LOGARITHMIC ORDER OF DEATH 47 



cussed elsewhere, but the principle should be discussed 

 here. In those instances in which this outstanding resis- 

 tance of a few individuals was oberved, the number of 

 resistant cells was always very small, usually much small- 

 er than 0.1%. Even if these cells were morphologically 

 different, it would be difficult to find them under the 

 microscope among the mass of normal cells. But it is 

 probable that they are not different, for their progeny 

 differs in no way from the parent culture (not even in 

 resistance, as will be seen below). 



It may be that the few survivors are very old cells 

 which failed to come out of their resting stage when 

 transferred to a new medium. In general, one loopful of 

 a grown culture transfers between one million and ten 

 million cells. If some of these remain in the resting 

 stage for 24 hours, a new transfer from this culture 

 would carry over a considerable number of cells much 

 older than the average, and they would not multiply 

 readily, but would more probably remain in the rest- 

 ing stage. It would require at least three successive 

 transfers at short intervals to eliminate such cells. This 

 explanation does not require that the resistance increase 

 with age. It merely assumes that all other cells are much 

 younger and therefore more sensitive. 



The Or d e r of Death in the Case of 

 Low Death rates. When death is very slow, the 

 survivor curves seem to be generally concave downwards. 

 However, most of Knaysi's experiments (1930, II) with 

 very weak disinfectants show a decreasing death rate and 

 survivor curves which are concave upwards. This indi- 

 cates some inhomogeneity in the material, which was 

 probably due to the use of agar-grown bacteria instead of 

 broth-grown cells. Woerz (1931, unpublished) observed 

 that Torula cremoris, which does not form clusters, died 

 in logarithmic order when held at a temperature well 

 above the maximal temperature of growth; but at tem- 



