We also separate silver from gold by the same method when we assay them. For this purpose the alloy is first rubbed against a touchstone, in order to learn what proportion of silver there is in it; then as much silver as is necessary is added to the argentiferous gold, in a bes of which there must be less than a semi-uncia or a semi-uncia and a sicilicus[14] of copper. After lead has been added, it is melted in a cupel until the lead and the copper have exhaled, then the alloy of gold with silver is flattened out, and little tubes are made of the leaves; these are put into a glass ampulla, and strong aqua is poured over them two or three times. The tubes after this are absolutely pure, with the exception of only a quarter of a siliqua, which is silver; for only this much silver remains in eight unciae of gold[15].
As great expense is incurred in parting the metals by the methods that I have explained, as night vigils are necessary when aqua valens is made, and as generally much labour and great pains have to be expended on this matter, other methods for parting have been invented by clever men, which are less costly, less laborious, and in which there is less loss if through carelessness an error is made. There are three methods, the first performed with sulphur, the second with antimony, the third by means of some compound which consists of these or other ingredients.
A little of the regulus is taken from the crucible, but not from the gold lump which has settled at the bottom, and a drachma of it is put into each of the cupels, which contain an uncia of molten lead; there should be many of these cupels. In this way half a drachma of silver is made. As soon as the lead and copper have been separated from the silver, a third of it is thrown into a glass ampulla, and aqua valens is poured over it. By this method is shown whether the sulphur has parted all the gold from the silver, or not. If one wishes to know the size of the gold lump which has settled at the bottom of the crucible, an iron rod moistened with water is covered with chalk, and when the rod is dry it is pushed down straight into the crucible, and the rod remains bright to the height of the gold lump; the remaining part of the rod is coloured black by the regulus, which adheres to the rod if it is not quickly removed.
If when the rod has been extracted the gold is observed to be satisfactorily parted from the silver, the regulus is poured out, the gold button is taken out of the crucible, and in some clean place the regulus is chipped off from it, although it usually flies apart. The lump itself is reduced to granules, and for every libra of this gold they weigh out a quarter of a libra each of crushed sulphur and of granular copper, and all are placed together in an earthen crucible, not into a pot. When they are melted, in order that the gold may more quickly settle at the bottom, the powder which I have mentioned is added.
Although minute particles of gold appear to scintillate in the regulus of copper and silver, yet if all that are in a libra do not weigh as much as a single sesterce, then the sulphur has satisfactorily parted the gold from the silver; but if it should weigh a sesterce or more, then the regulus is thrown back again into the earthen crucible, and it is not advantageous to add sulphur, but only a little copper and powder, by which method a gold lump is again made to settle at the bottom; and this one is added to the other button which is not rich in gold.
When gold is parted from sixty-six librae of silver, the silver, copper, and sulphur regulus weighs one hundred and thirty-two librae. To separate the copper from the silver we require five hundred librae of lead, more or less, with which the regulus is melted in the second furnace. In this manner litharge and hearth-lead are made, which are re-smelted in the first furnace. The cakes that are made from these are placed in the third furnace, so that the lead may be separated from the copper and used again, for it contains very little silver. The crucibles and their covers are crushed, washed, and the sediment is melted together with litharge and hearth-lead.
Those who wish to separate all the silver from the gold by this method leave one part of gold to three of silver, and then reduce the alloy to granules. Then they place it in an ampulla, and by pouring aqua valens over it, part the gold from the silver, which process I explained in [Book VII].