A BIOMETRIC STUDY OF PHAGOCYTOSIS WITH 
SPECIAL REFERENCE TO THE "OPSONIC INDEX." 
SECOND MEMOIR. ON THE DISTRIBUTION OF THE 
MEANS OF SAMPLES. 
By M. greenwood, June., and J. D. C. WHITE. 
{From the Statistical Laboratories of the Lister Institute of 
Preventive Medicine and the London Hospital.) 
In a recent memoir* we communicated the results of an analysis undertaken 
with the object of throwing light on difficulties associated with the ordinary 
method of estimating opsonic indices. The main results of that investigation can 
be summarised in the following way : 
(1) Phagocytic counts have pronouncedly skew frequency distributions and 
are good examples of Pearson's skew curves. 
(2) The means of small samples have likewise a skew distribution. 
(3) The use of a thicker bacillary emulsion while diminishing the skewness, 
does not, under possible experimental conditions, do so sufficiently to allow of 
testing on the basis of a Gaussian curve. 
With respect to (2), we pointed out that results based, as were ours, upon 
samples extending in no case to more than 2000 cells, were provisional and that 
we intended to re-investigate this part of the subject with the help of more 
adequate material. 
The present memoir comprises a study of the distribution of a large sample 
of phagocytic cells, the distributions of the means of sub-samples and some 
preliminary contributions to the problem of mean distributions in general. 
The basis of this work is a count of 20,000 cells made by one of us (J. D. C. W.) 
who had had 18 months' previous experience in the Inoculation Department of 
the London Hospital. 
* Biometrika, Vol. vi. p. 377. 
