If the chart had been made the full width of the page it would have been possible to get room enough to show the figures for single years at ages under five by using a space only one-fifth of the horizontal distance used for the five-year intervals. The large number of deaths at ages five to nine inclusive is very striking on the curve. Possibly the large death rate from five to nine may be due to the lessening of parental care at an age when exposure becomes more frequent. By ten years of age, the children have learned better how to take care of themselves and the number of deaths from pneumonia comes down to about the lowest point. Though the foregoing explanation of the large number of deaths from five to nine years may be correct, it is probable that the figures are more or less in error, due to the tendency to state ages in numbers

FREQUENCY CURVES

per looo

which are multiples of five. The peaks for the period five to nine may be due largely to parents giving the age roughly as "five years". In Fig. 148 the data in which the reader is interested are shown at the peaks of various triangles. The shaded triangles on the chart give a geometrical figure which at first glance might be considered as a curve. It is not until after a considerable amount of puzzling that one notices that the triangles have absolutely no significance and that they are only a means of showing the distance from the base line to the various points representing decrease or increase. It would have been better if plain black bars had been used for Fig. 148 instead of the triangles. Bars are so familiar to everyone that there would be no danger of error in interpretation.