354 HADIATIO.N UIOLOGY 



The denominator in Fa\. (9-2) is the fraction of l)acteria receiving two or 

 more particles; the luimerator is the prol)ahility that the group of k par- 

 ticles infecting a given i)acterium contains one full set of active units. It 

 is possible to determine x and r experimentally. A comparison of ij with 

 the experimental frecjuency w of active phage production gave, at first, 

 results compatible, with some limitations, with E(\. (9-2). The analysis 

 was therefore pushed further along these lines, and estimates were given 

 for the values of n for different phages. 



The tendency of the experimental ratios w/y toward unity for small 

 values of r (low doses) and high values of .r (high multiplicities) suggested 

 furthermore that any mechanism of recombination, if responsible for reac- 

 tivation, should be an exceedingly efficient one in order to allow an essen- 

 tially full utilization of needed units derived from many different phage 

 particles in the formation of active phage. This led to the "gene-pool" 

 hypothesis, according to which each unit reproduced independently of the 

 others, and the resulting new units reassembled to form the new particles. 

 This hypothesis could explain a number of features of the phage repro- 

 duction process. 



Additional evidence, however, has forced revision of one basic assump- 

 tion of the theory that multiplicity reactivation is due solely to a highly 

 efficient mechanism of genetic recombination. According to theory, 

 the minimum requirement should be the integrity of at least one full 

 set of units in the infecting particles. If the frequency of reactivation is 

 plotted against the dose of ultraviolet received by the particles, for high 

 doses the curves should tend to an ultimate slope equal to the slope of the 

 inactivation cur\-e for the free phage since both these slopes represent the 

 probability of persistence of one full complement of active units (Dulbecco, 

 1952). Analytically, it is easily seen that, for very large values of r, 

 Eq. (9-2) tends to the form 



Dulbecco (1952), having by a special procedure obtained data on the 

 frequency of recombination at very high radiation doses, found that for 

 phage T2 the curves for w versus dose reach their ultimate slope much 

 sooner than expected and that this slope is not the same as that of the 

 inactivation curve of the single particles but only about one-fifth of it (see 

 Fig. 9-1). This result indicates that the simple theory is inadequate. 



The situation may be summarized as follows: jMultiplicity reactivation 

 represents the result of a cooperation among inactive phage particles in 

 producing active phage. In this cooperation each particle contributes in 

 a more than additive measure. For high ultraviolet doses, for example, 

 bacteria with three particles have a probability of reactivation several 



