350 BACTERIOPHAGES 



time in a manner predictable from the proportion of recombi- 

 nants in the yield, indicating that both recombinants and hetero- 

 zygotes are drawn from the same vegetative pool, and that hetero- 

 zygotes must be formed continuously during phage replication. 

 Further evidence in support of model II was obtained froin an 

 analysis of the genotypes of the segregants from heterozygous 

 particles. The heterozygotes were obtained from the hrl X 

 r2 cross by premature lysis at an average burst size of one. 

 The r character of 106 mottled plaques was determined by 

 plating a sample from each mottled plaque, picking 5 r plaques 

 from each sample and back-crossing each with both r2 and rl 

 stocks. The types obtained were 54 rlh; 7 r7A+; 34 r2h; 

 11 r2A+. The proportion of the h alleles among the r2 hetero- 

 zygotes was thus 75 per cent as compared with an expected value 

 of 82 per cent from model II after allowing for recombination 

 that had already taken place in the vegetative pool. As was 

 pointed out above, the r2 heterozygotes must have been r+ at 

 the rl locus in order to have formed mottled plaques, and this is 

 confirmed in the back crossing of the segregants. Therefore 75 

 per cent of the particles heterozygous for the central r2 locus were 

 recombinants for the end markers h and rl . As lysis is delayed 

 and the recombinant frequency increases, the fraction of the 

 heterozygotes showing recombination for the end markers de- 

 creases, approaching the equilibrium value of 50 per cent as 

 shown in Table XVIII. Because model II has proved correct 

 it is evident that formation of heterozygotes may be a major 

 mechanism of genetic recombination in phage. For instance all 

 segregants of heterozygotes at the r2 locus from an hrl X r2 

 cross must be recombinants for the h and rl loci, and half of them 

 will be recombinants for the h and r2 loci as well. On the basis 

 of the known frequency of formation of heterozygotes and an 

 estimate of the length of the overlap region, Levinthal (1954) 

 calculated the frequency of recombinants between the closely 

 linked h and rl3 markers that could be formed by the mechanism 

 of the heterozygotes. The calculated frequency of recombinants 

 was the same as the observed frequency leading to the conclusion 



