The call for using farmer knowledge in designing appropriate research, development and extension strategies has become common in recent years (eg Richards, 1985). However, in practice involving farmers in research is generally limited to them participating in questionnaire surveys and providing lists of 'indigenous technical knowledge'.

This paper argues that this approach misses much of the deeper understanding developed by farmers as managers of their farming system and local environment. Understanding farmers' strategies allows a framework for posing technical, scientific questions in research. It also provides the basis for evolving development strategies that are not imposed as alien 'packages' that contradict with existing practices.

I illustrate this theme with an example from research carried out in Zvishavane District, southern Zimbabwe. It illustrates how an understanding of farmers' livestock management strategies, their environmental classification and their understanding of ecological processes led first to the questions of a scientific study being posed and secondly to the answers and their development implications being understood.

The questions relate to how livestock utilise different habitat patches in a dryland environment. This includes the use of two different ecological zones and, within these zones, the use of 'key resource' patches for providing fodder. This study leads logically to recommendations for grazing land and livestock development that, in certain respects, conflict with standard recommendations emanating from a process of research and development divorced from local social and ecological context.