Androgens affect cognitive processes in both humans and animals. The effects of androgens may be limited to certain cognitive domains, specifically spatial memory, but this hypothesis remains elusive. Here, we tested castrated and sham-operated mice in various behavioral tasks to ask whether androgens affect multiple or specific cognitive domains in male mice. Castration impaired spatial working memory performance in the delayed matching to place water maze task following a 1-h, but not a 1-min, retention interval, as has been reported for rats. In contrast, castration had no effect on novel object recognition memory, spatial reference memory in the water maze, motor coordination, or passive avoidance memory. Castration increased anxiety-like behavior in the open field test, but not the elevated zero maze. Finally, we assessed the effects of androgen replacement with non-aromatizable dihydrotestosterone on spatial working memory following various retention intervals. Dihydrotestosterone recovered spatial memory performance following a 24-h, but not a 1-h retention interval, and had no effect at other retention intervals. These data support that in male mice androgens specifically affect spatial working memory performance, and that the neurobiological processes underlying spatial memory formation may be differentially affected by androgens.