Page:Popular Science Monthly Volume 20.djvu/252

240 shown in the numerous excavations made for building purposes about Brooklyn. The bowlders upon Long Island have all been brought from the northwest, and those familiar with the ledges in Connecticut and up the Hudson can readily recognize the fragments in the various parts of the island—the Palisade traps in Brooklyn, and the New Haven red sandstones to the southeast of their place upon the mainland.

We can recognize the inner moraine in Ohio, while its remnants may be found in New York State by future investigators. Dr. Newberry describes a line of kames occupying the water-shed between the affluents of the Ohio and Lake Erie, which can be easily correlated with the modified drift fringing our terminal moraine. This line extends entirely through Ohio, and bends sharply at Fort Wayne, Indiana, following St. Joseph's River northeasterly, so that its whole course is parallel to the shore of Lake Erie. N. II. Winched thinks there are several of these moraines in the northwest part of Ohio. Professor T. C. Chamberlin has generalized the facts about the course of these moraines between Pennsylvania and Minnesota, and supposes there are two morainic lines parallel to the shore of Lake Erie, the outer reaching to middle Indiana. Likewise there appear moraines in the form of loops following the course of the shores of Lake Michigan, Green Bay, Keweenaw Bay, and the southwest end of Lake Superior. The glaciers seem to have followed the several valleys, continuing to flow as long as the material lasted. The Green Bay and Lake Superior streams did not cover the area in the lee of the highlands of northern Wisconsin; and hence there was a large tract of land, occupying essentially what is known as the "Lead-region of the Northwest," over which we search in vain for erratics or glaciated surfaces. Mr. Warren Upham has communicated to us many facts for Minnesota, Dakota, and Iowa. They indicate two looped moraines west of the Mississippi: one reaching nearly to Yankton, Dakota, having the celebrated Coteau de Missouri for its western border, and part of the Coteau de Prairie for its eastern; the other taking the eastern line of the Coteau de Prairie for one side, and pointing east of south to Des Moines, Iowa. An inner loop connecting the Wisconsin moraine with that of the Leaf Hills may have been pushed there by the Lake Superior stream, and the more southern loops may have had some connection with glaciers starting in the Dominion portions of the Rocky Mountains. Hence the Labrador and Rocky Mountain sheets may have been confluent, and, owing to the great masses of ice thus accumulated along the upper part of the tributaries of the Mississippi basin, we may understand why the glacier extended so much farther south in the interior than upon either coast.

—The earlier writers accounted for the glacial cold by supposing the land had been elevated sufficiently to lower the temperature, and subsequently