programme, the entire research team (breeder, agronomist, pathologist, anthropologist) were
present to select by consensus genotypes to keep for further stages of testing and multiplication.
In a novel initiative, farmers were invited to make their own selections from the station fields, and
to explain their reasoning to scientists. Researchers found that they had previously assumed too
narrow a range of local acceptability in traits such as tuber colour, size, shape and uniformity.
Whereas some researchers, for example, had for years favoured red-skinned clones, farmers
found red, white and pruple skins all to be acceptable. The only skin types farmers strongly
rejected were russets, which they believed to be diseased.

In other words, the scientists were more conservative than the farmers, and their misconceptions
led to unnecessary rejection of some potentially useful potato germplasm. Formal farm surveys
of existing varieties and preferences confirmed these findings. Incorporating farmers into on-
station germplasm screening can produce useful information at little cost.

Participatory research, then, can become a two-way flow that both takes scientists to farmers'
fields and brings farmers to the scientists' fields. CIAT's bean research programme in Rwanda
subsequently adopted this approach. Female bean seed experts now participate in on-station bean
varietal assessment (Sperling, 1988). Women farmers (since they rather than men tend the crop)
visit on-station germplasm trials at two or three critical points in bean growth (at flowering and
formation of pods, at maturation, and at harvest). Also valuable to both the scientists and visiting
farmers are the observations of station field labourers (themselves usually small farmers) who see
the scientists' trials through the entire crop cycle. In the Rwandan potato research programme,
local scientists knew that some station labourers were both very keen observers of experimental
germplasm, and experimenters with promising plant material on their own farms. These labourer-
farmers were among those who assessed potato germplasm in the exercise mentioned. This
technique is a useful complement to farmer-managed trials in farmers' fields.



 Recommendations

Plant breeders cannot respond to every quirk of farmers' circumstances. Their task becomes more
complicated, costs increase, and progress slows as the number of selection criteria increases.
Breeders require general guidelines based on accurate prior identification and ranking of cultivar
traits that particular categories of producers and users find important, discarding less relevant
screening criteria, and assessing farmers' capacities to change existing practices. Crop breeding
is a long-term investment; decisions taken at the outset have implications for many years to come.

If farmers in Africa, Asia and Latin America are to influence agricultural research more directly,
researchers and extensionists need better incentives and improved ability to address farmers'
needs. Skills to bridge the social distance between 'authoritarian' scientists and 'deferential'
farmers are essential, so that "when farmers experiment with low fertilizer applications to find out
what works and pays best for their conditions", researchers will see them as experimenters rather
than as "deviants who do not adopt recommended practices" (Chambers and Jiggins, 1985; Ashby,
1986).

Social science skills are often underutilised in the design and analysis of on-station and on-farm


GATEKEEPER SERIES NO. SA30