In selecting a scale for Fig. 110 the draftsman is torn between a desire to show the San Francisco fire peak at its correct height, and an opposing desire to show on a large scale the data for a whole series of years so that the fluctuations from month to month maybe clearly defined. It seemed best to cut off the top of the San Francisco peak so as to show the monthly data on a scale large enough to assure clearness. To get the correct slope for the two sides of the San Francisco peak, a piece of paper was laid down adjoining the chart and a point was located in the correct position for the top of the San Francisco peak. The sides of the peak were then drawn so that they would meet at that point if extended. Even though the figures for the San Francisco peak are given at the top of the chart where the peak is

Edwin D. Dreyfus, in Jndyslrial Engineering

Fig. 107. Mean Temperatures for Each Month in the Year at Different Cities. Plotted from Records of the United States Signal Service and of Blodgett's Climatology of the United States

Information for eight different cities is so given that comparisons are easily made. Note the different Hncs used here instead of colors. Since Pittsburgh was the city under prime consideration, the Pittsburgh line was made heavy ttiat it might be most clearly seen. This chart gives a great amount of data in smaller space than would be required to show the facts by any method other than the use of curves