W. K. Macdonkli. 



137 



'l'he stalisdcs also ciiablc nie to lind Llic cüetKcieut of correlatiou betwtjcn agc 

 am! severif-y of attack. 



TAÜLK Vir. 



VacciiKited Cases. 



»•=•3693 + -0386. 



If tho divisioii be made at 25 years instead of 20, r = -3297 + -0314 ; if at 35 yeara, 

 r = •3180 + •0293. Caiculating this cocffieieiit for the Glasgow epidemic of 1892-05, 

 I find ?• = •33Gß + •0715, for a division at 20 yoars ; I am nnablo to calculate it for 

 a division at 25 or 35 y(,'ars, as thore an? iio divisions at these periods in the 

 statistics of the earlier epidemic. 



(2) For the sake of comparisoii, the coefficieiits for the two Glasgow epidemica 

 are collected in the table below. 



TABLE VIII. 



It will be observed that in the recent epidemic the coefficient of correlation 

 between vaccination and strength of resistaiice, while less than in 1892-95, was 

 very much the same as in previous epidemics in other towns*. On the other 

 hand, the correlation between vaccination and degree of severity of attack is very 

 much less than in 1892-95, which points to a marked difFerence in the character 

 of the two epidemics, the earlier being the milder. Also the correlation between 

 degree of severity and (1) foveation, (2) area, and (3) nuniber, of scars is less even 

 than in 1892-95; but I understand from Dr Brownlee that in conclnding whether 

 a scar is good or bad, clinically, he would take into consideration its area. 



* Biometrika, Vol. i. No. 3, p. 380. 



Bioriieti'ika it 



18 



