Alice Lee 
543 
in the phthisis deatlirate has been substantially different for the different age 
groups, and is especially marked in the case of children, seems to indicate that 
recovery, at least from puerile phthisis, is more frequent now than formerly. 
However, not to spend more time on these assumptions — which, it appears to us 
that Dr Newsholme has by no means justified — let us examine whither this fourth 
method of approximately measuring segregation leads ns. Table V gives the 
necessary coefficients. 
TABLE V. 
Correlation and Difference Correlations of lO'^/P and 100pij<j) or 
Variate 
with 
England with Wales 
Scotland 
Ireland 
100pJ<p 
Crude 
Ai 
A2 
A3 
A4 
As 
Ao 
A7 
-•760+ ^046 
- -868 + •OSS 
- -879+ -0.35 
- -895 + -034 
- -895 + ^037 
- ^898+ •OSS 
- -907 + -037 
-•917+035 
+ •976+ •OOS 
+ ^848 + -OSS 
+ ^875 + ^037 
+ -874±^041 
+ •860+ -048 
+ ^847 ± ^056 
+ •850+ ^058 
+ ^835 + -067 
- -861 + ^028 
- ^755+ ^058 
- -824+ ^050 
-•809+^059 
-•811+ -064 
-•786+ ^076 
- -788+ ^079 
- ^792+^082 
+ •944+^012 
+ •772+ •055 
+ ^834 + ^047 
+ ^824 ± -055 
+ ^805 ± ^065 
+ ^788+ ^075 
+ •794+ -077 
+ •791 + ^082 
-•712+ ^054 
- ^819+^045 
- -922 ± ^023 
- ^954+ ^015 
- •964+^013 
- ^970+ -012 
-•973+ •on 
+ •eee ± -oei 
+ ^707 ± ^068 
+ •755 + -067 
+ ^791 + ^064 
+ -805 + ^065 
+ ^831 + ^061 
+ ^848 + ^059 
+ -857 + ^056* 
Now this table at any rate demonstrates a very high correlation between (p/P 
and pi/^, while the previous table for Dr Newsholme's third approximate segre- 
gation ratio led in the case of England with Wales to the value — •5S7, and in 
the case of Scotland and Ireland to negligibly small values ! Dr Newsholme 
himself writes : " Any of these indirect forms of segregation ratio has therefore 
to be verified wherever possible by the application to the same community and 
period of one or more other forms of the ratio, and checked where practicable 
by a special examination of sample constituent communities whose figures are 
included in the total. This has been done so far as the information obtainable 
allowed. It will be seen that the results obtained by applying different ratios to 
the experience of the same country and period are usually, though not invariably, 
in good agreement " (p. 268). 
What is quite clear from the above results is that, while in the case of 
Dr Newsholme's two chief measures of segregation, there is very sensible difference 
in the case of England with Wales, there is an absolute discordance in the cases of 
both Scotland and Ireland. Accordingly on the basis of his own axiom, that we 
must check our results by application of one or more other forms of the ratio, 
* This correlation continues to rise until it reaches ^929 with the thirteenth difference, but with such 
high differences the "population " is so reduced that the method ceases really to be reliable. 
