there are similar boulder phenomena to those observed forty miles west among the Camden hills. These mountains which tower up near the eastern border of the great fiord of the coast attain the altitude of nearly two thousand feet. They rise very abruptly from the sea, and are about a dozen or fifteen in number. From the summit of these hills a view is presented of the great archipelago of the Penobscot Bay. From Bar Harbor on the northeast, in Frenchmen's Bay, to Bass Harbor on the southwest, in Blue Hill Bay, journeying around these mountains on the south, close to their base, for the distance of twenty miles, there is one continued display of drift striæ and granitic boulders. The scratches run boldly up to the foot of these hills; and the harbors and coves trending north and south, exhibit these striæ everywhere. I have never seen in any part of Maine, not even in the granitic region of Mount Katahdin, an elevation nearly six thousand feet high, a picture of greater desolation than is presented about the south side of these mountains. Standing amid these vast ruins, it is apparent that only the irresistible grasp of a glacier could have broken them off and carried them far away toward the south. Here, as elsewhere, the granitic boulders are larger and less angular as we approach the hills from the south. Among these boulders, rocks of other formations are rarely found; and these are from the slates to the north of the hills, having been carried over these elevations, not around them. Upon the northern declivities of these hills, the boulders are not so numerous nor so large as they are on the south, but are generally much worn; and like those on the south, they are syenitic, like the formation of the hills.

But besides the coves and harbors of this great island of Mount Desert, its ledges, headlands and ponds, (some of which last are large and deep,) all trend north and south in conformity to the course of the striæ. More than one of its deeper ponds are gouged out of the solid syenite, and we find no evidence of this having been done by any other agency than that of glaciers.

Beyond these hills to the north, for fifteen or twenty miles, over a comparatively level country, the Taconic slates again appear; and the detritus of these rocks have been but sparingly transported over the granitic formation of Mount Desert. I think it a safe conclusion that the rocks torn from the hills, the valleys and the plains of the country, were not generally removed to any great distance southward. We find indeed the fossiliferous rocks from the region beyond Katahdin, a hundred and fifty miles north, scattered over the islands of the coast; brought doubtless upon the more elevated parts of the glacier, rather than attached to or near its under surface, during their distant transportation. But these rocks are by no means abundant.

Upon Vinalhaven, as we leave one of the granitic quarries on

the western side of Carver's Harbor and pass along north toward the highest granitic hill in the town, there rises a series of terraces, one above another, along a north and south course, till they attain the altitude of 150 feet above the water. The highest ridge but one is seven hundred yards long, and fairly unbroken except in one place, where there is a breach forty yards long. This granitic wall nearly resembles some of the "Horsebacks" in the country, and a well trained horse might be safely led along its entire length. This wall descends toward the east and the large salt-water mill-pond, at various angles, from thirty and sixty to ninety degrees. Between this ridge and the one above, which is the highest, there is a depression fifty yards wide and from ten to twenty feet deep. The rock of many thousand tons has all been swept away to the south of it for 150 feet, where we meet again another slope, rising gradually toward the south till it attains the usual height of this elevated ridge.

I consider this dell as having been gouged out of the rock, as the most of our harbors, coves, and ponds have been, by glacial action. We see the hills not only curving easily to the north, and steep on their southern sides, but we find also that their east and west sides are abrupt. We know that this abruptness, in the southern part of Maine, must have been caused by denudation; not such a wearing away as slow moving icebergs would be likely to make, though hundreds of thousands had struck in the same place. And those east and west sides both low, and high up the hills, are often seen beautifully, never roughly, scored. I suppose that when an iceberg touches along the side of a submarine hill, it would be deflected like any floating body, by the continuation of the current around the hills; it could not uniformly chafe and scratch the rock in all its inequalities of wall; for we know that the sides of the bergs are not abundantly sup plied with those stone-grooving tools necessary for the smoothing and scratching.

And, let me ask, by what means were those oblong and wedgeshaped boulders deposited in the peculiar manner in which we find them? They do not lie with their longer diagonals across the striæ, but were left by some agent head on in their course toward the south. Icebergs could not have dropped them thus, while the movements of a glacier would have compelled the boulder to take this wedge-shape form in many cases, and would have kept its base always directed forward in the line of its

course.

If then there is evidence of a power so great, acting against the highest hills of the coast, even leaving over their summits indisputable marks of extensive denudation, we have rea son to believe that the glacier which swept across them was of vast thickness. Had the glacier reached barely above the top

of the highest hill of the Mount Desert group, an elevation of 2000 feet, its action for ages might have indeed accomplished the rounding and sloping of the northern sides of those mountains; but the ice of a few hundred feet in thickness above their summits, could not have produced that vast amount of denudation required to form their southern brows. It is altogether probable that the glacier far overtopped the highest of the hills, and it is not unreasonable to suppose it to have been at least twice two thousand feet in thickness.

Moreover there is evidence that this glacier was not limited to this great fiord of Maine. It must have extended far toward the east and covered the country on the west. In fact, it was probably a part of the universal glacier which covered the continent wherever the drift striæ have been observed. But upon the discussion of this point I will not now enter.

ART. XXX-Contributions from the Sheffield Laboratory of Yale College. No. VII.-On the Indirect Determination of Potash and Soda; by PETER COLLIER, B.A., Assistant in the Sheffield Laboratory.

THE method customarily employed in estimating potash and soda, viz: by the precipitation of the former as platinchlorid of potassium and reckoning soda from the loss, though sufficiently accurate in patient and skillful hands, is yet open to many sources of error and at the best is exceedingly tedious and troublesome.

The indirect method does not yet appear to possess the confidence of chemists, at least, it is rarely mentioned in published investigations. I have therefore, at the suggestion of Prof. Johnson, made a number of experiments to ascertain the limits of error in this process.

The volumetric estimation of chlorine as perfected by Mohr offers by far the best basis for an indirect determination of the alkalies. It is in fact requisite in employing the usual direct method, to procure the alkalies in the condition of pure chlorids before precipitation.

When the alkali chlorids are obtained free from all foreign matters, it is but the work of a few moments to ascertain their content of chlorine.

The silver solution used for this purpose is best prepared by weighing off in a porcelain crucible about 4.8 grm. of clean crystallized nitrate of silver, fusing it at the lowest possible heat, and then ascertaining its weight accurately. After fusion it should weigh a little more than 47933 grm., the quantity, that, contained in a liter of water, gives a solution of which 1 c. c. =

001 grm. of chlorine. The fused salt is dissolved in a little warm water, the solution brought into a liter flask and filled to the mark, observing the usual precautions as to temperature, &c. When thus adjusted, add to the contents of the flask, from a burette, enough water to bring the excess of nitrate of silver above 4 7933 grms. to the requisite dilution. In this way it is easy with a burette and liter flask to make perfectly accurate standard solution, while this would be hardly possible should the operator weigh off less than 4-7933 grm. of nitrate of silver.

This solution, which may be preserved in a well corked bottle indefinitely, without change, is next tested by means of a solution of pure chlorid of sodium or chlorid of ammonium, a quantity, say about 2 grams, of one of these salts being dissolved in a liter of water and 20 c. c. of the liquid taken for the comparison. The solution being ready, the estimation of chlorine is conducted as described by Mohr, Fresenius, Sutton and others, chromate of potash being employed to indicate the completion of the reaction. The use of Erdmann's float in a burette (which may hold 70 c. c.) graduated to fifths ensures the needful accuracy of reading. In my determinations th c. c. of silver solution were deducted as the excess needed to produce a visible quantity of chromate of silver.

The appended table gives the results I obtained in the analy ses of the chlorids of potassium and sodium. The salts were perfectly pure and the quantities were weighed out in each case. In order to test the method thoroughly I have varied the proportions of the mixtures from one extreme to the other.

Summary of Volumetric Chlorine Determinations.

It may be seen from the above list of analyses, which includes all the determinations I have made, from first to last, for the purposes of this paper, that in no case does the difference between the quantities taken and found of either alkali chlorid exceed two milligrams, and in most instances it is less than one milligram. The correspondence between the amounts of chlorine

AM. JOUR. SOI.-SECOND SERIES, VOL. XXXVII, No. 111.—May, 1864.

as taken and found is of course still more near. The errors that appear in the estimation of the chlorids would be considerably reduced, if, as usually happens, they were calculated as oxyds.

Here follow the formula which I have employed for calcu lating the quantities of NaCl and K Cl, or of NaO and KO, contained in or corresponding to any mixture of alkali-chlorids whose total weight and amount of chlorine are known.

The results I have obtained thus demonstrate that the indirect method is in all cases equal in accuracy to the ordinary separation, while in the matter of convenience and economy of time there is no comparison between them.

ART. XXXI.-Contributions to Chemistry from the Laboratory of the Lawrence Scientific School; by WOLCOTT GIBBS, M.D., Rumford Professor in Harvard University.-No. I.

1. On the relations of hyposulphite of soda to certain metallic oxyds.

THE employment of hyposulphite of soda as an analytical reagent was first suggested by Himly,' who in a short paper, intended only as a preliminary notice, pointed out the relations of this salt to solutions of arsenic, antimony, copper and platinum, and suggested its use as a general reagent in place of sulphydric acid. Himly's paper attracted little attention and was soon forgotten. The subject was again taken up at about the same time by Vohl' and Slater' who appear to have been unacquainted with Himly's results. Vohl's investigation embraces the action of the hyposulphite upon solutions of arsenic, antimony, tin, copper, mercury, silver, gold, platinum, lead, bismuth, and cadmium. Slater studied the action of the salt upon solutions of chromium in the form of chromic acid, arsenic, antimony, copper, bismuth, lead, and mercury, as well as upon ferrideyanid, ferrocyanid, and cyanid of potassium, and upon sulphocyanid of iron and hypermanganate of potash. The results obtained by these chemists differ in some particulars, especially as regards copper and lead. More recently Chancel' has employed the hyposulphite for the separation of alumina from iron, and Stro1 Annalen der Chemie und Pharmacie, xliii, 150. Chemical Gazette, 1855, 869.

2 The same, xcvi, 287.. • Comptes Rendus, xlvi, 987.