Origin. These ancient channels were developed during the presence of the glacier, and served to carry off the water from the front of the ice sheet in its different stages. The arrangement of the channels is evidence of the former existence of an ice sheet over this region. The size, and particularly the course, of some of the channels and the amount of coarse material found in them can not well be explained in any other way.

The order in which these channels were occupied may be learned from the map, where they are numbered; but it should be remembered that it is impossible to represent the order of their occupation with minute accuracy. In some cases the same channel has been used in opposite directions, and this change has perhaps occurred more than once. It will be noticed that the numbers are generally wide apart, but are sometimes found near each other, especially along the course of James River. In such cases it is intended to indicate that the channel had been used at times corresponding to the occupation of channels bearing those numbers and that the altitude of the formation does not clearly distinguish the order of time. Usually the later terraces are lower than the older, but this is not always the case. During the rapid unloading of débris from the ice the streams might at one time erode and at another time fill, so that waters are conceived to have been at nearly the same level at widely different times.

To give an example of the way in which the channels record the presence and different stages of the ice sheet and at the same time the varying directions in which the water flowed, attention is called to the channels west and south of Mitchell. (See Pl. III.)

When the ice occupied and was building the outer member of the Gary moraine, the drainage about the western edge was discharged along a channel which follows the course of Firesteel Creek and a southern tributary, and which leads west of Mount Vernon; then through the main branch of Enemy Creek to the center of T. 102 N., R. 61 W., when it passed by another intermediate channel into the north branch of Twelvemile Creek, thence by another crosscut to the south branch of the same stream. Into this the water from the ice sheet discharged by several outlets through the moraine. Since the course of this stream was on a higher level than the land farther east a short distance within the edge of the ice sheet, as soon as the ice began to recede the waters at several points, particularly near the north line of Aurora County, turned eastward through the outlets and followed for a time the edge of the ice, while it formed the second member in a way analogous to that already sketched. When the ice receded to the third member a similar change took place by the main stream, following the present course of Firesteel Creek to a point northeast of Mount Vernon before turning south. The course of Firesteel Creek was not occupied as at present until the ice had withdrawn beyond. the junction of this stream with James River.

ANCIENT LAKES.

In connection with the channels, certain extensive areas which may be called ancient lakes should be considered. The use of this term does not imply that they were occupied at one time by water, although this may have been the case. It is probable that as the ice receded over them from the south toward the north the southern portion was first occupied by more or less ponded water and filled by the accumulating sediment from the streams draining the adjacent ice sheet; as the ice uncovered one portion after another these areas were successively filled with sediment in the same way, until, on the recession of the ice, they became flat plains, covered with sand or clay, with points of the underlying till rising above it like islands, and shallow channels winding about irregularly upon it. In some cases shallow lakes may have continued for an indefinite period. The largest and best example of this sort lies between Woonsocket and Forestburg, and extends from the outer portion of the third moraine to the fourth member of the second moraine. It covers an area of perhaps 50 miles. Over this area the surface is poorly drained and covered either with fine clay, commonly called gumbo, or with sand which is sometimes gathered in dunes by the wind. The gumbo seems to have been formed by the settling of clay in comparatively quiet water. The sandy areas from their position overlying the gumbo and till seem to have been formed as deltas by Sand Creek and Redstone Creek.

A similar area, not so clearly defined, surrounds Huron, especially on the south and east. This does not present so much of the gumbo surface, and the sand is also less marked.

Between Letcher and Mitchell and extending eastward across James River is another flat area which presents more of the usual surface of the till and less of the sand and gumbo.

GEOLOGIC HISTORY.

The granite which underlies most of James River Valley is considered to be part of the very old surface which existed before life appeared upon the earth. Whether it may be, in some cases, part of the original substance of the globe or whether it was first laid down in an ocean and afterwards changed to a crystalline rock by heat and pressure has not been satisfactorily ascertained.

The Sioux quartzite, which is exposed in the region east of Mitchell, belongs to a later epoch, and certainly was deposited in the ocean, probably upon the margin of a land area which existed farther east in Minnesota and Wisconsin. From that land, the waves, ocean currents, and tides, which were probably more vigorous then than now, derived the sands and spread them out over the bottom, much as now takes place on the seashore. As the material accumulated there was a gradual sinking of the ocean bottom till a thickness of several hun