ESTIMATION OF BACTERICIDAL SUBSTANCES 73 



tion (which see). In such a case a much larger quantity of bacteria 

 may be injected (1 oese) with impunity, and it will be observed 

 that notwithstanding the large dose no bactera will be found in the 

 peritoneal cavity at the expiration of one hour (Pfeiffer's experiment). 

 The same result will be obtained if instead of using a normal guinea- 

 pig and injecting some immune serum together with the bacteria, 

 these are introduced by themselves into a previously immunized 

 animal. 



The reason why bacteriolysis will be so much more extensive 

 in the presence of immune serum is the fact that as a result of infec- 

 tion (vaccination, immunization) the amboceptor content of the 

 blood-serum is materially increased. 



The complement which is necessary for the experiment is normally 

 present in the living animal. In its absence, of course, bacteriolysis 

 could not take place, and as complement readily becomes inactive 

 outside of the body, after standing even for a relatively short time 

 at body or room temperature, it is essential if the experiment is 

 conducted in vitro that only perfectly fresh serum be used. Other- 

 wise bacteriolysis will not occur, even though the serum be rich in 

 natural amboceptors. 



Quantitative Estimation of Bactericidal Substances. The quanti- 

 tative estimation of the content in bactericidal substances of a given 

 serum is most conveniently carried out by starting with a suspen- 

 sion, of known number, of the organism to be examined, and inocu- 

 lating tubes containing known amounts of serum, after which these 

 are incubated for a certain period of time and plates are prepared 

 in which the number of surviving organisms is finally determined 

 by a direct count. As serum is in itself an admirable culture medium 

 for most organisms it is, of course, essential to reduce this factor 

 as much as possible in the experiment. To this end one can either 

 determine the total number of bacteria which is completely killed 

 by a given amount of serum in a given length of time, or one can 

 determine the extreme degree of dilution in which a given serum will 

 still exercise a bactericidal effect, or one may determine the maximal 

 bactericidal effect, which is observed after different intervals of 

 time. The general arrangement of such a test is apparent from 

 the following example, which is taken from Wright, and which 

 represents the titration of a given serum against cholera vibrios 

 on the one hand and typhoid bacilli on the other: 



