Agroforestry development programmes frequently rely on knowledge from a number of different sources. In particular, there is a growing recognition amongst development professionals of the value of augmenting partial scientific and professional understanding with the detailed knowledge held by local people. Taking advantage of the complementarity of local, scientific and professional knowledge demands the development of effective mechanisms for accessing, recording and evaluating knowledge on specified topics from each of these sources. The research described in this thesis developed a methodology for the acquisition, synthesis and storage of knowledge. The defining feature of the approach is the explicit representation of knowledge. This is achieved through the application of knowledge-based systems techniques. AKT2 (Agroforestry Knowledge Toolkit), a software toolkit developed in Prolog, an artificial intelligence programming language, provides the user with an environment for the creation, storage and exploration of large knowledge bases containing knowledge on a specified topic from a range of sources. The use of diagramming techniques, familiar to ecologists and resource managers through systems analysis, provides an intuitive and robust interface. This knowledge-based system drives incremental knowledge acquisition based on an iterative evaluation of the knowledge bases created. The iterative approach to knowledge acquisition provides a coherent, consistent and comprehensive, and therefore more useful, record of knowledge. Once created, knowledge bases can be maintained and updated as a record of current knowledge. Techniques for the exploration and evaluation of the knowledge base may be useful in : " giving research and extension staff access to a concise and flexible record of the current state of knowledge; " providing a resource and mechanisms for use in planning and prioritising research objectives; and " providing a resource and mechanisms for the generation of extension materials tailored to the needs of particular clients.