Page:Graphic methods for presenting facts (1914).djvu/239

 *ing, as they appear to the eye, mean practically nothing. States having shading number two could vary from each other in the amount of horse power by more than the whole quantity of power in States having shadings from three to nine inclusive. If the steps in the shading scale had been so made that there were nine increasing classes of shading, each representing 1,000,000 horse power, it can be seen that all the States here numbered five to nine inclusively would have the same shading. Such a map made with a scale of uniform steps would appear so entirely different from the map shown here that no one would ever recognize the maps to have been made from the same data. The selection of scale for maps of this kind is important in order that the map may tell a truthful story. Wherever feasible, the numerical scale intervals for colorings or shadings should be uniform. Frequently the data are of such kind that there is much concentration at some portion of the scale, but with important facts to show at other portions of the scale above and below the point of greatest concentration. It may then be desirable to use uniform scale-steps for the shading in the portion of the scale with greatest concentration, and to have smaller scale-steps at either end or at both ends of the scale. In Fig. 177 small scale-steps would seem desirable for the lower portion of the scale.

It will be noticed in Fig. 177 that the numbers denoting the shadings are arranged with the smallest numbers to represent the largest quantities. This arrangement was made purposely. On a map of the kind seen in Fig. 177 there may be any number of different shadings, from one to a dozen or more. In order to simplify matters for the reader it seems best to assign the smaller numbers to represent those conditions which are considered most desirable or commendable. The reader may then see instantly which areas are first, second, third, etc., in rank simply by observing the figures inside the small circles.

In Fig. 178, if it be assumed that there is a likelihood of corn crops going up to 60 bushels per acre, the scale intervals are shown as uniform. Fig. 178 is of interest chiefly because it shows what can be done by hand ruling when it is necessary to produce an illustration for a report in which an expensive cut by the Ben Day method cannot be justified.

Fig. 179 does not do justice to the possibilities of the Ben Day method of shading. The cut was made by photographing a page of the Census Abstract which was printed on rather rough paper. With an original cut made directly by the Ben Day process the distinction