LOGARITHMIC ORDER OF DEATH 33 



after 6 seconds, it could be claimed that no cell died 

 in the first 3 seconds. Though such a claim may be 

 entirely unjustified, it is obviously impossible to disprove 

 it. 



But those who profess a constant death rate in bac- 

 teria can disprove the identity of the two orders of death 

 by pointing to the survivor curves. If the mortality 

 curves of bacteria and those of higher organisms are es- 

 sentially similar, the semilogarithmic graphs of the sur- 

 vivor curves of higher organisms should be straight lines 

 when the first two or three survivor counts are omitted. 

 A glance at Figure 2 will show that this is not the case. 



Far more helpful in one's decision as to the nature of 

 the curve is the death rate constant K, because each 

 value of K is significant in itself independently of the 

 previous or of the subsequent trend of the curve. It has 

 been shown that, with higher organisms, (or with clus- 

 tered bacteria, or also in the case of very slow death, as 

 will be explained below) K is not constant, but increases 

 with time; that, with uniform bacteria, it is constant; 

 and that, with bacteria of unequal resistance {e.g., with 

 very young cultures), it decreases consistently. To illus- 

 trate these differences of death rate, a few character- 

 istic data (where K is multiplied by 1,000 for easier com- 

 parison) have been gathered in Table 7. They represent 

 many of the organisms commonly used in experiments 

 on death. Each set gives the K-values for all successive 

 time units for which survivors were counted. 



Subsequent to the objections of Loeb and Northrop, 

 of Knaysi and of many others, Watkins and Winslow 

 (1932) made a very extensive study of the order of 

 death of Bacterium coll killed by heat and by dilute 

 NaOH. (For the complete 150 sets of data, see Watkins, 

 1932). They concluded: "The studies . . . seem to indi- 

 cate that when vegetative cells {Esch. coll) are exposed 

 to the toxic action of sodium hydroxide, reduction pro- 

 ceeds at a generally logarithmic rate and does not ex- 



