Kurland, / Robertson This dissertation research project supports a cultural anthropology student from Pennsylvania State University studying the social networks and patterns of support that low income African American men provide to their children. The questions to be studied involve the variation among low income African American men in the amount, kind and frequency of resource transfers to children; whether this variability is related to formal features of their social networks, specifically the density, size and strength of ties; and how the attributes such as age, socioeconomic status, gender and kinship relation of network members influence contributions of these men to the support of their children. Methods include a snowball sample of men with open-ended and structured interviews, participant observation, and a survey of 307 men. The hypotheses to be tested, among others, are that younger unwed fathers will have denser, more closed networks than older men whose networks will contain many valuable "weak ties"; and that the support will often flow indirectly through intergenerational care-giving networks. This research is important because it will provide a better understanding of familial roles assumed by these men, and will analyze the actual levels of support provided by these fathers, directly and indirectly, to their children. This in-depth ethnographic information will reveal the extant behavior to correct the misinformation produced by census surveys, which have tended to underreport the activity of unmarried fathers.