BACTERIOPHAGE GENETICS 351 



that formation of heterozygotes is probably the mechanism for all 

 recombinations occurring within Hnkage groups. 



5. Pseudoallelism 



The gene is defined by the different operations used in studying 

 its properties. Thus the gene may be defined in terms of its 

 inutabiUty or of crossover distance from some other marker or 

 perhaps eventually in terms of a specific nucleic acid structure. 

 It has becom-e evident that these various definitions are not de- 

 scriptive of the same physical entity and hence that the gene 

 concept may vary with the operations used to study its proper- 

 ties. This is particularly evident in phage genetics. It was 

 demonstrated by Benzer (1955, 1957) that the rll region in the 

 genetic structure of phage T4 can be divided into subunits in 

 terms of function, mutation, and genetic recombination. 



The rll cluster of mutants is located in a definite region of 

 linkage group II of phage T4 and related phages. The mutants 

 are identified by plaque morphology and by a unique host range 

 peculiarity that distinguishes them from other r mutants. The 

 rll mutants are unable to form plaques on the host K12 lysogenic 

 for phage lambda whereas the corresponding r+ phage particles 

 plate on K12 with the same efficiency as on host strain B. Mixed 

 infection of a susceptible host strain by two phage particles with 

 distinct but closely linked rll markers will result in the liberation 

 in the progeny phage of some wild type particles which can then 

 be detected by plating the progeny on the indicator host K12. 

 The resolving power of this indicator system for closely linked 

 rll markers is limited mainly by the background counts of r+ 

 particles produced by the spontaneous mutation of the rll 

 parents. Recombinant frequencies between loci separated by a 

 linkage distance as short as 0.01 per cent have been measured 

 and the resolving power is at least 0.0001 per cent for most rll 

 loci. Benzer isolated a large number of independently occurring 

 rll mutants and arranged in a linear map those loci that were 

 resolvable (Figure 12). The map distances are not additive in 

 part because of low negative interference of the kind predicted 



