PACR of a project (7/83-6/92) to construct minor irrigation schemes (MISs) in Madhya Pradesh, India. Restructuring of the project after the mid-term evaluation resulted in construction of fewer irrigation systems, increased participation by farmers and State departments (especially the Departments of Agriculture and of Tribal Welfare and the Water and Land Management Institutes [WALMIs]), and several major policy and practice changes by the Government of Madhya Pradesh (GOMP) aimed at creating institutional capacity and replicating the project in non-project areas. Of particular importance was the adoption, in June 1988, of a performance-based disbursement (PBD) mechanism, which improved implementation significantly: while only two MISs had been completed by 1990, by the PACD 20 MISs had been completed, with another 26 under construction. The GOMP plans to have 15 of these completed by June 1993 and the other 11 by June 1994. The project's four pilot activities produced major technical improvements/innovations. (1) The scheme at Khor introduced buried polyvinyl chloride (PVC) pressurized pipe technology (PPT) as an alternative to conventional surface channel network below the outlet; similar pipe systems are being extended to non-project MISs, including 10 planned for construction in 1992-93. (2) The Gadigaltar subproject introduced, with farmer participation, a major innovation -- an "on-demand" flexible water supply system utilizing a level-top canal in addition to buried low pressure pipes. (3) The Raipura scheme demonstrated the impact of land development on irrigation efficiency and agricultural productivity. (4) The Ghorapachhar subproject served as an adaptive research site for the GOMP's Irrigation and Agriculture departments, and for the WALMI at Bhopal. Performance testing of the ability of the new irrigation systems to deliver an adequate and timely supply of water (a practice first introduced in the Maharashtra Minor Irrigation Project and required in the revised benchmarks for this project) was reassuring to farmers and resulted in their immediate subscription to the new systems. Also, through use of a "bottom down" approach, the project realized (as did the Maharashtra and the Hill Areas Land and Water Development projects) the previously elusive goal of involving command area farmers in opting for construction of their respective systems. Reliability of water supply, another long-standing problem, as achieved by basing reservoir capacity on the extent of demand given a variety of cropping patterns and dependable climatic data. Also contributing to project success was the increased responsiveness of the GOMP to project innovations and the increased involvement of the private sector. Prospects for sustainability are good. Institutional capacities have been upgraded at various levels in the GOMP; 53 officers were trained in the U.S. and 2,246 locally in water management and irrigation related issues. The GOMP has implemented a policy shift away from administering water to managing water per farmers' needs and has developed a strategy for increasing farmer participation in decision-making and irrigation system operation and maintenance. Finally, the GOMP Water Resources Department has adopted the irrigation system design criteria developed under the project for the entire Madhya Pradesh minor irrigation sector. Lessons learned included the following. (1) The project demonstrates the value of interim redesign and revision. At the time of initial design, it can be difficult to ascertain a project's complexities. (2) Criteria for measuring performance should be quantified in a way that is appropriate and practical for the implementing agency. (3) Targeting of underserved and remote traditional tribal areas, while commendable, had anthropological implications that project designers failed to consider sufficiently. Specifically, the design did not call for sufficient involvement by social scientists and as a result farmer organizations were not as successful as they could have been. (4) Delays were largely caused by lack of technological capacity in both the public and private sectors, as well as by problems in land acquisition and by changes in GOMP personnel, chronic to India's bureaucratic system.