294 PHYSIOLOGY OF BACTERIA 



order would not appear at all if the first plate counts had 

 been made sooner. 



This belief is not based on facts. It has never been 

 proved that a logarithmic order of death would become 

 established e.g. in the experiments p. 275 with seeds, 

 fruit-flies or protozoa, if the first observations would 

 have been omitted. In fact, no logarithmic order would 

 be established if this were done, as any calculation or 

 graphic test will easily show (e.g., see Fig. 33) . The death 

 rate of the higher organisms increases, and the curve 

 continues to bulge, while with bacteria, the death rate 

 is always decreasing if not constant, except in the case 

 of some spore-formers and of staphylococci, which will 

 be discussed on p. 296. 



It is true, that in most experiments on the order of 

 death, or on the theory of disinfection, almost half 

 of the bacteria or more have died during the first time 

 interval investigated. But this very fact points out 

 that the order of death is essentially different from that 

 observed with higher organisms. 



Table 84 gives a survey of all experiments known to 

 the author referring to the order of death. They are 

 divided into three groups according as the death rate 

 is dominantly constant or decreasing or increasing. 

 Nearly one-half of all experiments cannot be used to 

 prove or disprove the logarithmic order of death because 

 the most important part of the experiment, the initial 

 number of bacteria, has been omitted. Of the remaining 

 one hundred and fifty-four experiments, only thirty-two 

 have a constant death rate. But we have just seen 

 (p. 290), that a decreasing death rate must be expected 

 to be the most common occurrence because the cultures 

 under test are rarely very uniform in resistance. It 

 should be remembered, too, that the decreasing death 



