11. REPLICATION OF DNA IN CHROMOSOMES 83 



another exchange. The nietaj)hase chroniosoine derived from tlie other 

 sister chromatid of this originally labek'd j)air will show an exchange 

 at the same locus (the twin exchanges referred to hereafter). The two 

 metaphase chromosomes descended from the other liomolog in these 

 diploid cells will have twin exchanges at different loci if an exchange 

 occurred during its labeling, i.e., the exchanges will be at different loci 

 because they are independent events. 



Unpaired (single) exchanges will appear as a result of exchanges 

 during or following the duplication at the second interphase. These will 

 be independent events in all four chromosomes and will be revealed im- 

 mediately when the chromosomes reach metaphase. 



The tendency of exchanges to occur as twins was evident in the first 

 sample of cell examined (Table I). In fact these were more numerous 

 than expected compared to the single, unpaired exchanges. If exchanges 

 at the second interphase occurred at the same frequency as in the first 

 interphase, i.e., the one in which the label was incorporated, the ratio 

 would be one twin pair (product of a single reciprocal exchange) to each 

 two single exchanges produced at the second interphase. Since there are 

 four reproducing chromosomes at second interphase if all descendants 

 of the original two homologs are considered, the chance for single 

 exchanges is increased by a factor of two compared to first interphase. 



When the two subunits of a pair of chromatids are unlike as assumed 

 in the prediction of the 1 : 2 ratio of twin and single exchanges, all of the 

 exchanges in the first interphase will produce twins and all of those in 

 the second interphase will produce singles. On the other hand, if the two 

 subunits were alike as would be true if each were represented by a 

 Watson-Crick double helix, only one-fourth of the exchanges of the first 

 interphase would yield twins and, of course, no twins would be produced 

 from exchanges at the second interphase. Therefore, the pi'edicted ratio 

 would be one twin pair to each 10 singles (Taylor, 1958a). 



The results of the first experiment revealed a 2:1 ratio, i.e., more 

 twins than predicted but a ratio not at all compatible with results pre- 

 dicted in a chromosome consisting of identical subunits. The deviation 

 appeared to be the result of an increased exchange frequency at the 

 interphase when labeling occurred or a decreased one at the second 

 interphase. The latter proved to be correct. Colchicine was present at 

 the second interphase, but not duiing most of the first. It was applied 

 just before the labeled cells, which were being analyzed, reached division. 

 By placing the roots in colchicine at the same time labeling began, i.e., 

 along with the thymidine-H'', the ratio was shifted in the predicted 

 direction, i.e., 1:1 (Table I). In a third experiment the colchicine was 

 added 2 hours before labeling began. In this instance the ratio was 



