1. The frequency of miniature end-plate potentials (m.e.p.p.s) was recorded from neuromuscular junctions in rat diaphragm phrenic nerve preparations in vitro after preparations had soaked in solutions containing Ca in concentrations between 10(-10) and 10(-2)M and a similar range of [Mg].2. Ethylenediamine tetra-acetate (EDTA) and ethyleneglycol bis (beta-aminoethyl ether) tetra-acetate (EGTA) buffers were added to prepare solutions with [Ca] and [Mg] below 10(-4)M. A computer program was used to estimate the free [Ca(2+)] in these solutions, and it was shown that the effects of Ca could be attributed to the free [Ca(2+)] in the bathing solution.3. M.e.p.p.s could still be detected without difficulty after soaking preparations for 6-8 hr in solutions containing EDTA or EGTA buffers and no added Ca. The basal frequency was unchanged upon exhibition of Ca in concentrations up to 10(-5)M and/or Mg in concentrations up to 10(-3)M.4. Ca in concentrations of and above 10(-4)M accelerated m.e.p.p. frequency from the basal level. This effect reached a maximum in [Ca] of 10 mM and raising the [Ca] above this level did not further change frequency. These effects were explained by the combination of Ca molecules with a nerve terminal receptor site. It was postulated that this combination allosterically activated the spontaneous release mechanism.5. Mg could accelerate m.e.p.p. frequency in the absence of added Ca. The interactions of Ca and Mg upon m.e.p.p. frequency indicated that Ca and Mg competed for the same sites.6. Raising the [H(+)] of the bathing medium accelerated m.e.p.p. frequency. This effect was thought to be exerted partly by combination with the same receptor sites as Ca and Mg and partly by variation of the ionization of the CaCl(2) of the bathing solution.