The dependence of transmitter release and synaptic effectiveness on the size of a neuron's peripheral field was studied using neuromuscular junctions in sartorius muscles of adult frogs (Rana pipiens). The size of the peripheral field (motor unit size) was reduced by crushing the sartorius nerve and surgically removing half of the muscle fibers. Synapses thus formed were compared with those formed when crushed nerves reinnervated intact whole muscles, as well as with synapses in normal unoperated muscles. Indirect observations suggested that all motor axons participated in reinnervation and that motor unit size was indeed smaller in half-muscles. Synaptic safety margins, as measured by the sensitivity of nerve stimulus-evoked twitching to low Ca2+, were substantially higher in muscles with reduced motor units. These higher safety margins were due to enhanced evoked transmitter release. In Ringer solution containing Mg2+ and lowered Ca2+, total evoked release and evoked release per unit nerve terminal length were approximately 2-fold higher in muscles with reduced motor units, when studied 7 to 18 weeks postoperatively. A similar difference was seen when unblocked release was measured in a normal physiological solution, after blocking excitation-contraction coupling and muscle fiber action potential generation with formamide. Miniature endplate potential frequency in half-muscles was 2 to 3 times higher than in controls when tested in normal physiological solution, but was not significantly different in low Ca2+, Mg2+-containing solution. By 34 weeks postoperatively, there was no longer a difference in evoked release, even though the difference in motor unit size persisted.(ABSTRACT TRUNCATED AT 250 WORDS)