120 S. E. LURIA AND R. DULBECCO 



critical test. In postulating a phase in phage growth in which particles, as we 

 know them in the extracellular phase, are not present, the hypothesis accounts 

 for the repeated failures to obtain active phage by premature artificial dis- 

 ruption of infected bacteria. It agrees with the observation made by Foster 

 (1948) in our laboratory that in the presence of proflavine the reactions leading 

 to production of active phage proceed normally for a part of the latent period, 

 but, upon lysis, no active particle is liberated. According to Foster (1948) 

 active phage particles are present in the infected bacterium after the proflavine 

 sensitive stage is passed — beginning 12 to 14 minutes after infection for phages 

 TZ or T6. A similar conclusion was reached by A. H. Doermann on the basis 

 of experiments on the effect of other inhibitors on the growth of T3 and T4r 

 (Doermann 1948). That viruses multiplying inside the host cell may not have 

 the same organization as in the extracellular form has been suggested before 

 for certain animal viruses (see Bland and Robinow 1939). 



Our theory does not assume any degree of linkage among units, although 

 linkage may be compatible with the theory. Each group of strongly linked 

 units would behave as one unit, possibly as a particularly sensitive one. Weakly 

 linked units would probably reduce the probability of reactivation for high 

 doses, when lethal units present in one group would hinder the utilization of 

 the linked active units for reactivation. 



In their work on the r and h mutants of phage TZ H, Hershey and Rotman 

 (1948, 1949) found evidence for a series of determinants exhibiting various 

 degrees of linkage, from those apparently unlinked (high frequency of inde- 

 pendent transfer, no correlation between the frequencies of complementary re- 

 combinant types in the yield) to others rather strongly linked. For the latter 

 ones, the authors considered that their results suggested the possibility of re- 

 ciprocal exchanges. 



As a working hypothesis we may assume, together with Hershey and Rot- 

 man (1949) that unlinked determinants may be located in different reactivation 

 units, possibly transferred by a gene pool mechanism, while linked determinants 

 may be located in the same reactivation unit. If reciprocal exchanges were 

 found to occur, then homologous units or groups of units should be supposed to 

 pair or group together at some stage in the growth process. Techniques recently 

 developed in our laboratory should soon permit a study of the inactivation of 

 individual genetic determinants and a solution of some of these problems. 

 It appears, therefore, advisable to refrain from further discussion at the present 

 time. 



The formation of active phage from inactive by transfer of discrete units 

 requires some revision of the interpretation of experiments on irradiation of 

 phage inside infected bacteria with ultraviolet light (Luria and Latarjet 

 1947). In case of multiple infection, the survival curves for phage-producing 

 ability immediately after infection indicated suppression by damage of a num- 

 ber of centers, with the sensitivity of individual centers lower than that of 

 extracellular phage particles. It now seems clear that what was measured was 

 the rate of inactivation of individual units rather than of whole particles. 

 The curves for suppression of the phage-producing ability of multiple-infected 



293 



