PRODUCTION OF ANTIPNEUMOCOCCIC SERUM 571 



known serum, at least two complete tests must be made, using ten 

 mice for each serum dilution. Different laboratories have modified 

 the details of the test, particularly with regard to the doses of se- 

 rum and culture employed. It has been found that satisfactory re- 

 sults may be obtained by using dilutions of serum and culture 

 higher than those proposed in the American Drug Manufacturers' 

 Association method. The details of a similar modification which, in 

 the experience of the authors, has proved satisfactory may be 

 found in the Appendix, pages 651-653. 



Limitations of the mouse protection test. Although the mouse 

 protection test is generally accepted as the most satisfactory 

 means of assaying antipneumococcic serum for therapeutic use, it 

 has been subjected to criticism on the basis of its inaccuracy and 

 the irregular results not infrequently obtained. Aside from the 

 personal equation in the technical manipulations of the test, the 

 most important variables are those connected with the cultures 

 and the mice used and the ratio between the amounts of serum and 

 culture injected. 



Variations related to the culture may be referable to fluctua- 

 tions in the virulence of the organisms, their ability to neutralize 

 or combine with antibody, and the number of pneumococci present 

 in the test inoculum. Following a study of the standardization of 

 antipneumococcic serum, Enlows (1923) 366 stipulated that the cul- 

 ture should be used in its active or early stationary phase of 

 growth. Felton (1925) 397 observed that the unit value of serum 

 might vary depending upon the strain of Pneumococcus employed 

 in the test, even though only slight differences in virulence of the 

 cultures could be demonstrated. Variable results occurred also 

 when the same strain of Pneumococcus was used under standard 

 conditions. According to Felton, 402 the change from larger to 

 smaller doses of organisms gave more satisfactory results in pro- 

 tection tests. 



In studies dealing with the lethal power of the test culture in 

 potency tests, Morgan and Petrie (1931, 1933) 917 ' 1085 concluded 

 that some control over this factor could be obtained by choosing 



