METHODS The experiments were done on human intercostal muscle obtained at thoracotomy from patients with no evidence of muscular or neuromuscular disease. The specimens were dissected and mounted as previously described (Elmqvist et al. 1964). For stimulation of the local branches of the motor nerve fine bipolar platinum or tungsten electrodes were used. These were insulated with glass to the tips. The stimulating pulses were biphasic, 0-01-0-05 msec in duration, and 2-50V in amplitude. The usual techniques for intracellular recording with KCl-filled glass capillary microelectrodes were employed (Fatt & Katz, 1951). The micro-electrodes used were 3-8 MQ in resistance measured in 3M KCI. Potentials were recorded both by photographing oscilloscope traces and on paper using the Mingograf 81, which is an ink-writer with a flat frequency response up to 500 c/s. In most experiments a dc recording at low amplification and ac records (time constant 0 5 sec) at three higher amplifications were registered simultaneously.

The bathing fluid had the following composition (mM): NaCl, 135; NaHCO3, 15; Na2HPO0, 1-0; KC1, 5 0; CaCl2, 2-0; MgCl2, 1-0; glucose, 11-0; and was bubbled with 95% 02 and 5% C02 with a resulting pH of 7-2-7-4. Choline 5x 10-6 g/ml.(3 5 x 10-5M) was always added to the solution in orderto avoid change in quantum size due to choline deficiency (Elmqvist and Quastel, 1965). Tubocurarinewas used to block neuromuscular transmission. Changes in magnesium and calcium concentrationwere accompanied by appropriate alterations of sodium concentration to maintain iso-osmolarity. The preparations were always allowed at least l hr to equilibrate with the bathing solution. Experiments were performed at a temperature of 320 C. Calculations. The size of all end-plate potentials (epps) were corrected for non-linearity of the end-plate response and the resting potential (RP) of the muscle fibre according to the following formula: