plant breeders to small farmers' circumstances.


Traditional western agricultural curricula, for example, discount inter-cropping, though many
plant scientists today recognize that insufficient research has been done on possible positive
interactions of species and cultivars planted in mixtures (Altieri, 1985; Willey 1979). Comple-
mentary effects involving the uptake of soil nutrients or water, for example, are poorly
understood, as is the degree to which crop and cultivar mixtures may slow the spread of pathogens
and pests. Yet intercropping research in Africa is often considered a retrograde step.

Attempts in the last decade to institutionalise processes for agricultural researchers (both national
and international) to learn directly from farmers, and for farmers themselves to do more than react
to scientists' proposals have been dominated by various types of farming systems research (FSR)
(Byerlee et al., 1982; Collinson, 1985, 1988; DeWalt, 1985; Horton, 1986; Merrill-Sands, 1986;
Norman et al. 1982; Rhoades, 1985; Eicher and Baker, 1982; Hildebrard, 1981). Though the term
FSR itself has become controversial, its basic principles are of lasting importance. These include:

- the need for close collaboration among technical scientists (both physical and biological) and
 social scientists;

 the usefulness of multi rather than single-commodity approaches (since farmers themselves
 pursue multiple enterprises and evaluate technical innovations in any one crop in the context
 of the systems they operate);

 and explicit recognition that the farmer and other agents in the food system are the primary
 clients of agricultural research, and that farmers' current production systems must be
 understood in order to design and assess on-farm and on-station experimental programmes
 intended to improve production.

The less effective alternative has been for researchers to seek the optimal way to grow crops and
to expect farmers to adjust to these requirements. When scientists' selections of new crop varieties
are based solely on features of the natural environment (such as rainfall, soils and temperatures),
farmers may reject the high-yielding varieties scientists most admire. More than maximum yield,
African cultivators often favour yield stability, short maturation periods, suitability for intercrop-
ping, storability and particular taste or cooking characteristics.


How African Farmers Use Cultivar Diversity

Breeding programmes have rarely exploited small farmers' sophisticated knowledge of differ-
ences among cultivars, and their use of these differences in cropping strategies. Cultivators
classify varieties, and value particular characteristics, for different purposes. They often manage
a combination of cultivars in the production process, and multiply or eliminate varieties as they
evaluate their performance over time (Brush et al, 1980; Conklin, 1988).

Farmers themselves are expert experimenters with new plant materials (Johnson, 1971; Ninez,
1984; Rhoades, 1987; Richards, 1985). When testing promising new plant genotypes, scientists


GATEKEEPER SERIES NO. SA30