15Y R. GREIG SMITH. 81 



to eliminate from the fluid traces of precipitable salts. Generally 

 a slight precipitate was obtained, and above this a uniform sus 

 pension of the bacteria. The nitrates of potash, soda and 

 ammonia failed to produce a flocculation when added to the 

 suspension, even when the emulsions were centrifuged (2,500 

 revolutions per minute). The continued addition of dilute silver 

 nitrate in place of the alkali nitrates produced very slight 

 precipitates. Strong silver nitrate, however, produced complete 

 precipitation. 



The failure of the dilute silver nitrate to effect complete 

 flocculation shows that either no silver salt had formed on the 

 surfaces of the bacteria, or, if one had formed, the silver nitrate 

 or the alkali nitrates were too weak to induce flocculation. It is 

 probable that no surface film had been formed. The absence of 

 flocculation by so strong a flocculating agent as dilute silver 

 nitrate emphasises the fact that bacteria when freed from bouillon 

 salts and their by-products, are not coagulated like inorganic 

 particles. 



The bacteria after being flocculated by the strong silver nitrate 

 were seen to be in clumps. Careful examination of these clumps, 

 and especially after they had been exposed to the light, showed 

 that the bacteria were enclosed in a matrix which undoubtedly 

 consisted of a silver compound of the intracellular salts which had 

 diffused out from the cells under the influence of the strong 

 silver nitrate. It is evidently impossible to obtain a pure floccu- 

 lation or agglutination of bacteria, and when such an appearance 

 is presented the failure to reveal the presence of a flocculated 

 matrix is due entirely to our instruments or methods of demon- 

 stration. 



Some experiments of Malvoz (io)are frequently quoted to show 

 that agglutination of typhoid bacteria may be obtained by the 

 addition of certain chemical reagents and stains. Since I have 

 shown that true chemical agglutination does not occur, it seemed 

 advisable to repeat his experiments. In one of these experiments 

 clumping occurs when strong alcohol or strong formalin is added 



