Listening t(}Earmers 'f'erceptions through Participatory Rice Varie/al Selec/ion Mrs. Yadav is 53 years old, iIliterate, and a fuIl-time farmer. Her husband is a full-time worker in the 110ur and oil milis. This makes her ¡he de jacto head ofhousehold. She supervises ¡he farm and makes decisions regarding what crops and varieties to grow. Three years ago, she grew mostly local varieties because of a lack of irrigation facilities. We gave her seeds of NDR 97, a new variety, which she planted on 0.10 ha of land. Later she increased the area planted lo this variety lo 0.5 ha. She told us the positive traits she likes in this variety, such as suitability ro her land type, good taste, shorter duration, good milling recovery, ease of threshing, and medium height, and negative traits such as less rice straw: 1 tried many varieties since the las! jour lo five yeors such as Saket4 and NDR80. but be- cause they were damoged by drought and disease, 1 slopped growing Ihem. 1 shifted back lo a local variety [ARIj although it does not laste good, has poor míllíng recovery and coarse grains. But 1 like NDR97 because of its suitobilíty lo my land, good laste, and shorter dura- lion. The only problem is Ihat it produces less biomass [strow}, which is no! enough for my two bul/ocles and five buffoloes. We need more straw for Ihe animols Ihroughoul the year. We a/so growcurbi [greenfodderj ond harvest them green during the kharif season. Due 10 the early duration ofNDR97. we can cultiva te our land for early rabi crops such as oilseed and vegetables before wheot. lolso like the taste ofNDR97 and 1 am satisfied with its mill- ing recovery. Jt is also easy lo thresh; il is neither very tall nor short. Mrs. Savitri Devi is 45 years old, illiterate, ¡md a full-time farmer oflhe backward caste. She culti- vates 1.1 ha of land in Mungeshpur. She has two types of land, up1and and lowland. She grew NDR359, Sruju52, and Jallahri in .1998. We gave the new seeds ofNDR359 to her in 1996. She prefers Ihis variety because it has a good taste and short duration. She describes Iheir use of NDR359: 1 don 't like the taste of Sarju52. lt is coarse and does not rema in sofi afier cookíng. Jt also does not have many broken grains afier milling. So we sold Sarju52 and used NDR359 for home consumption. One thing 1 noticed with the straw ofNDR359 is that it is sofi. so instead of storing it for a long time, we had lO feed ít immediafely fo our anima/s. 1/ we keep the straw for two to three months, it will not be very easy lO cut and the animals will refuse to eat ít. lnstead ojleaving the rice stalles to dry in the fie/d. which is our usual practice, we imme- diately thresh afier harvesting. lis short duration also enab/es me to grow another crop dur- ing the rabi season. Mrs. T. B Singh, 50 years old, belongs to the upper caste. Due lo labor shortages during the peak season and Ihe lack of male labor (her husband is fully engaged in a nonfarm job), she has been forced to provide physicallabor in most of Ihe rice operations. She was able to finish five years in school. She is the decision maker in !he household and is quite knowledgeable about farming. In 1997, she was one of Ihe collaborators of!he project. Afier testing 13 genotypes on her field, she obtained 5.2 tons per ha from PVS5 (NDRSB9730015), so she decided to continue to grow Ihis va- riety and expand the area during Ihe 1998 kharif season. She expecled to get six tons per ha, but be- cause of drought, Ihere were many unfilled grains. She told us about Ihe variety's positive traits aside from its high yield: 1 prefer PVS5 because of Its medium duration; medium bold, cylindrical grain; resistance to pesls and diseases; and better mil/ing recovery. In 1995, we gave her new seeda of BKP246. 190 T.R. Paris el al. 1 like this variety too because it is suítable for the lowland rainfed area, has good yields, and is not susceptible to diseases. I like the size and the shape ofthe grain-medium and boldo It a/so has the best milling recovery and commands a high price in the market. In 1998, 1 sold four quíntals of paddy at Rs 400 per quintal, while the o/her varieties are Rs 50 less than BFK246. We use Sarju52 and Saket4 for home consumption. Saket4 has fine graíns and ma- tures early, a trait ideal for the uplands. Our agricultural workers prefer coarse graíns, which last longer in the stomach than paddy with finer grains. I observed that the quantity ofstraw ofBFP346 is less, but grain quality is more important 10 USo Mr. Bansat Lal , 42 years old, an ilIiterate father !rom the backward caste, is a full-time fanner. Hís sons are fully engaged in nonfann activíties and his daughter-in-Iaw supervises fann activities and takes part in decision making. In 1997, he was a collaborator in the plant vanetal-selection program and obtaíned good yields. After threshing and mílling, the female members of his household al so agreed that the PVS5 (NDRSB97300 15) and PVS6 (NDRSB9730020) should be grown the follow- ing year. Both Mr. Lal and his daughter-in-law have the same criteria for selection, such as better yield, good qualíty of straw, medium height, resistance to pests and diseases, longer and fine grains, no broken grains after milling, softness and expansion after cooking. My daughter-in-law observed that PVS5 is easy to hull through hand pounding afier par- boiling. Jt is a/so good for puffed rice. Mr. Lal shared the seeds ofPVS5 with other fanners. In 1998, he cultivated PVS5 and PVS6 on his 3 bigha (0.3 ha) land area. He was able to obtain a yieldofsix quintals per bigha inone pIot and four quintals in another plot. These yielils were higher than those in nearby fields. Conclusions Socioeconomic surveys revealed that a major determinant of vanetal choice is the eonsCÍous attempt of fanners to match vaneties with the land type. Each field position in the topo-sequence corresponds to a risk of drought or submergence. In Mungeshpur (shallow and submergence- prone) fanners' eritena for selecting rice vaneties are associated mainly with duration (short to medium), for growing rabi crops after rice in the upland fields, and with better yield. A second determining factor is the adaptation to different user needs: food, livestock fodder, thatching, and cash. A third detennining factor is related to different postharvest operations like ease ofthreshing, good taste, high mil!ing recovery (above 65%), good storage capacity, and premium market price. Gender-specific roles and responsibilíties also determine vanetal preferences. For example, women prefer medium or semi-tal! vaneties that are easier to thresh, as well as vaneties that have a good quantity and quality of rice straw for livestock feed. Moreover, they prefer vaneties for the specific rice products that they make. While it may be difficult to combine all their preferred traits into one unique vanety because of genetic correlations, it is ímportant that both men and women have a "basket of choices" of vaneties suited to their needs and agroecosystems. elearly, Iistening to fanners' perceptions and involving both men and women fanners in selecting rice varieties at the early stage of breeding can lead to faster adoption of varieties suited to their specific rice ecosys- teros and diverse needs. 191 ｌｩｳｴ･ｮ￭ｮｧｴｾ＠ .. f.armers· Perceptions through Participatory Rice Varietal Seleerían References Adesina A.A. and J.B. Forson. 1995. Fanner,' perceptíons of new agricultural technology: Evidence from analysis in Burbna Faso and Guinea, West Afrie •. Agricultural Economics 13 (1995). CIAT. 1997. New frontiers inpartícípatory research and gender analysis. Proceedings ofthe interna¡ional seminar on parrícipatory research and gender analysisfor technology developmen(, Sept 9-14, 1996. CGIAR Systemwide Program on Partícipatory Research .nd Gender Analysis for Teehnology Deyelopment and Institutional Inno- vation Publication No. 294. e.lí, Colombia: Centro lotemacional de Agricultura TropicaL Courtoís, B., B. Bartalome, D. Chaudhury, G. MeLaren, C.H. Misra, N.P. Mandel, S. Pandey, T. Paris, C. Piggin, K. Prasad, A.T. Roy, R.K. Sahu, S. Sarkaruog, S.K. Shanna, A. Singh, H.N. Singh, O.N. Singh, R.K. Singh, R.K. Síngh, S. Singh, P.K. Singh, and B.V.S. Sisiodia. 2000. Participatory varietal seleelÍan for low input environ- ment$: A case study in eastem India. Euphytica (forthcoming). Eyzaguirre, P. and M. Iwanaga. 1996. Fanoers contribution to maintaining genetic d;versity in cropsand its role within the total genetie resources system. In Participatory plant breeding. proceedíngs of a workshop. 26--29 July 1995. Wageningen. Netherlands, edited by P. Eyzaguirre and M. Iwanaga. Rome: lntemalIonal Plant Genetic Resources Institute. Gupta A.K., K. Pate!, P.G. Chand Vijaya Sherry, A.R. Pastakía, S.S. Shukla, D. Koradíya, V. Chauhan, A. Rayal, C. Srinivas, .ud R. Sinha. 1997. Particípalory research: WiII the koel hatch the crow's eggs? In New frontiers in participatory research and gender analysis. Proceedings of the international seminar on porticipatory re- search andgender analysis for technologydevelopmen/, Sept 9-14, 1996. CGIAR Systemwide Program on Par- tieipatory Research and Gender Analysis forTechnology Development and Institutional InnovalÍon Publication No. 294. Cali, Colombia: Centro Intern.cional de Agricultura Tropical. Kshírsagar K.G., S. Pandey, and M.R. Bellon. 1997. Farmer perceptions. varietal charac/eristics and technology adoptivn: The C(ISe of rainfed rice iñ a vil/age in eastcm India. Social Seienees Discussion Paper 5/97. Makati Ciry, Plúlippines: Inlemational Ríce Research Instituto. NDUAT. 1993. Progress report offarmer participatory plant breeding in the key sites of Faizabad and Siddharth Nagardistricts, easlem India, 1997-98. Faizabad, India: Narendra Deya University of Agriculture and Tech- nology. P.ris, T.R., A. Singh, J. Luis, M. Hossain, H.N. Singh, S. Singh, andM.N. Síngh. 2000. Ríeeplant breeding and varietal .eleelion: Prelíminary results from eastem India. In Proceedings ofthe workshop on impact ofparticipatory re- seareh and gender analysis, September 1998. Quito. Ecuador, edited by J. Ashby and L. Sperling. Cali, Colum- bia: Centro Internacíonal de Agricultura Tropical. (Forthcoming.) Satheesh, P.V. 1996. Genes, gender and biodiversity: Deccan dov.loprnent socioty'. eommunity genebanks. In Using diversity: Enhancing and mainfaining genetic resources on"¡arm: Proceedings of a workshop heM on 19-21 June, 1995, New Del"i. India, edited by L. Sperling and M. Loevinsohn. New Dellú: IntemationalDevelopment Researen Cenler. Sperling L. 1996. Resu1ts, methods and institutional issues in participatory selection: The case ofbeans in Rwanda.ln Participatory plan/ breeding, proceedings of a workshop. 26--29 July 1995, Wageningen. Netherlands, edited by P. Eyzaguirre and M. Iw.nag •. Rome: lnternaliona! Plant Genet;c Resources Institute. Traxler, G. and D. Byerlee. 1993. Joint-product analysis ofthe adoption ofmadem cereal varie!Íes in developing coun- tries. American Joumal of Agricultural Economics 75: 981-989. Weltzien, E., M.L. Wlútaker, and M.M. Anders. 1996. Fanoer participation in pear! mille! breeding for marginal envi- ronments. InParticipatory plan! breeding. proceedings of a workshop, 26--29 July 1995, Wageningen, Nelher- lands, edited by P. Eyzaguirre and M. Iwanaga. Rome: Intemational Plant Genetic Resources Institute. !92 Opportunities and Constraints for Participatory Plaut Breeding: Farmers' Seed-Management Strategies and Their Effect on Pearl Millet Populations in Rajasthan, India Kirsten vom Brocke, Anja Chrístinck, and Eva Weltzien Abstract This paper presents information from a study on farmers' seed-management practices growing pearl mil- let in Raíasthan, India. It describes farmers' own crop-improvement activities in regard to yield, quality, and diversity of pearl millet, with emphasis on seed-management stralegies, such as introgression of modem varieties, selection, slorage, proeessing, e"chango, .nd proeurement. It .150 e"amines lhe farm- ers' definition of "variety" as compared lO lhe definition used by professional plant breeders. Far the study, farmers were divided into four graups, based on lheir seed-management practices. Dala were eol- lecled on specific trailS and correlated with grain yield under different c1imatic conditions. The pOlential and constrainls offarmers' practices are discussed, with emphasis on areas where researchers could con- centrate on specifie weaknesses that farmers' own seleclion practice. caonot effectively address. Introduction In many regions of the world farmers routinely produce seeds for their staple crops. 111is i5 partícu- lady cornmon in regions where agricultural production is affected by frequent and unpredíctable droughts, as in most areas where pearl millet (PennÍsetum glaucum [L.] R.BL),a cross-pollínating crop, is grown. Under these harsh climatie conditions, fanners have developed landraces that tend to show good levels of tolerance to these environments. 111e farmers have also evolved strategies for maintaining seed during drought years in order to safeguard food production and animal foddeL Given the fuet that formal plant-breeding programs have failed to develop superior varieties for marginallands and low-input conditions, the main objective ofthe study presented here is to better understsnd farmers' own seed-management pmctices as a basis for planning and implementing par- ticipatory strategíes that capitalize on fanners' local knowledge. This approach would aliow researchers to then conCentrate on specific weaknesses that farmers' own selection pmctices cannot effectively address. To date, these local strategies, including the fanners' needs and preferences, along with details of their cropping systems, are not familiar to scientists involved in conventional breeding programs. Kirsten vom Brocke ;, • PhD student al the Institute of Plant Breeding, Seed Science and Populatíon Generics, University of Hohenheim. Stuttgart, Germany; Anja Christinck is a PhD ,lUdent at !he In't;tule for Social Seiences of the Agricultural Se<:tor, Department ofCommW1ication and Extension, University ofHohenheim; and Eva Weltzien is a principal scientist witb !he Interna- lÍan.! Crops Researeh Institute for lhe Sem;-Arid Tropios (lCRlSA n, in Bam.ko, Malí (West Amea). The work jll'esented here is part of rhe project "Enhancing qu.lity, diversity and productivíty of fanners' peorl millet genetic 650 mm in the southeast (figure 2). In this study, we refer only to the westem part ofthe state, where farmers must make do with ¡ess than 350 mm of annual rainfall, with high variabilíty!Tom year to year. Experienced farm- ers often talk of a 10-year cyc1e in which two seasons have good raíns, two have severe drought with crop failures, and the rest usually have fair to good seasons. Soils are mainly sandy, and sand dunes are common. VilIages are typically scattered across wide areas. Pearl míllet is grown tbree to four months during the monsoon season, mostly in mixtures with other crops, such as legumes and cucurbits. Animal husbandry is another important par! of tbe farmíng system. Social conditions in tbe víllages are govemed by the caste system. Even today, the caste system stilllargely determines people's social status, occupation, income, and access to education and information. TM ｾｭ｡ｬ＠ ｾｲｩ･･＠ of lndle dep!cted in this l1Ul9 a/'é l'I#li1tJer (;(ItNid l\Of ｡ｾ＠ Figure 1. The state of Rajastban in the north- west ofIndia 194 Annual rajnfall in RaJasthan \ I , chUIIJ I ｡ｩｫ｡ｾ＠ J I 350-550 mm Sikar " , / NaQaut I ｾ＠ ,Jodhput 1/ • ｓ｡ｷ｡ｩｾ＠jI ＯＯｾｔｲｩ［Ｇ＠ ｡ｾ＠ Pia" , / The extemaI boutldario!$ ｯｬｾ＠ ｾ＠ in ｾ＠ mep are neMher t.tlI'f'tId: ｮｯｲｾＮ＠ Figure 2. District capitals and zones of mean annual rainfall in the study area K. vom Brocke, A. Chrlstinck, ami E. Welttien Farmers' seed-management strategies Farmer's concept 01 a "variety" Farmers' seed management can only be evaluated if one fully understands the farmers' concept of a "variety." Ims term, as understood by plant breeders, does not seem lO be fully appropriate for the farmers' pearl míllet seed system in west Rajasthan. In order to learo how farmers perceive "variet- ies," informal interviews as well as classification and ranking exercises were carried out during workshops with farmers from the study villages. Care was taken to inelude both female and male farmers in the interviewing process. The results demonstrate tha! environrnental adaptation was the main eriterion for farmers' c!assification of pearl millet plants in westem Rajasthan. Potential uses and quality aspeets further eontributed to the farmers' method of grouping different plant types (Christinck and vom Brocke 1998). Traditionallandraces that have adapted to the environment show a high basal and nodal tillering ability, indicating toleranee to drought and low requirements for soil fertilíty. If these eharaeteris- ties are combined with tmn stems, narrow leaves, and thin, compact panicles with srnall grains, farmers will conclude that sueh a plant will grow under low-input eonditions (Le., in their fields) and produce grain and straw of high nutritional quality. In contrast, the characteristics of modem varieties are low basal and nodal tillering ability, tmck stems with broad leaves, and Iarge panicles with relatively large grains Iha! are mostly round in shape. From the farmers' experience, this plant type is not toleranl lo drought stress, requires higher soil fertility, and has inferior food and fodder qualities. Farmers, however, are aware that pearl millet plants showing such characterístics can produce higher yields under favorable conditions (Chrístinck and vom Brocke 1998). Farmers are therefore concemed about the composition oftheir seed stocks, i.e., wmch plant types and, thus, which properties are present. Farmers expect plant types to change over time, in reaction to envi- ronmental conditions such as soil quality and raínfall, so Iha! the seed stock generated in one year cannot be exactly reproduced the next season. They have a strong concept of continuous interac- tions between plant type and environmenl, as evidenced by their belief, or experience, that any pearl millet cultivar, including modern varieties, that is grown in their fie1d for sorne years will eventually become like their local cultivars. Contrary to the views of professional plant breeders, the farmers' concept of a "variety" is not that of a population with more or less uniform and stable plant characterístics based on its genetic back- ground; the term "variety" is applied to a plant type that is evolving under or adapting lo certain environmental conditions. This concept is reflected in fue farmers' seed-management strategies. What is seed management? Seed management comprises all activities of a farming faroily that influence their seed stock, in- cluding introgression of modern cultivars (open-pollinated varieties or hybrids), seed selection, processing, storage, exchange, and procurement. In this paper, we refer mainly 10 seed selection and processíng, and the ways in wmch farmers deal with modem varieties frorn the market. Ways 01 se/ecting or processing seed Farmers in Rajasthan generally employ two main selection methods. The first is winnowing or grading, which entails cleaning and separating seed grains. The rate of selection can vary greatly. It may be lhat only 10% of the threshed and stored grain will be rejected (rnainly husks and broken 195 Opportunities and Constraints for Participatory Plant Breeding and insect-infested grains), or more than 50% ifthe grains, for example, are small and not fully de- veloped. Generally, the smaller grains are be used for food. The second method, which is also very cornmon, is the selection of panicles that show preferred traits. Farmers usually select for panicles on the threshing ground afier the panicles have been sepa- rated from the straw, although sorne farrners prefer to select for panicles in the field before harvest- ing, taking the entire plant into consideration, e.g., number oftillers, height. Even by inspecting the panicle, farrners can envisage what the plant's other characteristics looked like (or would look like when regrown). Many farmers do not perforrn panicle selection every year, but only in the better seasons, which usually occur every two to four years. In harsher years, they are most likely to use the winnowing/grading method. A third, less cornmon, forrn of selection is to use the harvest of a preferred field-a field considered to be more fertile than others-for sowing the following year. Using "improved varieties" or hybrids from the market If a farrning family does use pearl millet seed from the market, in most cases it will be mixed into the family's own seed stock. In western Rajasthan, farrners without access to irrigation facilities generally do not grow improved varieties or hybrids in pure stands. Market seed is mostly certified or "truthfully labeled" seed. Further advanced generations of such seed can be optioned from the market or from other farrners. This grain is not labeled and its origin is ofien unknown. There are two ways in which farmers use seed from the market: l. Occasional introgression of new seed from the market into the previous year' s seed stock: the resulting crop consists of many different plant types (traditionallandrace, market variety, and several generations of progeny). Mixing ratio and frequency can vary widely, ranging from 1:10 up to 50:50. 2. Regular introgression ofnew seed from the market into the previous year's seed stock, select- ing for desired plant types among outcrosses: One or more new plant types will become domi- nant, and the variability of plant types is less than in the first example. The amount and frequency of mixing new seed, as well as selection intensity, can differ greatly from farmer to farrner and from year to year. It is important to understand that most farmers do not use improved varieties to replace their own seed, as is ofien assurned. Rather, they use new seed to increase the variability of plant types in their fields, thereby creating new options for their strategies of selecting for preferred plant characteris- tics, including grain and straw yield, food and fodder quality, storability, drought tolerance, early maturity, tolerance to adverse weather conditions (heat, sandstorrns, thunaerstorrns), and resistance to bird or locust damage. Social aspects of seed management The availability of seed grain at the onset of rains is very important for farmers in western Rajasthan. The success of a crop depends very much on sowing irnmediately afier the first rains of the monsoon. For centuries, farmers have had to deal with crop failures due to severe drought con- ditions. Therefore, "taking care ofthe seed" is considered to be of great importance. Farmers who can successfully maintain their own seed, or be in a position to provide other villagers with seed in times of scarcity, are considered to be good farmers and are respected by al!. There is a special caste in most villages for whom maintaining seed and sharing it with others is considered to be a tradi- tional obligation. Nevertheless, other farrners can also build up a reputation for owning good seed, 196 K. vom Brocke, A. Christinck, and E. Weltzien and "lending" or sellíng il to others. Seed management is, therefore, related to aspects of caste and status in vilIage life. Furthermore, ít ís a gender-related actívíty. Selecting the seed, storíng it, and processing it before sowing is traditionally done by women, whereas soil preparation and sowing ís usually done by men. Men also often participate in harvesting, and depending on the family, they can be equally involved in selecting seed. Buyíng seed from the market and obtaining information about market varieties is done almost exeJusively by men. Diverse seed-management strategies co-exist ín villages in western Rajasthan, reflecting the diver- síty of socíoeconomic conditíons: farmers who grow traditionallandraces with or without selec- tion; families who mix, sometimes orregularly, seed from the market ínto the landrace seed with or without selectíon; and familíes who sow the pure seed of markct varieties. All these seed-manage- ment strategies can be found in one village. Even though pearl millet is a cross-pollinating crop, it seems to be possible for a village cornmuníty to maintain a diversity of plant types. The reasons for a farming family using a certain strategy can only be partly explained by soíl conditions and c1ima- tic factors. Other important factors seem to be the size ofthe landholding (market-oriented or sub- sistence-oriented), the number and species of animals and their fodder requirements, the aecess to cash income or loans to buy seed, the family tradition and knowledge, and access to information on new varieties, e.g., literacy and mobility. Most ofthese socioeconomic conditions are related to the caste system in Rajasthani villages. Quantification of the effects oC farmers' seed-management strategies Material and methods To quantif)' the effects of farmers' seed management, 69 graín stock samples were collected from 16 farmers located in four different villages in westem and central Rajasthan during 1995-1997. Samples were characterized by the farmer, e.g., as separated seed grain and food grain, and were classified into four main seed-management strategies (rabie 1). These grain samples from farmers, along with 12 modem varieties known to be grown in these víllages, were evaluated under varyíng drought-stress conditions at three research stations in westem Rajasthan (Mandor, Jodhpur, Palí) between 1997 and 1998. Climatíc conditíons in 1997 were generally favorable, whereas in 1998 severe drought affected the plant growth, especíally at Mandor. The fie1d trials comprised 81 entries and were laid out in lattice designs with five replications. The different plant traits that are used by farmers and scientists 10 describe the performance of pearl millet were recorded in order to assess productivity and characteristícs of entries. These plant traits inc1uded noda! tillering, leaf shape, stem diameter, panicle girth, number of productíve tillers, grain weight, straw and grain yield, as well as diversity of plant types withín one entry. Table 1. Farmers' Seed-Management Strategies as Represented in Field Trials LR Mainlains only locallandrace seed without introgression 01 modem malerial SeleClion method mainly winnowlng IGR1 Occasionally introgresses modem varieties into iandrace Seiection method malnly wlnnowlng IGR2 Introgresses modem malerial more regularly than strategy IGR1 Seleels regularly/frequenlly for panicles MV Modem varietíes 197 Opportunitifi!3.. and Constrainls lar Participatory Plan( Breeding Separate analysis of the five test environments revealed a significant phenotypic relationship be- twcen grain yield and plant characteristics (table 2). The number ofpanicles and basal tillers, plus nodal tillering and phenotypic diversity ofplant types within one entry, were al! positívely assocí- ated with grain yield in the stress envíronments and negatively associated in the non-stress environ- ments. Conversely, entries with large stems, large leaves and panicles, and bold grains showed negative correlatíon coefficients with grain yield under stress conditions and positive coefficients in the non-stress environments. Table 2. Phenotypic Correlation of Observed Traits witli Grain Yicld Traits Graio weighl Panicle girth Lcafwidlh Slcm diameler No. of panicles Tíllers Nodal lillering Plan! type diversity 'p < .OS. "p<.OL Favorable MAN97 J0091 0.69*· 0.75" 0.70** 0.83" 0.38- 0.62" 0.69- -0.54" -0.46" -O.54*" -0.58- -0.65" -0.57" Environments Mild terminal drought PAL97 0.42** . 0.42" 0.33" 0.41·· -0.41 0.01 -0.41- -0,36'" Early drought MAN98 J0098 0.08 -0.25' -0.60- -0.24' -0.62- -0.24' -0.65" -0.14 0.90- 0.48" 0.67- 0.36- 0.56" 0.27* 0.32'- 0.11 A genotype X envITonment (GE) analysis based on grain-yield data was carried out in order to gain an overall view ofthe effects ofthese strategies on the adaptation offarmers' seed stocks to differ- ent environments. For this purpose pattem analysis was used to c1assify environments and to assess relatíonships between the entries and between environments, as well as 10 analyze the interrelation between entries and environments. To generate the analysis, the statistícal packet GEBEI was used (Watson et al. 1996). The details ofthis calculation will be published elsewhere. Results and discussion The phenotypic relationship described in table 2 shows the effectíveness of fanners' seed-manage- ment strategies. Entries with plant characteristics that farmers associated with adaptation 10 stress proved to be more productíve under stress conditions than other entries. These findings were sup- ported by the results of the pattem analysis. The analysis indicated that most of the entries classified as LR showed dose interactíon with the preflowering drought stress at Mandor and Jodhpur. Compared 10 the LR entries, entries classified as IGRI tended to show a less specific interaction with the stress envITonments. In contrast to the management groups LR and IGRI, a change in the adaptatíon pattem seemed to be obvious in entries derived from IGR2. The positive interactíon of the samp!es exc\usively with the preflowering drought environments was mostly eliminated. Entries also tended to show relatívely high productivity in more favorable environments. Samples grouped in IGR2 thus tended to perform fairIy well in al! the test environments. Entries labeled as modem varieties (MV), indicated almost no positive associatíon with the preflowering drought 198 K vom Brocke. A. Christinck, and E. Wellzien envirorunents. Ihe exceptions were sorne modero varieties with pedigrees based on landrace mate- rial from westero Rajasthan. Fanners who practice IGR2, which includes introgression and selection for contrasting plant types, are generaJly successful with this method. In the one seed stock, the IGR2 method produced traits indicating adaptation to stress as well as potential for high yield under favorable conditions. In terms of potential grain yield, this method appears to be effective. Sorne of the fanners' grain stocks generated by this strategy even yielded better with increased rainfall compared to the "pure" land- races (LR). It was the fanners' aim to introgress modero varieties so as to produce seed stocks that "take advantage" of good rains and it appears they have met their objective. Although "pure" landraces are not as productive under favorable conditions, they are more resílient under conditions of stress. For centuries they have been grown in heterogeneous envíronments. They therefore have the capacity lo adjust to the erra tic climatic conditions that occur in this region. Seed samples from farmers practicing introgression, in combinatíon with regular panic1e selection, seem to indicate that it is possible to improve a landrace population through newly introgressed variability. It also appears !hat if farmers use panicle selection to separate seed from food grain, they can improve their control over seed-stock performance. Summary and conclusions Potentials and constraints offarmers' own crop improvement The present study has revealed opportunities, as well as constraints, for farrners' own crop im- provement. Olher studies have assumed that landraces are mainly a product of natural adaptation and that fanners ofien do not, or only "unconsciously," select landrace seed (Damania 1996). How- ever, direct observation and interview data from this study have revealed that this view does not ap- ply to the case of pearl millet in westem Rajas!han. The results of this study confirm !hat different seed-management strategies are practiced in the one village. Sorne fanners malntain the localland- raee with superior quality and yield stability, while others create variability through introgression ofmodero varieties. Furthermore, previous studies carried out in westero Rajasthan also show!hat fanners use theír own sophisticated strategies for seed managernent and crop ímprovement (Dhamotharan et al. 1997; Weltzien et al. 1998). Quantitative data from field trials proves that these farmer strategies lead to popuiations with díverse plant types. Ihis diversity offers possibilíties of recombination in the population and natural selection, and also increases the gains of fanners' selections. Seed selection, especialIy intensive selection of plant type or panicle, enables fanners to exert con- trol over the negatíve effects of introgression. F anners select aeeording to their various breedíng goals, such as yield stability under stress conditions and higher productivity in regard to straw and grain yield in the target environment. These selection strategies are largely guided by theír concept of a variety. Mainly fanners who practíce introgression along with panicle or plant type selection are able to ímprove the productivity of their landraces without losing yield stability. However, other results show that traditional methods of seed selection practiced before the introduction of ex- otic material, such as winnowinglgrading, can lead to a decrease in the expression ofadaptive traits and characteristics in the typicallandrace phenotype. This is due to seed winnowing and the use of Ihe bolder grain for seed purposes. Smaller grains (representing adapted landrace types) are re- jected, whereas bigger grains (representing less adapted modero material) remain in the seed stock. 199 Opportu"ities and Constrai"ts ror Participatory Plan! Breeding It should also be taken into consideration that the fanners who benefit from the higher yield poten- tial of the introgressed cultivars are mainly those who have relatively good Iand and resources. These farmers are traditionally those who distribute seed material to other, poorer, farmers in times of scarcity. As poor fanners usually have less fertile land and less manure, the properties of the originallandrace pearl millet are ideal for them. If better-off farmers continue lo use introgressed seed, which requires better land and continuous selection lo assure yield stability, the availability of landrace seed may decrease for poor farmers with marginal lands unless measures are taken to maintain the originallandrace plant type. Finally, farmers often show a lack oftechnical knowledge conceming the genetic material tha! is available on the market. For instance, most farmers are not aware ofthe differences between hybrid varieties and open-pollinated varieties, nor are they aware of the consequences of using hybrids in seed production. Role of researchers These constraints point to several possible ways in which researchers can help to improve farmers' seed stocks in westem Rajasthan. Researchers could take on an advisory role and support fanners in their own crop-improvement strategies, for example, with technical knowledge or explanations of the effects of different management strategies. The plant breeder could recommend material that has the ability to combine with local material and has the potential to achíeve genetic gains in farm- ers' preferred traits. Material should no! rnere\y be handed out to the farmer by the breeders. A material exchange between farmer and breeder should also be supported. Breeders could help to improve those traits that farmers have difficulty working with, e.g., specific resistance or seed-set improvement. Where farmer and breeder both provide material and resources, intellectual property rights should be respected. Results from this study show that farmers in westem Rajasthan are acti vely working on developing and improving their seed stocks, and that many opportunities exist for fruitful collaboration be· tween farmers, plan! breeders, and other scientists. References Christinck, A. and K. vom Brocke. 1998. Evaluating pear! millet cultivars with fanners. In Participatory Plant lm- provement. Proeeedings of MSSRF-ICRlSAT Workshop, edited by V. Arunachalam. Chennai, India: M.S. Swaminathan Research Foundatíon. Dhamotharan, M., R.E. Weltzien, M.L. Whitaker, H.F.W. Rattunde, M.M. Anders, L.C. Tiagi, V.K. Manga. and K.L. Vyas. 1997. Seed management strategíes of farmers ín westem Rajasthan in Iheír social and envíronmental contexts: Results from a workshop usíng new communication techníques for a díalogue between farmers and scientL¡ts. 5-8 February 1996, Digadi village, Jodhpur district, Rajasthan, india. Integmted Systems Project Progress Report No. 9. Andhra Pradesh, India: International Crops Research Institute for Ihe SemÍ-Arid Tropics. Damanía, A.B. 1996. Biodiversity conservation: A review of options complementaty to standard ex situ methods. Plan/ Gene/ic Resources Newsletter 107:1-18. Weltzien R.E., M.L. Whitaker, H.F.W. Rattunde, M. Dhamotharan, and M.M. Anders. 1998. Particípatory approaehes in pearl millet breeding. In Seeds 01 choice: Makíng the most of new varieties for small farmers, edited by J.R. Witcombe, D.S. Virk, and J. Farnngton. New Delhi: Oxford & IBH Publishing Ca. Watson S.H., l.H. DeL.ey, D. W. Podlích. and K.E. Basford. 1996. GEBEl: An analysÍ1i package using agglomerative hierarchical classijicalory and SVD ordination procedures for genotype x environment data. Research Report No.57. Brisbane: Department ofM.them.tics, The University ofQueensland. 200 Strengtb of Farmers' Knowledge and Participation in Crop Improvement and Managing Agrobiodiversity On-Farm P. Chaudhary, SP. Khatiwada, and K.D. Joshi Abstract This paper highlights the role offanners iD crap improvemeDt and managiDg agrobiodiversity. The fiDd- iDgS are mostly based 00 focus-group discussions aDd tield observations. Documentation of faooers' knowledge and experienees in erop ímprovement and managíDg agrobiodiversity may serve as a refer- enee for individual breeders or inslitutions involved in participatory erap improvement through differeDt strategies like particípatory plant breedíDg, particípatory varietal seleetion, Of partieipatory geooplasm enhaneernenl. The strength of participatory crop improvement is that there is rnultistage inyolvemeDt of farmers, from parent seleclion through to eultivation and selectian ofplanting materials, because faooers have a wide range ofknowledge .nd experience, and they are the end-users as well. Sine. ancien! times, fanners have been dependent upon lhe Iraditional seed-supply system, which slill aceoun!s for over 90% of Ihe secd requirement in NepaL A variety of mechanisms like varietal selectinn, seed seleetion, seed proeessing and storage, and Ihe seed-flow system have contributed lo crop development, creating agrobiodíversity on-faoo. More reeently, participatory germplasm enhancemen! has arisen as a new stra!egy to enhance lhe germplasm ofloeallandraces, whieh will no! only empowerfarmers in improving lbeir landraces bu! .Iso strengthen in situ conservation of sueh landraces on-fann. The curren! need is to incorporale farmers' relevanl knowledge and use it in lhe oyerall crop-improvement process. Key words: Participatory, crop improyement, agrobiodiversity, germplasm, on-fann, and knowledge Background In many developing countries, farmers playa pivotal role in the conservation of gene tic resourees, thus maintaining biodiversity, sinee they hold the bulk ofthese resources (Worede 1992). From time inunemorial, farmers have experimented with naturally existing genetic variations in their own environments to produce present-day landraces (Sthapit and Joshi 1998). Farmers have grown, tested, utílized, developed, and finalIy, selectednew varieties and crop combinatíons to suit particu- lar ecosystems. The role of farmers in creating agrobiodiversity is also evident from their involve- ment in seed storage and seed exehange. Of eourse, the need and preferences of individual farm families have driven them in ,the selectíon of crop species. For this reason, they have acquired a profound knowledge about landraces and niche-specific placement. Given the inherent advantages oftraditional practiees, on-farm landrace conservation and enhanee- ment provides a valuable option for observing genetic diversity (Worede 1992). A large number of subsistenee farmers still use traditional methods. Those using modern teebnology account for approximately 40% of global agriculture, while rest is under traditional agriculture, which provides between 15% and 20% of the world' s food (Franeis 1986; Sthapit and Joshi 1998). The mos! impor- tant factors that motivated fanners to diversifY crop and Iivestock in the past were probably ensur- ing Iivelihood and meeting qualitative preferences and requirements (Roder 1995; Sperling and Berkowitz 1994; Sthapit and Joshi 1998). Roder (1995), has reiterated the faetors motivating farm- ers in maintaining diversity as follows: p, Chaudhary is a site officer and KD. Jomi is a program officer with Loca) Initiatíves for Biodiversity Research and Development (U-BIRO). S.P. Khatiwada is a senior ",ieDtlSt with NRRP, Nepal Agricultura! Research Councíl (NARCl. 201 Strength o{Farmers' Knowledge and Participation in Crop lmprovement • Ihe need for high self-sufficiency due lo communication problems • reduction of risk factors • labor considerations • lack of availability of suítable ímproved varieties • market fluctuations • traditional food preferences • specíal requírements for ceremoníes and rítuals One of the commítments made in Leípzig in 1996 during the NGO eonference on the access and control of agricultural biodiversity was to enable the formal sectors, through trsiníng, to recognize the value offarmers' and indigenous peoples' knowledge and practices in conserving and strength- ening agricultural diversity, The following statements further stress tha! the documentation of farrners' knowledge and participation in crop improvement ís essentiaL • To be able to define precísely the objectives, límits and means for implementing in situ con- servation, it ís necessary lo obtain a better understanding of the structure of polymorphism witbin farmers' varieties, ways it evolves with farrners' practices and the methods and mech- anism for managing Ibis source of diversity, (FAO 1989; Brush 1992; Louette and Smale 1996) • Recognizing farmers' knowledge and the farmers' role in developing landraces and main- laíning theír genelic diversity through Ihe partnership of farmers wíth formal science insti- tutes ís an important step in enhancíng Ihe maintenance ofbiological diversíty, agricultura) sustainabilíty and food security at the farrn, regional and globallevels. (Teshome 1997) This paper bigh1ights the role of farmers in crop improvement and agrobiodiversity management The different stages of crop development and different approaehes applíed to bring about current agrobiodiversity are explicated in the following chapters, The examples are maín1y from one ofthe sites of the project "Strengthening the Scíentific Basis of in situ Conservation of Agrobiodiversity On-Farrn" being implemenled in Nepal jointly by the Nepal Agricultural Research Council (NARC), Local Inítiatives for Biodiversity Researchand Development (LI-BIRD), and the lntema- tional Plant Genetíc Resources lnstitute (IPGRl), Farmers' role in crop improvement Crops have trsveled through different stages of natural evolution and systematíc crop breeding, Breedíng by different groups, such as routine seed selectíon by farrners and formal breeding in pub- líe and private institutíons, has played an important role in bringing erops and varieties to their pres- ent status, Crop specíes have been adapted to different agroecological conditions while evolving from a wild to a cultivated fonn through more refined landraces because of farrners' selectíons, Farrners' landraces have been extensively used to develop improved varieties through breeders' efforts and again through diffusion through formal and informal institutíons. Gene flow from wild relatives to farrners' landraces and from landraces to improved cultivars is a dynamic process and should be maintained if plant breeding is lo meet the growing needs of the world's populatíon (Vaughan and Stich 1991; Slhapit and Josbi 1998). Thís ís why the conservation ofplant genetic resources in situ has very recently been widely accepted by several formal and informal institutions 202 P. Chaudhary, s.P. Khatiwada, and K.D. Joshi worldwide. The inclusion ofa landrace as one ofthe parents in participatory plan! breeding and the involvementoffarmers in several stages ofits development is imperative ifthe needs offarmers are to be accurately met, leading to a successful conservation strategy. The figure below outlines the stages and the processes through whích crops have traveled and the important role played by farm- ers to make the story successful. Stages of deveJopmant Process Wlld relativas Ni ｾ＠ rl OomesticatKm u! RI A: LI Farmers' varAR H Experimentatlon Experimentation Oif!usion Farmers' role ｾ＠ Exploratlon, coUection, testing : for local adaptation and vbhty Expeñmentation. testing for adaptation, verification, "tedion, nomendahJre ｾ＠ exchang$, purchase, 9ift, _ing, theft s ｾＭｴｻｾｾｾｾ］ｾ＠EL ' -----------, Conv, breeding :-- , H)lbridiZation. Improvement, Pareru selection. need and praference, Sék!ctlon, regeneration E mass, and Pl setectlon e ｉＭｾｾｐＺＺＺｃＺＧ＠ ＨｾｐｾｐｂｉｐｾｖｓＯｐＺＺＮＢＺｾｇ］ｅＩｾ＠ｔｾ＠ I ｾ＠ ＱＭｬＭｌｾｎＢＢＮｷＺｶｾ｡ｾＢＬＬｾｴｹＬＭ｟ｾ＠ Extension (E il1$t.fdevl NGO} , Adaptation. diffusíon Figure 1. Farmers' roles in the crop-improvement process Nomenclature 01 traditiona/ varieties Farmers have given names to their traditional varieties of different crop species based on their ¡den- tifYing characteristics, whích can either be external appearance or internal quality. For sorne ofthe landraces, one can easíly dístinguish one from another on the basis of their names. Farmers' nomenclature has a scientific basis since words lhat constitute the name have an important meaning lhat reflects the characteristics ofthat variety. For instance, lal tengan is one landrace; it has been named for its red (lal) lemma and palea color and a long, stout tentacle/spur (a type offish called a tengar has spur like this). A few examples ofthe names offarmers' varíeties and theirmeanings are presented in table 1. On-farm varietal diversijication Varietal replacement has been taking place with the introduction of modern varíeties for several years, starting from the Oreen Revolution in Asia during the early 1970s. In many regions ofthe world, farmers have economic incentives to replace the varieties that have evolved witrun theír own ecosystems with ímproved, introduced varieties (Louette and Smale 1996). Landraces seem so 203 Strength o(Farmers' Knowledge and Participa/ion in Crop Improvemenl Table 1. Name and Meaning of a Few Selected Landraces S.No. LR Nome Type Name & meanlng Easy way to identíty/distinguísh 1 Nakhí sara Bhaasiya Nakh¡=awn, Long awn; yellowtsh lemma and palea (UP) Saro=bhadaiya type 2 Bhadaíya basmati · 8hadalya= .arly seasoned, Slightly ｲ｡ｮｧｯｾｬ￭ｫ･＠ color; fine grained 8as;:aroma 3 BaSInatJ' Aghani 8as=aroma Like B, basmati: long panicle length; fine graio; aroma; awn on a few grajns 4 Lal tengar · Lal=red, Reddish UP color; boid graín INith long stout awn; Tenga,. type 01 fish with grown rn shallow water stout spur 5 Amaghauj · Ama=guava, Ghauj=cluster Yeflowish grain; two to fOUf grains originating from a single point giving ｣Ｇｵｳｴ･ｲｾｬｩｫ･＠ look; long and strong stalk 6 Dudhraj · Oudh=milk Whitish UP color; m¡¡ky-white grain 7 La/ka fararo · Lal=red, Yellowish UP color with minute reddlsh stripes Faram= research ¡nstitutian 8 Harinker · Harin=spotted deer i IJP during milking and dough stage lookslike .pot-I ted deer; small round grajo. 9 Parewa pankh · ｐ｡ｲＮｷ･ｾ＠ plgeon, ｾｴ･ｲｩｬ･＠ lemme l. long, ooverlng tIle graio Irom Pankh=feather both sides 10 Kariya keroadh · Karlya=black fine grein; bladaIQ. nakhí sara, mansa,., I I -! Bloticlablotle stress Bhathi, nakhl ssro, snga +- aH oIher varieties I ｾ＠ SocloeconomlesJ ReSOOice l'lCh Bflsmali I - -,.,.. I Re$0Uf'(';0 poor No basmati cuJtivated I -l AvallabUIty Bhltthi, amagMuj, sakhat, faJ ｴ･ＬＬｾｲ＠ (1)rought fTóm n.lghboring dlstrtctS) : aJl modm1 v,atietias I Figure 2. Agroecology and human-induced selection pressures on crop genetic resources Participatory plant breeding Partícipatory plant breeding (PPB) is widely used by different institutions, both government orga- nizations and nongovernment organizations, and even by farmers. However, farmers' participatíon in PPB varies. The approach and methods ofPPB are described in detail by IPGRI (1996: 57-65), Sthapit, Joshi, and Witcombe (1996), and Witcombe el aL (1996). However, the stages where farm- ers' involvernent is most ímportant are plant selection, germplasm enhancement, seed selectíon, and management (Joshi et al. 2000). Table 2 summarizes the range offarmers' partícipation in the PPB process. Prospects for germplasm enhancement with farmers' empowerment The germplasm of local landraces can be improved through pure-Hne or mass selection with the active partícipatíon of farmers and modest technical backstopping from formal institutions for most of the processes. This can be achíeved through farmers' active partícipation, with mínimum costs and Iittle effort for breeders. At the same time, the genetic potential oflocallandraces can be con- served by encouragíng in situ conservation. Farmers at Begnas, Kaski and Kachorwa Bara have recently taken the initiative for participatory germplasm enhancement (PGE) through pure-line selection. In these areas, farmers' knowledge about seed selection and storage were first documented. Gn the basis ofthis information, the farm- 205 Strength o( Farmers' Know/edge and Participation in Crop Improvement Table 2. Level of Participation in Different PPB Processes Cilation Modes .of partieipation Level 01 participation byfarmers Witcombe (1996) Consulta!ive Rasearcher con su lIs farmers lo assess needs, sel breeding 90a1s, and choose lesling sites, but researcher retains key decision making CollaboraUve Expert larmers grow early, variable generations and seleet best planls on !helr own fields McGuire, Maoiead, Farmer-Ied PPB External agents support larmers' own system 01 crop and Sperling (1999) development Forrnal-Ied PPB Farmers joio in lonmally initialed process 01 crop development ers were next given an orientation on seed selection and gennplasm improvement. Finally, an agreement was made to follow a pure-line selection process in which fanners' participation in Ihe process was assured. Ihis process was designed to help impart a selection of skills to fanners and improve their cop varieties through pure-line selection if they wished. Ihey would also feel empowered through their own participation in the process. This process rnay be proven to be a holistic, less time-consuming, and more cost-effectíve approach to ímprove the quality of land- races, thus making them competitive with improved varieties and, eventually, invigorating in situ conservation on-farm. The traditional seed-supply system The role of farmers in crop improvement and managing agrobiodiversity can best be explained by the traditional seed-supply system (figure 3). Approximately, 60% of global agrieulture ís per- fonned by subsistence fanners using traditional methods-providing between 15% to 20% ofthe world'g food (Francís 1986; Sthapit and Joshi 1998). Diffusion in most parts ofNepal happens through the infonnal seed-supply system; the contrihution ofthe fonnal seed sector is less than 5% in major staple crops (Baniya et al. 2000). The traditional seed-flow system ineludes variety selec- tion and adoption. seed seleetion, seed exchange, processing, and storage (Shrestha 1998), and al! of these processes are responsible for local crop improvement and creating agrobiodiversity. A review of ease studies from Bángladesh (Mazhar 1997), Indonesia (Winarto 1997), Nepal (Joshi el al. 1997; Sthapit et al. 1998), and Ethiopia (Worede 1992) shows a wide range of examples in dif- ferent eountries where farmers-either independent1y or in collaboration with fonnal or infonnal institutions-have played an important role in erop improvement through seed production and rus- semination (see also figure 1). Variety selection and adaptation From time immemorial, farmers have been observing and selecting their crops and crop varieties, saving and maintaining the seeds for next season, and experimenting with new seeds exchanged with neighbors and relatives (Shrestha 1998). It is noteworthy tba! farmers have med to seleet the best available portion of the harvest for growing the subsequent year and also to meet the require- ments offood and tradition. Fanners introduce new varieties in their localities to suit the different needs of 80il fertility, moisture, family, and society, and to spread labor and reduce risk. Hardon (1995) and Joshi et al. (1997) reported tha! farmers pos ses s the ability and knowledge to selee! crops and species that suit their environment and meet quality and other consumer requirements. 206 .. Threshing "Wínnowing Ｂｃｾ｡ｮｩｮｧ＠ .. Ory]ng • CuUing .. Treatmenl !ll!!J&llm> I1I>o ｾ＠ Kothi .. Gtain .. Gunny bag .. Panicle • Woodoo structore ｾ＠ Cob • Metal structure .. Fruit .. aome .. Pod • Mud structure ;. Clotbes p, Chaudhary. S,P. Khatiwada. and K.D, Joshí .. Purchase • Malerial exchaoge • B