We have seen that the Chinese system of weights is based upon natural seeds of plants, and we have actually found the wild hillsmen of Annam and Laos weighing their gold dust by grains of maize and rice. But it may be urged by the advocates of a Babylonian scientific origin based on the one-fifth of the cube of the royal ell, which in turn is based upon the sun’s apparent diameter, that the Chinese names of weights are merely conventional terms taken from the name of certain seeds, and on the other hand that the mere fact that a very barbarous people like the Bahnars of Annam weigh their gold dust by grains of rice is no evidence that people in a higher stage of culture were content with such rude metric standards. I propose to show in this chapter that it has been the actual practice of peoples as far advanced in civilization as the ancient Greeks or Italians, to employ seeds as weights down to the present day in Asia, that it was the general practice in the middle ages, that it was likewise the practice of the Romans of the empire, of the Greeks, and finally that such too was the practice of the Assyrians themselves at a period long before the bronze Lion weights were ever cast, or the stone Duck weights were carved. If I succeed in proving this proposition, the doctrine that the art of weighing was scientific must give place to the contention that it was purely empirical.

As we have found among the barbarians of Asia the first beginnings of the art of weighing by the employment of grains of rice and maize, it is best for us to take first in order some other Asiatic countries lying towards the same region.

The great islands of the Indian Archipelago, singularly rich in all endowments of nature, have for ages enjoyed a high degree of culture. Conveniently placed, they have received all the advantages of contact with the civilization of China, India, and even that of the Arabs from the distant west of Asia. Never were people more favourably situated for obtaining foreign systems of weights and measures, if they felt so disposed, than the Malays of Java and Sumatra and the other islands of the Indian Archipelago. That admirable observer, John Crawfurd, writing in 1820 says[225]: “In the native measures everything is estimated by bulk and not by weight. Among a rude people corn would necessarily be the first commodity that would render it a matter of necessity and convenience to fix some means for its exchange or barter. The manner in which this is effected among the Javanese will point out the imperfection of their methods. Rice, the principal grain, is in reaping nipped off the stalk with a few inches of the straw, tied up in sheaves or parcels and then housed or sold, or otherwise disposed of. The quantity of rice in the straw which can be clenched between the thumb and the middle finger is called a gagam or handful, and forms the lowest denomination. Three gagams or handfuls make one pochong, the quantity which can be clenched between both hands joined. This is properly a sheaf. Two sheaves or pochongs joined together, as is always the case, for the convenience of being thrown across a stick for transportation, make a double sheaf or gedeng. Five gedengs make a songga, the highest measure in some provinces, or twenty-four make an hamat, the more general measure. From their very nature these measures are indefinite and hardly amount to more accuracy than we employ ourselves when we speak of sheaves of corn. In the same district they are tolerably regular in the quantity of grain or straw they contain, but such is the wide difference between different districts or provinces, that the same nominal measures are often twice, nay three times as large in one as in another. For the hamat or larger measure perhaps about eight hundred pounds avoirdupois might be considered a fair average for the different provinces of Java. This may convey some loose notion of the quantities intended to be represented. For dry and liquid measures they may naturally have recourse to the shell of the cocoanut and the joint of the bamboo which are constantly at hand. The first called by the Malays chupa is estimated to be two and a half pounds avoirdupois. The second is called by some tribes kulch and is equal to a gallon, but the most common bamboo measure is the gantung, which is twice this amount. To those exact and business-like dealers, the Chinese, and in a less degree to the Arabs and people of the east coast of the Indian Peninsula, the Indian islanders are chiefly indebted for any precision we find in their weights. In all the traffic carried on between the commercial tribes and foreigners, the Chinese weights, though occasionally under native names, are constantly referred to. The lowest of these, called sometimes by the native name of Bungkal, but more frequently by the Chinese name of Tahil [tael], varies from twenty-four pennyweights nine grains to thirty pennyweights and twenty grains. Ten of these make a kati [catty] or about twenty ounces avoirdupois; one hundred katis make a pikul or 133⅓ lbs. avoirdupois, and thirty pikuls make one koyan. Of these the kati and the pikul, because they are constantly referred to in considerable mercantile dealings, are the only well-defined weights. The koyan by some is reckoned at twenty pikuls, by others at twenty-seven, twenty-eight and even at forty. The Dutch are fond of equalizing it with their own standards and consider it as equal to a last or two tons.

“The Bahara, an Arabic weight, is occasionally used in the weighing of pepper, but its amount is very indefinite, for in some of the countries of the Archipelago it amounts to 396 lbs., and in others to 560 lbs.”

Elsewhere he says[226], “The picul is strictly a Chinese weight as its amount shews, though the term happens in this case to be native. Its meaning in the vernacular languages is a natural load or burthen, and when used in this primitive sense it, without reference to the Chinese weight, is not found to exceed eighty pounds avoirdupois.” This is a fact of great importance as we shall see when we come to the development of the mina and talent of Graeco-Asiatic commerce.

Finally Crawfurd says, “The nice question of weighing gold, the only native commodity which could not be estimated by tale or bulk, has given rise to the use of weights among the natives themselves. Grains of rice are still occasionally used in the weighing of gold in the neighbourhood of the gold mines in Sumatra” (p. 274).

I have quoted at full length these passages in order that the reader may accept with fuller confidence statements so instructive as regards the origin of weight, the first object to be weighed, and the origin of the picul, or as we may call it the talent of Eastern Asia. Nine years before Crawfurd wrote there had appeared William Marsden’s admirable History of Sumatra[227]. He gives us far fuller information on the subject of gold than Crawfurd has done. Thus he writes: “In those parts of the country where traffic in this article (gold dust) is considerable, it is employed as currency instead of coin; every man carries small scales about him, and purchases are made with it so low as to the weight of a grain or two of padi. Various seeds are used as gold weights, but more especially these two: the one called rakat or saga-tim-bañgan (Glycine abrus L or abrus maculatus of the Batavian trans.), being the well-known scarlet pea with a black spot, twenty-four of which constitute a mas, and sixteen mas (mace) a tāil (tael): the other called saga puku and kondori batang (Aden anthera pavonia L), a scarlet or rather coral bean much larger than the former, and without the black spot. It is the candarin weight of the Chinese, of which one hundred make a tāil and equal, according to the tables published by Stevens, to 5·7984 gr. Troy, but the average weight of those in my possession is 10·50 Troy grains. The tāil differs however in the northern and southern parts of the island, being at Natal, Padang, Bencoolen and elsewhere twenty-six pennyweights six grains. At Achin the bangkal of thirty pennyweights twenty-one grains is the standard. Spanish dollars are everywhere current and accounts are kept in dollars, sukus (imaginary quarter dollars) and kepping or copper cash, of which four hundred go to the dollar. Besides these there are silver fanams, single, double and treble (the latter, called tali), coined at Madras, twenty-four fanams or eight talis being equal to the Spanish dollar, which is always valued in the English settlements at five shillings.”

He adds that copper is sold by weight (picul), and that tin, which was accidentally discovered in 1710 by the burning of a house, is exported for the most part in small pieces or cakes called tampangs, sometimes in slabs (p. 172), and furthermore they purchase bar iron by measurement instead of by weight (p. 176).

Several points of great importance are to be noticed in the foregoing statements. Firstly, that whilst for foreign trade with the Chinese they employ the Chinese weight, which we know always by its Malay name of picul, a well-defined weight standard of 133⅓ lbs. avoirdupois, they had evidently a native unit of weight, their own picul, which simply means and actually was as much as a man can carry on his back, and which, as we saw, rarely exceeds 80 lbs. avoirdupois. This seems to give us an insight into the manner in which the most primitive highest weight unit is arrived at. A man’s load is one of those natural standards which will vary according to race and climate, and the conditions under which the load has to be borne. Thus, the average weight of the load borne by a dock porter who has to endure the strain for only some few yards, will of course be far higher than that carried by the porters of travellers in Central Africa, where the load has to be borne day after day on a march of several hundred, or a thousand miles. Thus in the case of the Madis, a pure negro tribe, the average load seems to be about 50 pounds, which they can carry “20 miles a day for eight or ten consecutive days without shewing any signs of distress[228].” The Chinese, the superiors in science of all Eastern Asia, have carefully adjusted this “load,” and it makes, as we have seen above, their highest weight unit. Its particular amount is probably due to the fact that, having carefully fixed the weight of the smaller units, the candarin, the mace, the liung or tael, and the catty, their pound, they simply took the hundredfold of the chang or catty as the standard for their highest unit, and thus that which at an earlier stage was just as vague and fluctuating as the picul, 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 that some liquid or substance far heavier than rice was used to fill the cocoanut, when the value of its contents was being ascertained by weighing so as to serve as a general unit. The same variation in weight, owing to the different nature of its contents, has, as mentioned before, given rise in Ireland to barrels of various weights. Thus a barrel of wheat contains 20 stone avoirdupois, a barrel of potatoes 24 stone, a barrel of barley 16 stone, and a barrel of oats 14 stone. This diversity simply arose from comparative lightness or heaviness of the different commodities which were measured by one and the same unit of capacity: the barrel itself, having been fixed by a process of measurement, similar to that by which the milk-pan was regulated among the Welsh, and the pannier among the natives of Laos. The principle by which higher units of capacity or weight are formed is likewise well illustrated by the instance given above of the cartload of rice, which is simply regarded as the multiple of the pannier or bag, which forms the smaller unit for rice. The size of the cartload would be conditioned by the size of the cart usually employed, which in turn would depend on a variety of other things, such as the nature of the country, or its roads, or the kind of animals employed for draught. The vagueness in amount of the koyan or multiple of the picul noticed by Crawfurd, may thus meet with a reasonable explanation.

We may now return to the mainland of Asia, where we shall find in the weight system of the Hindus at least one remarkable point of affinity with that of Sumatra. Marsden has told us that the rakat or scarlet pea with a black spot is one of the chief weights employed for gold in Sumatra. This rakat is none other than the ratti, which is usually taken as the basis of the modern Hindu weight system. “This weight,” says that eminent scholar Colebrooke[229], “is the lowest denomination in general use, commonly known by the name ratti, the same with rattika, which, as well as ṛaktika, denotes the red seed as kṛishnala indicates the black seed of the gunjá-creeper.” Mr Thomas has shown the true weight of the ratti is 1·75 grains[230].

Many different standards have been used in India for various purposes, one for the weighing of gold, another for the weighing of silver, another used by jewellers, and yet another by the medical tribe, but all alike start from the ratti.

“The determination of the true weight of the ratti has done much both to facilitate and give authority to the comparison of the ultimately divergent standards of the ethnic kingdoms of India. Having discovered the guiding unit, all other calculations become simple, and present singularly convincing results, notwithstanding that the bases of all these estimates rest upon so erratic a test as the growth of the seed of the gunjá-creeper (Abrius precatorius) under the varied influences of soil and climate. Nevertheless the small compact grain, checked in early times by other products of nature, is seen to have the remarkable faculty of securing a uniform average throughout the entire continent of India, which only came to be disturbed when monarchs like Shîr Shâh and Akbar in their vanity raised the weight of the coinage without any reference to the numbers of rattis, inherited from Hindu sources, and officially recognized in the old, but entirely disregarded and left undefined in the reformed Muhammadan mintages[231].” We shall learn shortly that in its uniformity the ratti does not differ from other seeds such as wheat and barley. Probably, however, the fact that the gunjá-creeper was found everywhere in India gave it its position of a universal standard. Those who wish to study the elaborate systems of later times employed in India can consult the works of Colebrooke and Thomas already referred to.

The legislators Manu, Yájnavalkya, and Nárada trace all weights from the least visible quantity which they concur in naming trasareṇu and describing as the very small mote, “which may be discovered in a sunbeam passing through a lattice.” Writers on medicine proceed a step further, and affirm that a trasareṇu contains 30 paramáṇu or atoms. The legislators above-named proceed from the trasareṇu as follows:

But as we want to learn what was the actual usage of the Hindus, instead of dealing with the mere theoretic statements of late authors, I shall at once quote in full the tables given in the Līlāvati of Brahmegupta, who wrote his Algebra and Arithmetic about 600 A.D.[232]

Money (by tale). Twice ten cowries[233] are a cácíní; four of these are a pána, sixteen of which must here be considered as a dramma, and in like manner a nishká as consisting of sixteen of these.

Weight. A gunjá (or seed of Abrus), is reckoned equal to two barley-corns (yavas). A valla is two gunjás and eight of these are a dharana, two of which make a yadyanaca. In like manner one dhataca is composed of fourteen vallas.

Half ten gunjás are called a másha by such as are conversant with the use of the balance; a karsha contains sixteen of what are called máshas, a pala four karshas. A karsha of gold is named suvarṇa.

This is quite in harmony with the weight of gold as given by the legislators:

Yájnavalkya adds that according to some 5 suvarṇas = 1 pala.

All the authorities seem agreed in regarding the term suvarṇa as peculiar to gold, for which metal it is also a name.

We learn thus that the Hindu standards were fixed by means of natural seeds, and at no period do they, clever mathematicians as they were, seem to have made any effort at obtaining a mathematical basis for their metric systems.

We also observe that the weight known as the suvarṇa or gold weight par excellence is the weight of a karsha or 80 gunjás, which, if we take the gunjá = 1·75 grains Troy, gives the weight of the suvarṇa as 140 grains. I have already (p. 127) taken the original Hindu gold unit as not far from this amount. From the Līlāvati we may now with little misgiving assume it to have been such.

Lastly, let us observe that the barley-corn appears as the basis of the system in the tables of Brahmegupta and Bhascara, although the ṛaktika evidently overmasters it in the course of time. This is very interesting, for it indicates that the Hindus had learned the art of weighing in a comparatively northern region, where barley was the chief cereal under cultivation. If the system had been invented in the more southern parts of India, the grain of rice, the staple of life in the southern regions, would certainly have appeared as the sub-multiple of the ṛaktika, instead of the barley. As a matter of fact, rice-grains seem to have been occasionally used locally, for Colebrooke remarks that “it is also said that the ṛaktika is equal in weight to four grains of rice in the husk.” This supposition is completely in accord with what we found in Persia, where the modern weight system for gold, silver and medicine runs thus:

Although the miscal and habba denote Arabic influence, we may, without straining probabilities, conjecture that the use of the barley-corn here as well as in India, where we found it at a period anterior to Muhammadan conquest, indicates that in Persia it existed likewise from the earliest times. The close relationship between the ancient Hindus and ancient Persians makes it all the more likely. It is also pointed out that formerly the nashod was divided into three instead of four grains. As the Arabs divide their karat into four habbas, it is all the more likely that the 3 barley-corns = 1 nashod belong to the ancient system.

The Arab weight system is based on the grain of wheat, four of which make a karat (the seed of the carob or St John’s Bread)[235]. Occasionally in the Arab writers mention is made of a karat divided into 3 habbas[235]. The weight of the karat remains unchanged, but the grains in this case are barley grains, since, as we shall see presently, 3 grains of barley are equal to 4 grains of wheat (·063 × 3 = ·047 × 4).

It will now be most convenient for us to begin in the extreme west, and once more from that work back towards the coast of the Aegean Sea, in which our chief interest must always be centred.

Whether the Kelts of Ireland had any indigenous weight system or not, we have no direct evidence, although we do know as a fact that when Caesar landed in Kent he found the Britons employing coins of gold and bronze, and bars (or according to some MSS. rings) of iron adjusted to a fixed weight. However the earliest Irish documents reveal that people using a system of weights for silver directly borrowed from the older Roman system (although it is likely that they had a native standard for gold). As the solidus and denarius became the chief units of Europe from the time of Constantine the Great (336 A.D.), the Irish probably received their system at an earlier date.

When we pass to England, the very word grain which we employ to express our lowest weight unit, would of itself suggest that originally some kind of grain or seed was employed by our forefathers in weighing, but as the grain in use among us is the grain Troy, and as we have not yet learned its origin, it will not do to argue vaguely from etymology. But a little enquiry soon brings us to a time when the grain Troy did not as yet form the basis of English weights, and when a far simpler method of fixing the weight of the kings coinage was in vogue. It was ordained by 12 Henry VII. ch. V. “that the bushel is to contain eight gallons of wheat, and every gallon eight pounds of wheat, and every pound twelve ounces of Troy weight, and every ounce twenty sterlings, and every sterling to be of the weight of thirty-two grains of wheat that grew in the midst of the ear of wheat according to the old laws of this land[237].” Going backwards we find that in 1280 (8 Edward I.) the penny was to weigh 24 grains, which by weight then appointed were as much as the former 32 grains of wheat. By the Statute De Ponderibus, of uncertain date but put by some in 1265, it was ordained that the penny sterling should weigh 32 grains of wheat, round and dry, and taken from the midst of the ear. Going back a step still further we find that by the Laws of Ethelred, every penny weighed 32 grains of wheat[238], and as the pennies struck by King Alfred weigh 24 grains Troy, we may assume without hesitation that they were struck on the same standard of 32 grains of wheat. Thus from Alfred (871-901) down to Henry VII. (1485-1509), we find the penny fixed by this primitive method, and the actual weight of the coins, as tested by the balance at the present day, affords proof positive of the method.

But all the standards of mediaeval Europe (with the exception of the Irish) were based on the gold solidus of Constantine the Great[239]. The solidus (itself weighing 72 grains Troy or ⅟₇₂ of the Roman pound) was divided into 24 siliquae. The siliqua, or as the Greeks called it keration (κεράτιον, from which comes our word carat), was the seed of the carob, or as it is often called, St John’s Bread (Ceratonia siliqua L). Thus the lowest unit in the Roman system, as it is usually given, is found to be the seed of a plant. The same holds of the Greek system, for the drachma is described as containing 18 kerata or keratia, whilst according to others “it contains three grammata, but the gramma contains two obols and the obol contains three kerata, and the keras contains four wheat grains[240].” From this we see that the keration or siliqua was further reduced to 4 sitaria, or grains of wheat, whilst from another ancient table of weights[241] we learn that the siliqua likewise equals 3 barley-corns (siliqua grana ordei iii). Hence it appears that 3 barley-corns = 4 wheat grains. Thus both Greek and Roman systems just like the English and Irish take as their smallest unit a grain of corn. This also throws important light on the origin of that mysterious thing, the Troy grain. We saw above (8 Edward I.) that at the time of its introduction into England that 24 grains Troy = 32 grains of wheat, that is the Troy grain stands to wheat grain as 3:4. But as we have just seen that the siliqua = 3 barley-corns, and also = 4 wheat-corns, it follows that 3 barley-corns = 4 wheat-corns. And as 3 Troy grains = 4 wheat-corns, it likewise follows that 3 Troy grains = 3 barley-corns, or in other words, the barley-corn and Troy grain are the same things. It thus appears that the Troy grain is nothing more than the barley-corn, which was used as the weight unit in preference to the grain of wheat in some parts of the Roman empire. Furthermore this relation between barley-corns and wheat-corns can be proved to be a fact of Nature. In September, 1887, I placed in the opposite scales of a balance 32 grains of wheat “dry and taken from the midst of the ear,” and 24 grains of barley taken from ricks of corn grown in the same field at Fen Ditton, near Cambridge, and I thrice repeated the experiment; each time they balanced so evenly that a half grain weight turned the scale. The grain of Scotch wheat weighs ·047 gram, the Troy grain = ·064, ·047 × 4 = 188, ·064 × 3 = 192. Practically 4 wheat grains = 3 Troy grains.

Before passing from the Greek and Roman standards I may add that even higher denominations than the siliqua were expressed by the seeds of plants. The Romans made the lupin (lupinus) = 2 siliquae and under its Greek name of thermos (θερμός), it was assigned a like value (Metrol. Script. I. 81). In the Carmen de Ponderibus (Metrol. Script. II. 16), 6 grains of pulse (grana lentis) are made equal to 6 siliquae, and a like number of grains of spelt are given a similar value.

We next advance towards the East and take up the Semitic systems. We have already had occasion to touch upon that of the Arabs when dealing with the modern Persians. “There can be little doubt,” says Queipo (I. 360), “that the Arab system of weight was based on the grain of wheat.” The habba was their smallest unit. Four habbas are equal to 1 karat, the latter of course representing the keration or siliqua, and the former the 4 sitaria or wheat-grains, which we saw were its equivalent. This is the most ordinary value given to the karat in Makrizi and the other Arabic writers on Metrology, but occasionally we find the karat made equal to only 3 grains, which of course are barley-corns. We saw above that in the Persian system the nashod was formerly divided into 4 habbi of ·048 gram (which is plainly the weight of the wheat-grain), whilst now it is divided into 3 grains each of ·063 which represents the barley-corn, or in other words the Troy grain of ·064 gram. Of course the objection might be raised that as the Arabs had borrowed their higher denominations such as the dirhem (δραχμή) and dinar (denarius, δηνάριον), from the Greeks and Romans, and as their standard weight the mithkal is nothing more than the sextula or ⅙ of the Roman ounce, employed in the eastern Empire under the name of exagion (ἐξάγιον, whence comes the saggio of Marco Polo), so too their wheat-corns and barley-corns were not of their own devising, but likewise adventitious. After what we have seen above (p. 166) to be the practice of primitive people in the selling of gold, a traffic in which the Arabs had been engaged for many ages, it would seem hardly necessary to reply to such an argument, but as a more complete answer can be given in the course of the last portion of this enquiry, we shall deal with it in that place.

We now come to the Assyrians themselves, from the discovery of whose weights in the shape of lions and ducks, the whole modern theory of a scientific origin for all the weight standards of the Greeks as well as Asiatics and Egyptians has had its origin. But even within this sacred precinct of à priori metrology the irrepressible grain of corn springs up vigorously, although almost choked by the abundant crop of tares which have been sown around it. If we find that a Semitic people, who were the ancients of the earth before Pelops passed from Asia into Greece, or Romulus had founded his Asylum, employed the wheat grain as their lowest weight unit, we may then well argue that ages before the birth of the Prophet and the Arab conquest of Egypt and Syria, the Semitic folks employed grains of corn to form their lowest weight unit.

M. Aurès[242], a well-known Assyrian metrologist, has recently set forth the Assyrian system in its latest and most advanced stage. Following the veteran Assyriologist, M. Oppert, he finds that the Assyrians used a denomination lower than the obol. In the Museum of the Louvre there is a small Assyrian weight of the “duck” kind, which bears on its base the Assyrian character of 22 grains ½. The ideogram translated grain is evidently meant to represent some kind of corn with a rounded end. The weight of this object is ·95 gram (14⁶⁄₇ grains Troy). The weight is a ¾ obol, and therefore 30 grains went to the obol. This is the obol of the heavy Assyrian system, of which we shall presently speak. For the sake of clearness, I take M. Aurès’ table.

For our present purpose it is quite sufficient to call attention to the fact that this grain which forms the lowest unit of the Assyrian scale weighs ·042 gram (·95 ÷ 22·5) which is a very close approximation to the weight of the wheat-grain (·047). Making allowance for some loss which the weight may have sustained, it seems impossible to doubt that we have here the wheat-grain being used to form the smallest unit as it is in the modern Arabic system. The double obol of the Assyrians weighs 30 grains; we shall also find that the Hebrew gêrâh or obol (twenty of which made a shekel), weighed exactly 15 grains of wheat, that is the Hebrew gêrâh is the light obol which stood side by side with the heavy obol of 30 grains in the Assyrian system. Let us treat the matter from a slightly different point of view: As the light Assyrian obol contained 15 Assyrian grains, the light shekel contained 180 Assyrian grs. But as we know that this light Assyrian shekel weighed 8·4 grams, or 131 grains Troy, and as we know that the Troy grain is really the barley-corn and likewise that 3 barley-corns = 4 wheat grains, it is obvious that 131 grains Troy = 175 wheat grs. nearly, a very close approximation to the 180 Assyrian grs. Again as 180 Assyrian grs. = 8·4 grams, the Assyrian grain weighed ·046 gram, that is almost exactly the weight of a wheat grain (·047 gram).

But let us see for a moment in what fashion M. Aurès accounts for the presence of corn-grains in a system so elaborately scientific as he and his school maintain.

Starting as usual with the old assumption that all weight standards come from the measures of capacity and all measures of capacity in their turn are derived from the linear measures, he proceeds thus: The Assyrian ideogram which represents tribute, likewise represents talent. Tribute being paid in corn, no doubt the idea of weight first arose as the people carried their quota of corn on their backs to the receipt of custom. They accordingly weighed the measure (bar), which contained the proper amount of corn and took it as their weight unit, and then proceeded to make subdivisions of it. When their weight system was thus fixed, for convenience instead of going to the trouble of adjusting weights they took 30 grains of corn which would be just equivalent to the weight of an obol. After the many historical instances quoted in the preceding pages in which the methods of appraising the value of corn and other dry commodities have been set out, and also the manner in which corn grains have been employed for fixing the higher standard, as for instance in the adjustment of the English bushel in the reign of Henry VII., the reader will feel that M. Aurès has simply inverted the true order of events, and that as we found the natives of Annam and the Malays of the Indian Archipelago making their first essay in weighing by means of a grain of maize, or rice, or padi, so the ancient inhabitants of Mesopotamia made their first beginning, and as we have found everywhere that gold, the most precious of objects, was the first thing to be weighed, and as it only existed in small quantities, thus requiring but a very small unit of weight, so the Assyrians likewise began to weigh gold first of all, employing the natural seeds of corn, and only in process of time arrived at higher units by multiplying the smaller.

To all the evidence collected from Asia and Europe we can likewise add a fact of great importance from Africa. We saw that it was highly probable that the Carthaginians traded for gold to the West Coast of Africa, and beyond all reasonable doubt the natives of the Gold Coast have for ages been acquainted with that metal. Now it can be proved that these peoples, whilst employing no weights for any other mercantile transaction, used the seeds of certain plants for weighing their gold; thus Bosman writing two centuries ago says, “Having treated of gold at large, I am now obliged to say something concerning the gold weights, which are either pounds, marks, ounces or angels.... We use here another kind of weights which are a sort of beans, the least of which are red spotted with black and called Dambas; twenty-four of them amount to an angel, and each of them is reckoned two stiver weights; the white beans with black spots or those entirely black are heavier and accounted four stiver weights: these they usually call Tacoes, but there are some which weigh half or a whole gilder, but are not esteemed certain weights, but used at pleasure and often become instruments of fraud. Several have believed that the negroes only used wooden weights, but that is a mistake; all of them have cast weights either of copper or tin, which though divided or adjusted in a manner quite different to ours; yet upon reduction agree exactly with them[243]”.

I am informed by Mr Quayle Jones, Chief Justice of Sierra Leone, that at the present day, a seed called the Taku, (with a black spot) is employed by the natives of the Gold Coast for weighing gold. He also tells me that small quantities of gold are measured by a quill in ordinary dealings in the market[244]. I learn from another private source that 6 Takus = 1 ackie (20 ackies = 1 ounce). From Bosnian’s equating the bean with the red spot to 2 stiver-weights, we can deduce its weight as 2 grs. troy; this result combined with the colour of the bean would make us a à priori conclude that the Damba was the Abrus precatorius, so familiar to us already under its Hindu name of ratti.

Here we have a primitive people with a weight system of their own based on the Damba and Taku, just as the Hindu is based on the ratti, and here too we have another proof that the first of all articles to be weighed is gold. From Bosman we also learn that gold in small quantities was not always weighed, for he says of the inferior gold which was mixed with silver or copper, that it is cast into fetiches (small grotesque figures). “These fetiches are cut into small bits by the negroes of one, two, or three farthings. The negroes know the exact value of these bits so well at sight, that they never are mistaken, and accordingly they sell them to each other without weighing as we do coined money[245].” This recalls the practice as regards silver among the Tibetans at the present day.

Crossing to the eastern side of Africa we find the natives of Madagascar employing a system, the basis of which is a grain of rice. “The Malagasy have no circulating medium of their own. Dollars are known more or less throughout the island: but in many of the provinces trade is carried on principally by an exchange of commodities. The Spanish dollar, stamped with the two pillars, bears the highest value. For sums below a dollar the inconvenient method is resorted to in the interior, of weighing the money in every case. Dollars are cut up into small pieces, and four iron weights are used for the half, quarter, eighth, and twelfth of a dollar. Below that amount, divisions are effected by combinations of the four weights, and also by means of grains of rice, even down so low as one single grain—“Vary vray venty,” one plump grain, valued at the seven hundred and twentieth part of a dollar”[246]. The grain of rice therefore weighs ⁵⁄₉ gr. troy (·036 gram). As gold is not found in Madagascar[247] the natives could not weigh it first of all things; but they have carried out the principle of taking silver, the most precious article they possessed, as the first object to be weighed.

In this chapter, therefore, we have sought the method by which weight standards are fixed among primitive and semi-civilized peoples; we have studied the system or systems of China, Cochin-China, Cambodia, Laos and the great Islands of the Indian Ocean. Everywhere we have received the self-same answer, everywhere the lowest unit is nothing more than a natural seed or grain. We found in two places in the area studied, amongst the Tapaks of Annam and the Malays of Sumatra, the art of weighing in its earliest infancy; only one product, gold, as yet being weighed, and the weight unit employed for it being a grain of rice or maize. We found that this smallest natural unit of gold was amongst the Bahnars equated to the smallest unit of barter in use among them, the hoe, whilst their highest unit was the buffalo; and that by a simple process based on the known relation existing in value between the hoe, the muk, the jar, and the buffalo, there was no difficulty in arriving empirically at the exact value in gold of a buffalo. We found also that the two higher units of weight the picul, and the catty, which in almost every case were found to be confined to the ordinary merchandise, were beyond reasonable doubt not originally multiples of the lower the tael, but were really natural units obtained by a totally different process; the picul being the amount which an average man can conveniently carry on his back, the catty, as seen especially in the case of the neal of Cambodia, being nothing more than the cocoa-nut shell used as the ordinary measure of capacity, as a gourd of a certain kind is employed at Zanzibar, as the hen’s egg was employed by the Hebrews and also by the ancient Irish, as the cochlea or mussel shell was taken by the Romans as the basis of their measures of capacity, and as possibly the gourd itself under its name of Kyathos formed the lowest unit of capacity among the Greeks. We saw clearly that the catty has never become a weight-unit for precious metals among the Chinese, Annamites or Cambodians; the first named never having used any higher unit for such purpose than a bar of ten taels, and at the present day for the most part contenting themselves with the tael or ounce, whilst the two latter still use the nên or bar with its subdivisions into 10 denhs, or in other words, use as their highest monetary unit the tenfold of the tael or ounce. We likewise found that in Annam among the less advanced peoples there was considerable evidence to show that the bat or tical was originally the highest unit used for gold, and that this name bat was applied to weights of different amount; thus the chi which in commercial weight is only the quarter of a bat, is itself called the gold bat. The bat itself was the third of the tael. We also found the bar of silver, the common monetary unit at the present moment, equated to the buffalo, the common unit of barter among the Bahnars, and finally we had a distinct tradition that not so long ago the wild tribesmen who win the gold dust from the sands of their native brooks did not as yet even weigh the metal by means of the grains of maize which are now employed, but that they measured off a small rod of gold an inch long as the equivalent of a buffalo.

From all these facts it seems easy to trace the history of the development of weight standards in Further Asia; the first stage in trafficking in gold seems to be one purely by measure, then comes that of weighing by means of grains of corn, the weight in gold of one or more grains of corn being taken in the ordinary way of barter like other articles in the common scale of exchange. A multiple of the higher unit the bat was formed, possibly based on the slave as the multiple of the buffalo. This multiple is threefold of the bat, in that respect offering a strange analogy to the gold talent of Sicily, Magna Graecia, and Macedonia, which is the threefold of the Homeric ox-unit, and which, as I have conjectured, may have represented the value of a slave, as we certainly know as a fact that the highest unit in the Irish system, the cumhal, which represented the value of three cows or three ounces of silver, was neither more nor less than an ancilla (or ordinary slave-woman): the tenfold of this tael was the highest unit employed for either gold or silver by the most advanced peoples in this region, and is very well known as the nên or bar. All other goods were long appraised by measurement, the lowest unit of capacity being the cocoa-nut or the joint of the bamboo, the former known certainly to the Cambodians, the latter to the Chinese, whilst both are equally familiar to the Malays. The weight of the contents of the bamboo or cocoa-nut was presently taken, the standard employed being the tael, or highest unit yet employed for the precious metals. The weight of the contents would depend on the nature of the substance or liquid employed, for instance rice or some other kind of grain, or water. Thus the Chinese equate their catty to 16 taels; no doubt too convention came in at a later stage, and even though the contents might not actually weigh 16 taels, it was found convenient for practical purposes to regard some suitable multiple of the tael, such as 16, as the legal weight of the catty. A similar process was carried out in the case of the picul in the more advanced communities; a load was equated to the most convenient multiple of the catty, and as it was found that 100 catties gave a sufficiently near approximation to the ordinary load which a man could carry on his back, 100 catties were made the legal contents of the picul of trade.

We also learned how currency in baser metals such as copper or iron takes its origin. The history of the ordinary copper cash of the Chinese, which can be clearly traced step by step, brings us back to a time when a bronze knife, one of the most requisite articles of daily life, formed the ordinary small currency of the Chinese, just as the Greek obolos originally was an actual spike made of copper or iron, and just as the Bahnars of Annam still use the hoe as their lowest monetary denomination, an implement likewise similarly employed by the Chinese at an early period, as miniature hoes at one time used as true currency put beyond doubt. We also saw the negroes of Central Africa employing iron made into pieces ready to be cut into two hoes, and we also found those on the West Coast of Africa and the Hottentots employing bars of iron in a raw state, as a kind of currency. We also saw one most important feature possessed by all those in common, viz. the fact that in the determination of the value of the bar, the ingot, the piece of iron made in the shape of two hoes, and the bronze knife, not weight but linear measurement based on the parts of the human body, was the method invariably employed.

We then advanced to Western Asia and Europe and found everywhere alike the weight standards fixed by means of the seeds of plants. The process likewise was made perfectly plain. We did not find the highest denomination taken as the unit and the lowest reached by a long process of subdivisions, and finally for convenience sake described as consisting of so many grains of corn, as the brilliant French savant assumes in the case of the Assyrians: on the contrary we found that the bushel of Henry VII. was reached by first fixing the weight of the penny sterling by means of 32 grains of wheat, round and dry and “taken from the midst of the ear of wheat after the old laws of the land.” Again the Irish Kelts did not say that the unga or ounce must contain so many screapalls, and each screapall so many pingiuns, but they proceeded in quite the reverse way first fixing the weight of the pingiun by eight grains of wheat. We may then well assume that such too was the process among Greeks, Romans, Arabs and Hindus. Brahmegupta, and the legislators quoted above support this view by starting always with the smallest unit. It is only when we come to the system of Babylon we are asked to reverse the process, to admit that the idea of weights began with corn, the very commodity of all others which, according to all the instances previously quoted, was the last to be valued by weight, and which even amongst ourselves at this present moment can hardly be said to be regarded as an article appraised by weight. But furthermore if the Assyrians regarded the Talent as their unit, and their lesser denominations as its subdivisions, why did not the maker of the weight mentioned above inscribe it as ¾ obol, or by some other term to indicate that it was essentially regarded as a fraction of a higher denomination, and not as a multiple of a lower? But the ancient Assyrian who made the weight must plainly have regarded it in the latter light, for otherwise he would not have engraved on it 22 grains ½, actually resorting to the fraction of a grain. The only reasonable explanation of his conduct is that he was as firmly impressed with the idea that the basis of his system was the grain of corn (wheat) as were Brahmagupta, or Henry VII.’s parliament with the idea that the barley-corn and wheat-corn were the bases of their respective systems. If the objection be raised that the grains of corn were only devised in days long after the scientific fixing of weight standards, my answer is that if it was necessary to employ natural seeds as a means of determining the accuracy of scientifically obtained units, à fortiori it was necessary for mankind to have employed such seeds as their first step in the establishing of a system of weights.

No simpler idea connected with weight could have struck the primitive mind. The difficulty experienced by savages in counting beyond 3 or 4 is met by them by the use of counters. We are all familiar with the use of pebbles or small stones among the Greeks and Romans. Our own word calculate is simply an adaptation of the Latin calculare to count by pebbles (calculi). Some nations, probably all, have been unable to form abstract names for their numerals, and the name of the concrete object which they habitually employed as a counter has become firmly embedded as a suffix in the names of their numerals. Thus the Aztec numerals end in tetl, a pebble, because they employed small stones as counters. Similarly the Malays whom we found weighing gold by means of grains of padi employ that word as a numeral suffix, because they employed grains of rice for their calculations or, to speak more accurately, seminations. In the case of this people we find coincident the most primitive forms of numeration and of weighing, both processes being carried on by means of the same simple instrument, which Nature put ready to hand in the corn which formed their daily sustenance.

If any one still maintains that the Indian Islander or Tapak of Annam learned the art of weighing by grains from the Chinese, and would maintain that the latter either invented for themselves or borrowed from Babylonia a scientifically devised weight system, I will go a step further and try to produce some evidence of the process by which weight standards are arrived at, by seeking instances in a region so isolated as to be beyond the reach of all suspicion of having borrowed from Babylon.

From what I have said above, we cannot expect to find any such community in the Old World. The New World on the other hand supplies us with what we desire. When the Spaniards under Cortes, conquered the Aztecs of Mexico, that people, although in a high state of civilization, had as yet no system of weights. In consequence of this want the Spaniards experienced some difficulty in the division of the treasure, until they supplied the deficiency with weights and scales of their own manufacture. There was a vast treasure of gold, which metal, found on the surface or gleaned from the beds of rivers, was cast into bars, or in the shape of dust made part of the regular tribute of the southern provinces of the empire. The traffic was carried on partly by barter, and partly by means of a regulated currency of different values. This consisted of transparent quills of gold dust, bits of tin cut in the form of T, and bags full of cacao containing a specified number of grains[248].

From this we get an insight into the first beginnings of weights. Some natural unit (and by natural I mean some product of nature of which all specimens are of uniform dimension) is taken, such as the quill used by the Aztecs. The average-sized quill of any particular kind of bird presents a natural receptacle of very uniform capacity. These quills of gold-dust were estimated at so many bags containing a certain number of grains. The step is not a long one to the day when some one will balance in a simple fashion quills of gold dust against seeds of cacao, and find how much gold is equal to a nut. Nature herself supplies in the seeds of plants weight-units of marvellous uniformity. If any one objects to my assumption that the Aztecs were on the very verge of the invention of a weight system, my answer is that another race of America, whose political existence ceased under the same cruel conditions as that of their Northern contemporaries, I mean the Incas of Peru, who were in a stage of civilization almost the same as that of the Aztecs, had already found out the art of weighing before the coming of the Spaniards, although they were inferior to the Mexicans in so far as they had not a well-defined system of hieroglyphic writing, nor of currency such as the latter possessed. Scales made of silver have been discovered in Inca graves[249]. The metal of which they are made shows that they were only employed for weighing precious commodities of small bulk.

Unfortunately I can find no record of weights having been found along with the silver scales in the Inca graves. If the weights were simply natural seeds, they would easily perish, or even if perfect when the tombs were opened, would be simply regarded as part of the ordinary supply of food placed with the dead in the grave. But I forbear from laying the slightest stress on negative evidence of such a kind.

But beyond doubt we have on the American continent, far removed from connection with Asia, a series of facts closely harmonising with what we have found in Further Asia, and also among the peoples of Hither Asia, Europe and Africa. The Aztecs are still measuring gold, but the Incas have invented the balance. The Incas have no alphabet, the quipus as yet being their greatest advance towards a means of keeping a record of the past. It follows that it is possible for the human race to invent a system of weighing before it has made any advance in letters or science. Hence it is logical to infer that the civilized races of Asia and Europe could have discovered a means of weighing gold long before the Chaldean sages made a single step in their astronomical discoveries, or a single symbol of the cuneiform syllabary had as yet been impressed on brick or tablet.

Weights of various grains.