1. The amplitude of recurrent inhibitory postsynaptic potentials (RIPSPs) was examined in pairs of lumbosacral motoneurons that were separated by a known distance and were identified by antidromic stimulation of muscle nerves. One motoneuron was stimulated by injecting depolarizing current pulses, and postsynaptic responses were recorded and averaged in the second motoneuron. Input resistance, rheobase, and conduction velocity were determined for many motoneurons. Most motoneurons innervated extensor muscles. 2. RIPSP values as large as -283 microV were recorded, but most were between -10 and -40 microV. RIPSPs from individual motoneurons of a pool are distributed to several heteronymous motor nuclei and have a range of amplitudes comparable with homonymous RIPSPs. 3. A specific spatial distribution of RIPSP amplitudes was found whereby the largest RIPSP amplitudes (> 40 microV) occurred in motoneurons located within +/- 1.4 mm of the stimulated motoneuron. A significant correlation was found between RIPSP amplitude and the distance between motoneurons for all motoneuron pairs. This correlation was also found within individual groups of motoneuron pairs that innervate the lateral gastrocnemius, medial gastrocnemius, anterior-middle biceps femoris, or soleus muscles. 4. The dependency of RIPSP amplitude on the motoneuron species, which is the particular muscle a motoneuron innervates, is less distinct than the dependency of RIPSP amplitude on topography. Pooling all motoneuron species of close motoneuron pairs indicated that RIPSPs measured in homonymous motoneuron pairs were greater in amplitude than RIPSPs measured in heteronymous pairs. In addition, homonymous RIPSPs of anterior middle biceps femoris or lateral gastrocnemius motoneurons were greater than heteronymous RIPSPs of those motoneurons in all heteronymous combinations. However, homonymous and heteronymous RIPSPs were not significantly different when heteronymous pairs were restricted to individual combinations of species. These findings indicate that RIPSP amplitudes within a set of motor nuclei interconnected by recurrent inhibition are dependent in some cases on the species of motoneurons, but this effect is less important than the effect of topography on RIPSP amplitude. 5. These results indicate that recurrent inhibition in motoneuron pools that innervate hindlimb extensor muscles has a strong topographic organization, such that the strongest recurrent inhibition is produced by each motoneuron in a restricted rostrocaudal zone that includes both homonymous and heteronymous motor nuclei. This suggests that recurrent inhibition is organized for the control of several motor nuclei engaged in common motor activity as well as regulation of activity within individual motor pools.