neolite, and allied in composition to the zeolites, and these by subsequent metamorphism have been changed into anhydrous silicates. The production of harmotome, chabazite, and apophyllite by the waters of a spring at Plombières, at temperatures not above 160° F. as observed by Daubrée, lends probability to such a view. But while we admit the possible direct formation of double silicates in water at ordinary temperatures, there is not wanting evidence that the reaction which we long since pointed out, (Proc Royal Society of London, May 7, 1857) between silicious and argillaceous matters and earthy carbonates, in presence of alkaline solutions intervenes in the metamorphism of sedimentary rocks and in the production of many silicious minerals. The blue Silurian limestones of the island of Montreal, when treated by acids leave an insoluble residue, which contains about ten per cent. of soluble silica, mixed with an argillaceous matter whose analysis gave silica 73·0, alumina 18.3, potash 5.5, and only traces of lime and magnesia. In the vicinity of an intrusive dolerite, however, the limestone is changed in colour, and leaves by the action of acids a greenish matter which consists of silica 40-2, alumina 9.3, peroxyd of iron 5.2, lime 36·6, magnesia 3.7. The free silica and that of the intermingled aluminous silicate, has thus been saturated with protoxyd bases, still however, retaining the alumina in combination. A similar reaction with more aluminous matters, would give rise to epidote, garnet, magnesian mica, scapolite or feldspars like labradorite and anorthite, and it is not impossible that in such reactions a portion of alumina may sometimes be set free, and give rise to corundum, spinel, diaspore or völknerite. In the ordinary modes of decomposition of minerals containing alumina, this base separates in the form of silicate, and the conditions required for its elimination in a free state are but imperfectly understood. We have elsewhere pointed out the decomposition by alkaline and earthy carbonates, of solutions of sulphate of alumina or native alum, as one source of free alumina, and insisted upon the existence of pigotite, a native compound of alumina with an organic acid, as an evidence that this base is sometimes like oxyd of iron, (and oxyd of manganese,) taken into solution by water aided by organic matters. A hydrate of alumina gibbsite, is found associated with limonite, and the aluminous minerals from the south of France described by Berthier and Deville, show that free alumina is much more common in nature than was formerly supposed. Berthier long since gave the name of bauxite to an earthy pisolitic ore which occurs either massive or imbedded in limestones of tertiary age, at Baux, and many other localities in the departments of Gard and Var, and also in Calabria, and the Grecian Archipelago, forming in some places an abundant rock.* This substance is a variable mixture of a hydrate of alumina, apparently approaching diaspore in composition, with hydrous peroxyd of iron, sometimes constituting a workable iron ore, and at other times a veritable ore of alumina. It contains besides small portions of silica, titanic, vanadic, and phosphoric acids, and occasionally encloses grains of corundum. A compact dark red variety gave Deville, alumina 57-6, peroxyd of iron 25-3, and water 10.8, besides 3·1 of titanic acid, and 2.8 of silica. In other specimens the proportions of alumina and iron oxyd are nearly equal, or the latter predominates, as in one example where the proportions were 48.8 of iron oxyd, and 32.2 of alumina; and another, 60 of iron, and 18 of alumina and titanium. In these analyses the carbonate of lime, generally present, was first removed by a dilute acid; the prolonged action of stronger, acids completely dissolves the hydrated oxyds. By an intense heat this substance is converted into crystalline corundum, resembling emery in its physical character, but the presence of grains of corundum in the hydrated mineral seems to show that the transformation may take place at ordinary temperatures. The emery of Greece and Asia Minor, which is associated with variable proportions of oxyd of iron, is according to Dr. J. Lawrence Smiths always more or less hydrated. The argillaceous matter enclosing some varieties of this bauxite or impure diaspore, is white, without plasticity, and very rich in alumina; one specimen freed from the red ferruginous portions,gave alumina 58.1, silica 217, peroxyd of iron 3.0, titanium 3-2, water 14.0. This substance approaches in its composition to collyrite, and the dillmite which is the gangue of the diaspore of Schemnitz. These materials however contain from 20 to 40 per cent. of water. Scarbroite, schrotterite, and allophane are similar matters; the latter, unlike a clay in its structure, appears to have been deposited from solution. The subsulphate of alumina, known as websterite or aluminite, is often met with in layers and concre * Deville. Ann. de Chime et Physique, (3) Ixi. 309. tionary masses in tertiary clays,* and is sometimes mingled with a silicate having the composition of allophane. This frequent occurrence of alumina still retaining a portion of sulphuric acid, confirms the view which we have elsewhere expressed, that solutions of native alums have by their decomposition furnished the alumina for many of the minerals in question, while the con. ditions under which this base is taken into solution by organic matters, still require investigation. The careful examination of unaltered sedimentary deposits, is calculated to throw great additional light upon the origin of the crystalline rocks. ARTICLE XII.-On Canadian Caverns. BY GEORGE D. Gibb M.A., M.D., F.G.S. London, England. (Extracts from a Paper read before the British Association, Sept. 1859.) The prominent feature of a large portion of the province of Canada is the presence of various limestone rocks belonging to the Silurian formations. Until lately, the existence of caverns in these rocks, as well as in those lying subjacent-namely, the Laurentian of Sir William Logan, was almost unknown; as, with the exception of an isolated account here and there, no regular description of any cavern had appeared. Owing to the labours of the Canadian Geological Survey, and of several private individuals, a number of caverns have been discoved at distances remote from one another; some of these have received but a passing notice in the publications of the Survey, and are not, therefore, useful as a means of reference. The present communication it is hoped, will supply that deficiency, as in it I propose to embody, short descriptive accounts of all the caverns of Canada which are known up to the present time. The details of some of them are not so full as could be desired; nevertheless, with all the available sources of information within my reach, together with personal observation in some, on the whole the general descriptions may be relied upon as accurate, and as containing a correct account of the geological formations in which they lie. The caverns of Canada may conveniently be divided into two classes; the first comprises those which are at the present time washed by the waters of lakes, seas, and rivers, including arched, In this connection we may notice apatelite, a basic persulphate of iron, which occurs in conditions similar to aluminite. perforated, flower-pot, and pillared rocks, which have at one time formed the boundaries or walls of caverns, and all of them unquestionably the result of aqueous action. The second comprises caverns and subterranean passages which are situated on dry land, and so far as we know, not attributable to the same causes in their origin as the first, or at least not applied in the same manner. In the first class are included the following: 1. Caverns on the shores of the Magdalen Islands, 2. Caverns and arched rocks at Percé, Gaspé, 3. Gothic arched recesses, Gaspe Bay. 4. The "Old Woman," or flower-pot rock, at Cape Gaspé. 5. Little River Caverns Bay, of Chaleur. 6. Arched and flower-pot rocks of the Mingan Islands. 7. Pillar sandstones, north coast of Gaspé, 8. Niagara Caverns. 9. Flower Pot Island, Lake Huron. 10. Perforations and caverns of Michilimacinac, L. Huron. 11. The Pictured Rocks, Lake Superior. 12. St. Ignatius Caverns, Lake Superior, 13. Pilasters of Mammelles, Lake Superior, 14. Thunder Mountain and Paté Island Pilasters, L. Superior, In the second class are: 15. The Steinhauer Cavern Labrador. 16. The basaltic caverns of Henley Island. 17. Empty basaltic dykes of Mecattina, 18. Bigsby's Cavern, Murray Bay. 19. Bouchette's Cavern, Kildare. 20. Gibb's Cavern, Montreal. 21. Probable caverns at Chatham, on the Ottawa, 22. Colquhoun's Cavern, Lanark, 23, Quartz Cavern, Leeds. 24. Probable caverns at Kingston, Lake Ontario, 25. Mono Cavern. 26. Eramosa Cavern. 27. Cavern in the Bass Islands, Lake Erie. 28. Subterranean passages in the Great Manitoulin Island, Lake Huron. 29. Murray's Cavern and subterranean river, Ottawa. 30, Probable caverns in Iron Island, Lake Nipissing. The majority of those in the first class are on a level with the water, whilst the remainder are elevated above, varying from a few to upwards of sixty feet. In the second class the level varies, but nearly all are above that of the sea, as will presently be described; none penetrate the earth to a considerable depth, but this may be found to be otherwise as the explorations are continued. In none have animal remains been found, excepting in one instance, and they were discovered loose and not imbedded in stalagmite; and so far as I am aware, not a single object, such as a flint arrow-head or spear, used by the ancient inhabitants of the country, has been observed. This circumstance may in some measure detract from the value of the present communication; that part of the enquiry has still to be worked out, as many of the caverns have been but very partially explored, indeed some have scarcely been examined; and as several of them branch off by means of fissures and galleries, running from distinct chambers (most of the latter containing stalag. mite), we may yet hope for interesting discoveries, particularly in that district of country in which exist the huge caverns of Mono and Eramosa in the Niagara limestone rocks of the Upper Silurian formation. The researches of Mr. Sterry Hunt, of the Canadian Geological Survey, have shown that these limestones are essentially dolomitic, and thus perhaps favourably constituted for the development of caverns. (As examples of the caverns noticed by Dr. Gibb we take the following:- EDS.) CAVERNS ON THE SHORES OF THE MAGDALEN ISLANDS. On passing the interesting group of islands in the Gulf of St Lawrence, known as the Magdalens, the observer is struck with their beautiful and picturesque appearance, which is suddenly presented to his view. The cliffs, which vary in height, present equally various colours of red in which the shades predominate; these contrasted with the yellow of the sand-bars, and the green pastures of the hill-sides, the darker green of the spruce trees, and the blue of sea and sky, produce an effect, as Captain Bayfield describes, extremely beautiful, and one which distinguishes these islands from anything else in the Gulf. Such an agreeable picture it has been my own good fortune to witness and admire. The striking feature in their formation is the dome-shaped hills rising in the centre of the group, and attaining a height of from two hundred to five hundred and eighty feet. They are composed of the |