380 On the Influence of Vaccination 
The foregoing results may now be collected in the following Table. 
TABLE X. 
Coefficient of Correlation, r, between effectiveness of vaccination and strength 
to resist the disease. 
For the 6 towns enumerated -6561 ± -0092 
Sheffield -7694 ±-0124 
Leicester ... ... ... ... •6112 + -OT^S 
Gloucester -5897 ±-0198 
Homerton and Fulham Hospitals -5760 ± -0089 
„ „ „ -6615 + -0083 
Glasgow -7783 ± '0365 
London, 1901 Epidemic ... "5779 ±-0311 
London, 1892-3 Epidemic* ... -5954 ±"0272 
Doubtful cases 
Included in vaccinated 
Excluded 
In Shetheld and Glasgow the correlation is nearly the same, and considerably 
higher than elsewhere ; in the other towns it is remarkably uniform, the coefficient 
approximating to '6. It will also be noted that the correlation in the present 
epidemic is nearly the same as that in the epidemic of 1892-3. We have clearly 
in this coefficient a fairly stable statistical constant for smallpox epidemics. 
2. Coming next to the correlation between degree of effective vaccination and 
type of disease, I divide the types into two classes, (1) Mild, = mild, varioloid, and 
discrete, and (2) Severe, = coherent and confluent, and exhibit in Table XI. the 
statistics of the cases whose types were observed in the Sheffield, Dewsbury, 
Leicester and Warrington epidemics. The London figures for 1892-3 are excluded 
because a somewhat different classification was adopted there. No figures appear 
to be available for Gloucestei-. (See Report of Commission, pp. 66 — 69.) 
TABLE XL 
Sheffield, Dewsbury, Leicester and Warrington. 
Mild 
Severe 
Totals 
Vaccinated . . . 
2229 
505 
2734 
Unvaccinated 
229 
804 
1033 
Totals 
2458 
1309 
3767 
h 
= •39212, 
k = 
= -60009, 
whence 
•001160?-' + •0o0798r" + •01G967r= + -073650?^^ + ■090250?-=' 
+ •ll7624r2 + ?'= •959775, 
r= -7935 + -0093. 
* Pearson : Phil. Trans. Vol. 195, p. 43. 
