The T-maze has commonly been used to investigate the mechanisms underlying spatial learning in vertebrates and has yielded much information about how animals use response and place cues to orient toward a goal. We designed a T-maze to study the spatial learning abilities of crayfish (Orconectes rusticus), using tactile stimuli as a place cue and escape from warm water for reinforcement. An initial experiment found that most animals did not display a side-turning bias when first placed in the maze, and hence animals were randomly assigned to escape from the left or the right arm, one of which contained a smooth floor and the other a rough floor. We found that, over repeated trials, the latency to escape and the number of turns made prior to escaping significantly decreased indicating that crayfish learned to escape from the maze more rapidly and efficiently. Learning occurred over the course of six trials on a single day, and over 5 days of testing, providing evidence for spatial memory lasting 24 hr. In probe trials, in which experienced animals started the maze in an arm opposite to that used during training trials, crayfish did not display a preference for either response-based learning or place-based learning. Instead they engaged in renewed exploration of the entire maze. These findings suggest that, in addition to remembering the location of the exit, crayfish also remembered the overall configuration of the maze.