is incorrect. It needs at least an amount of saltpetre equal to that of the ore when clean, and if the slag is not very liquid, a part of the gold will remain in it. The method of roasting or oxidising the pyrites, and removing the oxide by washing, is so imperfect that it deserves no further notice. If gold is contained in metals, such as refuse iron, tin, zinc, brass, or copper, under the form of works of art, it is scraped off with a scraper or file. This labour may be accelerated by oxidising the metals at a low heat. The gold thus obtained, as well as jewellers' sweepings, are either melted with saltpetre, or with saltpetre and lead; in the latter case, some lead is obtained which contains all the gold, and may be cupelled in the usual manner. In this case tin or zinc is frequently present in the refuse, and as the oxides of these metals do not melt in saltpetre, it is advisable to add some common salt to the assay, to remove them from the lead, for these render the operation of cupelling slow and tedious; and if zinc or antimony is present, some of the gold is carried off by the evaporation of these metals. If sufficient saltpetre is used, and a little common salt, there will not be any other metals than lead and gold or silver in the button. It has been proposed to free gold from other metals by means of sulphuret of antimony; but this operation, which is generally practised by jewellers for refining their gold, is not applicable to an assay, because it is never correct. It has been also proposed to oxidise artificially all the metals which may be combined with gold, by using black manganese or oxide of iron. This method is not sufficiently correct to deserve any notice; for in all instances the oxides which are formed contain gold in small particles which never can be recovered. The true principle of an assay for gold, in all cases, is to form a very fusible slag, which absorbs all the metals; and as the gold will adhere readily to melted lead, it should be brought in contact with that metal in a state of fusion, and all the particles will be obtained. Common salt, or borax, remarkably promotes the solubility of metallic oxides, and in fact dissolves them very readily. In all cases, therefore, where we have to deal with metals or metallic oxides, either one or the other should be present; but, as borax causes other metals than lead to be formed, and salt produces only the precious metals and a limited quantity of lead, the presence of the latter is, in all cases, preferable to borax, when the precious metals are the object of the assay. There is no substance which dissolves metals more readily and converts them into oxides, than chlorine, and when this is present in a sufficient quantity, the precious metals are not excepted; but when metallic lead is once produced in a crucible, it is only necessary to bring the dissolved gold in contact with the lead, and it is absorbed by it. Metals and Gold.-When other metals are alloyed with gold, we separate the first either by cupellation, or if the kind and quantity is not suitable, we convert the first into oxide, and divide the latter. The alloy of lead and gold is the most suitable for cupellation, and it is in all cases advisable to remove all other metals and convert the alloys into this combination. As gold is very volatile, these conversions must be performed by as low a degree of heat as possible. The cupellation of lead and gold is perfectly safe and easy: this is not the case with other metals and gold. Copper and gold combine very readily, and in fact more so than lead and gold, but this alloy cannot be cupelled: it is therefore melted with lead, and the latter causes the oxide of copper to be absorbed by the cupel. The operation is uncertain, for the oxide of copper will absorb gold and carry it off. Such an assay is never correct, the refined gold containing copper, and the oxide of copper, gold. The same is the case with all other metals, and it is an object of peculiar importance not to cupel a test wherein other metals or substances than lead are contained. Other alloys than those of lead should be dissolved in muriatic acid, and precipitated by carbonate of soda, and the precipitate smelted like a mineral. In smelting, an excess of litharge is used; the quantity of lead is determined by the carbon which is added. An alloy of copper and gold requires a large quantity of lead for cupellation. We may assume that 100 parts of lead to one of copper is not too much, considering their relation, and the uncertainty of success in such an assay. It is easily understood that the solution of the alloy, and its precipitation with an alkali, is the only safe way of proceeding. As an alloy of gold always contains some silver, and as gold is not soluble in muriatic acid, the remaining parts which do not dissolve in the acid must be smelted with the precipitate. There would be no necessity of precipitating the solution, and merely smelting the residue would answer the purpose, but for the circumstance that chloride of silver is soluble to a certain extent in other or soluble chlorides and as in most cases the amount of foreign metals is small, this method does not increase the labours of the assayer. : If the quantity of a metal in a test specimen is large, it is advisable to make a sulphuret of the alloy. This is accomplished in melting the specimen, by adding gradually sulphuret of iron, coarsely pounded. The native pyrites contain sufficient to produce this effect on a large quantity of other metals. Galena may be used also, but it is not so effectual as iron pyrites. If neither iron nor lead pyrites can be obtained, pure sulphur is used, which is added to the melted or heated metal in small quantities and in the form of a coarse powder. The resulting sulphuret is then treated as native sulphurets, as has been stated in former pages. Platinum, Silver, and Gold.—The presence of the first of these metals causes the cupellation to work slowly, and in some cases, particularly when copper is present, the gold does not form a globule at all, but is carried with the oxide of copper and the platinum metal over the cupel. In such cases we melt the test with some silver, of which at least twice the amount of the gold alloy must be added, and also some lead to make the mass fusible; it is then cupelled in a strong heat, to obtain the metal in a perfect globule. The quantity of lead is variable; if the heat in the cupel is strong, which is the most advantageous, the quantity of lead may be 12 or 14 times that of the alloy. If the heat is low, but sufficient to melt the alloy, the quantity of lead must be at least 20 times that of the alloy. In all cases the resulting refined metal is adulterated by copper, that metal is present, which is more likely to be the case if the heat has been low in the cupellation. If no copper is present, less lead may be used in refining the metal. The presence of platinum in a test is recognised by the globule, which assumes a flattened instead of a round form,-this is pre-eminently marked in pure gold; silver, or an alloy with much silver, is also liable to form a flat globule in the shape of a hemisphere, while that of platinum is distinctly flattened. One per cent. of platinum will cause the globule to be rough and rugged on the surface, while pure gold and silver is perfectly smooth and shows a mirror-like lustre. In all cases the presence of platinum causes the globule to be dull and lustreless, often showing crystals on the surface. When the alloy contains more than ten per cent. of platinum, the cupel is covered with a silver coating, under the influence of a strong heat, which is often yellowish or grayish, and consists of platinum metal. The Moist analysis.-When a globule of gold is obtained in the process of cupellation, it is never pure, particularly when derived from a mineral. globule, for the purpose of testing it, is flattened on a steel anvil with a hammer; at first it is placed directly on the naked steel, and afterwards between strong and thin paper. It is next exposed to the influence of pure nitric acid, which must be free from muriatic acid, which will not cause a precipitate with nitrate of silver. If the gold is nearly pure it is not attacked by the acid; we melt it then with some pure lead, about equal to twice its weight, and dissolve this alloy. The nitric acid, gently heated, will now hold in solution the lead and silver, and leave gold, platinum, and the platinum metals at the bottom of the vessel. If pure silver is at hand it may be melted with the globule, of which an equal weight is sufficient. The result is more certain than by alloying it with lead. The residue of the solution is treated with aqua regia, which dissolves the metal. The solution thus obtained is evaporated in a flat porcelain dish, over a basin of boiling water, which causes the evaporation of all the acid. The dried chloride of gold is now dissolved in water, and to this is added a fresh solution of proto-sulphate of iron, which in a short time precipitates the gold in |