198 



but it is difficult to make the experimental conditions so 

 uniform that they can be more accurately defined. 



Other common characters are as follows: poorly de- 

 veloped power of fermentation of all the carbo- 

 hydrates and alcohols investigated ; complete absence of 

 dark coloration of medium containing a r but in; absence of 

 the ability to liquefy gelatine, and serum coagulated by 

 heat. ' 



Agglutination and complement fixation exhibit 

 great complexity, but, as far as can be seen, there are innu- 

 merable continuous gradations, so that there is no basis for a 

 division into serological sub-groups. 



In agreement with other authors I have found that Pfeif- 

 fer's bacillus may be fairly toxic (for rabbits), but my ex- 

 perience of the pathogenicity of this organism is too small and 

 scattered for me 1 to discuss it further. 



While in all the tests, the results of which have just been 

 shortly recapitulated, there is a series of smooth gradations 

 in the group, the i n d o 1 test forms a marked exception, as 

 the result of this test divides the strains into two sharply 

 defined groups: indol-producers and non-indolproducers. 



The question now arises whether this is the expression 

 of a fundamental difference between strains of Pfeiffer's ba- 

 cillus, and one which therefore must make itself felt in the 

 other characters of the strains, or is it to be regarded as an 

 isolated phenomenon without obvious connection with other 

 facts? 



The reactions most used in the classification of bacteria 

 in the smaller groups are the serological ones, particularly 

 agglutination. We will therefore first investigate whether any 

 relation can be shown to exist between the indol reaction 

 and agglutination. 



Even a casual glance at the large agglutination table will 

 show that indol-positive and indol-negative strains are not 

 sharply distinguished from one another by agglutination, strains 

 belonging to the one group often being agglutinated by sera 

 corresponding to the other. To further investigate the pro- 

 blem the number of group agglutinations of different grades 

 among indol-positive and indol-negative strains for each serum 

 are given in the table below ,,0-f", „1+", etc. being denoted 



