Recent studies based on 20th century US data conclude that abortion access raises children's average socioeconomic outcomes. We generalize a model of fertility, highlighting assumptions under which these abortion predictions can be reversed. Using 19th century abortion restrictions, we empirically demonstrate these points. Despite a more than 5 percent increase in birth rates among abortion-restricted cohorts, we find little evidence of negative selection at birth. Longevity was affected nevertheless; in the first ten years of life, children in these larger cohorts died of infectious disease more frequently. These mortality effects diminish with age, potentially reversing at older ages as a result of disease immunity or other offsetting factors.