THE OLDEST KNOWN FAUNA. 71 



period, in any province of distribution, was made up of forms 

 either identical with, or very similar to, those now living in 

 that area. 



For example, the elephants, tigers, bears, bisons, and hip- 

 popotamuses of the later tertiary deposits of England are all 

 closely allied to members of the existing Arctogaeal fauna ; 

 the great armadillos, anteaters, and platyrrhine apes of the 

 caves of South America, are as closely related to the existing 

 Austrocolumbian fauna ; and the fossil kangaroos, wombats 

 and phalangers of the Australian tertiaries to those which 

 now live in the Australasian province. 



No remains of elephants occur in Australia, nor kangaroos 

 in Austrocolumbia; nor anteaters and armadillos in Europe 

 in Tertiary deposits. 



But, as we go back in time from the Tertiary to the Sec- 

 ondary, this law no longer holds good. Most of the few ter- 

 restrial mammals of secondary age which have been dis- 

 covered belong to Australasian and not to Arctogaeal types, 

 and the marine fauna resembles that of the existing Pacific 

 more than it does that of the Atlantic area, but differs frcm 

 both in the presence of numerous wholly extinct groups. It 

 looks as if, in the latter part of the Cretaceous epoch, a 

 great change in the limits of the then existing distributional 

 area had taken place, and the types now characteristic of 

 the Arctogaeal province had invaded regions from which 

 they had before been shut out. And the assumption of a 

 process of a similar character appears to me to be the only 

 rational explanation of the rapid advent of types absent in 

 the Palaeozoic deposits known to us, in the earlier Secondary 

 rocks. 



Yet other results of first-rate importance have come out 

 of the study of the chronological relations of fossil remains. 

 Cuvier's investigations proved that the hiatuses between 

 existing groups of ungulate mammals tend to be filled up by 

 extinct forms. Later investigations have not only confirmed 

 this conclusion, but have shown that, in several cases, an 

 existing much-modified form can be shown to have been pre- 

 ceded in time, in the same distributional area, by exactly 

 such forms as it is necessar}? should have existed, if the much- 

 modified existing animal had proceeded by way of evolution 

 from a simpler form. 



For certain groups of animals, then, there is as much and 



as good evidence of their having been evolved by successive 



modification of a primitive form as the nature of the case per- 

 4 r 



