250 WORLD-POWER AND EVOLUTION 



and must be plotted with more reliance upon the judgment of the 

 investigator as has been explained in Appendix A. 



After the normals had been obtained, the third step was to find the 

 percentage by which the actual mortality exceeds or falls short of the 

 normal. For example, in April, 1914, the actual number of deaths in 

 Chicago was 3,298. Since April has only thirty days this number 

 needs to be corrected to 3,408 so as to show how many would have 

 occurred if there had been thirty-one days. The normal mortality 

 for 1914, as shown by the straight line in Figure 29, is 2,998. If we 

 reckon this as 100 per cent, the true mortality was 118.8 per cent. 



80,000 



Figure 28. Annual Deaths in Chicago, 

 1900-1915 



When the true mortality for each month at each place had thus 

 been reduced to percentages of the normal the fourth step was to 

 tabulate the percentages according to the temperature and humidity 

 of the months in which they occurred. To take an actual example, 

 among the twenty-eight cities of France and Italy a month having a 

 mean temperature of 19°C and a mean humidity of from 61 per cent 

 to 65 per cent occurred ten times. The places where these conditions 

 occurred are shown in column A of the following table. Column B 

 shows the mortality in percentages, and C the weight to be assigned 

 to each place according to the number of inhabitants. The weighted 

 average of column B, 91.4 per cent, indicates the mean mortality rate 

 in French and Italian cities when the temperature averages 19°C and 

 the humidity 61 per cent to 65 per cent. The average may equally 

 well be expressed as a departure of — 8.6 per cent (100 per cent — 91.4 

 per cent) from the normal. 



