STATISTICAL STUDY OF ANTI-TYPHOID 
INOCULATION. 
By C4. D. MAYNAED, F.R.O.S.E., Pretoria. 
In a paper read before the Soutli African Medical Congress of 1908, Major 
Buist, R.A.M.C., has collected a considerable number of figures relating to anti- 
typhoid inoculation. The conclusion that he considers these figures lead to is 
that inoculation is of decided benefit, both in preventing attack and in lowering 
the case mortality. It seemed that in view of the sparseness of the statistics at 
present available, it was more than ever desirable to examine such as we have by 
modern statistical methods. To base deductions on percentages, apart from their 
probable errors, is likely to lead to wrong conclusions ; and it is only by the help 
of methods such as those here employed that we can hope to learn all that the 
data signif3^ 
The statistics used in this paper have been obtained from the above-mentioned 
source, with the exception of the table containing the Transvaal figures, which 
were kindly supplied to me by Major Buist at a later date. 
The coefficients of correlation — between inoculated and freedom from attack, 
and between attacked and recovery (in the inoculated) — have been calculated by 
the " four-fold table " method of Professor K. Pearson. The probable error of 
these values has been assumed to be three times the error found by the formula 
^^, = -67449 (1 — ?'-)/V()i — 1). In two cases the long method was employed to 
check these results, but in neither case was the difference material. 
The figures contained in Tables I and II were obtained from a census of the 
whole army, taken by the military authorities. The data are admittedly in- 
complete owing to the absence of some of the troops on manoeuvres or on board 
ship, and further it is not certain that all cases of inoculation were correctly 
returned as such. The period covered is from March 1, 1906, to February 28, 1907. 
Although Major Buist quoted data for the inoculated in the British Isles, these 
