210 



appears firstly that there is a marked difference between 

 Bordet's bacillus and Pfeiffer's bacillus both as regards agglu- 

 tination and complement fixation, and secondly that the diffe- 

 rent strains of Bordet's bacillus usually resembled one another 

 fairly well. 



As in my investigations strains of Pfeiffer's bacillus were 

 found to be distinctly different from one another serologically, 

 it was taken for granted to begin with that at any rate most of 

 the strains in this group must also be serologically distinct 

 from whooping-cough bacilli, and this aspect of the case was 

 therefore only superficially investigated. 



A rabbit serum prepared by the injection of whooping-cough 

 bacilli agglutinated 4 different strains of the whooping-cough bacillus 

 at 1 : 10000, while it agglutinated 4 strains of Pfeiffer's bacillus (from 

 healthy persons) at most at 1 : 25 and incompletely at 1 : 50. At the 

 same time the 4 same strains of each species of bacterium were 

 tested as regards complement fixation with sera from 3 whooping- 

 cough patients, and 3 control sera from other persons. Each of 

 the 3 whooping-cough sera gave complement fixation with the 1 

 whooping-cough cultures, and all the strains reacted practically 

 equally with the same serum. But none of the strains of Pfeiffer's 

 bacillus gave a stronger reaction with the 3 whooping-cough sera 

 than with the 3 control sera. 



It was of greater interest to compare a number of 

 whooping-cough strains one with another, particularly by agglu- 

 tination, for if decidedly more uniform relations were found 

 than among Pfeiffer's bacillus, it would indicate that the great 

 differences observed among the latter could not be accounted 

 for by special peculiarities of the technique employed. 



Agglutination experiments were undertaken with the rabbit 

 serum referred to, against 75 strains of the whooping-cough ba- 

 cillus in dilutions of 1:2500, 3500, 5000, 7000, 10000, 14000, 20000 

 and 28000. Cultures grown on a mixture of equal parts of Bordet's 

 potato glycerin, agar (see Bordet & Gengou), ordinary peptone 

 broth agar and horse blood were employed. For the serum dilutions 

 and the suspensions of the cultures 0.3o/o salt solution with the 

 addition of l°/ 00 formalin was used. Each strain was further tested 

 with serum corresponding to Pfeiffer's bacillus I 6, and with pure 

 formalin salt solution. In these two control tubes there was weak 

 agglutination in 3 cases only and then merely in one of the two. 

 The experiment was arranged in 3 parts. Some sera which' on first 



