96 J. HERBERT TAYLOR 



throu<ili till' I'ontiiiiK'd ^i-owtli of the 3' ends. One of the most interesting 

 consequences of such a mechanism of reunion of broken ends would be 

 the elimination of single strands of DNA over a part of the replicating 

 unit. Perhaps more interesting is the idea that it would j)n)vide a mecha- 

 nism for copy choice ovvv short segments in combination with breakage 

 and exchange between homologous chromatids. Such a mechanism of 

 reunion could exjilain all of the types of aberrant recombination reported 

 in ascomycetes (Olive, 1959; Kitani et a/., 1961; Case and Giles, 1958; 

 Rizet et ai, 1961) and perha|)s also in phages (P>igar, 1961 ». 



The basis for the above hypothesis is somewhat weakened by subse- 

 quent studies of reunion of radiation-induced breakage. Chromatid ex- 

 changes apjiarently can occur in the presence of the block by FUDR. 

 This would suggest that many of the breaks do not involve DNA or 

 that another component, probably protein, is able to bring about I'eimion. 

 Such a mechanism is suggested by the model of chromosome structiu'e 

 presented in the following section. 



VII. Molecular Models for Organization of DNA 

 INTO Chromosomes 



Models which allow us to explain data, to express more concretely 

 certain concepts, and to plan and design future experiments are useful. 

 However, models have the disadvantage that they tend to freeze our 

 thinking in such a way that significant data may be neglected. When 

 we are fully aware of such disadvantages models can be useful, even 

 when the information for making them is limited. Several recent attempts 

 have been made to construct models of a chromosome (Schwartz, 1955; 

 Bloch, 1955; Taylor, 1957; Ris, 1957, 1961; Gall, 1958b; Freese, 1958; 

 Taylor, 1958d; Steffensen, 1959; Stahl, 1961). All of these are useful in 

 explaining some data and in ciystallizing concepts, but all are deficient 

 in other respects. 



Any model of a chromosome based on the present information should 

 be composed of one, or at most a few, DNA double helices that extend 

 the length of a chromatid. These must function as two subunits during 

 replication at all levels of organization. There is also considerable 

 evidence that the chromatid functions as two subunits in non-reciprocal 

 recombination (Olive, 1959) and in the induction and segregation of 

 mutants induced by base analogs (Strelzoff, 1962). However, in chemi- 

 cally and radiation-induced breakage and exchange as well as I'eciprocal 

 recombination and sister chromatid exchanges it acts as a single unit 

 except during prophase. On the other hand, the pattern and frequency 

 of sister chi'oinatid exchanges indicate that, even though the exchanges 



