There is a burgeoning literature on what has become known as 'farming systems research' (FSR) I Norman 1978, 1980; Gilbert, Norman and Winch 1980]. It can be seen as a response by the agricultural research community, and particularly its social scientific element, to repeated criticisms of new agricultural technology that was ill-suited to the requirements of certain types of user, especially small farmers: high yielding varieties that were unacceptably risky [Lipton 1968]; monocrop recommendations inappropriate to mixed cropping systems ]Norman 1972]; input-intensive packages where inputs were expensive, inaccessible or uncertain [Griffin 1974, Dasgupta 1977]; innovations that did not fulfil in the field the promise of the research station [IRRI 19771and so on ]Bunting 1982; ICRISAT 1980].

Despite some differences in emphasis, most major agricultural research institutes are now moving towards similar FSR methodologies that are distinguished from traditional methods by two redeeming characteristics: in the first place, they begin with the complete farm as an integrated system and move towards specific changes in management practices only when the detailed context of constraints and opportunities has been subjected to multi-disciplinary scrutiny; and secondly, they place the farm family firmly at the centre of the research and development process with the tasks of helping to define priorities and of managing system-specific onfarm trials I Byerlee et al 1979. Byerlee, Collinson et al 1980, Shaner et al 1981, Zandstra 19801. Although the new FSR methodologies have yet to be subjected to the acid test of widespread application outside the international research institutes, they do appear to offer a much stronger framework than has previously existed for multidisciplinary, farmer-oriented and relevant research.