This article describes the process of recruiting research subjects for a natural history study of illicit stimulant use in rural Ohio using respondent-driven sampling and ethnographic methods. Participant observation, qualitative interviews, and focus groups were used to establish the project and to evaluate and modify the sampling process as it unfolded. We outline the steps taken in several different rural counties, using ethnographic data to illustrate local differences and obstacles that were faced. The article concludes that respondent-driven sampling is a promising method for identifying and recruiting members of hidden populations in rural areas. However, adequate time must be allotted to establish ethnographic footholds and to reach various networks in separate communities.