To determine the membrane mechanisms underlying the interactions between inhibitory postsynaptic potentials (IPSPs) and excitatory inputs, we investigated, at the membrane potential level, the combined influences of low-frequency (0.05-0.50 Hz) imposed sinusoidal transmembrane currents (termed sine currents), representing the excitatory drive, and trains of regular (3-30/s) IPSPs. The two simplest possible neuron systems exemplified by the slowly and rapidly adapting stretch receptors of crayfish (RM1 and RM2, respectively) were used. At constant elongation the RM1 and RM2 behaved as a pacemaker and a neuron without self-sustained oscillations, respectively, but in dynamic conditions uninhibited controls and IPSP sine current interactions were essentially identical in both RMs. Controls showed the usual smooth variation of the RM firing rate in response to the gradually varying excitatory input. IPSP effects were characterized by the expected overall reduction of the postsynaptic firing rate. More important, special effects were also present, such as the simple fixed alternations of IPSP and postsynaptic spikes (e.g., 1 IPSP, 1 postsynaptic or 1:1; 1 IPSP, 2 postsynaptic or 1:2; 2 IPSPs, 1 postsynaptic spike or 2:1), where interspike intervals were more constant than uninhibited controls and where the sensitivity to the excitatory input was reduced to small values, and the sudden firing rate discontinuities consisting of instantaneous discharge accelerations or decelerations (termed 'jumps') between successive alternation ratios, where sensitivity increased to large values. Therefore with inhibition the RM firing rate varied discontinuously in response to the gradually changing input, and the discharge rate could take one of several discrete values by switching between different alternation ratios. At the alternations the times elapsed between an IPSP and the closest spike before (phase, ∅) or after it (cophase, Θ) increased and decreased, respectively, with increasing excitation. The major membrane potential modification that accompanied the interactions at the alternations was the gradual increase of the post-IPSP slope as a function of excitatory drive, which reduced the time to reach the firing level or Θ. Inhibition introduced subtle and complex nonlinear modifications in the coding of convergent excitatory input. The most notable nonlinearity was the discontinuous variations of the firing rate as a function of the gradually changing excitatory input. Effects were due to voltage interactions occurring at the extrasynaptic membrane, with a decisive involvement of the spike generator and insignificant participation of the shunting action of IPSPs. The results provide yet another example of the predominant influence of intrinsic membrane properties in determining the effects of synaptic-evoked activity. Evidence exists indicating that the complex IPSP effects operate in the natural situation and are not an experimental peculiarity. The dependence of IPSP effects on ∅ provides instantaneous feedback modulation of IPSP strength that does not involve a change in synaptic efficacy and that occurs in the absence of feedback connections. A model is proposed that explains several major features of our data.