1. Whole-cell and patch-clamp techniques (Hamill, Marty, Neher, Sakmann & Sigworth, 1981) have been used to make quantitative measurements of the transient inward sodium current (INa) in single cells from bullfrog atrium. This preparation is particularly suitable for the study of INa: (i) the current density is relatively low, (ii) the cells lack a transverse tubule system, (iii) isolated myocytes can be maintained at reduced temperatures (approximately 8-12 degrees C); therefore kinetics can be studied quantitatively. 2. INa was pharmacologically and kinetically isolated from other transmembrane currents by blocking ICa with CdCl2 (0.2-0.5 mM) or LaCl3 (5 x 10(-6) M), and by using only relatively short voltage-clamp depolarizations which did not activate IK (the delayed rectifier). 3. The voltage dependence of INa in bullfrog atrium is similar to that in amphibian node of Ranvier or fast skeletal muscle. The threshold for activation is approximately -50 mV. The peak of the INa vs. membrane potential relation is near -5 to -10 mV. The reversal potential in 'normal' (115 mM-Na+) Ringer solution is +59.0 mV (S.D. +/- 3.4, n = 10). Reduction of external Na+ concentration to one-third of normal resulted in an approximately -27 mV shift of the reversal potential, close to that expected for a highly Na+-selective conductance. 4. Steady-state inactivation of INa (h infinity), measured with a conventional two-pulse voltage-clamp protocol, spanned the membrane potential range from -90 to -50 mV. The potential dependence of h infinity was well described by a single Boltzmann function with half-inactivation at -71 mV and maximum slope of 6.0 mV. 5. Steady-state activation of INa (m infinity) was determined from fits of INa records to a Hodgkin-Huxley model. The potential dependence of m infinity was fitted to a Boltzmann function with half-activation at -33 mV and maximum slope of 9.5 mV. Thus at temperatures around 10 degrees C there was very little overlap of the m infinity and h infinity curves, and only very small steady-state 'window' currents are predicted. 6. The activation time constant, tau m, had a 'bell-shaped' dependence on membrane potential. The peak value of tau m was about 4.2 ms, at a membrane potential of -35 mV (9 degrees C). 7. The time course of inactivation of INa was consistently better described by the sum of two exponentials than by one exponential.(ABSTRACT TRUNCATED AT 400 WORDS)