shaking the whole together for a few minutes. This liquid, from its not crystallising when concentrated by evaporation, is very clean and agreeable to use, but does not seem so adapted for researches made beyond the reach of laboratories as does the borotungstate of cadmium solution. Methylene iodide can be prepared of a density of 3.33, which gives it a further advantage over all but Rohrbach's solution.

J. W. Retgers has shown how methylene iodide can be raised to a density of 3.65 by dissolving iodoform in it and afterwards iodine.* He utilises for the extraction of rutile, &c., from other heavy minerals various nitrates, which become liquid at about 70° Č., and are as dense as 5·0 (see p. 120).

Herr W. Muthmann † proposes the use of acetylene tetrabromide, and shows how it may be very cheaply prepared. It is diluted with benzene, or, as Mr. C. R. Lindley informs me, still more conveniently with petroleum spirit, known commercially as "deodorised benzene." Its maximum density is 3.01.

It will be seen that the dense liquids named will serve, by proper dilution, to determine the specific gravity of most of the rock-forming minerals, though they mostly fail to discriminate between garnet and ruby, topaz and diamond, &c. It may be noted, however, that beryl will float easily in a solution in which green tourmaline sinks, while the great mass of gems can be divided off by similar observations from quartz and other worthless matter. The specimens tested should be examined with a high-power pocket-lens or a microscope in order that their purity may be guaranteed; and it is obvious that abundance of enclosures, solid or fluid, will seriously affect the results. But in practice even closely-allied felspars can be distinguished as to specific gravity by this method, which has become of increasing value with the researches of each successive year.

Undoubtedly the happiest development of the method has been the diffusion-column invented by Prof. Sollas. A small test-tube, say inch in diameter, is half filled with the liquid at its maximum density; water or benzene, according to the dense liquid used, is then poured on the top, no special care being necessary. The tube is set aside for twelve hours or so, by which time a column will have been produced by diffusion, the density of which increases regularly downwards. Indexes are

*Neues Jahrb. für Min., &c., 1889, ii. Band, p. 185; also Min. Mag., vol. ix. (1890), p. 46.

+ Zeitschrift für Krystallographie, Bd. xxx. (1899), p. 73.

Nature, vol. xliii. (1891), p. 404; and T. D. La Touche, ibid., vol. liii., p. 199; also ibid., vol. xlix., p. 211.

dropped into this, either in the form of mineral fragments of known specific gravity, or of glass beads the latter, in coloured varieties, have a considerable range, and may have their densities determined in a diffusion-column side by side with known mineral indexes. These indexes, beads being the most convenient, float in the diffusion-column at levels corresponding to their specific gravities; hence the density of any mineral fragment dropped into a column may be found by measuring off the distance between two known indexes which lie respectively above and below it, and also measuring the distance of the mineral from one or other index. The matter is merely one of simple proportion, and the same column can be used for many fragments, and in experiments extending over several days. Mr. La Touche has devised an accurate mode of measurement, by drawing a thread horizontally across both a graduated mirror at the side of the tube and the tube itself; this thread is carried by a sliding piece of metal, fitting round the wooden support in which the test-tube is fixed. The graduated mirror is fixed vertically on the support at one side of the tube, and the position of any object in the liquid is read off by making the thread coincide with the centre of gravity of the object, the reading being given by the division cut by the thread when the eye views it as coincident with its reflection in the mirror. The note in Nature referred to contains figures which will show the details of construction. In many cases a millimetre-scale, held by the hand against the side of the tube, will suffice as a means of measurement.

Prof. Sollas points out that even gelatinous precipitates, if left long enough in the liquid, will lose their water and will sink to their proper level.

CHAPTER IV.

SIMPLE TESTS WITH WET REAGENTS.

THE test of solubility in water may be important in agriculture, where mineral salts of potassium are applied to the land. The taste of some minerals, as rock-salt, nitre, &c., is characteristic.

The test of solubility in acids has been very freely applied to minerals, though with results varying according to the strength

of the acid, the temperature employed, and the time allowed for the attack. Hydrochloric and sulphuric acids are those most commonly required; nitric acid may be useful if to hand. Various forms of stoppered bottles enclosed in cases with screwcaps have been devised to meet the requirements of the traveller. It is well not to keep a small sulphuric acid bottle too well filled, on account of the highly hygroscopic character of the liquid. Any neglect or defect in stoppering will allow it to take up water and overflow if left out of use for any length of time.

The mineral to be tested should be roughly powdered and placed in a small test-tube, a few drops of acid being poured upon it. Water should be added, since solution does not always take place in the concentrated acid. The results may be noted both in cold acid and after boiling. In all cases the time of immersion in the acid and the other conditions of the experiment should be noted where comparison is desired. As these facts are rarely stated in books on mineralogy, typical and known specimens should be compared with the doubtful one under the same conditions. Should complete solution take place, further qualitative tests may be applied. Ammonia, which is often carried by travellers, will thus serve to precipitate alumina and iron from solution in hydrochloric acid; and a number of other reactions will readily suggest themselves.

Some silicates are decomposed by boiling in hydrochloric acid, particularly those that are hydrated or with a low percentage of silica. The silica separates either in a powdery or a gelatinous condition, the jelly of silicic hydrate being often well seen after partial evaporation and cooling of the liquid. The mass clings to the test-tube, but may be removed by boiling with a strong solution of sodium carbonate.

Good examples for observing this gelatinisation are natrolite, nepheline (or elæolite), wollastonite, and ilvaite. The great majority of olivine crystals also gelatinise easily, and may be thus distinguished from pale pyroxenes, which are not decomposed.

But it must be remembered that the greater number of natural silicates are not decomposed by acids. In such cases it is necessary to fuse the powder for some time with sodium carbonate in a platinum spoon, on platinum foil, or, less conveniently, on charcoal, and to treat the resulting mass with water and hydrochloric acid in a dish or test-tube. The silica now separates out, while the bases go into solution. To extract the whole of the silica, the mixture must be evaporated to dryness, lumps being broken up with a rod or spatula, at a temperature somewhat above 100°, overheating being liable to cause recombination.

Again add hydrochloric acid and water, and heat; the silica can now be filtered off, and the bases in the solution are ready for any further experiments.

The commonest and most important use to which acids are put by the geologist is, however, in the examination of carbonates. A free effervescence occurs, carbonic anhydride being given off, when a carbonate is placed in hydrochloric acid. The acid should be slightly diluted, and in many cases must be heated before the reaction will take place. Sulphides of certain metals, as zinc, lead, and iron, are decomposed similarly with evolution of bubbles of sulphuretted hydrogen; but, provided the mineral examined be itself free from included sulphides, there is little danger of any confusion being caused. The smell of the sulphuretted hydrogen is, moreover, noticeable, even among the fumes of the hot acid.

The use of the acid-bottle in the field itself is very limited, owing to the occurrence of dolomite and other carbonates which do not effervesce until heated. Thus the rough and ready test of putting a drop of acid directly upon the mineral or rock is of service in indicating calcite, but by no means decides that the substance is not a carbonate when no effervescence is obtained. Heating in the test-tube is the only sure method; a few granules of the substance, a small tube, a match or so, and the acid-bottle, being all the apparatus required. The occurrence of dolomite is often overlooked, and some hard dolomites have even been regarded as quartzites on account of their non-effervescence in cold acid.

In 1877 Dr. H. Carrington Bolton read a paper urging the use of organic acids in the examination of minerals,* and in this and subsequent publications he has described a series of very successful experiments, showing that in particular citric, tartaric, and oxalic acids effect decompositions for which hydrochloric acid has generally been thought necessary. Citric acid may

thus be carried about in a solid form, a saturated solution in cold water may be made at any time, and the ordinary tests for the presence of carbonic anhydride, or sulphur in certain sulphides, may be performed with this, hot or cold, in a test-tube. Some silicates are decomposable, with or without gelatinisation, and in many cases the solution does not require to be heated. Ordinarily a rather longer time must be allowed for the action of the acid than is the case with hydrochloric acid.

Application of Organic Acids to the Examination of Minerals," Annals New York Acad. of Sciences, vol. i. (1879), p. 1. See for this and later work Chemical News, vols. xxxvi., xxxvii., and xliii.

Dr. Bolton has tabulated his results with citric acid, which is the most useful reagent;* he employs also a boiling solution of citric acid to which sodium nitrate is added, and imitates the reactions of hydrochloric acid by introducing iodine in the form of potassium iodide, which is decomposed by the hot citric acid. + The value of these results obviously consists in the fact that the reagents are solid, and are dissolved only as required.

From the series of minerals examined we may quote the following:

DECOMPOSED IN FINE POWDER BY A SATURATED SOLUTION OF CITRIC ACID.

A. Without evolution of gas.-Brucite (cold solution). Gypsum fon boiling).

B. With evolution of carbonic anhydride.-Calcite and aragonite easily in cold solution; dolomite and ankerite far less easily; chalybite and magnesite only on boiling. It must be noted in testing for carbonates with hot citric acid that the oxides of manganese (hausmannite, pyrolusite, manganite, psilomelane, and wad) evolve carbonic anhydride by decomposition of the citric acid; but the character of these minerals is not likely to allow of confusion with carbonates.

C. With evolution of sulphuretted hydrogen.-Galena, zincblende, and pyrrhotine, in cold solution. Iron pyrites resists until boiled with citric acid and sodium nitrate, when it readily decomposes, whether in the cubic or rhombic (marcasite) form. Copper pyrites requires similar treatment.

D. With separation of silica.-Nepheline, analcime, stilbite, and wollastonite yield silica in a cold solution (long standing is desirable), becoming partially decomposed; natrolite and hemimorphite are decomposed with gelatinisation. Serpentine and ilvaite are decomposed only on boiling, without gelatinisation.

Olivine, augite, epidote, almandine, and hornblende (slightly) have been decomposed by boiling the solution and adding potassium iodide.

E. Among the minerals that are not decomposed by the above attacks, we may quote diopside, asbestos, zircon, idocrase, zoisite, the micas, leucite, sphene, talc, the felspars, barytes, celestine, and anhydrite.

* Chemical News, vol. xliii. (1881), p. 40.

+ See also Chem. News, vol. xxxviii. (1878), p. 169.