In thinking about the diffusion of innovation, one tends to overlook the obvious fact that not all innovations are adopted by, or are intended to be adopted by, individuals. In the first place, different sorts of innovations may "require" different units of adoption- for example, "it takes two to tango." In the second place, different cultural or situational norms may "prescribe" different units of adoption for an innovation. Most empirical research on diffusion has focused exclusively on the individual as the unit of adoption. This is because the innovations that have attracted modern sociologists have tended to be appropriate for individual adopters. Still, it is altogether obvious that certain recommended contraceptive practices, for example, "require" joint adoption by husband and wife or that middle-class culture "prescribes" a family decision concerning the purchase of a new car. Focusing only on the individual in such cases is misleading if one is to understand the diffusion process completely. Here there is something to be learned from anthropological students of diffusion who often treat the tribe or the group as the unit of adoption even for such ostensibly (to us) individualistic innovations as Christianity in cases where the decision of the chief or the elders is binding upon all. Moreover, many of the innovations in our society are adopted not by individuals or even by families but by organizations. The city-manager idea and the kindergarten were adopted by cities and by school boards respectively; automation is adopted by factories.2 This paper proceeds on the assumption that it is worth exploring the process of innovation front the point of view of the social units which adopt them. As a beginning, let us assume that there are three distinguishable units: individuals, informal groups or collectivities, and formal organizations of all kinds. Innovations can then be classified in terms of the extent to which they "require" one or another type of unit. Culture and subcultures can be classified in terms of their preference among the types of unit for given kinds of innovation.