very full and valuable disquisition on "Ancient Indian Weights and Coins," that they are now easily accessible; and I am glad to be able to refer the reader to that Essay, in which he has so successfully upheld the independent origin of the ancient Indian coinage, which I have always advocated.

It now only remains to notice the relative values of the three metals, gold, silver, and copper, of which these coins are made. In the time of Alexander the price of gold was ten times that of silver; and the gold stater, or didrachmon, was equal to 20 silver drachmas. In India, where gold was found in considerable quantities, while silver was comparatively scarce, the price of gold was only eight times that of silver under the native rule. This is shown by the valuation of the gold suvarna of 140 grains at 25 silver kárshas of 44.8 grains pure metal; as 44·8 × 25 = 1120 grains, which, divided by 140, gives 8 rates exactly. In treating of the relative values of the Greek and Indian money, I have assumed that the silver coins of Alexander were quite pure. This is not exactly the case; as the result of several assays shows that they contain only 96 per cent. of silver, and 3 per cent. of alloy. But as gold has been found in these coins to the extent of 24, or one-quarter of a grain,32 it seems to me almost certain that the silver money of Alexander was as pure as the scientific skill of his workmen could make it, that is always supposing the presence of the gold to have been unknown. If, however, the presence of the gold was known, the value of grain would be 2 grains of silver, which would partly cover the deficiency in value of the alloy. But I fully believe that the presence of the gold

32 Hussey, Essay on the Ancient Weights and Money, p. 71.

4

was quite unknown, and that the silver was honestly esteemed to be quite pure.

The coins of the Greek kings of Bactria appear to follow the same standard; but with the use of the ArianoPali alphabet, the silver coins of the Greek kings of Kabul and India become somewhat heavier, 16 good didrachmas of 7 different kings averaging 146.6 grains, and numerous hemidrachmas of 17 kings averaging 36-35 grains. The full weight of the hemidrachma was therefore not less than 36.5, or perhaps 37, grains. But this was not all pure silver, as I found that 70 hemidrachmas of Apollodotus and Menander, assayed at five different times, gave an average weight of 35.58 grains in weight, but only 32.78 grains of silver. Assuming the full weight of the hemidrachma at from 36.5 to 37 grains, the amount of pure silver in each coin, at the above rate, would have been from 33.6 to 34 grains, which agrees with the Attic standard of 33.6 grains for the hemidrachma, and 67.2 grains for the drachma, which I have adopted in this disquisition. I have recently melted 106 hemidrachmas, from the Sonpat find, of Heliokles, Straton, Antimachus II., Antialkidas, Apollodotus, and Hermæus, besides 475 hemidrachmas of Menander, which gave almost the same result as the previous assays. The actual value of the later coins was therefore the same as that of the earlier ones, the alloy having no doubt been purposely added, as in our modern European coinage, for the purpose of hardening the silver. The amount of alloy was probably fixed at one-tenth, which would have increased the weight of the hemidrachma from 33.6 grains of pure silver to 36.96 grains of hardened silver, which agrees with the full weights of 37 grains of the best preserved specimens.

After the Greek occupation, the relative values of gold

and silver in North-west India must have changed from 8 to 10 rates. This was only the natural consequence of the redistribution of the great hoards of silver money obtained by Alexander in Persia, where the rate of gold to silver was 13 to 1. The result of this change was a slight fall in the value of the silver karsha of India. Before the time of Alexander it had been worth 4th of 44.8, or 5.6 grains of gold; but after the Greek occupation it was worth only 'oth of its weight, or 4.48 grains of gold; and as the silver karsha was only equal to two-thirds of the Greek drachma, the value of the stater in Indian money became 30 silver karshas, at 1 karsha to the drachma.

Of the price of copper in Greece the learned Böckh was "unable to find any definite statement."33 But from the value which I have now assigned to the lepton of seven-tenths of a grain of silver, or 33.6 grains of copper, which is exactly half a drachma in weight, the relative proportion between silver and copper in Greece was 1 to 48. In India at the same time it was 1 to 50, the karsha of 44.8 grains of silver being worth 16 panas of copper of 140 grains each, or 16 x 140 = 2,240 grains of copper were equivalent to 44.8 grains of silver, which gives exactly 50 rates. The small difference of 2 rates between 48 and 50 is caused by the difference of weight between the Indian pana of 140 grains, and the Greek dichalkon of 1344 grains. The copper coins of the Greek kings of Bactriana adhere to the Greek standard, but those of the Greek kings of Ariana and India would seem to have been raised to the Indian standard. The following list shows this result very clearly :

33 Public Economy of Athens, p. 30.

The Greek coinage of India would thus appear to have been assimilated very early with the indigenous copper money of the country. I have already pointed out that the square copper money of Pantaleon and Agathokles of about 180 grains weight was an actual Indian coin mentioned by Manu, and equal to 1 pana.34 To this evidence I will now add the large copper pieces of Demetrius, of which three specimens weigh respectively 364, 359, and 357 grains, giving an average of 360-16 grains, or exactly 2 panas, which is another of the coins mentioned by the Indian lawgiver Yajnavalkya as a fine.35 Similarly my Horse coin of Menander, which now weighs 679 grains, must originally have weighed about 700 grains, or just 5 panas, a sum which is also mentioned by Yajnavalkya. The Dolphin coin of Menander, weighing 343 grains, and the Ox-head coins weighing 341, are, I think, further examples of the 2 pana pieces. The large Victory coins of 246 grains are perhaps intended for 2 pana pieces, named dripana, of 280 grains, although it is not improbable that they were hemiobols, or pieces of 4 chalki of the Greek standard of 268.8 grains. But as this mixture of standard would have been extremely

34 In fact I possess several old Indian coins of this very weight.

35 English Translation of Code, ii. 297.

inconvenient, I conclude that the chalkous, which was the Greek copper unit, must have been raised from the Attic standard of 67-2 to 70 grains, so as to assimilate the two systems by making the Greek chalkous exactly equal to half an Indian pana.

In conclusion, I may mention that two at least of the Indian names of coins were not unknown to Western authors, as Hesychius calls the κέρσα, Ασιανὸν νόμισμα, and the κορσίπιον, νόμισμα παρ' Αιγυπτίοις, τὸ κερσαῖον λεγόμενον. The first of these is evidently the Indian karsha, and the second is the harshapa or karshapana. For παρ' Αιγυπτίοις I would therefore propose to read παρὰ Γυπτίοις, and to refer the name to the powerful family of Gupta kings. Now the work of Hesychius is generally considered to have been abridged from the larger lexicon of Diogenianus, who flourished in the second century of the Christian era, at the very time that the Guptas were at the height of their power, under Chandra Gupta II. and Kumara Gupta, whose rule extended to Surashtra and Bharoch, or Syrastrene and Barygaza, where their silver coins are still found.

As a means of convenient reference, I add a Table of Ancient Indian Coins, showing their relative values to each other, and their weights in English grains.