The ontology of Evans-Pritchard (1940)'s characterization of Nuer society is essentially identical to that of Wynne-Edwards (1962, 1986) within evolutionary biology. When compared, each solves an internal problem for the other. The recursive nature of Evans-Pritchard's definition of group is eliminated upon application of Wynne-Edwards' definition; and a mechanism for group selection compatible with Wynne-Edwards' intuitions is provided by Evans-Pritchard. Both are explicated by an experimental manipulation of cooperation among ant queens, providing an evolutionary analysis of punishment which recovers the heart of Wynne-Edwards' thought and documents a process assumed by Evans-Pritchard. By focusing on a manipulatable system which isolates a key phenomenon derived from study of a less tangible system (ethnography) we suggest a partial test of conclusions derived from the latter.