soonest of them all, so that it becomes necessary to conduct the roasting of the ores in such a manner that cobaltous oxide only is produced, which, on being fused with silicic acid and potash, gives the beautifully colored smalt. The foreign metals should not become oxidized, but remain combined with arsenic or sulphur, forming a cobalt speiss.

As, in roasting the above-named arsenides and sulphides, cobalt is first oxidized, and iron and bismuth earlier than copper and nickel, ores containing copper and nickel must not be roasted too strongly, and least of all dead-roasted; this may, however, be done with entirely pure ores of cobalt, or such as contain only iron, as, in the latter case, ferric oxide is formed, which only slightly affects the cobalt color. If impure ores are roasted too little, a beautiful smalt is formed, but much cobalt is lost in the speiss.

The object of the smalt assay is either, a, the determination of the coloring power of a sample (assay for determining the intensity of the color), or, b, to acertain the degree of roasting to which the ore must be subjected in order to obtain a color of pure quality (assay to determine the color tone), or, c, to determine how much of an already known sample must be taken to produce a certain shade of color.

Several lots of ore, each from 1 to 5 grammes (15.43 to 77.16 grains) according to the richness of the ore, are weighed off. The separate lots are each roasted for a different length of time (for instance, the first lot of an hour, and the subsequent lots each from 10 to 15 minutes longer), while one of the samples is left unroasted. Each sample is divided by weighing into two equal parts, of which one is tested with fluxes to determine the quality of color, and the other its intensity.

A. Assay to determine the quality of color.—Each sample is mixed with three times the quantity, by weight,

of quartz free from iron and manganese, and with a quantity of pure potash corresponding to half the combined weight of quartz and ore. The mixture is placed

in flat dishes of white fire-clay (smalt-dishes). It is then fused in the muffle (Fig. 25, p. 51), which should be heated as strongly as possible, until a completely homogeneous glass has been formed (this will require 4 hours or longer). A sample is then taken from the charge by means of a pair of tongs and cooled off in water. The mass is dried and pounded in a bright steel mortar, in order to obtain angular fragments (powdered smalt is apt to appear dirty). The fragments are sifted upon white paper, and, without taking intensity into consideration, a judgment is formed at which degree of roasting the most beautiful color has been obtained.

B. Assay to determine the intensity.-The assay sample is fused in the above manner with different quantities of quartz (for instance, 1 to 10 times the quantity), and half the quantity of both of potash, to a homogeneous cobalt glass. Generally 2.5 grammes (38.59 grains) of the assay sample are taken when it has been mixed with once to twice the quantity of quartz, and 1.25 grammes (19.29 grains) if with more quartz, to prevent the crucibles from becoming too full. Too much quartz makes it difficult to fuse the mixture, and too much potassa gives smeary colors. A sample is taken from the smelted mass, cooled off, and dried. It is then pounded and sifted, or washed in spitz-glasses (Fig. 3, p. 28), and its color compared with that of a standard (muster), as to color and grain. This is done by spreading some of the standard color with a knife evenly upon a board, and placing a quantity, as large as a pea, of the assay sample upon it and pressing the latter into the former, when an experienced eye can tell in a well lighted room (not exposed to the direct

rays of the sun) whether the assay sample corresponds with the standard in color-tone and grain, or not. If it should correspond in all respects, a confirmatory test is made by pressing some of the standard into the assay sample, when the same conditions must exist, and by examining the grain in both cases with a magnifying glass.

As moist smalt appears darker than dry, the standard and assay sample must stand for from 6 to 8 hours alongside of each other in a somewhat damp place, before the examination is made. The intensity of the color also increases with the coarseness of the grain, and an experienced eye and skilful hand are required to give the assay sample taken from the fused mass, the same grain as the standard with which it is to be compared. If the color of the assay sample is lighter than the standard with which it is compared, the assay must be repeated with a larger quantity of ore, and vice versa. If the color is not exactly the same, recourse is frequently had to mixing the product with lighter or darker varieties of smalt. An ore is the more valuable, the more quartz it requires for the production of a certain intensity of color.

VIII. ZINC.

45. ORES.

Smithsonite (zinc carbonate), ZnCO3, with 52Zn; calamine (zinc silicate), Zn,SiO+3H2O, with 53.7 Zn; willemite, Zn,SiO4, with 58.1 Zn; zinc bloom, Zn,CO¿+ 2H2O, with 57.1 Zn; zinc blende, ZnS, with 67.01 Zn; zinkite, ZnO, with 80.24 Zn; franklinite, (Zn, Fe) (Fe, Mn,)O, with 21 Zn.

46. DRY ASSAYS.

These are inaccurate, but give approximately the quality of the metal which may be expected from an ore (distillation assay), or the approximate percentage of zinc.

For many purposes they are sufficiently accurate (indirect assay).

A. Assay by distillation.-A mixture of 400 to 500 grammes (6172.94 to 7716.17 grains) of the comminuted assay sample with 80 to 100 per cent. of powdered charcoal, and, in case calamine is to be assayed, with 80 to 100 grammes (1234.58 to 1543.23 grains) of potash or calcined soda, is placed in a retort of refractory clay and heated in a furnace with a strong draught (Fig. 30, p. 56). The neck of the retort should project about 10 centimeters (3.93 inches). In it is luted a tube of glass or porcelain about 30 centimeters (11.81 inches) long, which is kept cool by moistened rags wrapped around it. The burning gases and the fumes escape from the end of the tube, while the zinc, mixed with oxide, which is distilled over, is mostly deposited in the neck of the retort and but little of it in the tube. After the "flaming" ceases, the tube should be frequently poked with an iron wire. The retort, after having been exposed to a white heat for several hours and the "flaming" having entirely ceased, is taken from the furnace and allowed to cool off. The metallic zinc is scraped from the neck of the retort, is placed in a crucible and fused with black flux and a covering of common salt. It is then poured out into an ingot mould and weighed. The retort is carefully broken into pieces, and all those having adhering to them either zinc or zinc oxide are collected together and placed in a porcelain dish, and the metal dissolved off by digestion. with nitric acid. The solution is filtered and evaporated to dryness. The residue is heated to redness, the zinc calculated from the resulting zinc oxide, with 80.24 per cent. of zinc and added to the metallic zinc.

B. Indirect assay.-5 grammes (77.16 grains) of zinc blende are placed in a covered crucible (Fig. 39, p. 66) and

.

heated under the muffle in order to expel the volatile substances. After it has been cooled off, it is weighed and the ignition repeated until a constant weight is obtained. The accurately weighed refractory residue is mixed with 3 grammes (46.30 grains) of iron filings free from rust, and 2.5 grammes (38.58 grains) of blastfurnace slag. The mixture is brought into a charcoallined iron crucible (Fig. 43, p. 67), covered with 2.5 grammes (38.58 grains) of blast-furnace slag, and the remaining space filled up with powdered charcoal. A loosely fitting cover is luted on. The charge is then heated at a white heat in the blast-furnace for to 1 hour, whereby iron sulphide is formed, while zinc volatilizes and the earthy substances fuse together with the iron slag. After the charge has become cool, the button consisting of brittle matt and slag is weighed, and its weight deducted from that of the refractory residue of the zinc blende plus 3 grammes (46.30 grains) iron filings, plus 5 grammes (77.16 grains) of iron slag. The difference will give the percentage of zinc. The fewer the metallic impurities in the blende, which would be separated by the iron and volatilized (such as antimony, lead, bismuth, etc.), the more accurate will this method be.

47. WET ASSAYS.1

Gravimetric and volumetric methods are used. Although the latter assay is somewhat less accurate than the former, it is nevertheless available in many cases for practical purposes.

A Gravimetric assays.

1. Determination of zinc as zinc sulphide.-1 gramme (15.43 grains) of the finely powdered assay sample,

1 B. u. h. Ztg. 1876, p. 148, 173 (Laur).