The present investigation assessed the effects of the selective GABAB receptor agonist baclofen (1, 3, and 6 mg/kg) on spatial learning in the Morris water maze, an aversively motivated spatial learning task. Potential anxiolytic and sedative effects of baclofen were also assessed in an open field. Baclofen dose-dependently reduced locomotion in the open field but had little effect on thigmotaxia (anxiety). In the water maze, baclofen dose-dependently impaired spatial learning and reduced swim speed. During the probe trial given after training, only rats treated with the highest dose of baclofen (6 mg/kg) failed to show a bias for the correct quadrant. Following four additional retraining trials, a second drug-reversal probe trial was given and it was found that rats switched from saline to the highest dose of baclofen (6 mg/kg) showed a bias for the correct quadrant, as did rats switched from the two lowest doses of baclofen (1 and 3 mg/kg) to saline. Rats switched from the highest dose of baclofen (6 mg/kg) to saline failed to show a quadrant bias. Performance on a visible platform task was not impaired by baclofen at any dose. Together these results suggest that baclofen resembles GABAA agonists/positive modulators in that it impairs spatial learning, but not performance of a previously acquired escape response; but differs in that it does not reduce thigmotaxia (anxiety). Potential mechanisms by which baclofen impairs mnemonic processes are discussed.