326 
Smallpox and Vaccination 
five, after which it is moderately constant. One point, however, of interest 
becomes manifest here, and that is the way in which the correlation coefficients 
for the different ages between vaccinated and unvaccinated tend to assume higher 
values, at the ages of 10 to 20 years*, as the century has gone on. Taking the 
first set of statistics which exist, namely those of Dr Marson, passing from these 
to the great epidemic of 1870-72 in London up to the last epidemic of 1902, we 
see the tendency for those coefficients to become larger continuously manifest, 
and the corresponding figures for the other epidemics for which they have 
been calculated seem to fit into place much as might be e-xpected from their 
date of occurrence. This affords some confirmation of what has been suggested 
earlier in this paper in connection with the rise in age at which vaccinated 
persons are attacked with smallpox, namely, that vaccinal lymph has gradually 
gained potency through being passed for many generations through the human 
subject f. 
The question now arises as to whether in different epidemics the extra protec- 
tion which good scars are known to afford to their possessors compared with bad 
scars assumes greater relative importance in mild or severe epidemics. There 
seems, however, to be little difference. In the accompanying table will be seen 
TABLE XL 
Congelation between Recoveries and Deaths in the Epidemics 1902 and 
1892—95, and 1902 and 1870—83 for each Vaccinal Group. 
Nature of Group 1902 & 1892-95 1902 & 1870-83 
One Scar -22 -43 
Two Scars -18 -37 
Three Scars -24 -43 
Four Scars -21 "52 
the correlation between cases and deaths for the epidemic in 1902 and those 
in 1892-95 and 1870-83. These coefficients indicate that there is no great differ- 
ence in a severe epidemic or in a mild between the manner in which each different 
group is affected in comparison with the same group in a severe or in a mild 
epidemic. The figures, it is true, show some little difference, but not much when 
it is considered that the change in the age distribution of the different groups 
has been proceeding not by any means equally in them all. 
Another point which may be noted is the manner in which the relative 
protection against death between those with good and those with bad marks 
varies. It is found to differ not very considerably below the age of twenty, to 
* The whole correlation coefficients are given in the table, but only the early years are referred to 
in the text, as adult primary vaccination was very common in the first half of the century. 
t It is this change in the correlation coefficients which serves to differentiate the two interpretations 
which one puts on the rise of the mean age : (1) that due to the lymph gaining potency which is the 
cause of the rise of the general mean age of vaccinated persons attacked by smallpox, here the corre- 
lation between vaccinated and unvaccinated has increased as the century has gone on : and (2) the 
higher mean age of indifferently vaccinated persons taking smallpox than of well vaccinated, here 
the correlation between recoveries and deaths is in favour of the well vaccinated. It is thus seen that 
the mean age may be raised in two different ways. (Cf. footnote, p. 319.) 
