centre the
elaphines spread westward and eastward, resulting in Europe in the red
deer, which penetrated southward into north Africa at a time when there
was a land connection across the Mediterranean. In the opposite
direction, the nearer we get to Bering's Straits the closer is the
resemblance to the American wapiti, until the splendid species from the
Altai Mountains (_C. canadensis asiaticus_), and Luehdorf's deer
(_C. c. luehdorfi_) from Manchuria, are regarded only as sub-species
of the eastern American form, which they approach through _C. c.
occidentalis_ of Oregon and the northwestern Pacific Coast.

This evidence is conclusive in itself, and is further confirmed by the
geological record, from which we know that the land connection between
Alaska and Kamtschatka was of Pliocene age, while we have no knowledge
of the wapiti in America until the succeeding period.

While there is not the least doubt that the smaller American deer had an
origin identical with those of the old world, the exact point of their
separation is not so clear. Two possibilities are open to choice:
_Mazama_ may be supposed to have descended from the group to which
_Blastomeryx_ belonged, this being a late Miocene genus from
Nebraska, with cervine molars, but otherwise much like _Cosoryx,_
which we have seen to be a possible ancestor of the prong-horn; or we
may prefer to believe that the differentiation took place earlier in
Europe or Asia, from ancestors common to both. But there is a serious
dilemma. If we choose the former view, we must conclude that the
deciduous antler was independently developed in each of the two
continents, and while it is quite probable that approximately similar
structures have at times arisen independently, it is not easy to believe
that an arrangement so minutely identical in form and function can have
been twice evolved. On the second supposition, we have to face the fact
that there is very little evidence from palaeontology of the former
presence of the American type in Eurasia. But, on the whole, the latter
hypothesis presents fewer difficulties and is probably the correct one;
in which case two migrations must have taken place, an earlier one of
the generalized type to which _Blastomeryx_ and _Cosoryx_ belonged,
and a later one of the direct ancestor of _Mazama_. There is
little difficulty in the assumption of these repeated migrations,
for evidence exists that during a great part of the last half
of the T