rapid, its rate having been accelerated by the erection of saw-mills, dams, &c. From the accounts of various travellers who have visited the falls in the last 200 years, the author endeavoured to obtain an estimate of the true rate of recession. Between the visit of Father Hennepin in 1688 and that of Carver in 1766 he finds a recession at the rate of 3.49 feet annually, between Carver's visit and 1856 a mean annual recession of 6.73 feet, and between Hennepin and 1856 one of 5-15 feet. The time-estimates for the cutting of the gorge would be, according to the above means, 12, 103, 6, 276 and 8202 years. The author considers the data upon which the second of these numbers is founded the most reliable.

April 17.-Henry Clifton Sorby, Esq., F.R.S., President, in the

Chair.

The following communications were read :

1. "On the Geological results of the Polar Expedition under Admiral Sir George Nares, F.R.S." By Capt. H. W. Feilden, R.A., F.G.S., and C. E. De Rance, Esq., F.G.S.

The authors describe the Laurentian gneiss that occupies so large a tract in Canada as extending into the Polar area, and alike underlying the older Palæozoic rocks of the Parry Archipelago, the Cretaceous and Tertiary plant-bearing beds of Disco Island, and the Oolites and Lias of East Greenland and Spitzbergen. Newer than the Laurentian, but older than the fossiliferous rocks of Upper Silurian age, are the Cape-Rawson beds, forming the coast-line between Scoresby Bay and Cape Cresswell, in lat. 82° 40'; these strata are unfossiliferous slates and grit, dipping at very high angles.

From the fact that Sir John Richardson found these ancient rocks in the Hudson's-Bay territory to be directly overlain by limestones, containing corals of the Upper Silurian Niagara and Onondaga group, Sir Roderick Murchison inferred that the Polar area was dry land during the whole of the interval of time occupied by the deposition of strata elsewhere between the Laurentian and the Upper Silurian; and the examination by Mr. Salter, Dr. Haughton, and others of the specimens brought from the Parry Islands have hitherto been considered to support this view. The specimens of rocks and fossils, more than 2000 in number, brought by the late expedition from Grinnell and Hall Lands have made known to us, with absolute certainty, the occurrence of Lower-Silurian species in rocks underlying the Upper Silurian; and as several of these Lower-Silurian forms have been noted from the Arctic Archipelago, there can be little doubt that the Lower Silurians are there present also. The extensive areas of dolomite of a creamy colour discovered by M'Clintock around the magnetic pole, on the western side of Boothia, in King William's Island, and in Prince-of-Wales Land, abounding in fossils, described by Dr. Haughton, probably represent the whole of the Silurian era and possibly a portion of the Devonian.

The bases of the Silurians are seen in North Somerset, and

consist of finely stratified red sandstone and slate, resting directly on the Laurentian gneiss, resembling that found at Cape Bunny and in the cliffs between Whale and Wolstenholme Sounds. Above these sandstones occur ferruginous limestones, with quartz grains; and still higher in the series the cream-coloured limestones come in. The Silurians occupy Prince-Albert Land, the central and western portion of North Devon, and the whole of Cornwallis Island. The Carboniferous Limestone was discovered, rising to a height of 2000 feet, on the extreme north coast of Grinnell Land, in Feilden and Parry Peninsulas, and contains many species of fossils in common with the rocks of the same age in Spitzbergen and the Parry Archipelago, being probably continuously connected with the limestone of that area, by way of the United-States range of mountains. The coal-bearing beds that underlie the Carboniferous Limestones of Melville Island are absent in Grinnell Land; but they are represented by true marine Devonians, established in the Polar area for the first time, through the determination of the fossils, by Mr. Etheridge. In America a vast area is covered by Cretaceous rocks. The lowest division, the Dakota group, contains lignite seams and numerous plant-remains indicating a temperate flora; overlying the Cretaceous series are various Tertiary beds, each characterized by a special flora, the oldest containing subtropical and tropical forms, such as various palms of Eocene type. In the overlying Miocene beds the character of the plants indicates a more temperate climate; and many of the species occur in the Miocene beds of Disco Island, in West Greenland, and a few of them in beds associated with the 30-feet coalseam discovered at Lady-Franklin Sound by the late expedition. The warmer Eocene flora is entirely absent in the Arctic area; but the Dakota beds are represented by the "Atane strata" of West Greenland, in which the leaves of dicotyledonous plants first appear. Beneath it, in Greenland, is an older series of Cretaceous plantbearing beds, indicating a somewhat warmer climate, resembling that experienced in Egypt and the Canary Islands at the present time. In the later Miocene beds of Greenland, Spitzbergen, and the newly discovered beds of Lady-Franklin Sound, the plants belong to climatal conditions 30° warmer than at present, the most northern localities marking the coldest conditions. The common fir (Pinus abies) was discovered in the Grinnell-Land Miocene, as well as the birch, poplar, and other trees, which doubtless extended across the polar area to Spitzbergen, where they also occur.

At the present time the coasts of Grinnell Land and Greenland are steadily rising from the sea, beds of glacio-marine origin, with shells of the same species as are now living in Kennedy Channel, extending up the hillsides and valley-slopes to a height of 1000 ft., and reaching a thickness of from 200 to 300 ft. These deposits, which have much in common with the "boulder-clays" of English geologists, are formed by the deposition of mud and sand carried down by summer torrents and discharged into fiords and arms of the sea, covered with stone- and gravel-laden floes, which, melted by the heated and turbid waters, precipitate their freight on the mud

below. As the land steadily rises these mud-beds are elevated above the sea. The coast is fringed with the ice-foot, forming a flat terrace 50 to 100 yards in breadth, stretching from the base of the cliffs to the sea-margin. This wall of ice is not made up of frozen sea-water, but of the accumulated autumn snowfall, which, drifting to the beach, is converted into ice where it meets the seawater which splashes over it.

2. "On the Paleontological results of the recent Polar Expedition under Admiral Sir George Nares, K.C.B., F.R.S." By Capt. H. W. Feilden, R.A., F.G.S., and Robert Etheridge, Esq., F.R.S., F.G.S.

XI. Intelligence and Miscellaneous Articles.

AMMONIO-ARGENTIC IODIDE. BY M. CAREY LEA.

WHEN silver iodide is exposed to ammonia-gas it absorbs 3-6 per

cent., and forms, according to Rammelsberg, a compound in which an atom of ammonia is united to two of AgI. Liquid ammonia instantly whitens Agl; every trace of the strong lemon-yellow colour disappears. The behaviour of the ammonia iodide under the influence of light differs singularly from that of the plain iodide, and will be here described.

The affinity of AgI for ammonia is very slight. If the white compound be thrown upon a filter and washed with water, the ammonia washes quickly out, the yellow colour reappearing. If simply exposed to the air, the yellow colour returns while the powder is yet moist; so that the ammonia is held back with less energy than the water. So long, however, as the ammonia is present, the properties of the iodide are entirely altered.

AgI precipitated with excess of KI does not darken by exposure to light even continued for months. But the same iodide exposed under liquid ammonia rapidly darkens to an intense violet-black, precisely similar to that of AgCl exposed to light, and not at all resembling the greenish-black of Agl exposed in presence of excess of silver nitrate. (This difference no doubt depends upon the yellow of the unchanged AgI mixing with the bluish-black of the changed, whereas in the case of the ammonia iodide the yellow colour has been first destroyed.)

When the exposure is continued for some time, the intense violet-black colour gradually lightens again, and finally quite disappears; the iodide recovers its original yellow colour, with perhaps a little more of a greyish shade. This is a new reaction, and differs entirely from any thing that has been hitherto observed. It has been long known that darkened AgI washed over with solution of KI and exposed to light, bleached. This last reaction is intelligible enough; for KI in solution exposed to light decomposes, and in presence of AgI darkened by light gives up iodine to the AgI, and so bleaches it. The above experiment is quite different. The darkened substance may be washed well with water (during which

operation it passes from violet-black to dark brown), and may then be exposed to light either under liquid ammonia or under pure water: in either case the bleaching takes place, though in the latter case more slowly.

If the experiment be performed in a test-tube, the bleaching under ammonia requires several hours, under water from one to three days. But if the iodide be formed upon paper, and this paper be exposed to light, washing it constantly with liquid ammonia, the darkening followed by the bleaching requires little more than a minute. In this case, however, the bleaching is not so complete, perhaps because of the influence of the organic matter present. The bleaching appears to depend upon the escape of ammonia; for if the darkened ammonia iodide is covered with strong liquid ammonia and the test-tube well corked, the bleaching does not take place.

It became a matter of interest to know whether the darkening under ammonia was accompanied by any decomposition-whether the ammonia took up iodine from the silver salt under the action of light. For this purpose AgI was precipitated with excess of KI, and subjected to a long and thorough washing; it was then exposed for several days to light under strong liquid ammonia. As AgI is not wholly insoluble in ammonia, the mother-water was first evaporated to dryness at a heat but little over ordinary temperatures. The traces of residue were washed with water; and this water gave distinct indications of iodine. The iodine present is in so small a quantity that it may easily be overlooked; but it is certainly there. The washing given to the AgI was so thorough that it seemed impossible to admit that traces of KI remained attached to the Agl; but in order to leave no room for doubt, the experiment was repeated, using an excess of silver nitrate in making the precipitation, followed by thorough washing. Iodine was still found in combination with ammonia; and under these conditions there could be no doubt that AgI had been decomposed.

When AgI is blackened under ammonia in a test-tube, and the uncorked test-tube is set aside in the dark for a day or two, the AgI assumes a singular pinkish shade. It thus appears that AgI under the influence of ammonia and of light gives indications of most of the colours of the spectrum. Starting with white, it passes under the influence of light to violet, and thence nearly to black this violet-black substance washed with water passes to brown. The brown substance covered with ammonia and left to itself in an open test-tube becomes pinkish in the dark, yellow in sunlight. These curious relations to colour which we see in the silver haloids, from time to time exhibiting themselves in new ways, seem to give hope of the eventual discovery of some complete method of heliochromy.-Silliman's American Journal, May 1878. Philadelphia, March 25, 1878.

ON THE PRODUCTION OF PLATEAU'S FILM-SYSTEMS.

BY A. TERQUEM.

M. Plateau, by the use of a mixture of soap-water and glycerine, has produced liquid films of a certain extent, and has thus been able to verify most of the laws respecting the form of the surfaces which constitute the boundaries of liquids whose molecules are subjected only to their reciprocal actions.

I pointed out, some years since, that for the glyceric liquid a mixture of soap-water and sugar might be substituted, the latter substance having, especially, like glycerine, the effect of augmenting the viscosity of the liquid, and preventing it from flowing away too rapidly. The production of the film-systems of M. Plateau demands the employment of a great quantity of liquid if polyhedra be used the edges of which are of large dimensions.

I have thought that large films of liquid might be easily obtained by bounding them in part by flexible threads instead of using for the purpose rigid wires only.

Thus, if two horizontal rods be joined by two vertical and equidistant flexible threads, on dipping the system in the saponaceous liquid and slowly lifting it out again, we get a vertical plane film bounded above and below by the two rods, and laterally by the flexible threads, which take the form of arcs of a circle. The radius of the circle evidently depends on the stretching weight. It is easily demonstrated that, if we designate by the distance between the two threads, by R the radius of the arc constituted by them, by the angle made with the vertical by the tangent of the arc at the point of attachment of the thread to the lower rod, by f the superficial tension of each of the two surfaces of the liquid film, and by p the stretching weight, we have the relation

p=2f(l+2R cos p).

Every thing happens, therefore, as if the distance between the threads were equal to that between the centres of the two arcs, the tension of the threads being omitted.

I of course submitted this formula to a series of experimental verifications, by measuring with the cathetometer the diminution of the vertical distance between the two horizontal rods produced by the existence of a liquid-film between them.

If H is the initial length of the threads, and H' the new vertical distance of the horizontal rods, to find the radius of the circle and the angle we have only to solve the two equations

whence we derive

As the angle is generally very small, this transcendental equation can be solved with sufficient approximation by substituting