There is no doubt that fluctuations in prices of securities, cereals and staples are frequently used as a basis for mere gambling transactions. But the most remarkable feature of the whole problem is the fact that the percentage of loss in transactions is _greater_ than the mechanical percentage. In the work already mentioned, the writer undertook to establish this. In 500 accounts examined, there was a loss of $1,245,000, and profits of $288,000, leaving a deficit of $957,000. The commission charges and interest amounted to only $275,000. There thus appeared a loss of $682,000 which could not be attributed to a gambling percentage. It may be added that the period considered in the computations was from July, 1901, to March, 1903, and that the price of the stock considered (U. S. Steel Common) was the same at the beginning and the end of the period. This tends to again refute the theory of mere gambling, with a ruinous percentage against the player, for no mechanical device could by any possibility operate against the player to a greater extent than the fixed percentage in favor of the machine. A gambling machine will stick to its knitting. If, for example, we take the simplest form of gambling device--two dice thrown from a cup,--we know that certain numbers formed by adding the total spots which appear uppermost will show more frequently than others. Thus the number two can be effected in but one way, the number three in two ways, the number four in three ways, and so on up to the number seven, which can be formed by six different combinations, thus: 4 and 3 5 and 2 6 and 1 3 and 4 2 and 5 1 and 6 from which point the chances decrease until the number 12 can be formed in only one way--two sixes. This proposition applies to all forms of mechanical gambling, and is so simple in principle, and so distinct in operation that if we make a fair number of casts, say 1,000, and do not make more sevens than any other one number, we may be positive that the dice are defective, or loaded. Therefore, if percentages hold true, we must attribute the surplus loss in speculation to mental operations. In the total results mentioned, these mental operations were so erroneous as to cause a loss greater than the percentage itself; but, on the other hand, a certain number of accounts showed profits; that is to say, the percentage was overcome, which is again an obvious impossibility in true gambling.