1. The effect of Ba2+ on the evoked secretion of individual acetylcholine (ACh) quanta was studied on frog neuromuscular junctions using conventional electro-physiological techniques. 2. In solutions containing 1.8 mM-Ba2+ and no added Ca2+, 1 Hz stimulation for less than 1 min elevated miniature end-plate potential (m.e.p.p.) frequencies to 5-20 times the control level (seven experiments). Similar results were obtained when a Ca2+-chelating agent was added to the Ba2+ solution. 3. Repetitive nerve stimulation at frequencies greater than 1 Hz in concentrations of Ba2+ greater than or equal to 1.8 mM elevated m.e.p.p. frequencies to unmeasurable levels (greater than 100/sec). Such high m.e.p.p. frequencies were accompanied by a steady depolarization of the post-synaptic membrane, which was used to estimate the number of ACh quanta released. 4. The number of ACh quanta released asynchronously by nerve impulses was directly related to the external concentration of Ba2+ in a non-linear fashion. 5. Ba2+ was two orders of magnitude more effective than Ca2+ in supporting the evoked discharge of m.e.p.p.s. Ca2+ was a potent antagonist of asynchronous release in Ba2+ solutions. 6. Mg2+ and Co2+ both competitively antagonized evoked release in Ba2+ solutions. The equilibrium dissociation constant for each ion as an antagonist of asynchronous, Ba2+-dependent release was similar to its corresponding value as an antagonist of synchronous, Ca2+-mediated release. 7. It is suggested that Ba2+ supports dispersed, quantal ACh release directly by acting through the same conductance pathway normally traversed by Ca2+.