244 MANUAL OF BACTERIOLOGY. 



the organisms, and, if a powerful germicide, in sufficient 

 amount to prevent growth, and thus give erroneous results. 

 In methods Nos. 2 and 3 this factor may be partially obviated 

 by washing in sterile water after exposure to the germicide. 

 This does not remove another source of error, namely, the 

 chemical action that may take place between the substance and 

 the protoplasmic contents of the bacterial cell. This action 

 may extend deeply enough to restrain the growth of an organ- 

 ism for a very long time without actually killing it. When 

 placed under suitable conditions, such union may be broken 

 up and the organism regain its power to develop. It has been 

 suggested that, to remove errors in the above methods, the 

 bacteria after exposure to the germicide be inoculated into 

 susceptible animals; but Sternberg's experiments in this direc- 

 tion have shown that bacteria may become so attenuated in 

 virulence by the action of germicides insufficient to kill that 

 the value of animal inoculation experiments is limited. More- 

 over, it sometimes happens that it is desired to test germicides 

 on bacteria which are not pathogenic for animals. 



Geppert suggested a valuable modification of these methods 

 while determining the germicidal value of bichloride of mer- 

 cury. After exposing his test-organism to bichloride of mer- 

 cury, and before inoculating into bouillon to determine death 

 of the organism, he treated with a dilute ammonium sulphide 

 solution, thus effectually neutralizing any mercury-salt 

 remaining. 



Sedgwick developed this method still further, and to him 

 we are indebted for demonstrating its accuracy and practica- 

 bility. Sedgwick proceeds as follows: 



4. To 15 c.c. of sterile water in a 60 c.c. Erlenmeyer flask 

 add 2 c.c. of a virulent culture of the test-organism. Then add 

 a solution of the substance under investigation in the propor- 

 tion necessary to give the dilution wished. Mix thoroughly, 

 and allow this "action-flask" to stand as long as it is desired to 



