A test known as Darton's is believed to be a valuable means of detecting minute quantities of gold in rocks, ore tailings, etc.
"Small parts are chipped from all the sides of a mass of rock, amounting in all to about ¼ ounce. This is powdered in a steel mortar and well mixed. About half is placed in a capacious test tube, and then the tube is partly filled with a solution made by dissolving 20 gr. of iodine and 30 gr. of iodide of potassium, in about 1½ ounces water. The mixture thus formed is shaken and warmed. After all particles have subsided, dip a piece of fine white filter paper in it; allow it to remain for a moment; then let it drain, and dry it over the spirit lamp. It is next placed upon a piece of platinum foil held in a pincers, and heated to redness over the flame. The paper is speedily consumed; and after again heating to burn off all carbon, it is allowed to cool and is then examined. If at all purple, gold is present in the ore, and the relative amount may be approximately deduced. This method takes little time, and is trustworthy."
Black sand, which is iron, often with some platinum and iridium, sometimes interferes with the result of a gold assay. Attwood recommends the following method as applicable to such a case:
"Take 100 to 1000 grains and attack with aqua regia in a flask; cool for about thirty minutes or more; dilute with water and filter. If gold is present, it will now be held in solution in the filtrate. Remove the filter and evaporate the filtrates to dryness; then add a little hydrochloric acid, evaporate and re-dissolve the dry salt in warm water; add to the solution so formed proto-sulphate of iron; which will throw down the gold in the form of a fine, dark precipitate. The precipitate is seldom fine, being mixed with oxides of iron, and must now be dried in the filter paper, and both burned over the lamp in a porcelain dish. Then mix the dried precipitate with three times its weight of lead; fuse, scorify and cupel. In case platinum, iridium, etc., are found associated with the gold, an extra amount of fine silver should be added before cupellation, and the gold button will be found pure."
In one of his reports the State Mineralogist of California gives a most lucid description of a mechanical assay of gold-bearing sands, stamped ore, etc., etc. He states:
"It must be understood that this is only a working test. It does not give all the gold in the rock, as shown by a careful fire assay, but what is of equal importance to the mine-owner, mill-man, and practical miner, it gives what he can reasonably expect to save in a good quartz mill. It is really milling on a small scale. It is generally very correct and reliable, if a quantity of material be sampled. The only operation which requires much skill is the washing, generally well understood by those who are most likely to avail themselves of the instructions. These rules apply equally to placer gravels. Take a quantity of the ore—the larger the better—and break it into egg-sized pieces. Spread on a good floor, and with a shovel mix very thoroughly; then shovel into three piles, placing one shovelful upon each in succession until all is disposed of. Two of the piles may then be put into bags. The remaining pile is spread on the floor, mixed as before, and shovelled in the same manner into three piles. This is repeated according to the quantity sampled, until the last pile does not contain more than 30 pounds of ore. As the quantity on the floor becomes smaller, the lumps must be broken finer until at last they should not exceed one inch in diameter. The remainder is reduced by a hammer and iron ring to the size of peas. The whole 30 pounds is then spread out, and after careful mixing portions are lifted with a flat knife, taking up the fine dust with the larger fragments, until about 10 pounds have been gathered. This quantity is then ground down fine with the muller, and passed through a 40-mesh sieve. If the rock is rich, the last portion will be found to contain some free gold in flattened discs, which will not pass this sieve. These must be placed with the pulverized ore, and the whole thoroughly mixed, if the quantity is small, but if large must be treated separately, and the amount of gold allotted to the whole 10 pounds and noted when the final calculation is made.
"From the thoroughly-mixed sample, two kilogrammes (2000 grammes) must be carefully laid out. This is placed in a pan or, better, in a batea, and carefully washed down until the gold begins to appear. Clean water is then used, and, when the pan and the small residue are cleaned, most of the water is poured off and a globule of pure mercury (which must be free from gold) is dropped in, a piece of cyanide of potassium being added with it. As the cyanide dissolves, a rotary motion is given the dish, best done by holding the arms stiff and moving the body. As the mercury rolls over and ploughs through the sand, under the influence of the cyanide it will collect together all the particles of free gold. When it is certain that all is collected, the mercury may be carefully transferred to a small porcelain cup or test tube, and boiled with strong nitric acid, which must be pure. When the mercury is all dissolved the acid is poured off, more nitric acid applied cold, and rejected, and the gold is then washed with distilled water and dried.
"The object of washing with acid the second time is to remove any nitrate of mercury which might remain with the gold, and which is immediately precipitated if water is first used.
"The resulting gold is not pure, but has the composition of the natural alloy. Before accurate calculations of value are possible, the gold must be obtained pure and weighed carefully. To purify the gold it should be melted with silver, rolled out or hammered thin, boiled twice with nitric acid, washed, dried, and heated to redness.
"The method of calculating this assay is simple. It will be observed that 2000 grammes represent a ton of 2000 pounds; then each gramme will be the equivalent of one pound avoirdupois, or one 2000th part of the whole, and the decimals of a gramme to the decimals of a pound. Suppose the ore yielded by the assay just described, fine gold weighing .072 gramme, it must be quite evident that a ton of the ore would yield the same decimal of one pound. Now one pound of gold is worth $301.46, and it is only necessary to multiply this value by the weight of gold obtained in grammes and decimals to find the value of the gold in a ton of ore—$301.46 × .072—$21.70. The cyanide solution should be kept rather weak, as gold is slightly soluble in strong solutions of cyanide of potassium. Cyanide is a deadly poison."