Note on Square-Root Charts
F. R. MACAULAY' and other writers2 have noted a tendency for changes in the square roots of common-stock prices to be constant regardless of price level. This phenomenon has naturally suggested that when such prices are represented graphically the charts be designed so that vertical distances from the origin are proportional to the square roots of the prices indicated in the margins. There is some reason to believe that charts of this kind might also be useful in plotting other kinds of data. Assume that n sales are distributed at random over 1/p firms during some interval of time, and that u1, u2, *, u1/, are the actual numbers of sales made by the different firms F1, F2, , Fil, Then, a priori, the probability that a particular firm will make one of these sales is p, and the mean and variance of the u's will tend to be