Such has

ice will gradually withdraw until it entirely disappears. been the case with the vast ice sheets which comparatively recently covered this region.

As the ice proceeds the débris under the thicker central portions will be pushed along, producing scratches and grooves upon the underlying rocks, but toward the edge, where the ice sheet is thinner because of rapid melting above and below, the débris below, consisting of bowlders and clay, will cease to be moved, and, like the sediment of a river when the current slackens, will build up a bar. Moreover, the material carried in the lower part of the ice sheet will gradually be dropped and become more or less assorted by the waters which result from the melting of the ice. The resulting deposits will be of fluvial character where drainage is free, and lacustrine where escape is hindered. In the latter case especially portions of the ice mass will become detached and buried in the débris and their melting will be delayed until after the surrounding material has become permanently settled, so that eventually pond holes, and even lake beds of considerable size, will result.

From this sketch of the glacial action one can readily understand how the first advance of the ice is preceded by a flooded condition of all streams. In case they flow from the ice, their waters will be swollen by the melting ice and their sediment increased by the mud and sand contributed by the glacier. In case their course is toward the ice sheet or parallel with the edge, they will become dammed and changed temporarily into lakes in which the material from the ice will be deposited.

From this explanation also it may be seen that the pre-Glacial surface will be more or less covered with a layer of sand or gravel before the bowlder clay or regular glacial till is deposited upon it.

It is also seen how the pre-Glacial surface divides the course of the ice sheet and decides the character, in a measure, of the deposits laid down by it. The axis of the ice flow will coincide with the central portion of the pre-Glacial valley, and over it the ice will be thickest and its surface most elevated. The divides between the valleys and other elevated points will be the last to be covered by the ice and the first to be uncovered. From this fact it might be expected that the glacial deposits are thinner there, but such is not often the case, because of the tendency for the deposits to accumulate more rapidly under the thinner portions of the ice, as before explained. These facts are well illustrated in this area.

TILL OR BOWLDER ('LAY.

This deposit is the accumulation formed underneath the ice, as has already been explained. It is generally of unstratified character, being a heterogeneous mixture of sands, clay, gravel, and occasional bowlders. It is usually distinguished from ordinary stream deposits

by its unassorted character and the absence of stratification. It frequently contains pockets or lenses of sand of considerable extent, which is often stratified. In some cases these are portions of the accumulations of subglacial streams, which have become broken up and buried in the till. They sometimes contain abundant water, but are often dry and filled with air, which is sometimes under pressure. The ingredients of the till vary in different localities according to the rocks over which the glacier has passed. In this area the clay comprises perhaps 90 per cent of the whole.

Surface of the till.--The surface of the till in lower places is more or less covered with silt, laid down by the waters escaping from the ice soon after its deposition, and also by the wash in more recent times, especially upon hillsides. In some localities, also, considerable silt has been deposited from the atmosphere; but both of these influences have not modified the till of this region to any great extent, except in the lower portions bordering James River, of which mention will be made in a later section. The surface of the till in general may be said to be much more even in this area than is frequently the case in other regions. This is mainly the result of its having been deposited far within the principal moraines upon a surface already smoothed by former and more vigorous ice sheets.

The till of this region exhibits the usual basin and swell surface, but in a very mild degree. Few of the basins are of sufficient extent to hold water through the summer time. As a whole, however, they have a most important influence in retaining the surface water. When snow melts, or after rains, they hold the water so that over much of the surface very little of the rainfall escapes into the streams. An area around Huron, covering perhaps 100 square miles or more, seems to have been, during the recession of the ice, a shallow lake. So, also, at an earlier stage there was an area of similar character and equal extent east of Woonsocket, and also another northeast of Mitchell. All of these will be discussed more in detail under the heading "Ancient channels."

Thickness of the till.-The average thickness of the till can be roughly estimated to be about 75 feet over the whole area. Southwest, west, and south of Mitchell, within a radius of 10 or 12 miles, the thickness is often less than 50 feet. This includes the region in which the Cretaceous strata outcrop. So, also, in the valleys of all the principal streams the thickness of the till is often less than 20 feet. On the contrary, in the elevated region about De Smet depths of 200 or 250 feet may occur, although but few drillings have been made to show this fact. Southward from that region the till gradually diminishes in thickness, until in the vicinity of Spencer it is generally less than 50 feet. In the region south of Plankinton, also, wells show a frequent thickness of 150 feet, while some attain over 200 feet.

In the flat lands east of Woonsocket the depth of the till is nearly