ho is better to judge which variety of wheat or barley best suits a particular area or climatic zone than the farmer with years of hard earned experience and knowledge of his fields and his crop?
        The dangers of remotely-based plant breeders and scientists imposing the new crosses they think are ideal have long been realized at ICARDA which tries to work closely with local partners when helping improve yields of cereals and other crops. Whereas research-station-based plant breeding is often quick to reject the results of some crossings as being unsuitable, the farmer approach can be more considered and pragmatic.
         A project funded in northern Syria by Germany's Federal Ministry for Economic Cooperation (BMZ) has underlined the importance of allowing farmers access at an early stage to the considerable genetic variability created by plant breeders in the search for new varieties. It is important for adapting varieties to differing environments and for the maintenance of genetic diversity. Barley growers themselves favored many crossings using germplasm that was being discarded by a plant breeder.
         At the start of any breeding program, the scientists begin with a wide pool of genetic variability in the form of germplasm. In the past this has been narrowed down through field selections of crosses to produce improved but closely related cultivars which can be grown over a wide geographical area without too much variation in yield or performance. In the process of narrowing the selection down, many potential varieties are discarded, perhaps because they have advantages which are too specific to a particular area or growing conditions to be of wider use.

56% (Ebla). Six-row types are generally more lodging-resistant--a useful trait in wet environments.
        Dry and wet locations also influenced whether or not the landraces stayed in the farmer selections and the breeder shared this divergence of opinion over their value at Tel Hadya (wet) and Breda (dry). Farmers at Ebla agreed with those at Tel Hadya that the
landraces and the black-seeded types were unsuited to their wetter conditions and should go.
         In dry sites the frequency of black-seeded types increased almost two-fold during the two cycles of selection. However, local preferences for or against particular colors did also create a couple of extreme selections of one or other.
        The breeder and the farmers were able to agree on some traits. In their own fields and at Breda, farmers went for taller plants and higher grain yield--as did the breeder--while at Tel Hadya the farmers and breeder agreed on selection for lodging resistance, higher grain yield, larger kernels and shorter plants. Later heading ability was also used by the breeder alone to select types at Tel Hadya.
         Nor are the farmers slow to pick out lines that yield significantly better in their own fields. They were just as good as the breeder at selecting the high yielders at Breda, but at Tel Hadya the breeder selections always outyielded significantly the population mean. Farmers got it right five times out of nine, and it must be realized that the real issue in participatory plant breeding is whether the farmers' selections in their own fields are reliable and not whether they can select efficiently in the unfamiliar environment of a research station in a high-rainfall area.

          One of the expected results of farmers selecting for specific adaptation is that, even in a relatively small geographical area, this decentralized participatory selection should produce many different cultivars instead of the few bred conventionally and selected centrally.
          The BMZ-funded project, which is continuing into 1999, demonstrates that bringing farmers early into the equation does indeed affect genetic diversity. Two cycles of selection--in 1996/97 and 1997/98--were run at ICARDA's Tel Hadya and Breda research stations and on farm sites. Starting with the same initial population of 208 barley types, a professional breeder and nine farmers were given access to the same mixture of germplasm types differentiated by being two-row or six-row, modern or landraces, fixed or segregating, or varying in seed color. On one location trials were conducted both on barley-barley and barley-vetch crop sequences.
         Including varieties known to be unsuitable for the trial locations helped stimulate farmer interest in the project by providing a talking point for the 'judges.'
         Where farmers were involved there were twice as many entries left after two cycles than for breeder selection alone at Breda, and three times as many as remained when the breeder was left to work alone at Tel Hadya (See table). The breeder eliminated all the landraces and black-seeded types at Tel Hadya and all six-row types at Breda.
         Some germplasm types also disappeared when the farmers took part but they were of different types in differing locations. Six-row barley was rejected at all nine village sites except at Ibbin and Ebla, both categorized as 'wet' sites where the frequency of six-row types rose from 24% in the original population to 50% (Ibbin) and

        The project has also demonstrated another advantage of including growers in the selection of barley varieties which may have to fit in with changing agronomic practices or rotations. The top 5% of the lines in a barley-barley rotation were all different from the 5% most preferred types in the barley-vetch trial. One line which scored a 54% preference in the barley-barley trial came just seventh in order of preference in the barley-vetch selections. Similarly, the most preferred lines from the barley-vetch trial scored no higher than 26% with farmers growing barley followed by barley.

         This shows that by including farmers as partners in breeding, it opens up the possibility of rapidly adapting breeding material to the changes in agronomic practices and farming systems of the target environments. It is very difficult and very expensive to incorporate differing rotations and agronomic practices in a centralized breeding program. By decentralizing plant breeding, there is no additional cost--only additional benefits--provided farmers are included among the participants.


        Dr Salvatore Ceccarelli is Barley Breeder in the Germplasm Program at ICARDA.

.