from the village. They are believed to agree in altitude with those represented on Plate XXXVII, estimated to be respectively 10, 20, 40, 60 and 80 feet high. There has been no attempt as yet to show the causes acting for the formation of each terrace. They were probably connected with variable ice barriers. None of these smaller terraces are to be found outside of the glacial lake area.

THE CONNECTICUT ESKER.

Our quadrangle contains the greater part of the famous esker extending from Thetford to Windsor and first understood by Dr. Warren Upham and described in the New Hampshire Report under the name of "kame." The present usage of geologists in regard to the terminology of these two classes of deposits was not fully understood in 1878. Dr. Upham followed the leading of Dr. James Geikie. As this ridge has been so thoroughly described it is not necessary to repeat what has been written already, but to refer to the map where the course of the ridge has been portrayed.

Some suggestions have come as to the relations of this esker to the till. No section has been discovered as yet where the till and the gravel meet each other. Excavations in Hanover look as if the esker were underlaid by an early clay, which has been crumpled by some unknown force, perhaps the motion of the ice. This clay appears in Girl Brook, at the cut for a sewer near the Mary Hitchcock Hospital, near the Ledyard bridge and between the houses of Professors Patten and Dow.

There seems to have been a system of drainage in the village of Hanover prior to the formation of the esker, interrupted by the deposition of the gravel. Thus an arborescent system of valleys converge at the cut by the Ledyard bridge. But their outlet was blocked by the incursion of the esker; for it would appear that the esker was once continuous from brink to brink above the present road which has been used only a little more than a century. The former crossing of the Connecticut was a mile higher up. The Honorable Daniel Blaisdell informed me that he saw a pine tree bridging this chasm when it had been cut down, the stump being on one side and the top on the other. If so, the cut could not have been very deep. This ancient system of drainage raises important questions which cannot be answered.

Plate XXXI shows the appearance of this esker as it has been excavated of late for road material near White River Junction.

The following is a list of the stones found in it:

Red and drab porphyry from the White Mountains.

Granites allied to those of Barre and other localities southeast from Montpelier.

Granite or protogene from Fairlee.

Diabases and diorites from the Chloritic Group.

Jasper from the same group.

Amphibolites from eastern Vermont.

Limestone from the Conway schist.

Coös mica schists and quartzites.

White quartz veinstone.

Augitic rock.

Blue Mountain granite.

Cambrian quartzite from the Green Mountains.

Hornblende schists of variable composition.

Epidotic hornblende.

Decayed limestone.

Decayed rocks stained by manganese.

Chloritic schists and sandstones.

Argillites.

It is an important omission that none of the protogene granite of the Hanover-Lebanon area has been found in the esker. This shows that nothing came into it from the east. The absence of serpentine is more likely to be accidental.

THE HIGHER FLOOD PLAIN.

Various expressions are appropriate for the extremely high mass of sand and clay brought down at the time of the melting of the ice. It may be the higher flood plain or the highest normal terrace, and both may be modified by the extra high accumulations brought down by tributaries where for special reasons the amount of water discharged by them was excessive.

We cannot say that the highest tributary delta represents the highest normal terrace. Thus there is thirty feet difference between the delta of Mink Brook where it touches the edge of the flood

plain on the Lebanon road and the plain in the village. If the water of the Connecticut ever reached the higher level, it must have been only for a moment, as it were, and is not to be considered as the true height.

There is not time to consider a sand plain which appeared at a higher level than the highest terrace in Lebanon and Hartford. It reaches 560 feet east of Wilder, and on the hill west; also at the fair grounds and at the turn of the road southwest toward Farnum Hill, about a mile east of the electric power house of West Lebanon. These are somewhat connected with the dunes of Plainfield. It is worthy of note that the altitude of the high sand is the same as that of the delta of Mink Brook. The material of the flood plain in the west part of Lebanon and Hanover and across the river is a loamy sand often carrying claystones. It shows occasionally thin strata of a vegetable character, as if a season of growth, or possibly a time of freshet were indicated. Plate XXXII was taken to illustrate the presence of two of the layers. The locality is in the northwest part of Lebanon near the east edge of the plain, and the underlying till is also shown just at the top. Better exposures were seen later farther north, in one of which, in the Mink Brook Valley, seven of these peaty seams were counted. The phenomenon is a common one.

THE PROTECTION OF TERRACES.

It is only recently that we have arrived at a clear understanding of the reason why the lower terraces occur and why their number is variable. It is purely a local affair. The original high flood plain extended across the valley and the several lower terraces have been I carved from it. Sometimes the entire mass has been removed by erosion where the current happened to be stronger than usual.

This appears to have been the case in the lower Coös meadows at Haverhill. Following the river to Hanover, about two miles north of the college, the high plain suddenly appears. Apparently this is due to the fact that the presence of the esker protected the silt from being worn away when the Connecticut changed its course and turned northwesterly across the esker. This ridge of gravel has maintained in its place the whole of the flood plain as far as Wilder: and it may be remarked incidentally that the original course of the river must have been beneath this protected plain, because the present stream runs over ledges at Wilder.