In order to investigate the properties and pharmacology of the excitatory synaptic drive received by motoneurons during swimming in the lamprey, propriospinal excitatory interneurons were activated as a population by the regional application of N-methyl-D,L-aspartate (NMA) to either the 6-8 rostral-most or the 6-8 caudal-most segments of lengths of isolated spinal cord. This caused a rhythmic motor output to be generated in these regions. Synaptic potentials that were phase-locked to, and dependent on, the rhythmic motor activity of the segments exposed to the agonist could be recorded in motoneurons lying outside the activated regions. The synaptic drive to motoneurons located rostral or caudal to the activated regions was studied. Motoneurons received both descending and ascending synaptic input, which consisted of alternating excitatory and inhibitory phases. The inhibition could be reversed by chloride injection and blocked by strychnine, leaving an oscillating excitatory phase. The descending excitatory drive could extend 1-9 segments from the active region, while the ascending excitatory drive was recorded only in motoneurons that were 1-3 segments rostral to the active region. Both types of drive occurred in phase with the ipsilateral ventral root discharge: The peak depolarization of the descending drive occurred at the same point in the swimming cycle as that of the depolarizing phase seen during fictive swimming, while that of the ascending drive occurred significantly later. Both ascending and descending drives were partially reduced in amplitude by 2-amino-5-phosphonovaleric acid or Mg2+. The blocking action of Mg2+ was, in both cases, voltage dependent. Cis-2,3-piperidine dicarboxylic acid or kynurenic acid caused a much greater reduction in the amplitude of the oscillations. These results suggest that a major part of the excitatory drive for swimming in lamprey motoneurons is generated by populations of propriospinal interneurons with relatively long descending and/or short ascending axons, which fire rhythmically during swimming and release an amino acid transmitter that excites motoneurons through N-methyl-D-aspartate (NMDA) and non-NMDA receptors. This information will allow these important neurons to be identified in future experiments.