The male - female earnings differential in Canada is investigated with particular attention to the effect of trade unions on the gender earnings gap. An augmented version of the human capital theory based earnings equation is used as the empirical framework of the study. Gender-specific earnings equations are estimated using cross-sectional aggregate data for metropolitan cities and regions in Canada for 1971 and 1981. The empirical results for 1971 indicate that unionization increases the annual earnings of males more than of females, whereas the 1981 results indicate that trade unions decrease the male-female earnings gap. However, the latter evidence is not statistically significant. The study suggests that the increase in female participation in trade unions was an important factor in reducing the gender earnings gap over the period 1971-1981.