310 C. LEVINTHAL 



is thought of as being responsible for the production of some substance; let 

 us say all or part of a protein molecule. In order to produce viable phage, 

 the infected cell must produce both the A substance and the B substance 

 intact. A cell infected by a mutant will, in general, not make a functioning 

 substance; the protein molecule may not be made or it may be made in a 

 sufficiently altered form so as to be nonfunctional. Therefore, when two 

 different A mutants are added to a K12 (A) cell, the infected complex does 

 not have a good copy of the A substance, whereas if an ^ mutant and a B 

 mutant are added, one produces the B substance and the other the A sub- 

 stance, so that the infected complex has the material necessary to grow new 

 phage. 



Some of the mutants tested are capable of undergoing a certain amount of 

 growth in the cell K12 (A). These mutants, which are designated as "leaky", 

 would under this picture produce a defective, but not totally nonfunctional 

 substance, while the nonleaky mutants and, presumably, the deletions would 

 be ones which were unable to make a product which functioned at all. 



D. Negative Interference 



In many organisms a crossover occurring at one point on the genetic map 

 reduces the probability that another crossover will occur in the same region 

 of the map in the same mating event. In Drosophila (Weinstein, 1958), if very 

 closely linked markers are used, a double crossover is much less probable 

 than the product of the individual crossover probabilities. In phage crosses, 

 on the other hand, there is an apparent negative interference which applies 

 even to distantly linked markers. This means that in a three-factor cross 

 abc X a'^b^c'^ a higher frequency of recombination betw^een the markers b 

 and c is observed in that fraction of the progeny which is also recombinant 

 between a and b than if aU the progeny particles are examined. A double 

 recombinant is thus more probable than would be predicted if the two re- 

 combinational events were independent. However, Visconti and Delbriick 

 (1953) pointed out that this is an expected consequence of the multiple 

 random matings. It is due to the selection of particles which have mated 

 more frequently rather than one recombinational event affecting the prob- 

 ability of another. Since the number of rounds of mating in T2 is large and 

 the estimate of it is rather uncertain, it is difficult to determine if any residual 

 interference exists in this system. But wdth the phage A the total number of 

 rounds of mating is only about one-half, and Kaiser (1955) was able to show 

 quite clearly that interference does not exist in this system — at least 

 for markers which are not closely linked. 



If very closely linked markers are studied, then it is found (Streistnger, 

 1956; Chase and Doermann, 1958) that a pronounced negative interference 

 does exist. Using several closely hnked rll mutants. Chase and Doermann 



