In recent years the use of genetic manipulations to investigate the molecular mechanisms underlying learning and memory has become a common approach. In a great many cases, the spatial learning ability of mutant mice has been assessed using the Morris watermaze task. The performance of these mice may, however, be strongly influenced by their genetic background and, therefore, the interpretation of their phenotype requires a preliminary characterization of the parental strains. The present study compared 129S2/Sv and C57/BL/6J inbred mouse strains, which have been widely used in deriving lines of genetically modified mice, on the hidden platform version of the watermaze task. During acquisition, the C57 mice displayed shorter escape latencies to find the platform than the 129S2s. Further analysis revealed, however, that the C57 mice also swam faster than the 129S2s. The analysis of path lengths was thus a more reliable measure of spatial learning, and revealed an equal level of performance in the two strains. This conclusion was confirmed during the two probe trials with both strains showing a similar spatial preference for the training site. These results suggest that the 129S2 substrain is no less proficient than the C57 substrain in terms of spatial learning in the watermaze, and also demonstrates the importance of not relying solely on escape latency as a measure of watermaze performance.