A strain of
Bacterium lactis aerogenes
giving normal growth when inoculated from bouillon into a standard glucose-phosphate-ammonium sulphate medium yields many long snake-like forms when inoculated into a similar medium with a much lower glucose concentration. The size distribution in the cultures giving the snake-forms is quite different from normal, and is represented by the equation
n
l
=
n
exp ( —
l/l̄
), where
n
l
is the number of cells of length greater than
l, n
is the total number and
l̄
is the mean length. This holds well over most of the range but does not take into account an occasional excess, in the later stages of growth, of exceptionally long cells. This formula suggests that we are dealing with a condition where the cells elongate, but where division is delayed, and depends upon a favourable conjunction of certain independent events in the cell. (If the probability of division becomes
too
small, this law itself will break down.) During the growth cycle
l̄
passes through a maximum and then decreases, most of the snake-forms disappearing again—though some occasionally persist in excess of expectation: these may be forms of low viability. A size coefficient, σ, is defined which gives a good representation of the abnormality of the appearance of the culture under the microscope, and serves to characterize the distribution of lengths. σ and
l̄
decrease as the osmotic pressure of the medium is increased by the addition of salts or of erythrite. With successive passages of a culture through the ammonium sulphate medium the power to give the snake-forms shows a regular decline and is finally lost. It does not appear to be easily restored by several passages through bouillon. The tendency to give snake-forms is enhanced by one preliminary passage through a medium containing asparagine, but growth in an asparagine medium gives normal forms. Experiments on the effect of centrifuged medium from old cultures, of inoculum size and age, and of glucose concentration on σ and on the lag, taken in conjunction with previous work on the lag phase of
Bact. lactis aerogenes
, lead to the following hypothesis: Two separate factors
L
and
D
, one of which,
L
, is diffusible into the medium, the other,
D
, being probably retained by the cells, are responsible respectively for elongation and for division.
D
may be consumed or diluted in the process of division: and its formation may be accelerated or impeded by the presence of other substances in the medium. If the cells are transferred to a new medium, the rates of formation of
L
and
D
may be out of balance and snake-forms appear. Successive passages restore the balance by a process of ‘training’. The mechanism of the training is discussed in the light of a hypothesis in which a crude model of enzyme synthesis having certain analogies with crystal growth is used.
Analysis of the Mechanism by Which Glucose Inhibits Maltose Induction of MAL Gene Expression in Saccharomyces