sales increased over 50%, but total trade, as shown by Fisher, increased only 5%. In the following year, stock sales _decreased_ over 21%, but total trade, on Fisher's figures, _increased_. The following year, 1901, stock sales virtually doubled, but Fisher's final figure shows only an increase around 13%. Two years later, in 1903, stock sales fell off about 40%, from the figures for 1901, but again, as compared with 1901, total trade on Fisher's figures shows an appreciable gain. The influence of stock sales on Fisher's index is, virtually, negligible. The dominating factor is the _receipts_ of selected staples, cattle, cotton, rice, pig iron, etc., in the principal cities of the United States. There is not a _single year_ in which his final figure for T does not move in harmony with this factor (p. 479). He gets, thus, for the volume of trade through the fourteen years under consideration, a surprising steadiness, and a pretty uniform progressive development. In defence[247] of his method of weighting, Fisher says, simply: "These weights are, of course, merely matters of opinion, but, as is well known, _wide differences in systems of weighting make only slight differences in the final averages_." (Italics mine.)[248] Are these figures valid? Well, first one is struck with the absolute magnitude assigned to T. The figures seem vastly greater than would have been anticipated. The method of calculating it, for 1909, I shall discuss in detail in the chapter on "Statistical Demonstrations of the Quantity Theory." For the present, it is enough to note that the absolute magnitude is derived from figures collected by Dean David Kinley for the National Monetary Commission,[249] of deposits, exclusive of deposits made by one bank in another, made in about 12,000 banks (out of 25,000) on March 16, 1909. These deposits were classified as (1) money (with subdivisions) and (2) checks and other credit instruments. A cross-classification divided them into (1) retail deposits; (2) wholesale deposits; (3) all other deposits. Kinley's object was to determine the extent to which checks are used, as compared with money, in payments, particularly in wholesale and retail business. Fisher's total, briefly, was obtained as follows: Kinley's figures, for the one day, were increased to make an allowance for the non-reporting banks; they were further increased on the assumption that March 16 was below the average for the year; the figure finally obtained for the day was then multiplied by 303, assumed as the number of banking days in the year, and the product, 399 billions, was taken as representing the total circulation of money and checks in trade. For some reason not made clear, this total was subsequently reduced to 387 billions. Counting the average price, P, as $1, T was considered to be 387 billions.[250]