Ethel M. Elderton and Karl Pearson 
553 
Thus both periods show significaut if not very large correlation*. The difference 
(•187 + '093) between the coefficients for the two periods is, however, probably 
not significant. Thus in towns with greater isolation-rate there is certainly a 
higher attack-rate, and equally certainly no argument can be based on the crude 
figures to prove that tiie more the isolation the less prevalent is diphtheria. We 
will now turn to the death-rate 31, and we find : 
First Period : 1904-1908, r,^„ = + -153 + '075, 
Second Period: 1909-1912, -012 ± -075. 
In the first period isolation was associated with a higher diphtheria death- 
rate, in the second period with a lower diphtheria death-rate, but neither ai'e 
of any real significance. Thus all we can conclude from the crude figures is 
that they show no evidence that isolation has reduced the general death-rate from 
diphtheria. 
We next take the case death-rate (m) and we have for the two periods : 
First Period: 1904-1908, ?v,„ = - -509 ± "057, 
= - -527 + -056, 
Second Period: 1909-1912, ?v,„ =- -534 ± -0.54, 
r = - -49.5 + -057. 
is - -153 and of the excesses + •134 ; this shows very fair accordance, 17 deviations being positive and 
IG negative. The greatest deviations occur in Horusey, Bath and Brighton, where residential neighbour- 
hoods show fewer children, and in Edmonton, Walthamstow and Khondda where there are probably 
Observed 
Calcu- 
lated 
A 
Observed 
Calcu- 
lated 
A 
Derby 
4-42 
4-25 
-•17 
Edmonton 
1-36 
1-80 
+ •44 
Southampton ... 
2-67 
2-70 
+ •03 
Bath 
1-18 
■87 
- 31 
Hornsey 
2-48 
1-86 
-•82 
Newport 
1-14 
1-25 
+ •11 
Bristol 
2-33 
2-23 
- •lO 
Rhondda 
MO 
1-34 
+ •24 
Reading 
2 •13 
1-95 
- -18 
Bury 
1-08 
•91 
- -17 
Nottingham . . . 
2-09 
1-99 
- -10 
Rothcrliaui 
1-06 
1-18 
+ ■12 
Salford 
2-04 
2-16 
+ •12 
Dewsbiiry 
1^01 
1-01 
- ■oo 
Ilford 
1-98 
2-09 
+ •11 
Blackburn 
•99 
•91 
- ■OS 
Brighton 
1-94 
1-68 
- •26 
Manchester 
•94 
•97 
+ ■03 
Stockton 
1-90 
2-12 
-+-•22 
Oxford 
•89 
■79 
- •lO 
Ipswich 
1-87 
1-89 
+ ■02 
Bolton 
•86 
■85 
-■01 
Grimsby 
1-85 
1-90 
+ •OS 
Rochdale 
•82 
•75 
-•07 
Walthamstow . . . 
1-85 
2-10 
+ •25 
Northampton ... 
•78 
•76 
-■02 
Coventry 
1-84 
1-93 
+ •09 
1 Barnslej' 
•75 
•88 
+ •13 
Plymouth 
1-61 
1-56 
-•05 
Wigan 
■58 
•64 
+ ■oe 
Wakefield 
1-40 
1-39 
- -01 
W. Bromwich ... 
•45 
■53 
+ -08 
Smethwick 
1-38 
1-55 
+ •17 
excess of children. On the whole the general order is very well maintained, and the general attack-rate 
closely fixes the juvenile attack-rate. In any further collection of material, it would of course be well 
to have the age-distribution of cases. 
* We endeavoured to see whether the correlation of isolation- and attack-rates would be modified if 
we took the attack-rate on children under 15 years. This made little difference, ?• being raised only 
from + •290± •069 to + -31.5 ± -068. 
