This study examines the effects of (1) current individual parental assets, (2) assets held by each parent before marriage, (3) transfers made at the time of the parents' marriage, and (4) family background characteristics of parents on the morbidity of preschool boys and girls in rural Bangladesh. The approach is unique in that it simultaneously tests alternative models of household decisionmaking and investigates gender bias within the household. Moreover, it is one of the few investigations we know of that provides a formal test of the intrahousehold model in the South Asian context. Results indicate that higher father share of current assets benefits boys’ health, but does not affect girls’ health. A greater proportion of pre-wedding assets held by the mother lowers the number of morbidity days experienced by girls. A larger share of wedding payments directed toward the husband’s side at the time of marriage reduces illness for preschoolers of both sexes later in the union. The finding that maternal and paternal shares have different impacts leads to a rejection of the unitary model of the household. Extended families also appear to play an important role in the production of child health, especially the number of living brothers of each parent, which reduces preschooler morbidity days.