On-station trials began to consider the comparative performance of the improved cultivars in
these various crop associations. Trials of potatoes intercropped with maize in 1987 showed that
land equivalent ratios increased with increasing plant densities, even when plant densities of
associated crops were those normally used in pure stands (7.2 potato plants and 8.0 maize plants
per square metre; preliminary results are reported by Jeroen Kloos in the 1987 CIP regional
progress reports). Trials to test the performance of field mixtures of improved potato cultivars
were also recommended, after farm surveys showed that most farmers grow three to five different
potato cultivars at once, most of their fields being planted with cultivar mixtures.


 Participatory Breeding Research

It is easy to assert that defining appropriate varietal screening priorities and experimental
conditions require frequent and direct communication both between farmers and researchers and
between researchers of different disciplines (economists, anthropologists, breeders, agronomists,
phytopathologists, soil specialists). Few biological scientists, however, are trained in techniques
to elicit and to apply knowledge from farmers (Richards, 1985; Brokensha et al. 1980).

 Although it sounds straightforward for scientists to learn from farmers, and to convene groups
 or panels or innovator workshops, how to do this is rarely part of scientists' training, and good
 methods are anyway not well known. Nor has discussion of such methods penetrated the harder
 professional literature (Chambers and Jiggins, 1985).

After a decade of rhetoric about feedback of farmers' problems to extension workers and
scientists, a large gap remains between the ideal and the reality., Innovations in both training and
methods are required to bridge this gap. To the usual on-station and on-farm trials, and formal
and informal surveys must be added less familiar approaches such as panels of farmers who
regularly meet with and advise scientists, one-shot group interviews, the training of scientists in
role reversal, workshops with innovative farmers, and village meetings in which farmers decide
on the design of on-farm trials. Farmers included in the design of on-farm trials can "contribute
to defining evaluation criteria, before researchers [have] screened out most of the options by
fixing the experimental design" (Ashby, 1986). When setting up on-farm variety trials, scientists
can begin by asking farmers how they themselves would test a new cultivar on their own land
(Biggs, 1988). In addition, researchers can track farmers' own innovations, which take them
beyond the limitations of reductionist methods of on-station trials, as they adapt new cultivars to
complex intercropping, rotation and agroforestry practices, and as they exploit diverse microen-
vironments (Chambers and Jiggins, 1985). Large-scale formal surveys, with their well-known
problems of data reliability, sampling biases, logistical costliness, and lengthy processing
requirements, are also increasingly replaced by less formal and more innovative techniques. In
such attempts, team work, rather than 'lone ranger' research (Robert Rhoades' term) increases the
credibility of results.


Farmer Participation in On-Station Germplasm Screening

Normally when on-station germplasm plots were harvested in Rwanda's national potato research


GATEKEEPER SERIES NO. SA30