The purest and most transparent, because free from salt, is made if it is drawn off at the thickening stage, according to the following method. There are poured into the caldron the same number of amphorae of the solution as of congii of the lye of which I have already spoken, and into the same caldron is thrown as much of the already made saltpetre as the solution and lye will dissolve. As soon as the mixture effervesces and forms scum, it is transferred to a vat, into which on a cloth has been thrown washed sand obtained from a river. Soon afterward the plug is drawn out of the hole at the bottom, and the mixture, having percolated through the sand, escapes into a tub. It is then reduced by boiling in one or another of the caldrons, until the greater part of the solution has evaporated; but as soon as it is well boiled and forms scum, a little lye is poured into it. Then it is transferred to another vat in which there are small rods, to which it adheres and congeals in two days if there is but little of it, or if there is much in three days, or at the most in four days; if it does not condense, it is poured back into the caldron and re-boiled down to half; then it is transferred to the vat to cool. The process must be repeated as often as is necessary.
Others refine saltpetre by another method, for with it they fill a pot made of copper, and, covering it with a copper lid, set it over live coals, where it is heated until it melts. They do not cement down the lid, but it has a handle, and can be lifted for them to see whether or not the melting has taken place. When it has melted, powdered sulphur is sprinkled in, and if the pot set on the fire does not light it, the sulphur kindles, whereby the thick, greasy matter floating on the saltpetre burns up, and when it is consumed the saltpetre is pure. Soon afterward the pot is removed from the fire, and later, when cold, the purest saltpetre is taken out, which has the appearance of white marble, the earthy residue then remains at the bottom. The earths from which the solution was made, together with branches of oak or similar trees, are exposed under the open sky and sprinkled with water containing saltpetre. After remaining thus for five or six years, they are again ready to be made into a solution.
Pure saltpetre which has rested many years in the earth, and that which exudes from the stone walls of wine cellars and dark places, is mixed with the first solution and evaporated by boiling.
Thus far I have described the methods of making nitrum, which are not less varied or multifarious than those for making salt. Now I propose to describe the methods of making alum,[10] which are likewise neither all alike, nor simple, because it is made from boiling aluminous water until it condenses to alum, or else from boiling a solution of alum which is obtained from a kind of earth, or from rocks, or from pyrites, or other minerals.
If vitriol forms part of the aluminous ore, the material is dissolved in water without being mixed with urine, but it is necessary to pour that into the clear and pure solution when it is to be re-boiled. This separates the vitriol from the alum, for by this method the latter sinks to the bottom of the caldron, while the former floats on the top; both must be poured separately into smaller vessels, and from these into vats to condense. If, however, when the solution was re-boiled they did not separate, then they must be poured from the smaller vessels into larger vessels and covered over; then the vitriol separating from the alum, it condenses. Both are cut out and put to dry in the hot room, and are ready to be sold; the solution which did not congeal in the vessels and vats is again poured back into the caldron to be re-boiled. The earth which settled at the bottom of the caldron is carried back to the tanks, and, together with the ore, is again dissolved with water and urine. The earth which remains in the tanks after the solution has been drawn off is emptied in a heap, and daily becomes more and more aluminous in the same way as the earth from which saltpetre was made, but fuller of its juices, wherefore it is again thrown into the tanks and percolated by water.