This paper reports on a detailed sociological study carried out as the NGO (or change agent) was completing three-year projects in two sites, Nagadeepa and Pimburettewa. The study describes the change agent's strategy, and analyzes its impact and the perceptions of farmers and government officials regarding its impact. The case study is placed in a wider context, in terms of both the participatory management policy of the Government of Sri Lanka, and the lessons learned that are relevant for NGOs working in other countries as well.

Argues that single irrigation systems managed by autonomous system-specific organizations accountable to their customers, perform better and are more sustainable than those managed by agencies dependent on the government, or by agencies responsible for multiple systems. Selected cases are reviewed and the plausibility of this hypothesis established. General recommendations are made for policy makers designing irrigation reform programs.

Research results: performances assessment; Research results: design and operation of irrigation systems; Research results: policy, institutions, and management; Research results: health and environment; Training and institutional strengthening; Conclusions: outputs, impacts, and future directions.

On the Accelerated Programme of Mahaweli Development, hydroelectric and irrigation project in Sri Lanka; evaluation and analysis by an inter-disciplinary group which met at a week-long seminar in Ascona, Switzerland.