Study of the excitatory sodium current (INa) intact heart muscle has been hampered by the limitations of voltage clamp methods in multicellular preparations that result from the presence of large series resistance and from extracellular ion accumulation and depletion. To minimize these problems we voltage clamped and internally perfused freshly isolated canine cardiac Purkinje cells using a large bore (25-microns diam) double-barreled flow-through glass suction pipette. Control of [Na+]i was demonstrated by the agreement of measured INa reversal potentials with the predictions of the Nernst relation. Series resistance measured by an independent microelectrode was comparable to values obtained in voltage clamp studies of squid axons (less than 3.0 omega-cm2). The rapid capacity transient decays (tau c less than 15 microseconds) and small deviations of membrane potential (less than 4 mV at peak INa) achieved in these experiments represent good conditions for the study of INa. We studied INa in 26 cells (temperature range 13 degrees-24 degrees C) with 120 or 45 mM [Na+]o and 15 mM [Na+]i. Time to peak INa at 18 degrees C ranged from 1.0 ms (-40 mV) to less than 250 microseconds (+ 40 mV), and INa decayed with a time course best described by two time constants in the voltage range -60 to -10 mV. Normalized peak INa in eight cells at 18 degrees C was 2.0 +/- 0.2 mA/cm2 with [Na+]o 45 mM and 4.1 +/- 0.6 mA/cm2 with [Na+]o 120 mM. These large peak current measurements require a high density of Na+ channels. It is estimated that 67 +/- 6 channels/micron 2 are open at peak INa, and from integrated INa as many as 260 Na+ channels/micron2 are available for opening in canine cardiac Purkinje cells.