RECOMBINATION IN BACTERIOPHAGE 61 



Some of the tubes contain one or more mixed bursts plus one or more unmixed 

 bursts. Such tubes would tend to show less than average values of k and less 

 than average yields of recombinants together with greater than average burst 

 size. The number of such tubes is about five percent of the total in our experi- 

 ments and these tend to be concentrated in the class k^0.2. This probably 

 explains the negative correlation between proportions of recombinants and 

 burst size among tubes selected for small k. 



A larger proportion of the tubes, between 10 and 20 percent in our experi- 

 ments, contain two or more mixed bursts. These tubes would tend to have 

 greater than average k and greater than average burst size, but will not show 

 exceptional proportions of recombinants. These tubes have the effect of weak- 

 ening any correlation that may exist between burst size and proportion of re- 

 combinants, especially among the class k^0.21. The data previously consid- 

 ered do not, therefore, exclude the possibility of a weak positive correlation 

 between proportion of recombinants and size of bursts. 



Advantage was taken of the finding that high pH reduces size of bursts, to 

 examine the relation between burst size and yield of recombinants in another 



Table 7 



Small Bursts from the Cross hXr7 in Broth of pH 9.0 



See legend table 4. The proportions of recombinants have not been corrected for dispropor- 

 tionate yields of parental types of virus. The individual cultures contain an average of about 1.4 

 mixed bursts. 



way. Since increased pH was found also to cause some of the bacteria to fail 

 to liberate virus, the single burst technique was chosen. 



Bacteria were infected in the usual way with mutants h and r7 in broth of 

 pH 6.8. At the end of the five minute adsorption period, and without treatment 

 with antiserum, single burst cultures were prepared after diluting in broth of 

 pH 9.0. Preliminary experiments showed that the yield of virus was complete 

 under these conditions within one hour, and than no inactivation occurred 

 during two additional hours at 37°C, or overnight in the refrigerator. Entire 

 samples were plated on single mixed indicator plates. One successful experi- 

 ment yielded 66 bursts, of which 54 were mixed, among 104 tubes receiving on 

 the average 1.4 bacteria per tube. Evidently most of the bacteria liberated 

 some virus. The average burst size, after subtracting the virus carried over 

 from the input (totalling 36 particles per tube), was only 114 per tube, or about 

 70 per bacterium corrected for the probable multiples. The average proportion 

 of recombinants was nevertheless of normal size (table 7). 



It will be noticed also that the correlation between the numbers of the two 

 recombinants in these bursts is exceptionally good. The correlation is, however, 



255 



