APPLICATION OF THE PHENOMENON 399 



observations "which indicated tlie complexity of the problem. 

 Having established a reliable killing test of living meningococcus 

 for mice, he found on several occasions that the immune serum 

 not only failed to protect the animals but quite commonly ac- 

 celerated or increased the death rate, as comj^ared with control 

 mice which received normal horse serum and others receiving- 

 no serum at all. At other times he found that the sample bleeding- 

 yielded a serum with quite marked protective power biu at the 

 conclusion of this test, a further bleeding yielded serum with no 

 apparent protective power. Gordon, and also Griffith, had similar 

 experiences. It seemed to these authors that the period elapsing 

 bet-\veen immunization and bleeding ^vas of considerable impor- 

 tance. Amoss and Wollstein (1916) obtained active sera in eight 

 to twelve weeks. Hitchens and Robinson (1916) claimed that a 

 suitable serum is only produced by prolonged immunization. By 

 their test, sera resulting from t-\velve to fifteen months' treatment 

 of horses were most active. Murray (1929) states that for the 

 preparation of protective sera of therapeutic \alue in horses there 

 is no agreement about the method to be employed. "All ^vho have 

 tried it have experienced very inconsistent results, and it is still a 

 matter for experiment. There is no doubt that occasionally re- 

 markably poAverful sera are produced, but imtil a reliable method 

 of titrating their value in the laboratory is arri\ ed at there is little 

 likelihood of progress." 



In order to determine -whether there existed any serological 

 specificity of the active principles of meningococcus, a monovalent 

 goat serum prepared by the injection of filtrates and cultures of 

 strain 44D (Group I) ^vas tested against homologous and heterol- 

 ogous active principles. 



In these studies it ^vas found that the neutralizing potency of 

 this serum -vvas considerably lower against reacting factors of 

 Group III. 



Moreover, there appeared to exist considerable differences in 

 neutralizability of filtrates obtained from various strains of the 

 same serological groups. Thus, in experiments with Group I 

 meningococcus where a standard serum ^vas used the neutraliza- 

 tion titers of filtrates derived from various strains fluctuated be- 

 tween 55 and 150 units; and uith meningococcus Group III be- 

 tween 75 and 120 units. 



In Chapter iv, there were described the results of studies on 

 the effect of bacterial variation on the reacting factors of B. ly- 



