This paper reports on a study, commissioned jointly by the Evaluation Department (ODA-EvD) and the Research Task Group (RTG) of ODA, to review a selection of renewable natural resources (RNR) research projects in order: to identify features common to projects generating successful/unsuccessful uptake and impact; to identify the features which might in future be used as indicators of successful uptake and impact; and, to make recommendations for the improvement in the procedures for management of research projects. A total of 21 projects was examined, of which 8 had earlier been comprehensively reviewed by ODA-EvD. The study team was asked to select the remaining 13 projects from those undertaken by the ODA-supported Natural Resources Institute (NRI) in consultation with ODA advisers, NRI staff, the RTG and ODA-EvD. An effort was made to achieve a balance between those projects commonly perceived to have been successful, and those which were not. Information was assembled by review of project files, and through semistructured interviews with project staff and key informants. An analytical framework was constructed, which took account of recent literature which stressed: (a) that uptake is strongly influenced by the quality of project design and management; (b) that research often fails to produce useable outputs because of inadequate familiarity with the socio-economic and natural resource conditions facing intended users, and inadequate participation by users in the prioritisation, design, conduct, review and dissemination of research; (c) that strong "user constituencies" are an important prerequisite for generating responsiveness by researchers to users' requirements, that such constituencies should comprise a range of users (farmers; processors; NG0s; parastatals; commercial organisations) appropriate to the requirements of the specific circumstances, and that conventional notions of "institution building" relying heavily on staff training are unlikely to generate adequate productivity and responsiveness in the absence of user constituencies. The analytical framework also drew on a review of concepts and recent experience in the application of the Project Cycle and Project Framework to RNR research projects which, because of the greater uncertainty attaching to them, have management requirements which differ from those of other RNR projects. The analytical framework generated a number of key questions, at each stage in the progression from project preparation through research to uptake and impact, which were used as the basis for the design of section headings in the narrative assessments of individual projects, and for the classification of information in summary form. An important distinction was drawn between the intended end-users of research output (eg farmers, but in some cases also the commercial sector, NGOs and governments) and intermediate users (generally other researchers). Various channels of dissemination were also identified (eg extension services; NG0s) which transmitted the research output essentially without modification. Factors were identified which were associated with the successful production of research outputs and their uptake, and which inhibited the successful production of research results and their uptake. Practically all the shortcomings which related to the Project Cycle were already being addressed through the tighter project management procedures, including the preparation of detailed Project Memoranda and Project Frameworks, which have been introduced since the late 1980s. No general recommendations in these areas were therefore made. However, a number of specific recommendations were made in relation to continuing shortcomings in the current use of Project Framework and Memorandum. Further recommendations were made on a number of matters related to the arrangements for resourcing research and the management of research, the organisation of research, training of research staff, and the arrangements for reviewing research programmes.