542 
Tuberculosis and Segregation 
Thus we see that apart from the time-factor there is no relation whatever 
between either pauperism or phthisis and population. In the relation between 
total pauperism and phthisis deaths, no further correction for population is 
needful than that obtained by the annulment of the time-factor as in Table IV. 
Table IV his shows us that neither pauperism nor phthisis is organically related 
to population, although we might well have anticipated that greater density of 
population would influence pauperism and provide greater chances of infection, 
and so of deaths, in the case of phthisis. 
(5) We now come to Dr Newsholme's fourth and last measure of segregation. 
It is " the ratio in which the number of paupers treated in workhouses and work- 
house infirmaries stand to the total number of deaths in the community" (p. 276). 
In our notation this is pi/<^, or as an index 100pi/(f). But iu the figures actually 
given in Table LXV (p. 277), and headed Segregation Ratio, Dr Newsholme 
appears to be using 100<p/pi. The same remark applies to Tables LXVIII and 
LXIX (pp. 280—281). Thus it is difficult to be certain of what Dr Newsholme 
intends to be taken as his fourth measure of segregation. In our discussion below 
we have used both li)0^/pi and 100pi/(f> to provide for both contingencies and 
to check our results. 
Unfortunately Dr Newsholme makes little attempt to justify either his third 
or fourth ratio as an approximate measure of segregation. It will be remembered 
that he has defined the true method of measuring segregation to consist in 
forming the ratio "stating how many of the total days of sickness (number of 
patients and number of days of sickness) are passed in institutions" (p. 267). In 
this fourth index of segregation he replaces phthisical patients in institutions by 
indoor paupers, and total of phthisical patients by total deaths from phthisis, 
dropping any question of the number of days of sickness. At the very least this 
seems to involve two assumptions, (a) either that all indoor paupers are phthisical 
or that for the period in question the proportion of indoor paupers who are 
phthisical has remained constant, (h) that for the period in question the number 
of deaths from phthisis has remained a constant fraction of the total number of 
cases of phthisis. It is difficult to see how, without such assumptions, such figures 
can " measure with approximate accuracy the ratio which states how many of the 
total days of tuberculous sickness are passed in institutions " (p. 267). 
Yet in another paragraph Dr Newsholme quotes with apparent approval the 
statement of Mr Fleming, who speaks of the "great change in the character of 
workhouse inmates during recent years.... The able-bodied inmates are gone and 
the sick inmates have come" (p. 273). Such a statement is absolutely inconsistent 
with the assumption (a) above. 
To justify (ft) we must assert that for the last fifty years of the nineteenth 
century there has been no change in efficiency of treatment in the case of tuber- 
culosis, for without this we cannot assume that deaths from phthisis are even an 
approximate measure of the number of cases (p. 267). The fact that the reduction 
