992 HISTORICAL GEOLOGY. 



three levels ; and part may have been occasioned by the contributions of 

 side valleys, and unequal resistance to Avear. On account of this feature 

 the formation is often called the Terrace formation. 



On the Connecticut the upper flood plain or terrace is, for the most of the way, 150' 

 to 250' above the river. Large deposits of clay occur in the lower part, and others of less 

 extent at various levels to the top. The stream owes its abundant waters to the high 

 mountains about its sources, of which the White Mountains were the higliest. The depth 

 of water may have been 50' or 100' ; there is no basis for a satisfactory estimate. 



Along the Hudson River the height of the upper terrace is 100' to 280', and finally 

 340' between Albany and Schenectady. As before stated, the heights increase to the 

 northward, where the Cliamplain subsidence was greatest. 



The Connecticut River had a dam at the narrows below Middletown, Conn., as the fall 

 off in the terraces below it shows ; another, as stated by B. K. Emerson, near Northamp- 

 ton, Mass., between the opposite trap ridges. Mount Tom and Mount Holyoke ; and possibly 

 others. Just above the Northampton dam, where the upper terrace-plain is about 200' 

 above the river, a portion of the flood-waters escaped over the west bank near Florence, 

 passed to the west of Mount Tom, and took a southward course along the Farmington valley, 

 as the levels of the terrace-plain show, to New Haven, Conn., where it was discharged by 

 the bay into the Sound — resuming thus a route followed by the whole Connecticut stream 

 or estuary in Triassic time. An ice-dam, or drift-dam, closed a narrow gorge through the 

 trap ridge above Hartford, Conn., which was the channel of the Farmington River, and 

 another deep gorge through sandstone above Cheshire, Conn., that of the Quinnipiac River, 

 so that the new discharge-course of the Connecticut secured the upper parts or heads of 

 both the Farmington and Quinnipiac rivers as its tributaries, and took possession of the 

 valley of the small stream called Mill River to reach the Sound. On the terraces of the 

 Connecticut valley, see the author's papers of 1870 to 1884 ; also E. Hitchcock, 1841, 1857 ; 

 C. PI. Hitchcock, Vermont Geol. Bep., 1861, and New Hampshire Geol. Bep., 1878; W. 

 Upham, New Hampshire Geol. Bep., and later papers. 



W. B. Dwight states that at the clay-beds, near Newburg, north of the Narrows, the 

 clay fills large conoidal depressions in the sand-beds. One of the three there observed is 

 elliptical in section, about 80' by 50' in diameter at bottom, 150' in longer diameter at top, 

 and 90'-100' in depth. The clay is straticulate, the layers concave, with the wall of the 

 mass rather firm ; and the sand and gravel beds outside bend downward at the wall. 

 The clay contains a few bowlders. 



In western Pennsylvania, plains of great extent have a height of 275' to 300' along the 

 Monongahela, and of 300' on the Ohio 5 miles below Pittsburg ; their height above 

 the sea level is about 1050'. Nothing of marine origin, however, has been found in the 

 region to suggest the presence of the sea. On the lower Ohio occur terraces 160' to 100' in 

 height above the river. They exist also along the Mississippi in Kentucky, and farther 

 south. 



Kettle-holes, although characteristic of many moraines, also occur at 

 times over the stratified fluvial deposits of the Champlain period. An exam- 

 ple occurs in the plain on which the city of New Haven, Conn., is built, one 

 to four miles north of the center of the city. The terrace formation of the 

 region consists chiefly of sand and fine gravel. The small depressions repre- 

 sented on the accompanying map. Fig. 1555, are the kettle-holes. They are 

 often 100 to 150 feet in diameter, and 30 to 40 feet deep. The ice had left 

 the region long before deposition of the beds had taken place. On the map, 

 the depression in the plain lettered Beaver Pond Meadows has a depth of 25 



