or back-loads in use still among the less-advanced peoples of the Indian Archipelago, became a fixed scientific unit. Secondly, we must notice that the Malays have not followed the Chinese in the subdivisions of the catty. For whilst in China 16 taels or ounces go to the catty, the Malays follow more strictly the decimal system, and make their catty simply the tenfold of the tael or ounce. This same method of division we found already in Annam, and not only in Annam but also in Cambodia and Laos we found the silver nên or bar, invariably consisting of ten such parts, corresponding in weight to the Chinese tael, sixteen of which go to the catty.
It would appear, then, that here we have a combination of units of weight and units of capacity. The higher gold and silver unit, the nên, is simply the tenfold of the lower unit, the tael or ounce, while the catty, which is never employed in China in estimating gold or silver, but is a genuine commercial unit, was probably originally some natural unit of capacity. We saw strong evidence of this in Cambodia, where the name for this weight is neal or cocoanut, and we have just found the cocoanut as the chief unit of dry measure amongst the Malays of the Indian Seas. It was probably found that 16 times the tael or ounce came nearer to the weight of the contents of a cocoanut or bamboo joint (whatever kind of matter they may have weighed in it for this purpose, whether rice, or water), than the original 10 ounces, which formed the bar, the highest genuine weight unit. Sixteen was likewise a convenient number, its factors being numerous, and it could be divided in four portions, each of which contained four other units. It will presently be a question as to whether similar influences have not produced our pound avoirdupois, with its 16 sub-multiples.
M. Moura found a difficulty regarding the Cambodian neal or cocoanut catty; because a neal of rice only weighs half the weight, at which the neal is rated as a weight. But we saw in Java that the chapa or cocoanut measure is estimated at 2½ pounds avoirdupois. It is then not improbable