This study aimed to provide empirical verification to the hypothesis that unwanted fertility could lead to mortality, with underinvestment as an intervening mechanism. Unwanted fertility was defined as a live birth that deviated from parental expectation in some distinguishable characteristics or set of characteristics. Births defined as unwanted were: 1) births which respondent indicated she preferred to have occurred later or births which she did not want at all, 2) those whose birth order exceeded the mother’s desired family size, and 3) those which exceeded the number of desired boys of the number of desired girls of the mother. Underinvestment as defined by Scrimshaw (1978) involved the idea that mothers would not go to extremes to save the life of a child and might invest more time and resources in other children. To measure underinvestment, a scale was constructed made up of the following items: 1) source of prenatal care, 2) frequency of prenatal care, 3) age started supplementary feeding, 4) length of breastfeeding, 5) source of medical treatment, 6) length of time before treatment was sought, and 7) time spent cuddling and playing with child. Mortality was classified into two types: postneonatal mortality and childhood mortality. Postneonatal mortality involved deaths from ages one to eleven months. Childhood mortality involved deaths at ages one to four years. Thus, analysis proceeded in two ways: 1) that involving all live births born at least a year before the time of the interview and 2) that involving only live births born at least five years before the time of the interview. Results of the study did not support the hypothesis that unwanted fertility could lead to mortality, with underinvestment as an intervening mechanism. In general, no significant association was found between unwantedness and mortality. However, underinvestment and mortality (postneonatal and childhood) were significantly associated for both wanted and unwanted births. Underinvestment and childhood mortality were significantly related for births in lower class families and for births in middle and upper class families. Postneonatal mortality was significantly related to underinvestment among lower class births only. Postneonatal mortality was not significantly related to SES. Childhood mortality was significantly related to SES with underinvestment as an intervening variable. The study underscored the difficulties involved in measuring the variables of interest -- unwantedness and underinvestment. Several limitations of the study were pointed out and recommendations for further research were stated. / Ph. D.