54 MOLECULES, VmUSES, AND BACTERIA 



the paired segments. In this second process, it is supposed that as the 

 chromosome dupHcates, the dupHcate chromatid is copied first oflF one 

 parental strand and then ofiF the other. Reciprocal recombinant struc- 

 tures are formed because two new chromatids cannot be copied ofiF the 

 same parental strand at the same time. Thus if one chromatid switches 

 in the copying process, the second is obliged to make a reciprocal 

 switch. There takes place a fairly frequent switching back and forth in 

 the copying, and this is what we call negative interference. It is the 

 switching process that really describes the intimate process of crossing- 

 over and its frequency. If we study events which involve distances 

 longer than the e£Fectively paired segments, we underestimate the 

 switching back and forth, bcause we include regions which are un- 

 paired and which, therefore, cannot contribute to recombination. 



This theory of recombination is satisfying in many respects, and 

 no experimental results are in critical conflict with it. Most crosses in 

 classical genetics have dealt exclusively with the probability of ef- 

 fective pairing occurring in regions delimited by genetic markers. 

 Only in systems in which "high resolution" is possible can we study 

 the second process. While the probability of efiFective pairing may be a 

 fairly simple function of the linear dimension of a given segment of 

 chromosome, it remains to be demonstrated that there is a linear 

 relationship between distance and recombination frequency within 

 the effectively paired region. In general, map distances between 

 mutated sites within a single functional gene are only poorly additive. 

 This suggests that there may not be a simple relationship between map 

 distance and recombination frequency within the paired region— a 

 point to which we shall return. 



Setting this difficulty aside for the moment, it should be noted that 

 the model has several qualities. Not only does it provide an explanation 

 for the fluctuations of the genetic yardstick in a single kind of organism, 

 according to the scale of the events under study, but it also explains 

 why constancy of the genetic yardstick throughout the plant and ani- 

 mal kingdoms— which might be expected in view of the apparent uni- 

 versality of DNA as the coding substance— is not, in fact, observed. In- 

 deed, as mentioned above, calculations show that the length of nucle- 

 otide sequence per map unit is 1,000 for phage, and 40,000 for the mold 

 Aspergillus. If, however, experiments are set up to measure recombina- 

 tion in effectively paired regions alone, these values become 100 and 

 270 respectively, which is surprisingly close agreement ( see Pritchard, 

 1960). 



Non-reciprocal recombination 



Before we consider more closely the nature of recombination in 



