Jevons’s method is based on the assumptions that the proportions of coins issued at different dates found in the given samples roughly correspond to their proportions in the circulation at large, and that the numbers in circulation of the latest issues do not much differ from the numbers issued from the Mint. In short, if we know the relative proportions of coins of 1860 and of 1912 in the circulation, and if we know, approximately, the absolute number of coins of 1912, we can calculate the absolute number still circulating of the coins of 1860. In applying this method to the Indian data, we are assuming that the proportions of rupees of each date found in the bags examined in a great number of scattered Government Treasuries are a fair sample of the proportions still in circulation throughout the country. In a country such as India, however, there may be great stagnancy in a part of the circulation, and the coins finding their way to the Government Treasuries may be a sample rather of the floating surplus of coinage, which has a relatively high velocity of circulation, than of the total stock, which includes semi–hoards passing from hand to hand comparatively seldom. Since these samples are likely, therefore, to contain an undue proportion of recent issues, estimates of the total circulation, which are based on them, may be expected to fall short of the truth rather than to exceed it. There is reason, also, for supposing that in some cases the officials charged with the duty of examining the samples did not always deal with them conscientiously. A tendency was noticed for the returns of one year to resemble those of the previous year more closely than they should, and not infrequently a batch of coins would be attributed to a year in which it is known that none were minted. Nevertheless the calculations of Mr. Harrison and Mr. Adie, and the data on which they are based, seem on the whole coherent, and bear, so far as one can judge, the marks of substantial accuracy.