Some make twenty-five needles, in order to be able to detect the two scripula of silver or copper which are in a bes of gold. Of these needles, the first is composed of twelve duellae of gold and six of silver, and the same number of copper. The second, of twelve duellae and one sextula of gold and five duellae and one and a half sextulae of silver, and the same number of duellae and one and a half sextulae of copper. The remaining needles are made in the same proportion.

Pliny is our authority that the Romans could tell to within one scripulum how much gold was in any given alloy, and how much silver or copper.

Needles may be made in either of two ways, namely, in the ways of which I have spoken, and in the ways of which I am now about to speak. If unequal portions of silver and copper have been mixed with the gold, thirty-seven needles are made in the following way:—

Since it is rarely found that gold, which has been coined, does not amount to at least fifteen duellae gold in a bes, some make only twenty-eight needles, and some make them different from those already described, inasmuch as the alloy of gold with silver and copper is sometimes differently proportioned.

These needles are made:—

Next follows the fourth kind of needles, by which we test silver coins which contain copper, or copper coins which contain silver. The bes by which we weigh the silver is divided in two different ways. It is either divided twelve times, into units of five drachmae and one scripulum each, which the ordinary people call nummi[41]; each of these units we again divide into twenty-four units of four siliquae each, which the same ordinary people call a grenlin; or else the bes is divided into sixteen semunciae which are called loths, each of which is again divided into eighteen units of four siliquae each, which they call grenlin. Or else the bes is divided into sixteen semunciae, of which each is divided into four drachmae, and each drachma into four pfennige. Needles are made in accordance with each method of dividing the bes. According to the first method, to the number of twenty-four half nummi; according to the second method, to the number of thirty-one half semunciae, that is to say a sicilicus; for if the needles were made to the number of the smaller weights, the number of needles would again be too large, and not a few of them, by reason of the small difference in proportion of silver or copper, would have no significance. We test both bars and coined money composed of silver and copper by both scales. The one is as follows: the first needle is made of twenty-three parts of copper and one part silver; whereby, whatsoever bar or coin, when rubbed on the touchstone, colours it just as this needle does, in that bar or money there is one twenty-fourth part of silver, and so also, in accordance with the proportion of silver, is known the remaining proportion of the copper.

The other method of making needles is as follows:—

So much for this. Perhaps I have used more words than those most highly skilled in the art may require, but it is necessary for the understanding of these matters.

I will now speak of the weights, of which I have frequently made mention. Among mining people these are of two kinds, that is, the greater weights and the lesser weights. The centumpondium is the first and largest weight, and of course consists of one hundred librae, and for that reason is called a hundred weight.

The various weights are:—

This libra consists of sixteen unciae, and the half part of the libra is the selibra, which our people call a mark, and consists of eight unciae, or, as they divide it, of sixteen semunciae:—