1 We used single human atrial myocytes to study I f occurrence, properties and pharmacological modulation. Cells were obtained by chunk enzymatic digestion from samples of right atrial appendages of patients undergoing corrective cardiac surgery. 2 Patch-clamped cells in the whole-cell con®guration were superfused with a modi®ed Tyrode solution to reduce contamination by interfering currents and to amplify I f . The average cell membrane capacitance was 85.06+2.41 pF (n=531). Data were consistent with the geometrical dimensions of the cells (length 94.2+1.89 mm, width 17.9+0.42 mm, n=126). 3 When hyperpolarizing to 7120 mV from a holding potential of 740 mV, 252 of 306 tested cells (82%) expressed a hyperpolarization-activated inward current (I f density =3.77+0.25 pA pF 71 ); the current was considered to be present in a given cell if its density at 7120 mV was larger than 0.5 pA pF 71 . 4 Current activation was sigmoidal and ®tted a Boltzmann model; the average activation curve (n=25) showed a maximum current amplitude of 205.97+19.94 pA, corresponding to 3.87+0.63 pA pF 71 , voltage of half-maximal activation (V 1/2 ) at 786.68+2.19 mV and a slope of 711.39+0.69 mV. The reversal potential of I f measured by tail-current analysis was 713.07+1.92 mV (n=6). The addition of CsCl (5 mM) fully and reversibly blocked the current. 5 In the presence of the b-adrenoceptor agonist isoprenaline (Iso, 1 mM), V 1/2 was signi®cantly shifted toward less negative potentials by 6.06+1.96 mV (n=16, P=0.0039). The selective A 1 -adenosine receptor agonist cyclopentyladenosine (CPA, 1 mM) caused a statistically signi®cant shift of V 1/2 toward more negative potentials with respect to the control curve, both in the absence (77.37+1.83 mV, P=0.0005, n=11) and in the presence of 1 mM Iso (74.97+1.78, P=0.031, n=6). 6 These results demonstrate that a current with the properties of I f described in cardiac primary and secondary pacemakers occurs in the majority of human atrial cells. While the pathophysiological relevance of I f in human atrial tissue remains to be de®ned, our data clearly show that it is modulated through stimulation of b-adrenoceptors and A 1 -adenosine receptors.