LOGARITHMIC ORDER OF DEATH 17 



Table 1 gives other examples. The difference in the 

 order of death between bacteria and multicellular organ- 

 isms is here shown by the computation of the value 



j^ _ _1_ |q„ initial number 



t survivors 



This value, the ''deathrate constant," is constant if the 

 order of death is logarithmic (see p. 30). The table 

 shows that, with bacteria, the successive values for K 

 either fluctuate around an average, or eventually de- 

 crease with time, while, with higher organisms, they in- 

 crease. 



When no special effort is made to experiment with a 

 homogeneous material, the survivor curve^ may not be 

 straight but concave upwards. This is in fact the most 

 frequent occurrence. This type of curve differs from 

 that of higher organisms still more than the rectilinear 

 curve. Its concavity is due to lack of uniform resistance. 

 The more sensitive bacteria w^ill die rapidly, causing a 

 steep decline in the number of survivors. When most of 

 these bacteria are dead, the remaining ones die at a lower 

 rate, and as the less resistant individuals are gradually 

 removed, the survivor curve becomes less and less steep. 



A simple, theoretical example is given in Table 2. 

 Three groups of bacteria, each containing 1,000 indi- 

 viduals, are supposed to be mixed and exposed to heat. 

 Each group represents a different degree of resistance, 

 the individuals of one group dying at the rate of 90% 

 per minute, those of the other two groups at 50% and 

 10%, respectively. The death rate of the composite 

 sample is not increasing as it would be with higher or- 

 ganisms of graded resistance, but it is decreasing. The 

 survivor curve (Fig. 3) is concave upwards, not down- 

 wards. 



1. Unless otherwise indicated, the discussion of the characters 

 of survivor curves will always refer to semi-logarithmic plots. 



