146
B. M. SHAFFER
tive. Or streams may form but be limited in extent because fresh stream
cells cannot be supplied faster than the existing ones move inwards.
But there are conditions in which a secretion front can be relayed indefinitely : a large part of a culture plate may be covered by a single
stippled aggregation; and when a plate has been inoculated at one side,
continuous streams may grow for several days, as they capture the
cells left behind the advancing feeding front. What can limit the size
of an aggregation in cultures in which the density and state of the cells
favour the chemotactic relays? Apart from fragmentation of wellestablished streams, the most important factor is presumably the rate
at which triggers become active, because the earlier ones tend to prevent
others appearing later. We might initially assume that cells were competitively debarred from developing into triggers when they became
stream cells. This would necessarily involve the relay mechanism in
defining the boundaries. But actually some diffusible product of an
aggregation can prevent cells from becoming founders without turning
them into stream cells, at least in P . violaceum (Section III, C). And the
fact that mean aggregation area in homogeneous cultures is remarkably
constant over a wide range of cell densities (Section III, B) suggests that
it is determined by a trigger inhibitor that is made by established triggers and spreads out from them by nonrelayed diffusion. Its production
might be permanently confined to these cells; alternatively, if all the
triggers developed before the first fronts were propagated, it might be a
factor later made by the peripheral cells. Paradoxically, this could even
be the inducer of attractor secretion if this actually inhibited trigger
differentiation when present in a concentration or gradient too small to
start the relays; though it hardly seems possible that, during a preliminary phase of activity, a trigger could secrete an amount of inducer
unable to excite even the nearest relays and yet able to inhibit over the
requisite distance—in D. discoideum (Section III, B), an order of magnitude greater than the maximum direct range of a mature centre
(Bonner, 1947). But it may be that the relays after all are involved in
inhibition, and that the aggregation area may remain constant with
increasing density because any tendency for more triggers to become
active in a given area per unit time is balanced by an increase in the
secretion of inhibitor and in the velocity, and where repetitive the
frequency, of the secretion fronts. Further analysis of inhibition in P .
violaceum should prove very informative.
A final question is whether all cells would eventually show trigger
activity if not inhibited. At present all we can say is that the probability
of a cell's doing so depends on the species and the environment, both
cellular and noncellular. In P . violaceum, it has yet to be shown that a
single, isolated cell will differentiate like a founder, however long it is
B. M. SHAFFER
tive. Or streams may form but be limited in extent because fresh stream
cells cannot be supplied faster than the existing ones move inwards.
But there are conditions in which a secretion front can be relayed indefinitely : a large part of a culture plate may be covered by a single
stippled aggregation; and when a plate has been inoculated at one side,
continuous streams may grow for several days, as they capture the
cells left behind the advancing feeding front. What can limit the size
of an aggregation in cultures in which the density and state of the cells
favour the chemotactic relays? Apart from fragmentation of wellestablished streams, the most important factor is presumably the rate
at which triggers become active, because the earlier ones tend to prevent
others appearing later. We might initially assume that cells were competitively debarred from developing into triggers when they became
stream cells. This would necessarily involve the relay mechanism in
defining the boundaries. But actually some diffusible product of an
aggregation can prevent cells from becoming founders without turning
them into stream cells, at least in P . violaceum (Section III, C). And the
fact that mean aggregation area in homogeneous cultures is remarkably
constant over a wide range of cell densities (Section III, B) suggests that
it is determined by a trigger inhibitor that is made by established triggers and spreads out from them by nonrelayed diffusion. Its production
might be permanently confined to these cells; alternatively, if all the
triggers developed before the first fronts were propagated, it might be a
factor later made by the peripheral cells. Paradoxically, this could even
be the inducer of attractor secretion if this actually inhibited trigger
differentiation when present in a concentration or gradient too small to
start the relays; though it hardly seems possible that, during a preliminary phase of activity, a trigger could secrete an amount of inducer
unable to excite even the nearest relays and yet able to inhibit over the
requisite distance—in D. discoideum (Section III, B), an order of magnitude greater than the maximum direct range of a mature centre
(Bonner, 1947). But it may be that the relays after all are involved in
inhibition, and that the aggregation area may remain constant with
increasing density because any tendency for more triggers to become
active in a given area per unit time is balanced by an increase in the
secretion of inhibitor and in the velocity, and where repetitive the
frequency, of the secretion fronts. Further analysis of inhibition in P .
violaceum should prove very informative.
A final question is whether all cells would eventually show trigger
activity if not inhibited. At present all we can say is that the probability
of a cell's doing so depends on the species and the environment, both
cellular and noncellular. In P . violaceum, it has yet to be shown that a
single, isolated cell will differentiate like a founder, however long it is