26 L. A. JULIANELLE 



Theoretically we would expect that the amount of leukocidin pro- 

 duced would bear a relation to the virulence of a strain, for the latter 

 would depend to some extent on the former. Since virulence and 

 hemolysis were found to be individual characters, it was hardly supposed 

 that hemolysis would show any dependence on leukocidin production. 



III. SEROLOGIC REACTIONS 



As a final analysis, recourse was taken to differentiate the hemolytic 

 staphylococci on a serologic basis. The impression is that although 

 biochemical reactions may vary, serologic reactions if once positive will 

 always remain positive. So, for example, the agglutinability of an 

 organism may fluctuate quantitatively, but not qualitatively. For no 

 other reason, then, this part of the work seemed to have the greatest 

 promise. Both deviation of complement and agglutination tests were 

 made, and the agglutination tests were supplemented by absorption tests. 



In preparing immune serums, strains Al, A5, PI, T9 and LI were 

 employed. Salt suspensions were made from agar slants and rabbits 

 were injected intravenously in 3 day periods, with 2 days between each 

 period. Five-tenths c c of the suspensions was injected the first period, 

 and this was increased 0.5 c c each period until a serum of sufficiently 

 high titer was obtained. 



COMPLEMENT FIXATION 



The literature on the complement fixation of staphylococci is scant. 

 The one reference available was that of Kolmer, Trist and Yagle 53 in 

 relation to influenza. Using a Staphylococcus aureus antigen, they were 

 unable to get fixation with either normal or influenza serum. 



The antigens used in these experiments were suspensions of 24 

 cultures to which were added 0.1% formaldehyd. The preparation of 

 the serum has already been described. 



After going through the preliminaries of obtaining antigenic and 

 complementary doses, the tests were made by incubating at 37 C. It 

 was found that all 5 serums gave fixation with all of the antigens. 

 There appears to be no qualitative differentiation of the different 

 strains. 



One more step was taken, and that was to determine whether there 

 might be quantitative separation into groups by complement fixation. 

 Four strains were picked at random, and the serum used in dilutions 

 of 1 : 50, 1 : 100, 1 : 150. The results did not warrant extending the 



Jour. Infect. Dis., 1919, 24. p. 583. 



