310 N. S. FERRY AND H. C. KLIX 



vious experiments already described, but as it was found that 

 all strains of the same organism gave similar reactions it was 

 deemed advisable to cut down the number to three of each in 

 order to save time. Of the B. broncMsepticus, no. 36 (dog), no. 

 123 (monkey) and human strains were tested; and of Bad. 

 pertussis, no. 0363 (Bordet), no. 109 and no. 248 (Povitzky). 



Technic. For the volume of the complement fixation tests 

 it was found more satisfactory to use 2 cc. than 5 cc. as advised 

 for the Wassermann test or 0.5 cc. suggested by Olmstead and 

 Povitzky in a serological comparison of the Bordet-Gengou 

 bacillus and hemoglobinophilic bacilli. The hemolytic system 

 was composed of sheep cells in a 2 per cent suspension, guinea-pig 

 complement in a 1 to 10 dilution and rabbit amboceptor in 1 to 

 1500 dilution. Complement titration was made by using 0.1 cc. 

 of amboceptor and varying amounts of complement. 



In determining the relationship of the various strains two units 

 of both amboceptor and complement were employed. All 

 titrations were incubated one hour before and one hour after the 

 addition of the sensitized cells, at 32°C. The dilution was 

 chosen in which complete hemolysis was produced, readings being 

 made at the end of the hour's incubation. All serum was inacti- 

 vated by heating in water bath one half hour at 56°C. 



The antigen was titrated by mixing it in varying amounts 

 with one unit of the hemolytic system. 



Preparation of antigen. After trying out several methods of 

 antigen preparation it was finally determined that filtered auto- 

 lysates gave the most stable and satisfactory products. 



The antigens were prepared as follows: The organisms were 

 grown on agar for forty-eight hours at 37.5°C., then taken off and 

 suspended in distilled water and shaken for forty-eight hours in 

 a mechanical shaker. This suspension was then heated at 56°C. 

 for one-half hour, incubated twelve hours, after which enough 

 sodium chloride and formalin was added to make an 0.85 per cent 

 and 0.5 per cent solution respectively. Filtration was carried 

 on through asbestos. 



Preparation of immune serum. The same serums were used 

 for this work as for the previous experiments, a description of 

 which has already been given. 



