The failure to show the zero line at the bottom of a chart is so common a fault, found in nearly all publications, that some typical examples are shown here in the hope that a bad practice may be somewhat reduced.

Fig. 247 gives the data of Fig. 246 redrawn so that the zero line is shown at the bottom of the chart. It is believed that this illustration will prove conclusively how great an error may be made if charts are read hastily on the assumption that the bottom line of the chart is the zero line. Since some persons are almost sure to read a chart from the bottom line, it seems desirable that all charts should be so made that the reader may interpret from the bottom line as a zero line, or else receive positive warning that he should not do so.

temimi Indicating INCREASE In Deaths from Degenerative Diseases end DECREASE fivm Tubenulosjs

ferinm

Data of Elmer Rittenftouse

Fig. 248. Comparison of Death Rates in the United States, Showing Reduction in Death Rate for Tuberculosis and Increase in Death Rate for Degenerative Diseases

This illustration is' identical with Fig. 246 except that here a wavy line is used at the base showing that the bottom of the chart is not at the zero of the vertical scale. It is always desu-able to have the bottom line at zero. If that is not possible the wavy line should be used as a warning to the reader