The meta analysis reported in this paper considers the implications of combined correlational evidence for the nature of the relationships of mathematical and spatial skills, and for the possibility that spatial skill underlies gender differences in favor of males on mathematical tasks. In all, 136 studies reported in 116 articles and dissertations were considered. Results indicate that corrected space-mathematics correlations are not high. Geometry-space correlations are surprisingly low. Orientation correlations are generally lower than visualization correlations. Verbal-mathematics correlations are usually numerically higher than space-mathematics correlations; same-study differences are frequently significant, especially for two-dimensional spatial tasks. Thus, correlational evidence does not indicate that spatial skill plays a special role in mathematical achievement as mathematics is taught and tested today. Gender patterns sometimes vary. Females' correlations do not differ according to the cognitive level of the mathematics task as often as do those of males. Females have higher verbal-mathematics than space-mathematics correlations in more categories than do males. Correlational evidence does not support the hypothesis that spatial skill underlies gender differences in mathematics. Seven tables and four figures present data from the analysis. There is a 158-item list of references. An appendix contains 16 additional tables of study data. (SLD)