It is a mistake, indeed, to extend the results of statistical research very far beyond the time to which the facts and figures belong. It would be easy to multiply instances of the incorrectness of such a process. To take a single case.—When cholera has been extending its ravages in this country, the statistics of mortality from that cause, if studied with reference to four or five successive weeks, have indicated a law of increase, which is very readily expressed so as to accord with the mortality during those weeks, and perhaps two or three following weeks. But if such a law were extended indefinitely it might be found to imply nothing short of the complete desolation of the country by cholera, within the space of a few months. Thus, if the deaths (from cholera) in five successive weeks were 20, 27, 35, 47, and 63,—numbers corresponding with the general characteristics of cholera mortality in the earlier stages of a visitation,—the weekly mortality a year later, estimated according to the observed percentage of increase, would be more than 173 millions! Now this method of estimation, though leading to this preposterous conclusion as respects a more distant epoch, would probably lead to tolerably correct results for the next week or two after that in which 63 persons died,—the estimated numbers being 84 and 110 for the next two weeks respectively. It seems to me, therefore, that we are not justified, by the observed seeming fulfilment of Mr. Jevons’s anticipations, in concluding that a hundred years hence the consumption of coals will be 2,000 millions of tons, or that the total consumption during the next 110 years will be 100,000 millions of tons. We might almost as safely infer that because a growing lad requires such and such an increase of food year by year, the grown man will need a similar rate of increase, and the septuagenarian require so many tons and hogsheads of solid and liquid food _per diem_.