It is assumed in most theories of excitation (e. g. Hill 1936
a
) that, when a critical state is reached at a point of the nerve, “excitation” occurs and then automatically propagates over the whole length of the nerve fibre. Since, however, the self-conduction of an impulse involves stimulation of each element by the simultaneous activity of a finite adjacent region of nerve, it is more natural to suppose that initially also a certain
minimal length
of nerve must be excited by an applied shock, in order to give rise to a propagated disturbance. A subthreshold stimulus, therefore, by exciting too small a region, might produce a localized response, the spread and size of which are not enough to excite resting points further on. If such a non-conducted response can actually occur, it might be expected (i) to facilitate excitation by a second shock, if applied to the same region shortly after the first, and (ii) to be accompanied by a small local action potential. Both possibilities can be investigated experimentally, though both might be expected to be complicated by the direct effects (electrotonus,“local potential”) of the first shock.