540 Tuberculosis and Segregation 
annulment of the time-factor (see our p. 538), has no bearing on the segregation 
problem of Dr Newsholme. 
TABLE III. 
Total and Partial Correlation Coefficients of Crude Nmnhers of Indoor and 
Total Paupers (pi and pr), Total Population (P), and Phthisis Deaths {(f)). 
Coefficients 
England with Wales 
Scotland 
Ireland 
-•934 ±-014 
-•718f053 
+ •4571-087 
Total 
Coefficients 
r 
r 
+ •955 + -010 
-•950 ±-011 
- -544 ±-077 
+ -831 ± ^034 
-•896±^022 
- -528 ± ^079 
+ -763 ± -046 
+ ^479±^084 
- 251 ±-103 
+ •577 ±-073 
+ •780 ±-043 
+ •070 ±109 
r 
PjP 
- -674 ± -060 
- -805 ± •OSS 
-•684f058 
Partial 
Coefficients 
-•287 ±-100 
- ^905 ± -020 
+ ^lllfl08 
-•575f073 
+ -162fl07 
+ •4921 •083 
PPt'p& 
-•189±^106 
-•017fl09 
- ^305 ± ^099 
To approach nearer to the meaning of the relation between total pauperism 
and phthisis we determined the correlation between j)j and 0 for constant P, and 
found 
pr,,^^= - 277 ±-101, 
which is barely significant having regard to its probable error. 
Now after elimination of the time-factor, we found for the correlation of ^/P 
and Ii at the eighth difference — •G14 + •143, but this is the same as the corre- 
lation of — ^"^^ 4'/^- Hence the correlation of Pr/P and cj)jP must be very 
significant, positive and of the order •G. Now if p^. and (p after the removal of the 
timefactor were practically independent of each other, there would be a high 
positive correlation between pr/P and ^/P, due to the fact that P when it 
takes — after annulment of the time-factor — any random deviation appears in 
both variates' denominators. In other words, we are inclined to believe that 
the high negative correlation between (j)/P and is solely due to spurious 
correlation arising from the nature of the indices used. 
To throw still more light on the matter we have investigated the correlation 
between the total number of paupers and the total number of deaths from phthisis 
when the time-factor approaches annulment. It will be seen from the table 
below that for both Scotland and Ireland there is finally no relationship at all 
