1. A single-microelectrode technique has been used to study the voltage dependence and the kinetics of inactivation and reactivation of a tetrodotoxin-resistant inward current (ICa) in single cells from bull-frog atrium. 2. In most cases the kinetics of both inactivation and reactivation can be well described as a single-exponential process. 3. Several different observations indicate that inactivation of ICa in these cells is controlled by both voltage-dependent and current-dependent processes, as has been demonstrated previously in heart (Kass & Sanguinetti, 1984; Lee, Marban & Tsien, 1985) and in other tissues (Hagiwara & Byerly, 1981; Tsien, 1983; Eckert & Chad, 1984). 4. Evidence in favour of a voltage-dependent inactivation mechanism included: (a) In paired-pulse measurements of steady-state inactivation ('f infinity') a 'conventional' steady-state f infinity vs. membrane potential (Vm) relationship was obtained in the range of membrane potentials from -60 to 0 mV. (b) Increasing [Ca2+]o from 2.5 to 7.5 mM, which resulted in a 2-3-fold increase in ICa, did not produce any significant increase in the amount of inactivation. (c) Using a 'gapped' double-pulse protocol non-monotonic U-shaped inactivation relationships were obtained, i.e. positive to approximately +20 mV some removal of inactivation occurred. However, f never approached a value near 1.00 at very depolarized potentials; it reached a maximum between 0.5 and 0.6. (d) In constant [Ca2+]o and at fixed Vm, the kinetics of ICa inactivation were independent of peak size of ICa. This was demonstrated by: (i) varying the holding potential (-90 to -30 mV), (ii) using paired-pulse 'recovery' protocols, and (iii) partial block by La3+ (1-10 microM) and Cd2+ (0.1 mM). (e) Influx of Ca2+ ions was not an obligatory prerequisite for development of inactivation. In all ionic conditions (Ca2+, Sr2+, Ba2+, Na+-free and Ca2+-free Ringer solutions) currents displayed inactivation phenomena, although the extent and kinetics of inactivation were dependent upon ionic conditions. Outward currents recorded above the reversal potential for ICa exhibited time- and voltage-dependent inactivation, and could be inactivated by brief depolarizing pre-pulses that produced no net inward current flow. Evidence against a possible role of the electrogenic Na+-Ca2+ exchanger in producing inactivation of these outward currents was obtained.(ABSTRACT TRUNCATED AT 400 WORDS)