V. MOLECULAR MECHANISM OF MUTATIONS 261 



The directional effect of these mutagens can be observed when the 

 rates of induced r(>version are deterniined for many transition nuitants. 

 If one obtains a bii^artition into highly and weakly inducible mutants 

 and if the results with different mutagens agree with the expectations, it 

 seems likely that the hypothesis about the direction of base pair changes 

 is correct. 



It should be remembered, however, that the frequency of revertants is deter- 

 mined by a selective technique. The mutagenic chemical alters only one base 

 in a pair, and the resulting hybrid may be able to act functionally hke a rever- 

 tant either immediately or only after one or two DNA duplications. Hence the 

 result of measurements on induced reversions depends on the number of DNA 

 duplications which the selective method allow^s after the first base has been 

 altered. This functional expression may depend, furthermore, on the particular 

 mutant (which of the two DNA strands is attacked, the functional or the other 

 one) and on the mutagen (how much the altered base differs functionally from 

 a normal one). The rate of reversion induction has, therefore, been measured 

 in two ways, at least for nitrous acid (E. B. Freese and E. Freese, 1961) and 

 hydroxylamine (Freese et al., 1961b). On the one hand, the chemically treated 

 T4-r// phages were plated directly on the selective bacteria K (rate a//3). On 

 the other hand, they were preadsorbed onto (UV-killed) bacteria B in which all 

 live phages can replicate about five times; the infected bacteria were then plated 

 on the selective bacteria K (rate a'/fi). The results in Table V show that for 

 HA plating with preadsorption merely enhances the values found for direct 

 plating, while for nitrous acid some rates, which were smaller than others for 

 direct plating, were much larger after preadsorption. The cause of these differ- 

 ences has yet to be resolved. 



The rates a/fS are quite different when the same mutant is induced to 

 revert by different mutagens. This reflects the varying degree of the 

 mutagenic versus the lethal effect. The lethal effect is especially small 

 for HA. If the lethal effect is large, as for EES, the difference between 

 highly and weakly inducible mutants becomes much less pronounced. 



Different mutants show greatly different rates of reversion induci- 

 bility by a given mutagen. With HA as mutagen a clear bipartition has 

 been observed into highly and weakly (or not at all) inducible mutants; 

 the a/(i values of these two classes differ by more than a factor 100 (see 

 Table V). It should be mentioned that one more AP mutant (AP 12) 

 has been found highly inducible but has not been included in Table V 

 for reasons given later. The same AP mutants that are most inducible 

 by HA are also most inducible by all other agents used (though for 

 nitrous acid only after preadsorption). But with these other agents the 

 bipartition is not as clear-cut. It is improbable that the higher nmta- 

 genic specificity can be responsible for the observed bipartition since 

 HA (and nitrous acid) attack C while low pH and EES attack G and 

 yet the results agree. Thus, most of the rarely or non-inducible mutants 

 have an A-T pair at their mutant site since they have been induced by 



