722 THE ECONOMIC JOURNAL followed which is in any way ori?nal lies in the attempt to utilize facts that can be deduced from their serial ? nature. If it is asstuned that the circulation from 1876 to 1890 is approximately constant, it follows that the variations of the per- centages representing the several coinages which make up the mass circulating are approximately identical with their absolute move- ments. It is of course evident that any such constancy is not in complete accordance with the facts; for, assuming the circulation to be in equilibrium with requirements, and the annual loss to be a quantity varying according to an ascertained law, it is obvious that any fresh mintage other than normal will, as the case may be, swell or contract the quantity circulating in other words, the varying quantities of each coinage in existence are not precisely represented by the variations in their respective contributions to the circulation as shown by the percentages in my first table. There is also a disturbing element in the reappearance of coin from hoards in times of distress, and for other reasons from abroad. It is hoped that all arguments hereafter employed will be considered as subiect to these qualifications and defects. Before entering into the heart of the subiect it will be best to. give a short account of the manner in which these examinations have been conducted. In India there are, excluding native States, eight principal centres of Government namely, Assam, the North- ?Vest Provinces (including Oudh), the Punjab, Bombay, Madras, the Central Provinces, Burma, and Bengal. These centres are again divided into administrative units called districts. In each of these there are public treasuries, while each in turn has one or more branch treasuries, subordinate to it. In 1875 directions were issued to each head of a district to open a bag of one thousand or two thousand rupees at random, and to report its contents according to year of mintage. It was also suggested that a bag containing coin received from a rural part of the district should be selected when convenient, in order to better arrive at the composition of the ordinary census reduced tive result of this (Table A). The same results circulation. The to 100 is given collec- below are repeated in Table B, but the highest figure in the row corresponding to each year has been raised to 100, to admit of a comparison of the relative rates of disappear- ance of each year's mintage. ? l?[y best acknowledgments are due to Prof. F. Y. Edgeworth for valuable hints in this and other parts of the argument.
