upon a first visit to southeastern Dakota is the extent of the apparently level area where the till comes to the surface. This impression is heightened probably by the absence of forests, and would very likely be the same in portions of Ohio and Indiana were it not for the timber. James river valley in Dakota is depressed in the centre about five hundred feet below the edge, but it is, roughly speaking, seventy miles across; so that the slope does not strike the eye. So level is the country that every special line of glacial accumulation is a prominent feature in the landscape, and the various halting-places of the ice in its retreat are readily discerned, and were accurately delineated, so far as my observations have extended, by Professor Todd in the paper referred to. For a long time a lobe of ice filled the James river valley, running parallel with that occupying the upper Minnesota valley, and extending southward into Iowa. The edges of these lobes thinned out along the north and south line which runs near the east margin of southern Dakota, and favored the accumulation of the lines of hills known as the "Coteau des Prairies." Professor Todd and others speak of this as a series of "terminal moraines" formed along the sides of the reëntrant angle whose apex penetrated to the vicinity of the Sisseton Agency. I am not sure but that it would facilitate an understanding of the subject to speak of the Coteau des Prairies as a medial moraine toward which the glacial debris (carried upon the deeper portions of the ice) gravitated in both directions. But certain it is, that, starting from a point at the Sisseton Agency different lines of glacial accumulations stretch southward at varying angles,-the later accumulations forming the more obtuse angle. Coming up the valley of the James from Yankton one crosses the oldest of these accumulations (the Altamont moraine) in the neighborhood of the city itself, and the second (or Gary moraine) in the neighborhood of Mitchell, sixty miles to the north, having run parallel with it, however, for about thirty miles. The third, or Antelope moraine, is encountered near Huron, about sixty miles north of Mitchell, and continues visible upon either side of the river, about twenty miles distant, beyond Aberdeen.

On the western side of this lobe there are corresponding lines of accumulation, the outer of which is the vicinity of the Missouri river and on its eastern side. Everywhere in coming up from the Missouri on to the plateau. which is in most places from four hundred to five hundred feet above the river, one encounters two or three terraces covered with boulders, the two highest of which are about three hundred and four hundred feet respectively. These moraines rise out from this general level one hundred or two hundred feet higher, and are marked features of the landscape. The streams entering the Missouri from the east are all of them short, the longest not being more than forty miles in length. These streams are in all cases that I observed marked by broad and elevated local terraces, the edges of which where they overlook the immediate trough of the stream are crowded with granitic boulders. In some cases, as Professor Todd has shown, these valleys terminate abruptly at the water-parting, as if being the continuation of glacial streams from the east which had originated upon the ice lobe while it filled the James valley. The boulders 14

A. A. A. S. VOL. XXXVII.

which are superabundant over all this region are largely of granite or gneiss, and must have been transported over a long distance, in some cases probably more than four hundred miles, from the northeast.

Owing to the difficulty of penetrating the Indian reservation, Professor Todd had not been able to extend his observations to any great extent west of the river and north of Pierre. But through the courtesy of Rev. Thomas L. Riggs, and of his assistant, Rev. James F. Cross, missionaries to the Sioux, I was permitted to enter the reservation at Fort Yates, and traverse with them, on the west side of the river, the region intervening 'between Fort Yates and Oahe, in the vicinity of Pierre, making the distance travelled over about 150 miles. Leaving Fort Yates, and going four miles in a south-southwesterly direction, we reached an elevation of 425 feet, all of which was made during the last mile, and represents the general height of bluffs on either side of the trough of the Missouri, which is here about four miles wide. The elevation of Fort Yates above tide is 1694 feet, making the elevation of the bluff 2119. The fort is situated on a glacial terrace about fifty feet above the present high-water mark. Eight miles farther on, the elevation was 2294 feet, which is about the general level for many miles. The elevation of Grand River was 1819 feet. The divide between the Grand river and the Moreau is almost exactly the same elevation (namely, 2294 feet) with that between the Grand and Fort Yates. South of the Moreau river the divide between that and the Cheyenne rises to 2404 feet. Over the whole of this distance from Fort Yates, boulders of granite and gneiss were very abundant, being as numerous upon the higher as upon the lower points. Frequently the boulders were several feet in diameter. But, two or three miles to the south of the highest land between the Moreau and the Cheyenne rivers this northern drift suddenly ceased at an elevation of 2354 feet, or fifty feet lower than the previous reading. From this point south and east to Oahe (a distance of about sixty miles) no signs of northern drift were seen, except in the valley of the Cheyenne river, upon coming down to a level of 2100 feet. The elevation of the higher land between the Cheyenne and the Bad river was from 2200 to 2300 feet. Upon coming again into the valley of the Missouri, northern drift began to appear immediately upon the western bluff, at about 2100 feet above tide.

The western limit of drift upon the Northern Pacific railroad is fixed by Professor Chamberlin at Sim's Station, about forty miles west of Bismarck. Prof. N. H. Winchell, in his report upon the Custer expedition to the Black Hills, in the summer of 1874 (see pp. 22 and 23) found the limit of drift to be in the Dog Teeth Buttes, thirty or forty miles south of Sim's Station, and about twenty miles west of the 101st meridian. Putting these observations with my own, it would appear that the limit of

1As I had no barometer with me, the elevations given were determined by Rev. Mr. Riggs upon a subsequent journey, and forwarded to me. Being the result of single readings, they are of course liable to a considerable margin of error, which, however, may be eliminated in subsequent journeys. But I am confident that the margin of error is not large, since the readings are nearly correct at the river crossings.

glacial drift between the Northern Pacific railroad and Pierre runs from Sim's Station southward, crossing in a pretty direct line the valleys of the Cannonball, the Grand, and the Moreau river, to the divide between the Moreau and the Cheyenne, where it is certainly ten or twelve miles west of the 101st meridian. At this point it is running nearly east and west between the Moreau and the Cheyenne; but just where it strikes the trough of the Missouri I am at present unable to tell. It certainly cannot be much west of the mouth of the Cheyenne.

Over the glaciated portions traversed, I was unable to find any typical deposits of till, but this may have been owing to the absence of favorable places for observation, since the rolling prairie region afforded few opportunities for erosion. Frequently, however, knolls and short ridges projected above the general level, and were literally covered with boulders, pebbles, and gravel, all of which appeared to be more or less washed. Nor was I able to find on any of the boulders distinct marks of glacial scratching. About halfway between the Grand and Moreau rivers, upon the very summit, we traversed an extensive shallow depression resembling a kettlehole about two miles wide. This was bordered by hills about one hundred feet above the bottom. The rim was marked by numerous knobs which were perfectly covered with granitic and gneissoid boulders. Many large boulders were also scattered over the bottom, these being frequently from three to four feet in diameter. The boulders seem to be limited to a su perficial deposit. On approaching the Moreau, while still upon the high land, numerous boulders from four to eight feet in diameter were observed. On striking into the valley of the Cheyenne river, about twelve miles above its mouth, and coming down upon a terrace, about three hundred feet below the general level, a few small gneissoid boulders appeared. Again on crossing the river at about the same elevation, we struck into an old river-bed, which seemed to be perfectly level, and about two miles wide, and to extend as far as the eye could reach both up and down the valley. Granitic pebbles were abundant in this deposit and they were all well waterworn, and evidently occupied a depth of a number of feet. Here and there over it were scattered a number of small granitic or gneissoid boulders a foot or more in diameter. Whether the source of this granitic material was the glaciated region to the north or not, I am at present unable to tell, since the Cheyenne has access to granite in the Black Hills where it rises. Still, from the similarity of the boulders to those found on the highlands to the north, I think a connection will be found between this old river-bed and the glaciated region to the north. I am inclined to believe that this is the line of marginal drainage by which the water (which both before and since occupied the trough of the Missouri) worked around in front of the ice, meeting the Missouri valley near the mouth of the Cheyenne river. Since my return, Mr. Riggs has crossed the Cheyenne about six miles farther up stream, and reports this old river-bed as there on the north side, and coming in from the north at an angle of about twenty degrees. I shall look for further evidence of this stream having worked around Fox Ridge from the glaciated region where we left it on the divide between the Moreau and the Cheyenne rivers.

About ten miles south of Grand river, however, we found Long Butte to be not like the buttes which had occurred at frequent intervals in previous parts of the journey, capped with sandstone, but to be a gravel deposit evidently an old river bottom containing much material brought down from the Black Hills, and all well waterworn. This ridge extended for some miles, being a very marked feature in the landscape, and is, I suppose, a remnant left by the erosive forces which had operated for a long time upon the underlying cretaceous formations, wearing them away faster than they did the superincumbent gravel deposit.

The further study of this portion of the glacial border, will be important for its bearing upon current theories concerning the glacial period. The thinness of the deposit, and the washed appearance of the material of which it is composed, will be interpreted by some to indicate that it is the eroded and wasted margin of a first glacial period, separated by an enormous lapse of time from the last glacial period. This, however, does not seem to me to be necessarily the case. In interpreting the marginal deposits of the glacial period, we should remember that the motion of a glacier constantly diminishes towards the margin until it finally reaches a line of stability; so that the deposits at the very margin are those which have been carried upon the ice, rather than shoved along under it, which would account for comparative absence of abrasion. Furthermore, the marginal deposits are, of course, the older deposits on any theory, and have been longest subjected to the action of erosive and disintegrating agencies. Thirdly, in the present case, these marginal deposits were peculiarly subject to erosive action from the vast drainage of the Missouri river which had been pushed out of its natural channel. These considerations seem to me still to leave it an open question, so far as these deposits are concerned, whether they do not belong to an earlier stage of the same epoch during which the heavier deposits had been made upon the east side of the river. At any rate, it would seem that the abruptness of the termination of these deposits, shows that they were due to direct glacial action, rather than to that of floating ice. It would be difficult to conceive of a subsidence in that region which should have produced a body of water with a shore corresponding to this drift limit. Now that the country is likely soon to be opened to settlement, we shall look to investigations in that quarter for much light upon many problems of glacial theory.

SOME THOUGHTS ON ERUPTIVE ROCKS WITH SPECIAL REFERENCE TO THOSE OF MINNESOTA. By Prof. N. H. WINCHELL, State Geologist of Minnesota, Minneapolis, Minn.

GEOLOGISTs know of but two sources for all the rocks which constitute the crust of the earth, viz.: (1) cooling from fusion, and (2) solidifying after being disintegrated and distributed by water. Under accepted hypothesis the former were first to be formed, and the latter have resulted by a long continued series of dissolution and selection through the agency of oceanic water, from the former.

Those that have cooled from fusion, known as igneous, or eruptive, rocks have been divided broadly into basic and acid eruptives. Though it is obvious that this distinction may fade out when subjected to universal and rigorous application, yet for the purpose of this paper it may be accepted in advance, since these terms do afford a classification which covers a multitude of facts and sufficiently characterize the two evident great divisions of eruptives.

In offering to the Section a few thoughts on these two classes of eruptives I wish to disclaim any desire to ignore the views and facts that have been published by others. I shall simply base some statements on some facts which have come under my own observation, to which I have given considerable reflection during the past ten years, derived from my work on the geological survey of Minnesota, referring to my reports for details of the evidence. Time has not yet been given to make an exhaustive study of similar facts from other parts of the world, and the tentative hypotheses that I shall give may have to be modified or abandoned on making wider comparisons.

The basic eruptives are sometimes distinguished as volcanic, in distinction from the acid which at the same time have have been called plutonic. They are such as are known commonly to have flowed in recent times from volcanic vents, and to have been spread in extensive sheets and streams over the face of the earth. They are heavy, dark-colored, often bearing iron-peroxide (and in rare cases metallic iron): as a group they are dolerytes,' and embrace diabase, basalt, some of the diorytes, and those rocks that in Wisconsin and Minnesota have been designated gabbro. They contain less than sixty per cent of silica.

The acid eruptives (so called) are such as have been supposed to have constituted the earliest super-crust of the earth, and are now exposed at the surface because they have been kept uncovered by denudation. They occupy, where seen on the continental areas, the lowest of the rock strata; and from these ancient bosses are seen various offshoots that penetrate in forms of velns, sheets and dikes, the rocks immediately overlying. In the higher rock horizons this class of eruptives is comparatively rare, but isolated small areas are known in Europe and America. In most instances, however, in which they are found in Silurian, Carboniferous or Tertiary time, they seem to exhibit a diversity of minor characters of structure, by which they are distinguishable, though their chemical composition, in its essential ingredient of silica, shows their general alliance to this class of eruptives. The acid crystalline rocks embrace granite, syenite, apparently some diorytes, felsyte, trachyte, quartz-porphyry and all such as have a silica content of sixty per cent or more.

In addition to the distinction of content of silica these two classes exhibit other differences, when compared in their original and unweathered state, and when not modified by immediate contact with other rocks. The basic rocks are dark colored, gray, varying to hornblendic-brown or chloritic-green. The acidic as a class are of lighter colors, as required by the 1 Dana: Am. Jour. Sci., Nov., 1878.