The effects of the phototoxic K+- channel blockers 8-methoxypsoralen (8-MOP) and 5-methoxypsoralen (5-MOP) on Ranvier nodes were compared to those of 5,8-diethoxypsoralen (5,8-EOP) by means of the Hodgkin-Huxley formalism. When these test substances were added individually to the bathing solution (8-MOP: 100 micromol/l; 5-MOP: 50 micromol/l; 5,8-EOP: 10 micromol/l) the following completely reversible effects were observed: 1. 8-MOP, caused a nearly potential-independent decrease of the sodium permeability, P'Na, by ca. 17%. 5-MOP and 5,8-EOP merely decreased the maximal value of P'Na, by ca. 12 and 8% respectively, whereas with weak depolarisations P'Na was unchanged. 2. In the tested potential range the potassium permeability, P'K, was caused to decrease by ca. 9% by 8-MOP, ca. 21% by 5-MOP and ca. 19% by 5,8-EOP. 3. The potassium currents acquired a phasic time course previously described for 8-MOP and 5-MOP. They reached a relative maximum and approached a lower steady-state value, kinfinity, with a time constant tauk at V = 120 mV of about 16 ms (8-MOP), 20 ms (5-MOP) and 94 ms (5,8-EOP). To obtain dose-response relations the drug-induced effects on peak P'K and on the steady state value, kinfinity, were measured. The corresponding apparent dissociation constants (in micromol/l) were 66.6 and 80.1 (for 8-MOP), 87.6 and 25.8 (for 5-MOP), and 13.5 and 6.5 (for 5,8-EOP). In view of the similarity of the actions of 5-MOP and 5,8-EOP as well as the fact that 5,8-EOP is not phototoxic, in future 5,8-EOP may well prove to be a particularly suitable K+-channel blocker for the symptomatic therapy of multiple sclerosis and other demyelinating diseases.