it that are represented imperfectly,, 

 if at all, in our records, and that there may be important parts of it 

 which have left little or no record, owing to their habitat, small size or 

 other circumstances. We may, with some reserve, conclude that the en- 

 tire absence from the record of a group which is abundant In other faunae 

 indicates its real absence from the fauna. But we are not justified in so 

 concluding in the case of rare or inconspicuous races. It is fair to as- 

 sume that the absence of Perissodactyla from the Oligocene fauna of 

 Egypt or the Miocene fauna of Patagonia was real, and not a matter of 

 defective record. The same assumption would be unjustified in the case 

 of didelphid marsupials and dilambdodont Insectivora respectively. But 

 the most conclusive evidence of the absence of a certain group from a 

 given fauna is that while it is not found fossil, another group is found to 

 have become adapted on parallel lines, taking its place in the fauna. The 

 absence of Perissodactyla and Artiodactyla from the Miocene of South 

 America is confirmed by our finding Litopterna, Toxodontia and Astra- 

 potheria, which parallel in adaptation the horses, rhinoceroses, tapirs, 

 camels, etc., of the North ; the absence of Carnivora by the parallel adapta- 

 tion of marsupials to take their place. The evolution of lemuroid pri- 

 mates in Madagascar into large quadrupedal forms apparently paralleling 

 certain groups of Ungulates,"^ affords some evidence that the Tertiary 

 hoofed mammals were unable to invade Madagascar. 



The absence of fissiped Carnivora from the recorded Oligocene fauna 

 of Egypt would not be conclusive in itself ; but, coupled with the excep- 

 tional variety and abundance of the more archaic creodonts of the family 



"* It may be noted in illustration of this point that a natural cast of the entire carcass 

 of a mammal would afford far less secure information as to its real affinities than would 

 a fossil skull, and less even than a lower jaw with reasonably perfect teeth. The parallel 

 adaptations so frequently recognized among mammals lead to superficial resemblance of 

 distantly related types whose true affinities are readily recognized by the internal struc- 

 ture. If. as among most invertebrates, we had only an external skeleton to guide us, 

 the real affinities would not be so securely recognized. 



»5 The skull and the short limbs of Megaladapis are very suggestive of such types as 

 Promerycochwriis. The feet do not, however, indicate a terrestrial habitat, nor are the 

 teeth efficient in grinding. The resemblance in teeth and skull of Archaolemur to the 

 Anthropoidea is very marked. 



