Page:Graphic methods for presenting facts (1914).djvu/204

 *venient for general use. Here the paper is used to indicate 100 per cent in the vertical direction, and the horizontal scale is chosen for data unrelated to time.

Fig. 157. Cumulative Frequency Curve Study of the Number and the Size of All Orders Handled During a Ten-days Test in a Certain Kind of Freight-handling Work

The orders are sorted into various classes or groups according to the number of pieces called for by each order. The limits for the various classes were fixed by the exercise of a little judgment and are shown by the vertical lines on the chart

Curve "D" shows by small circles the actual number of orders found in each group. The curve is read from the right-hand scale

Curve "C" shows by small circles the total number of pieces (or packages) carried by the various orders which are found in each group. Curve "C" is read from the left-hand scale

Curve "A" shows the percentage of the total orders handled which contain more than any given number of pieces considered on the horizontal scale of the chart

Curve "B" shows the percentage of the total number of pieces carried by those orders containing more than any given number of pieces considered on the horizontal scale of the chart

Curves "D" and "A" refer to the number of orders or the amount of clerical work while curves "C" and "B" relate to the number of packages handled or the actual volume of business transacted

In Fig. 157 we have an application of the cumulative frequency curve to a class of work which would be extremely difficult to understand if the graphic method were not used. In handling large quantities of freight the different orders cover many diverse kinds of material, and the jobs vary from one package up to many thousands of packages on one order. The pieces or packages referred to here may be anything from a single casting weighing 20 tons to a shipment of canned goods with 5,000 boxes or cases on a single shipping order. In the long run, however, the different classes of goods could be averaged, and it will be found that in any one locality there would be but slight