88 
The Opsonic Index — A Medico-Statistical Enquiry 
When these curves are tested for goodness of fit to the observations, it is found 
that this is much better for the case in which couplets and triplets are counted as 
2's and 3's than for the case where they are counted as single ingestions, P = '51* 
as against P = 'Oo. It is quite conceivable that a still better fit would be obtained 
if we were to regard a couplet for purposes of estimation, say as 1"5 or some other 
intermediate value between 1 and 2. If such were the case we might conclude 
that, provided the criterion is sound, this would be the proper value to assign to 
a couplet inclusion. 
-861 O 123456789 10 
Number of bacteria ingested. 
Fig. 5. Couplets and triplets counted as single ingestions. 
Other points than those illustrated or mentioned will occur to every worker in 
these fields. Further, those that are illustrated are single trials, which require 
repetition both with regard to the particular organism used and to other organisms. 
They are given here for the purpose of focussing attention on the points which 
must be considered in discussing the question of the reliability or comparability 
of the opsonic index as determined by different observers. When, however, such 
points as the possibility of there being aiiy specific opsonic variation demonstrable 
in tuberculous sera and still more in the case of other sera are disputed, it seems 
to me that the proof that they are is so overwhelming as to be incontrovertible. 
All honour then to those who have elaborated so important a method as that here 
considered ; for they have added to the list of quantitative estimations applicable 
to medical diagnosis and therapeutics one of incalculable value. Its chief draw- 
back is the difficulty of its determination. 
It remains now only to consider some of the variations which have been 
proposed upon the original and very natural measure pi'oposed by Wright and 
Douglas. Much more criticism seems to have been levelled at the method of 
measuring the effect of vacci no-therapy than upon the efficacy of vaccino-therapy 
itself, which seems somewhat of a reversal of their relative importance. One would 
* P = l represents perfect fit. 
