The adaptation of Bact . coli mutabile to lactose

When Bact. coli mutabile not previously exposed to lactose is plated on lactose-ammonium sulphate agar the number of normal sized colonies $(\text{lac}^{+})$ eventually formed is a complicated function of the inoculum size. For small numbers all the cells plated eventually form colonies; for large numbers the colony yield is determined not by a number of mutants in the parent culture but by plate exhaustion (for which the earlier developing colonies are chiefly responsible). The time of appearance of the $\text{lac}^{+}$ colonies is much longer than with a culture previously grown in lactose. Thus $\text{lac}^{+}$ mutants could not have been present from the start unless their growth is inhibited by an excess of normal cells. When, however, a small number of previously adapted cells are mixed with an excess of unadapted cells the presence of the latter does not impede the development on agar of $\text{lac}^{+}$ colonies from the former. When cells are first placed in a liquid lactose medium and samples are transferred at intervals during the ensuing lag phase, the time needed for colony development on lactose-agar progressively diminishes, once again showing that an adaptive process is occurring during the lag in the liquid medium. In certain special circumstances the adaptation to the liquid lactose medium may occur with abnormal speed. The growth rate of newly-adapted strains is at first variable. If interpreted by a mutation theory the observations would demand the assumption of a complex polygenetic system for which current applications of the Luria-Delbruck and Lea-Coulson theories would be invalid. Recent arguments about the mutational nature of these phenomena are criticized in the light of the new evidence.