294 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1912. 



ver}^ soft, marly, salt-bearing beds of the Salina formation, generally 

 200 or 300 feet in thickness. The selective processes of erosion led 

 the streams to attack the softer strata with greatest effect, whUe the 

 harder limestone resisted and formed the great escarjDment which 

 now characterizes it, from New York to Wisconsin. ExtensiA^e val- 

 leys were eroded m the soft rocks below it and the limestone ledge 

 was driven back as fast as it was undermined. Other valleys were 

 also excavated m the soft shales above. 



Thus, Lake Ontario, Georgian Bay, the northern channel of Lake 

 Huron, and Green Bay were excavated out of the soft rock below the 

 limestones of the Niagara group, while Lake Erie, .the main body 

 of Lake Huron and all of Lake Michigan were excavated out of the 

 soft strata above the limestones. Lake Superior appears to be some- 

 what exceptional. It is thought to be largely an original rock basin, 

 or perhaps a syncline out of which the soft rocks have l^een eroded. 

 These softer rocks were probably mainly those that lie below the 

 limestone of the Niagara group. 



Thus, the shape and size and arrangement of the lake valleys 

 were primarily dependent upon the geological structure — upon the 

 relative position and thiclaiess of the soft beds and the distribution 

 of their exposed parts. Where the soft beds were exposed to effective 

 stream erosion they were removed more rapidl}^ than the harder 

 rocks, and thus became the main valleys of the region. 



In the present attitude of the land the Paleozoic strata dip dis- 

 tinctly but gently southward in the basm of Lake Ontario; south in 

 the eastern part of Lake Erie, and southwest and west in its western 

 part; toward the southwest in the main part of Lake Huron, but 

 toward the south in the northwestern part of this basin; toward 

 the south m the northeastern part of Lake Michigan, toward the 

 east in the southern part, and toward the south in the peninsula 

 east of Marquette, while farther west the older rocks bordering Lake 

 Superior on the south dip steeply northward toward the axis of the 

 basin and the dips are various m other parts. 



That these valleys were going through the process of development 

 by erosion during practically all the time from the close of the Paleo- 

 zoic to the beginning of the glacial period seems not improbable. 

 Indeed, the time must have been very^long to have made such extensive 

 valleys by so slow a process. It might be thought that some move- 

 ment of elevation or tilting had turned these old valleys into lake 

 basins long before the time of the Ice Age, but no certain evidence 

 indicating such a change has been found. Up to, or nearly to, the 

 beginning of the Ice Age the valleys appear to have had complete 

 drainage by rivers and held no lakes. 



