specified change occurs, if production becomes easier or more difficult, if a tax be imposed, or a regulation of any kind affects previous conditions, what changes will be necessary to restore the equilibrium? These are the main problems of Political Economy. To solve, or attempt to solve them, we have to describe accurately the existing mechanism, and to suppose that it will regulate itself on the assumption which I have indicated as to demand and supply, the flow of capital and labour, and so forth. To go beyond these assumptions, and to justify them by psychological and other considerations, may be and is a most interesting task, but it takes us beyond the sphere of Economics proper. I must here diverge for a little, to notice the view of the school of economists which seems to regard scientific accuracy as attainable by a different path. Jevons, its most distinguished leader in England, says roundly, that political science must be a "mathematical science," because "it deals throughout with quantities"; and we have been since provided with a number of formulæ, corresponding to this doctrine. The obvious general reply would be, that Political Economy cannot be an exact science because it also deals throughout with human desires. The objection is not simply that our data are too vague. That objection, as Jevons says, would, perhaps, apply to meteorology, of which nobody doubts that it is capable of being made an exact science. But why does nobody doubt that meteorology might become an exact science? Because we are convinced that all the data which would be needed are expressible in precise terms of time and space; we have to do with volumes, and masses, and weights, and forces which can be exactly measured by lines; and, in short, with things which could be exactly measured and counted. The data are, at present, insufficiently known, and possibly the problems which would result might be too complex for our powers of calculation. Still, if we could once get the data, we could express all relevant considerations by precise figures and numbers.