Multifocal motor neuropathy (MMN) is associated with serum autoantibodies to gangliosides, but their pathogenic role is uncertain. We have used a novel approach to study the effects of serum and plasma from 8 patients with this syndrome, 6 of whom were anti-GM1 positive. The nerve stimulus required to evoke muscle contraction and endplate potentials (EPPs) was measured in the mouse phrenic nerve-diaphragm preparation during 4 to 6 hours of direct application (plasma at 1:1 or serum 1:2 dilution) and following intraperitoneal injection of plasma (1 ml/day) for 1 to 5 days ("passive transfer"). Direct application of MMN serum or plasma produced a progressive increase in stimulus threshold, followed by complete block of nerve-evoked muscle contraction in 3 cases, and an associated decline to about 50% of the EPP amplitude followed by sudden loss of EPPs. These effects were complement independent. Even with complete block of nerve-evoked EPPs, miniature EPP (MEPP) frequency could be increased by raising external K+ to depolarize the nerve terminal directly. Passive transfer of 1 ml of MMN plasma (n = 5) for 3 days caused similar but less marked changes. These results demonstrate that serum factors in MMN can block nerve conduction at distal motor nerves.