ACETYLCHOLINE (ACh) is released from stimulated nerve terminals in packets or quanta, each containing roughly 10,000 molecules1,2, but we still do not know how the quanta are formed. A widely accepted hypothesis is that the ACh contained in vesicles found in the nerve terminal is released by exocytosis, while an alternative suggestion is the gated channel hypothesis. This takes into account the fact that the ACh which seems to be in free solution in millimolar concentrations in the terminals4 is probably the source of the ACh which leaks molecule by molecule from the resting terminals (only a small fraction of the spontaneous release is quantal5,6). Other evidence against the vesicle hypothesis is that when nerves are exposed to radioactive precursors, the free ACh is labelled first, and with stimulation the newly synthesised ACh is released preferentially7. The free ACh could also be released in packets if stimulation opens gates on channels through which ACh diffuses from the terminals; with each opening of a channel, about 10,000 transmitter molecules might escape by moving out of the terminal down a concentration gradient. One way to test the gated channel hypothesis is to change abruptly the concentration of free ACh in the terminal, and so change the number of ACh molecules per quantum. This has been achieved in a molluscan, cholinergic neurone by injecting acetylcholinesterase into the cell body. The enzyme diffuses down to the terminals, hydrolyses the free ACh, and produces a transmission block8. Unfortunately, in this preparation it is not possible to determine the quantal size, so it is not certain that the block occurs because the quanta become smaller. Quantal size is readily estimated at the vertebrate neuromuscular junction, but it is not possible to inject substances into the fine nerve terminals and there does not seem to be a feasible way to increase or decrease abruptly the amount of ACh in the terminal. An alternative approach, which I have used here, is to change the concentration of free ACh by moving water in or out of the nerve terminal, using osmotic pressure as the driving force9,10.