530 On the Distribution of the Means of Samples 
8. Even in the case of samples of 100, it is doubtful whether single determi- 
nations which give values even beyond the limits I'S to "7 can be regarded as 
satisfactory evidence of differentiation. 
9. The two previous conclusions are based on the analysis of four experimental 
frequencies. 
10. The Tubercle Bacillus is not necessarily an unfavourable test of the 
opsonic method. 
11. The opsonic index is less useful as a means of controlling treatment than 
for diagnostic purposes, since in the latter case it may be possible to make several 
comparable determinations. 
12. Differentiation in respect of phagocytic power (Tubercle Bacilli) can 
possibly be established by the opsonic index in some cases. 
