LOGARITHMIC ORDER OF DEATH 35 



were approximately constant throughout the experi- 

 ment. We have estimated the ratio of K for the two 

 time-halves of an experiment and find this ratio to aver- 

 age 0.97 for 109 experiments with NaOH, 1.17 for 16 

 experiments with a borate-buffer mixture and 0.96 for 

 26 experiments with hot water (50° and 55°)." 



The second group of explanations comes from biolo- 

 gists who accept the logarithmic order of death as an ex- 

 perimental fact and try to explain how the resistance is 

 graded in a different way in bacteria and higher organ- 

 isms. Eeichenbach (1911) and Henderson Smith (1921) 

 are the only authors who seem to have presented really 

 original ideas on the subject. 



Eeichenbach assumed that, in a bacterial culture, a 

 definite proportion of cells from each generation ceases 

 to multiply and becomes increasingly more resistant dur- 

 ing the ensuing rest period. Such a culture would ex- 

 hibit an exponentially graded resistance. This assump- 

 tion, it must be recognized, has no experimental foun- 

 dation. However, it would lead to the same result to as- 

 sume that the bacterial cell does not divide into two equal 

 parts, but that the mother cell produces a daughter cell 

 as in the case of yeasts, and that the mother cell is defi- 

 nitely older and more resistant than the daughter cell. 

 This mode of multiplication would result in an expo- 

 nentially graded resistance which under certain condi- 

 tions could yield a perfect logarithmic order. But the 

 theory, aside from its unverified morphological impli- 

 cations, fails to explain the logarithmic order of death 

 in the case of spores. It seems impossible that the as- 

 sumed type of gradation of resistance could occur in the 

 case of spores. 



Henderson Smith (1921) points out that a variation in 

 resistance may be brought about by differences in the 

 thickness of the cell membranes. Penetration by poison 

 may be proportional to the square of the thickness, and 

 this would result in an order of death approximating the 



