DeveloEing Local Organizational Capaci/)' ror Participalory Seed Managemenl • Finally, the success of community action to manage development processes will depend fun- damentally on the community' S ability to control the processes ofknowledge production, de- sign, and implementation of actions. The practícal ímplícations ofthís methodology can be summarized as the need to search for ways in which partícipatory research can be part of an ongoing process. Inherent to the process is the acknowledgrnent that power relations between researchers and the researched is problematic and that there ís a need to develop a process of critical reflection that situates the production ofknowl- edge and action withín a specific context of a negotiated process, emphasizing community actíon (see also Koning and Martin 1996). The setting The major ethnic group ínhabiting the research sites is an ethnically distínct but heterogeneous group ofpeople known as the Raí. Together with a related group of people known as the Limbu, the Raí refer to themselves as Kírals, a term employed as much to unify aH the various "tríbes" and c1ans as it is a political statement employed to dístinguísh them from the dominant Hindu majority. Having until the recent past practiced a distinct system of cornmunal land tenure known as Idpat, the Kirats constitute one of the oldest ethnic components of the regíon. Yet in decades following their integration into Nepal after the "unification" in the mid-18th century, the Kírats have been confronted with numerous chaIlenges to their traditional way of life. Dominant lowland influences have resulted in changes in sociocultural practices associated with traditional land-management practices and given rise to the ubiquitous rain-fed and irrigated terraces (bari/khét) that suil wetland paddy and other lowland crops. In the process, engineered landscapes have replaced extensive areas of forest cover where traditional swídden (slash-and-burn clearing) was practiced. Compounding the asymmetry ofhistorically derived center/periphery relations are constraints im- posed by the harsh mountaín environment. Typical ofthe eastem Himalayan region (see Shrestha 1989), human sertlements are siruated in elevations ranging from 500-2000 melers, where land- distríbution partems combine with steep slopes and shallow soil depths lo severely constrain agri- cultura! activities. The land-distríbution figures ofTamku VDC (table 1), where the research sites are located, demonstrate the environmental constraints that the inhabitants are confronted with. From the total avaílable Jane!, ónly 10.6% is suitab!e for agriculture, and from this total arable area, 54% has slopes of 40 degrees and soil depths of not more than 20 cm (Goldsmith, 1982). Asymmetrícal center/periphery relations embedded in historical processes have contríbuted signif- icantly to the present deteriorating ｳｴ｡ｴｾ＠ of local institutional capacities to negotiate and orlen! Table 1. Land Clusification ofTamku VDC AgricuHurallands 10.6% Grazingl.ands 14.6% Shrubs 7.8% Deciduous forests 35.6% Subtropical forests 10.8% Rack ice 20.6% Source: Khan.1 (1992). 150 B. Gurung and P. Gurung development servíces to their benefit, espccially to counter the period of food deficít that typically lasts for four to five months ayear. Unable to support their subsistence needs through crop yields alone, many households have male family members migrating in increasing numbers to urban cen- ters in search of employment, leaving women and children to manage and care for the farm. An ad- ditional outcome of prolonged periods of food deficit is the inabilíty of households to save seeds from consumption in times of stress. Thís, along with deteriorating local knowledge about seed-management practices and the absence of organizational capacíties to access external sources of improved seed technologies has profound implícations for the long-term subsistence of house- holds in the region. It also significantly determines the narure and type of research methodology to be adopted for particular sites. The research process: An interactive methodology The objectives of the project evolved in several stages of a diagnostíc process that sought devolu- tion by emphasizing cornrnunity participation in íncreasing stages during the research process. In order to facilitate cornrnunity control and ownership, the methodology was developed &om the principIes of problem posing, dialogue, and refleetion based on the Freirean (1972, 1973, 1978) notion that cornrnunity ínvolvement in the development process can be generated through develop- ing a critical awareness of the causes of problems. The diagnostic process ínvolved the following steps: l. A survey was conducted to establish the need for a participatory seed-management initiatíve, based on the following research themes: • assessing the capacity oflocal cornrnunity-based organízatíons • determining existing pattems of food sufficiency • identifyíng appropriate crap(s) for enhancing improved seed-management strategies • determining factors for farmer participation through gender-differentiated varietal assess- ment ofidentified crop(s) • determining the source of germplasm, either in exísting local vaneties or through external means 2. Analysis was done through a critical examination ofbaseline data to determine how the prob- Iem of food deficit i5 contexrualized by cornmunity members. Ibat ís, are problems of food def- icit línked to just econolTÚc issues of subsistence or are they affected by social dynamics of decision-making? And to what extent are these embedded in tbe value5 and cultural constructs of the cornrnunity? Conceprualized problems in this way necessítates posing the following questions: • Do the issues deal mainly with problems of subsistence, decision-making, or values? • Where will action most likeIy come from? • What will most effectively motivate people? 3. Problem-posing material was prepared through the development of codes, which are represen- tations of existing problems in the form of stories, dramatized enactrnents, pictures, results of particípatory rural appraisal (PRA), etc. Fundamental to the preparation of codes is tbe need to ensure that they present a scene showing a concrete experience ofthe problem, which is famil- iar to the participants. 151 Developing Local Organizational Capacity [or Participa¡ 4. Discussíon was dírected through an interactive workshop whereby cornmunity members par- ticipated in defming tbe problem offood deficit and searchíng for solutioos. The primary objec- tive ofthis process was 10 develop a critical awareness oflhe problem offood deficit through tbe search for potential solutions. Additionally, the process also creates a context for the com- munity 10 provide cornmenls on the research results and to define the direction oflhe process. The process begins witb a description ofcedes, followed by a firsl analysis, which is then re- lated to reallife and followed by a deeper analysis, ending in self-reliant action planning. Farmer participation in the research process The degree and type of farmer participation depends principally on the objeclives for participation, as well as the context, as determined by the particular stage of tbe process. Thus, the diagnostic phase, consisting oftbe survey, anaiysis, code preparation, and discussion, involved varying levels of farmer participation. In tbe survey, three members of tbe cornmuníty and two projeet members comprised the researeh team. Clan elders and farmers selected on the basis tbeir knowledge related to seed managemen! were consulted abeu! tbe relevance of tbe project. In addition, tbe executive body of cornmunity-based organizations were consulted to establish interest in developing a work- ing partnership to conduct the project. The survey was conducted to establish (1) a crop inventory, (2) to determine tbe needs and priorities of different groups, based on gender and wealth considerations, and (3) lo identify erop for improv- ing seed-management technology. At tbe same time, farmers were selected for consultation on the basis of tbeir knowledge, financial.status, and gender. The subsequent analysis of the data lo de- velop appropriate cedes was conducted in collaboration witb local researchers and farmers. The main objective ofthe workshop !ha! followed was to present tbe codes to tbe larger cornmunity Table 2. Types of Farmer Participatíon A B e D Survey x Analysis i x Code preparation x Discussion x Souree: Adapted from Biggs (l989l Note: A = contractualj B = consultative; e ｾ＠ collaborative; D = collegiate. to understand tbe root causes and potentíai solutions to problems of food deficit in the regíon. The selectíon of cornrnunity members was based on tbe eriteria developed in prior consultation witb local members of the research team. During this stage of tbe interface, farmers were more exten- sively involved in tbe direction of!he discussion of research findings, as well as decís ion making lo determine tbe level of participation in setting tbe agenda for future action. User differentiation The selectíon of participants was deterrnined by tbe following eritena: 152 S, Gurung and p, Gurung • demonstrated instances of innovation in seed management and knowledge of causallinks be- tween problems of food scarcity and gaps in existing seed-management practices • gender-differentiated knowledge and gendered experiences • farming for subsistence as a full-time subsistence activity Innovatíon The participants selected for participation in the research process demonstrated varying degrees of innovation in crop management. The type of innovations ranged from pre-harvest selection prac- tices to post-harvest storage practices. In sorne instances, the praetices were leamed from experi- enee gained extemally.asin the case of selecting for desired traits of rice during the pre-harvest period or experimentíng with new strategies as in the case of post-harvest storage of maize mixed with millet to reduce pest attack. While post-harvest selection practices were common for crops such as maize and millet, pre-har- vest selectíon was practiced only on paddy, One farmer, selecting specifically for larger panicles, denser grain quality, and tal! height in a landrace (punche dhan) was successful in producing a "va- riety" subsequently named afier him (changkhu dhan,literally "Changkhu's rice"), This "varieti' is currently widely adopted by other farmers in the commnnity, with Changkhu presently selecting for early maturation to coincide with the planting of winter wheat. In seed-storage technology, sorne innovative farmers experiment with the leaves of a locally avail- able plant (bajo) to ward offpestattacks on maize seeds. Dried leaves ofthis plant are placed in the bottom of the seed container and altemately in several layers approximately every three to four inches, then fue container is sealed by additionalleaves at the topo Sealed in September or early Oc- tober, the relatively airtight spaces and the toxic nature ofleaves sufficiently wards off pest attacks. In another example, one woman farmer, noticing that millet grains were free of pests that attacked maíze seeds, began mixing a handful of millet grains in the container where maize seeds were stored. This relatively simple practice was based on her observation that millet seeds were free from fue pests lhat attacked ｴｨｾ＠ maize seeds that were stored in close proximity to fue millet. Knowledge and gendered experiences In varietal assessments of maize, conducted separately between women and men farmers during the inÍtial research phase, women and men listed different categories of preferences based on their roles and experiences. Men listed four varieties of maize, mostly modem varieties that had been intro- dueed into the commnnity in the last several years. W omen, on the other hand,listed eight varietíes, mostly landraces whose use had been discontinued in the project site but existed in the women's na- tal villages. Women cited fodder quality, ease in grinding, and taste as fue primary eritecia for their preference oflandraces. Men, on fue other hand, cited high yields, early rnaturation, resistance to drought conditions, and market prices as important in their preference for modern varieties, An ad- ditional ranking of maize varieties among farmers revealed differential knowledge and preference priorities between women and men (table 3). Farmingfor subsisten ce That participating farmers be involved in farming as a full-time subsistence activity was an impor- tant eriteria for selection for two reasons: the first was prompted by fue project need for the uninter- rupted involvement of participants for two production seasons (for most farmers in the area, 153 Developing Local Organizational Capacitv (Or ｐＮＺＺＺ｡ＢＬｲｨＢＬﾷ｣ＢＬｩｰＢＬ｡ＺＮＺＺｴｯＢＬＭｲｶＮｲＮＮＮＺＺＺＮｓ･ＢＬ･ＢＬ､ＭＢｍｾ｡ＺＺＺｮ｡］ｧ＼＾Ｚ･ＺＺＺＺｭＢＬ･ＢＬｮＬＭＭｴ＠ __ ｾ＠ ________ _ Table 3. VarietaI Knowledge and Preference Ranking of Maize for Men and Women Women Men 1. bhote' pahelí 1. manakamana-l (MV) 2. pahelí 2. dhude' selí 3. dudhe' selí 3. paheli 4. bhole' selí 5. lamlunge' seti 6. arun-2 (MV) 7. manakamana-l (MV) 8. chepti seti .. -"----1 _ food-scarcity periods necessitated involvemcnt in off-farm activities for supplementing household ineomes); the second was because those farmers who were involved in farming as a "full-time" activity showed a greater inclinalion to be relative\y self-sufficient in food production, even during the scarcity período Of the nine farmer participants in Tamku VDC who were included in the "iuno- vative" category, aIl claimed sufficient food security during the year and could be counted upon by other cornmunity members for food loans during periods of food deficit. Out-migration ofmen to urban centers in search of employment is one ofthe primary strategies em- ployed lo counter food deficits. In the past, it was cornmon for men and women to become involved in recíprocal arrangemenls within t/le corpmunity during times of food shortage. UsualIy this in- volved providing labor for wealthier farmers in return for food provisions during times of scarcity. Increasíngly, however, Ihe presenl trend is for!he majority ofyoung men to migrate to urban cen- lers to work as porters for trekkíng companies, perform meníal jobs in restaurants and hotels, or migrate lo Ihe MiddIe East (arab) through the numerous employment agencies !hat have sprung up in Nepalese townships. In addition to out-migration, people a1so forage for a variety offorest foods (kandamul), although a degree of social stigma surrounds foraging activities, príncipally through the perceived notion that it ís part of the "prímitive" past. At the household level, food-preparation strategies a1so play an important role in "making it lasl longer." Grains are boiled with excess water, creating a porridge-like consistency to ínerease !he quantity. "Visitors and guests" duríng !he time of scarcity are actively discouraged from visiting, though sorne women particípants cited visiting relatives (preferably fue natal home, for married women) as an option to combal food shortages. A seasonal calendar for food production reveals a period of severe food scarcity between !he monlhs begiuning in late February and lasting till early luIy. The relationship between food produc- tion and out-mígration, especially of males to urban centers in search of employment, is direct1y proportional to !he íncreasing number of female-headed households as well as the additional, "gendered" burden of farming responsibilities that this trend implies. Moreover, !here was a strong relationship between decreasing food produetion and poor aceess lO seed sources and deteriorating seed-savíng practices. Research suggested !hat !he deterioratíon of seed saving was not necessarily related lo loss ofknowledge but was, ralher, determined lO a large extent by food scarcity and the additional burden of farm households to do "other things." Increasing trends in food scarcity over the last few generations have resulted in people consuming ínslead of saving seed materíaL 154 B. Gurung and P. Gurung Though there were many reasons for food scarcity, research demonstrated a causal relationshíp be- tween decreased crop yíelds and the ínability to manage seed, in terms of both maíntaíning seed purity (saadha biyu) and poor seed storage practices. Moreover, access to the Agriculture Input Sector (AIC), a public-sector undertaking responsible for seed supplies was dífficult, sínce ít is sit- uated in district headquarters a day' s walk from the village and using il ofien proves to be a dífficult bureaucratic process beyond the reach of individual farmers. The consequenees of low yields, the ínabílity lo maíntain seed purity, and lack of access to reliable sourees of new germplasm aH con- tribute to food scarcity in Tamk:u. Lessons learned: Reconceptualizing participation and knowledge In order to address the objective of developing improved seed technologies in marginal mountain environments while emphasizing community control ofthe rnanagement of the process, it becomes important lo conceptualize farmer participation in the research process as an instrument of empow- erment. Dne principie way forward in tlús direction is lo situate farmer participation in the context ofIocal knowledge. In doing so, however, it becomes important to view knowledge, or indigenous technical knowledge, beyond common representatíons of its beíng produced as a raliona! response to environmental contingencies (e.g., Matlúas-Mundy et al. 1991; Howes and Chambers 1980; Brokensha, Warren, and Werner 1980). Instead, it becomes important to sitnate indigenous techní- cal knowledge within cultural categories of meaning, which can then become an empowering base for participation in the interface with more powerful externa! categories of knowledge. The workshop discussions revealed how empírical experiences cannot be separated from cultural experience, especiaHy in the way Rai farmers talk about food scarcity and place the phenomenon in a mytlúc context. Local discourse offood scarcity finds expressíon both in the dominant Nepali lan- guage as well as the various díalects of the Raí group. The words to describe food scarcity range from anikal (foad shortage), bhakmari (to kilI hunger), mahamari (the great killer), and sisawa (famine) in the Kulung dialect ofthe Rai. lt also finds expression through simple expressions such as "khana ka abab hunu" (to be short of edíbles), "dhayrailchitto bhok lagnu" (lo experience hun- gerpangs sooner and more frequently than normal), "chasum na hunu" (to lack prosperity), as well as more abstraet expressions, such as in tlús lament in the Kulung dialect "Etenay sisawa udanai [ay tay ha wumche " (dear friends. and brothers, ... how do we survive the sisawa [food shortage] tlús year?) or the more common instructional verse admonishing people to save seeds to combat food shortages "Almal ma jiyu bachhaunu, Aníkal ma biyu bachhaunu " (save oneself in times of confu- sion, [but] save seeds in times of[food] shortage) or "Chha geda sabai mera Chhaina geda sabai tenda" (having seeds, all is mine, [not1 having seeds, all is not mine [Le., lost]). In the indigenous schema, food scarcity is a condition of cultural "disorder" that has its genesís in the curse that one warring ancestor castes upon another for perceived treachery. In cultural terms, the condition becomes inevitable and requires annual propitia!ion of the ancestor through ritual ap- peasement. The myth, consisting of ancestral deeds tbat include the settling of present territories, serves as a metaphor for the sacred relationship that exists between the Raí and the delimited terri- tory they occupy. Traditional Kirati notions of ethnicity canno! be separated from this relationshíp and are symbolized by an ancestor stone that is sitnated in every village and propitiated in annual agricultura! ceremonies (ca:ri). 155 Developing Local Organizational Capacitl'Ji:J!J'articipatory Seed Management What such a view of knowledge irnplies is that by granting legitirnacy to cultural episternologíes, indigenous explanations for ernpírical categories are not subjugated by ratíonalist scientific expla- nations and thereby becorne an ernpowering elernent for farrner participation. Wíthín such a con- text, the transferoftechnical skills to enhance seed technology neither dirninishes nor disernpowers indigenous systerns of rneaning. References Biggs, S.D. 1989. A multiple source of innovation model of agricultural research and techn%gy promotion. Agricul- tural Administration (Research and Extension) Network, Paper No. 6. London: ODl. Brokensha, D., D.M. Warren, and O. Werner (Eds). 1980. Indigenaus knawledge systems and development. Lanham, Maryland: Universíty Pross of America Freire, P. 1972. Pedagogy af/he oppressed. New York: Seabury. F reire, P. 1973. Educa/ion for critical consciousness. New York: Seabury. Freíre, P. 1978. Pedagogy in process. New York: Seabury. Goldsmith, P.F. 1982. The lond ond soil resources olthe KHARDEP area, Koshi Hill Area Rural Development Pro- gramo Report no. 16. Kathmandu: Koshi HiII Area Rural Development Program, Howes, M. and R. Chambers. 1980. lndígenous technícal knowledge: Analysís, implícatíons and issues. In Indigenous Knowledge Systems and Development, edited by D. Brokensha, D.M. Warren, and O. Werner. Lanham, Mary- land: University Press of America Khanal, N.R. 1992. Study of geo-hydrologJ, landuse and population in the Makalu Barun Conservation Project area. MBCP Working Paper Series 14. Kathmandu: Woodlands Mountain Institute. Koning, K. de and M. Manín. (Eds). 1996. Participatory research in health: Issues and experiences. New Delhi: Vistaar PIlblicatíons. Matbias-Mundy, E" O. Muchena, G. McKierman, and P. Mundy. 1991. Indigenous technical knowledge ofprivate tree management. Ames, lowa: lowa State Universíty Press, Shrestha, T.B. 1989. Development ecology oflhe Arun River Bastn. Kathmandu: lnternalional Centre for Integrated Mountaín Development. 156 Participatory Approaches to Crop Improvement at the Community Level in Vietnam Nguyen Ngoe De Abstract Crop improvement has been one of lhe strong, continuous programs in the Mekong Delta for major crops, especially rice. However, most breedíng programs have been set and desígned by breedors, neglecting the role of users: farmers and farming communities. Farmers have been Ibe passive users, receiving finishcd breeding lines/varieties for their production. The dissemination process of"technol- ogy tr.nsfer" has becn very slow and costly for both beeeders and farmers. The USe ofparticipatory approaehes in crop improvement have ensured the involvement offarmers in the whole process or, at least, in the evaluation process. This has resulted in • better understanding and ac- ceptability of new erop varieties generated through the breeding programo C,n Tho University, as the leading researeh institution for adapting participatory approaches to rice im- provemenl, started on-farm breeding programs as early as 1975, afterthe war, by sending out their staff and students lO work elosely with farmers on erop-improvemen! programs. In !994, with Ibe inception of Ibe Cornmunity-Based Biodiversity Developmenl and Conservation (CBDC) project, participalory plant- breeding (PPB) and panicipatory varietal-selection (PVS) approaches were introduced as melhods lO develop .nd identiry erop vari.ties specific lo ruche enyironmen!s and farmers' preferences. These partieipatory approaches are also being used in one oflhe study ,ites, Tra Cu, of the global in situ conservation project implemented in Vietoam in collaboration with Ihe lntern.tional Planl Genetíe Re- sourees Institute (IPGRI). The resul! has becn very positive, wíth many promising erop varieties selecled from these programs and used in larger-scale production. Farmers have becn successfui in segregating material seleclion and many farmers have beeome well known lhrough Ibese activitios. Participatory approaches are very important for erop improvement at the cornmunity level in Vietnam. PPB and PVS approaches are the key tool for erop improvement. Suceessful results from farmer selec- tions have strongly proven thal Ibese approaches are right. This experienee has been very useful for nationa! crop-improvemenl programs. Introduction Crop improvement has been one ofthe strong, continuous programs in the Mekong Delta for major crops, especially rice and beans. However, most breedíng programs have been set and designed by breeders neglectíng the role of users: farmers and farming cornmunities. Breeders have set their own breeding objectives and conducted crop-improvement programs based on their own analysis ofproblems and on-station research findings (COWI 1999). At the end oftheír breeding programs, promising breeding materials are released to fanners as so-called "technology transfer." Fanners are passive users, receiving finished breeding línes/varieties for their production. In many cases, fanners, especially the poor, refuse to try new varieties because they do not want to take the risk. Resource-rich fanners are the Iirst to try such varieties. Participatíon is Iimited to providing a piece ofland to the breeders for on-fann trials. The dissemination process of"technology transfer" has been very slow and costly for both breeders and farmers. As a result, the adoption of recornmended varieties, in many cases, has been very slow, doubtful, or has even failed. Local adoption of new technologies is dependent not only on technical suitability and economic viability but abo on social Nguyen Ngoc De is at the Rice Research Department. Mekong De1ta Farming Systems Research and Development Institute, Can Tho Universjty. Vietnam, !57 Participatory ApprQaches lo Crop lmprovemenl al Ihe Community Leve! in Vietnam acceptance. The use of participatory approaches in crop improvement assures farmers' involve- menl in the whole process or, al least, in the evaluation process. This has resulted in better under- standing and greater acceptability of new crop varieties generated through breeding programs. Can Tho University, as the leading research institution for adapting participatory approaches in rice improvement, started on-farm breeding programs as early as 1975, afier the war, by sending out their staff and students to work closely with farmers on erop improvement programs (Xuan et al. 1993). In 1994, with the inception ofthe Cornmunity Biodiversity Development and Conservation (CBDC) project, participatory plant breeding (PPB) and participatory varietal selection (PVS) were introduced as methods to develop and identifY crop varieties specific to niche environments and farmers' preferences (CBDC 1996, 1997). Witcombe and Joshi (1996) defined PPB as involving farmers in selecting genotypes frem geneti- cally variable, segregating materials and PVS as involving the selection by farmers ofnonsegrega- ting materials, characterized as products from plant-breeding programs. However, they also agreed lhat PPB is a logical extension ofPVS. In our view, PVS is only a lower leve! ofPPB. PPB, there- fore, should be understood in its broader meaning and implications as the involvement offarmers in the whole process of plant breeding, no! only the selection of segregating and nonsegregating mate- rials. Farmers can be involved at the very beginning, when strategies and objectives are se! forplant breeding, in identifYing parents, making crosses (of course with training from the formal sector), and selecting both segregating and nonsegregating materials. The experiences from the CBDC project in Southeast Asia have proven lhat peint, especially in the Mekong Delta in Vietoam and in Bohol, Philippines, for rice (CBDG 1998). These participatory approaches are also being used at one of the study sites, Tra Cu, of the global in situ conservation project implemented in Vietnam in collaboration with the Intemational Plant Genetic Resources Institute (IPGRl). Methods used in participatory crop improvement The participatory cropcímprovement program uses pvs andlor PPB approaches, depending on fanners' varietal needs and their breeding knowledge and technical skílls. The pvs approach has been used to improve locallandraces and to evaluate the finished breeding materials, obtained frem research institutions, on farmers' field. When varietal options avaílable to farmers through PVS are limited or exhausted, PPB is initiated (CBDC 1998). Farmers with knowledge ofand interest in breeding are involved in PPB activities, i.e., activities frem crossing desired parent lines to select- ing and evaluating the segregating genetic materials (De and Tin 1998). A flow diagram showing the methods used in participatory erop improvement is presented in figure 1. The methods used in implementing PPB and PVS are discussed below. Methods used lor PPB Particípatory plant breeding involves the following steps and activities. Need assessment and seledion of cooperating farmers, Cornmunity meetings are organized to identifY farmers' problems and needs and to come up with suitable crop-improvement strategies and plans. A group offarmers (Group 1 farmers), with knowledge ofand interest in breeding, are selected as cooperating farmers in consultation wíth the community. Breeding acti vities are then formulated and decided upon with these cooperating farmers. 158 Technicai ＬＭＭＭＭＭｾ＠ asslstance Breeders : lit Project Sta" r" '--------' MooitOfing Materíals !Crop.improvement sedor ; , PGR conservation & development ¡ Supporting sectors ｾ＠ Experience shafing • Traíning - Farmers' field days - Field visrt ·IPM1. EPM2 .1 Group 1 i------------to>I¡ (2·3 rafmOn;) I PartlclpatQry plant breeding ·Broeding . Segregaq materials selection .1 Group 2 I '------------+1. 1 (5·10 rarme,.) ｾ＠ J Group 3 I '------------+1 1 ＨＳｾｏｲｾｭＩ＠ ｾ＠ On·farm trials ｾ＠ PVS • Varietal yeUd triai - Seed selecOOn • Seed muttipfication • Seed distribution ! FarmenJ In the eommunity I • Planting material$ & saed production - Olssemination I Farmet'$ In oth$l' communltfes I Note: Group ｉｾａ､ｶ｡ｮ｣･､＠ fanners with good knowledge and skili in breeding. Group Ｒｾｆ｡ｮｮ･ｲｳ＠ with good knowledge and .kiU in seed seleclion. Group ＳｾｆＧｬｬｦｬ･ｲｳ＠ with good knowledge and ,km in seed production. 1. ｬｐｍｾｉｮｴ･ｧｲ｡ｴ･､＠ pest management. 2. ｅｐｍｾｅ｣ｯｬｯｧｩ｣｡ｬ＠ pest management. Figure 1.Community-based netwarklng diagram lar PPB and PVS Setting breeding objectives and identifying donor parents. Breeders work c10sely with farmers to agree on breeding objectives. Farmers have been found to use both quantitative and qualitative criteria to detennine these breeding objectives. Sorne of the examples of such eriteria are high yield, short duration, resistance to major pests and diseases, stickiness of cooked rice, and so on. Based on the breeding objectives, breeders then assist farmers in searching for suitable donor parents for crossing. These donors may be found among the avaílable genetie materials at the local level or from research institutions and are made available to the cooperating farmers. Making crosses and selecting segregating materials. The Group 1 farmers are given additional training on crossing techniques and assisted in making the desired crosses. In other cases, breeders provide seeds of segregating lines at very early generations (F2, F¡, and F 4) to the farmers ror selec- tion of desired lines based on their own eriteria. Farmers have been found to handle segregating materials from generations as early as F2• In the process, farmers apply their own crop-management practices. Based on breeding objectives, farmers observe, evaluate, and harvest the selected plants individuaIly. This process is repeated until stable Hnes are obtained. For management reasons, the 159 Participatory Approachl!!...tt? Crop lmprovemenl al the Community Level in Vietnam number of individual plants selected each season is limited, depending on farmers' capacity for seed handling and the land assigned as a breeding plot. Therefore, the genetic variation in farmers' selections is usually narrow. Only Group 1 farmers are involved in the selection process, while field operations are done with the help of other farmers in the community. Observation test. Pure lines selected from the segregating material s are planted in observation test plots to check for adaptation and yield, with common local varieties used as local checks. Farmers compare the performance of new varieties/lines with the local check and select promising ones for further evaluation in yield trials by Group 2 farmers. Monitoring. The Group I cooperating farmers take close field observations with technical assis- tance from breeders and agricultural extensionists. These farmers also keep records on field condi- tions and crop performance for later analysis in determining the suitabilíty of the new erop varieties under selection. Methods usedfor PVS Participatory varietal selection involves the following steps and activities. Need assessment and selection oC cooperating farmers. As in PPB, eonununity meetings are organized to identifY farmers' problems and needs in relation to their current erop varieties. F armers may want to improve their current varieties or ehange for promising new varieties. A sepa- rate group of farmers (Group 2 farmers), with good knowledge of and skills in seed seleetion and management, are also selected as eooperating farmers in consultation with the conununity. PVS activities are then fonnulated and decided upon with the cooperating farmers from both Group I and Group 2. Provision of genetic materials and participatory selection. Three sources of genetic material s are used to obtain seeds for participatory selection of desired erop varieties: • PVS with improved locallandraces. The improvement oflocallandraces is done through mass as well as pure-line seleetion. Since the mass-seleetion method does not require very specialized skills, Group 2 farmers, afier a simple orientation, have been able to undertake tbis selection. On the other hand, pure-Iine seleetion for erop improvementrequires speeial- ized skills and care on the par! of the farmers. For this reason, only Group 1 farmers have been used to do pure-liÍle selection, afier adequate training and with inlensive monitoring. The improved locallandraces are then given to a large number of farmers within the eonunu- nity, as PVS material s, for their own testing and seleetion. • PVS witb reintroduced locallandraces. PVS also reintroduces landraces from genebanks back to the conununity when local materials have been destroyed by disaster. Usually the eollected local varieties from different locations within and outside of the eonununity are evaluated in the conununity to give farmers more choiees. • PVS with modern crop varieties. Modem erop varieties from research institutions and fin- ished products from PPB are also given lo the eooperating farmers for testing their suitability under farmers' own management conditions and household requirements. Yield trials oC successful PVS varieties. The erop varieties preferred by farmers under the PVS program are then put into varietal yield trials in the conununity for farmers to observe directly and make selections of their choices. Conunon varieties in the conununity are used as local cheeks in these trials. Farmer field days are organized just before harvesting to bring farmers in the conunu- 160 Nguven Ngoc De nity to tbe trial plots for ajoint evaluatíon ofthe tested varieties. Desírable varieties (usually two to three varieties) are then selected for seed multiplicatíon. Seed multiplication. Varieties selected by farmers from yield trials are rustributed to a group of farmers (Group 3 farmers), with consíderable knowledge of and interest in seed production, to mul- tiply large quantities of seeds for use by olher farmers in Ihe community. Seed multiplication fields are closely monítored and used as final checks for large-scale productiori. Monitoring. Field visits and farmer field days are tbe most appropriate tools for participatory mon- itoring and evaluation ofPVS activities. Breeders, field staff, extension workers, and farmers par- ticipate in such activities. Data collection depends on farmers' objectives and ineludes common traits such as growth duration, plan! height, tillering capacity, grain yield and quality, and tolerance to insects and díseases. Field experiences with rice Participatory varietal selecnon (PVS) Rice is tbe major food crop in tbe Mekong Delta. PVS actívities on rice have been undertaken in different forms in tbe Mekong Delta starting as early as the 19708. The most common of tbese actívíties was varietal yield trials. The main objectives ofthe varietal yield tríals were to generate farmer-preferred crop varieties and faster disseminatíon of tbese varietíes. Can Tho University has been a leading research instítution in ínítiating and implementing on-farm research activitíes. In the beginning, breeders and researcherS cooperated witb advanced farrners individually throughout tbe Mekong Delta (De 1997). During the period 1975-1995, hundreds ofpromising rice varietíes were tested in farmers' fields, and a number of varieties were identified and released. Sorne of tbese rice varieties are IR36 (later named NN3A), HT6 (NN6A), MTL30 (NN7 A), HT19 (NN2B), IR42 (NN4B), MTL58 IR 13240- 108-2-2-3), and MTL87 (IR50404-57-2-2-3). These varieties have made great contributions to tbe ímprovement of rice production in the Mekong Delta. Many farmers, such as Mr. Raí Ruu (Long An provínce); Mr. Raí Chung, Mr. Tu Tai, Mr. Ba Chuong (Tien Giang province); Mr, Ba Cung (An Gíang province); Mr. Muoi Tuoc, Mr. Muoi Than Nong (Vinh Long province); and sorne others, were known as tbe "rice-selection kings." Farmers were also found lO use pure-line selec- tion to itnprove tbe formally released varieties for grain quality and adaptation to specific conru- tions in tbeir areas. 1bis process has, in fact, strengthened on-farm conservation of crop diversity. Later, since 1994, witb tbe inception ofthe Community Bioruversity Development and Conserva- tion (CBDC) project, PPB and PVS have been íncluded in their current form in the crop-im- provement programo There has been a shift from dealing witb advanced, inruvidual farmers to farmer groups and farming communities (CBDC 1998). As a result, more farmers have been involved, the degree of participation has improved, and more work has been organized at the grass-roots level by communities tbernselves with help from many local autborities. Four farming communities used as pioneers are Nhut Ninh community (Tan Tru district, Long An province), My Thanh community (Ba Tri district, Ben Tre province), Ke Sach community (Ke Sach distríct, Soc Trang province), and Long Thanh community (Vinh Loí rustrict, Bac Lieu province). The results of PVS activíties in tbese communities are presented in tables 1 and 2. 161 Par¡icipatory Approaches¡(}.Crop Improvement al ¡he Cammunity Level in Vietnam Tabla l. Number ofRice Varieties Tested and Selected from PVS Activities at Four Communities in the Mekong Delta , Nhut Ninh III.Y Thanh Ke Sach long Thanh Vea, Testad Selected Tested Selected Testad Salected Tested 1994 TR 252 8 DWR 20 6 MR 18 4 HYV 5 1 5 1 1995 TR 23 3 HYV 1 i 5 4 5 3 ＭｾＭＭＭＭＭ 1996 TR 1 MR ! 22 1 HYV 9 9 34 9 89 _, 1997 ,TR 222 2 i MR 7 Lost' 32 29 HYV 20 9 16 8 1998 MR 11 HYV 12 6 I 18 8 19 9 Source: CBOC (1998), " Note: TR= Tradition.1 rice; DWR= Deep-wate, rice; MR= Medium rice; HYV= High-yie1díng rice (early). 1 , No data availabl. al the lime of writing. 2. Due lo a typhoon al Ihe las! .lage ofthe lrial, no result was possible, 22 169 9 25 20 12 24 Selected I 2 16 1 1 - 3 5 Table 2. Common Varieties Selected from PVS Activities at Four Communities in the Mekong Delta Rice varieties i Nhut Ninh My Thanh KeSaeh long lhanh TR Nep Thom, Tal TaiNguyen Nguyen, Me Huong MR MTL83, MTL 124 MTLa3 HYV IR49517, IR64, IR54883, 8976B, MTL99.101. IR64. MTL 138, MTl156. 157, MTL138,205 MTL 142, 157, MTL142,147. MTL159,199 MTL 164, 190, MTL 149, 150. MTL 199,201, MTL 156, 157. MTL202 MTL159,199 Souree: CBOC (1998), Note: ｔｒｾｔｲ｡､ｩｴｩｯｮ｡ｬ＠ rice; DWR= Deep-water rice; ｙＡｒｾ＠ Medium rice; HYV= Hígh-yleldíng rice (early), Participatory plant breeding (PPB) In tbe 1996/97 dry season, tbe project decíded to start providing segregating breeding materials from 63 F 2 populatíons of 12 crosses made by tbe Rice Research Department ofCan Tho University for fanner selection in the four cornmuníties 1ísted aboye (table 3), The names of these crosses are L245, L246, L247, L248, L249, L250, L251, L252, L253, L254, L255, and L256, Many farmers 162 __________________________________________________________ ｾｎｾｧｵｾｾｾｮＮｾｧｯ｣ｄ･＠ Tabla 3. Number ofSegregating Populations Distributed and Selected by Four Communities from PPB Acthities in the Mekong Delta, by Year Number of populations selected by generation (F" F" F., F,) ... - Community F, F, F4 F, i Farmers' selection Nhul Ninh 13 : 13 MyThanh 20 8 3 1 L246-10-1-B Ke Sach 10 4 2 1 L246-7-3-B (SiC-1) L247-1·5-B (SiC-2) Long Thanh : 20 11 i Total 63 36 5 2 were interested in seleeting individual plants from segregating populations based on theír own eriteria and under their own management conditions. Sorne of the farmer-selected varieties are now stable lines and are being tested in yield trials. L246-7-3-B, and L247-1-5-B, the two promising farmer se1ections and noted by farmers as SiC-l (Soc Trang Selection, no. 1) and SiC-2 (Soc Trang Selection, no. 2) respectively, were purified by bulk selection method afier F4. Farmers in Ke Sach cornmunity (Soc Trang province) are now mul- tiplying it for distribution arnong themselves. Mr. Canh is the leader of this farmers' group who has led the selection activities in this corhmun1ty. Simílarly, L246-IO-¡-B, a promising Jine selected by farmers in My Thanh cornmunity (Ba Tri district, Ben Tre provincc) is also now under yield test and seed multiplication. Besides four cornmuníties the initíal1y selecled, the PPB and PVS prograrns were also expanded to include other advanced, individual farmers in Ihe Mekong Delta. One of these was Mr. Hai Triem from An Giang province, who was well-known as "farmer of the era" and was awarded the Third Labour Medal by the central government for his contribution to rice improvement. Problems and lessons Problems • The low ｾ､ｵ｣｡ｴｩｯｮ｡ｬｬ･ｶ･ｬ＠ ofthe farmers means they require more training and the adoption ofPPB is slow. • Few farmers are interested in working with breeding and selectíng segregating materials. F armers are more willing to multiply promising varieties than to select from segregating ma- terials or make crosses. • The number of farmers collaborating in PPB is limited, especially in pedigree selection and selection of segregating material because these are time-consuming activities. • Agricultura! policy is more favorable to cornmercial production than tp conserving diversity. • Due to the fasl turnover of rice varieties by farmers (every three to four seasons), il is difficult to keep their Ínterest and get their cooperation for the entire process of selecting segregating lines, which takes time 10 get results. 163 Participatory Approaches to Crop Improvement at the Community Leve! in Vietnam Lessons • Support from local authorities and organizations in term of organization, management, addi- tional funds, and facilitation is very important. Cooperation with groups and eornmunities on PPB and PVS gives better results than work- ing only with individual farmers. • Farmers' field schools and farmers' field days for PPB and PVS are good ways to motivate the farmers' participation at the eornmunity leve!. • Farmers conserve and maintain the diversity of plant genetie resources to meet their own needs for home consumption, marketing, and adaptation to local environments and farm resources. • Biodiversity development should be considered on a temporal and spatial basis at the level of speeies, erop, and agroeeosystem. PPB and PVS inerease plant genetic resources at the level of the gene pool and not at the level of speeific varieties. • In situ and ex situ conservation and development are eomplementary. • Biodiversity in the Mekong Delta is eurrently under pressure but integrated farming systems and diversifieation of plant genetic resources could help to eorrect the situation. Participatory approaehes are very important for erop improvement at the eornmunity level in Viet- nam and are efficient ways of aehieving crop improvement at this leve!. PPB and PVS are the key tools for this. Sueeessful results tTom fármers' selections have proven that these are the right approaehes, providing a very useful lesson for national erop improvement programs. References CBDe. 1996. Annual CBDC pro}ect report o[ 1996. Can Tho, Vietnam: Cornmunity Biodiversity Development and Conservation Project. CBDC. 1997. Annual CBDC pro}ect report o[ 1 997. Can Tho, Vietnam: Cornmunity Biodiversity Development and Conservation Project. CBDe. 1998. Annual CBDC pro}ect report o[ 1998. Can Tho, Vietnam: Cornmunity Biodiversity Development and Conservation Project. . COWI. 1999. Seed sector study, Vietnam. Volume J: Findings, Conclusions and Recommendations. DanidaIMARD draft reporto March 1999. Hanoi, Vietnam: Danida. De, N.N. 1997. Data colleelion and analysis in the Mekong Delta Cornmunity Bioruversity Development and Conser- vation Project of Vietnam. Paper presented at the 2"" !PGRl meeting, Rome, Italy, 25-29 August 1997. De, N.N. and H.Q. Tin. 1998. Participatory plant breeding in Vietnam CBDC pro}ect. Technical report to CBDC Regional Coorrunation Unit. Quezon City, Philippines: South East Asia Regional Institute for Cornmunity Edu- carion. Xuan, V.T. and N.N. De. 1993. Present status of agricultural extension in Vietnam. Paper presented at the first South- east Asia workshop on formulation of project proposals on technology transfer for major food crop production, FAO and UAF, Ho Chi Minh City, 6-9 Dec. 1993. Witcombe, J. and A. Joshi. 1996. Farmer participatory approaches for varietal breerung and selection and linkages to the formal seed sector. In Participatory plant breeding: Proceedings o[ a workshop, 26-29 July 1995, Wageningen, Netherlands, edited by P. Eyzaguirre and M. lwanaga. Rome: Intemational Plant Genetic Resources Institute. 164 Using Farmer Knowledge for Participatory Sweet-Potato Variety Selection in Garut, West Java, Indonesia Caecilia Aji-a Widyastuti and MinantyorÍni Abstract Thís paper describes trials usíng sweet-potato gennplasm from lrian Jaya, where sweet potatoes are a sta- pIe foed in the highlands. During the coIlection of sweel-potato gennplasm, fannecs' knowledge ofthose sweet potatoes has a1so been coIlecled. Fanners' knowledge about sweet potatoes in Irian Jaya will be used as a hasis for trus prcjecl and includes information on yields, the use of sweet polatoes as human foed or reed for !ivestock, and the condition of Ihe environmenl. Varieties are selected on the basis offarmers' criteria, including market orientation and table consump- tion: skin color, l1esh color, unifonnity, and other criteria. The project is also collecting information on farmers' cultivation practices, such as using high ridges in the rainy season and reducing the leaves dur- ing Ihe growing period, as well as how lo choose healthy cuttings. Methodology The objective of tbis research is not only to get a bigh-yielding sweet potato fuat is adaptable in Garut, but also to get new variety/ies wifu fue agrononllc characteristics required by different user groups (Le., fanners, traders, consumers). The study was set up in fue village ofDesakolot, Cilawu District, Garut Regency ofWest Java Provinee in a rainfed field that had been used for brick makíng six years before and had remained fallow fOT five years. The year befare fue trials took place, fue field was planted wifu yambean. One week prior to planting, ISO sacks ofmanure were applied in order to improve fue soíL Thís ís al- ways done in this area, especially for land has been used for brick makíng. Thls field is Iypical of places where sweet potatoes are grown. The nearest field to this site is planted wifu com, sweet po- tatoes, and ginger. Thls neighboring field was also used for brick making, and fue vigor offue plants grown on it is good. Prior to establishing lhe field trials, planting material s were multiplied in Cibadak, Pacel, about 3.5 hours away trom Garut, since ít was very dry in Garut. A total of 64 cultivars, including five eheeks (BISI83, SQ27, CIP-I, Jahe, and Keleneng) were tested (fue last two offue checks are well-knowu local eultivars in fue area). There were 36 hills per plol. The date of planting was 26 F ebruary 1998. The experimental design ls a randomized complete bloek wifu three replíeations. The size of indi- vidualplotsís 1.6 mx 3.0m. Spacingis 80 cm betweenrowsand 15 cm to 18 cm betweenbills. Har- vesting ls done aceording lo fue fanners' sehedules. During fue harvest, we invíted fanners, traders, and extensionists to select sweet potatoes based on their eriteria. By using participatory lools sueh as flags, they walked around lhe tria! field and chose what fuey Iíked. After fuat they ranked the selected varieties based on produetion, skin and flesh color, uniformity, skin smoothness, and general acceptance (table 1). Figure l shows participants ranking fue selected varieties. Caecilia Afta Wídyastutí is a rese.rche, al the Intcm.tional Potato Cenler (CIP)-ESEAP Region. Bogor, Indonesia. Minantyorini 1S a researcher al the Research Institute for Bioteclmology, Bogor. Indonesia. 165 Using Farmer Knowledge for Participatory Sweet-Potato Variety Selection in Garut, West Java, Indonesia Table 1. Selection Criteria and Rank of Sweet-Potato Varieties Rank 01 seleclion 11 111 IV V VI VII Criteria of selection Production Ski n color Root shape Flesh color Unilonmily (shape and size) Skin smoothness General acceptance Resulls (in arder) Kinta, Toweko, Lemekuara, Umakmbi, Pipombi Toweko, Pipombi, Lemekuara, Umakmbi, Kinta Umakmbi,Toweko, Kinta, Pipombi, Lemekuara Toweko, Umakmbi, Lemekuara, Kinta, Pipombi Umakmbi, Toweko, Pipombi, Lemekuara, Kinta Toweko, Pipombi, Lemekuara, Kinta, Umakmbi Toweko, Umakmbi, Lemekuara, Pipombi, Kinta Figure 1. Farmers, traders, and extensionists ranking selected sweet potatoes Results and discussion The experimental field was harvested on 22 August 1998, according to the farmers' schedule. No check varieties were select by farmers-not even Racik, the most popular local cultivar. Five new cultivars, i.e., W0139 (Toweko), W033l (Kinta), WOlll (Umakmbi), W01l3 (Lemekuara), and WOlO9B (Pipombi), were selected by the farmers, traders, and consurners (table 2). Toweko ap- pears to be the most preferred cultivar in this area. F armers in Desakolot plant sweet potatoes for cornrnercial purposes. They have several require- ments, such as high yield, smoothness of skin, skin and flesh color, uniforrnity in shape and size, and root shape. High yield is one important requirement for cornrnercial purposes. The idea of "high yield" includes early maturation. Farmers prefer to plant sweet potatoes that with a high yield but they also require other criteria such as smooth skin, good skin and flesh color, etc. Table 2 shows that Kinta, 166 C. A. Widyasluli and Minant}'Orini Table 2. Farmers' Selections from the Irian Jaya Sweet-Potato Trial Uniformity Skin ACCQS- Produc- Ski n Roo! Flesh (sh.pe ｳｭｯｯｴｨｾ＠ Generat No sion No. Local name tion color shape color and slle) neS$ acceptance W0139 Toweko . ｾ＠ .. H·U' • ..... 1> ..... - 2 W0331 Kinta 3 WOlll Umakmbi ＢＧｾ＠ .. 4 WOl13 Lemekuwara 5 WOl09B .. Note: Ranking is inmeated on a scale from 1 lo 5, where ••••• indic.tes highly accept.ble and • inmeates I()w acceptabílity . which had the híghest yield was gíven low acceptance overall because it did not have acceptable skin color, unifonnity, or skin smoothness. Smooth skin color refers to skin that has not been damaged by weevils or nematodes and thal exhíb- íts no cracking. Skin should be thick enough to withsland peelíng during transportatíon and to be resistant to weevils or nematodes. The smoothness of the skin has a considerable effect on the príce of sweet potaloes. Farmers always refer lo good-tasting sweet potatoes as ubi ketan (stícky sweet potatoes) ifthey see a sweet potato with purple flesh. According to them, these sweet potatoes get a good price. Toweko (W0139) was given eight flags because it meets the criteria ofbigh yield, good skin color, unifonnity in shape and size, good flesh color (dark yellow), and is suitable for fresh consumption and for snack food (keremes). According to fanners, the mínimum príce for Toweko should not be less !han Rp 500. After tastíng the raw Toweko, the fanners predicted that tbis cultivar would be well received in the market. The particípatíng farmer wanted to plant Toweko 30% in the first sea- son and increase it to 50% for the next season. They said they would plant 100% ifthe market could absorb that mucho Two participating farmers, Haji Sumama and Amin, will be responsible for mul- tiplyíng this sweet potato as asource ofplanting material. Umakmbi (WOlll) was chosen with four flags because the skin is very smooth and thick, meaning it could resist weevil attacks. The flesh color is dark pUIple, meaning it will taste good (ubi ketan-sticky sweet potato), and the roots are very uniform in shape and size. With these critería, the farmers predicted that this sweet potato would command a good price in the market. According to the farmers, they can increase the production oftbis variety. Farmer Unang will be responsible for multíplying this sweet potato as a source of planting material. Kinta (W033 1 ) was given six flags because of its high yield and purple flesh, meaning it will laste good (ubi ketan-sticky sweet potato). The skin is very smooth, with no evidence ofnematode attack. Farmer Agus will be responsible for muItiplying this sweet potato as a source of plantíng material. Lemekuwara (WOI13) was chosen with two flags because of ¡ts rounded shape and smooth, red skin, whích mean it will be easier to seU in the market. Farmers chose this from replication III, whích indicated high productíon. Farmer Eman will be responsible for multíplying this sweet po- tato as a source of planting material. 167 Using Former ｋｮｯｷｬ･､ｧ･ｦｯＮｾＮｦＮ｡ｲＨｩ｣ｩｰ｡Ｏｯｲｹ＠ Swee/-Po/ato Variety Selee/ion in Garu/, West Java, indonesia Pipombi (WO 1 098) was chosen with eight flags because the size is uniform, il has smooth skin color, and it can be sold fresh. Farmer Eneek will be responsible for multiplying Ihis sweet pOlato as a souree of planting material. ConcIusions Based on our experienees with this trial, we have formed the following conc1usions: • Using farmers' knowledge about sweet potatoes from Irían Jaya will help researchers lo do preliminary se1ections for the trial. • The partícípation offarmers in the arca where the trial was set up will help in seleeting sweet potatoes based on farmers' enteria, such as marketability and table consumption. • Farmers selected sweet potaloes based on their marketabílíty and farmers' own eritena. References Schneider, J., CA Widyastuti, and M. Djazuli. 1993. Sweetpotato in the Baliem Valley orea. Irian Jaya. Bogor. Indo- nesia: Intemational Potato Center (CIP)-ESEAP Region. Widy.stuti, C.A. 1994. Women role of Dani tribe to maint.in the sustainability of sweetpotato as staple food in Jayawíjaya Regeney, lrian Jaya. In Proeeedings 01 application sweetpototo production technology and post harvesting ro support agro industry. Malang, Indonesia: Research Institute for Legnme and Tuber Crops. Widyastutí, C.A. 1995. The collection of associa!ed knowledge during short gennplasm eolleclÍans: Field experiences in Java and lrian Jaya. In Indigenous knowledge in conservaríon 01 crop genetic resources, edited by 1. Schnei- der. Bogor, Indonesia: Intem.tional Potato Cen!er (CIP)--ESEAP Regíon and Central Researeh Institute for FoodCrops. 168 C. A. Widyastuti and Minantyoríni Table 3. Yield of Varieties Tested and Fumers' Ranking for Marketabílity Yield (Ton/Ha) Marketable Not marketable No Accession No Cultlv3r I 11 111 X I 11 111 X 1 W0131 ,Son 0.56 . 0.14 1.94 0.88 1.81 1.67 0.83 1.44 2 WOI94 Yaronambiri 5.83 12.5 ¡ 8.47 8.93 2.36 1.39 2.92 • 2.22 3 W0116 Helalekue 7.08 7.08 7.22 7.13 2.92 • 1.11 1.94 1.99 4 W0113 Lemekuara 2.36 7.78 9.44 6.53 1.11 1.25 2.36 , 4.72 5 W0323 Womln 4.44 9.17 7.36 6.99 1.94 1.53 4.03 2.50 6 W0045 Poniai 5.00 6.39 6.25 5.88 2.08 : 2.36 . 1.67 2.04 7 W0061 Ti 6.81 3.61 5.00 5.14 0.14 0.69 0.97 0.60 8 9 Senggol 2.92 1.39 1.67 1.99 0.28 0.56 1.39 0.74 W0033 Sengkerengke 5.14 8.06 3.06 5.42 1.81 1.94 3.19 2.31 10 W0350 lIoka 11.11 12.22 ! 7.50 10.28 1.11 1.25 0.97 1.11 11 W0104 Gelakue 2.36 3.61 0.26 • 2.06 2.06 1.39 1.67 1.71 , 12 W0158 Musanaken baru 15.14 1028 2.50 9.31 5.42 3.19 3.19 3.93 13 W0220B Helalekue lama B ＭｾＬＮＧＱ＠ 0.37 - - 0.69 0.23 14 W0220A Helalekue lama A 1.25 3.47 3.05 0.14 ¡ - 0.28 0.14 15 W0008 Esipalek - - 0.83 0.28 - - 0.28 0.09 16 W0124 Naulupe 5.83 11.39 5.14 7.45 2.22 0.83 1.94 1.67 17 W0204 Korwambi - 0.69 - 0.23 0.42 - 0.14 0.19 18 W0181 Walegeln 2.50 • 2.36 • 0.83 1.90 2.50 0.69 0.97 1.39 19 W0084 Kuruparambi 3.61 4.44 1.67 3.24 2.22 0.97 0.97 1.39 20 W0187 Mugulele 3.06 4.03 2.64 3.24 1.67 .:s. 1 '" 3.61 2.82 21 W0048 : Giniagalo 7]8 5.14 3.06 5.33 1.39 0.56 0.56 0.84 :;, 22 W0139 Toweko 12.08 8.33 10.28 10.23 2.22 2.22 2.50 2.31 23 W0130 Siknimbi 4.58 7.92 1.25 : 4.58 0.83 0.97 1.11 0.97 24 W0197 Mukolele 5"56 4.31 3.89 4"59 1.94 • 2.78 2.64 2.45 25 W0223 Umakmbi 6.25 10.00 5.56 7.27 1.94 1.53 0.97 1.48 26 W0111 Umakmbi 8.19 3.33 6.25 5.92 2.22 2.22 1.81 2.08 27 W0316 Ketfelale , 5.00 5.00 9.44 6.48 : 0.97 1.11 2.36 1.48 28 W0018 Mai/ongge 17.08 10.83 12.22 13.38 0.69 1.53 0.97 1.08 29 ,W0300 Musan 9.03 • 3.75 ¡ 6.53 6.44 1.53 2.22 1.94 1.90 30 W0201 Gilikue 0.56 12.22 - ! 4.26 0.14 - 0.05 31 W0331 Kinta I 13.19 12.22 8.61 11.34 1.67 2.22 1.81 1.90 32 W0339 Kuning 10.97 5.69 9.17 8.61 I 1.53 : V8 0.97 : 1.76 33 W0253 Yoban 4.58 • 4.72 I 5.28 , 4.86 1.39 2.22 1.67 1.76 34 .W0041 Pusemangken 0.42 - 1.53 0.65 0.83 1.39 0.74 169 Using Farmer Knowledge (or Participatory Sweel-Potato Voriery SelecUon in Garo!. Wes! Java. Indonesia Table 3. Yield ofVarieties Tested and Farmers' Ranking for ｍ｡ｾｫ･ｴ｡｢ｩｬｩｴｹ＠ (Continued) Yield (Ton/Ha) Marke!able No! marketable No Accession No Cultivar I 11 111 X 1 11 111 X 35 WOO10 I Musan 2.50 - 2.22 1.57 1.67 0.56 1.94 1.39 36 WOl84 Lía-lia 8.19 9.17 7.36 8.24 2.08 2.50 292 2.50 37 W0125 Linggoara 4.31 1.67 1.67 2.55 0.56 1.39 0.83 0.93 38 W0241 Sahoma 11.25 8.33 10.28 9.95 1.25 1.81 0.69 1.25 39 W02BO Tuwembi 8.75 8.33 9.17 8.75 1.94 2.64 2.36 2.31 40 WOO14 Kentang 7.36 8.89 4.31 6.85 j.53 1.53 1.67 1.58 41 W0141 Gelakue Putih 2.92 6.53 2.22 3.89 1.94 2.22 0.97 1.71 42 W0021 Kila 1.25 1.94 - 106 1.53 2.92 0.28 1.58 --- 43 W0227 Kentang 0.83 2.50 0.97 1.43 1.11 0.56 0.97 0.88 - 44 W0109 Pipombi 3.06 3.47 0.28 2.27 2.92 0,97 I 0.69 1.53 45 W0109 S Pipombi S 1.25 4.44 3.06 2.92 0.69 1.39 2.36 1.48 46 W0220 Helalekue Lama 5.69 9.86 5.14 6.90 4,17 2.78 1.53 2,83 47 W0134 Nasimbi 1.39 2.78 4,86 3,01 1.25 ＧＮｾｾ＠48 W0156 Soepak Saru 4.17 4.31 I 10.28 6.25 3.61 49 W0206S Andelan B 4,72 0,56 0.42 1.90 153 • 097 1. O 50 W0206C Andelan C 1.67 1.25 1.25 1.39 1.11 • 0.42 0.69 0.74 51 W0167 Anewun 0.83 - - 0.28 0.56 0.28 0,42 • 0.42 52 Tabimbi 4.03 5.69 5.28 5.00 0.83 0.14 1.11 0.69 53 WOO05 Hoboak 8,19 2.22 6,53 5,65 0,97 i 0.83 1.25 • 1.02 54 W0206D Andelan O 3,61 1.11 2.92 2.55 0.97 0.97 2.22 1.39 55 W0260 Mikmak 7,64 8.75 14.72 10.37 1.94 1.25 2.64 1.94 56 WOO55 Mikmak 4.31 7.22 7,78 6.44 1.39 0,83 1,94 1,39 57 WOOO2 Mikmak 6,81 0.83 10.97 620 1.67 0.14 0,97 0,93 56 W0017 Wortel 6.81 4.86 1.53 4.40 1.81 0.97 1.94 1.57 59 WOO39 Tinta Kuning - 3.33 - 1.81 1.71 0.83 0.56 1,39 0.93 60 Bis 183 12,36 13.06 13,61 13,01 I 403 028 4,44 2.92 61 ! SQ27 5.69 10.91 10,97 9.21 1.39 0.14 2,92 1.48 62 CIP-l 8.47 9.03 7,08 . 8,19 1.39 264 I 2,92 2.32 63 Jahe 1,94 9.31 9,31 6.85 1.81 2.22 1,25 1.76 64 Keleneng 2,78 4,17 8,19 5.05 1.25 1,39 4,58 2.41 65 Racik 6.11 0,42 6.33 ＴＮＹＵｾ＠ 4,58 3,33 4.44 170 Understanding Agroecological Dornains: The Key to a Successful Participatory Plant Breeding Prograrn R.B. Rana, B.R. Sthapít, A. Subedi, D.K. Rijal, and P. Chaudhary Abstract Farmers have an intricate koowledge of Iheir agroecological domains. The empirical evidenees from Kachorwa (Ieral) and Begnas (mid-hill) sites in Nepal suggesl that farmers dislinguish domains for rice primarily on Ibe hasis of moislure and fertility. Farrners also differentíale the number, relative size, and specific eharaeteristics of each domrun wíthin a given geographíe area. Símilarly, Ibey allacat. individ- ual varietiesllandraces to each domain, indieatíng lhat the competítion between varietíesllandraces accurs within the domain and Ibat transgression of domain was ralber limited. These deductions need to he verified at a wider level. A fuller understanding by researehers ofspecific agroecologieal domains is a prerequisite ror them to contribute substantíally in planning and executing effective participatory plant breeding (PPB) programs. Only with a sound knowledge of agroecological domains and the varietal dis- tribulion within domains can a program on diversity deployment and biodiversity conservation be effec- tively implemented. Likewise, justifying Ihe cosl-effectiveness of PPB, targeting researchlexlension activities, and measuring Ihe contribution of PPB to foad security demands a detailed Wlderstanding of agroecological dom.ins. Simple and practical ways lO ilIieit inforro.tion on agroecoJogical dornaios and assaciated varietiesllandraces tbrough farmem' group discussion al Ibe víllage level have been suggested as a pre-projeet activity for PPB, which could enhance Ibe suecess of PPB programs. Introduction The importance of agroecological dornams can be found in earlier work on defining and delineating . recornrnendation dornains (RDs), whích is c10sely associated with the farrning systerns research of the late 1970s (Wotowiec, Poats, and Hildebrand 1986). Initial work on RDs concentrated on a few relatively easily identifiable factors (bíological variables), such as land and soil types, agro ecologi- cal zones, and erop types and rnanagernent (Harrington and Tripp 1985). The exercise on RD was híghly complex sinee the process was to identify farrning households, based on the sirnilarity in their practiees, rather than farrns. But the delineation of agroecological domains was rnueh less eumbersorne with rice because rice is very sensitive to changes in agroecological conditions and its adaptation is Iirnited, as compared lo sorne other crops such as maíze. Moreover, rice is the rnost im- portant cereal crop in the regíon, so farrners have an in-depth knowledge ofrice-growing environ- rnents and varieties suitable to different agroecological dornains. The current endeavor on refining the definition of agroecological dornaíns for rice in parts ofNepal is the case of"sharpening the focus" fur better targeting of participatory plant breeding (PPB) work, including diversity deployrnent, eonservation of landraees in different dornains, and planning stra- tegic erop rnanagement research. The methodology adopted is quite simple and can be replícated in other areas for wider use by the researchers and deve10pment workers. R.B. Rana, A. Subedi, D.K. Rijal, and P. Chaudhary .te with Local lnítlatives fot Bíodlversity Re.eatch and DeveJopment (LI-BlRO). B.R. 8thapit is with the Intemalional Plan' Genetíc Re.ources Institule, posted in Nepal. 171 Farmers define and characterize agroecological domains Field exercises for delineating agroecological domains have largely been influenced by the meth- odologies on RDs advocated by Collinson (1980), Franzel (1985), and Vaidya and Floyd (1997). Ihey emphasized the use of secondary sources of information, followed by preliminary surveys supplemented later by a formal survey lo refine the domains. However, later views on the subject hold lha! the refining process should take place only after researchers have a clear understanding of the variabílity inherent in the local farming systems (Cornick and Alberti 1985). The current study embodies the thoughts from both the methodologies for delineating domains and associated rice landraces/varieties. In the process of delineating agroecological domains, two group meetings were organized in the Kachorwa and Begnas eco-sítes. The first meeting was held with field-based staff; the second, with farmers from the project area. Ihis was followed by a transect walk by researchers and farmer rep- resentatives lo jointly validate farmers' statements. Ihe exercise took about two days, including field visits in each site. Interactions with field-based staff Sínce field-based staff are stationed in villages, it was expected !hat they would have a fairly good understanding of the agroecological domains and the farming systems of their respective eco-siles. Hence, the first level of group discussions was organized in field offices, with the field officer, tech- nical assistants, and motivators part.icipating. Afier discussions, the participants were able to come up with four major agroecological domains, mainly defined on the basís of water regímes. They also broadly classified the soíl type and fertility status of soils from each domaín, based on scientific knowledge of soil classification and character- ization. Participanls were also asked to estímate the size of each domain and place different land- races/varieties in their right domains. Estimating tbe relative size of each domain was straíght- forward because tbe pok:harilman occupied only a limited area within the eco-site. But placing each landrace/variety in its right domain proved more difficult. The team could place tbe majority of landraces/varieties in their domains, but the number of landraces/varielies per eco-síte was too large for them to rernember aH the names and tbeir right enviromnents. The process was also com- plicated by the fact lhal sorne of the landraces/varieties are grown in more than one domain. The whole process was reviewed by the participants, and once they were satisfied with the sleps and outputs, the field officer was asked lo facilitate tbe same process for tbe farmers' group discussion. Group discussion with farmers A group discussion was held with farmers witb the specific objective of delineating agroecological domains. Fíeld officers/sile coordinators facilitated the discussion and tbe whole exercise was repeated witb farmers' groups. Both female and male farmers participated in the discussion and put forward tbeir opinions. Farmers identified four agroecological domains within the eco-site (ucha, samta/, nichaJkhalar, and pokharllman), based on the major criteria of moisture regime and fertility status/gradient (tables 1 and 2). They could easily identifY the relative size of each domain, but there were dis- agreements among about soil c1assification. Perhaps this reflected the variability of the soíl types and soil fertility slatus in each domain. Placing landraces/variety in the domains initiated a lively 172 Table l. Agroecological Domains at Kachorwa Eco-Site Domaln Ucha (bhadalya rice culli- vated on availability 01 water, good wlnter crops) Samlal (Good erop 01 bhadaiya rice and winter erops, aaghanl rice can be grown) Nicha/Khalar (Good erap of aaghani rice and medlum winter crops) PokharllMan (can only graw aaghanl rice) Source: Chaudhary (2000). 5011 type Balaute " sandy (ujar " whitlsh) Domat" Loam Balaute domat " sandy lcam (whltish and brown) Matiyar " Clay1 (Piyar" Yellowi5h) Matiyar " Clay? (kalolkariya = black) Production potentlal low (111) HIgh (1) Hlgh (11) Low (IV) Cultlvated landraceslvarletles Mutmur, Sotwa, Sokan, Sara,., No modem varleties grawn, Lalka larm, Nakhi sara, Sathl, Bhadaiya Basmatl, Khera, Aanga, Ujala laram, Sotwa, SOkan, Dudhi sara, Kamod, Madhumala, Basmati, Karma ... (China 4, Philips, Jiri, 1V, Chandina, Sabetri, .. ) - Modern varieties Ba5matl, Lajhl, Mansara, Karma, Balsar, Rat ranl, Faram, Kamod, Madhumala (Mansula, Sabetri, Pankaj, Nat masula, Jaya, K, Mansuli ... )-Modern varieties Bhati, Megraj, SilahouL No modern varieties grown. Table 2. Agroecological Domaln! at Begnas Eco-Site Slzeof Domains doma!n Productivity Cuitivated landraceslvarletles Mula khallBhale I Kalo Jhinuwa, Pahenlo Jhinuwa, Jhinuwa, Lamcho Jhuluwa, khetlKhule !2 ha) 8% 4% Ave. operational size 1.00 ha 0.70 ha Literacy rates (%) Male head 72% 51% Female head 40% 14% Average family size 7 7 Note: Figures in parentheses are percentages oftotal mate and femate labor inputs in rice production. The gender division of labor in rice production The majority of the respondents belong to the lower social class, with small-sized landholdings. Females are younger and have lower literacy rates, compared to males, and have over 20 years of farming experience. The extent of female participation in rice production is high in both villages. Sorne tasks in rice production and postharvest operations are gender specific. Land preparation and the application of chemicals are men's responsibilities in both villages (10% of fertilizer applica- tion is done by women in Basalatpur). In Mungeshpur, women from the lower social status domi- nate in the work of pulling seedlings (100%), transplanting (70%), weeding (80%), applying farrnyard manure (60%), harvesting (82%), and threshing (82%). In Basalatpur, more men than women participate in pulling seedlings and harvesting. Women do the transplanting of seedlings (100%) and most ofthe weeding (75%), with men doing most ofthe spraying (90%). Women are also mainly responsible for postharvest activities such as cleaning and selecting the seeds for the next season, storage, and processing rice into other food products for home consumption and for sale. They are the primary end-users of rice byproducts and biomass for livestock and other farm use. A village study in eastem India revealed that women from the lower castes provided 60% to 80% ofthe total labor input in rice production (Paris et al. 1996). Aside from their significant con- tributions in rice production, women also provide labor in non-rice crops, collect green animal 182 TR. Paris el al. fodder, and feed and tend Iivestock. Thus, men's and women'g preferences for specifíc traits in rice varieties may differ, based on gender-specific roles and responsibilities. With inereasing male migration lo cities, women are laking on more responsibilities as farm managers, aside from theír normal household and childcare responsibilities (Paris el aL 1996). Rice varieties Varieties grown by ¡armers The rice varieties eurrently grown by farrners are shown in table 3. Traditional varieties are more cornrnon in Basalatpur than in Mungeshpur. Although modern varieties (MVs) show higher adop- tion rates in Mungeshpur, these varieties ofien suffer from submergenee, drought, and stress al re- produetive and ripening phases when the erop is planted late. Most farrners felt that traditional varieties are more tolerant to drought, submergenee, pests, and diseases, while MV s performed well under irrigated conditions. The majority of the farrners indieated that they felt that MVs needed better management lhan traditionaI varieties. Modero varieties need more labor, higher lev- els of fertilízer, and more irrigation, but more farmers prefer to grow MV s because of their higher yields. Table 3. Popular Rice Varieties Grown by Farmers According to Land Type Landtype Variety Basalatpur Mungeshpur Uplandlmidland Traditional Bengalía. Sarya, Kuwan Ari. Bagrí, Balbagra, Chaini Mashurí, Oríswa, Malwa Improved NDR-97, ｓ｡ｾｵＭＵＲＮ＠ PNR-381 Saket-4, NDR-80, 91,118 NDR-359, Pant-4. Pant-10, Pant-12, Sarju-52 Shaliow ｬｯｷｬ｡ｮ､ｾｯｷｬ｡ｮ､＠ TradHional Kalamanak, Motibaddam, Bilaspurí, lodrasan Malwa, Malasia Improved Mashuri, Rajshree, Sambha Mashuri, Madhu, BKP-246, Mashuri Dwarf Mashun Topographical adaptations Farrners generally match varieties wíth their environment. For rainfed rice, this means an adapta- tion to the hydrological conditions of their fields, Each field position in the topo-sequence corre- sponds to a risk of drought or submergence. The drought risk inercases frorn the bottom to the top of the topo-sequence, while submergence risk decreases along the same path, assocíated with progres- sively lower water depths and earlier recession of the water. This translates into different ideotypes for the different situations. Table 4 shows varietal diversity according to land type/topography. In Basalatpur, varieties such as Bengalia, Sarya, Oriswa, Kuwari Mashuri, Malwa, and Ghanbhanan are the major traditional rice cultivars grown in the uplands, and Kalamanak, Malasia, Motibad- dam, and Malwa are the major varieties grown in the lowlands. Improved varieties, such as NDR-97, PNR-38 1 , and Sarju 52 are grown in the uplands by a few farrners, but the improved vari- ety, Mashuri, occupied more area in the lowlands. In Mungeshpur, the cornrnon local varieties grown on upland fields are Ari, Bagri, 90 days, Sonia, Lalmati, Punjab, Lalbagra, Ashwani, lndrasan, and Bilaspuri. The improved varieties are Saket-4, NDR-80, and NDR-118 in upland and medium fields and Sarju 52, Mashuri, and dwarfMashuri mostly in lowland fields, 183 Listening lo Farmers' Perceplions Ihrough Partícípatory Rice Varíetal Se/ecUon Table 4. Farmers' Perceptions ofUseful Traits in Selecting Rice Varieties According to Land Type Mungeshpur ----- Upland lowland Upland lowland Traits Male female Male Fomale Mal. Female Male Female Grain yi.ld 36.67 39.50 48.67 49.67 41,67 35,96 42,06 40.45 Ouration 25.83 34.5Q 0,67 1.00 20.56 25,84 20.56 15,QQ Grain príce 0.00 0,00 15.67 16.00 1.67 2.81 2.97 1.82 Resistance ID abiotic stress 8.33 6.10 0.67 0.33 6.10 6.18 ! 5.10 5.00 Biomass quality 3.33 2.50 5.33 4,61 5.00 2.25 5.52 8.64 Taste 1,67 0.50 10.33 12,33 2.78 2.81 2.12 3,18 Bold and pura graln 7.61 1.50 1.67 0,00 4.44 4.49 3.40 5.00 Adaplation lo speciflc soillype 3.33 3.00 2.33 0,67 5.00 4.49 5.52 6.36 Postllarves! quaHIy 0.83 3,00 6.67 7.67 0.00 5.06 0.00 2,27 Resistance to bioHc stress 4.17 2.50 1,00 1.33 3.89 1,69 4.25 3.18 Cooking characteristics 0.83 1,00 1.67 2.00 3.89 3,92 3.40 5.00 Response to fertilizar 2.50 1.00 2.67 1.33 5,00 2.25 4.25 1.82 Competitiveness with weeds 0.00 0.00 0.00 2.33 0.00 2.25 0.00 2.27 Resistance to lodging 1.67 0.00 2.65 0.67 0.00 0.00 0.85 0.00 Adaptation to several preparations 2.34 4.00 0.00 0.00 0.00 0.00 0.00 0.00 TOTAL 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100,0 Note: Traits are Iisted in order of importance. Graín yield ¡neludes tillering, panide length, and number of graios. Resistance to biotic stress ineludes resistanee to pests and blast Resisrance to abiotic stress ¡neludes resistance to zinc deficiency and drought. Biomass quality ineludes heigllt and qualily and quantity ofstraw. Postharvest quality ineludes case ofhulling and milling recov- ery. Cooking characteristies inelude cooking time, elongation ability, aspect afier cooking, and impression in the stornach. Medium-duration fields are grown mostly in medium land. Varieties such as Sarju-52. Ashwani. NDR-359. Pant-4. -10, and-12. andIndrasan are grownon the fields thatare located in between up- per and lower levels oí land type. Fanners of Mungeshpur prefer to grow these varieties on the these land types on the belief that they need optimum moisture during the growth period. Fields dif- ter in agrohydrological ｣ｨ｡ｲｾ｣ｴ･ｲｩｳｴｩ｣ｳ＠ in Basalatpur; therefore, sorne farrners prefer to grow medium varieties on upland fields also. Farmers' perceptions of usefol fraits in varietal adoption To determine whether there are gender differences in perceptions of useful traits in varietal adop- tion, we used graphic illustrations of traits. We first showed cards that illustrate useful traits in selecting rice varieties. We then asked each farmer what traits he or she consider in selecting rice varieties for specific land types-upland and lowland fields. To assess how farmers valued each trait, we asked the question, "If you had 100 paisa, how much would you pay for each trait? The value in paisa allocated to a particular trait corresponded to the importance given by the fanner. Because many traÍts are interrelated, we rec1assified them in consultation with a plant breeder. For example, we grouped traits such as ease in hullíng and mílling recovery under postharvest quality. Table 2 shows the seleetion eriteria of male and female fanners for different land types and villages. 184 T.R, París el al. F avorahle rainfed low/ands (Basalatpur, Siddathnagar district) In the lowland areas in Basalatpur, yield and duratíon are the most important trait5 maJe and female farmers consider in selection rice varieties, In this village, the popular traditional varieties are Bengalía, Oríswa, and Kuwari mashuri. These are short-duration (90-110 days), medium-height varieties, The average yields are 2.5 tons per hectare, Farmers prefer short-duration rice varíeties in the uplands because of the importance of growing early winter crops such as oilseed, linseed, pulses, peas, and potatoes. They prefer to parboil Bengalia; otherwise, its grains break easíly. Women in Basalatpur use traditional rice vari- etíes for making puffed rice and churra, beaten rice Iike cornflakes .. For women who continue to use the traditional method ofhand-pounding rice, postharvest qualities such as ease ofhulling and mgh milling recovery are additional useful traits. The men did not mention these. The finding that women are more concerned !han men with postharvest traits and milling recovery are similar to the findings in a participatory breeding project in the hígh altitudes in NepaL Sthapít, 10sm, and Wítcombe (1996) also observed that women farmers are particularly skillfuJ in assessing post- harvest traits, such as milling recovery, and the cooking and eating quality of rice. They found that the evaluation scores between maJe and femaJe farmers in Chhomrong village showed significant agreement. Women farmers reported ¡hat they would like to decide on varíety selection after the postharvest evaluation. Consumers preferred wmte-grained rice to red-pericarped rice because it saves women time in milling. In Basalatpur, both male and female farmers agreed upon the important traits fo! 10wland rice vari- eties. Grain price is an important cohsiderlltion for farmers here because they seU traditional varíet- ¡es in the market. These, like Kalamanak, command a higher price because oftheir good taste and aroma. Kalamanak gives Iow yields of 1.5 to 2 tons per hectare. In contrast, grain price is not an important consideration in Mungeshpur because rice ís mainly used for home consumption and is seldom sold in the market. Shallow, suhmergence-prone uplands (Mungeshpur, Faizahad distríd) In Mungeshpur, both male and female farmers agreed upon important traits in selecting varieties for the uplands. Women gave more importance to postharvest qualities and grain quality such as bold and pure graíns. For the lowlands, both males and females cited better grain yield, medium duration (125-135 days), bioniass, and resistance to abiotic stress as their selection critería for low- land rice varíeties. Women gave greater weight to better adaptation to specific soH types and to grain quality. Women mentioned additional useful traits for varíeties in the uplands and lowlands that were not mentioned by men: competitiveness with weeds and postharvest quality. Weeds are the major problem in the uplands, particularly when rice is direct-seeded. In the lowlands, weeds are more prevalent during drought. These additional traits are related to the roles and responsibili- ties of female farnily members (e.g., hand weeding and feeding rice straw to livestock). Farmers' evaluation of new rice genotypes grown in farmers' fields During the 1999 monsoon season, two farmers from each of the villages of Mungeshpur and Saríyawan (rainfed neighboring village) ofthe Faizabad district and from Basalatpur were selected to check the performance of rice genotypes in their fields. The genotypes were (1) advanced lines from a shuttle breeding project from Uttar Pradesh, (2) released varieties, and (3) the most common local varieties. Of the 14 genotypes screened in Basalatpur, two are scented varíetíes (Kamini, 185