Do employers substitute adults for children, or do they treat them as complements? Using data from a Mexican schooling experiment, I find that decreasing child farmwork is accompanied by increasing adult labor demand. This increase was not caused by treatment money reaching farm employers: there were no significant increases in harvest prices and quantities, nonlabor inputs, or nonfarm labor supply. Furthermore, coordinated movements in price and quantity can distinguish this increase in demand from changes in supply induced by the treatment's income effects. Thus, declining child supply caused increasing adult demand: employers substituted adults for children.