G. D. Maynard 
378 
evidence as to the value of vaccination even when small-pox has been contracted, 
yet the two processes are not identical, and it is I think safe to say that there is 
no a priori reason for assuming that such a state of affairs might not be. 
In addition to the figures already dealt with, there is in Major Buist's paper 
another table apparently compiled to show that the protection conferred by the 
process does not attain its maximum efficiency at once. Thus we find it recorded 
that out of 112 men attacked after inoculation 57 contracted the disease within 
one year, 17 within two, 10 within three, 7 within four, 4 within five, 4 within six, 
and 2 after the lapse of seven years. As the majority of the men inoculated left 
for foreign stations shortly after the event, I sought for a comparison amongst the 
uninoculated in foreign stations to see if there were sufficient grounds for this 
suggestion. The only figures that I could find that were at all comparable were 
given in Major Roberts' Enteric Fever, obtained from the Army Medical Records, 
for the troops in India for the years 1900-3. Those given by Major Buist covered 
a period from 1900-7. The largest proportion of the inoculations took place in 
the early part of this period. It is true that many of these cases were contracted 
in South Africa, and that the Indian statistics include some inoculated men, but 
I do not think that either of these sources of error robs the comparison of all value ; 
for the relative number of the inoculated to the uninoculated is small, and both in 
South Africa and in India the disease is endemic in a way that it is not in the 
British Isles, and, further, towards the end of the Boer War enteric fever was 
probably as rife in South Africa as it ever is in India. 
For this comparison to be entirely satisfactory it would be necessary that all 
the men inoculated should have completed seven years. As this was not the case, 
I attempted to adjust the figures so as to compensate for this deficiency. Un- 
fortunately each year did not supply an equal number of men, the largest 
proportion occurring in the first two years of the period, i.e. 1900 and 1901. It 
was also found impossible to alter the last two groups except by inspection of the 
plotted figures, as the rmmber of cases occuriing in them was so small that the 
error of one case would make too great a proportional difference. I think the 
result is an over-correction, and the truth will lie between the crude figures and 
the corrected ones. It will be seen that neither curve is different in type from 
that given by the Indian data, and that all the constants only vary by an amount 
that might certainly be due to random sampling in a homogeneous population. 
The type is that of Professor Pearson's Type I, and the following are the equations 
to the curves. 
Indian , = 2:?-62 - 1) - I 
^ X - -6521 , X -3307 
Buist's 1/ = 8-726 — 11 I 1 
V15-669 ) \ 8-09 
, X--4246 / ™ N-7793 
