Relatively little is known about the long-term impacts of agroforestry extension projects. Assessment of technology adoption (particularly agroforestry given the gestation time of trees and woody perennials) is difficult in the immediate aftermath of an intervention. Projects – due to their time-bound and resource-constrained nature – rarely possess the capacity to monitor and evaluate outcomes occurring after operations cease. Where ex-post studies have been conducted, low adoption rates and post-project abandonment of introduced technologies are frequently observed outcomes. Another issue relates to the implementation and management of introduced practices. Figures purporting to quantify the scale of adoption and diffusion may be misleading since they can give the impression that N farmers have adopted X technology in a uniform and standardised manner. Evidence suggests, however, that technologies are frequently adapted and modified by farmers to achieve a best fit within their own household circumstances and land management objectives.

For these reasons, there is a strong case to be made for revisiting agroforestry project sites years after interventions have ended. This paper reports upon a follow-up case study conducted in 2010/11 in one south Indian village, where a DFID-funded agroforestry project was implemented between 2002 and 2004. Using a combination of geospatial mapping, plot surveys and farmer interviews, the fate of an introduced agroforestry technology was analysed in a temporally and spatially explicit manner. In this case, 97% of adopting households (n=34) were found to have retained the technology – albeit in differing forms and under varying management regimes – and 21% had subsequently expanded the practice on to additional areas of their landholding. Limited diffusion to non-project households was also found to have occurred. At village-landscape scale, fruit-based agroforestry now covers around 15% of suitable agricultural land (excluding low-lying areas of rice cultivation) representing a patchwork effect rather than a wholesale transformation. Reasons for success will be discussed along with constraints to more widespread adoption.