A large literature documents the significant gap in average test scores between blacks and whites, while a related literature finds a substantial narrowing of the gap during the 1980s, and a stagnation in convergence during the 1990s. The authors use two data sources the Long Term Trends NAEP and AFQT scores for the universe of applicants to the U.S. military between 1976 and 1991 to show that most of the racial convergence in the 1980s is explained by relative improvements across successive cohorts of blacks born between 1963 and the early 1970s and not by a secular narrowing in the gap over time. Furthermore, these across-cohort test score gains occurred almost exclusively among blacks in the South. They then examine the potential causes of these large composition effects in the test score gap and their significant variation across U.S. states. They demonstrate that the timing of the cohort-based AFQT convergence closely tracks the convergence in measures of black and white infant health for those cohorts. For example, the cohort-specific AFQT gaps (adjusted for age and year effects and selection into test taking) and the racial gaps in post- neonatal mortality rates deaths between one month and one year of birth exhibit very similar patterns across states and birth cohorts. The authors show that the black-white convergence in AFQT scores appears to have been more closely linked with post- neonatal mortality rates than with earlier health measures such as neonatal mortality (deaths within one month of birth) and low birthweight. They also find little evidence that other potential confounders (e.g., schooling desegregation, family background) can explain these patterns in AFQT scores. Investments in health at very early ages after birth appear to have large, long-term effects on human capital accumulation. They authors also discuss preliminary evidence that the staggered timing of hospital integration across the South is consistent with the patterns of gains in black test scores 17 to 18 years after birth more so than other hypotheses for progress in black infant health (e.g., Food Stamps, AFDC, Medicaid).