118 S. E. LURIA AND R. DULBECCO 



According to our analysis, active phage is reconstituted from inactive by re- 

 incorporation of active units derived, directly or indirectly, from the inactive 

 particles in a kind of hybridization. As in hybridization, the possibility of 

 recombinations between particles of different wild-type phages suggests the 

 existence of common genetic determinants and the absence of complete incom- 

 patibility. It is possible that various interference phenomena among different 

 phages may result from such an incompatibility. 



Reassembly of material from inactive particles into active ones is a remark- 

 ably efficient process. Genetic material from several inactive particles may be 

 brought together, although the relative efficiency of cooperation diminishes as 

 the number of particles involved in this pluriparental reproduction increases. 



We have given estimates for the minimum number of transferable units per 

 particle for several phages. It is interesting to notice that phage T4, more resist- 

 ant than the related phages T2 and T6 to ultraviolet light (figure 1), appears 

 to have fewer units. This may indicate absence of a portion of genetic material 

 present in the other T-even phages. 



Lea and Salaman (1946), analyzing the dependence of the rate of inacti- 

 vation of phages by X-rays as a function of the density of ionization, and 

 assuming that the radiosensitive material consisted of spherical units, arrived 

 at the conclusion that a large phage contained 14 such units, whereas a small 

 phage contained one only. Although the hypotheses involved were probably 

 oversimplifications, the conclusion receives qualitative support from our re- 

 sults. 



We must consider next the possible mechanisms of genetic transfer. We may 

 divide the mechanisms into two groups, those in which the reproducing element 

 is supposed to be at all times the phage particle as a whole, and those in which 

 the reproducing elements are assumed to be component parts of the particle. 



The simplest hypothesis of the first group would be that reactivation 

 results from pairing (or grouping) or the initial infecting particles, followed 

 by reciprocal exchanges such as occur in chromosomal crossing-over; if these 

 exchanges lead to formation of an active particle, the latter proceeds to multi- 

 ply. This simple hypothesis can easily be disproved. The high efficiency of 

 reactivation would require very large numbers of successive reciprocal ex- 

 changes to bring together all active genetic material before multiplication takes 

 place. Mixed infection with one active and one inactive particle should also lead 

 to the occasional loss of active phage, since we know that genetic recombina- 

 tions occur between active and inactive particles. Finally, the fact that in- 

 fection with one particle each of inactive T2 and inactive T2r yields a mixture 

 of active T2 and active T2r disproves this hypothesis, since any number of 

 reciprocal exchanges between the original particles before multiplication could 

 never lead to formation of active particles of both types. 



Another interpretation based on reciprocal exchanges would be that inactive 

 particles reproduce, and that exchanges occur at various stages of the repro- 

 duction among the original particles or their inactive offspring. These ex- 

 changes should be numerous enough to make the probability of incorpo- 

 ration of all active units into an active particle close to unity, and an ac- 



272 



