CORRESPONDENCE OF CURVES 



257 



between the curves of temperature or rainfall and thofee of flies or 

 deaths from diarrhoea. In addition he points out the manner in 

 which the curve of fatal cases or deaths falls away from the curve 

 of flies in the middle of the decline. 



In pointing out the intimate correspondence of the curves of 

 the cases and deaths to the number of flies captured week by 



WEEK ENDING 

 MAY JUNE 



JULY 



AUGUST 



SEPTEMBER 

 4 ! 19 26 



Fig. 100. Chart illustrating the relation of the numerical abundance of house-flies 

 to summer dian-hoea in the city of Manchester in 1909. Prepared from the 

 statistics given by Niven. 



week, Niven states : " The number of cases begins to increase until 

 the flies captured attain a maximum. The maximum number of 

 cases commencing in the years 1904 and 1905 is in the same week 

 as the maximum number of flies. The figures for the deaths are 

 still more striking. The shape of the curves at an interval of 

 a week to a fortnight near the maximum point is practically 

 identical in the two. The errors of the curves will be subjected 

 to examination afterwards. But, even with their manifest and 



H. H.-F. 17 



