ther, it is the Boulder-clay connected with this S. W. striation that has hitherto proved most rich in marine shells.

If, however, we pass from the St. Lawrence Valley up the valleys which open into it from the North, as for example the gorge of the Saguenay, the Murray Bay River, or the Ottawa River, we at once find a striation nearly at right angles to the former, or pointing to the South-east.

At the mouth of the Saguenay, near Moulin Bode, are striae and grooves on a magnificent scale, some of the latter being ten feet wide and four feet deep, cut into hard gneiss. Their course is N. 10° W. to N. 20° W. magnetic, or N. 30° to 40° W. when referred to the true meridian. In the same region, on hills 300 feet high, are roches moutonnees with their smoothest faces pointing in the same direction, or to the North-west. This direction is that of the valley or gorge of the Saguenay, which enters nearly at right angles the valley of the St. Lawrence. At the month of the Saguenay the Lark Shoals constitute a mass of debris and boulders, both inside and outside of which is very deep water; and many of the fragments of stone on these shoals must have been carried down the Saguenay more than fifty miles.

In like manner at Murray Bay there are striae on the Silurian limestones near Point au Pique, which run about N. 45° W. but these are crossed by another set having a course S. 30° W., so that we have here two sets of markings, the one pointing upwards along the deep valley of Murray Bay River to the Laurentide Hills inland, the other following the general trend of the St. Lawrence valley. The Boulder-clay which rests on these striated surfaces is a dark-coloured Till, full of Laurentian boulders, and holding Leda truncata, and also Bryozoa clinging to some of the boulders. In ascending the Murray Bay River, we find these boulder-beds surmounted by very thick stratified clays, with marine shells, which extend upward to an elevation of about 800 feet, when they give place to loose boulders and unstratified drift. About this elevation, the laminated clays meet a ridge of drift like a moraine, crossing the valley, which forms the barrier of a small lake, Petite Lac, and a second similar barrier separates this from Grand Lac. If the valley of Murray Bay River was occupied with a glacier descending from the Laurentian hills inland, which are probably here 3000 to 4000 feet high, this glacier or large detached masses pushed from its foot, must have at one time extended quite to the border of the St. Lawrence, and at

another must have terminated at the borders of the two lakes above mentioned.

On a still larger scale the N. W. and S. E. striation appears in the valley of the Ottawa, and farther west between the head of Lake Ontario and Lake Huron. In these places there is no elevation capable of giving rise to local glaciers, and therefore, as in New England and Nova Scotia, we must ascribe the glaciation either to general ice-laden currents from the North-west, or to the great continental glacier imagined by some geologists.

A most important observation bearing on this subject appears in the Report of Mr. R. Bell, in the region of Lake Nipigon, North of Lake Superior. He observed there the prevailing South-west striation, but with a more westerly trend than usual. Crossing this, however, there was a southerly and S. E. set of striae which were observed to be older than the South-west striae. In some other parts of Canada these striae seem to be newer than the others, but there would be nothing improbable in their occuring both at the beginning and end of the Boulder-clay period.

In summing up this subject, I think it may be affirmed that when the striation and transfer of materials have obviously been from N.E. to S.W., in the direction of the Arctic current, and more especially when marine remains occur in the drift, we may infer that floating ice and marine currents have been the efficient agents. Where the striation has a local character, depending upon existing mountains and valleys, we may on the other hand infer the action of land ice. For many minor effects of striation, and of heaping up of moraine-like ridges, we may refer to the presence of lake or coast ice as the land was rising or subsiding. This we now see producing such effects, and I think it has not been sufficiently taken into the account.

As to the St. Lawrence valley, it is evident that its condition during the deposit of the Boulder-clay must have been that of a part of a wide sound or inland sea extending across the continent, and that local glaciers may have descended into it from the high lands on the north and possibly also on the south. During this state of the valley great quantities of boulders were brought down into it, especially from the Laurentide hills, and were drifted along the valley, principally to the south-west. Extensive erosion. also took place by the combined action of frost, rain, melting snows, and the arctic current and the waves, and thus was furnished the finer material of the Boulder-clay.

It is further to be observed that oscillations of land must be taken into account in explaining these phenomena. Elevations increasing the height and area of land might increase the space occupied by snow and land ice. Depressions, on the other hand, would bring larger areas under the influence of water-borne ice and marine deposits, and these might take place either in a shallow sea loaded with field and coast ice, or in deeper water in which large icebergs might float or ground. There is reason to believe that such alternations were not infrequent in the Postpliocene, and that their occurrence will explain many of the complexities of these deposits.

If we adopt the iceberg hypothesis, we must be prepared to consider in connection with this subject a subsidence so great as to place the Laurentides and all but the highest summits of the Appalachians under water. In this case a vast volume of Arctic ice and water would pour over the country of the great lakes to the S.W., while any obstruction occurring to the south would. throw lateral currents over the Appalachians to the eastward. If we adopt the glacier hypothesis, we may on the other hand. imagine a great movement of land ice to the S.W., westward of the Appalachians, and a separate outward movement eastward from these hills and down the Atlantic slope of America. On either hypothesis there are difficulties in accounting for some sets of striae, but on that last-mentioned I believe them to be insuperable.

It is evident from the descriptions of Smith, Geikie, Jameson, Crosskey, and others, that the Boulder-clay of Scotland and Scandinavia corresponds precisely in character with that of Canada, and there, as in America, the theory of a continental glacier has been resorted to for its explanation. The objections to this hpyothesis are very ably stated by Mr. Milne Home in a paper on the "Boulder-clay of Europe," in the Transactions of the Royal Society of Edinburgh, 1869.

To this period and these causes must also be assigned the excavation of the basins of the great American lakes. These have been cut out of the softer members of the Silurian and Devonian Formations; but the mode of this excavation has been regarded as very mysterious; and like other mysteries has been referred to glaciers. Its real cause was obviously the flowing of cold currents over the American land during its submergence. The lake basins are thus of the same nature with the deep hollows

intervening between the banks cast up by the Arctic currents on the present American coast, and like those deep channels of the Arctic current in the Atlantic recently explored by Dr. Carpenter. Their arrangement geographically as well as their geological relations, correspond with this view.

Another consideration with regard to the great lakes deserves notice. Dr. Newberry has collected many facts to show that the lake basins are connected with one another and with the sea by deep channels now filled up with drift deposits. It is therefore possible that much of the erosion of these basins may have occurred before the advent of the glacial period, in the Pliocene age, when the American continent was at a higher level than at present. Dr. Newberry has given in the Report in the Geology of Ohio a large collection of facts ascertained by boring or otherwise, which go far to show that were the old channels cleared of drift and the continent slightly elevated, the great lakes would be drained into each other and into the ocean by the valleys of the Hudson and the Mississippi, without any rock cutting, and if the barrier of the Thousand Islands were then somewhat higher, the St. Lawrence valley might have been cut off from the basin of the great lakes.

I shall close the discussion of this subject by quoting from one of the papers above referred to, my views in 1864; reserving, however, some points respecting the present action of floating ice, to which I shall refer in the sequel.

"Our American lake-basins are cut out deeply in the softer strata. Running water on the land would not have done this, for it could have no outlet; nor could this result be effected by breakers. Glaciers could not have effected it; for even if the climatal conditions for these were admitted, there is no height of land to give them momentum. But if we suppose the land submerged so that the Arctic current, flowing from the northeast, should pour over the Laurentian rocks on the north side of Lake Superior and Lake Huron, it would necessarily cut out of the softer Silurian strata just such basins, drifting their materials to the southwest. At the same time, the lower strata of the current would be powerfully determined through the strait between the Adirondac and Laurentide hills, and, flowing over the ridge of hard rock which connects them at the Thousand Islands, would cut out the long basin of Lake Ontario, heaping up at the same time in the lee of the Laurentian ridge, the great mass of boulder

clay which intervenes between Lake Ontario and Georgian Bay. Lake Erie may have been cut by the flow of the upper layers of water over the Middle Silurian escarpment; and Lake Michigan, though less closely connected with the direction of the current, is, like the others, due to the action of a continuous eroding force on rocks of unequal hardness."

"The predominant southwest striation, and the cutting of the upper lakes, demand an outlet to the west for the Arctic current. But both during depression and elevation of the land, there must have been a time when this outlet was obstructed, and when the lower levels of New York, New England, and Canada were still under water. Then the valley of the Ottawa, that of the Mohawk, and the low country between Lakes Ontario and Huron, and the valleys of Lake Champlain and the Connecticut, would be straits. or arms of the sea, and the current, obstructed in its direct flow, would set principally along these, and act on the rocks in north and south and northwest and southeast directions. To this portion of the process I would attribute the northwest and southeast striation. It is true that this view does not account for the southeast striæ observed on some high peaks in New England; but it must be observed that even at the time of greatest depres sion, the Arctic current would cling to the northern land, or be thrown so rapidly to the west that its direct action might not reach such summits."

"Nor would I exclude altogether the action of glaciers in eastern America, though I must dissent from any view which would assign to them the principal agency in our glacial phenomena. under a condition of the continent in which only its higher peaks were above the water, the air would be so moist, and the temperature so low, that permanent ice may have clung about mountains in the temperate latitudes. The striation itself shows that there must have been extensive glaciers as now in the extreme Arctic regions. Yet I think that most of the alleged instances must be founded on error, and that old sea-beaches have been mistaken for moraines. Even in the White Mountains the action of the ocean-breakers is more manifest than that of ice almost to their summits; and though I have observed in Canada and Nova. Scotia many old sea-beaches, gravel-ridges, and lake-margins, I have seen nothing that could fairly be regarded as the work of glaciers. The so-called moraines, in so far as my observation extends, are more probably shingle beaches and bars, old coast

VOL. VI.

с

No. 1.