Vital statistics data for the United States from 1922 to 1936 and from 1950 to 1972 were used to analyze fetal and early neonatal mortality. This analysis corroborates the previously established pattern of the sex ratio of fetal deaths--highest from months 3 to 5, lower from months 6 to 7 or 8, and increasing at term. It also indicates a postponement of late fetal deaths into the early infant period. Whereas earlier research reports have described the pattern of the sex ratio of fetal deaths, this report repeats this analysis for a recent national data base. This line of analysis is extended by using the patterns observed in the data to produce an empirical estimate of the primary sex ratio. For 1950 to 1972, this ratio (male to female) is conservatively estimated to be 120:100.