Do the characterisitics of local social structures affect fertilizer adoption among rural households? This paper extends the model of technology adoption of Feder and Slade (1984) to incorporate social capital, and then tests the model with household data from two agro-ecological zones in rural Tanzania. Probit estimates of the model show that the probability of adoption of improved fertilizer in 1994-95 in the Central Plateau region in increasing in land under cultivation, cumulative adoption patterns, ethnically-based social affiliations, the adoption of improved seeds, the availability of credit and extension services, and the average years of residence in the village. In the Plains region, this probability is increasing in land under cultivation, ethnically based social affiliations and consultative norms. Overall, these results, which are robust after testing for the likely reverse causality of land under cultivation, support the finding that ethnically based and participatory social affiliations act as forms of social capital in the adoption decision.