6o2 SCIENCE PROGRESS 



then reached the very climax of their evolution ; Artiodactyla, 

 Perissodactyla, Proboscideans, and Carnivora swarmed into 

 South America, and it went ill with the feebler indigenous 

 quadrupeds of the southern continent. In this way the existing 

 uniformity of the five continents was brought about. 



At this point it is convenient to summarise the foregoing 

 argument. In the Cainozoic we can discern, I think, five 

 periods. The first period is one of general unity, though, as 

 already pointed out, there may never have been a point at which 

 all the land-bridges existed absolutely contemporaneously. 

 Then South America was separated. The second period was 

 therefore characterised by the isolation of South America and 

 the union of the other four continents. Then Africa was sepa- 

 rated. In the third period therefore three enormous but distinct 

 areas of placental evolution existed. This period was long ; 

 it corresponds to the greater part of the Eocene and most of 

 the Oligocene of the stratigraphists. Then Africa was re- 

 united. The fourth period therefore resembled the second in 

 the independence of South America and the union of the other 

 four continents. Finally, South America was at last reunited, 

 and this event ushered in the fifth and last period, an epoch 

 of general unity ; this general unity seems to have been 

 surprisingly perfect in the Pliocene, though seriously interrupted 

 for brief periods during the cold phases of the Pleistocene and 

 at the present day. 



These separations and reunions are the outstanding events 

 of the Cainozoic Era. They profoundly influenced mammalian 

 evolution. They are, it seems to me, to be regarded as literally 

 " epoch-making." 



I have endeavoured to show these periods in the diagram. 

 By abandoning all attempt at depicting the actual form of 

 land, it is possible (in this simple case) to show land and time 

 on one diagram. 



Within the limits of the present article it is impossible to 

 enter into the thorny problem of Madagascar, and, since there 

 are no Anthropoidea there, it is fortunately unnecessary to 

 do so. The old theory put forward by Wallace and Sclater 

 was that the Madagascan fauna (other than the civets) was a 

 " sample " of the pre-Miocene mammalian life of Africa. This 

 theory was invented, however, before the recent discoveries 

 in Egypt, and it is obviously incorrect. Matthew now thinks 

 that the fauna is entirely derived from chance voyagers from 

 Africa. The assemblage has not an African appearance ; but, 

 from wherever the lemurs came, and whenever they arrived, 

 they have not had to compete with their higher relatives, and 

 Madagascar has been the chief centre of lemurine evolution. 



We pass now to a brief examination of the actual fossils 



