JOSHUA LEDERBERG 

 mutagenic treatments; but if there is any specificity, it has so far been of a 

 second order. In this respect, induced resemble spontaneous mutations, whence 

 some genes may mutate more frequently than others, but not in such a way 

 that the environment can be said to direct the mutation of a specific gene pre- 

 ferentially to the exclusion of others. 



Spontaneous mutation. — We may return at this point to the mechanism of 

 "spontaneous mutation," keeping in mind that the study of experimentally 

 controlled variables on mutation erases the distinction between spontaneous 

 and induced. So long as the concept connoted by these terms is kept clearly 

 in mind as one distinct from that of directed mutation, there need be no 

 confusion. 



Evidence bearing on the relationship between growth and mutation is 

 especially paradoxical, for mutations to phage resistance, for instance, ap- 

 parently do not accumulate in a resting culture ( i ) . On the other hand, cultures 

 whose growth is regulated in a special steady-state, controlled-flow culture 

 vessel ( "chemostat" ) mutate at nearly constant rates per unit time, whether 

 the cells are proliferating slowly or rapidly (106). One interpretation is that 

 spontaneous mutations are due not to intramolecular accidents or reproductive 

 errors, but rather to the action of intracellular chemical mutagens formed by 

 metabolic processes. In this connection it is worth noting that formaldehyde 

 and hydrogen peroxide are both fairly common metabolic intermediates, and 

 that two other mutagens, caffeine and allyl isothiocyanate (mustard oil), are 

 also natural products. Information on the effect of temperature changes on 

 mutation rate would be especially valuable if it could be dissected from effects 

 on growth or metabolism. Under conditions of steady growth a temperature 

 increment of io° accelerated mutations in E. coli by a factor of about 2, and a 

 similar increment is reported in other systems (85) . 



Bacterial populations. — The necessity for thinking of bacterial cultures 

 always in terms of populations, which may have genotypically diverse com- 

 ponents, can scarcely be over-emphasized. The process reviewed to this point, 

 mutation, is the fundamental source of genetic variation, but in view of the 

 smallness of spontaneous mutation rates, it is obvious that the occasional 

 change of a cell from one genetic condition to another can make little impression 

 upon the composition of bacterial populations. The forces that determine which 

 genetic types will predominate in bacterial cultures are the subject of population 

 dynamics. In diploid sexual organisms, population genetics is greatly compli- 

 cated by recombination and by the concealment of genetic variation in the 

 heterozygous condition, so that the most drastic culling may have to be carried 

 out for a great many generations to have a marked effect on the relative 

 frequency of different gene forms. Selection in bacteria is, as a rule, more 

 straightforward, as shown, for example, by the quantitative isolation of phage- 

 resistant mutants by a single application of the virus. The physiological inter- 

 actions of bacteria in dense cultures lead to less trivial problems in population 

 dynamics. Such interaction may involve an obvious competition for nutrients, 



