In the leech Hirudo medicinalis inhibitory motor neurons to longitudinal muscles make central inhibitory connections with excitatory motor neurons to the same muscles. We have used a variety of physiological and morphological methods to characterize these inhibitory connections. The efficacy of the transmission between the inhibitors and the excitors was measured by using three intracellular electrodes, two in the inhibitor (one for injecting current and one for measuring voltage) and a third electrode in the excitor for measuring the resultant voltage changes. We have determined that delta Vpre/delta Vpost, or what we have called the transmission coefficient, is X = 0.51, as measured in the somata of the two cells. Evidence which we have obtained leads us to propose that these inhibitory connections between motor neurons are probably monosynaptic. The synaptic latency is consistent with a monosynaptic connection. In addition, a double-labeling technique, whereby one neuron was filled with Lucifer Yellow and the other with horseradish peroxidase (HRP), was used to determine the anatomical relationship between inhibitors and excitors in whole mounts. This revealed varicosities on the processes of inhibitor motor neurons which appear to make contact with processes of excitor motor neurons. A second double-labeling technique, whereby one neuron was filled with HRP and the other with an electron-dense particulate marker, revealed adjacent processes between an inhibitor and an excitor in electron microscopic thin sections which could be the sites of synaptic contact between the neurons. The connections appear to be mediated largely by graded transmitter release from the inhibitory motor neurons. Three different methods were used to demonstrate that synaptic transmission remained in the absence of impulses in the inhibitory motor neurons. These included eliminating the impulse-supporting portion of the motor neuron by pinching off its axon, abolishing impulses by replacing Na+ with Tris in the medium bathing the nerve cord, and increasing the threshold for impulse production by raising the Mg2+ and Ca2+ concentrations in the medium bathing the nerve cord.