The effects of prenatal care utilization on birth weight (BW) may vary by unobserved fetal health endowments. This heterogeneity will be masked by estimating the effects at BW mean but can be evaluated by estimating the effects at BW quantiles as fetal health endowment is a strong correlate with the BW quantile order. We estimated the effects of prenatal care visits and delay before prenatal care initiation, on BW mean and quantiles using a sample of infants from Argentina. Self-selection into prenatal care was modeled using 2SLS and instrumental variable quantile regression. Results suggest that the 'mean' effect of prenatal care utilization largely underestimates the effects at lower BW quantiles. About 35 and 77 g increase in BW mean and 0.1 quantile respectively, per visit and about 30 and 139 g decrease in BW mean and 0.1 quantile respectively, per week delayed, were estimated. Ignoring self-selection into prenatal care resulted in underestimation of mean and quantile effects. Results highlight the limitation of analyses focused on 'mean effects' in the presence of treatment heterogeneity and emphasize the importance of identifying women at risk for having infants at lower BW quantiles as they may benefit most from earlier and more intensive prenatal care.