[NTIIODUCTIOX. 
xxvu 
beds of Xecrolestes, apparently a close ally of the Cape Golden Moles, and of the 
Sparassodoiita, which, after all, seem to be Creodonts and not Marsupials. Further- 
more, light is also thrown on the numerous points of similarity between Struthioncs 
and the llhem, especially when it is remembered that a large Ratite bird, Ereinoj^tezvs, 
existed in the Eocene of Africa. As to the Ungulates, it seems likely that the 
separation of the two areas took place when the main divisions were only just 
beginning to be differentiated, and that groups like the Pyrotheria and the Arcluuo- 
hyracidac are not ancestral to the Proboscidea and llyracoidea of the Old 'World, 
but more probably represent terms of partly parallel scries which had a common 
ancestry on the common land-surface before the separation of the two regions took 
place. If this were so, we should expect to meet with a general resemblance between 
the various groups rather than a close similarity of structure, and this, in fact, is what 
we find. In the case of the occurrence of the primitive Sirenian Prorastomus in the 
West Indies, and of the Water-Snake Pterosplienus in the Eocene beds of Alabama, 
it seems likely that these animals passed either along the southern coast of the 
Eocene Atlantic or across the bridge of shallow water between the chain of islands 
above referred to as probably lying between West Africa and Brazil. The fact that 
the mammalian fauna of Madagascar is a comparatively poor one and entirely 
lacking in many of the groups that must have inhabited the Ethiopian mainland, is 
considered by Tullberg to be accounted for by supposing that the eastern part of 
Africa with Madagascar was separated from the main South-west African continent by 
an arm of the sea, and that it was not till after the isolation of Madagascar (probably 
in the late Oligocene) that the two portions of Africa became united. At this time 
East Africa was probably united to South-western Asia by continuous land, along 
which the Proboscidea reached India, and perhaps thence penetrated to North 
America. In both these regions, as well as in Europe, the group seems to have 
undergone the further series of modifications which gave rise to the modern type of 
Proboscideans. 
Another consideration which adds to the importance of Africa as a centre of mam- 
malian evolution has been pointed out by Stromer (42), namely, that part of it at 
least has probably never been submerged since the Palaeozoic period, and formed a 
portion of a vast Permo-Triassic land-area inhabited by a great variety of mammal-like 
Theriodont reptiles from which the Mammalia may have actually arisen. This being 
the case, it is not only the Tertiary, but also the Secondary, deposits of this region that 
may be expected to yield most important data for the history of the Mammalia. 
