193. It has been stated by some authors that there is never much loss of gold in roasting ores, unless the roasting is performed too rapidly, and that the addition of salt makes no difference in this respect. I differ with them on this point, even when the statement is supported by so formidable a name as that of Plattner. But I think that illustrious chemist would yield to the logic of a pecuniary loss of three thousand dollars, especially when backed by other facts.

When I first entered on the business of gold chlorination on the large scale, I had a class of concentrations to treat which, while consisting chiefly of iron pyrites, and presenting no visible peculiarity, had nevertheless baffled many old operators, a fact of which I was not at the time aware. The ore was roasted in a three-hearth reverberatory furnace, with the addition of from I to 2 per cent of salt, on account of the presence in it of a considerable quantity of silver.

I was surprised to find that although the assays of the tailings were satisfactory, the gold, when collected, fell alarmingly short of the results which I had guaranteed, and I was of course obliged to make up the deficit.

Relying on the statement referred to, which I found in the only handbook on the subject within my reach, and being then a novice in this branch of metallurgy, I did not dream that a serious loss was taking place in the roasting furnace, especially as I had an expert metallurgist in reduced circumstances employed on the roasting, who had no more suspicion of the truth than I had.

Finding that I was sustaining some loss by inadequate tub room for the gold solution, and having consequently to draw the liquid from the gold-tub too soon after precipitation, I thought that I should find the whole loss to be in the leaching and precipitating department, and each time that an improvement was made, expected better results, so that I was led on from trial to trial, until the total loss reached the sum named.

I am not quite certain that I should ever have discovered the truth, but for the following accident. One day, I so far checked the draft of the furnace as to cause some fumes to come through the airholes and working doors, and a yellow sublimate on the masonry attracted my attention. On examination I found the sublimate to be very rich in gold, although there was none to be seen in it by the most careful washing. It also contained iron perchloride, and copper chloride, with some lead and other substances. This aroused my suspicion, and I at once did what I ought to have done sooner.

Knowing that the material could be roasted with little or no loss, if no salt was used, because some assays had been made in that way, I weighed two half ounces of a sample, and roasted them in the muffle side by side, under precisely the same conditions, except that to one of them I added 4 per cent of salt. The roasting was purposely pushed to an extreme as to heat and time, and when the two tests were assayed, under ex-actlv similar conditions, that which was salted was found to contain less than half as much gold as the un-salted one.

I then took some light fluffy sublimate from the flue of the roasting furnace, an assay of which gave me a value of some $600 per ton, chiefly gold. The quantity of this material was, however, very small, and the bulk of the matter in the dust chamber was not much richer than the average of the ore treated, a circumstance which indicates that the gold was actually to a great extent volatilized in some not easily condensable form. I also found that the ore sustained a loss of weight in roasting, equal to about 18 per cent, consequently the roasted ore ought to have been more than 18 per cent richer than before roasting, which was not the case.

If this is not considered to be sufficient proof that gold may be volatilized in the roasting of some ores with salt, the deficiency is supplied by the fact that, as soon as I made the necessary change by reserving the salt until the nearly dead roasting of the ore was finished, not only did the roasted ore assay 20 per cent richer than when raw, but the yield overran my guarantee, while the tailings nevertheless contained considerably more gold than before.

The moral of this is, never to neglect any precaution in the way of tests and assays, nor to place implicit reliance on rules laid down, or inferences drawn by others who have worked under different conditions.

I afterwards found that a very small quantity of salt, not more than three pounds to the ton, might be mixed with the crude ore without detriment to the gold, and with decided advantage to the extraction of the silver.