This study of health care allocation to children in northern Puno, Peru, utilizes quantitative and qualitative data to explore differential resource allocation to children in rural Andean households. As part of a broader ethnographic study of health in two communities, quantitative data on reported health status, symptoms, and treatments (both lay and specialist) were collected for 23 children under the age of 7 over a one year period. Additional data were collected from local health post records. Data were analyzed by gender, and by three age groups (birth to 1 year, 1-3 years, 4-6 years) to determine if differences existed in the allocation of health care. The data suggest a pattern of discrimination against females and younger children, especially infants under age one, despite the fact that these groups were reported to be sicker. Differences were especially significant in the allocation of biomedical treatments, the most costly in terms of parental time, effort, and money. Ethnographic data on child illness, gender, and developmental concepts help to explain why children of different genders and ages may be treated differently in the rural andes. They provide a context in which to interpret health care allocation data, and, in the absence of a population-based study, reinforce findings based on the limited study sample. Female children are valued less because of their future social and economic potential. Females are also regarded to be less vulnerable to illness than male children, meaning that less elaborate measures are necessary to protect their health. Young children are thought to have a loose body-soul connection, making them more vulnerable to illness, and are though to be less human than older individuals. The folk illnesses uraña (fright) and larpa explain child deaths in culturally acceptable ways, and the types of funerals given to children of different ages indicate that the death of young children is not considered unusual. Health care allocation and ethnographic data suggest that selective neglect (passive infanticide) may be occurring in rural Peru, possibly as a means of regulating family size and sex ratio. It is important to go beyond placing blame on individual parents or on culture, however, to address the underlying causes of differential health care allocation, such as poor socioeconomic conditions, lack of access to contraceptives, and female subordination.