The growth of 'the community' as a inajor focus of development through which collective social action can take place has spread rapidly through development ideology since the 1970s. Yet the existence of communities and their structure and degree of coherence is an empirical matter, to be described through ethnography, rather than something that can be assumed a priori. The paper takes the view that uniform notions of community are often imposed on diverse rural populations both for administrative convenience and for sentimental reasons: outsiders would like to see communities in the developing world precisely because they appear to have become fragmented in the North. To this end they may find it necessary to construct communities where none exist through various 'participatory' approaches. The reasons why communities may not exist are diverse: this may be the result of climatic uncertainties and thus a patchy resource base; it may result froin war and insecurity,. or it may simply reflect changing economic opportunities. Individuals and households are obliged to seek subsistence opportunistically and cannot afford the luxury of a stable community. The paper takes the example of the situation in the Rwenya Basin in north east Zimbabwe, where elements of all the above factors are present, to suggest that interventions can only be effective when the notion of community is deconstructed and a greater level of ethnographic realism is introduced.