We have previously shown that the Lambert-Eaton myasthenic syndrome (L.E.M.S.) is an autoimmune disorder (Lang, Newsom-Davis, Prior & Wray, 1983). IgG autoantibodies interfere with release of acetylcholine from presynaptic nerve terminals of the skeletal neuromuscular junction. The experiments reported here were designed to investigate the mechanism of action of these antibodies at depolarized nerve terminals.Mice were injected with control or L.E.M.S. IgG (30 mg/day, i.P.) for 2 days, and the diaphragm muscle removed on the third day, as previously described (Lang et al. 1983). Miniature end-plate potentials (m.e.p.p.s) were recorded using microelectrodes in Krebs solution (23-26 'C) containing high K+ concentration (15-9 mM) and a range of Ca2+ concentrations (0-25-2-5 mm).In controls, m.e.p.p. frequency increased with Ca2+ concentration from a value of 5 03 + 0 69 s-1 (n = 19 end-plates) at 0-25 mM-Ca2+ concentration to a plateau value of 79-3 + 6-8 s-1 (n = 18) at 1-5 mM-Ca2+ concentration. Further increase in Ca2+ concentration produced no further increase in this plateau value. The curve of m.e.p.p. frequency versus Ca2+ concentration for L.E.M.S. IgG-treated animals was similar in shape to that for controls, but with significantly reduced (P < 0-002) m.e.p.p. frequency values at all Ca2+ concentrations. The extent of these reductions was to 34 46 % of control values. Thus the curve was shifted to the right with a reduced plateau. The plateau value of m.e.p.p. frequency was reached at similar Ca2+ concentrations as for controls but was reduced to 36-4 + 4-9 s51 (n = 19, Ca2+ concentration = 1-5 mM).These reductions in m.e.p.p. frequency at potassium-depolarized nerve terminals by L.E.M.S. IgG indicate that it is unlikely that this antibody acts solely (if at all) on action potential mechanisms, since no action potentials were involved in the present experiments. On the other hand, voltage-dependent Ca2+ channels would be expected to be opened by the depolarization produced under the above experimental conditions. L.E.M.S. IgG may act on these latter channels to cause their loss of function as suggested previously (Lang, Newsom-Davis, Prior & Wray, 1984). The present results are consistent with such an action, and the reduced plateau value of the above curve produced by the L.E.M.S. IgG shows further that the mechanism is not one of competitive antagonism.