NORTH AMERICA IN THE ICE PERIOD. 241 



pressed to its original level. Later authors suggest that the formation 

 of ice-caps at the poles would draw away enough water from the lower 

 latitudes to increase the amount of land in the north. Either view 

 would afford conditions adequate to produce the results ; but the evi- 

 dences of such an oscillation, from the nature of the case, are very- 

 difficult to obtain. The most obvious are derived from the existence 

 of ancient river-channels now submerged beneath the ocean. Upon 

 the south sides of Long Island* and Cape Codf there are several 

 ravines channeled out of the till and modified drift, too large to have 

 been excavated by the present drainage system. They are fiords, and 

 many of them are filled by linear fresh-water lakes, kept in position 

 by bars of beach-sand near the ocean's level. Similar facts have been 

 reported in connection with the former entrances of tributary streams 

 into the Great Lakes. J In all these cases the excavations must have 

 been made in connection with the disappearance of the ice-sheet. 



A more extensive series of excavations occurred in the larger riv- 

 ers, though it is not so easy to fix their date. By studying the sub- 

 marine contours off the mouths of the Connecticut and Hudson Rivers, 

 we can follow the outlines of their valleys for many miles out to sea. 

 The first named flowed around the east end of Long Island, leaving 

 the Housatonic to join the Hudson just below Manhattan Island, as 

 shown by Professor Newberry. The Hudson continued southerly for 

 seventy or eighty miles, as first pointed out by Professor Dana. As 

 no deltas appear in connection with the present mouths of these 

 streams, their submergence must have been comparatively recent, while 

 they may have existed as channels of erosion for millions of years. 



A further examination of coast-charts reveals the fact that there 

 is a belt of shallow land bordering the continent from New Jersey to 

 Newfoundland, and that it is as wide as the extensions of the Hudson 

 and Connecticut. Hence, if there has been a submergence off New 

 York, the same oscillation occurred along the whole coast ; and thus 

 a tract of land, as large as Pennsylvania, New York, and the maritime 

 provinces combined, has been lost to the continent, probably since the 

 glacial period. The corresponding area in Northern Europe, which 

 seems to have been elevated at the same time, may be found indi- 

 cated in Geikie's "Ice Age." 



More impressive proofs of a former elevation of the continent ap- 

 pear from a careful study of the lower Mississippi. The contour-line 

 of one thousand feet depth suggests the continuance of the river-bed 

 for a distance of forty miles into the Gulf of Mexico. The Tertiary 

 rocks of the river-basin have been excavated to the depth of two 



* " American Journal of Science," III, vol. xiii, p. 142 (Lewis). 



\ "American Naturalist," vol. xiii, p. 555 (Upham). 



\ The discovery of the pre-glacial outlet of Lake Erie by Professor J. W. Spencer, 

 and the former northerly drainage of the tributaries of the Alleghany River by the Penn- 

 sylvania geologists, have been published since this article was written. 

 vol. xx. 16 



