From 1996 to 2006, Nepal experienced a substantial fertility decline, with the total fertility rate dropping from 4.6 to 3.1 births per woman. This study examines the associations between progress towards universal primary and secondary schooling and fertility decline in rural Nepal. Several hypotheses regarding mechanisms through which education affects current fertility behaviour are tested, including: the school environment during women's childhood; current availability of schools; knowledge of educational costs; and women's own educational attainment. Data for the analysis come from the 2003–04 Nepal Living Standards Survey, a nationally representative random sample of households, which includes detailed data on fertility, household expenditure, educational attainment, demographic characteristics and the use of social services. Census and administrative data are also used to construct district-level gross enrolment ratios for primary and secondary schools during the women's childhood. Discrete dependent variable modelling techniques are used to estimate the effects of the following variables on the probability of women giving birth in a given year: district-level gross enrolment ratios for primary and secondary schools during women's childhood; having had a child previously in school; women's own educational level; current school availability; and other covariates. Separate models are estimated for the overall sample of rural women of reproductive age, and for parity-specific sub-samples. The results suggest that district-level gross enrolment ratios for secondary schools and, in some instances, having had a previous child enrolled in school are significant determinants of fertility in rural areas. These results are highly independent of women's own educational levels. Overall, the results suggest that, in the rural Nepal context, mass schooling influences the fertility transition through both community- and household-level pathways.