Gender differences in spatial aptitude are well established by adulthood, particularly when measured by tasks that require the mental rotation of objects (Linn & Petersen, 1985; Voyer, Voyer, & Bryden, 1995). Although the male advantage in mental rotation performance represents one of the most robust gender differences in adult cognition, the developmental trajectory of this male advantage remains a topic of considerable debate. To address this debate, we meta-analyzed 303 effect sizes pertaining to gender differences in mental rotation performance among 30,613 children and adolescents. We found significant developmental change in the magnitude of the gender difference: A small male advantage in mental rotation performance first emerged during childhood and then subsequently increased with age, reaching a moderate effect size during adolescence. Procedural factors, including task and stimulus characteristics, also accounted for variability in reported gender differences, even when controlling for the effect of age. These results demonstrate that both age and procedural characteristics moderate the magnitude of the gender difference in mental rotation throughout development. Public Significance Statement This meta-analysis documents the development of gender differences in spatial reasoning, finding that boys first outperform girls on mental rotation tasks during primary school and that this male advantage subsequently increases with age into adolescence. Because children's spatial skills predict their later science, technology, engineering, and math (STEM) achievement, addressing the gender difference in mental rotation early in childhood may help to minimize later gender disparities in STEM success.