98 J. in:RRERT taylor 



two chains and located opposite a gap in the complementaiy chain. For 

 modifications and other features of the model see papers by Taylor 

 0958(1; 1959a. h; 1962) and Kcllenbcrgcr (1960). Linkers have not yet 

 been identified in analyses of DNA and indeed this can hardly be 

 expected if they separate segments with molecular weights of several 

 millions. Since such linkers can only be found by a search directed 

 specifically to the problem and then probal)ly only after improvement 

 in present te('hni(iues, a little sj^eculation on the nature of the linkers 

 may be useful in the interim. 



In addition to a capacity for folding and efficient packaging which 

 might be provided for by irregularities such as odd bases, for example, 

 methylaminopurine (Dunn and Smith, 1955 and 1958), a chromatid must 

 have many growing points for replication and yet be closed to replication 

 for long periods of function. The initiation and control of replication 

 would appear to require operator sites to explain the observed control in 

 order or sequence and to prevent more than one replication at each 

 locus. To provide growing points breaks might be made by an enzyme, 

 but to explain how the two subunits of a chromatid can remain intact 

 or undergo only a few exchanges during replication the breaks must be 

 produced in a very precise manner which would require regular structural 

 sites for initiation of replication. To provide the necessary controls one 

 may have to imagine several specific inducers for replication, the speci- 

 ficity of which need not reside in the type of linkers but in the neigh- 

 boring nucleotides. In addition, a minimum of four types of linkers 

 appears to be required. The first type of linker needed is one to join 5' 

 OH or 5' phosphate groups at the ends of two polynucleotide chains 

 with a reversal in polarity (Fig. 7). This will be referred to as a 5' 

 linker. Their complementary chains will be assumed to be joined by a 

 3' linker. The chromatid is assumed to consist of a series of tandemly 

 linked segments (replication units) of DNA with a 3' OH group linked 

 to a 3' OH group and a 5' end linked to another 5' end at each operator 

 site. Any irregularity in single chains of the DNA double helix wouM 

 appear to be inadequate, for these would tend to be redistributed by 

 inversions or other chromosomal aberrations in such a way as to prevent 

 the subunits from maintaining their alternating sequence of linkers. 

 Without such a sequence there would be numerous mistakes in sorting 

 during replication and the subunits would not maintain their integrity. 

 With linkers of the type proposed above, chromosomal aberrations might 

 change the length of segments but would not destroy the alternating 

 sequence of 3' and 5' linkers. 



The 5' linkers would always be closed and in non-replicating DNA 

 the 3' linkers would also be closed. One of the first steps preparatoiy to 



