Kinetics of the cardiac pace-maker current (if) were studied using high K+, low Na+ solutions under conditions where the current time course could be dissected from other components. Activation of if during relatively large negative pulses is S-shaped, and is approximated by an exponential function of time to the third power. Less-pronounced S-shaped activation occurs at potentials close to the middle of the activation curve (near -70/-80 mV). Here, allowing for the presence of a very slow component, the power required to fit the current activation approaches 1. The comparison between current activation and deactivation at the same potentials shows that although deactivation can be approximated by a single exponential, the two processes have a quite different time dependence, and this difference depends on the membrane potential. This behaviour is not compatible with Hodgkin-Huxley kinetics. While near the half-activation range the current decays with an apparently single exponential time course, at more positive potentials the current deactivation becomes sigmoidal. At least the third power of an exponential is required to fit its time course at potentials positive to about -40 mV. These data imply that both open and closed states correspond to several distinct channel configurations. The 'delay' in the current onset during a hyperpolarization is decreased by applying large, short hyperpolarizations before activation. Suitable pre-pulse durations and/or amplitudes can reduce the subsequent current activation to a single exponential. Records with and without a pre-pulse do not always superimpose. After the activation 'delay' has been removed by a suitable hyperpolarization preceding an activating pulse, the time course of its recovery can be studied by applying depolarizations of given amplitude and variable duration. The time course of the delay recovery does not seem to be linked to the time course of current deactivation recorded at the same voltage. Reduction of the activation 'delay' by conditioning pre-hyperpolarizations does not affect current decay during a subsequent depolarizing pulse. The current decay appears to depend only on the current amplitude reached before a deactivating pulse is applied. This, and the evidence in the preceding paragraph, suggest that the delay recovery and the current deactivation are independent processes. A reaction scheme is proposed, which has been developed on the basis of the experimentally determined kinetic properties of if. The channel model is composed of five gating subunits of three different types, not all independent in their movements.(ABSTRACT TRUNCATED AT 400 WORDS)