RECOMBINATION IN BACTERIOPHAGE 



51 



linked to each other, but rl is probably situated on an independently ex- 

 changing structure. 



The results of table 2 show further that the two recombinants appear in 

 equal numbers in any one cross, and that pairs of reverse crosses yield equal 

 numbers of recombinants. It is these relations, which increase the resemblance 

 to simple types of Mendelian segregation, that we wish to examine by the 

 single burst technique. In the remainder of the experimental part of this 

 paper we describe the results of this examination, but limit our comments 

 chiefly to the technical problems encountered. The general implications of the 

 data will be considered in the discussion. 



VARIATIONS IN YIELDS OF PARENTAL TYPES OF 

 VIRUS AMONG INDIVIDUAL BACTERIA 



Yields of virus from single bacteria show large fluctuations in size (Del- 

 BRiJCK 1945b) and, in our experiments, variations in relative yields of the 

 two infecting viruses. The variations in total yield are shown in fig. 5 which 

 includes the complete data for six experiments in which the proportion of 

 multiple bursts is small (100 bursts out of 484 tubes, or 11 probable multi- 

 ples). The burst sizes range from 150 to several thousand, with a mean of 

 520, or 470 corrected for probable multiples. The distribution is the same for 

 the mixed bursts, and for the bursts containing only one viral type. Owing 

 to these variations, it is convenient to describe the individual bursts in terms 

 of the fractional yield of the several viral types. 



400 500 600 

 Burst size 



>1000 



Figure 5. — Distribution of total viral yields among 100 single bursts from mixedly infected 

 bacteria. The solidly shaded areas refer to unmixed yields. 



Each cross yields four types of virus, wild type, the single mutants h and 

 r, and the double mutant h r, of which two are parental and two are recombi- 

 nant types. In order to examine the variations in relative yields of parental types 

 independently of variations in yields of recombinants, it is convenient to ex- 

 press the former in terms of the proportion of virus containing a specified 

 allele, viz: 



158 



