has earned for it the name of Walker's balance, and it remains at present the most convenient and portable instrument of which the geologist can avail himself.

The steel bar, A, in fig. 5 is supported in the rest, B, by a knife-edge piece fixed through it about 3 inches from one end. The remainder, some 18 inches long, is graduated into inches and tenths, starting from the point of support.

The short arm of the bar is notched upon its upper surface, and a heavy weight, C, can thus be hung from it at a variety of distances from the fulcrum.

The long arm passes through a looped upright, D, which checks undue swinging, and, by a mark scratched on it, serves to indicate when the bar comes to a horizontal position.

The specimen, which may weigh several ounces, is hung by a cotton thread, a loop of which passes over the long arm. It is then slid along the arm until it counterbalances the weight C, which has been suspended near to or far from the fulcrum, according to the weight of the specimen used.

When the bar indicates by its swing that it would come to rest in a horizontal position, the reading a is taken; i.e., the distance from the fulcrum of the point of suspension of the specimen.

The weight C is kept in the same position, and the specimen is immersed in a tumbler of water; to restore equilibrium, the specimen must now be carried farther out along the beam. Let this new position be b. Then, a and b being, by the principle of the lever, inversely proportional to the weights in air and water respectively, G

=

b b

α

The results are accurate to the first place of decimals, and often compete with the ordinary balance in the second place; while for mineral or rock specimens of a fair size they may be held to be entirely satisfactory.

The earlier forms of the instrument had a spare hammer-head as a weight, a shaft being also supplied. This hammer might be

used and worn down without affecting the value of the results, since all we require is that C should be the same in the two experiments made upon any one specimen. The division of the bar into centimetres and millimetres will give more delicate readings and also a useful scientific scale. The supports are made to unscrew from their bases, and all is packed away into a light box, which for travelling can be reduced to a baize wrapper with pockets, such as is often used for tools.*

The somewhat similar balances devised near the beginning of this century present several ingenious features, but involve greater difficulties in manufacture. Thus Lukens † used an equipoised beam, suspending the specimen from the shorter and thicker arm, and running a weight, which might be a smaller suspended specimen, along the other and graduated arm to restore equilibrium. Coates proposed a similar beam, but introduced a graduation that enabled the specific gravity to be read off without calculation. "The shorter end is undivided; but on the longer is inscribed a scale, of which every division, reckoning from the extremity of the lever, is marked with a number, which is the quotient of the length of the whole scale, divided by the distance of the division from the end. Thus at half the length is marked the number 2, at one-third 3, &c. Also at two-thirds the length is marked 1, at two-fifths 21, &c., the pivot of the instrument represents unity, and a notch is made at the further end." Any convenient weight is hung by a hook from this notch, A; the specimen is slung from the other arm by a horse-hair or thread and slid along till equilibrium is attained. The reading A B, where B is the fulcrum, is obviously constant for all experiments. Immerse in water; the small weight must now be slid in from A towards the fulcrum B; let this reading in water be CB; then A B A B G= The graduation adopted gives this result. AB-CB A O' Ο at once, for we have only to read the figure coincident with the point C.

=

Mr. Roswell Parish § has described a balance resembling in some points that of Coates. Two pans are hung one above the other from a fixed point on one arm of the beam, the lower pan.

* Walker's balance is made by Mr. G. Lowdon, Reform Street, Dundee. Price 31s. 6d.

+ Philosophical Magazine, vol. lviii. (1821), p. 108. From Journ. of Acad. of Nat. Sciences, Philadelphia, vol. i., Part 2.

+ Ibid., p. 109. From same source.

American Journ. of Science, ser. iii., vol. x. (1875), p. 352.

being immersed in water. The beam is then equipoised by a small sliding weight, clamped by a screw, working on the arm that bears the pans. The specimen is laid in the upper pan, and balanced by the addition of a light pan, into which sufficient sand is thrown, suspended from the point corresponding to A in Coates's instrument. The specimen is now transferred to the lower pan, and the balancing-pan is slid inwards, care being taken not to disturb the sand. The reading now made gives the specific gravity without calculation, the graduation being on the plan employed by Coates.

One merit of this instrument is that fragmentary materials can be determined, as no suspending thread is required; but in practice it is probable that the results obtained by it are not superior to those given by Walker's balance, while it is more complicated in construction.

Prof. Jolly's spring-balance or Federwage is, however, simple and yields excellent results. A long brass spiral spring, which may be exchanged for one of greater delicacy if the specimen is exceptionally small, is hung from a sliding rod, set in a pedestal some 3 feet high. One end of the spring may thus be brought 5 feet above the table. The base of the instrument is pierced by three levelling-screws, and a long slip of looking-glass, with even graduations marked on it, is let into the face of the pedestal. Two light pans are hung, one below the other, from a wire hooked to the lower end of the spring, and on the wire is fixed a little bead, acting as an index.

The lower pan is sunk well in a tumbler of water, the support of which can be slid up and down the pedestal; and the slidingrod is carried so high that the pans come to rest somewhere opposite the upper divisions on the graduated mirror. Looking along the top of the index-bead until it appears to coincide with its image in the mirror, the position of rest, a, of the spring is noted, in terms of the fine graduations used. It will be seen that this reading corresponds to the determination of the sinking-weight of the aræometer, only in this case the figure will vary according to the adjustment of the spring at starting.

Place the specimen in the upper pan, having previously drawn the tumbler to a lower position to avoid the wetting of both pans. Readjust the tumbler until the pans swing freely and as much of the lower suspending-wire is immersed as before. Take a second reading, b; then b - a = w, the value in air. Transfer the specimen to the lower pan, and readjust. The

new reading, c, will be less than b, and G

w

=

Though not so suitable for travellers, this makes an admirable laboratory-instrument,* and, the readings being merely proportional, the utility of the spring as a weight-measurer is not affected by expansion due to change of climate.

We must conclude the present section with an account of the use of dense liquids in determining the specific gravity of mineral particles. If a solution of known density is to hand, and a specimen, though it has been completely freed from bubbles, floats upon the surface, while others sink with more or less rapidity, some idea of their relative specific gravities may be obtained.

Further, if the liquid is diluted until a particular specimen swims about in it and remains sluggishly wlierever it is placed, the liquid and the mineral will be of the same specific gravity.

That of the liquid may be determined by throwing in a series of specimens already determined, until one is found that will neither float nor sink to the bottom; or by suspending a weight from a chemical or Jolly's balance, and comparing the readings given when it is immersed in water and in the liquid respectively. Prof. Sollas ("Granites of Leinster," Transactions of the Royal Irish Academy, vol. xxix., 1891, p. 430) has even employed a minute hydro

meter.

This method of determining specific gravities, which can be used even in the case of very small specimens, was brought into prominence by Mr. E. Sonstadt + as recently as 1874, and has since been largely utilised.

Sonstadt's solution consists of a saturated solution of potassium iodide in water, in which is stirred up as much mercuric iodide as it will dissolve. "It will then dissolve more iodide of potassium, then more mercuric iodide, and so forth. The iodides dissolve very slowly at the last, and as it is best not to accelerate the solution by the application of heat, considerable time must be allowed when a liquid of maximum strength is required. The solution, after filtering, is fit for use. It may be diluted to any extent, and then concentrated by heat, without injury." The maximum density obtainable falls just short of 3.2, and is about 3.17 in hot climates, these figures being higher than those first given by Sonstadt.

In addition to its use in determining specific gravities, Sonstadt pointed out that his solution would serve to separate Supplied by Krantz, Rheinisches Mineralien-Contor, Bonn, at 378. + "New Method of taking Specific Gravities,” Chemical News, vol. xxix., p. 128.

mineral particles of one kind from others with which they might be mixed, as in the case of diamonds occurring in quartz sand. This application has been so far extended by Thoulet in France, and Goldschmidt in Germany, that the solution has often been named after these workers instead of after its original dis

coverer.

*

Rohrbach's solution of iodide of mercury and iodide of barium has a density as high as 3-588, but decomposes on addition of water, and must be reduced to the density required by a specially prepared dilute solution. Neither of the foregoing liquids are satisfactory for the traveller, or even for laboratory use, on account of their dangerously corrosive and poisonous character. They have been largely superseded by the solution of borotungstate of cadmium, first prepared by D. Klein,† and now very widely used. This is also a pale yellow liquid, with a density of 3.28; it can be diluted with water and again concentrated by heating over a water-bath until a hornblende crystal just floats upon the surface. Any overheating will cause the salt to crystallise out on cooling down, when a fresh dilution will be necessary. Though poisonous, the borotungstate is not irritant like the mercury solutions; it can be carried about in a stoppered bottle in the solid state, and dissolved in distilled water when required. A few ready-made solutions of known density, kept carefully stoppered, will be very useful in the discrimination of gems. The only objections to this liquid are that it decomposes carbonates, so that specimens before use should be treated with a mild acid; and that it tends to crystallise readily upon the stoppers of bottles or the glass rods used in stirring. The rods and vessels used should always be washed with distilled water, the resulting very dilute solutions being kept together in a bottle, to be concentrated by evaporation when time allows.

Another liquid that is of great utility has been brought forward by R. Brauns.‡ He uses methylene iodide, which must be diluted with benzene and not with either water or alcohol, and which, to preserve its pale straw-colour and transparency, must be kept as much as possible from the light. When it has become darkened, as must eventually happen, the colour can be restored by putting a few globules of mercury into the bottle and

Neues Jahrbuch für Mineralogie, &c., 1883, p. 186.

+ Comptes Rendus, tome 93; August 8, 1881. The solution, at maximum density, is sold by Marquart, of Bonn, at about £3 per kilogramme. On its manufacture, see W. Edwards, Geol. Mag., 1891, p. 273.

+ Neues Jahrbuch für Mineralogie, &c., 1886, ii. Band, p. 72. The liquid is sold by chemical dealers at about 4s. per oz., three or four ounces being a fair quantity to begin with.