BOOK VII.

249

with the mouth and then a blue flame will be emitted. In the end the tubesare weighed, and if their weights prove equal, he who has undertaken this workhas not laboured in vain. Lastly, both are placed in another balance-pan andweighed ; of each tube four grains must not be counted, on account of thesilver which remains in the gold and cannot be separated from it. From theweight of the tubes we learn the weight both of the gold and of the silverwhich is in the button. If some assayer has omitted to add so much silver tothe gold as to make it three times the quantity, but only double, or two and ahalf times as much, he will require the stronger quality of aqua whichseparates gold from silver, such as the fourth quality. Whether the aquawhich he employs for gold and silver is suitable for the purpose, or whetherit is more or less strong than is right, is recognised by its effect. That ofmedium strength raises the little bubbles on the tubes and is found to colourthe ampulla and the operculum a strong red ; the weaker one is found tocolour them a light red, and the stronger one to break the tubes. To puresilver in which there is some portion of gold, nothing should be added whenthey are being heated in the cupel prior to their being parted, except a besof lead and one-fourth or one-third its amount of copper of the lesser weights.If the silver contains in itself a certain amount of copper, let it be weighed,both after it has been melted with the lead, and after the gold has been partedfrom it; by the former we learn how much copper is in it, by the latter howmuch gold. Base metals are burnt up even to-day for the purpose of assay,because to lose so little of the metal is small loss, but from a large mass ofbase metal, the precious metal is always extracted, as I will explain inBooks X. and XI.

We assay an alloy of copper and silver in the following way. From afew cakes of copper the assayer cuts out portions, small samples from smallcakes, medium samples from medium cakes, and large samples from largecakes ; the small ones are equal in size to half a hazel nut, the largeones do not exceed the size of half a chestnut, and those of medium size comebetween the two. He cuts out the samples from the middle of thebottom of each cake. He places the samples in a new, clean, triangularcrucible and fixes to them pieces of paper upon which are written the weightof the cakes of copper, of whatever size they may be ; for example, he writes,“These samples have been cut from copper which weighs twenty centum-pondia.” When he wishes to know how much silver one centumpondium ofcopper of this kind has in it, first of all he throws glowing coals into theiron hoop, then adds charcoal to it. When the fire has become hot, the paperis taken out of the crucible and put aside, he then sets that crucible on thefire and gradually heats it for a quarter of an hour until it becomes red hot.Then he stimulates the fire by blowing with a blast from the double bellowsfor half an hour, because copper which is devoid of lead requires this time tobecome hot and to melt; copper not devoid of lead melts quicker. Whenhe has blown the bellows for about the space of time stated, he removes theglowing'charcoal with the tongs, and stirs the copper with a splinter of wood,which he grasps with the tongs. If it does not stir easily, it is a sign that the