BY R. GREIG SMITH 79 



bouillon or serum. It is the jwecijntate that is clumped/ the 

 bacteria are carried with it mechanically. By adopting this view 

 we have an agreement between the numerous observers. It is 

 also easily understood why dead typhoid bacteria agglutinate 

 like living ones. The products of metabolism being in contact 

 with the dead cells are precipitated by the active serum upon 

 their surfaces and in the medium. The precipitate is flocculated 

 by the saline constituents of the solution, and both dead cells and 

 precipitate gather together into floccules. The fact that dead 

 typhoid bacteria may be employed in Widal's test bears out the 

 theory of a coagulable surface precipitate, and agrees with Nicolle's 

 experiments, which showed that foreign bacteria in a filtered 

 culture of Bact. typhi were clumped by active typhoid sera. 



Gruber (7), writing recently, considers that Kraus' precipitate 

 is quantitatively too small to explain agglutination, and thinks it 

 probable that in the act of agglutination certain substances in 

 the bacterial membranes are made more insoluble. A shrinkage 

 and separation follow whereby glutinous masses are formed on 

 the bacterial surfaces. This appears to be very similar to his old 

 hypothesis of the swelling of the membranes, to which Bordet 

 pointed out that there was no reason given for the approach of 

 the bacteria. Again, the shrinkage and formation of sticky 

 masses on the bacteria is an idea, while Kraus' precipitate is a 

 fact, 



Radzievsky (8), in a preliminary paper, objects to the precipi- 

 tate idea apparently because he obtained no precipitate in young 

 cultures in which the bacteria clumped normally. Both writers 

 apparently forget that the bacteria must be saturated with the pre- 

 cipitable substance before it is given off into the culture medium, 

 and in a young culture while the organisms are saturated there 

 may be but an infinitely small amount in the culture fluid. 



I have tested the validity of Gruber's and of Radzievsky's 

 objections and cannot agree with them. The precipitate may be 

 and undoubtedly is very small in amount, but it is still appreci- 

 able. A twenty-four hours' bouillon culture of Bact. tyj>hi was 

 filtered through a Kitasato filter, and an agar culture of Bact. 



