Electrical coupling has been found in many brain areas and proposed to be important in generating certain oscillatory or rhythmic activities (Loewenstein et al. 2001). Xenopus tadpoles at stage 37/38 can swim at a rhythm of 10-25 Hz. Their swimming circuit in the spinal cord and hindbrain has less than 10 types of neurons. Previous study (Perrins & Roberts, 1995) showed electrical coupling between motoneurons in the spinal cord. Our recent experiments have revealed the existence of extensive electrical coupling among one of the interneuron types: premotor excitatory interneurons with descending axons (dIN). Paired whole-cell recordings from neurons identified with neurobiotin staining showed that dINs were electrically coupled exclusively to other dINs (35 in 51 dIN-dIN pairs). In contrast, apart from weak coupling in one dIN-motoneuron pair, no clear electrical coupling was found in 29 pairs of one dIN and another central pattern generator neuron of other types (23 motorneurons, 3 commissural interneurons and 3 ascending interneurons). The coupling was bi-directional with a small coupling co-efficient (7.3 ± 2.5%, n=27) and with little evidence for dye coupling. Gap junction blockers carbenoxolone (n=12, 300 – 500 μM) and heptanol (n=3, 2mM) only blocked the coupling at high concentrations and this block was not reversible. Flufenamic acid reversibly blocked the electrical coupling (n=4, 100 – 400μM). All three chemicals produced side effects such as affecting miniature EPSP frequencies, resting membrane potentials, spike shape and dIN firing to current injection. The lack of specificity of gap junction blockers makes interpreting the role of electrical coupling among dINs during tadpole swimming difficult. The function of this type of gap junctional connection may involve synchronizing dIN activities and produce more accurate swimming activity pattern.