F. A. Dodge

1. The quantitative dependence of transmitter release on external calcium concentration has been studied at the frog neuromuscular junction, using intracellular recording and taking the amplitude of the end-plate potential (e.p.p.) as an index of the number of packets released.2. The relation between [Ca] and the e.p.p. is highly non-linear. The initial part of this relation on double logarithmic co-ordinates gives a straight line with a slope of nearly four (mean 3.78 +/- 0.2 S.D. in 28 experiments). Addition of a constant amount of Mg reduces the e.p.p. without altering the slope of the log e.p.p./log Ca relation.3. The slope of this logarithmic relation diminishes as [Ca] is raised towards the normal level.4. The results are explained quantitatively on the hypothesis that Ca ions combine with a specific site X on the nerve terminal forming CaX, and that the number of packets of acetylcholine released is proportional to the fourth power of [CaX].5. The analysis suggests that a co-operative action of about four calcium ions is necessary for the release of each quantal packet of transmitter by the nerve impulse.

The oscillatory behavior of the cephalopod giant axons in response to an applied current has been established by previous investigators. In the study reported here the relationship between the familiar "RC" electrotonic response and the oscillatory behavior is examined experimentally and shown to be dependent on the membrane potential. Computations based on the three-current system which was inferred from electrical measurements by Hodgkin and Huxley yield subthreshold responses in good agreement with experimental data. The point which is developed explicitly is that since the three currents, in general, have nonzero resting values and two currents, the "Na" system and the "K" system, are controlled by voltage-dependent time-variant conductances, the subthreshold behavior of the squid axon in the small-signal range can be looked upon as arising from phenomenological inductance or capacitance. The total phenomenological impedance as a function of membrane potential is derived by linearizing the empirically fitted equations which describe the time-variant conductances. At the resting potential the impedance consists of three structures in parallel, namely, two series RL elements and one series RC element. The true membrane capacitance acts in parallel with the phenomenological elements, to give a total impedance which is, in effect, a parallel R, L, C system with a "natural frequency" of oscillation. At relatively hyperpolarized levels the impedance "degenerates" to an RC system.

Intracellular recordings from Limulus eccentric cells suggest that the generator potential arises from the superposition of numerous discrete fluctuations in membrane conductance. If this is so, a relation between frequency response to flickering light and noise characteristics under steady light may be predicted. This prediction is verified experimentally. If a discrete fluctuation model is assumed, the data indicate that increased light has two major effects: (i) the discrete events are strongly light-adapted to smaller size, and (ii) the time course of each event becomes briefer.