42 LOGARITHMIC ORDER OF DEATH 



IV. CONDITIONS RESULTING IN A NON-LOGARITHMIC 

 ORDER OF DEATH 



Cell Clusters a n d Chain Formation. 

 The logarithmic order of death of bacteria can be estab- 

 lished only by counting the survivors, and the only prac- 

 tical method in use is the plate count method which gives 

 the number of colonies developing from a known volume 

 of bacterial suspension. The number of colonies repre- 

 sents the number of original bacteria only when the bac- 

 teria are single. If they are clustered like Staphylococ- 

 cus, or in long chains like Bacillus mycoides or some 

 Streptococcus strains, one colony may represent a large 

 number of cells. As long as not all cells from such a clus- 

 ter die, the rest will still give rise to a colony, and death 

 becomes evident only when the last of the cells in that 

 cluster is dead. Such clusters *'die" like multicellular 

 organisms, and their survivor curves must be concave 

 downwards. This is commonly observed with Staphylo- 

 coccus (see Fig. 7). 



Substantial support was given to this explanation by 

 Wyckoff and Rivers (1930) who irradiated bacteria 

 spread of an agar surface (see Fig. 8). In one experi- 

 ment the bacteria were exposed immediately after hav- 

 ing been spread ; the survivor curve was then nearly rec- 

 tilinear. In another experiment the bacteria remained 

 several hours on the agar surface before irradiation, 

 and had multiplied to aggregates of perhaps 4 to 16 

 cells. As long as not every cell in such an aggregate 

 was killed by X-rays, a colony developed, and the sur- 

 vivor curve was similar to that of a multicellular organ- 

 ism. Many recorded bacterial survivor curves resem- 

 bling those of higher organisms can be interpreted in this 

 way (Rahn, 1930). 



The Effect of the Age of the Cell. 

 It has been demonstrated by Sherman and Albus (1923) 



