Although technology adoption has been the subject of a great deal of economic research, that focused on the economics of adoption of low-input "sustainable" systems has been much more limited and recent. This paper attempts to explain the recent decline in the use of cover crops using in maize farming in the Department of Atlantida, Honduras. In the early 1970's, farmers in the region began rotating maize with the velvetbean (mucuna ssp.), a system learned from Guatemalan immigrants. Tohe mucuna-maize system decreased the labor required for maize farming even as it increased yields, prevented erosion, and conferred a variety of other agronomic benefits. By 1992, estimates show that the system had diffused among more than 60% of farmers in the Department. Both due to this widespread dissemination, and the fact that diffusion was largely spontaneous (unassisted by extensions and NGOs), the maize-mucuna system has become a widely acknowledged "success story" of sustainable agriculture diffusion. However, recent anecdotal evidence, confirmed by the survey research reported here, shows that by the late 1990s, use of the system had begun to decline sharply. Various hypotheses about the cause of this decline were investigated in this research, including whether the abandonment of the mucuna-maize system is attributable to a generalized decline in maize cultivation, changes in land tenure and distribution, a burgeoning cattle industry, infrastructural improvements, widespread infestations of noxious weed (rottboellia cochinchinensis), or limitations in farmer management. Modeling techniques evaluated two land-use decisions: whether to adopt mucuna-maize and the contingent decision of whether to abandon the system, once adopted. Bivariate probit analysis is used in the econometric analysis. Descriptive statistics and econometric results indicate that age, level of income from non-maize sources, the presence of rottboellia, and access to a road or highway are significantly related to the abandonment of overcropping. Meanwhile, greater dedication to maize, diversification into high value crops, greater experience with the system, and annual reseeding of mucuna are associated with continued use of the mucuna-maize rotation. The empirical results overall demonstrate that the phenomenon of maize-muchuna adoption and abandonment is a highly complex process. The results have policy implications for the "farmer to farmer" model of extension as well as the promotion of mucuna-maize as a sustainable agriculture technique. In the first case, less emphasis on diffusion and greater attention to farmer-to-farmer teaching of crop system dynamics may be important for the durability of cover crop systems. Regarding the second, cover crop species like mucuna should not be viewed as "silver bullet" solution to sustaining low-input agriculture: indeed, exclusive rotation of mucuna with maize may eliminate critical sources of plant and animal species diversity, ultimately undermining the system itself.