The already serious food security situation in semi-arid areas of Kenya has been rendered more precarious through the implementation of economic liberalisation measures as part of structural adjustment. Household food insecurity is almost an endemic problem in the semiarid areas of Lower Embu and Tharaka-Nithi Districts of eastern Kenya. The Dryland Applied Research and Extension Project (DAREP), supported by the UK Department for International Development carried out adaptive research in these districts using participatory methods. During informal diagnostic surveys conducted between November 1993 and May 1994 food insecurity was identified as a priority problem. It subsequently became a central theme in the project's research programme. Three specific aspects of household food security were identified . First, even in good years when most households produce sufficient quantities of basic food crops, they still experience a hungry period because they sell some of their output in order to raise cash for other needs. Second, all households lack green vegetables during the dry months, which cover about seven months of the year. Third, periodic droughts create a crisis of hunger for all households. This paper describes the process and methods used for diagnosing food security problems; analysing the causes offood insecurity and the coping strategies which farmers have been using; and planning a range of adaptive research interventions designed to address the priority food security problems. For each problem related to household food security one or more research strategies were outlined and for each strategy several possible activities identified. Fifteen out of sixteen identified research activities addressing the food security problem have been implemented by the project. An implicit assumption of the research approach was that improvements in household food security would be best achieved by improving the overall livelihood base, rather than by concentrating solely on food-specific technologies. The adaptive research intervention strategy was therefore broad based and aimed at increasing the range of technical options available to farmers. Case studies document progress with some interventions, including research on early maturing and drought tolerant varieties of existing food crops, the introduction of new food crops and improvements in animal health. Lessons learned so far are highlighted and the implications for policy and future research are spelt out.