216 BAENETT COHEN 



are properly selected, the deflections may be reduced considerably. 

 As a matter of experimental fact, they have never yet been com- 

 pletely eliminated (except possibly for radioactive substances). 



These considerations are not held to affect the validity of the 

 monomolecular law. They lead us rather to investigate the 

 disturbing influences which, when discovered and eliminated, 

 furnish further support to the law. 



We now possess a fundamental concept of the origin of the 

 monomolecular law as related to chemical reactions and may 

 turn to the phenomena observed in disinfection. If the disin- 

 fection process be followed by noting the numbers of surviving 

 bacteria at successive intervals we find that the rate at which 

 disinfection proceeds is iii general proportional to the number 

 present. This proportionality is not absolutely true but gener- 

 ally so, and the graph representing the course of the process is 

 more or less a straight Une with deflections at both ends. Under 

 appropriate conditions these deflections may be eUminated so 

 that the curve becomes very nearly a straight Une. The mortal- 

 ity process as observed by Chick and others, when strong disin- 

 fecting agents are used, has been shown conclusively to follow 

 the logarithmic rate, or a sUght modification of it. 



It is needless to stress the self-evident analogy between the 

 course followed by a monomolecular chemical reaction and the 

 course of the mortality process of bacteria. They represent 

 actual observations. 



In this connection must be mentioned the statistical deduction 

 of Yule (1910) based upon the theory of probability, that in a 

 population exposed to a single lethal influence, the rate of death 

 will follow a logarithmic course, and when there are a number of 

 sub-lethal causes, the death rate wiU be a modification of the 

 logarithmic one. 



The above facts have been fully explained by Chick and it 

 would seem superfluous to repeat them were it not for the criti- 

 cisms of Chick's conclusions by Loeb and Northrop (1917), 

 Brooks (1918) and Smith (1921). These authors urge that 

 individual variations among bacteria may be distributed on a 

 frequency curve such that the mortality rate will proceed appar- 

 ently logarithmically. 



