grs. and an ounce of 390 grs, which is certainly not far from the weight of the small gold talent. It follows from this that we may expect pounds of different weights in Italy, according as the foot-unit varies in different districts.
In later times, besides the pound of 12 unciae, there were several commercial pounds on Italian soil, the pound of 16 ounces (from which our own avoirdupois is probably descended), that of 18 unciae, and that of 24. The last two are easy of explanation, since one is simply the double, the other one and a half times the Roman pound. But perhaps a different explanation must be sought for the 16 ounce pound. The foot was divided by Greeks and also by Italians into 16 fingers as well as into 12 thumbs. Was therefore the pound of 16 ounces simply derived from the division of the foot bar into 16 fingers, the weight of the finger being however equated to that of the Roman thumb or inch of copper?
The as, having been once subjected to weight, its hundredfold, the centumpondium or "hundred weight," became the highest Roman weight-unit. Thus the as and the centumpondium of the Italians correspond to the mina and talent of the Greeks. But it will be observed that the Italians obtained their higher unit by the old decimal system, whereas the Greeks had borrowed the mina and its sixtyfold from Asia. The centumpondium must be regarded as a true-born Italian unit, not one borrowed from Greece or Asia, and of this there is further proof. We saw by the ancient Roman law that the cow was estimated at 100 asses, the sheep at 10 asses. No doubt from time out of mind 100 of the bars of copper, which formed the chief lower unit of barter, made one cow, just as in Annam 280 little hoes make one buffalo (p. 167). When copper came to be weighed, the amount of copper which formed the equivalent of the highest unit of barter, the cow, was taken as the highest weight-unit. From what I have said above it is not improbable that the Roman libra and the Sicilian litra of copper were almost equal in weight. The fact that the Greek writers always employed the Sicilian word litra ([Greek: litra]), to translate the Latin libra, likewise indicates that in the Greek mind there was a tra-
