Karl Pearson 
215 
The analytical constants are : 
Fleming 
Greenwood and White 
Mean 
1-0056 
1-0089 
Mode 
•9744 
•9743 
Standard Deviation 
•1104 
•1291 
P3 
•000377 
•000570 
Pi 
•000361 
■000673 
01 
■0787 
•0700 
0, 
2 4338 
2-4200 
Equation to the curve : 
/ r \l-!;662 / r \ 2-5850 / r \ 1-5092 / r \ 2-4031 
^ 1TO6 ( 1+ «) l 1 -^) ■ ^ M ( 1+ 5Wo) O-MOgo) • 
Unit of x, -05 of opsonic index Unit of x, -05 of opsonic index 
and a total frequency of 380. and a total frequency of 380. 
In ascertaining the indices here from the 20 means of Table V (see p. 223), no 
mean has been used with itself, as the indices are supposed found from independent 
sets of 50. 
(7) I have now reached what I consider the most difficult point in the 
matter. I have shown : 
(i) that the frequency distribution of the bacilli per leucocyte is the same 
for Fleming and for Greenwood and White, if we take populations of 1000 or 
more. 
(ii) that the 20 independent means of Fleming and the 20 of Greenwood and 
White, both from single slides, show also the like degree of variation. 
(iii) that the possible opsonic indices deducible from these means for the two 
investigations do not differ more than is compatible with a difference of material, 
or (as shown by Greenwood and White's long series) by the difference of two 
slides from the same material. 
But notwithstanding these arguments, which would lead to actually the same 
ranges of opsonic indices, if we take mathematically at random 50 leucocytes 
from those counted by either of these investigators, there is a very curious and 
marked difference between their method of counting or recording the bacilli per 
leucocyte. But the remarkable point about the matter is that Strangeways is in 
complete practical agreement with White in the results of his counting: see my 
Diagram III. Now Strangeways and White counted, using a mechanical stage, 
every available leucocyte as they worked across the slide and recorded them in 
succession. It is difficult therefore, since Strangeways worked on a slide prepared 
in Wright's Laboratory, to believe that the difference arises from technique in 
the preparation of the slide. It must lie in the method of counting. 
The difference between the records of Fleming and White is of the following 
character : While the small numbers and large numbers of bacilli per leucocyte 
both occur with much the same frequency in the case of the two investigators 
