PRODUCTION OF ANTIPNEUMOCOCCIC SERUM 585 



question is a study of the relation between mouse protection and 

 the amount of polysaccharide antibody as determined by some 

 form of precipitation test. Brown (1933) 154 found a high degree of 

 correspondence between mouse protection and precipitation with 

 the cellular carbohydrate, but it is difficult to concede any advan- 

 tage in its use over the purified forms of capsular polysaccharide. 



In a study of thirty-nine serums, Felton 407 observed correlation 

 coefficients of 0.93 and 0.91 between protection and precipitin 

 titer, and between protection and the amount of immune protein 

 precipitated with polysaccharide. Barnes, Clarke, and Wight, us- 

 ing fifty samples of unconcentrated serum, calculated the correla- 

 tion coefficients between mouse protection and their modified rou- 

 tine precipitin test, Felton and Stahl's combining equivalent 

 test, and Smith's optimal proportions test. Correlation coefficients 

 found for these three methods were 0.930 ± 0.012, 0.925 ± 0.014, 

 and 0.909 ± 0.016, respectively, in the case of Type I prepara- 

 tions. Between mouse protection and the first two precipitin tests, 

 values of 0.817 ± 0.031 and 0.739 ± 0.043 were obtained in the 

 case of Type II serums. Felton and Stahl 432 reported correlation 

 coefficients of 0.90 and 0.89 for Type I and Type II, respectively, 

 between the combining equivalent test and mouse protection for 

 both native serum and concentrates. Results of titrations by meth- 

 ods depending upon precipitation of antibody by SSS therefore 

 agree closely with those of the mouse protection test. It is not un- 

 likely that the correlation would be even higher were it not for the 

 inherent errors of the mouse protection test. 



The data on methods of standardization may be summarized 

 thus: The mouse protection test is inconvenient, expensive, and 

 subject to variations difficult to control. It appears likely that 

 some in vitro method of titration may ultimately supplant the ani- 

 mal test, and of those discussed, some form of precipitin test with 

 purified capsular polysaccharides offers the greatest promise. The 

 adoption of any new method depends largely upon its simplicity, 

 accuracy, and close correlation between estimated potency of the 



