BACTERIOPHAGE GENETICS 305 



elementary mating event involves the pairing of two particles, which co- 

 operatively produce a third which contains the overlap region where the 

 contributions from one parent end and that of the other parent begins, as 

 in Fig. 4 (c). We will assume that this is the only way recombinants can 

 arise, but that particles can also duplicate by themselves without mating. If 

 a heterozygous particle duplicates without mating, it is assumed to segregate 

 and produce its two daughter types within the same cell in which it was pro- 

 duced. We will also assume that the mating is random within the vegetative 

 pool and that mixing is complete so that no particles are excluded from the 

 mating process. We will further assume that if a mating event of the type 

 described here takes place, the probability of the overlap and the average 

 length of the overlap will be the same for the phages T2, Tl, and A, but that 

 the difference between these systems will be in the frequency with which the 

 mating occurs, as compared to the growth without mating. Granting these 

 assumptions, the fraction of the total progeny w-hich will be produced in a 

 mating event can be estimated for T2 as about one-half. This estimate can 

 be obtained in several ways, but none of them is very accurate. One is from 

 the frequency of double heterozygosis for two distinctly linked markers, 

 which Hershey found to be 6 % as compared to the 2 % which is the fre- 

 quency for a single marker. By selecting a plaque that shows mottling, one 

 is selecting a phage which arose in a mating event between two particles 

 which differed with respect to the r marker involved. Thus, we are excluding 

 from consideration those matings which are incestuous in that they involve 

 pairing between genetically identical particles. At the first mating, two 

 particles which differ with respect to the r marker must also differ with 

 respect to all other markers, and one would expect the frequency of observed 

 heterozygosity for an milinked h marker to be double the usual value of 2 %. 

 However, this effect would become less significant as more recombinants 

 were formed in successive rounds of mating. If the probability of the occur- 

 rence of a heterozygous region were equal for all parts of the genetic structure 

 and if every particle could be heterozygous, then we would expect to find 

 about 3 % of the mottled plaques also heterozygous for a distantly hnked 1i 

 marker. The fact that one observes 6 % would by this model imply that about 

 half of the observed particles in the progeny came from a mating event and 

 the rest did not. 



An analysis of similar recombinational data for Tl and A indicates that 

 there are far fewer mating events per growth cycle with these phages than 

 with T2 or with T4. Therefore, it is very much less likely that any particle 

 observed in the progeny will have arisen in a mating. This difference in the 

 number of mating events also reflects itself in the number of heterozygotes 

 observed in the progeny of the cross. Only those particles which have just 

 arisen in a mating can have received genetic information from two different 

 VOL. II — 20 



