MATTHEW, CLIMATE AXD EVOLUTION 203 



one of them are such as might possibly at least have reached the island 

 without continental union, whether by accidental transportation, by 

 swimming or by other means. 



Take for example the mammals of Sumatra, Java and Borneo. We 

 cannot reasonably suppose that the rhinoceroses, tapirs, deer, wild dogs, 

 felids and numerous other large animals common to them and the ad- 

 joining continents reached these islands except by land. They are too 

 large for transportation on "rafts" of vegetation such as occasionally 

 drift to sea from the mouths of tropical rivers. They are dry-land ani- 

 mals not given to swimming long distances. And we would not invoke 

 the agency of man to account for a whole fauna. But most important 

 is the fact that all the animals that we might fairly expect to find there 

 in view of a former land connection are really present. 



Contrast with this the fauna of Madagascar.'^ There are no ungulate 

 mammals there, except for the bush-pig, possibly introduced by man (in 

 accord with known customs of the Malays) and a pigmy hippopotamus 

 (now extinct) which might have reached the island by swimming, as 

 hippopotami are kno^wm to travel considerable distances by sea from one 

 river mouth to another. The great majority of the unguiculate groups 

 of the mainland are also absent. The only representatives are a few 

 very peculiar carnivores of the family Viverridge, a peculiar group of 

 insectivores (Centetidae) and a peculiar group of Cricetine rodents, each 

 apparently evolved on the island from a single type introduced long ago, 

 a species of shrew (Crocidura) of more recent introduction and a variety 

 of bats. There are numerous lemurs and no monkeys there ; and the 

 lemurs appear to have radiated out from a single group^^ into a 

 number of peculiar types, two of which, now extinct, paralleled the 

 ungulates and the higher apes in several significant features. The fauna 

 of the island does not resemble the present fauna of Africa, nor can it 

 be derived from any one past fauna, known or inferential, of that conti- 

 nent. The attempt to derive it from the present or from any known or 

 inferential past fauna of India involves still gi-eater difficulties. On the 

 contrary, the Malagasy mammals point to a number of colonizations of 

 the island by single species of animals at different times and by several 

 methods. Of these colonizations, the Centetidae are the earliest, perhaps 

 pre-Tertiary ; the lemurs, rodents and viverrines are derivable from one 

 or more middle Tertiary colonizations : and in both cases the ''raft" 



31 A. R. Wallace: Island Life, pp. 381-412. ISSl. See also Trouessart Catalogus 

 Mammalium and Suppl. Quinq. ; Lydekker. Geog. Hit. Mam., pp. 211-226. 1896. Lydek- 

 ker's arguments for continental union are mostly invalidated by more recent discoveries. 



^- See W. K. Gregory's studies upon the affinities of the Lemuroidea, forthcoming in 

 Amer. Mus. Bulletin. 



