82 J. HKIIHKHT TAYLOR 



durinp; replication has been propo.scd (Taylor, 1962) to provide a sorting 

 and control mechanism. Fragmentation of such a structure could give 

 the nmltistranded DNA rejwrtcd for rapidly proliferating cells (Cavalieri 



and Rosenbei'i;, 19()l;i.l)). Fui-thcr details of the model will he presented 

 in Section VII. 



IV. Exchanges between DNA Subunits and Genetic Recombination 



A. CHRO.MOSOMAL LEVEL 



Sister chi-omatid exchanges had been indicated by the behavior of 

 ring chromosomes in maize (McClintock, 1938; Schwartz, 1955) but the 

 autoradiograms of tritium-labeled chromosomes (Fig. 2) provided unmis- 

 takable visual evidence for their occurrence in other species (Taylor, 

 1958a, 1959a). Their frequency and pattern could be determined in well- 

 spread preparations of chromosomes at the second division after labeling 

 Exchanges were not revealed in the first division because all chromatids 

 were labeled. When the exchanges became visible they had to be 

 interpreted in terms of events during two duplication cycles — the one 

 in which labeling occurred and the succeeding one. 



An analysis was made from root cells of Bellevalia, which has four 

 pairs of large chromosomes (Taylor, 1958a). Attention will be directed 

 to the exchanges in the largest chromosome of the complement. Since 

 we are dealing with diploid cells, there are two of the large homologous 

 chromosomes, each composed of a single chromatid before duplication. 

 Each chromatid contains two subunits of DNA that extend throughout 

 the length— these are the units conserved during replication. Aftei' 

 duplication the chromosomes consist of two chromatids, each with two 

 subunits, one new labeled subunit and one unlabeled original. If the 

 subunits are analogous to the polynucleotide chains of the Watson- 

 Crick double helix, they will be unlike. Reciprocal exchanges between 

 the two chromatids will require the union of labeled subunits with 

 unlabeled subunits in each chromatid because only like chains can join 

 (Fig. 5). Since each chromatid will have a labeled subunit along the 

 entire length at the following metaphase, there will be no visible evidence 

 of the exchange. How(>ver, when the two subunits separate and replicate 

 at the next duplication (referred to below as the second interphase), one 

 chromatid will have the labeled subunit up to an exchange point and 

 beyond that locus its sister chromatid will contain the labeled subunit. 

 The exchanges will now be revealed by autoradiography at the succeed- 

 ing metaphase, because silver grains will be over one chromatid from 

 the end to an exchange and over its sister chromatid from that point to 



