Almost 60 years after the Renshaw cells were first identified (Renshaw, 1946), their contribution to motor control is still elusive. The complexity of the spinal recurrent network has led to some contradictory ideas as to how recurrent inhibition may actually influence the ± motoneurone firing times (cf. Windhorst, 1996). For instance, beside a stabilizing effect on motoneurone discharge variability, positive or negative effects on motoneurone synchronous activity have both been postulated. Recurrent inhibition is not homogenously distributed among motoneurone pools. Its presence is well established among the motoneurones innervating proximal limb muscles such as extensor carpi radialis and tibialis anterior, whereas it is lacking in the intrinsic hand muscles such as abductor digiti minimi and it is a matter of debate in the facial muscles such as masseter. If recurrent inhibition does take part in the regulation of the motoneurone firing properties, the effects obtained by altering the efficacy of the spinal recurrent network can be expected to differ depending on the muscle tested. In the present study, spinal recurrent inhibition was transiently enhanced while the discharges of pairs of motor units were being recorded in the extensor carpi radialis, the abductor digiti minimi, the tibialis anterior or the masseter muscles, during voluntary isometric contraction maintained for up to 30 min. The subjects were undergoing continuous intravenous saline (NaCl 0.9 %) perfusion interrupted by a 2 min injection of L-acetylcarnitine (L-Ac, 30 mg/kg diluted in 5 ml of saline), known to enhance cholinergic transmission from motoneurone recurrent collaterals to Renshaw cells in humans (Mazzocchio & Rossi, 1997). In control experiments involving other motor unit pairs, L-Ac injections were replaced by 2 min injections of saline performed in exactly the same conditions. The variability and synchronization of the motor unit discharges were analysed before, during and after the injection. In both the extensor carpi radialis and tibialis anterior muscles, the L-Ac injection led to (1) a significant decrease in the variability of the inter-spike intervals with no consistent changes in their mean duration and (2) a significant increase in the synchronous activity. In contrast, L-Ac injection did not alter in any way the pattern of motor unit discharge in the abductor digiti minimi and the masseter muscles. None of these changes occurred in the control experiments performed with saline injections. The contrasting effects observed with extensor carpi radialis and tibialis anterior motoneurones known to undergo recurrent inhibition and with abductor digiti minimi motoneurones known to lack recurrent inhibition strongly suggest that Renshaw cell activity enhanced by L-Ac injection had promoted the changes in variability and synchronization in the extensor carpi radialis and tibialis anterior motoneurone pools. These data support the hypothesis that recurrent inhibition may contribute to limit the variability of steadily discharging motoneurones and to synchronize motoneurone activity. In contrast, the absence of effect of L-Ac injection in the masseter muscle argues in favour of the notion that masseter motoneurones lack recurrent inhibitory collaterals.