GENETICS OF BACTERIOPHAGE 



unit 

 distance 



h ri3 



FIG. 2. Linkage relations among several genetic markers. 



mutant pair, and the numerical results can be summarized in the form of a genetic 

 map, as shown in Figure 2. On this map we have shown only a few markers that 

 we wish to refer to in this paper. The letter m (minute) stands for a small plaque 

 mutant. It is well established that the three loci, rl , h,tand m assort independent- 

 ly of each other, and that the loci linked to h are arranged in Hnear order. Dr. N. 

 Visconti (personal communication) has recently confirmed the earlier results on 

 these points by a new method of three-point testing. 



Triparental Recombination 



Important information about genetic recombination comes from experiments 

 in which the frequency of triparental recombination is measured. If bacteria are 

 infected with the three mutants h, m, and rl , the recombinant hmr can arise only 

 by interactions involving all three. The results for this triple infection and its re- 

 verse, hr X hm X mr, are shown in Table 1. About three per cent of the viral 

 yield consists of the triparental recombinant in these crosses. This shows that in- 

 teractions among three particles of virus occur with high frequency. A measure of 

 this frequency can be expressed in terms of some artificial assumptions. If we 

 suppose that multiplication precedes recombination, and that recombinants are 

 formed during random successive pairings between phage particles, each particle 

 would have to pair with three other particles to explain the results shown in Table 

 1. The genetic factors used in these experiments are unlinked. The data for tri- 

 parental recombination in experiments with linked markers (Hershey and 

 Rotman, 1948) lead to the same quantitative conclusion. 



This conclusion is a stumbling block to the further understanding of the mecha- 

 nism of genetic recombination, since it is extremely difficult to distinguish be- 

 tween successive interactions by pairs, and other types of interaction that might 

 involve larger groups. The available information is insufficient, therefore, to de- 

 cide between alternative hypotheses of viral interaction so far considered. 



181 



