M. Greenwood and J. D. C. White 
377 
We do not think it necessary to examine in detail the various conflicting 
statements which have appeared, because we hope to show that: — (1) the problem 
is essentially a statistical one : (2) it is statistically complex and cannot be solved 
without such a preliminary analysis of the data as has, up to the present, not been 
attempted. 
In order to establish these propositions to the satisfaction of a biometric reader, 
it will be necessary to describe in the fewest and simplest words the way in which 
an " opsonic index " is determined. 
In consequence of Wright and Douglas' work, it is believed that the power of 
ingesting and rendering harmless foreign organisms — with which certain of the 
white blood cells, the phagocytes, are endowed — depends upon the presence in the 
tissue liquids, especially the blood plasma, of a substance certainly complex and 
probably specific, termed an " opsonin " (feast-preparer). This opsonin renders the 
foreign organism apt to be ingested and destroyed by a phagocyte. The difference 
between the phagocytic powers of a normal and a diseased person is due far more 
to differing amounts of opsonin in their respective tissue liquids than to differences 
in the phagocytic cells. Thus it is said that, with certain relatively unimportant 
exceptions, if A be normal and B diseased, B's leucocytes mixed with A's serum 
are as strongly phagocytic as A's own leucocytes in A's serum, and conversely. 
If one wishes to determine the opsonic power of a man's blood with reference 
to, say, the tubercle bacillus, the following process is adopted*. 
A sample of the patient's blood serum is taken, further a sample of normal 
corpuscles which have been washed free of adherent plasma, and an emulsion of dead 
tubercle bacilli ; serum, corpuscles and bacillary emulsion being taken in equal parts. 
The three constituents are now mixed together and incubated at body temperature 
for a definite time. A second mixture is prepared and treated in the same way, 
except that normal serum (that generally of the operator, or a mixture of several 
presumably normal sera " pooled ") is substituted for that of the patient. 
Film preparations are then made from each mixture and so stained that the 
tubercle bacilli can be readily detected within or without the cells ; with the 
ordinary technique, the red colour of the bacilli is well shown up against the blue 
nuclei and faintly blue cytoplasm of the leucocytes. 
A variable number of leucocytes, in most laboratories neither less than 50 nor 
more than 100, is now counted on each of the two films and the number of bacilli 
per cell recorded. Let us suppose that the leucocytes mixed with the normal 
serum contained 100 bacteria in 50 cells, while those mixed with the patient's 
serum contained 50 bacteria in the same number of leucocytes, then the patient's 
opsonic index is said to be 50/100 or 0*5. 
* The whole of our work is based upon determinations made relatively to the tubercle bacillus, 
and the phrase " opsonic index," when used without qualification, is to be understood as meaning 
opsonic index for tubercle. We are not prepared to say how far our conclusions are valid for other 
opsonic determinations. 
Biometrika vi 48 
