Spinal interneurons are numerous and diverse. Some of their roles in the generation of swimming motor behaviors of Xenopus tadpoles and lampreys are well understood. Two major classes of problems regarding spinal interneurons have, however, proven more elusive. The first is the issue of how patterns of activity in populations of interneurons vary during different behaviors. Much is known about recruitment of motoneurons and their involvement in a range of behaviors, but much less is understood about recruitment of interneurons. The second, more difficult problem has been establishing links between the neuronal classes in relatively simple vertebrates and those in more complex ones, such as mammals. Both of these issues are now tractable through the application of optical and genetic methods in zebrafish. By labeling different classes of neurons with calcium indicators in transparent larval zebrafish we have explored the patterns of activity in groups of interneurons during different motor behaviors such as swimming and escape. Our work indicates that gradations in the rapid movements of an escape bend are accomplished largely through changes in the activity level in an active pool of interneurons rather than by recruitment of inactive cells. Changes in the classes of active interneurons are associated with the production of very different motor behaviors such as swimming and escape. One strategy for labeling spinal interneurons for these functional studies is to use the promoters of genes involved in the differentiation of spinal cord to drive expression of genetically encoded fluorescent markers or calcium indicators in subsets of cells. This approach has allowed us to show that Engrailed-1 expression marks a subset of spinal interneurons that are active in swimming and that serve, in part, to gate the flow of sensory information during swimming. These neurons probably play multiple roles both in rhythm generation and in sensory-motor gating. Interneurons with a similar morphology are labeled in transgenic mice in which the Engrailed-1 expressing neurons are targeted. These neurons comprise multiple functional classes in mammals, suggesting that a primitive cell type, similar to the Engrailed positive neurons in fish, may have differentiated into multiple types with more specialized functions in mammals. Optical and genetic tools along with the shared developmental history of vertebrate spinal cord thus allow us to begin to relate neuronal classes in fish to those in mammals. This should help to direct functional studies of the even more diverse array of interneurons in mammalian spinal cord.