GKNETIC RF.COiMRlNATION IN BACIFRI Al. \ IIIUSI'S c; 



of the theory, predictions about the result of some crosses have been 

 made. Up to now these predictions have been verified. 



The drift to\\'ard genetical equiHbrium can now be easily ex- 

 plained by assuming that mating among vegetative phage particles 

 goes on at a constant rate with time. The phage which matures at 

 any moment inside the bacterium is withdrawn from a population of 

 vegetative phage which has undergone a certain number of rounds 

 of mating. The later the process of maturation is stopped by delaying 

 the lysis, the higher will be the average number of matings and the 

 greater w^ill be the number of recombinants found. Two other impor- 

 tant facts are explained by this theory. One is the lack of correlation 

 bet\\'een recombinants. In a cross AB x ab, if the yield of a single 

 burst is examined, the number of recombinants Ab and aB should be 

 the same. A lack of correlation between the two frequencies means 

 that the process by which one recombinant is formed is independent 

 of the formation of the other one. This fact is accounted for by this 

 theory, as the maturation of a single vegetative phage into phage is an 

 event independent of the maturation of any other particle. The other 

 fact accounted for by this theory is an apparent negative interference, 

 which is that recombinants between some two markers show a higher 

 value of recombination between any other markers. It has already 

 been mentioned that, if the crude recombination data of a cross are 

 used to make a map, a linear arrangement of the markers belonging 

 to the same linkage group will be obtained. Let us suppose that there 

 are three unlinked factors. A, B, C. Considering any two of these 

 markers, we will get in a cross 60 per cent parental types and 40 per 

 cent recombinants. This fact can be explained as due to an incomplete 

 genetical equilibrium reached by the population of vegetative phage 

 at the moment of maturation. Suppose now that we compare a paren- 

 tal type to a recombinant, say AB to aB inside the class of recombi- 

 nants Be. Let the ratio be 1:1. Because of the random distribution 

 of the number of matings per particle and because any mating with a 

 similar type gives no possibility of recombination, we operate a 

 selection in favor of recombination when we choose particles which 

 have recombined for some other marker. When we apply the right 

 calculation, this apparent negative interference disappears if we con- 

 sider a single mating. Thus genetic maps can be made by calculating 

 the average number of matings. For normal lysis in phage T2 a value 



