The data for Fig. 130 and for Fig. 131 were plotted on the of giving, monthly, the average number of accidents in each department during the last twelve months for which records were available. Actual figures in tabulated form were used to show for immediate reference the number of accidents in any month. The curves on the twelve-months average basis were consulted to determine whether there was any great increase or decrease in any department.

In Fig. 132 the vertical scale was omitted, perhaps with the idea that the chart would thus appear more simple to the average indi-

COMPONENT PARTS SHOWN BY CURVES

vidual attending a large exhibition. If the scale, were given, however, it would be plotted on the basis of actual horse-power rather than on percentage, for the top curve here represents the total quantity. Percentage scales cannot well be used in diagrams of component parts if a fluctuating curve line instead of a horizontal line representing 100 per cent is given at the top. The reader may, however, get a fair idea of the percentages if he roughly calculates the height of the areas in question on any vertical line of the co-ordinate ruling and then, using that vertical line as a measuring rod, estimates the height of the areas as a percentage of the total height of the chart.

In Fig. 133 the straight line at the top of the chart does not have any significance, as it is due only to the co-ordinate ruling which serves as a background to the chart itself. The important part of the chart ends at the top of the shaded area. We may consider the top of the whole shaded area as a curve and read the values accordingly from the scales on the right- and left-hand sides.