In the present work, dynamic clamp was used to inject a current that mimicked tonic synaptic activity in the soma of cat lumbar motoneurones with a microelectrode. The reversal potential of this current could be set at the resting potential so as to prevent membrane depolarization or hyperpolarization. The only effect of the dynamic clamp was then to elicit a constant and calibrated increase of the motoneurone input conductance. The effect of the resulting shunt was investigated on repetitive discharges elicited by current pulses. Shunting inhibition reduced very substantially the firing frequency in the primary range without changing the slope of the current-frequency curves. The shift of the I-f curve was proportional to the conductance increase imposed by the dynamic clamp and depended on an intrinsic property of the motoneurone that we called the shunt potential. The shunt potential ranged between 11 and 37 mV above the resting potential, indicating that the sensitivity of motoneurones to shunting inhibition was quite variable. The shunt potential was always near or above the action potential voltage threshold. A theoretical model allowed us to interpret these experimental results. The shunt potential was shown to be a weighted time average of membrane voltage. The weighting factor is the phase response function of the neurone that peaks at the end of the interspike interval. The shunt potential indicates whether mixed synaptic inputs have an excitatory or inhibitory effect on the ongoing discharge of the motoneurone.