the lower part of the stem being at the same time kept cool by moist strips of linen or paper tied around it. The mercury adhering to the neck is removed by gently tapping and wiping out, and that from the receiver is dried with. absorbent paper and caustic lime, and weighed in a watchglass.

1

Idria: 140 grammes (2160.52 grains) of ore, with two or three spoonfuls of powdered lime, are placed in iron distilling tubes of about 52 millimeters (2.04 inches) diameter, lying in two rows, one above the other, on each side of a furnace. The receivers are luted on, and the process is finished when the tubes show a bright-red heat. The loss of mercury is considerable. Hungarian mercurial tetrahedrite:" The ores are heated with the same quantity of iron turnings-roasted ore at the same time with an equal quantity of lead oxide1-in glass retorts resting upon clay dishes. The neck of the retorts, when the operation is finished, are broken off by a blow, and the mercury is removed by means of a wiper of rabbit fur. It is then collected into a globule and weighed. Rose's method: A body of magnesite (or chalk with an equal quantity of sodium bicarbonate), 26 to 52 millimeters (1.02 to 2.04 inches) long is introduced into a glass tube, closed at. one end, and measuring 314 to 470 millimeters (12.36 to 18.5 inches) in length, and 9 to 13 millimeters (0.35 to 0.51 inch) in width. Upon this is placed an intimate mixture of the ore and quicklime in excess, upon this more lime with which the mortar has been cleaned off, then more quicklime, and upon all a loose plug of asbestos. The open end of the glass tube is drawn out and bent to an obtuse angle, and introduced into a narrow-necked flask, so that its end just touches the water contained therein. The horizontal part of the tube is gradually heated from front to back in a combustion furnace, such as is used for organic analyses (Fig. 58, without the cylinder on the right). When the operation is finished, the bent end of the tube, in which the mercurial vapors have been condensed, is cut off, and the mercury, which has been protected from oxidation by the current of carbonic acid which is developed, is collected in the matrass. This

B. u. h. Ztg. 1854, p. 394.

2 B. u. h. Ztg. 1854, 357.

B. u. h. Ztg. 1866, pp. 24, 262.

• Bergwerksfreund, v. 127 (Berthier). B. u. h. Ztg. 1879, p. 206 (Atwood).

is well shaken and allowed to settle. The clear water is then poured off, and the mercury placed in a previously weighed porcelain crucible. The water still adhering to it is removed with blotting paper. It is then dried under a bell-glass over sulphuric acid, or in an air-bath at 100° C. (212° F.), and weighed.

B. Assays in which the mercury is determined in combination with goid.-These are the most accurate assays. 1. Eschka's process.-The quantity of ore taken for the assay varies according to its richness. If the ore carries as much as 1 per cent. 10 grammes (154.32 grains), 1 to 10 per cent. 5 grammes (77.16 grains), and over 10 per cent. 2 grammes (30.87 grains) are used. The sample is placed in a porcelain crucible, the edge of which has been ground smooth, and mixed with half its quantity of iron filings free from grease, covered with a layer of iron filings 5 to 10 millimeters (0.19 to 0.39 inch) thick. A well-fitting concave cover, made of fine gold, and previously accurately weighed, and the concavity of which is filled with distilled water, is now placed on the crucible.

Oestr. Ztschr. f. Berg. u. Hüttenwes. 1872, No. 9. B. u. h. Ztg. 1872,

p. 173.

The lower part of the crucible is heated by a flame for about 10 minutes, during which time the mercury is volatilized, and deposits itself on the gold. The gold cover is now removed, the water in its concavity poured off, and the mirror of mercury washed with alcohol. The cover is then dried for about two or three minutes in the water-bath, placed upon a tared porcelain crucible and allowed to cool in the desiccator, and then both the mercury and crucible are weighed together. The most accurate results are obtained in the case of poor ores carrying up to 10 per cent. Hg.

2. Küstel's assay.'-This is executed in a manner similar to the above, before the blowpipe, with the difference that the heating is done in a tube, the front end of which is provided with a gold spiral.

C. Assay of cinnabar.-10 grammes (154.32 grains) of cinnabar ore are introduced into a glass retort (Fig. 35, p. 63), and heated. The sublimate of mercury sulphide which deposits itself in the neck of the retort is collected and weighed. Some metallic mercury freed by organic substances, which may have been present, may be mixed with the sublimate. This is removed from the sublimate by nitric acid, and the quantity of mercury dissolved ascertained by the difference in weight. 86 parts of mercury correspond to 100 parts of cinnabar.

A. Gravimetric assay.-1 gramme (15.43 grains) of cinnabar is heated with aqua regia, and repeatedly evaporated with hydrochloric acid, in order to expel the nitric

B. u. h. Ztg. 1874, p. 70.

acid. The solution of chloride is boiled with stannous chloride in excess, the clear fluid is poured off, and the beads of metal are collected into a coherent globule, by heating it again with some stannous chloride and a few drops of hydrochloric acid. The mercury is then washed by decantation, first with water containing hydrochloric acid, and then with pure water. It is now introduced into a small porcelain crucible, previously weighed, and the greatest part of the adhering water removed by means of filtering paper. The mercury is then dried in a desiccator (Fig. 17a, p. 43), with concentrated sulphuric acid, and weighed.

B. Volumetric assays.-These are mostly complicated, and possess no advantage over the ordinary analytical determination by weight, or, are not generally available, as they require the absence of certain metals.

XIII. ANTIMONY.

60. ORES.

Stibnite (antimony sulphide), Sb2S3, with 71.77 Sb; valentinite (antimony oxide), Sb2O3, with 83.56 Sb; pyrostilbite (antimony oxysulphide), Sb. O,.SbS,, with 77.21 Sb.

61. FIRE ASSAYS.

3

The assays of antimony sulphide, which is the principal ore, are inaccurate on account of the volatilization of antimony, incomplete decomposition by alkaline carbonates and even by potassium cyanide (4 parts), as well as on account of iron, which always passes partly into the

Mohr, Titrirmethode, 1874, pp. 236, 318, 436, 438, 441. Fresenius' Ztschr. ii. 381.

antimony. There are likewise no simple docimastic tests by the wet method. Sometimes the yield of antimony sulphide (antimonium crudum) contained in the ore is ascertained by the liquation process, by an assay on a small scale.

A. Liquation process for determining antimonium crudum.-1 to 1 kilogrammes (2.2 to 3.3 lbs.) of ore comminuted to fragments the size of a hazel-nut or walnut are introduced into a covered crucible having a perforated bottom and fitting air-tight in another crucible in such a manner that sufficient space is left between the two to allow the fused antimony sulphide to collect in the lower crucible. The lid and the joint between the two crucibles should be luted with fire-clay and sand. The lower crucible is surrounded with some poor conductor of heat (ashes), and the upper ore with live coals, kept in a glow with a bellows, and heated to a moderate red heat. The antimony sulphide will melt and collect in the lower crucible.

B. Determination of antimony in antimony sulphide. 1. Assay by precipitation.-5 grammes (77.16 grains) of ore are fused in a crucible with the same or double the quantity of black flux, or potassium carbonate and flour, about 2 grammes (30.87 grains) of iron filings, 0.75 to 1.25 grammes (11.57 to 19.29 grains) of borax-glass, and a cover of common salt. The fusion is continued for three-quarters of an hour. When the operation is finished, the brittle regulus is freed from slag, and the adhering particles of slag are removed by washing. The yield by this assay is about 68 per cent., and if more

'Fresenius' Ztschr. xvii. 185 (Becker's gewichtsanalytische Probe); Mohr, Titrirmethode, 1874, p. 267, 309; Muspratt's Chemie, i. 820; Fleischer, Titrirmethode, 1876, p. 299.