238 G. H. KNIBBS. 



earlier ages. Italy and Sweden shew a similar decrease of 

 frequency, the maximum being between the ages 60 and 70, 

 so also do Massachusetts and Buenos Ayres. With females 

 the maximum frequency is by no means well-defined, and 

 the aggregate numbers for Australia are too small to justify 

 any conclusion as to the age at which the tendency occurs 

 with females. 



It would appear from the preceding tables that the 

 measure of the stress of life is special to each country, and 

 that, measured against the capacity to endure this stress 

 it falls off in Australia at the age of about 60 to 65 for men, 

 becoming even at 90 as small as it was at 47 for that sex. 

 It also distinctly decreases for the later years of life, for 

 women. It is further worthy of note that the average 

 frequency among Australian women never reaches, at any 

 period of life, the average frequency at the age of 23 

 among men. 



5. Seasonal fluctuation of suicide.— The actual numbers 

 committing suicide are of practically little importance for 

 the purpose of disclosing the seasonal fluctuation. This 

 fluctuation is perhaps best shewn where, as in Australia, 

 the numbers are small, by computing for a period of years 

 how many persons commit suicide per month. Two cor- 

 rections are necessary, one for the equalisation of the 

 months (so that equal periods of time are represented), the 

 other to reduce the numbers to what they would be for a 

 constant population. The scheme of applying these correc- 

 tions is fully indicated in my paper on " Statistical applica- 

 tions of the Fourier Series," this Journal, Vol. xlv, pp. 

 76 — 110. Briefly it may be said that it is preferable to 

 reduce the monthly numbers to that value which would 

 correspond to the population at the middle of the period 

 under consideration, that is to say, all results from January 

 to June inclusive have to be increased (in decreasing 



