The method of reading this curve chart is as follows: at the end of six months, or 50% of the time period, there had been 82% of the total annual run-off of the St. John River. If the distribution was the same for every month, at the end of six months 50% only would have run off.

REFERENCES:

Lorenz, M. O., "Methods of Measuring the Concentration of Wealth," Publications of the American Statistical Association, Vol. IX, 1905. This issue of the Journal is so limited that the American Statistical Association cannot sell it. However, it is available in most libraries.

i|i=iii=iii=iii=i|i

GRAPHIC PRESENTATION

PERCENT OF TOTAL NUMBER OF STORES

U. S. Department of Commerce, Bureau of ForeiRn and Domestic Commerce, "Small Scale Retailing," 1938.

Correlation of Proprietors and Nunnber of Retail Stores in the United States in 1933.

1. Because this chart is plotted on probability paperj the "line of equal distribution"

assumes an "s" shape rather than a straight line.

2. It is clear from this chart that the number of proprietors of retail stores correlates

closely with the number of stores. The figures and curves show that nearly 33 per cent of all stores are operated by nearly 35 per cent of the proprietors -- who operate stores of less than $3,000 annual volume.

When certain series of observations showing frequency data are plotted on arithmetic probabiHty paper, the points do not fall in a straight line, but in a curve. Plotted on probability paper with a logarithmic scale as the ordinate, the points may fall approximately in a straight line or a gentle curve. In order to benefit from the use of probability paper, it is not necessary that the plotted points fall exactly in a straight line. If the curve is so gentle and uniform that it may be extended beyond the limits of the plotted points, it will usually be found sufficient.