It is well known that highly 'female' fields of study in tertiary education are characterized by higher fertility. However, existing work does not disentangle the selection-causality nexus. We use variation in gender composition of fields of study implied by the recent expansion of tertiary education in 19 European countries and a difference-in-differences research design, to show that the share of women on study peer groups affects early fertility levels only little. Early fertility by endogamous couples, i.e., by tertiary graduates from the same field of study, declines for women and increases for men with the share of women in the group, but non-endogamous fertility almost fully compensates for these effects, consistent with higher early fertility in highly 'female' fields of study being driven by selection of family-oriented students into these fields. We also show that the EU-wide level of gender segregation across fields of study has not changed since 2000.