Although several markers of synaptic efficacy are enhanced during proestrus, spatial water maze performance is impaired. Because levels of both estrogen and progesterone are elevated in proestrus, the nature of their individual and combined effects on spatial learning was examined. Long-Evans hooded rats were ovariectomized postpubertally and pretrained on a water maze with a visible platform (nonspatial). Following pretraining, rats were administered estrogen (5 μg sc) or oil 48 and 24 h prior to testing and progesterone (500 μg sc) or oil 4 h prior to testing. Rats were tested on a water maze in a different room with a submerged platform (spatial) for 16 trials with random start location in a single testing day. Latency and path length to the target platform were significantly greater in estrogen plus progesterone-treated animals than in controls. Neither estrogen nor progesterone alone significantly impaired performance relative to controls on either measure. Swim speed was not significantly affected by any of the hormone treatments. Performance on a nonspatial cue task was not significantly altered by ovarian steroids. Thus, the combination of estrogen and progesterone produces deficits in the acquisition of the Morris water maze that may be specific to spatial reference memory. These deficits could be due to hormonal influences on extrahippocampal structures or to detrimental effects on behavior resulting from the increased synaptic activity intrinsic to the hippocampus proper.