RECOMBINATION ANALYSIS IN MICROBL\L SYSTEMS 69 



small. Indeed, non-equivalent recombination frequencies in reciprocal 

 crosses would be obtained in crosses involving two point mutations, one 

 of which was situated very near an end. The data of Lacks and Hotch- 

 kiss ( 1960 ) cannot be explained in this way, though in crosses involving 

 a terminal marker ( c or / ) this factor may be operating for one of the 

 two terminal markers. The strongest reason for excluding the distance 

 from an end of the DNA molecule as the sole cause of the asymmetries 

 of the data on the amylomaltase locus is the fact that the center mutant 

 g gives asymmetrical results, and in the same direction, with both ter- 

 minal markers. Unless we suppose that the locus occupies the entire 

 DNA molecule, c and / cannot both be near ends. Current estimates of 

 the size of a locus argue against this. Thus the model taking the size of 

 m.utant sites into account has unique explanatory properties, in that it 

 explains asymmetries regardless of the positions of markers with respect 

 to the ends of the DNA molecule. 



A mathematical formulation of this recombination model has been 

 elaborated by Prevost ( unpublished ) , and its usefulness in mapping in 

 transformation is under investigation. The mutants discussed above are 

 too few in number to provide a criticabtest of the model. In particular, 

 the series contains too few "point" mutants to be very useful in a quan- 

 titative test. Therefore we must first obtain a new series of mutants 

 which, on the basis of as many criteria as possible, can be considered to 

 be point mutations. For the present, rather than enter into further de- 

 tails of the model, its limitations, and its points of oversimplification, it 

 seems more worthwhile to discuss to what extent this model may be ap- 

 plicable to other genetic systems. 



Applicability of the model to other systems 



The principal reason for believing that the model may generally 

 describe recombination on the molecular scale is that the basic assump- 

 tion made is essentially that advanced to explain the high frequency of 

 recombination observed in fine-scale analysis of chromosomes in other 

 genetic systems. This assumption is that when chromosomes or DNA 

 molecules are effectively paired, multiple cross-over events will occur 

 in the paired structure or in progeny patterned on it. The present 

 model goes farther only in giving a concrete form to these cross-over 

 events. It thereby provides a theory which can be tested exeprimentally. 

 It may seem strange that a detailed model has been presented only for 

 transformation, in view of the fact that recombination data in this type 

 of system are still very fragmentary. However, transfonnation ( or trans- 

 duction, for that matter) offers a unique analytical situation in two 

 respects. The "male" element of a cross is only part of a genome-an 

 isolated DNA molecule. Because of this, only one kind of recombina- 



