Plasma membrane was purified from roots of rye (Secale cereale L. cv. Rheidol) by aqueous-polymer two-phase partitioning and incorporated into planar bilayers of 1-palmitoyl-2-oleoyl phosphatidylethanolamine by stirring with an osmotic gradient. Since plasmamembrane vesicles were predominantly oriented with their cytoplasmic face internal, when fused to the bilayer the cytoplasmic side of channels faced the trans chamber. In asymmetrical (cis:trans) 280∶100 mM KCl, five distinct K(+)-selective channels were detected with mean chord-conductances (between +30 and -30 mV; volyages cis with respect to trans) of 500 pS, 194 pS, 49 pS, 21 pS and 10 pS. The frequencies of incorporation of these K(+) channels into the bilayer were 48, 21, 50, 10 and 9%, in the order given (data from 159 bilayers). Only the 49 pS channel was characterized further in this paper, but the remarkable diversity of K(+) channels found in this preparation is noteworthy and is the subject of further study. In symmetrical KCl solutions, the 49 pS channel exhibited non-ohmic unitary-current/voltage relationships. The chord-conductance (between +30 and-30 mV) of the channel in symmetrical 100 mM KCl was 39 pS. The unitary current was greater at positive voltages than at corresponding negative voltages and showed considerable rectification with increasing positive and negative voltages. This would represent 'inward rectification' in vivo. Gating of the channel was not voltage-dependent and the channel was open for approx. 80% of the time. Presumably this is not the case in vivo, but we are at present uncertain of the in vivo controls of channel gating. The distribution of channel-open times could be approximated by the sum of two negative exponential functions, yielding two open-state time constants (τo, the apparent mean lifetime of the channel-open state) of 1.0 ms and 5.7 s. The distribution of channel-closed times was best approximated by the sum of three negative exponential functions, yielding time constants (τc, the apparent mean lifetime of the channel-closed state) of 1.1 ms, 51 ms and 11 s. This indicates at least a five-state kinetic model for the activity of the channel. The selectivity of the 49 pS channel, determined from both reversal potentials under biionic conditions (100 mM KCl∶100 mM cation chloride) and from conductance measurements in symmetrical 100 mM cation chloride, was Rb(+)≥ K(+)> Cs(+) > Na(+) > Li(+) > tetraethylammonium (TEA(+)). The 49 pS channel was reversibly inhibited by quinine (1 mM) but TEA(+) (10 mM), Ba(2+) (3 mM), Ca(2+) (1 mM), 4-aminopyridine (1 mM) and charybdotoxin (3 μM) were without effect when applied to the extracellular (cis) surface.