If I were treating of the problem of an ideal standard of value, I think I should be inclined to agree with Professor Marshall that a standard that represents a gradually descending scale of prices to keep pace with the "real" cheapening improvements in industrial processes is better than one which represents an absolute constancy of prices. But it would be quite impracticable to discover the exact rate of fall of prices which would correctly register the improvement going on in industry, and, moreover, it would, I believe, be so small as not to depart much from the mutiple standard. This I infer is also the opinion of Professor Marshall. Professor Kinley makes the very interesting suggestion that we can suppose a more ideal standard than the tabular by making our unit a definite percentage of the national annual dividend. This appeals to me as a rough and ready way of fixing a unit more nearly ideal than that fixed by the tabular standard. But it would certainly not be practicable. It would not even be quite ideal. But if Professor Kinley will measure his standard, the compensated dollar plan will be able to take care of it. In fact, if we could find a more absolute standard than the tabular standard and could accurately measure it in statistics, precisely the same method of compensating the dollar could be employed to keep the dollar in tune with that standard as with the tabular standard. The only difference would be that the guiding index would be different. The plan for compensating the dollar does not in essence consist in selecting the multiple or any other standard. It consists in a method of making the monetary unit conform to any standard chosen. But there is convincing evidence that the multiple standard is usually near enough to the ideal for all practical purposes and infinitely nearer than the gold standard. _While individual goods may vary greatly in absolute value, the general mass of goods will vary comparatively little and seldom._ There may be some absolute change in the general mass of commodities, but it must usually be extremely small in comparison with changes in any one commodity like gold. It is clear from the theory of chances that this must be the case. The odds are hundreds to one that the variations in absolute value in several hundred commodities will offset each other to a large degree. We very seldom have world feasts or world famines. If the corn crop is short in some places it is abundant in others. If it is short everywhere the crop of wheat or barley or something else is practically certain not to be. We cannot expect that everything will usually move in one and the same direction. If there is a war in Japan, it is not likely that there will also be a war in India. A world war or even anything as near to a world war as the present conflict in Europe is a most unusual thing.