These methods are described here for convenience, as being more intelligible after the chemistry of the process has been discussed. The necessity of the presence of oxygen has already been dwelt on above. It has, however, been frequently pointed out that in the interior of a mass of ore undergoing treatment the conditions are not favourable for the maintenance of a sufficient quantity of oxygen in a free state. Both the cyanides and the pyrites of the ore tend to unite with it, and further absorption of free oxygen from the air is extremely slow. Hence the time required for the treatment of a charge is many hours, or even days, although under favourable conditions the gold could be dissolved in a few minutes, or at most in two or three hours. To supply the oxygen, various oxidising substances have been tried, such as the passage of a current of air through the solution, and the addition of various materials, some of which were tried in a unsuccessful attempt to find a process for treating refractory ores. These substances hasten the solvent action of cyanides on metallic gold, and have sometimes been used in practice, especially in the United States, although they act as “ cyanicides,” destroying some of the solvent by direct or indirect oxidising effects.

Engineers suggest that the action of oxygen is due to its strong electro-negative relation to gold in cyanide solutions and has experimented with various materials which are electronegative to gold in solutions of cyanide of potassium. In the list of such materials, given above, are carbon, iron, lead, and mercury. Engineers tried the effect of finely powdered carbon mixed with the ore, and continually agitated with it in a cyanide solution. He found that the gold was more rapidly dissolved than if no carbon had been present, and supposed that some of the particles of gold came into contact, with the carbon and formed galvanic couples, in which the electro-positive element, gold, was quickly acted on. Such contact, however, could not be attained in practice except to a small extent, and the method is therefore useless, whilst the use of lead or iron in this way is a fortiori impossible. On the other hand, mercury can be more readily subdivided and distributed through the ore, but, as it amalgamates with the gold, the conditions are changed, and, as a matter of fact, the gold in pasty amalgam is only slightly more rapidly dissolved by cyanide than is pure gold.

From the fact that mercury is electronegative to gold in cyanide solutions, we concluded that metallic gold in contact with a solution of mercury cyanide would rapidly dissolve and mercury be reduced. He found this to be the case alike with gold and silver, which dissolved with almost equal readiness.

The addition of a small quantity (2 ozs. to 12 ozs. per ton of liquid) of potassium mercuric cyanide, HgCy2. 2KCy, to ordinary cyanide solutions to quicken their action by enabling the presence of free oxygen to be dispensed with. The gold displaces the mercury from solution, and so is dissolved, whilst the mercury is precipitated on the surface of the particles of gold and forms amalgams.

If, for example, gold is digested with an aqueous solution of potassium cyanide and one or other of the chlorides of mercury, the action is represented, so far as weights are concerned, by the following equations:

2Au + HgCl2 = Hg + 2AuCl
2Au + Hg2Cl2 = Hg2 + 2AuCl

These equations do not, of course, represent the whole of the interchanges. The presence of the alkaline cyanide evidently assists in dissolving the gold, for a solution of HgCy2 alone in water has no action on gold. By using an excess of gold, the whole of the mercury can be removed from solution, and equivalents of gold dissolved as represented by the equations given above, 2 x 196.8 parts of gold being dissolved when 200 parts of mercury are added as mercuric chloride or 2 x 200 parts as mercurous chloride. In some experiments made by the author employing 0.5 gramme mercuric chloride and 1.0 gramme potassium cyanide, the amount of gold dissolved by the hot, strong (5 per cent.) solution in thirty minutes was 0.724 gramme, whilst theory requires 0.726 gramme. Under similar conditions, the amount of gold dissolved by potassium cyanide alone was only 0.010 gramme.

Some supposed that the gold would thus be rapidly dissolved, and the precipitated mercury would then be re-dissolved in the cyanide, and thus be ready to react as before. This view seems to be incorrect. According to the author’s experiments, the gold is at first rapidly dissolved, and the mercury precipitated. As the action proceeds, however, the dissolution of the gold becomes slower and slower, the mercury appearing to protect it more and more as the percentage of gold in the amalgam is diminished. If the particles of gold are only moderately fine (e.g., gold precipitated from the solution of the chloride by sulphurous acid) the action becomes extremely slow after about 85 per cent, of the gold has been dissolved, the amalgam then consisting of about three parts of mercury to one of gold. If, on the other hand, very finely-divided gold is used, such as gold leaf, the action is fairly rapid until about 95 per cent, of the gold is dissolved, and in one case 98 per cent, of such gold was dissolved in four days by a solution containing 1.5 per cent, of HgCy2. 2KCy. In the case of gold leaf, however, which contains both silver and copper, the rate of dissolution is higher than it would be for pure gold in a similar state of subdivision, as the presence in the alloy of either silver or copper favours the dissolution. The retarding effect exercised by metallic mercury when amalgamated with the gold is exemplified by the results of some experiments in which the solution contained 1.5 per cent, of HgCy2. 2KCy, the time of treatment was thirty-six hours, and the weights and states of aggregation of the metal treated were approximately the same. Under these conditions, about 86 per cent, by weight of some samples of pure gold were dissolved, and only 14 per cent, of the gold contained in amalgams consisting of two parts of mercury and one of gold. Dissolution of gold by solutions containing mercury cyanide is greatly expedited by heat. In the author’s experiments on ores, the quickening effect of mercury cyanide on solutions of cyanide was very slight. The use of bromine in the Sulman-Teed process has been already described. Sodium peroxide has also been tried, but it does not seem to be used in practice.