This paper investigates the intergenerational transmission of health using individual survey data on 2.24 million children born to 600000 mothers during 1970-2000 in 38 developing countries. These data are merged with macroeconomic data by region and birth cohort to create an unprecedentedly large sample of comparable data that exhibits massive variation in maternal and child health as well as in aggregate economic conditions. Our measure of maternal health is height, and our measure of child health is infant mortality risk. We find a substantial positive intergenerational correlation of health. This is undiminished when we purge pure endowment effects using environmental conditions in the mother’s birth year to instrument her height. The effects are nonlinear, being larger at the tails of the height distribution. They are asymmetric in that the penalty associated with having a short mother is much larger than the gain attached to having a tall mother. The higher fertility of shorter women in these data does not seem to be an important mechanism driving the correlation. Higher education and urban location at the individual level and wider economic development at the aggregate level are both shown to weaken the intergenerational coefficient, and these SES effects are most marked for short women. Overall, the analysis suggests that relaxing liquidity constraints and improving the supply or effectiveness of public services will limit the degree to which child health is tied to family circumstance and, accordingly, limit intergenerational persistence in inequality in health and wealth.