1. The contribution to the diastolic depolarization of the hyperpolarization-activated current, if, relative to other components was investigated in isolated rabbit sino-atrial (SA) node myocytes. 2. During the diastolic phase the membrane potential depolarized by 0.1096 +/- 0.014 V/s, which requires only about 3 pA of inward current in a cell with an average capacity of 30 pF. The problem of which ionic component is responsible for initiating the diastolic depolarization was investigated by analysing the composition and the properties of the net inward current in the diastolic range of voltages. 3. The measured instantaneous 'background' current activated during voltage clamp steps from a holding potential of -35 mV was outward positive to approximately -61 mV, and had a region of negative slope conductance from -45 to -35 mV. 4. The instantaneous component lost its rectifying behaviour in the presence of Ni2+ (100 microM) and nitrendipine (10 microM). These blockers of Ca(2+)-dependent currents modified the instantaneous I-V relation at voltages positive to -45 to -50 mV, thus implying that Ca2+ currents become important at less negative potentials than -50 mV, towards the very end of diastolic depolarization. 5. Possible errors introduced by voltage clamp analysis with the whole-cell method on the instantaneous current and on if measurement were evaluated. Leakage through the seal resistance caused the instantaneous I-V relation to be displaced in the inward direction at negative voltages. Correction for the seal leakage moved the reversal potential for the instantaneous current toward the negative direction from -61 to approximately -66 mV. Thus, no depolarization can be driven by the background current beyond -66 mV. 6. During voltage clamp analysis, lack of series-resistance compensation led to lack of intracellular voltage control, as was apparent using a second pipette on the same cell. This slowed activation of if and led to a 1.5- to 2-fold reduction of if size in the range -55 to -115 mV. Thus, uncorrected measurements of the instantaneous component and of if may concur to underestimate the role of if in pacemaking. 7. These results lead to the conclusion that in the SA node cells analysed, pacemaker activity is generated with the essential contribution of the hyperpolarization-activated current, if. Numerical computation of SA node cell activity using an extension of the DiFrancesco-Noble model shows that the if-activation hypothesis can account for the presence of spontaneous action potentials and their sensitivity to if changes.