84 BULLETIN: MUSEUM OF COMPARATIVE ZOOLOGY. 
further details concerning the African Permian sea it is possible to 
admit a shallow sea in the Atlantic area as far north in that district 
as the southern tropic, but this at a later stage than the epoch of 
principal glaciation in Africa. Both in Africa and South America, 
the marine beds of the Permian demand land near sea-level immedi- 
ately before and after the invasion by the sea. Little else can at 
present be argued from their occurrence. 
The volcanic islands which stake out the mesial line of the Atlantic 
and its connection with the Arctic Ocean hug the shores of Europe 
and Africa rather than those of the Americas. Since those voleanoes 
whose substructure is known are associated with crustal displacement 
involving either horizontal or vertical motion or both and particularly 
the latter mode of derangement, it is probable that we see in these 
islands the indirect evidences of a geologically recent movement of the 
Atlantic bottom which we know from Iceland was well under way in 
Miocene times. Along the shores of North and South America which 
trend northeast and southwest we find traces of downsinking of the 
land in Cretaceous times. In the North Atlantic region, the Pan- 
Appalachian mountain-chain extending from the dislocated structures 
of the Rocky Mountains across the southern United States to the 
coast in Newfoundland can be traced north of the Alps in Europe 
into Asia but is interrupted in the Mississippi embayment and by the 
North Atlantic basin which cuts across the folded structures as if they 
had sunk to form the present ocean floor. These and other indicated 
changes of depth and outline of the Atlantic province make it in- 
credible that in Permian times the basin had much of its present 
length, breadth, and depth. We may conclude therefore that the 
geologist is free to converge the coasts of Africa and South America in 
Permian and earlier Carboniferous time as closely as any biological 
facts and geological evidences of land may demand for their explana- 
tion. 
In south Brazil all the known facts from the Atlantic border demand 
an extension of land beyond the present coast in late Palaeozoic time, 
but how far towards the coast of Africa we can not say. The dis- 
covery of the Gondwana flora in the Falkland Islands makes it possible 
to effect the distribution of these plants from Australia into South 
America by way of the Antarctic continent. Possibly Africa also 
received its population by this route or more likely from India. 
The revealing of the geology of the Antarctic continent receives from 
this state of the problem of Gondwana-land a renewed interest. 
The Permian Glacial Problem.— In the foregoing account of the 
