SUICIDE IN AUSTRALIA \ A STATISTICAL ANALYSIS OF THE FACTS. 241 



Month 



Jan. 



Feb. 



March 



April 



May 



June July 



August 



Sept. 



Oct- Nov. 



Temp. Fait- 



71-1 



707 



68-4 



63-5 



577 



5361 51-8 



54-1 



577 



61-8 1 65-6 



Explanation of Graph. — The horizontal divisions denote not 

 calendar but equalised months and the vertical divisions denote 50 suicides 

 per month out of an assumed total of 10,000 per annum (833£ per 

 month). The rectangular lines denote the group results for the equalised 

 months on the basis assumed. The upper curve denotes the probable 

 instantaneous values for the whole of Europe, the lower the probable 

 instantaneous values for the whole of Australia. The results are so 

 corrected as to correspond to a population constant throughout the year. 



In the diagram the rectangular lines shew the rates for 

 the various months for Australia and Europe generally, and 

 the curves give the most probable form of the fluctuation. 



6. Correlation of seasonal fluctuation with tempersture. 

 If the monthly mean temperatures of the capital cities 

 of Australia be weighted in proportion of the populations, 

 the resultant mean is as follows: — 



Dec. 



69-0 



and these results may be regarded as approximately repre- 

 senting the temperature conditions influencing the rate of 

 suicides owing to the fact that the populations of the cities 

 have a preponderating place. 



It was shewn in a paper by the author, "On the appli- 

 cation of the Fourier Series to Statistics," see this Journal, 

 Vol. xlv, p. 109, that the frequency (q) of suicide per million 

 per diem in Australia could be put in the form 

 (5)...q == 0-329 + 0*00304 (t - 62°) 

 where t is the temperature Fahrenheit given by the curve 

 indicated in the above table: or say 



(5a). ..q = 0*33 + 0*003 t 

 where t is the temperature above 62° Fahr. 



Owing to the irregularity of the results in Australia 

 from month to month the remarkable correlation between 

 temperature and suicide frequency is best seen by combi- 



P— Sept. 6, 1911. 



