THE SYNAPTIC MECHANISM FOR POSTSYNAPTIC INHIBITION 83 



quence the diffusion of CI" outward would be greatly impeded. Cl~ could only 

 be restored to its initial electrochemical potential within the neuron when the 

 excess Na+ had been pumped out and replaced by K+. Thus, even if Cl~ 

 were the only ion that moved freely across the activated inhibitory patches 

 of the postsynaptic membrane, there should be a much slower recovery of 

 the £'i.p.s.p. after Na+ than after K+ injection. Results such as those of 

 Fig. 8 do not require the further assumption that 1C+ ions also move freely 

 across the activated inhibitory patches. 



Yet, if K+ flux does not play a considerable part in the generation of the 

 postsynaptic inhibitory current, it seems impossible to explain how the 

 ■^i.p.s.p. can be at about 5 to 10 mV more hyperpolarization than the resting 

 potential (Fig. 3) without requiring that ^c, should be maintained at this 

 level by a Ch pump (cf. Boistel and Fatt, 1958). No other ion species could 

 substitute for Cl~, for no other permeable anion would normally be in a 

 concentration to make any appreciable contribution to the inhibitory current. 

 Hence, despite the enigmatic nature of the evidence from cation injection into 

 motoneurons, it seems hkely that the postsynaptic inhibitory patches are 

 permeable to K+ as well as to Cl~. In the light of our experimental evidence 

 K+ permeabiUty must be much lower than CI", approximate values being 

 about 20% and 80%, respectively, of the total, as indicated by the relative 

 lengths of the arrows in Fig. 8f (Eccles, Eccles and Ito, unpubhshed observa- 

 tions). 



Investigations on synaptic inhibitory actions at peripheral junctions have 

 indicated that K+ and CI" are likewise the only two ions that are effectively 

 concerned in the postsynaptic inhibitory action (cf. Eccles, 1959). Several 

 examples can be cited: Tauc (1958) has concluded that the i.p.s.p. of the giant 

 cells of Aplysia results from an increased permeability to both K+ and CI"; 

 the inhibitory action on crustacean muscle fibres appears to be produced 

 almost entirely by an increase in CI" permeability (Boistel and Fatt, 1958); 

 the inhibitory action on crustacean stretch receptor cells is due almost 

 entirely to an increase in K+ permeability with Ci" ions playing a subordinate 

 role at most (Kuffler and Edwards, 1958; Edwards and Hagiwara, 1959); 

 finally, vagal inhibition of the heart is similarly due to a large increase in K+ 

 permeability with little if any increase for CI (Burgen and Terroux, 1953; 

 Trautwein and Dudel, 1958). 



Thus these diverse types of inhibitory action are all effected by an increased 

 permeability to either K+ or CI" or to their combination in varying degree. 

 The initial hypothesis has been that the inhibitory postsynaptic membrane 

 functions as a sieve, being permeable to all ions below a critical size in the 

 hydrated state. If, as suggested by Boistel and Fatt (1958), it be further 

 assumed that the pores are charged positively as in Fig. 9b, the membrane 

 would exhibit a selective preference for small anions as against small cations, 

 as occurs with crustacean muscle and mammahan motoneurons. If, on the 



