Samples of goldsmiths' sweepings (scrapings, fragments of crucibles, rags, etc.) are taken by incinerating the entire mass in order to destroy organic matter. The mass is then comminuted by stamping or trituration, and passing it through a sieve having meshes the size of a grain of sand (less than 1 millimeter, 0.039 inch). The iron in the coarse mass remaining in the sieve is extracted by using a magnet, and the residue fused with soda and borax, cast into a bar, weighed, and sampled by chipping from top and bottom. The portion that has passed through the sieve, freed from iron with the magnet, is then weighed, sampled from every weighing, then united, quartered, and triturated until everything passes through a very fine sieve. It is then assayed, and the yield of metal obtained from both coarse and fine material is calculated.

b. Sampling before weighing.-This is done by passing with a hollow sheet-iron cylinder in several places through the heap down to the bottom. The lower end of the cylinder is provided with a valve which when closed retains the charge, and with a handle. The samples are then mixed and reduced as above described (Freiberg).

In some of the German smelting works, the ore,' when it is bought, is weighed

to within 10 lbs. (5 kilog.; 11 lbs.) when it contains up to 0.5 per cent. Ag, or 0.01 per cent. Au.

66

66

66

1 lb. when it contains over 0.5-5 per cent. Ag, or 0.0105-0.1

per cent. Au.

0.1 lb. when it contains over 5-50 per cent. Ag, or 0.1005–1

per cent. Au.

0.02 lb. if it contains more.

10 lbs. if the ores contain no gold or silver.

3. Substances in a state of fusion.—An average sample may be obtained in the following manner: A dry tappingbar, previously heated, is held in the fluid mass (slag,

1 Bezahlungstarife für den Einkauf von fremden Erzen u. Gekrätzen auf den fiscalischen Hüttenwerken bei Freiberg. Clausthal, 1870. Oestr. Ztschr. 1869, No. 44. Engl. Standard f. Zinnerze, B. u. h. Ztg. 1862, p. 261; f. Kupfererze, 1862, p. 316. Spanische Tarife, B. u. h. Ztg. 1868, p. 26. Tarnowitzer Erztaxe in Preuss. Ztschr. xiv. 217.

etc.). When the bar has become cold, the adhering mass is broken off, comminuted, and then mixed and reduced. In this operation, however, the iron must not decompose the fused material (a possible separation of lead may occur, for example, from lead matt and lead speiss).

B. Alloys.-These are homogeneous when in a fluid state, especially after they have been stirred. But when solidified, they show a different composition in different places (edge, centre, top, bottom). In sampling, this must be taken into consideration.

1. Sampling by cutting.-The sample (2.5 grammes (38.59 grains) from every ingot of silver, 1.5 grammes (23.15 grains) from those of gold) is cut from the upper and lower sides on opposite ends of the ingot by means of a hollow chisel and hammer. (In England and the United States, the opposite edges are chipped off.) The samples are hammered or rolled out, the resulting sheets cut into shreds, and each sample is assayed by itself (0.5 to 1 gramme, 7.71 to 15.43 grains, of each is weighed off for the purpose). The yield is calculated, and either the average is given (gold assays), or the lowest yield (sometimes in silver assays). This method is best adapted for alloys of tolerably uniform composition, but is also employed for those showing a considerable difference in the lower and upper sampling.

For instance, the lower sample from refined Upper Harz silver is from Too to Too richer than the upper, the percentage of gold increasing towards the bottom. The centre, as a rule, contains more silver than the edge. In the "five-mark piece" the centre is Too richer than the edge, and the same is the case with the "thaler," as they are stamped from a bar poorer on the edge than in the centre. For this

' Kerl, Grundr. d. allgemeinen Hüttenkunde, 2 Aufl., 1879, p. 15. Dingler, cciii. 106; ccxv. 431. Ber. d. deutsch. chem. Ges. 1874, p. 1548, B. u. h. Ztg. 1874, p. 63; 1875, p. 251.

reason, when taking samples from such coins, it is best to cut out a quadrant, cut off the corners, and assay them. In this way the assay samples represent the composition of both periphery and centre of the bar from which the coins have been stamped. Fewer differences occur in gold than in silver coins.

2. Sampling by boring.'-This is done by boring through the edge and centre of the ingot. By this means a sample is obtained from the centre of the ingot, which is not the case in chipping a sample, but the ingot is made unsightly. It is very difficult to mix the borings uniformly, and it is therefore better to fuse them under a covering of charcoal powder. It is best to use a mechanical contrivance for boring through thick pieces. This consists of a lever weighted at one end, and a drill, operated by the hand in the centre.

3. Sampling by dipping.―This is obtained in the same manner as mentioned on p. 22, for instance, from refined copper. Another method of taking samples is, by dipping the curved, bright end of a pair of pincers, or of an iron rod, into the metal bath. When the pincers or iron rod has become cold, the crust adhering to the end is broken off. A sample is generally taken from the surface of silver while it is being refined, and one from the under side of the congealed refined silver.

1

4. Sampling by granulation.-This sample indicates in the most reliable manner the average value of the metal. It is obtained in the following manner: the bars of precious alloys are smelted in a black-lead crucible. The mass, while in fusion, is stirred with a rod of iron or clay, and a sample is scooped up with a small ladle from the bottom of the crucible. It is then poured in a thin stream into a copper vessel filled with warm water, to

1 Mitchell, Prac. Assaying, 1868, p. 565.

which a gentle rotary motion is given by means of a broom; or the sample is directly poured through a birch broom. The resulting granulated metal is then carefully dried. Alloys of base metal (as granulated lead') are fused under coal-dust and then directly poured upon an iron plate.

The following samples are taken for producing coins: ingot sample from the metals to be alloyed; granulated or crucible sample from the fused alloys; stock sample from sectors of the finished coins; and a sample from defective coins which have been thrown out and fused together, for instance inside of four weeks.

3. PREPARATION OF THE SAMPLE.

Alloys are prepared by rolling out and cutting up the resulting sheets; or the granulated metal is used without further preparation. The following operations may be required for non-alloys.

1. Determination of moisture.-The sample is divided, by weighing with reduced weights, into as many centners or kilogrammes as are actually contained in the lot (for instance, in Freiberg, 1 centner weight =75 grammes or 1157.87 grains). The weighed portion is heated in an iron or copper pan or directly in the removable scale-pan of the balance, by holding it over a heated stove or a brazier of charcoal, and constantly stirring it until a cold plate of glass or slate, when held over it, shows no deposition of moisture, and two successive weighings agree. The heating should be carefully conducted, so that, with sulphur compounds for example, no odor of sulphurous acid shall be developed, and, with organic substances, no carbonization shall take place (this may be guarded

1 B. u. h. Ztg., 1869, p. 278 (Brixlegg).

Fig. 1.

against by holding a piece of paper in the mass). Waterbaths are used for drying the sample at 100° C. (212° F.). These are copper boxes, with double walls, the intermediate space (Fig. 1) containing water; or, they consist of hemispherical copper or enamelled iron vessels, placed one within the other, leaving an intermediate space for water; or of two cylinders, one placed within the other (Scheibler's1 steam apparatus). Where a determi nate temperature, high or moderate, is required, it is best to use Fresenius' drying disk of cast iron, which is heated from below (Fig. 2). It is 25 centimeters (9.84 inches)

in diameter, and 4 centimeters (1.57 inches) thick; b is the handle, 36 centimeters (14.17 inches) high; c are small brass dishes with numbered handles fitting into suitable recesses; d, a case filled with copper-filings for

1 Dingler, ccxxiii. 312. Muspratt's Chemie, v. 1635.