Spontaneous mutations in bacteria are generally reversible. Consequently, in a medium where prototrophy confers no selective advantage, various auxotrophs (and other mutant types) should acclumulate until for each type the number is such that the lossby mutation to prototrophy should equal the gain by mutation to the corresponding auxotrophic condition. That is to say, mutational equilibria should be established.Nevertheless it is common experience that prototrophic bacterial populations continuously maintained on complex media remain essentially prototrophic and contain at any time a very small proportion of auxotrophs. Conversely, under the same conditions, populations started from an auxotroph remain essentially auxotrophic and contain at any time a very small proportion of prototrophs. The forward and reverse mutation rates for any pair of alleles can lead to only one equilibrium, yet everyday experience shows us that two stable equilibria exist for each-known mutant in bacteria. We may conclude therefore that for each mutant at least one of the equilibrium ratios cannot be attributed to a balance of mutation rates.The above considerations serve to emphasize the ubiquitous role of selection in controlling the stability of bacterial populations which would otherwise become extremely heterogeneous as a result of mutation pressure. Perhaps it is not too surprising then, that we have found as a result of experiments to be described below that there exists in Escherichia coli a mechanism for stabilization of the major component of the population by the simultaneous suppression of all mutants.Materials and Methods.-The four initial stocks of E. coli (strain 15) used in these experiments were histidine requiring, lactose non-fermenting (h-lac-); histidine independent, lactose non-fermenting (h+ lac-); histidine requiring, lactose fermenting (h-lac+); and histidine independent, lactose fermenting (h+ lac+). Stocks were kept on nutrient agar slants transferred every three months, grown for 36 hours at 370C., then kept refrigerated between transfers. The main experimental procedure was the serial transfer of populations comprised of various initial ratios of the above stocks through Gray and Tatum synthetic medium containing 25y of L-histidine monohydrochloride per cc. (H+). The growth rate and final level of growth of h-and h+ are indistinguishable on H+.' The serial transfer experiments (STs) consisted in transferring 0.5 cc. of cultures in stationary phase to 50 cc. of the above medium in a 125-cc. Erlenmeyer

A bacterial culture may contain a very large number of organisms. Consequently, even mutations occurring at very low rates per organism may not be uncommon. For this reason bacterial cultures are likely to be heterogeneous and their composition will depend upon mutation rates and selection pressure (Ryan, 1948). When mutations influence the growth characteristics of bacteria, the behavior of the whole culture may be drastically influenced. In the case of the so-called biochemical mutations, which affect the organisms' capacity to synthesize amino acids and growth factors, this influence is very evident. For example, by overgrowth during serial transfers in the presence of limiting concentrations of tryptophan, Wright and Skeggs (1945) have been able to select a strain of Lactobacillu8 arabino&u that does not require this substance. In the absence of tryptophan this new strain grows as fast and as far as does its tryptophan-dependent parent in the presence of tryptophan. Increasing concentrations of tryptophan, however, first depress and then allow this rapid growth of the tryptophan-independent mutant. Although the factors underlying this behavior are not yet understood in L. arabinosus, we have found a parallel situation in a histidineless mutant of Escherichia coli. This paper will present data that show that the depression of growth on intermediate concentrations of amino acid is the result of mutation and the consequent interaction between mutants and parents. The next two papers in this series will analyze the mechanism of this interaction (Ryan and Schneider, 1949a,b).

No abstract

The residue was cooled, extracted with four 100-ml. portions of chloroform, the organic layer dried with drierite and the solvent removed by distillation on the steam-bath. The solvent-free residue was distilled in vacuo, b. p. 144°( 0.06 mm.) in an all-glass-interjoint apparatus to give a light yellow oil, which could not be induced to crystallize; yield, 60%.The base was identified as the dipicrate, which was prepared in anhydrous diethyl ether and recrystallized several times from a large volume of methanol, m. p. 168-169°( cor.).

space, the dual being a discrete one. However, for infinite duration the dualism is a self-dualism iti the following sense.A Wiener differential process on the half line 0 < x < o is its own cosine and sine transform, as a process.Finally from our approach we are able to analyze Khintchine's stationary processes very systematically, although none of the results can be strictly new, since these processes are more or less included in the theory of Hilbert space.A full account will appear in another journal.Some of the variations which occur during the growth of bacteria are due to the selection of types which arise at random by mutation. In the case of resistance to phagel and radiations2 in Escherichia coli and penicillin resistance in Staphylococcus aureus3 these changes behave like genic mutations in that there is a finite probability for each individual bacterium to undergo an hereditary change in the course of its lifetime. These changes are not induced by the unfavorable agents but occur in their absence. Phage radiation and penicillin simply select those resistant organisms which arise at random. With respect to nutritional requirements the demonstration of the genic nature of mutation is not so clear. Roepke, et al.,4 found that mutants of E. coli with specific growth factor requirements appeared after x-radiation. However, they were unable to attribute the changes to the x-radiation because of the incidence of spontaneous mutations. This was also the experience of Gray and Tatum6 with E. coli but these authors showed that growth factor deficiencies were induced in Acetobacter melanogenum by x-radiation. Later, when more material was available, Tatum6 was able to show that x-radiation significantly raised the frequency of nutritional mutants in E. coli. By analogy with the production of biochemical mutations in Neurospora,7 a sexual organism, the latter authors suggest that biosyntheses in bacteria are controlled by specific genes. We have been able to show that a strain of Clostridium septicum mutates to a condition where it is no longer able to synthesize the pyrimidine, uracil, but requires it in the medium for growth. Conversely, the uracil-dependent strain mutates to a state where uracil is VOL. 32, 1946 281