Through a case-study of cocoa-farming in Ghana, this paper takes up the long-running but recently neglected debate about the ‘cash crop revolution’ in tropical Africa during the early colonial period. It focuses on the supply side, using quantitative evidence as far as possible, to test the much criticised but never superseded ‘vent-for-surplus’ interpretation of the export expansion as a substitution of labour for leisure. The paper argues that while the model captured certain features of the case, such as the application of labour to underused land, its defining claim about labour is without empirical foundation. Rather, the evidence points to a reallocation of resources from existing market activities towards the adoption of an exotic crop, entailing a shift towards a new, qualitatively different and more profitable kind of production function. This innovation is best understood in the context of the long-term search of African producers for ways of realising the economic potential of their resource of relatively abundant land, while ameliorating the constraints which the environment put upon its use.