204 BATLEYV, WiEEETS. 
completed, we may infer further that the dominant features of 
oceanic circulation have obeyed the conditions of atmospheric cir- 
culation and of rotation of the sphere which now govern the great 
oceanic eddies. We may introduce in the Atlantic and Pacific the 
dominant drifts from east to west in equatorial regions with the 
resulting circulation northward along the east coast and southward 
along the west coast of the continent. A circulation of the oceanic 
waters in the epicontinental seas must result from the great oceanic 
drifts, and the direction of flow will be determined by the configura- 
tion of the lands and the depths of the seas. 
From the geographic conditions thus developed inferences regard- 
ing the climate and the life habitats of the time may be drawn. If 
now we turn to the records of paleontology and compare the dis- 
tribution of faunas and floras with the conditions of distribution 
which should result from the inferred physical phenomena, we may 
check the whole line of reasoning and by a readjustment draw a 
step nearer to the truth. This is the method which has been pursued 
in making the maps of North America that are published with the 
papers in this number and that will appear in connection with 
further papers of the series. 
In a first essay of this kind (and I am not aware of any earlier 
attempt to combine the various lines of evidence in a similar manner) 
it is probable that important facts have been overlooked. The 
very broad scope of the discussion makes this probability almost a 
certainty, and it is not to be expected that the maps here presented 
should give a final or satisfactory solution of the problems. They 
are to be regarded as tentative and suggestive only. 
On one point they have been particularly criticized, it being said 
that each individual map covers so long a period of time and such 
diverse conditions that they do not truly represent any special geo- 
graphic phase of the continent. This criticism is valid, and one of 
the steps in the advancement of knowledge will be that of selecting 
narrower time limits and more precise correlations than have been 
attempted in these cases. We may undoubtedly make progress in 
this direction at the present time so that the fifteen maps which 
will accompany this series may be replaced by two or three times as 
many; but there is danger in carrying the refinements too far on the 
