riers, we are coming to a point where the efficiency of this factor can be safely taken into account. The outcome of free interbreeding, as Jordan has pointed out, is to unify species and obliterate variations. Per contra, isolation checks this process and gives freer play to tendencies arising from other factors in variation. The effect is thus, as a general rule, negative, but expresses itself freely enough in geographic provinces severed by some barrier or condition which has the effect of a barrier. Among existing species the formative effects of segregation have been very largely illustrated from restricted areas such as the subdivisional valleys and forests of Hawaii with its distinctive forms of the Helicidæ and other terrestrial snails—a case that is paralleled in paleontology by the snails of Steinheim. But the effect is to be reckoned with in larger or continental areas between which there has been at one time opportunity of interchange, especially in the case of marine species, with which we chiefly deal, along the epicontinents.
I have particularly in mind phenomena which have been brought to my notice by a somewhat extended study of the Devonian faunas of the southern hemisphere and the broader application of the factor is best enforced and illustrated by this instance. I may say that this broader notion seems to be that entertained by Darwin so far as he specified the conception of geographic segregation as an element in natural selection and it was his work in South America that formed the basis of his conclusions.
With other students we recognize the existence during the Devonian of austral continental lands which have been variously designated and variously outlined. By some this land has been posited as a north and south Atlantis lying in the meridional axis of the present ocean, by others a broken land mass partly crossing the southern Atlantic from east to west. But now we begin to see its continuity and the extent of its strands, with something of its changes in outline during its early history. It was the precursor and the nucleus of Gondwánaland. With it began, so far as we now know, the long history of that continental land and the successive records of life developing under continued conditions of geographic isolation from the northern strands.
From Argentina, Bolivia and northern Brazil we have very lucid evidence, on the basis of paleontology, that in the late Silurian the shore lines were continuous with those of the north. We have no dependable knowledge of these earlier faunas at the east and indeed their entire absence is indicated by stratigraphy; but with the submergence of the Silurian at the west, there entered from the African east upon this south Atlantic field, a positive diastrophism whose axis was well nigh normal to that of the present Atlantis, and along the shores of this growing land bridge entered an invasion of marine life
