xvi Introduction 



polynucleotide chains separate and each parental chain acts as the template 

 for the de novo synthesis of a complementary daughter chain, a pair of DNA 

 molecules would be generated, each half-old, half-new, whose specific purine- 

 pyrimidine base pair sequence is identical to that of the parent molecule. A 

 genetic mutation, from this point of view, would then be a rare copy error in 

 the replication process by which a nucleotide carrying an incorrect base is 

 introduced into the replica nucleotide chain, thus producing a change in the 

 genetic information. Even though a number of modifications of this replication 

 scheme of Watson and Crick were subsequently proposed (45), later experi- 

 ments have shown that in the replication of bacterial DNA the distribution of 

 the atoms of the parental molecules appears to proceed by the semi-conserva- 

 tive route (122), a central feature inherent in the Watson-Crick scheme. 



One of the first successful attempts to bridge the gap between chemistry 

 and genetics was made by Benzer (12) in his first paper of this collection. 

 Benzer discovered a method for scoring very rare recombinant viruses appear- 

 ing in phage crosses between parents bearing extremely closely linked mutant 

 loci. This allowed him to construct a jine structure map of a large collection 

 of mutants situated in a very restricted region of the phage genome. In 

 consequence of this work, the concept of the gene, traditionally regarded as 

 the unit of recombination, mutation and function, became clarified. For 

 Benzer showed that these three aspects of the genetic material are opera- 

 tionally separable and hence cannot share a common unit. Translated into 

 molecular terms, the unit of recombination appears to represent one, or a few, 

 nucleotide pairs along the DNA molecule, whereas the unit of mutation can 

 be of variable length, ranging from the alteration of a single nucleotide pair, 

 in case of a point mutation, to long-span alterations of the phage genome, 

 covering hundreds or thousands of nucleotide pairs. Finally, the unit of 

 function, or cistron, assumed to determine the specific chemical structure of an 

 enzyme protein, or more precisely, of a polypeptide chain, is of the order of 

 1000 nucleotides in length (13). 



After Benzer had arranged his set of closely linked spontaneous phage 

 mutants into a linear linkage map, it became obvious that there exists a great 

 variability in mutability of different genetic sites within a single functional 

 group, or cistron, since at some loci, or "hot spots," spontaneous mutations recur 

 with much greater frequency than at other, nearby loci. This differential 

 mutability of individual genetic sites very probably reflects the chemical 

 structure of the hereditary molecule corresponding to each locus; e.g. the 

 chance of making a spontaneous copy error at a given site in the course of 

 viral DNA replication might depend on which particular sequence of purine 

 and pyrimidine residues obtains there and which particular base substitution 

 will produce the mutant genotype in question (13). Benzer and Freese (14), 

 therefore, examined also distribution, or mutational spectrum, of mutants 

 induced by the action of chemical mutagens, in particular by replacement of 

 thymine by its analog 5-bromouracil in the viral DNA, which replacement, as 



