milk production each month and the total milk production for each calendar year. The notes in the left-hand margin show the dates on which calves were born.
In Fig. 215, also, curves are plotted on a card for five years by months. In this case there was no hope of getting the cost per ton below 50 cents, and the scale was accordingly so chosen that it began at 50 cents instead of at zero in order that fluctuations from month to month might be more carefully observed. As the zero line was not shown on the card, the wavy line was drawn at the bottom of the ruled portion, indicating clearly to the reader that he must not interpret the curve as though the bottom were at zero.
When using an arrangement showing five years by months on one card, considerable mental effort is necessary to get a clear interpretation of the fluctuations in the curve from year to year. In Fig. 215 certain peaks in the curve appear to have somewhat similar shape. Thus the peak for 1908 looks like the peak for 1907 until closer examination shows clearly that the low point for 1908 was in July, while the low point following the peak of 1907 came not in July, 1907, but in January, 1908. The waves themselves, although of somewhat similar shape, have peaks at entirely different times, the peak for 1908 being in February and the peak for 1907 in April. The foregoing examples may serve to point out the mental effort necessary in reading horizontally if the danger of misinterpretation is to be avoided. If the curves for these same years 1907 and 1908 were plotted on 4-by-6-inch cards, and one card placed above the other as in Fig. 207, there would be no possibility of error on the part of the reader. The eye would follow the vertical lines and see at once that there was no great similarity in the two waves. Having five separate cards causes the reader to take more time in handling cards in order that he may save mental effort and avoid error in interpreting the yearly waves. Having five years on one card saves handling the cards, but it takes more mental effort to be accurate in reading the curve horizontally. The choice between five cards of one fiscal year each and one card for five years must rest with the judgment or the habit of mind of each individual person.
It sometimes happens that the seasonal wave in a curve is almost completely obscured by the tendency toward a very great increase or decrease in the business. Conditions may give so large an increase in volume of business in any one year as to more than offset any de-
