THE SYNAPTIC MECHANISM FOR POSTSYNAPTIC INHIBITION 81 



with diameters smaller than 1-32 times K+ in the hydrated state, and were 

 impermeable to fourteen species of anion with diameters larger than 1-35 

 times K+. There was only one slight discrepancy from a size determination: 

 the formate ion exhibited permeability, while the slightly smaller bromate ion 

 (1-32 as against 1-35 times K+) did not. Hence the simplest assumption is that 

 the inhibitory transmitter has converted the specific inhibitory patches to a 

 sieve-like membrane having pores of a precisely standardized size (cf. Coombs 

 etai, 1955b; Eccles, 1957). 



Chloride is the only permeable anion that normally exists in a concentration 

 sufficient to contribute appreciably to the inhibitory ionic current. But, if 

 the i.p.s.p. is produced solely by the net movement of Ch ions down their 

 electrochemical gradient, the equilibrium potential for Ch ions (£^,,) must be 

 at from 5 to 10 mV more polarization than the normal resting membrane 

 potential. This value for E^,, could be maintained only if there were a chloride 

 pump (cf. Boistel and Fatt, 1958). With those nerve or muscle fibres where 

 accurate investigation has been possible, it has not been necessary to pos- 

 tulate a Ch pump (Hodgkin, 1958; Hodgkin and Horowicz, 1959). Particular 

 attention should therefore be given to the possible role of K+ ions in contri- 

 buting to the generation of the i.p.s.p., for there is independent evidence that 

 the equiUbrium potential for K+ ions {E^) is maintained at 20 mV (or even 

 more) polarization above resting membrane potential of motoneurons, for 

 K+ appears to be the only ion that contributes appreciably to the after- 

 hyperpolarization following an SD spike potential, which has an equilibrium 

 potential of from —90 to 100 mV (Coombs ef al., 1955a; Eccles, Eccles and 

 Ito, unpubhshed observations). 



In the original investigation of the postulate that the net flux of K+ ions 

 contributes substantially to inhibitory current. Coombs et al. (1955b) com- 

 pared the eff'ects of passing depolarizing currents out of intracellular micro- 

 electrodes that were filled either with Na2S04 or K2SO4. It was assumed that 

 the current was carried out of the microelectrode largely by the highly con- 

 centrated cations therein, Na+ or K+ as the case may be, and that it was 

 passed across the cell membrane partly by an outward flux of cations (largely 

 K+) and partly by an inward flux of anions (largely CI ). Thus an injection 

 out of a K2S04-filled electrode would add K+ plus Ch ions to the cell; and 

 after cessation of the current the normal composition of the cell would be 

 recovered by water moving in to restore osmotic equilibrium and by K+ 

 and CI' ions diffusing out. On the other hand, after Na+ injection out of the 

 Na2S04 electrode, there would be depletion of the intracellular K+ and 

 replacement by Na+, but also much the same increase in Cb as with the K+ 

 injection. It was assumed that in this case also the CI" ions would quickly 

 attain equilibrium by diffusion across the cell membrane, a process that is 

 almost complete in 1 min. However, as seen in Fig. 8b, there was a very large 

 change in the ^'i.p.s.p. (from —80 mV to —35 mV), and it took many minutes 



7 



