My interpretation of the phenomena would differ from that of Dr. Newberry in the following particulars-(1) I would refer the continental elevation and the deep erosion to the Pliocene period, before the advent of the glacial epoch. (2) I would refer the glaciated surfaces and the lower part of the Erie clay to the time of the Canadian Boulder-clay, and would regard it as an evidence of subsidence and an ice-laden sea, with the arctic current passing over the continent from the North-East. (3) I would regard the upper part of the Erie clay as equivalent to the Leda clay. (4) I would place the upper and confessedly water-borne drift as the equivalent of the Saxicava sand, and as belonging to the period of elevation.

It is a difficulty, both in Dr. Newberry's view and mine, that marine shells are not found in the Erie clay and surface drift. The following considerations, however, diminish this. (1) The greater part of the Leda clay is very poor in fossils, even near the ocean, and so is the boulder clay. (2) The submergence of a vast continental area under cold water might have continued for a long time before the marine animals could widely spread themselves over it, especially under the unfavourable circumstances of ice erosion. (3) The few and scattered marine remains to be expected in these deposits may have escaped observation. The occurrence of much drift-wood in the Erie clay is also, in my judgment, inconsistent with the occurrence of a general glacier immediately previous to the deposition of the clay.

We may now consider the several members of the Post-pliocene in succession, beginning with the oldest.

Throughout a great part of Canada there is a true "Till," consisting of hard gray clay, filled with stones and thickly packed with boulders. In some places, however, the clay becomes sandy, and in some portions of the carboniferous areas, the paste is an incoherent sand. The mass is usually destitute of any stratification. or subordinate lamination; but sometimes in thick beds horizontal lines of different texture or colour can be perceived, and occasionally the clay intervening between the stones becomes laminated, or at least shows such a structure when disintegrated by frost. The Boulder-clay usually rests directly on striated rock surfaces;

but I have observed in Cape Breton a peaty or brown coal deposit, with branches of coniferous trees, to underlie it, and in other places there are deposits of rolled gravel under the Boulderclay. At the Glen brick-work, near Montreal, a peculiar modified Boulder-clay occurs, consisting of very irregularly bedded sand and gravel, with many large boulders, and only thin layers of clay.

The stones of the Boulder-clay are often scratched and ground into those peculiar wedge-shapes, so characteristic of ice-worked stones. Very abundant examples of this occur in the Boulder clay of Montreal and its vicinity.

At Isle Verte, Riviere du Loup, Murray Bay, Quebec, and St. Nicholas, on the St. Lawrence, the Boulder-clay is fossiliferous, containing especially Leda truncata, and often having boulders and large stones covered with Balanus Hameri and with Bryozoa, evidencing that they have for some time quietly reposed in the sea bottom before being buried in the clay. This is indeed the usual condition of the Boulder-clay in the lower part of the St. Lawrence River. Further up, in the vicinity of Montreal, it has not been observed to contain fossils, but it presents equally unequivocal evidence of sub-aqueous origin in the low state of oxidation of the iron in the blue clay, which becomes brown when exposed to the weather, and in the brightness of the iron pyrites contained in some of the glaciated stones, as well as in the presence of rounded and glaciated lumps of Utica shale and other soft rocks, which become disintegrated at once when exposed to weathering.

The true Boulder-clay is in all ordinary cases the oldest member of the Post-pliocene deposits, and it is not possible to ascertain the existence of Boulder-clays of different ages, superimposed on one another. It may be observed, however, that in so far as the Boulder-clay is a marine deposit, that which occurs at lower levels is in all probability newer than that which occurs at higher levels. It is also to be observed that boulders with layers of stones occasionally occur in the Leda clay; and that the superficial sands and gravels sometimes contain large boulders; but these appearances are not, I think, sufficiently important to induce any experienced observer to mistake such overlying deposits for the true Boulder clay.

In some localities the stones in the Boulder-clay are almost exclusively those of the neighbouring rock formations, and this is

especially the case at the base of cliffs or prominent outcrops, whence a large quantity of material would be easily derived. In other cases material travelled from a distance largely predomi nates. Throughout the valley of the Lower St. Lawrence, the gneiss and other hard metamorphic rocks of the Laurentian hills to the north-east are very abundant, and in boulders of large size and much rounded. Occasional instances also occur where boulders have been transported to the northwards; but these are comparatively rare. I have mentioned some examples of this in Acadian Geology, p. 61. Similar instances are mentioned in the Geology of Canada, page 893.

Though the boulder clay often presents a somewhat widely extended and uniform sheet, yet it may be stated to fill up all small valleys and depressions, and to be thin or absent on ridges and rising grounds. The boulders which it contains are also by no means uniformly dispersed. Where it is cut through by rivers, or denuded by the action of the sea, ridges of boulders often appear to be included in it. Those on the Ottawa referred to in the "Geology of Canada," page 895, are very good illustrations, and I have observed the same fact on the Lower St. Lawrence and on the coast of Nova Scotia. It is also observable that these lines and groups of boulders are often not of local material, but of rocks from distant localities, and that a number of the same kind seem often to have been deposited together in one group.

Loose boulders are often found upon the surface, and sometimes in great numbers. In some instances these may represent beds of boulder clay removed by denudation. In other cases they may have been derived from the overlying members of the formation, or may have been deposited on the surface, without any covering of clay or gravel. In "Acadian Geology," p. 64, I have illustrated the manner in which large stones, sometimes 8 feet or more in diameter, are moved by the coast ice and sometimes deposited on the surface of soft mud, and I have had oocasion to verify the observations of the same kind made by Admiral Bayfield, and quoted by Sir C. Lyell in the "Principles of Geology." Lastly, on certain high grounds there are large loose. boulders, which have probably been moved to their present positions by means of land ice or glaciers.

The Boulder-clay not only presents, as above stated, indications of successive beds, but it occasionally contains surfaces on which lie large boulders striated and polished on the upper surface, in

the manner of the pavements of boulders described by Miller, as occurring in the Till of Scotland. These appearances are, however, rare, and few opportunities occur for observing them.

A very general and important appearance is the polishing and striation of the underlying rocks usually to be observed under the Boulder-clay, and which is undoubtedly of the same character with that observed under Alpine glaciers. This continental striation or grooving is obviously the effect of the action of ice, and its direction marks the course in which the abrading agent travelled. This direction has been ascertained by the Canadian and United States Surveys, and by local observers, over a large part of Eastern America, and it presents some broad features well deserving attention. A valuable table of the direction of this striation is given in the Geology of Canada, which I may take as a basis for my remarks, adding to it a few local observations of my own. The table embraces one hundred and forty five observations, extending along the valleys of the St. Lawrence and the Ottawa and the borders of the great lakes. In all of these the direction is south, with an inclination to the West and East, or to state the case more precisely, there are two sets of striae, a South-west set and a South-east set. In the table eighty-four are westward of South and fifty-eight are eastward of South, three being due South. It further appears, when we mark the localities on the map, that in the valley of the St. Lawrence and the rising grounds bounding it, the prevailing course is South-west, and this is also the prevalent direction in Western New York, and behind the great Laurentide chain on the North side of Lake Huron. Crossing this striation nearly at right angles, is a second set, which occurs in the neck of land between Georgian Bay and Lake Ontario, in the valley of the Ottawa and in the hilly districts of the Eastern Townships of the Province of Quebec, where it is connected with a similar striation which is prevalent in the valleys of Lake Champlain and the Connecticut River and elsewhere in New England. In New England this striation is said to have been observed on hills 4800 feet high, as for example on Mansfield Mountain, where according to Hitchcock there are striae bearing S. 30° E. at an elevation of 4848 feet. In Nova Scotia and New Brunswick, as

See also, for the Western districts, Whittlesey's Memoir in the Smithsonian Contributions, and Newberry's Report on Ohio.

in New England, the prevailing direction is South Eastward, though there are also South-west and South striation, and a few cases where the direction is nearly East and West.

It is obvious that such striation must have resulted from the action of a solid mass or masses of ice bearing for a long time on the surface and abrading it by means of stones and sand. It is further obvious that the different sets of striation could scarcely have been produced at the same time, especially when, as is not infrequent, we have two sets nearly at right angles to each other in the same locality. Hence it becomes an important question to ascertain the relative ages of the striation and also the direction in which the abrading force moved.

Taking the valley of the St. Lawrence in the first instance, the crag-and-tail forms of the isolated hills of trap, like the Montreal mountain, with abrupt escarpments to the north-east and slopes. of debris to the south-west, the quantity of boulders carried from them far to the south-west, and the prevailing striation in the same direction, all point to a general movement of detritus up the St. Lawrence valley to the south-west. Further, in some cases the striae themselves show the direction of the abrading force. For example, in a fine exposure recently made at the Mile-end quarries, near Montreal, the polished and grooved surface of the limestone shows four sets of striae. The principal ones have the direction of S. 68° W. and S. 60° W respectively, and the second of these sets is the stronger and coarser, and sometimes obliterates the first. The two other sets are comparatively few and feeble striae, one set running nearly N. and S., and the other N.W. and S.E. These last are probably newer than the two first sets. Now with regard to the direction of the principal sets of striae, this at the locality in question was rendered very manifest by the occurrence of certain trap dykes crossing the limestone at right angles to the striae. The force, whatever it was, had impinged on these dykes from the N. E., and their S. W. side had protected the softer limestone. The locality is to the North-east of the mass of trap constituting the Montreal mountain, and the movement must have been up the St. Lawrence valley from the N.E., and toward the mountain, but at this particular place the striae point West of its mass. This, I have no hesitation in saying, is the dominant direction in the St. Law rence valley, and it certainly points to the action of the arctic current passing up the valley in a period of submergence. Fur