We investigated the mechanism whereby ions cross dihydropyridine-sensitive (L-type) Ca channels in guinea pig ventricular myocytes. At the single-channel level, we found no evidence of an anomalous mole-fraction effect like that reported previously for whole-cell currents in mixtures of Ba and Ca. With the total concentration of Ba + Ca kept constant at 10 (or 110) mM, neither conductance nor absolute unitary current exhibits a paradoxical decrease when Ba and Ca are mixed, thereby weakening the evidence for a multi-ion permeation scheme. We therefore sought independent evidence to support or reject the multi-ion nature of the L-type Ca channel by measuring conductance at various permeant ion concentrations. Contrary to the predictions of models with only one binding site in the permeation pathway, single-channel conductance does not follow Michaelis-Menten kinetics as Ba activity is increased over three orders of magnitude. Two-fold variation in the Debye length of permeant ion solutions has little effect on conductance, making it unlikely that local surface charge effects could account for these results. Instead, the marked deviation from Michaelis-Menten behavior was best explained by supposing that the permeation pathway contains three or more binding sites that can be occupied simultaneously. The presence of three sites helps explain both a continued rise in conductance as [Ba2+] is increased above 110 mM, and the high single-channel conductance (approximately 7 pS) with 1 mM [Ba2+] as the charge carrier; the latter feature enables the L-type channel to carry surprisingly large currents at physiological divalent cation concentrations. Thus, despite the absence of an anomalous mole-fraction effect between Ba and Ca, we suggest that the L-type Ca channel in heart cells supports ion flux by a single-file, multi-ion permeation mechanism.