Psychological research shows that learning can be powerfully enhanced through testing, but this finding has so far been confined to memory tasks requiring verbal responses. We explored whether testing can enhance learning of visuospatial information in maps. Fifty subjects each studied 2 maps, one through conventional study, and the other through computer-prompted tests. For the tests, subjects were repeatedly presented with the same map with one feature deleted (e.g., a road or river), and tried to covertly recall the missing feature and its location. Subjects' map drawings after 30 minutes were significantly better for maps learned through tests as compared to the same amount of time devoted to conventional study. These results suggest that the testing effect is not limited to the types of memory that require discrete, verbal responses, and that utilizing covert retrievals may allow the effect to be extended to a variety of complex nonverbal learning tasks. (Contains 1 figure and 1 table.) [This article was printed in the in-press version of: "Psychonomic Bulletin & Review" (2007, 14, 474-478). Its contents may differ from the final published version.]