APPENDIX '1 89 



35" X e~^-^^ 



P(2) = = 0.04, or 2 plates out of 50 with 2 infected 



2X1 



bacteria 



35^ X e~'^-^^ 



P(3) = — = 0.005, or no plates in 50 with 3 infected 



^ ^ 3X2X1 ^ 



bacteria 



Presumably the actual distribution was 13 plates with 1 infected 

 bacterium and 2 plates with 2 each, or 17 infected bacteria in all. 

 These produced a total of 1750 plaques, or an average burst size 

 of 1750/17 = 103 phage particles/bacterium. 



Average plaque count/0.5 ml. sample taken from the sample 

 flask after the end of the rise period was 39, which amounts to 

 an average burst size of the bacteria in the sample flask of 39/ 

 0.35 = 111. 



e. Discussion. It must be remembered that in this case the aver- 

 age burst size from the single cell bursts is calculated on but 17 

 bacteria and from the average in the sample flask from about 32 

 bacteria, so that neither figure is a reliable estimate of average 

 burst size. This is particularly true since there is such a wide 

 distribution in individual burst sizes, which range from 8 to 319 

 in the foregoing example. Of course, 2 of the plates contain 

 phage yields from 2 infected bacteria each, but it is impossible to 

 tell which plates are involved and there is certainly no a priori 

 reason for suggesting that these are the plates with the highest 

 yields. This broad range in individual burst sizes is the inter- 

 esting fact brought out by this experiment and one which could 

 hardly be demonstrated in any other way. There is no ade- 

 quate explanation for this extreme range in burst sizes, but it can 

 hardly be correlated in any simple way with the size of the host 

 cell, since Delbriick found that the spread of burst sizes is much 

 greater than the spread of cell sizes (Delbruck, 1945a). Also, he 

 found that the distribution of burst sizes was not dependent on 

 whether the samples were plated early or late during the rise 

 period (Ellis and Delbruck, 1939); i.e., the burst size of a bac- 



