1. Intersegmental co-ordination in Xenopus embryos could be influenced by longitudinal gradients in neuronal properties or synaptic drive. To determine if such gradients exist intracellular recordings were made from putative motoneurones at different spinal levels. 2. No evidence was found of a longitudinal gradient in neuronal resting potentials. In a rostrocaudal direction the duration of current-evoked spikes increased and the amplitude of the spike after-hyperpolarization (AHP) decreased. 3. During fictive swimming the amplitude of the tonic excitatory synaptic input and the mid-cycle IPSPs declined in a rostrocaudal direction. The rise-time and fall-time of mid-cycle IPSPs increased in a rostrocaudal direction. 4. Rostral to the eighth post-otic segment mid-cycle IPSPs occurred on all cycles of fictive swimming episodes. More caudally IPSPs became irregular in occurrence and caudal to the twelfth post-otic segment no mid-cycle IPSPs could be detected, even during the injection of depolarizing current or when recording with KCl-filled electrodes. 5. The duration of spikes occurring during fictive swimming increased and the amplitude of spike AHP decreased in a rostrocaudal direction. A spike AHP was absent during fictive swimming activity in neurones caudal to the ninth post-otic segment even though it was present in current-evoked spikes in the same neurones. 6. On-cycle IPSPs (occurring shortly after the spike at phase values less than 0.4) were observed predominantly at the beginning of swimming episodes in neurones recorded rostral to the eighth segment, but were not detected at all in more caudal neurones. 7. If the rostrocaudal gradients in synaptic excitatory and inhibitory drive to putative motoneurones during fictive swimming are also present in premotor spinal interneurones they would be expected to have a strong influence on rostrocaudal delays. Such gradients could therefore be important components of the mechanism underlying intersegmental co-ordination.