We introduce a new instrument for family size, infertility, to investigate the causal relationship between children and female labor force participation. Infertility mimics an experiment where nature assigns an upper bound for family size, independent of a woman's background. This new instrument allows us to investigate the differential labor supply without restrictions on initial family size. Using the Demographic and Health Surveys from 26 developing countries we show that OLS estimates are biased upward. We find that the presence of children affects neither the likelihood of work nor its intensity, but impacts the type of work a woman pursues. (Contains 10 tables, 1 figure, and 14 footnotes.)