Great Ice-Da I lis. — Taylor. 37 
logical thinking could hardly carry unfavorable inference so 
far. For, as Mr. Leverett pointed out in a recent letter* the 
presence of the dams at the supposed places was not only not 
dispraved by the later beach extensions, but the fact that those 
extensions are all much weaker and fainter than the previously 
known parts of the same beaches is, in reality, favorable to the 
supposition of ice-dams. Otherwise, why should the beaches 
change their strength and character where they cross or in- 
tersect moraines? The thing that the advocates of ice-dams 
had apparently failed to do was to find a dam so situated that 
its withdrawal would cause the lake to fall to a lower level — 
they had not certainly located a dam that was critical in the 
lake historv. This is logically the most that Dr. Spencer could 
have claimed. In this connection it should be remembered, 
however, that it was not until 1896 that the outlet channels in 
southeastern Michigan were recognized and correlated with 
the beaches and moraines of the Erie basin. In the Ohio-New 
York area, Mr. Leverett found no outlet channels to guide him 
in his correlations. It was the finding of the channels in Mich- 
igan that brought out clearly the relations between the dams 
and the beaches, as the writer has endeavored to show more 
fully than before in this paper. We have seen above how the 
faint extensions came to be made and why it could hardly be 
expected that they could be traced to a clear connection with 
those moraines that mark the place of critical dams. 
In a letter in reply to Mr. Leverett, Dr. Spencer main- 
tains, further, that gravel floored channels across divides are 
not ''evidence per se of glacial dams."f No one would claim 
that they are, except in regions that are known to have been 
glaciated; and even there the channels alone can not be re- 
garded as full and final proof, but merely as strong evidence. 
If, however, such channels are found connected with shore 
lines in hydrographic basins favorably situated the presump- 
tion for ice-dams becomes all but conclusive. In such cases 
the theory of ice-dams is a more likely guide to a true inter- 
pretation than any other and ought not to be discarded until 
it is clearly disproved. Dr. Spencer appeals to a channel across 
*Am. GeoL, vol. XXI. March, 1898, pp. 195-199. 
tAm. Geo]., vol. XXI, June. 1898. pp. 393-396. See also "Chan- 
nels over Divides not Evidence per se of Glacial Lakes." Bull. G. S. 
A., vol. 8, 1891, p. 491. 
