396 M. R. SMIRNOW 



that are susceptible to variations, but for some unexplained 

 reason cannot be regarded as altogether constant. Indol for- 

 mation would invariably return when these altered bacteria 

 were transplanted into plain broth at frequent intervals. Lab 

 enzyme would also return in most of the altered strains but not 

 invariably so. The same can be said of the fermentative proper- 

 ties. Very often, however, these characteristics appear to be 

 entirely done away with, the change being permanent as far 

 as could be made evident by sub-culturing into plain broth. 

 (See Table IX.) In these cases observations were made up to 

 two months after the last exposure to the influencing substance, 

 making frequent transfers. There seemed to be no definite 

 rule of reversion, and no relation between the reappearance of 

 one enzyme and another. The reappearance of the ferment- 

 ing enzymes in one sugar was not necessarily accompanied by 

 those in other sugars. At times the fermentation of one sugar 

 might have returned to nearly normal, while others might show 

 httle or no presence of gas with the same strain of B. coli. 



ADDENDA 



The work under this heading was undertaken as supplement- 

 ary to the foregoing section for the purpose of verification and 

 also in response to comments made upon the report of this 

 paper at the meeting of the Society of American Bacteriologists, 

 held at Philadelphia, in December, 1914. 



To avoid any objection arising as to the possible existence of 

 mixed cultures, or of "weak members" at the start, each strain 

 of B. coli was plated, a single colony selected, and replated, and 

 from this latter plate, several cultures were selected, inoculated 

 on agar slants and after twenty-four hours cultivation, carried 

 through on all media. The strain that showed the greatest 

 amount of enzyme formation was selected as the "strongest" 

 and the one to be subjected to experimentation. 



In general, the technique of these experiments was identical 

 with that already described, observations being made with 

 special reference to indol production which was taken as the 



