Change of flagellar antigenic phase in the Salmonellas is considered as resulting from the mutation and back mutation.

A rapid method of determining the phase of numerous colonies of Salmonella typhi-murium is described; melted sloppy agar containing antibody for one phase is layered on tplates on which colonies have grown. On brief incubation, colonies of the agglutinated phase produce narrow dense sharp edged zones of opacity, while colonies of the other phase wide diffuse zones of swarming organisms.

A theoretical analysis shows that if the growth rates of the phases are equal the proportion of cells of ‘mutant’ phase in a culture initially all in one phase and maintained in continuous logarithmic growth will (at first) increase linearly with the number of generations at a rate proportional to the mutation rate. In one strain of Salm, typhi-murium experiments showed the rates of growth of the phases to be equal; the proportion of cells of mutant phase in cultures of this strain during long periods of continuous growth has been measured; linear rates of increase were found, and the mutation rates are calculated.

The variance in proportion of cells of mutant phase in replicate cultures has been measured; the findings are correlated with theoretical considerations.

Certain anomalous decrease in proportion of cells of mutant phase during prolonged growth are described, and a possible explanation is proposed.

Eight further strains have been investigated by the same methods; in the nine strains, the rates of mutation from group to specific phase range from 10-4 to 4.7 X 10-3per bacterial generation cycle, and the rates of the reverse mutation from 10-5 to 8.6 X 10-4. In all the strains, including two of variety binns, the rate of mutation from group to specific is higher than the rate of the reverse change; it is shown theoretically that such strains will tend towards an equilibrium composition in which the specific phase will predominate; in one rapidly mutating strain such an equilibrium has been demonstrated.

In a limited experiment, alteration of medium and of temperature of incubation caused to detectable alteration of rate of mutation per bacterial generation.

Stable O mutants, and poorly flagellated mutants which back-mutate to the fully motile form, are described.

The experimental findings are compared with reports in the literature on the phase of Salm. typhi murium as isolated. The behavior of strains of variety binns is discussed, and it is suggested that this description should not now be applied to strains in which a specific phase can be demonstrated. The biological funtion of mutablility of phase is discussed.