||Area based development programmes aimed at improving the incomes of small farmers are a major concern of governments in less developed countries. New technology is the foundation of these prograrmnes and the selection of this technology is a critical choice, often badly made. Institutional and infrastructural developments,'which may include roads, marketing, credit and extension services, water supplies and processing facilities, focus around the needs ,of the selected technology. If the wrong technology has been chosen for a programme, that is if farmers do not adopt it, not only has the research effort expended in producing the technology been wasted but a whole range of development resources have been misapplied. Scarcely less critical is a secon~ tvP? nf failure. The selected technology may have a social or economic impact contrary to government policy objectives. For example, if a 'programme intended for the mass of the small farmer population is only absorbed by the largest farmers it increases the income gap between rich and poor that it may have been government policy to reduce. Both types of failure emphasise the need to ensure that the right technology is selected for development programmes; that is technology that is acceptable to, and within the reach of, the population of farmers towards which the programme is directed. The classic small farmer problem is the need to deal with large numbers with limited funds and manpower. Neither policy formulation nor research and development programmes can be designed for individuals. How can small farmers best be grouped into target populations for which the same new technology will be relevant? Groups should have the same researchable problems and the same development opportunities, and should react in the same way to policy changes. Effective grouping depends on identifying the factors causing variation between groups of farmers and adopting a method of classification which effectively weights the influence of these factors on farming. Two types of division can be identified; a locational division by area and a hierachical division between farmers in the same areas. Three sets of factors are identified as contributing to these divisions: 1. Natural factors; climate, soil, topography. 2. Historical factors; food preferences and social custom, present technology and tenurial arrangements. 3. Institutional and economic factors; access to markets and access to inputs. The second and third sets are relevant to both locational and hierachical divisions, the first to locational divisions alone. Differences in any of these factors may be sufficiently significant to cause differences in the researchable problems and development opportunities relevant to subpopulations of farmers. Several grouping techniques have been widely used to help overcome this universal small farm problem. Administrative units such as Districts are often used as a base for action programmes. Such units do not, except fortuitously, "reflect differences between farmer groups. Agro-ecological zones have been developed in which climate and soil conditions and therefore crop growth characteristics are similar. Such zones only differentiate on natural factors, they omit economic, historic and institutional criteria. There are many situations in which ecomic and institutional factors overwhelm natural influences. For example, because of the market and transport problems poed by their perishable nature vegetables will be grown close to urban markets, even though climate and soil is more suitablefurther away. Similarly Land Use Zones only reflect a partial picture. In areas where land is plentiful the relatively scarce factors of labour capital dictate production patterns, ano the emphasis on land use as the criterion for grouping is inappropriate. The most satisfactory approach to grouping farmlrs with the same problems and possibilities is on the basis of tteir existing farming system. 1. The farming system is a manifestation of a weighted interaction of natural, economic, historical and institutional factors influencing farmers' decisions. It thus reflects the balance of those factors important in identifying homogeneous farmer populations. 2. the existing farming system is the starting point for development, the base on which productivity improvements have to be grafted. All farmers, but particularly small farmers with low risk thresholds and low capital availability, move in small steps away from their existing, tested, farming system. The prime task of research is to identify such small steps; changes which are consistent with farmer priorities, which are within their risk ceilings, within their resources capabilities and which are profitable for them. The prime task of a development programme is to. pass these on to the farmer. The choice of possible steps can only be evaluated in the context of the farmers' existing system which reflects their circumstances and priorities. Grouping farmers on the basis of their existing farming systems identifies target. populations for which the same research effort -and the same development programme, based on the new technology emerging from the research, will be appropriate. Cimmyt terms such groupings Recommendation Domains.