The card for 1910 in Fig. 211 has a broad line drawn by hand to show the zero line. On the card for 1911 the scale begins at 400, and, because the scale does not begin at zero, the bottom of the co-ordinate field is marked with a wavy line. This wavy line is made very rapidly with a pen and serves the very useful purpose of safeguarding the reader from interpreting the curve as if the lower line on the field were zero. Whenever zero is not shown at the bottom of the ruled field, this wavy line should be used. Any card can thus be read independently, with safety so far as its interpretation from the zero point is concerned. When several cards are laid out together, and the zero line is shown on the left-hand card, as in Fig. 211, it is a simple matter for the eye to imagine the zero line extended to the right below the other cards, thus permitting easy interpretation of all cards.
It sometimes happens, especially in plotting costs, that the desired direction of the curve will be downward instead of upward. In such a case, and to show small fluctuations from month to month, the scale on the curve cards may be so selected that the zero line does not appear even on the first card plotted. Should the costs be reduced so rapidly that the curve tends to run off the bottom of any card, the card for the succeeding year may have the joint line drawn in such a way as to allow an extension of the scale downward, exactly as Fig. 211 shows an extension of the scale upward. By drawing the left-hand joint line on the later cards above the bottom of the ruled space, and by putting the right-hand joint lines for the first cards at the bottom of the ruled curve field, the series of curves can be made to progress downward to any desired extent in exactly the same manner as the curves in Fig. 211 progress upward year by year. By the use of these joint lines a thoroughly universal arrangement of cards may be secured, allowing extra space for movement either up or down.
Joint lines were devised chiefly to permit of showing a large upward or downward progress in the curves for succeeding years without the necessity of laboriously replotting the curves for the earlier years. The desired result has been very satisfactorily attained. Should anyone, however, object to the presenting of a series of cards in steps as in Fig. 211, he need only replot the curve for the earlier years to some smaller scale. In general, however, it will be found that practically no replotting is necessary or desirable. By connecting the cards with joint lines and by using a wavy base line when the scale does not extend to zero, all necessary convenience and accuracy may be secured.
