^{page 722 note *} See Murray and Renard, Deep-Sea Deposits Chall. Exp., London, 1891.

^{page 722 note †} The following are some of the manganese-bearing silicates (the numbers after the names indicate the percentage of MnO):—Paulite, 0–6; diallage, 5–20; augite, 0–3; acmite, 1–3; rhodonite, 54; hermanite, 47; regerine, hornblende; pyrosinalite, 21; astrophyllite, 10; tephroite, 70; knabellite, 35; zephrolite; manganese-alumina garnet; pyrochlore, 7 (somewhat variable); tantalite, 1–6.

^{page 722 note ‡} Murray and Renard, op cit., p. 307.

^{page 722 note §} I. In the following rocks, chiefly from the Clyde drainage area, the manganese oxide was soluble in carbonic and dilute acetic acids, therefore presumably present as carbonate:—

II. In the following rocks the manganese occurred only partly in a soluble condition:—

III. In the following rocks the manganese was in an insoluble condition, probably silicate:—

Besides the above-named rocks, about fifty others were examined for manganese, which was found to be present in amounts varying from traces to 0·5 per cent, of MnO.

Fresh Blue Muds, from various places in the estuaries of the Clyde and Forth, were examined and found to contain manganese partly as silicate and partly as carbonate. There was a slight trace of manganese found in some samples of Globigerina Ooze, but none in corals, boiler deposits, or siliceous sinter (from Iceland). A piece of coral (Pleurocorallium johnsoni, taken by the “Challenger,” Station 3, 1525 fathoms), was coated on the outside with manganese dioxide, but internally there was no trace of manganese. Minute traces were found in kelp and in sea-weed ash.

The process adopted for the comparative estimation of manganese was the colour-test, obtained when the material under examination was fluxed with potassium and sodium carbonates and a little pure potassium nitrate. The fused mass, on cooling, gives the green colour characteristic of manganates—even when manganese is only present in minute traces. Comparison with a standard series of coloured fluxes, each containing a known amount of manganese varying from 0·01 to 1·0 per cent., was an easy, quick, and at the same time a comparatively accurate, method of estimating the quantity of manganese present. The proportion of manganese existing in the rocks under the various conditions of carbonate, silicate, and higher oxides might be roughly determined by thorough trituration and treatment with carbonic acid in aqueous solution; but the expenditure of time involved in the prosecution of this process renders the substitution of dilute acetic acid for carbonic acid advisable, the dilute acetic acid attacking the carbonate alone. The amount of peroxide present was determined as usual by the Bunsen process, i.e., by taking advantage of its power of liberating chlorine from hydrochloric acid. Crystalline silicates, even when reduced to the finest state of division, are only very gradually decomposed by carbonic acid. We found that all the manganese which is combined with carbonic acid in the rocks is thus rapidly extracted by the use of dilute acetic acid, and can be determined by evaporating the solution so obtained to dryness, and treating the dried residue with fluxing materials at a red heat, or by actual precipitation by the general methods in use for manganese determination. We can thus determine the proportion of manganese existing in a rock or mineral as carbonate. If, however, the manganese exists partly as carbonate and partly as silicate, we obtain the portion present as carbonate in the acetic acid solution, and the portion existing as silicate or the higher oxides in the insoluble residue.

The process adopted by us for the determination of the quantity and state of combination of the manganese in a rock sample was as follows:—

(1) Exactly 1 gramme of the sample, reduced to a fine powder, was intimately mixed with 4 grammes of the fusion mixture already referred to (20 parts Na₂CO₃, 26 parts K₂CO₃, 1 part KNO₃), and heated in a platinum crucible over the blow-pipe until tranquil fusion supervened. The liquid mass was then poured out upon a porcelain slab and allowed to cool. The quantity of manganese present in the fused magma was ascertained by comparing its colour with that of the series of standard coloured fluxes. This gave the total percentage of manganese in the sample.

(2) Another portion from the same pulverised sample, weighing 1 gramme, was exhausted with dilute acetic acid (to extract the carbonates), and the amount of manganese in the residue determined colorimetrically after fusion as in (1). The deficit (i.e., the difference between the above two determinations) due to exhaustion with dilute acetic acid we assume to be present as carbonate of manganese.

In many of the rocks and minerals examined in this manner we found the manganese wholly combined with carbonic acid, in others partly with carbonic acid and partly with silicic acid; in some cases it is present as peroxide; it may exist in all three forms in the same rock, whilst in the majority of minerals it exists combined with silicic acid alone.

The felstones of the upper area of the Clyde basin seem to contain all their manganese in a soluble condition, presumably as carbonate. These rocks effervesce on treatment with dilute acids. If the deposits of the Clyde Sea-Area contain more peroxide than the deposits of other similar areas, this may be due to the soluble condition of the manganese in these felstones. In the laboratory, when a portion of these rocks was simply fused or fritted without fluxing material, the carbonic acid was expelled, and the bases were found in the cooled mass to be combined with silicic acid alone, as a silicate or silicates insoluble even in strong hot hydrochloric acid. Even when ten per cent, of carbonate of lime was added to the pulverised felstone, and the mixture fused, the silicates in the fritted material were insoluble, showing the acid nature of this class of rocks. This experiment shows that in the presence of silicic acid (or acid silicates) carbonates are decomposed by heating, silicic acid taking the place of the carbonic acid expelled, and also that the manganese found in these rocks has really been infiltrated as carbonate into even the heart of the felstone. These felstones are hard compact rocks, and contain no water which can be expelled even at 400° F. (204°·4 C), nor do they absorb water, even when soaked in it for twenty-four hours.

We are indebted to John Young, Esq. LL.D., Hunterian Museum, Glasgow; Mr C. Maclaren Irvine, Lanarkshire; J. S. Dixson, Esq., Hamilton; Mr Durham, Newport, Fife; Mr Pearcey, and Captain Turbyne, for specimens of rocks, drainage waters, and deposits from the Clyde area.

^{page 724 note *} Professor Clarke estimates that manganese makes up about 0·08 per cent, of the earth's crust. (Bull. Phil. Soc. Washington, vol. xi. p. 138, 1892Google Scholar).

^{page 724 note †} The following are the principal manganese minerals:—Pyrolusite, (MnO₂); hausmannite, (Mn₃O₄); braunite, (Mn₂O₃); manganite, (Mn₂0₃H₂0); psilomelane, (MnO₂, united with some protoxide, as of Mn.Ba.K₂, or H₂); dialogite (MnC0₃).

The following are oxides–Jacobsite, crednerite, chalcophanite, franklinite, pyrochroite.

Sulphide (blende)—Alabandinc (MnS).

Anhydrous Carbonates—Breunnerite, siderite, mangano-calcite.

Hydrated Sulphates—Lankite, mellardite.

Hydrated Phosphates (the numbers are percentages of MnO in the mineral)—Fillowite (40), heterozite, dicksonite (25), fairfieldite (1C), ncddingite (46), eosphorite (24), childrenite (9), tuplite.

^{page 725 note *} L. de Launay, Formation des gîtes Métallifères, p. 232; M. Saralp, Le Manganese des Pyrénés, (Memoirs de l'Académie des Sciences de Toulouse, 1893).

^{page 725 note †} Stones and sand coated and aggregated with manganese dioxide, and often presenting a polished surface and black metallic lustre, were observed in the River Clyde (Stonebyres Falls, Rutherglen, Underbank) and its tributaries the Mousewater, Hallhill, Diller, Devon, Teiglam, Craignethan, Birkwood, Poniel, Powtrail and Shortcleugh Burns, and Cander and Fence Avaters; also in the streams flowing directly into the Clyde Sea-Area, as Sea Mill Water, Skelmorlie and Fairlie Burns (Wemyss Bay), and the streams falling into Loch Fyne, Loch Ranza, Loch Goil, and Campbelltown Loch. At Middleton Farm, Loch Fyne, there occurs a sandy deposit containing 0·7 per cent. MnO₂, and at Dundee there is found in alluvium, interlayered with red sand, a black sand containing 2·5 to 3 per cent. MnO₂.

^{page 725 note ‡} Water was examined in this way from the Clyde, Nethan, Mousewater, Hagshaw Burn, Loch Ranza Burn, and Glen Morag Burn.

^{page 725 note §} See Murray, and Irvine, , “Silica in Modern Seas,” Proc. Roy. Soc. Edin., vol. xviii. p. 243, 1891Google Scholar.

^{page 725 note ∥} The samples in which manganese was found in solution came from Clyde River and its tributaries, as, for example, Mouse Water, Hagshaw Burn, Hallhill Burn, Skelmorlie Burn, Glen Morag.

^{page 726 note *} To determine the limits of detection of manganese in sea-water, standardised solutions of pure chloride of manganese (in sea-water) were prepared, and after boiling with excess of bromine for some time the precipitated MnO₂ was estimated. With solutions containing from 1 part in 15,000 to 1 part in 100,000, the separation was rapid and apparently complete, and the precipitated dioxide was collected on a filter and weighed, after ignition, as Mn₃O₄; with solutions containing 1 part in 1,000,000, MnO₂ separated out after boiling with bromine for some time. On treating a solution of 1 part MnCl₂ in 10,000,000 of sea-water, the precipitate of MnO₂, after prolonged boiling with bromine, was quite distinct, but in this case appeared as a brown scum on the surface of the liquid, and formed a distinct brown ring round the walls of the white porcelain basin above the evaporating surface of the liquid.

Having established this point, we endeavoured to find manganese by this method in fresh, clear, filtered sea-water, obtained from the German Ocean. Two gallons were evaporated until the contained sea-salts began to crystallise out. The liquid was filtered clear of deposited sulphates and carbonates, and treated with bromine. There was not even the faintest trace of coloration. The basin in which this sea-water was evaporated was washed with hot hydrochloric acid, so as to decompose carbonates thrown down during evaporation, and the filtered liquid so obtained exactly neutralised with ammonia, and thereafter treated with bromine. In this case there was not the faintest trace even of coloration. We therefore conclude that in the sea-water samples examined by us, manganese, if present, could not have been there to a greater extent than 1 part in 10,000,000.

To confirm the above by more delicate tests, the two portions from the sea-water—viz., the strong brine and the residue of salts treated with hydrochloric acid—were boiled down until most of the salts had crystallised out. The crystals were separated, washed with pure hydrochloric acid, and the washings added to the mother liquor, which was then evaporated to dryness, and ignited to drive off ammoniacal salts. This residue—in which we assume any manganese present in solution in the original sea-water would appear—was then fused with carbonates of potash and soda and a little potassium nitrate. The fluxed mass when cold was absolutely milk-white. This result (when the extreme delicacy of this method of detecting manganese is taken into account) points to the conclusion that manganese is not present in solution in ordinary sea-water, at least within the chemical limits of observation at our disposal, and therefore ordinary sea-water cannot provide the material for the formation of manganese nodules.

^{page 726 note †} Murray, and Irvine, , “On the Chemical Changes which take place in the Composition of the Sea-Water associated with Blue Muds on the Floor of the Ocean,” Trans. Roy. Soc. Edin., vol. xxxvii. p. 481, 1893Google Scholar.

^{page 728 note *} See Murray, and Irvine, , Trans. Roy. Soc. Edin., vol. xxxvii. p. 485Google Scholar.

^{page 729 note *} See Irvine, and Gibson, , Proc. Roy. Soc. Edin., vol. xviii. p. 54, 1891Google Scholar.

^{page 729 note †} If we take the Blue Mud extending over an area of 1 square mile and 1 foot in depth as containing one half of its weight of water (equal to 867,700,000 lbs.), and holding 1 part of MnC0₃ in 95,000, we will have a total amount of 9134 lbs. MnC0₃ per square mile, and if we take the amount found in the water immediately overlying the mud in Granton Quarry as representing what occurs in the Clyde area, i.e., 1 part in 300,000, we have per square mile of water 1 foot deep (weighing 1,735,400,000 lbs.) 5785 lbs. MnC0₃, or a total, including that obtained from the mud and from the water overlying it, of 14,919 lbs., or 23 lbs. per acre of surface available for nodule formation. Of course there may be much more, as there must be a, continuous removal of MnC0₃ by tidal action; but, considering the extent of the floor of the Clyde basin, even the amount here estimated is very great.

Taking the amount of flow at Lanark at 15,000 cubic feet per second, holding 1 part manganese in 28,000,000 of water, this represents a daily quantity of over 1 ton carried by that river to the sea. This is manifestly a low approximation, as the flow is much augmented by tributaries between this point and the sea.

^{page 732 note *} Partial Analyses of Several Clyde Nodules (Anderson).

Partial Analysis of a Manganese coating on a stone from Loch Fyne (see Fig. 4).

The substance was broken, and the Fine Particles separated from the Coarse by levigation, and dried in air.

Manganese coating was divided into Fine Washings = 63·03 per cent.

The main difference between the two analyses is the larger quantity of iron and elayey matter present in the Fine Washings, and a corresponding increase of manganese in the Coarse Particles.

^{page 734 note *} See Buchanan, , Trans. Roy. Soc. Edin., vol. xxxvi. p. 459, 1891Google Scholar.

^{page 734 note †} Captain Turbyne, of Mr Murray's yacht “Medusa,” writes as follows as to the manganese dredging in the Clyde Sea-Area:—“Regarding the nodules being in pot-holes, I consider that I have absolute proof of that on the outer side of the barrier of Loch Goil. In the first place, we were dredging from the outside toward the barrier or up the loch, on the slope, as I thought, between the mud and harder ground. In this case we ought to have been shallowing our soundings, but after towing for a short time the dredge suddenly began to dip, which was seen from the angle of the wire, and more had to be run out; then we suddenly came fast, and had to heave up, and this was the haul in which the nodules were got. They differed from those at Skelmorlie Bank in being much larger, and to the unaided eye they seemed perfectly smooth and quite round. In the second place, after finding the nodules I tried to get more both on this and on several other occasions; but though we tried to strike the spot as nearly as possible, it was only after dredging up and down the loch and in close sections across it, that we again hit on the spot. For these reasons I have come to the conclusion that the nodules are found in a hole of small extent. This is the only instance in my experience in which I am certain that the nodules were taken from a small hole. In my opinion the slopes of Skelmorlie Bank and Minard Narrows are full of small holes containing mud and manganese nodules where the tide meets with an obstruction. The 106 fathom-hole has always been a mystery to me. I often thought that if the nodules were formed in that deep hole, why don't we get them off Brodick in 95–97 fathoms, which part is practically a continuation of the trough in which the 106 fathom-hole is situated? Surely this submarine tongue you mention and the small size of the 106 fathom-hole has got something to do with it. I can say nothing about under-currents at the deep hole, but there is a strong surface-current at spring-tides, as we rapidly got out of position, and it is well known to fishermen and others that, with the wind against the tide, a nasty sea is met with off Skate Island.”

^{page 740 note *} Gümbel, , Sitzb. d. k. Layer. Akad. d. Wiss., Bd. viii. p. 189, 1878Google Scholar; Forschungsreise S.M.S. “Gazelle,” Th. ii. p. 103.

^{page 740 note †} See Lockyer, , Nature, vol. xxxviii, p. 521, 1888Google Scholar; Hickson, , The Fauna of the Deep Sea, p. 38, London, 1894CrossRefGoogle Scholar.

^{page 741 note *} Murray and Renard, Deep-Sea Deposits Chall. Exp., pp. 327–336.

^{page 741 note †} Dieulafait, , Comptes rendus, tom. xcvi. p. 718, 1883Google Scholar.

^{page 741 note ‡} Murray and Renard, Deep-Sea Deposits Chall. Exp., p. 372, note.

^{page 742 note *} Fortnightly Review, January 1894, p. 73.

^{page 742 note †} On the Distribution of Volcanic Débris over the Floor of the Ocean, Proc. Roy. Soc. Edin., vol. ix. p. 255, 1877Google Scholar.

^{page 742 note ‡} Bischoff, Chemical Geology, vol. iii. p. 508, English edition.

^{page 742 note §} Annales de Chemie et de Physique, ser. 5, tom. xxvii. pp. 289–311, 1882Google Scholar.