Using a half-century of death records from San Antonio/Bexar County, Texas, we examine the timing and cause structure of Spanish surname and Anglo infant mortality. Our findings show that despite the substantial disparities between ethnic-specific infant mortality rates in the early years of the study, there have been consistent declines in overall, neonatal, and postneonatal mortality for both groups, as well as a major convergence of mortality rates between Spanish surname and Anglo infants. Further, we demonstrate that the convergence is of relatively recent origin and is due primarily to shifts in postneonatal mortality. Finally, we examine the transition reflected in the cause structure of ethnic-specific infant mortality and show that the convergence was largely the result of reductions in deaths from exogenous causes. Implications for research into the “epidemiologic paradox” are discussed. * The authors gratefully acknowledge the support provided for this research by the National Institute of Child Health and Human Development (Grants HD 22490 and 24643). We also wish to thank Starling G. Pullum for invaluable computing assistance, and Benjamin Bradshaw and Arlen Carey for helpful comments on earlier drafts.