574 BIOLOGY OF PNEUMOCOCCUS 



Although this method served to compensate for irregularities in 

 the mortality-rate of the test animals, it has been found to be un- 

 reliable, since a different value may be obtained when the number 

 of survivals is substituted for D and D'. A more accurate formula 

 has been derived and is included in the Appendix, page 657. 



It has been claimed that satisfactory end points may be ob- 

 tained by using serum doses more closely spaced (Park, 1930 ; 1053 

 Morgan and Petrie, 1933 917 ), but in the experience of the authors 

 of the present volume this plan is not reliable under routine labora- 

 tory conditions unless an inconveniently large number of mice are 

 used. 



In a recent series of studies on the mouse protection test, Good- 

 ner and Horsfall 537 and Goodner and Miller (1935) 540 decided that 

 the most important variable in the mouse is the number of mono- 

 cytes present in the peritoneum at the time of injection of serum 

 and culture. The mice were carefully selected from a strain inbred 

 for several years. It seems not unlikely that the phagocytic activ- 

 ity of the animals used may condition the progress of the pneumo- 

 coccal infection in the protection test. On the other hand, it is 

 questionable how this factor may be controlled in selecting mice, 

 even from an inbred stock, at the time of carrying out protection 

 tests. Irwin and Hughes (1931 ) 670 found in studying the bacteri- 

 cidal power of the blood of rats that, when matings were made of 

 susceptible animals, that is, those with low bactericidal powers, the 

 offspring were all susceptible; if a low X high mating was made, 

 two-thirds of the offspring were susceptible and the remainder re- 

 sistant. The breeding of a stock of mice of uniform susceptibility 

 to pneumococcal infection, therefore, would appear by analogy to 

 be possible. Such a program has been initiated at the Rockefeller 

 Institute under the direction of Webster and by Little at Bar Har- 

 bor. If the efforts are successful, it may be possible to establish a 

 supply of mice of uniform susceptibility for titrating antipneumo- 

 coccic serum. 



The outcome of protection tests is also dependent upon the pro- 



