MATTHEW, CLIMATE AND EVOLVTIOX 293 



limited and conditioned by distribution of land and water or by moun- 

 tain or desert barriers than is that of mammals. Climate and environ- 

 ment are much more important factors. Their dispersal is accordingly 

 much wider, and this is especially true of the more migratory and strong- 

 flying types. The general course of their dispersal from the northern 

 land masses is in some respects much more obvious than with the Mam- 

 malia, provided we allow for the extreme imperfection of their geological 

 record; but on this account, it is not supported by the mass of direct 

 evidence which we have among mammals. 



The most primitive living birds, the penguins, are Antarctic in their 

 distribution, and as fossils are known only from the Antarctic Tertiaries, 

 where they include gigantic terrestrial adaptations. It is of interest to 

 note that the only actually known land vertebrates of the Antarctic con- 

 tinental area are penguins. If this continent had been united during 

 the late Mesozoic and early Tertiary to Australia and South America, 

 we should expect to find a fossil mammal fauna, probably highly pro- 

 gressive and specialized before the spreading ice swept it out of existence. 

 We might, indeed, -hope to find a few marine adaptations from this mam- 

 malian fauna still haunting the edges of the Antarctic pack. But in 

 fact, the three items which to my mind have a bearing upon early Ter- 

 tiary conditions in Antarctica all point towards continued isolation and 

 obviously parallel the fauna of oceanic islands. These are, — 



1) Gigantic land-penguins in the ? Eocene deposits of Seymour Island 

 (also in Patagonia). Compare with the gigantic land birds of various 

 oceanic islands, correlated with paucity or absence of land mammals. 



2) The living marine penguins are not readily interpreted as a pri- 

 marily marine adaptation, but they are very easy to understand as modi- 

 fied survivors of a group formerly of terrestrial habits, altered to meet 

 the present conditions under which alone could life be maintained on 

 the Antarctic shores. 



3) The occurrence of Miolania, as interpreted on page 283, is sug- 

 gestive of the former presence of giant land-turtles in Antarctica, al- 

 though not explainable as evidence of former land connections with South 

 America and Australia. 



There may be other indirect evidence in the distribution of marine 

 Vertebrata and Invertebrata, which, if conservatively interpreted, would 

 confirm or disprove these indications. So far as they go, they suggest 

 that ground-birds and land-turtles were the large land vertebrates of 

 Tertiary Antarctica as in oceanic island faunae of to-day. 



The distribution of modern land birds is universally interpreted in 



