directed flame. The separated heavy minerals can now be cleaned by washing.

Messrs. Penfield and Kreider (Amer. J. Sci., vol. xlviii., 1894, p. 143, and vol. 1., 1895, p. 446) describe improved apparatus for conveniently effecting the separation.

Another but more risky method of isolating particular minerals from the powdered rock consists in the use of acids. It is easy thus to extract the silicates from cipollino or calciphyre by destroying the surrounding carbonates with acetic acid; but stronger acids are likely to produce surface-decomposition of the minerals that are to be ultimately examined. It is obvious that the nature and strength of the solvent used in each instance must be left to the judgment of the observer.

M. Fouqué employed hydrofluoric acid in the isolation of the minerals of the lavas of Santorin. He placed about 30 grammes of the rock-powder, from which the finest and the coarsest particles had been sifted off, in a platinum dish into which concentrated hydrofluoric acid had been poured. The materials were inserted cautiously and stirred together; the process of decomposition was arrested at any required stage by pouring in water and washing off thus the fluosilicates, fluorides, and gelatinous products that had been formed. The materials, when washed, should be rubbed with the finger under water to free them from the last traces of the jelly.

In this way the amorphous glassy matrix may be removed from around many minerals, though it may be difficult to free felspars completely from it without seriously attacking the crystals. The ferro-magnesian minerals are attacked only after long immersion; hence they can be isolated from quartz and felspar with comparative ease. The acid is thus found to attack first the glassy matter, then the felspars, then quartz, and lastly the ferro-magnesian group (pyroxene, amphibole, olivine) and magnetite.

The determination of the proportions in which particular minerals are present in a rock can of course be effected by weighing the original powdered material and the successive groups of isolated constituents. Delesse † long ago employed a rougher method, which is simple and very reasonably accurate. It is thus of especial value to observers far removed from refined apparatus. Delesse chose a plane or even polished surface of the

*Fouqué and Lévy, Minéralogie Micrographique, p. 116.

+"Procédé mécanique pour déterminer la composition des roches." Annales des Mines, 4me. sér. tome xiii. (1848), p. 379. Published also with trifling variations at Besançon, 4to.

rock, or in special investigations the six surfaces of a parallelopipedon cut from it. He covered each such surface with a sheet of goldbeater's skin or fine paper, increasing the transparency if necessary by soaking the covering and the face of the rock in oil. The covering was affixed with gum.

The outlines of the minerals were then traced through with a pencil or fine pen, and the various minerals were coloured with different tints. The tracing was removed from the rock and gummed to a sheet of lead or tin-foil. The outlines were cut round with a pair of scissors and the pieces of the same tint were sorted together. To avoid errors due to irregular thickness of the gum and paper, each sorted group was treated in water and the fragments of the foil were alone finally used.

These groups of fragments were then weighed and compared with the total weight of foil that corresponded to the area or areas of the rock selected, the proportions of each mineral being thus ascertained. Delesse found it convenient to estimate fine lamellar minerals, such as mica, by difference.

When a good balance is at hand, the paper may probably be cut out and estimated directly, without transference to the foil. The modern method has been to employ a thin section of the rock under the microscope, to draw the field seen with a cameralucida or neutral tint reflector, and to weigh the various parts of the dissected drawing upon a chemical balance.

One of the most interesting results of such investigations is that mentioned by Delesse, who found that minerals of a striking or rich colour are present in much less proportions than the appearance of the rocks containing them seems to indicate.

CHAPTER XV.

THE PETROLOGICAL MICROSCOPE AND MICROSCOPIC

PREPARATIONS.

THE microscope may be regarded as the one expensive piece of apparatus in the otherwise modest equipment of the geologist; but a good instrument will obviously last a lifetime. While the details of the stand required have been made the special care of certain well-known makers, it is possible to procure first-class objectives second-hand, and to fit them as one's needs extend.

The essential points of the microscope used by geologists are as follows:

(1) A good polariser and analyser, both so fitted as to be almost instantaneously brought into position or again removed; the analyser may be above or below the eye-piece, the former, or "eye-piece analyser," being most suited for observations with the quarter-undulation plate or the quartz wedge when crystals are studied in convergent light. The outer flange of the polariser, and of the analyser if this also rotates, must be graduated at every 90°, so that the position of "crossed nicols" can be easily found. In this position, when the shorter diagonals of the calcite prisms constituting the nicols are at right angles to one another, the field of the microscope should be dark until some crystalline substance is placed above the polariser.

(2) Either the stage of the microscope must rotate, or the two nicols must be arranged so as to rotate together, as in the remarkable instrument now made by Messrs. Swift & Son, at the suggestion of Mr. Allan Dick. In either case, the orientation of any crystal in a section with regard to the diagonals of the nicols must be ascertainable by means of a graduated circle and an index. If the stage rotates, as in most instruments, its edge is marked off in degrees (fig. 14).

(3) There must be cross-wires or "spider-lines" in the eyepiece, and a pin must project from the eye-piece and fit into a slot in the main tube, so as to prevent any rotation of the wires, which are parallel to the diagonals of the crossed nicols.

In addition, an easily removable achromatic condenser should be fitted in the aperture of the stage above the polariser, so as to converge the light upon any crystal brought into the

centre of the field.* Some means of focussing it within a short range should be supplied, and it is a great convenience when working with high powers if it can be retained as an ordinary condenser whether the polariser is in use or not. If the figures viewed in convergent polarised light are to be seen with the eyepiece, a lens of suitable focus is inserted when required into a slot above the objective.

A quartz wedge, the longer direction of which is parallel to the vertical axis of the crystal from which it was cut, is a very useful accessory. Such wedges are sold for about 20s., and should show, when placed between crossed nicols, a regular gradation of colours from the grey of the first order of Newton's scale at the thin end up to the pink and green fourth-order colours at the thick end. These colours should not, as in wedges of too steep an angle, be crowded together towards the thinner end.

In other points the geologist's microscope resembles the ordinary instrument. A nose-piece carrying two or three objectives is invaluable. Owing to the limits imposed by the thickness of rock-sections, very high powers are out of place. If only two objectives are at first purchased, there is little doubt that they should be those styled 2-inch and 1-inch. If a series is available, the following are recommended:-2-inch, 1, 1, 1, and

or-inch. The rack used in focussing should be long enough to allow of the use of a 3-inch objective, which is occasionally required, particularly when a slide illustrating rock-structure has to be studied. The higher powers are necessary for the study of the groundmass of rocks and for use with convergent polarised light.

If a rotating stage is used, some form of centring is desirable. The adjustment may be made by two screws in the collar into which the objective fits, or similarly by an arrangement beneath the stage itself In all cases it is necessary that an object viewed with an inch objective should remain in the field throughout a rotation of the stage, and should, if placed on the intersection of the cross-wires, barely deviate from that position.

The difficulties arising from this petrological requirement have been met by microscopists in two brilliant and different manners. Nachet of Paris divides the main tube of the instrument in two, supplies a double arrangement for focussing, and carries the objective with its adjustment on a pillar attached to the rotating stage. Thus the centre of the objective is always in precisely the same relation to any object in the field, since it rotates with the slide itself. This system has also been successfully adopted by Messrs. Crouch & Co., of London.

A 4-inch objective, supported inverted under the slide and above the polariser, serves to produce the characteristic figures.