sion of a prosperous year. Reference to the figures for the total yearly sales, however, shows that the sales for the third year increased only very slightly over the preceding j^ear. When we compare the sales of this branch house with the expenses, we get quite a different story from that read from sales curves alone. The increase in sales from the first year to the second year was very great and without a correspondingly great increase of expenses. Putting the figures for the first and second years into a ratio, sales increased 64 per cent, while expenses increased only 10 per cent. In the third year, however, the sales hardly increased at all, while the shape of the curve for expenses shows an almost constant increase. A little mental arithmetic will show further that with almost stationary sales the expenses were permitted to increase about one-seventh or, roughly, 14 per cent.

Notice that in Fig. 209 arrow-heads mark a desire that sales should go up. No arrow was used in connection with the curve for expenses.

270 GRAPHIC METHODS

It was permissible that expenses should go up in the second fiscal year, for the volume of sales increased very rapidly. The cost per unit of sales or the ratio of expenses to total sales had decreased greatly. An arrow pointing downward would have given a wrong impression as the total expenses had increased justifiably. It is better, therefore, to avoid using arrows than to use arrows which would mean nothing.