IORe I ~~ ~~ e .. H lo o ... DFID Plallt SCÜ'l lCC.s Re.search Prob'7Y1"lflW $ 1 DDS LI-BIRD (~, IPORI - ,• ISBN 958-694-040-3 An Exchange of Experiences from South and South East Asia Proceedings of the International Symposium on Participatory Plant Breeding and participatory Plant Genetic Resource Enhancement ($) DDS la ,t, Ll-SIRO IPaRI CGIAR Program for Participatory Research and Gender Analysis (PRGA) Apartado Aéreo 6713, Cali, Colombia ISBN 958·694·040·3 Press runo 300 Printed in Colombia September 200 1 An exehange of experiences from South and South East Asia: proceedings of the international symposium on Participatory plant breeding and participatory plant genetic resouree enhancernen!, Pokhara, Nepal, 1·5 May 2000, - Cali, Colombia: Participatory Researeh and Gender Analysís Program, Coordination Offiee; International Center for Tropical Agriculture, 200 I 459p, English descriptors: 1. Plant breeding, 2, Participatory rescarch, 3, Natural resources, 4, Farmers, 5. Eiodiversity. 6. Indigenous knowledge, Descriptores español: 1, Fitomejorarníento. 2, Investigación participativa. 3. Recursos naturaJes. 4. Agricultores, 5, Bíodiversídad. 6, Conocimiento indígena. AGRlS primary material category: F30 Genética vegetal y Fitomejorarníento. AGRlS sccundary material category: A50 Investigaci6n agraria, Classífication LC,: SE 123 ,E9 Copyright © 2001. CGlAR Systemwide Program on Participatory Research and Gender Analysís for Technology Development and Insntutionallnnovation (PROA Program). AH Rights Reserved. Table of Contents Introduction. . . 1 Inaugural Address 3 Mr. Dhruva Joshy OVERVIEW PAPERS Participatory Plant Breedíng: A Framework for Analyzíng Diverse Approaches . • • . . . . • . . 7 L. Sperling, JA. Ashby, ME. Smítn, E. Weltzien, and S McGuire Particípatory Vanetal Selection in Hígb-Potentíal Pruductíon Systems. . . . • . • . . . • • . • 19 JR. Witcombe Enbancing Biodíversity and Productíon through Participatory Plant Breeding: Setting Breeding Goals . . . . • . . • • • • . . . • . . . . . . . . . , . . , , . . . . • 29 Bhuwon Sthapit, Krisnna Joshi, Ram Rana, Afadhusudan Upadhaya, Pablo Eyzaguirre, and Devra Jarvis THE CONTEXT OF PARTICIPATORY PLANT BREEDING Cultivating the Landscape:Enhancing tbe Context ror Plant Improvement . . . . , • • • • . . • 55 Farhad Maznar and Daniel Buckles The Broader Instltutional CODtext oC Participatory Plan! Breeding in the Cbanging Agricultural and Natural Resourees R&D System in Nepal . . . • . . . . . • . . . . • • . . • . • • . . 61 Stephen Bíggs and Devendra Gauchan Partlcipatory Plan' Breeding in Diverse Produdion Environments and Institutional Settings: Experienee oCa Nepalese NGO. • • . . . . . . . . . • • . , , • . . . . . • • , . , • . . 75 A. Subedi. K.D. Joshi. R.B. Rana, and M. Subedi COMMUNITY GRASSROOTS MOVEMENTS FOR LINKING PPB ANO PARTICIPATORY PGR MANAGEMENT Landrace Renaissance in the Mountains: Exporíences of the Beej Bachan Andolan in lhe Garhwal Hlmalayan Region, India • • . • • . . • . . • . . • . • . • • . . . . . . • • . . 87 Vir Singh and Vijay Jardhar! Empowenng Farmers through Participatory Plant Breeding: An Initiative ofthe Green Foundation. . . . • . . . • . . • • , . • • . . • • . . • . . . . . • . . • • . . . . • 97 Vana}a Ramprasad tlnd Shibu M P Rethínking the Partlcipatory Paradigm in Plan! Breeding: A Nonbreeder's Perspective . . • . • . 105 Bishnu Ro} Upreti Adding Benefits lo Local Crop Diverslty as a Sustainable Means of On·Farm Conservation: A Case Study of 3n in Situ Projectfrom Nepal. . . . • . . • • . . . • . . , . . . . . . 117 D-K Rija/. R.B. Rana. MP. Upadhyay. K.D. Joshi, D. Gauchan, A. Subedi, A. Mudwari. s.r Khatíwada, and B.R. Slhopit Participatory Improvement of Rice Crops with Tribal Farmen in India. . • . . . . . . • . . . 127 V Arunachalam / Coments COl\iSERVE's Experience and Work on Participatofy Plan! Breeding in Rice . . . . . . . . . . 133 Gilda T. Ginogaling Enhancing Farmen' Particlpation in Planl Breeding: Communlty Biodiversity Development and Conservalion Program (CBDC), Bohol Project, Phllippines o o • • • . • • . o o . . o . o 141 Hide/isa M de Ramos Developíng Local Organization.1 Capacity ror Participalory Seed Management: Experiences rrom Ihe E.stero Himalayas o o o o • . • • • • . • . . . o o • • • • • • • • . • • . • . 149 Banm Gunmg and Prem Gurung Parlicipatory Approaches to Crop lmprovement al Ihe Community Level in Vietnam o o . . o • . 157 Nguyen Ngoc De STARTING FROM FARMERS' KNOWLEDGE WHEN PLANNING PPBIPARTlCIPATORY PGR PROGRAMS Using Farmer Knowledge for Participalory Sweet-Potato Variety Seleelion in Garnt, West Java, Indonesia • • . . • . • o . . . . . . . o . . . . . . . . o o o o o • o o o • 165 Caeci/ia Afra Widyastuti aad Minantyorini Understanding Agroecological Domains: The Key lo a Sueeessful Participatory Plant Breedlng Programo • . • . . . o • , , , , o • • • . . . . . o o • • • • • • . . • • • . . • . . 171 R.B. Rana, B.R. Sthapit. A. Subedi, D.K. Rija/, and P. Chaudhary Listenlng lo Farmon' Perceptions through Parlicipatory Rice Varietal Selection: A Case Study in ViIlages In Eastern Uttar Pradesh, India • • • • • . . • . o o Thelma R. Paris, Abha Síngh, Joyce Luís wíth Hari Nath Singh, Omkar Nath Singh, . . . 179 Sanjay Singh, Ram Kathin Singh, and Surapong Sarkarung Opportunllies and Constraints fOf Participatory Planl Breeding: Farmen' Seed-Management Strategies and Thelr Effect an Pearl Millet Populations In Rajasthan, India . . , . . . • . . . • 193 Kirslen vom Brocke. Anja Christinck, and Eva Weltzien Slrength of Farmen' Knowledge and Participation in Crop Improvement and Managing Agrobiodiversity On-Facm. • • • • . • , , . . , . . o o o o . . , o • o • o • . . o • • 20 I P. Chaudhary, SP. Khaliwada, and K.D. Joshi Need for Advocacy for Effeetlve Participatory Crop Improvement and Plant Genetic Resouree Enhaneement: Case Sludies on Rice-Breedlng Processe. from Khotang and Jajarkot Districts, Nepal. . , . . . , , • • • , • . . • • • , , • • • • . . o o o • . , , . • o • . . . . 213 Yamuna Ghale FOCUS ON METHODS IN PPB: BREEDING CONCERNS Beyond Taro Leaf Blight: A Particípatory Approach for Plant Breeding and Seleedan for Taro Improvement in Samoa. • • • • , • • • • • . • • • . . . . . . . . . . . . . • . . 219 D. G. Hunter, T. losefa, C.J. Delp. and P. Fanoli Participatory Plant Breeding in Malze for Ihe Chholanagpur Platean of Eastern India. . o . . • • 229 Arun Kumar, D.K. Ganguli, S.e. Prosad, J.s. Gangwar, D. S. Virle, and J. R. Witcombe Participatory Plant Breeding in Rice in Eastero India. • . . • . . . . . . . . . . , , , , • . 233 Ravi Kumar, D.N. Singh, S. C. Prosad, J.S. Gangwar, D.S. Virk, and J.R. Witcombe Participatory Crop Improvement in Maíz. In Gujarat, India. S.N. Gaya!, A. Joshi, and J.R. Witcombe Iv . . , . . . . . 237 Towards a Practical Participatory Planl-Breeding Strategy in Predominantly Self-Pollinated Crops. . . • . • • . . . • • . • . . . • • . . • . . . • • . . . . . . . . .• . . . 243 JR. Wítcombe. M Subedi. and K.D. Joshi Participatory Crop Improvemenl for Intercropped Maíze on Bar! Land Terraees with Trees . . . . 249 T. P. Tiwarí. Daljit S. Virk. and Fergus L. Sinclair LESSONS LEARNEO, EVALUATION, ANO IMPACT OF PPB , . . , , . . • • . . . . . • • • • . 261 Partidpatory Varielal Selecnoo in Finger Millol • D.H. Halaswamy, B.rS. Gowda. A. Seetharam, D.S. Virk, and JR. Witcombe Participatory Varie!al Selection, Food Security, and Varielal Diversity in a Hlgh-PoteDtial Production System In Nopal . . • . . . • . . • • . • • . . . • . . . . • . . . • . . . . 267 K.D. Joshi and J.R. Witcombe A Holisne Approach lO Parncipalory Crop lmprovernent In Wheat. . • . • • . . • . • • • . . 275 D.S Virk. D. Harris, B.S Roghuwanshi, A. G.D. Raj, P.S. Sodhi, and JR. Witcombe Participalory Varietal Sol.etion in Rice in the Punjab. . . . . . . • • . . • . • • . • . . . . 283 SS Mn/hí. JR. Witcombe, D.S. Virk, and K.B. Singh • Equlty lsoues in VarielaI Dissemination tbrougb Farmers' Fairs (Klsan Melas) in Punjab, India. . . 291 J' Singh, SS Malhi, JR. Witcombe. and D.S. Virk Particlpalory Varielal Seleellon in Rabi Sorgbum in India . • • . . . . . . . . • . . • . • , . 297 D.S Rana, SL Kaul, Chari Appaji, Parbhakar, NS Kalyanshetly. B. V,S Reddy, JR. Witcombe, and D.S. Virk Tbe Impact of Parncipatory Plant Breeding (PPB) on Landrace Diversity: A Case Sludy for Higb-Altitude Rice in Nepal . • • • . • • . . • • . . , • . . • • • • , • . • . . . • • . 303 K.D. Joshi, B.R. Sthapit, and JR. Witcombe FOCUS ON METHOOS IN PPB: SOCIAL SCIENCE TOOLS Role ofFarrners in Sctting Breeding Goals . . • . • M Subedi, P.K. Shrestha, S Sunwar, and A. Subedi • . • • . . • • • . . • . . • • • . 311 Sensory Evaluation ofUpiand Rice Varieties witb Farmers: A Ca•• Study . • • . . . . . • • • . 319 R.K. Singh, K. Prasad, NP. Mandal, R.K. Singh, D. Coartois, v'P. Singh, and T. París Incorporation ofUsers' and Gender Perspectives in Farmer-Led Participatory Plant Breeding Malze:Experiences from Ibe Western HilIs of Nepal . • • . . , • . • , • . . . . . . . , 329 00 Pratap K. Shrestha, Madu Subedi, Diwakar Poudel, and Sharmila Sunwar Underslanding Farmers' Seleelion Criteria for Rice Varietles: A Case in Madhya Pradesh, Eastern India • • • • • • . . • • • . • . . • . . . • . . . • . • . . . . . . R.K Sañu, V,N Sahu, M-L. Sharma, T. Paris, K. McAl/isler, R.K. Singh. and S Sarkarung . . . . 343 DEVELOPING NEW PPBIPGR PROGRAMS Sced Sccurity in Badakshan, Afghanistall. . . . • . . . • . • , • . . . , . . . . . . . . . 359 Iqbal Kermali v Invohing Farmers io the Dovelopmenl Process lo Improve Adoption ofVarieties Developed by National Maize-Breedlog Programs. . • • . • • • • . . . . • . . . . . • • • . • . . . 367 J.K. Ransom, K.B. Koirala, N. Rajbhandari, and K. Adhikari Participalory Plan! Breeding and Property Righ!s • . . . . . . • . . . . , . . • . 373 Projeet 01 SWP PRGA Increaslog Ihe Relevanee ofBreeding lO Small Farmers: Farmer Pllrlldpation and Local Knowledge In Breeding Bnrley for Specific Adaptation lo Dry Areas of Jardan. . . . . . • . . 381 S Ceccarellí, o. Kalawin, S, H Saoub, S. Granda. H Halíla, M. Ababneh, Y. Shakatreh. and E. Bailey Presenl Status of Partlcipatory Plant-Breedlng Research 00 Wheat at Ibe National Wheal Research Program of Nepal . . . . • • . . . . • • • . . . . • • . . . . . . . M.R. Bharta, G.O. Ferrara, B. Gurung. r.? Pokhrel, NR. Gautam. P. Gurung, and R.B. Neupane • . 391 FARMERS SPEAK FOR THEMSELVES Conserving Agricultural Blodiversity . . . . . . . . . . • . • . • • . • • • . . . . . . . 401 Sundaram Verma Expedence, Researeh, and Fads Related lo Local Species of Paddy. • • . . . • . • • . . . . . 403 Chape! Div Bhagat Malí Assísted by Pashupati Chaudhari and Parmananda Chaudhari Problems of M.lze Cultivation .nd the Role and Approach of Local Farmefs in Solving Ibis Problem. . • . • . . . . . . . • . • • • • . . . • . • . . . . . • • . . . . . . . 407 Sr; Har; Prasad Aryal Tbere Is the Possibility of Simichaur Becoming Makai ehaur: My Experienees in Crossing M.ize Species • • • • . • . . • . • . . • • . • • . • . • . . • • . . . . . . . • . . • 409 Mr•. Lal K"marí Basnet How Did the Fumers of Chhomrong lmprove Ibe Local Paddy Species? . . . . . . . . . . . . 411 Mr. Om Baltadur Gurung, Mr. Najarman Gurung, Mrs. Mi» Kumari Gurung, Mrs. Naulí Gurung Local Species: Method. uf CultívatioD, Some Successe. aud Some Problems .• 413 Ram Ashraya Saha Kalewar Role ofF.rmers in Selecting Crop Speeies . • • . . . . . . • . . . . . • • . . . • • . . . 417 .Jhapad Bahadur Bhandari Role of Farmers in Ihe Improvement ofCrops: New Species of Paddy in Maramche. • • . . . • • 419 Mr. Chandra Kanta Poudel Assisted by Mr. Di! Bahadur K.e. and Mrs. Sita Poudel The Importanee of Crup Improvement in Conservatlon of Diversity K. e. Adhikary, S.F. Adhikary, and K-R. Adhikary . . . . • . . • . . . • • • 425 Question-and-Answer Sesslon with Nepall Farmer•. 429 Diversity Versus Mono-Cropping 431 Bidakanne Sammamma Experiences Growlng a Modero Rice Cultivar. . . . . . . . . . . . . . • . • • . . . . . . 433 Raksya Begam v, Is There an lmminenl Crisis in Agricullur.? . . . . . . . • . • • • . . • • . . . . • • . . 435 Abu Taher Rahamani Acronyms. 437 ParticipanlS 439 Workshop Program: 445 vii Introduction In the past lOto 15 years participatory plant breeding (PPB) has captured lhe interest and curiosity of a growing number of development practitioners and agricultural scientists. Its main appeal has been lts promise to improve lhe chronically low adoption Tates of new crop varietíes developed by lhe research community to alleviate food scarcity, poverty, and natural-resource decline in developing countries. By bringing lhe formal breeding process closer to farmers and combining disciplines such as plant genetics, pathology, entomology, anthropology, economics, marketing, etc., with tradítional farmer knowledge, PPB has facílitated the crcation of more "relevant" varicties. However, olher equally significant goal s can drive lhe development ofPPB programs, for instance, lhe desire to empower farming communities, which gain greater control of breedíng and seed supplies, to enhance biodiversity and germplasm conservatíon, and to develop adapted germplasm for especiaIly dísadvantaged user groups (e.g., women, poor farmers). From May 1-5, 2000, over 100 scíentists, farmers, development practítíoners, and cornmunity organizers trom 7 Soulh and Soulheast Asian countries (with Jordan and Samoa added in for good measure) met in Pokhara, Nepal, lo discuss PPB methodologies, strategíes, actors, environments, and impacls. The objectíves of lhís symposium were to cxchange and compare lhe diversíty of experiences with PPB in the regíon, to identifY gaps in the research to data, and to network South and Southeast Asian ínstítutíonal plan! breeders, farmers, and others who are using particípatory methodologies. The symposium was al 50 an opportunity to further díscuss 1lIld contribute to lhe working document Guidelines for PPB ... drafted in a 1997 expert consultation and refined in subsequent meetings and electronic discussions wilhin the Plan! Breeding Working Group of the CGIAR Systemwide Program on Participatory Research and Gender Analysis (PRGA). The symposíum had two parallel programs--one for scientists and ¡he other for farmers-which carne together al different poínts in the four days through formal presentations, working-group discussions, and ímpromptu evening sessíons. The aims were to encourage interactions wílhin groups and well as between them. Farmers and plant breeders contributed from lhe wes! and central regíons ofNepal and three dispersed regions ofIndia (Ultar Pradesh, Hyderabad, and Rajaslhan) wilh the result Iha! díscussions were variously conducted in Englísh, Híndi, Nepali, and Bengalí. Exchanges were generally intense, informalive, illuminaling, and--occasíonally--conflictive, yet managed lo cross lhe linguíslíc, philosophical, and methodological divídes effectively. These proceedings are bul one product-and a critical one--ofthe four-day semínar. Several field prograrns developed directly from new collaborative relationshíps established al the meeting, as did an íncreased underslanding of the mutual contributions that instítutional plant breeders and farmer plan! breeders can make to sustaíning and enhancing the farmíng communities' contribution to agrículture. The strength of the workshop rested largely on lhe range of co-hosts and in the dyoamíc work of their partners. Special acknowledgements go lo the Nepal Agricultural Research Couneíl, which launched the workshop, and to the diverse regional networks, which carne together for four days to compare, contra~t, and debate appropriate goals, approaches, and methods to use for effective participatory plant breedíng. Co-hosts The System-wide Prograrn on Participatory Research and Gender Analysis (PRGA) The International Development Research Center (IDRC) The Department for Intemational Development (DFID) Using Diversity Network (lJD) Soutb Asia Network for Food, Ecology and Culture (SANFEC) Deccan Devclopment Socicty (DDS) Local Initiatives for Biodiversity Research and Development (LI-BIRD) The Eastem Himalayan Network The International Plant Genetic Resources Institute (IPGRI) 2 Inaugural Address Mr. Dhruva Joshy Participants and Ladies and Gentlemen, 11 is, indeed, a great pleasure for me to be associated with the inauguration ofthe lnternational Symposium on Participatory Plant Breedingfor South and Southeast Asia being held in NepaL 1 wish lo express my sincere gratitude to fue organizing committee ofthis symposium for ¡he invitalion extended to me to delíver an inaugural address in this opening ceremony, which, to my rnind, i5 of special importance, since Nepal i5 ¡he birthplace ofparticipatory plant breeding (PPB) as a methodology used in plant breeding. First of aH, 1 would like lo thank Ihe organizer for choosing beautiful Pokhara valley as the venue for this symposium. 1 would like to welcome you to Pokhara, fue valley of eight lalces, fue custodians of wild rice and 70 rice landraces, and Ihe place known for ils natural beauty with fue majestic Annapuma Himalayan rangc in the background and for ils great ethnic and cultural díversity. Annapuma Conservatíon Area and the study site ofin situ conservatíon ofagrobiodíversíty are also in thís valley. In the present paradigm of sustainable agricultural development initiatives, conservation of agricultural biodiversity'is the cornerstone of sustainable production and of locar and national food security. Seed is fue first link in a sustainable food chain of the human kind. In Nepal, still abou! 800/0-90% of farmers' seed demand is mel by the informal seed seclor for the majority of erops. In this system, farmers produce fue ir own seeds on their own farms, or obtain seed via exchange or purchase from other farmers, reJatives, or local traders. Particípatory plant breeding has its primary attraction in this system as il has tremendous polential to address fue needs of farmera, particularly in fue developing countríes offue region. 11 takes us closer lo marginal areas; it helps us to hamess the potential of many minor and neglected erops; and most important, il addresses Ihe livelihood needs ofpoor people and helps to alleviate poverty. Evidence shows thal tbe conventional plant breeding ofthe Green Revolulion has yiclded good results in the more favorable agricultural syslems. Most low-resource farmers in marginal areas, however, have no! benefitted from fuese modern cultivars as expected. As an alternative for these areas, participatory approaches to crop improvement and selection have been initiated with good results. It is quite appreciable that the COIAR system has recognized this gap and institutionalized the PRGA program to assess and devclop methodologies on participatory plant breeding. The Nepal Agricultural Research Council (NARC) is aware and very supportive of such initíatÍves on participatory approaches to crop improvement. 1 am proud lo mentÍon Ihat we are the first nalional agricultura! research system to release a producI ofPPB, Ihe Machhapuchre-3, a rice variety for rainfed lowland areas of mid-altítude (1300-2000 m) in 1996. Tills variety was bred by fue breeders ofthe Nepal Agricultural Research Council and later selected and lested by farmers from Chhomrong and Ghandruk villages under the guidance of scíentists from Lumle Agricultura! Research Centre. Local Initiatives for Biodiversíty, Research and Development (LI-BIRD) ís playing an importan! role in scaling up the products and approaches ofthese PPB inítiatives to wider areas.1 Me Dhruva lQshy is executive director of the Nepai Agricultural Research Council (NARC) - .... _ . - .... _ .... ~_ .... _-~---- 3 Inaugural Address am pleased to hear that sorne ofthe farmers ofthese cornmunities are participating in this workshop to share their experiences. There is now substantial evidence that farmers maintain and improve their landraces through a continuous process of selection. 1 am sure we can hear many such exampies trom Bangladesh, India, N epal, and other countries in this symposium. A major challenge in modern plant breeding is to devise appropriate ways to address the problems of resource-poor farmers in marginal environments who have ofien contributed important genetic diversity to the formal system with little benefit in return. In countries like Nepal where ecosystems, farming systems, and user preferences are so diverse and complex, participatory plant breeding can consolidate the impact of institutional plant breeding. This is not an expensive approach, but our plant breeders need a flexible mind-set to try new methods and approaches. Ifthis happens, it may provide an opportunity, no less important than that offered by biotechnology, to solve the food-security problem of the country to a certain extent. F or sustainable agriculture, increasingly uniform crops may be more vulnerable to pests and di seases. We have had recent examples ofBPH darnage in the terai region ofNepal. 1 heard that many participatory approaches to crop improvement have been initiated in both high-potential and marginal areas of this region, including Nepal, and sorne interesting results are emerging. 1 believed that participatory varietal selection and PPB can deploy new diversity to combat new pests and di seases and that participatory research willlead to rapid extension ofthe technologies. New genetic materials reach farmers' hands quickIy. Ifthe material is good, farmers take care ofit, as they have maintained many landraces such as Másuli rice over the years, and, hence, it will remain in the 'informal system. Distinguished Guests, Ladies, and Gentlemen, Nepal Agricultural Research Council has assigned a high priority to conservation and sustainable utilization of agrobiodiversity. Both, ex situ and in situ conservation strategies are being adopted in the national research system. The gene bank maintains nearly 10,000 accessions of 60 crop species at low temperature and low humidity. Nepal is the probable homeland ofthe Asian cultivated rice Oryza sativa. Wild species, namely, Qrjza nivara, o. rujipogon, o. granulata. and o. ojJicinalis, are distributed in different parts of the country. o. rujipogon is available in tbe various lakes of Pokhara. To allow the evolutionary process to proceed, in situ conservation provides a unique opportunity in the natural habitat. May 1 take tbis opportunity to mention a few words about enhancing partnerships in agricultural research between nongovernmental organizations (NGOs) and NARC. NARC has now started to initiate partnerships in research with NGOs and the private sector and is keen to support capable NGOs in participatory approaches. 1 am also pleased to inform tbis August gathering that NARC, LI-BIRD, and the International Plant Genetic Resources Institute (IPGRI) have jointly developed a PPB program in three in situ sites in Nepal with the objective of assessing PPB as a strategy for on-farm conservation and productive gains. Needless to say, the sharing of our field experiences trom across the globe will bring further refinement of these methodologies in the future. Wbile speaking of subsistence farmers and biodiversity relationsbips, these custodians of genetic diversity are still maintaining crop diversity and varietal diversity within species, in spite of the threats of erosiono This has been possible because ofthe result of cultural practices and local knowledge systems built through the input of millions of unknown and invisible farmers. Indigenous knowledge has been overlooked in the past but is increasingly being acknowledged as a cornerstone 4 Mr. Dhnlva Joshy for sustaínable development. The value of indígenous knowledge systems has become ever more important in our diverse mountain system where people live in great ecological, bíological, and cultural diversity. It would not be out of place to mention the liberalization poliey the country is foIlowing. This has prompted NARC to look Ínto the issues of trade-related intellectual property rights (TRlPR) under the General Agreement on Tariffs and Trade (GATT). The World T rade Organization (WTO), under the TRlPR agreement, requires member countries to provide legal protection to plan! varieties. As in many developing countries, protection of inteIleetual property rights (lPR) has not been taken seriously in the proccss oftechnology generation in Nepal. I believe the symposium will provide us some suggestions for dealing with the changing scenario of globalization with respect to IPR. What will be the impact ofIPR on biodiversity? How will IPR handle the issues ínvolving the varieties developed by PPB? What wiIl be the consequence ofhigh-yield adaptive terminator technology on agrobiodíversity? These are some ofthe issues that need answers in response to the globalization of agrículture. Today various forces are at work-popuJation pressures, a desire for a better quality oflife, urbanization, and the availability of modern technologies-which have started a chain reaction in the transfonnation of mountain agriculture. It is widespread. The lmfortunale oulcome of thís process, observed during the past few decades, is a negative impact on native agricultural biodiversity. The~efore, ways have to be fmmd to contain and reverse Ihis trend. We need 10 be realistic, as resources are limited. We need to understand what needs to be preserved and do the best we can. Ways mus! be found to motívate farrners to maintain farrn bíodiversíty; 1 beIieve PPB is one approach we can ¡ook forward lo. While conventional institulional efforts for the conservation of agricultural biodiversity must continue, it is equally important to find innovative ways of maintaining in sítu crop conservation. Local initiatives, people's participalion, and combining conservation wíth use are some ofthe important concepts for developing appropriate approaches that can combine agricultural biodiversity with sustainable agriculture development. PPB appears to carry lots of promises in Ihis regard but we slill need to develop our critical mass of researchers. 1 am sure Ihis is exactly what this meeting is also going to discuss. Over the next five days, 1 very much hope that each of you will be able to benefit from the sharing of combined knowledge and experiences in this ficld. This meeting is especially unique as 1 was told tha! farrners from various countries have also assembled here to share their wisdom and insights. The most interesting change is that plan! breeders and researchers are here to learn from Ihem. NARC will foIlow the outcomes of Ihe symposium with much interes!. Finally, 1 would like to thank once again the PRGA and other cosponsors for inviting me and NARC scientists lo participate in Ihis meeting. 1 hope the weather will be kind over ¡he next week or so, particularly on the field-trip days. J am Iooking forward to the outcome ofyollT productíve deliberatíons over the next five days and wish you a very fruitful and pIeasant stay in Pokhara and Nepal. Thank you. Dhanyabad. - - ..... _ - - - - - - - - - - - - - - - - - ..... _ .. 5 Participatory Plant Breeding: A Framework for Analyzing Diverse Approaches L. Sperling, JA, Ashby, ME Smith, E Weltzien, and S. McGuire Abstraet Participalory planl breeding (PPB) is a relalively new approaeh to germplasm development. Overvíew summaries of cases to date show thal most PPB programs were begun within the last 10 years, whether located in public-scctor or nongovemment (NGO) crop-improvement programs. Sorne have argued that coromercial, prívate-sector, plant breeding has long been c1ient-driven, or "participatory." However, when PPB is used to reaeh poor clíent groups, to breed for high-stress or heterogeneous conditions, and lo ineorporale diverse, specíalized elient preferenees, the result is a fundamental change in tbe way planl genetic resoureeS are managed by formal breeding programs and farmers. Tms paper oullines a framework for relaling different participatory plant-breeding approaches lo differenl outeomes .nd impacts. Based on a detailed analysis of 65 case studies of programs and projecls involving PPB, it suggests some ofthe wide variabilíty ofPPB programs and lays out key varíables that are crucial for díserintinating among PPB approaches. These inelude tbe inslitutional context, Ibe bio-social envíronmenl, Ihe kínd ofparticipation achieved, .ad lhe goal s sel forthe PPB work. Jt ís only when Ihese variables are clearly descríbed thal practitioners can start to link the type ofPPB employed (melhod .nd orgartizational forms) with the lype of impact achíeved. Suoh c1arity ís cssentíal ifPPB is also to have the scienlific and orgartizational foundations necessary lo judge its utility for a given abjeelive. Introduction Participatory plant breeding (PPB) is a relatively new approach to germplasm development. Overview summarics of cases to date (Weltzien/Smilh et aL 2000; McGuíre et aL 1999) show that most PPB programs were begun within the last 10 years, whether located in publíc-sector or nongovernrnent (NGO) crop-ímprovement programs,. Although sorne have argued that commercial, prívate-sector, plant breeding has long been client-driven, or "participatory" (Dr. Don Duvick, personal communication), PPB, when used to reach poor client groups, to breed for high-stress or heterogeneous conditions, and to incorporate diverse, specialized elient preferences results in a fundamental change in the way plant genetic resources are managed by formal breedíng programs and farmers. Ihis artiele aíms to set up a framework for relating different participatory plant-breedíng approaches to outcomes and impacts. Based on a detailed analysis of65 case studies ofprograms and projects involving PPB (Weltzien/Smith et al. 2000; McGuire, Marucad, and Sperlíng 1999; Hecht 2000), it suggests sorne of lhe wide varíability ofPPB programs and lays out the key variables that are crucial for discriminating among PPB approaches. These include thc institutional context, the bio-social environment, the kind ofparticipation achieved and the goals se! for the PPB work. lt is only when these varíables are clearly described that practitioners can start to link the type of PPB employed (method and organizational forros) with the type of impact achieved. Such clarity is L. SperHng and J.A. Ashby work with the Systemwlde Program on Particlpatory Research and Gender Analysis for T eclmology De~ me velopment and Institutional Innovation (PRGA), based atCIAT. M.E. Smith is at Departrnent ofPlant Breeding, Carnel! (Jniver~ E. Weltzien works with ICRISAT's Genetk Resources and Enhan<::ement Program, based in Bamako, Mall, and S McGuire 15 with fue Technology and Agrarian Developmem Group, Wagenlngen University in the Netherlands. sÍtj' in New York. We thank B. Sthapit for providing useful insights 00 this frame1.\'ork. 7 Participatory Plant Breeding: A Framework for Analyzing Diverse Approaches essential if PPB is also to have the scientific and organizational foundations necessary to judge its utility for a given objective. The institutional context One of the most important differences among approaches to participatory plant breeding is institutional, in lhe sense (following North 1990) ofthe mies for behavior, the norms and values, and the incentives that govern how farmers and scientists share the responsibilities, work, and benefits of a joint plant-breeding effort. The key institutional difference lies in the obligations that determine the locus of control or decision making about the objectives ofthe plant breeding and lhe kind ofresults and data required to support these objectives. We distinguish two main institutional approaches: one when farmers join in breeding experiments that have been initiated by formal breeding programs, which we term formal-Ied PPB, and the other when scientists seek to support farmers' own systems ofbreeding, varietal selection, and seed maintenance, which we callfarmer-IedPPB. The incentive structure and the rights and obligations that characterize these two approaches can be expressed in different types of organizational arrangements. Formal-led PPB Formal-Ied PPB has certain unique institutional characteristics. Researchers mn formal-Ied PPB programs and invite farmer participation in formal research. Researchers have an obligation and ofien a priority objective to feed information baek to the formal research sector, which means lhat the seientific standards of replicability and validity of results must be me!. PPB is expeeted to complement the formal-sector researeh system, e.g., either refining breeding strategies so that speeifie environments and varietal preferences are addressed orreorienting priorities. Generally, formal-Ied PPB programs also involve strong linkages to formal systems for variety release and seed production. Finally, scientists involved in formal-Ied programs are usually expected by lhe scientific community to extrapolate their methods, if not lhe varieties per se, beyond the individual cornmunity with which they work. They ofien need to show what the advantages ofPPB are, eompared to formal breeding approaches (WeltzienlSmith et al. 2000). Farmer-led PPB Researchers or other professionals involved in farmer-Ied programs are expected to facilitate a process in which farmers establish the breeding objectives. Farmers bear lhe main responsibility for and, ofien, lhe costs of conducting experiments and selecting materials for seed multiplication and dissemination. Researchers are expected to take a support role in this process. The objective of farmer-Ied PPB is to develop varieties orpopulations lhat suit specific local environments and local preferences; any broader applicability beyond local circumstances is fortuitous. Farmer-Ied PPB, with a few exceptions, tends to work for a specific client group or groups that have no obligation either to feed information back for wider geographical extrapolation or to feed products such as varieties into external formal systems (McGuire, Manicad, and Sperling 1999). It is important not to confuse lhe scale of a PPB effort (i.e., the size ofthe program or the extent of its geographical eoverage) with the institutional approach. The faet that PPB is carried out at the village or locallevel does not mean that it is, ipso facto, farmer-Ied. Case-study analysis indicates that there is a very wide range of collaborative arrangements in PPB carried out at lhe local or village level (PRGA 1999), sorne ofwhich can be described as using a farmer-Ied institutional approach; others are instead controlled by representatives of outside agencies, albeit small-scale ones like 8 L. Sperlíng el al. local NGOs. Since most PPB is still experimental and most initiatives in their early stages are conducted in a few sites, it is not yet clear whether there is an inherent difference in potential seale between the formal-led and fanner-led approaches. Bio-social environment of PPB Two types of parameters have proved heuristic for characterizing the environments in which PPB programs take place. 1 The first of these is the type of agroecological environment in which PPB programs develop. This has been plotted along a crop-specífic scale, ranging from high stress to low stress, based on actual versus expeeted yields, coupled with an index for incidcnce of crop faílure (thus combining yield leve! and stability) (figure 1) (WeltzienJSmith et al. 2000), Agroecological envíronments potentially range from those that are primarily subsistence-oriented and UnfaVOfable ..................... " ..... ,..... _..... _... _.............................. ", .... ,...... ,.,_,_ ",.Favorable • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • Figure 1. The distribution of participatory plant breeding cases by type of environmnet (For case identification, see Weltzien/Smith et al. 1999; for an extensive description of each case, see McGuire et al. 1999 and Welízien/Smith et al. 1999.) 1, This characterization has been done in collaboration with the Plant Breeding Working Group of the PRGA. This group em~ braces about 150 plan! breeders, social sdentists, development personnel, grassroúts activists, artd geneticists frorn a wide range ofpubJic- and private-sector insütut¡ons. both North and South, Thecommon link between members is a methodological interest ¡n PPB. ~~~~~ ................. ~ .... -~~~~~~ 9 Participato!)' Plant Breedíng: A Framework fOr Analvzíng Diverse Approaches highly unstable, implying that farmers' crop choices are govemed by their own adaptive and preference needs, to systems in which crop production is very controlIed and largely driven by the needs and preferences of urban consumers andJor commercial processors. The second parameter suggests the broad economic environmenl of PPB; thal is, the degree of "homogeneous demand versus heterogeneous demand" for varieties. Plotting was based on a nominal Reale of l lo 10, aeeording lo the "leniency/narrowness ol'varietal characteristics demanded by end-users" and Ihe similarity/discordance between varieties used for home consumption and for sale (WeltzienlSmilh et al. 2000). At the higher end (for example, 8, 9), the environmenl lends lo correspond to a high degree ol' homogeneity in product, ofien favoring a narrow range ol' graín, laste, and cooking types. Such a high degree ofuniforrnitylhomogeneity is ofien associated with an economic environmcnt where farrners are producing for highly specialízed markets. Because conventional breediug has been less efTective in difficult environments and in reaching farrners with few resources, sorne plant breeders consider PPB as most appropriate for high-stressl marginal environments, where agriculture is low-input. This would confirrn the rationale fortesting particípatory approaches that are ofien site-specific. Analysís of actual PPB cases, however, shows a more complex picture (figure 1). Not aH PPB is concentrated in hígh-stress environments with low-input agriculture. An unexpectedly large number ofPPB programs are being initiated in intermedíate arcas where agroclimatic stress is les s severe. On ¡he whole, these are cases where qualíty eoncerns, such as meeting end-user preferences, are the paramount challenge (WeltzíenlSmith et ato 2000), e.g., see cases ofPROINPA work in Bolivia and CIAT/CIALS in Colombia). PPB programs are also becoming more common in the favorable or so-called "Green-Revolution" arcas. Here, PPB approaches are being explored lo help inerease varíelal divcrsity and lo enable farrning communities to have greater control over their breeding and seed supply (McGuire et al. 1999). Participation and PPB Participation is a terrn with a number of different connotations, and it is essential to be cIcar about what the separate dimensions of participation are, which logether define what we call its qua/ity. With respect lo the "qualíty of participation" in PPB, it is useful to identifY tbree diffcrent dimenSlOns: • stage of participation • degree of participation • actors' roles in participation When researchers describe "participation" in PPB programs, Ihey are generalIy reterring lo the stage ofthe breeding cycle at which farrners are involved. It ís usually faír lo say lhat the earlíer user participation occurs in the breedíng process, the more opportuníty users are given to influence the objectíves, breeding strategy, and final outcome. Bu! the extent lO which users can realize this opportunity depends on the degree of participation. The third dimension of participation is the specific role taken by rcsearchers, farnlers, or others. Role refers lo the function perforrned: for example, management or providing inforrnation or field labor. 10 L. Sperling el al. 3. Information-giving role-Providing informatíon on varietal preferences, plant types, or desired traits to be maintained or introduced: Fanners can offer kcy insights into fue trade-offs fuey are wíl1ing lo make among characteristics in designing fue desired plant ideotype. Farmers often have strong preferences Ihat greatly shape adoption and which need to be integrated into potential varietal entries. 4. Trainerlskill-builder role-While thls role is often associated wifu researcher input (and can be key for empowering fanners to continue generating breeding materials themselves), fanners can also playa central role in skill building through farmer-to-farmer training and farmer-to-researcher training. 5. Field laborer role-Providing labor: Farmer labor may be needed when formal research canuot select with available resources. In all cases, farmers often do fue routine land preparation, weeding, etc. 6. Input supply role-Providing land for "realistic" bio-physical sites: Formal breeders sometimes have greater success by selecting directly in target environments. To do thls, they may use actual farmers' fields in the same way they use more standard experimental stations-as researcher-desígned and -managed testing sites. 7. Providing landrace or farmer material used for further breeding work: There is also a key role played by fanners in providing germplasm to the breeding process. While formal breeding approaches have used this farmer resource extensively, it has ofien been done without involving fanners specifically in the process of choosing germplasm, or in the subsequent processes of evaluation and selection. In sorne PPB cases, farmers have explicitly generated new base material for a shared breeding program by making or facilitating crosses between chosen parents. Whether they are directly involved or whether farmer germplasm is used with direct farmer advice, the ouocome of the programs should recognize fanners' contributions when attributing any property rights to the finished materials. Roles 5, 6, and 7, in isolation or as farmer-only roles in a program, do not make a program partici- patory. There probably isn't a breeding program in the world, or at least, in the developing world, lhat does not use skilled farmers as laborers. There is also a good deal of on-fann testing, where farmers provide land and other resources. For a program lo be participatory, it has to be linked to sorne degree ofreal decision making (i.e., roles 1 through 4). Goals ofPPB Overfue las! decade, PPB has been applied as a crop-improvement strategy primarily in response to the need fOI impact in noncommercial crops and in very unpredictable, stressed production environments. However, a range of other goals has also been defllled within PPB programa: for instance, enhancing biodiversity and germplasm conservation, deve10ping adapted germplasm for especially disadvantaged user groups (e.g., women, poor farmers), and making breeding programs more cost-efficient, particularly through decentralization of prograrns that target more niches. TabJe llists the broad goals around which PPB programs have been designed and sorne indícators that can be used to track whefuer these goaIs are being meL Close analysis of the set of PPB cases shows that some goals are explicit and often atrained (for instance, produetion inerease), while olhera are poorly artieulated and usually not addressed unleas 13 :':1 Table 1. Potential PPB Program Goals and Posslble Indicators for Montioring Progress towards Them PPB Goal. Posslble Indicalot'$ ------ Productlon galns (in eludes quality ¡ncrements, higher value products) · · ·· · ylekl ¡ncreases, stability I ·· • Effect!ve targeting of user needs ------ Cost~efficiencies Capacity building and knowtedge generation fO!' farming commun¡~ ties and the formal research and de~ velopmen! (R&D) sectors ••• _ _ _ _ _ · · · beneflts gained 1hrough higher mame! value of product (¡ncome generated) better ldentificatjon of farmer~preferred quality traits, such as taste, etc. communities get wider access to germptasm communifies get wider access to information/related knowledge more !ntravarietal diversity more !ntervarletal diversity compatibility of new materials with existing únes (Iess varietal replacement: more compatibllity with landracss) targeting o~"more micro-niches _ . ._ . . . .- · · · · · · · · ------ An objective moy be to monage '. pool' 01 dlversity versus 'a variety' Efforts might be aime<) e .. :1 CI' !!! 8 6 Al! trial mean u.. 4 2 Mean yield (kg ha·1 J Figure 2. Comparíson ofyields in tbe AlI-lndia Coordinated Sorgbum Improvement Project trials and in the dlstricts in which these tríals were conducted, 1989 (Packwood et al. 1998) testing the least important entries--those in the first year of the trials-than the more important entries undergoing the second or third year oftesting (Witcombe et al. 1998c). • The trials did not allow selection of specifically adapted varieties . For example, earliness i5 extremely important ¡n marginal areas because it allows the escape ofend-of-season drought. (Earliness ls prized by farmers in HPPSs as well because it increases the possible optíons in the cropping system and gives more time for the timely sowing ofthe following crop.) However, analysis of many trials showed that in nearly all there was selection against early- and later-maturing entries (Witcombe et al. 1998c). In selecting for wide adaptation, Le., the entries that yield best on average, there is selection formediocrity in flowering time (figure 3). • The selection system lo promote entries from one trial stage to the next did not allow a trade-off between different traits. The promotion criteria are heavily biased towards grain yield, and little or no consideration is given to other traits, such as early maturity, stover yield, and grain quality. OnIy if an entry survives tbree years in the trial can other traits be taken into account when it is considered for release. Traits other tban yield will have been ignored in the earlíer stages ofpromotion--initial to advanced trial, or promotion to a second year oftestíng in an advanced trial. Hence, in practíce, varietíes with advantages in non-yield traits can only be selected if they have a yield advantage in the first two years of testing (Witcombe et al. 1998c). In sununary, in marginal areas, the fullowing disadvantages ofmultilocational trials were seen: • • • • • • Trial sites poorly represented the crop area. Tria! sites poorly represented farmers' fields. Irials were unreliable. Resources were allocated inefficiently between varieties in different years of testíng. Selection for wide adaptation selected against specific adaptation. The selection criteria used rarely allowed trade-offs between traits. 21 Participatory Varíetal Selection in High-Potential Production Systems .. .. 2.6 2.4 ~ 2.2 . ~ , ~ J: , !::. " ;¡; >= ~óJ@ O 2.0 1.8 1.6 O O o O O o O o O 1.4 1.2 DDclb 1.0 50 60 70 80 90 Time to bloom (days) Figure 3. Time to bloom and yield of entrjes in the very early, direct-sown rice trial of the AHIndia Coordinated Rice Improvement Projeet of 1993 (The four highest yielding cnmes are all oC intermediate flowerjng time; early-flowering entries are eliminated.) Trials for favorable areas sbare tbe disadvantages of trials for marginal ones Although il is not the perceived wisdom, the drawbacks described for trials targeted at marginal agricultural environments are shared with those targeted at high-potential production systems (HPPSs). There are very few trials to represent the often extremely large areas of high-potential production systems. For example, in state-level trials there are only four trial sites for rice in the Indian Punjab to represen! a rice area of abou! 2.2 míllion hectares and only two siles in the AH-India coordinated trials. The Punjab does not represent a single target environment; there are marked differences in adoption of varieties by farmers from district to district; however, not all of the districts are represented in the formal trial system. High-potential production systems are nol uniform (Wítcombe 1999) but have great physical and socioeconomic diversity. Physical variation is often related to the cost and availability of irrigation water that can be supplied predominantly by tube well in some arcas and by canal in others. Variation in Boil and land type is significant. For example, in rice there are niches, such as more waterlogged areas, where long-duration rice is required (figure 4). In contrast, in some areas short-duration varieties are needed either because of physical variation (limited water) or temporal variation (a need to harvest Ihe crop early for timely sowing ofthe following crop). Unlike marginal areas, the disparity in the level ofinputs on the research station trials and farmers' fields is indeed much less and this is not a major reason why trials poorly represent farmers' fields in HPPSs. However, unlike marginal areas where the plantíng date used by both researchers and farmers is dictated by significant rainfall events, there can be a large disparity between the sowing dates offarmers and the sowíng dates of research station trials. Coordinated research trials require a 22 J Witcombe 1997 1999 Swarna Radha 4 Sabitri Masuli Figure 4. The adoption of Swarna, a late-maturing variety for wetter areas, in a village in Cbitwan, Nepal, after two seasons (data from K.D. Josbi, LI-BIRD) great deal of organization to assemble and redistribute lhe seed to lhe trial sites. Hence, it is common in a crop such as rice, where lhe seed is produced in lhe off-season just before the main season trials, for lhe trials to be sown later ralher lhan earlier in lhe season. Apart from lhe practical difficulty of representing earlier sowing dates, lhe low number of trial sites means lhat lhe range of planting dates used by farmers cannot be represented. For example, both the sowing and transplanting ofrice extend oyer a three-week period in Lunawada District, Gujarat (figure 5). It is a practical impossibility to haye all lhese sowing dates in a formal trial system, yet significant interactions between sowing date and yariety occur. Trials in HPPSs, allhough more reliable lhan those in marginal areas because of lhe existence of irrigation and more uniform land, can still suffer from high experimental error because of small plot sizes and limited replication. -- 100 ...111 80 / Q) ... E tcI 60 I ¡ /SOWin 9 s:::: o 1:: o C. ...o ,. , ,- /' o~ 40 20 Q. O / V 12 /, , 22 June 29 I I - ./ " ,- (" -- , Transplanting .1 "",,;; 9 , , , 18 26 July Figure 5. Sowing and transplanting dates of rice in Lunawada District, Gujarat (Virk et al., tbis volume) 23 Participatory Varíeta/ Se/eclion in High-Potential Pro The deficiencies in resource allocation, described for tríals in marginal areas, are caused by the promotion criteria used. These eriteria are used índependently of the targeted production system, so resource allocatíon is just as poor in trials for HPPSs as in those for marginal areas. In trials for HPPSs, the trade-offbetween multiple traits is no better than in trials targeted at marginal arcas. The value of shorter-duration crops is insufficiently reeognized in the trial system for HPPSs where seleetion is almost entirely for yield and division of the trials by maturity c1ass is lacking or inadequate. Early maturity can allow another erop to be grown during ayear, either a cash crop or a green-manure erop, and it can spread demands for labor at sowing, transplanting or harvest time. Trade-offs between yield and other important traits (e.g., fodder yield or grain qualíty) also reeeive msufficient attention. What are the roles of participation and decentralization in PVS? The deficiencies identified in the multilocational trial system can be removed by radically modifYing the design of the multilocational trials without significantly increasing farmer participation. Altematively, the problems can be addressed by idtroducing a major component ofparticipatory varietal testing (Witcombe and Virk, forthcoming). This rruses the questíon as to whethermodifications to the design of tbe trial system, a11 of wruch result m decentralization, are simpler and cheaper than employing participatory approaches. The six problem areas identified in tbe multilocational testing are examined to see if redesigning the trials by decentralization or mcreased farmer participation is tbe most efficient solution. Both decentralízation and participation help to solve these problems because they can do the following: 1. 2. 3. 4. 5. 6. allow trial sites to better represent the crop area allow betterrepresentation oftbe environments m farmers' fields increase the reliability oftbe trials allocate resourees more efficíently between varieties in different years of testíng allow varieties to be selected for specific adaptations allow trade-offs between traits In the first five of these, decentralizatíon or participation can provide a solution mainly by allowing more replíeation, particularly replícatíon that ínereases tbe number of test sites. Adding more researeher-managed test sites in a decentralized testing prograrn is expensive. Adding farmers in a participatory testing program is cheaper because there are many farmers who are willmg to collaborate wi th minimal cost. These six issues are considered m more detail below. 1. Allow trial sites to better represent tbe crop area Trials can be modified to better represent the target areas (or, indeed, tbe niches within areas) by having more trials divided mto more zones and types. However, clearly many more formal trials would be needed to do thls and the merease would consume many more resources. Participation provides a more cost-effective solution. Moreover, the participation offarmers does notjust allow varieties to be tested in more niches, it helps to identifY them. 2. Allow better representation of tbe environments in farrners' fields The formal trial system can be modified to reduce purchased inputs to farmers' levels. After surveying farmers' cultivation praetices, more realistic management can be adopted in 24 J Witcombe research-station trials. However, only participatory methods, which allow many farmers lo be sampled, can realistically accounl for the range of managemenl practices and sowing dates found in farmers' fields. Replication across sites is the key to representing the diversity ofthe envíronments offarmers' fields, and participatory methods would appear to be the only cost-effective way of achíeving the amount of replication required. 3, Increase the reliability oC the trials The overall relíability of a multílocational trial can be increased by increasing the number of sites, the number of replicates al each site, the size of plots, or any combination of these. Ofthe three components, the number of sites is the most critica!. The number offormal testing sites that can be controlled and managed by scientists can be increased but at considerable expense in both requirements for infrastructure and running costs. Increasing trial sites ís cheaper with participatory methods because farmers are interested in participating in varieta! trials without any financial incentive other than the provision of seed free of cost The major costs are then for data collection. Qualítative data are "scientific," analyzab!e, and more cheaply collected than quantitative data. Henee, ifbreeders and release cornmittees were prepared lo accept qualitatíve data on yield and other traits, rather than the current insistence on quantitative data, the costs of this data collectíon wou!d be considerably reduced. 4, Allocate resources more efficiently between varieties in different years oC testing Participatory approaches, because of the quantities of seed required, would concentrate on more advanced entries, which would automatically correct the imbalance that concentrates too many resources on varieties that are at an earJy stage of testing. With PVS, the number of sites, i.e., farmers' fields, in which a variety is tested can easily and systematícally be increased as a variety is promoted through the testing stages. 5. AUow varieties to be selected for specific adaptations The hígher the number of trial sites, the more accurately selection can be targeted to níches--either physical or socioeconomic. This allows specific adaptations to be exploited, as was seen for the example of Swarna rice in Nepal. Although a hígher number of tria! sites in the formal system would allow the selection of more specíficalIy adapted varieties, it is a more expensive altemative to increased participation. 6, AUow trade-ofIs between traits It ís certainIy feasíble to introduce a trade"'Üffbetween traits in a formal tríal system after consultative participation that determines the traits that farmers consider important and how farmers trade them off. Irials can then be split according to farmer-ímportant traits, e.g., trials for hígh grain yield, hígh stover yield, and dual-purpose varieties for grain and slover. Selection índices can also be constructed to allow the promotion of a greater range of varieta! types in any trial. These methods, however, are complex and requíre traits lo have standard weightings even though they differ from farmer to farmer and &om season to season. Collaborative participatíon that allows farmers to decide overall whích variety or varieties they prefer is a simpler and more effective solution. Six issues have been considered in this comparison ofthe roles of decentralization and participation. However, there is a seventh important íssue that only participation addresses. 25 Participatory Varíeta/ Se/eetion in High-Polentia/ Produetion Systems 7. Participation promotes the speed of adoption ofpreferred varieties No matter how decentralized a breeding program and its varietal testing system, if it does not involve farmers, it cannot directly promote adoption. Only participation can do this. Conclusions on PVS in HPPSs Other papers in these proceedings will attest to the efficiency ofPVS in more favorable agricultural environments (Virk et al., this volume; Malhi et al., this volume; Joshi and Witcombe, this volume). It is highly effective and has been demonstrated to achieve the following: • IdentifY and promote varieties that were not recornmended for the area in which the PVS was done (this means that the recornmendation domaíns of many varieties that are adapted to HPPSs are too small) • Inerease varietal bíodíversity (more varieties are adopted because farmers, when given choices, can identifY varieties for niches) • Promote acceptable recornmended varieties (recornmended varietíes are adopted more quickly in villages where PVS is done than in control villages) • IdentifY recornmended varietíes that are either not accepted by farmers or are poorly accepted However, PVS has certain Iimitations. It is dependent on a seed supply lo start the PVS trials, and often the seed of released varieties is surprisingly difficult to obtain. When nonrecornmended varieties are identified, the seed supply limits the speed of their adoption. The success of a PVS program depends on other externa! factors such as the timing and success of recent releases in the target area. PVS is much less Jikely lo be considered successful when introduced varieties compete against a very recently released variety that is líked by farmers !han when, perhaps for more than a decade, there has been no significant change in the variety grOWll. PVS is a1so dependent on pre-existing varieties. If there are no suitable varieties among those currently available, then it will not succeed. In contrasto PPB approaches that generate new variabiJity do not suffer from this limitatíon. In participatory approaches in maize and rice breeding in marginal areas (Goyal et aL, this volume; Kumar et aL, tbis volume), 30% gains in yield were obtained over the best varieties-about three times the rate uf genetic gain using conventional methods. Success in HPPSs is yet to be demonstrated but research in this area is underway (Witcombe et al., this volume). Participatory varietal selection in HPPSs is much more difficult to justifY to scientists and policymakers than it is in marginal areas where the need for and success of a different approach was evident. The need was clear from a lack of adoption of new varieties and the success of PVS has been convincingly demonstrated by many (e.g., Sperling 1996; Wítcombe et al. 1999). PVS in bigh-potential production areas is new research from which results are only just emerging. It ís an a1temative lo an entrenched system that can justifiably claim success-the adoption of modem varietíes ís, after all, almost universal in HPPSs. However, this success does not necessarily equal efficiency-a 100% adoption of modem varieties can be acbieved with or without extensive participation. However, could participatory methods be more cost-effectíve, produce better varieties, and create and maíntain greater varietal biodiversity in farmers' fields? The theoretical basis as to why tbis might be so has been presented here, and the evidence to justifY this theoretícal assumption is emerging. 26 J Witcombe It is extremely important for these issues to be thoroughly explored. HPPSs produce most of the world's food. I[the produetíon inereases from PVS of 10%-40% found so far in these produclÍon systems were to be widely replicated, this would have a considerable impact on improving food security and would directly, and indirectly, greatly benefit the poor. References Ashby, J.A., T. Gracia, M. del Pilar Guerrero, C.A. Quirós, J.1. Roa, .nd J.A. Beltrán. 1996. Innovation in the organizanon ofparticiparory planl breeding. In Participatory Plam Breeding: Proceedings ola workshop on participaloryplant breeding.16-29thJuly /995. Wageningen, The Netherlands, edited by P. Eyzaguirre and M. Iwanaga. Rome: Intemanonal Plant Genenc Resources Institule. Goyal, S.N., A. Joshi, and J.R. Wilcombe. Participatory Crop Improvement in Maize in Gujaral, India. This volume. Joshi, K.O. and J.R. Witcombe. Participarory variel.1 selection, food security and varietal diversity in. high-potential ptoducnon system in Nopal. This volume. Kumar, R., D.N. Singh, S.C. Prasad, J.S, Oangwar, D.S, Virk, and J.R. Wílcombe. Particípatory plant breedíng ín rice in eastem Indía. This volume. Malm, S.S., lR. Witcombe, D.S, Virk, and K.B. Síngh. Participatory varietal selection in rice ín the Punjab. This vo/~ ume. Maurya,D.M, 1989. The Ínnovatíve approach ofIndian fanners.ln Farmer First. Former Innovation and Agricultural Researeh, edítedby R. Chambers, A. Pacey, and L.A. Thrupp. London: Intennediate Technology Publicanons. Paekwood, A.J., D.S. Virk, and J,R. Witcombe. 1998. Trial testing sítes in the Al! India Coordinated Projecls-How well do !hey represent agro~ecological zones and fanners' fields? In Seeds oJchoice: Making the mas! olnew varieties lor smallJormers, edíted by J.R, Wítcombe, D,S. Virk, and J. Farringron. New Delhí: Oxford lBH; and London: Intennediate Technology Publications, Sperling, L. 1996. Resulls, methods and institutional issues ín partícípatory selection: The case ofbeans in Rwanda. In Particípatory Plant Breedíng: Proceedings ola workshop on participatory plant breeding, 26-29th July 1995. Wageníngen. The Netherlands, edíted by P. Eyzaguirre and M.lwanaga, Rome: Intemalional Plant Genetic Resourees Institute. Virk, D.S., D. Harris, E.S. Raghuwanshi, A.O.E. Raj, P.S, Sodhi, and J.R. Witcombe, A holislí. approach to participa~ tory crop improvement in wheat, This volume. Witcombe, J.R. 1999. Do farmer~participatory metbods apply more to high potenrial areas ¡han to marginal ones? Outlook on Agriculture 28:57-59. Witcombe, J.R., AJ. Packwood, A.G.B. Raj, and D.s. Vírk. 1998a. The extent and rate of adopnon of modern cultivars ín India. In Seeds olChoice: Making the Mosl oJNew Varietieslor Small Farmers, edíted by J,R, Wítcombe, D.s. Viril, and J. Farringron. New Delhi, Oxford IBH, and London: Inlennedíate Technology Publicatíons. Witcombe, J.R., D.S. Virle, and J. Farringron (Eds). 1998b. Seeds 01 Choice: Making the Most oJNew VarielíesJor Small Farmers. New Delhi, Oxford IBH, and London: Intennediate Technology Publícations. Witcombe, J.R., D.S. Virk, andA.G.B. Raj. 1998c. Resouree allocatíónand efficiency ofthe varietal testing system. In Seeds olChoice: Making theMosl olNew VarielíesJor Small Formers. edíted by J.R. Wílcombe, D.S. Vírk, and J. Farrington. New Delhi, Oxford lBH, and London: Intermediate Technology Publícatíons. Witcombc, lR., R. Petre, S. Jones, and A. Joshi. 1999. Farmer partícípatory erop improvemenl, IV: The spread and ímpaet ofa rice variety identified by participatory varietal selection. Experimental Agriculture 35:471-487. Witcombe, J.R. and D.S. Vírk. Forthcoming. The success ofparticipatory varietal selection in a high-potential producIÍon system in India. How important is the partícipation? In Portícipatory Piant Breedíng and Rural Development, edited by l.A. Ashby and L. Sperlíng. Calí, ColumbIa: COlAR Systemwide Program on Partícipatory Rese.rch and Oender Analysis, Centro Internacional de Agricultura Tropícal. Witcombe, J.R., K.D. Joshi, and M. Subedí. Towards a practical participa!ory plan! breedíng s!rategy ín predominantly self-pollinated crops. This volume. 27 Enhancing Biodiversity and Production through Participatory Plant Breeding: Setting Breeding Goals Bhuwon Sthapit, Krishna Joshi, Ram Rana, Madhusudan Upadhaya, Pablo Eyzaguirre, and Devra Jarvis Abstract Participatory plant breeding (PPB) is one ofthe on-fano conservation strategies designed to maintain or enhance the level of londrace genetic diversity deployed on-fano. The global in situ project .ims lO strengtben lbe skiIls and knowledge of local cornmunities in locating and understanding Ihe value of landraces, and also monitoring genetic erosion. Participation by fanoers and infonoal sectors in decentralized testing of materials can resolt in much greater diversity in the fields of coIlaborating fanners, as well as providing a broader range of varietal choiees and adoption. Germplasm exchange between farmers and farmers' selection criteria can also contribute to and enhance on-farm eonservation .nd biodiversity. Within the project, "Strengthening Ibe Scientific Basis of in situ Conservation of Agricultural Biodiversity," supported by the International Plant Genetie Resources Institute (IPGRI), case studies of ríce were mode in Nepal on consolid.ting the roles of farmer participation in PPB .nd seedexchange processes. Participatory methods, sueh as diversity fairs, diversity blocks, and cornmunity biodiversity registers, Viere used lO understand!he value oftocal diversity and .[so to strengthen the roles of fanners and informal sectors in tne local crop-development process. PPB programs in Nepa[ are designed to investigate (1) whetber f.rmers' cu[tivars per se can be eonserved, (2) ifPPB has contríbuted to Ibe enhaneemen! ofbiodiversity in terms of a broader genetie base tha! provides benefits to tbe cornmunity, and (3) if genetic improvement was been acltieved wilbout loss of genotic diversity. This paperdeseribes prelíminary resul!s of understanding genetic divergenee in terms of tbe use va[ue of loca[ biodiversity and lhe participatory methods used lO se[eel landrace parents. Methodological constraints of particípatory approaches in settíng breeding goal. in the context of biodiversity enhancement and production objeetives in biodiversity ríeh arcas are discussed, The paper .Iso documents how Ibe needs of farmers and ohjeetives of biodiversity enhancement can be integrated during Ibe setting of breeding goals aod supplying useful genetic diversity by bringing new, restoring o[d, and generating new genetie diversity (local x exotic) in Ibe agroecosyslem in three eco-sites ofNepal. Keywords: Particípation, PPB, PVS, rice, biodiversity, in situ (on-farm) conservation, diversity fairs, díversity block, cornmunity biodiversity regíster, diversity deployment Introduction Participatoty plant breeding (PPB) has been proposed as a strategy to enhance on-fann conservation tbrough use (Eyzaguirre and Iwanaga 1996; Jarvís and Hodgkin 1997; Jarvis, Sthapit, and Sears 2000) and thereby conserve the processes of evolution and adaptation of crops to their envíronments (Altieri and Merrick 1987; Brush 1991). PPB and on-fann conservation have a cornmon goal: both approaches encourage fanners to continue selecting and managíng local crop populations. Sperling (2000) reviewed the goals of 40 PPB case studies worldwide in which only a few cases addressed the objectives ofenhancing biodiversity on-fann. Most PPB programs aim to improve research efficiency and productivíty for the target environment and ofien do not have a transparent goal-setting process (WeltzienlSmith, Meitzner, and Sperling 2000). Consolidating the role of farmers in setting breeding goals and selecting segregatíng or variable materials in the process of planting, managing, harvesting, processing, consuming, and marketing their crops, PPB offers Bhuwon Sthapit, Pablo Eyzaguirre. and Devt. Jarvís are with the Internarlona! Plant Genetic Resources Institute (IPGRl)" Krishna Joshi and Ram Rana are with Locallnítíatives for Biodiversity. Research and Development (LI-BIRD), Madhusudan Upadhaya is wiili the Agricultural Botany Division, Nepal Agricultura] Research Council (NARC), 29 Enhancing Biodiversil:¡! and Producrion through Participatory Plan! Breeding farmers the opportunity 10 continue to select and manage local crop populations to better their liveIihoods and income. By definitíon, PPB involves deeentralization of the breeding process from research station to farmers' fields, bul it also ineludes significant farmer participation (Witcombe el aL 1996). With increasing levels of farmer particípation and decentralized testíng, PPB can enbance the deployment of genetic diversity and a1so broaden the base of landrace populatíons in a sustainable manner. Increased varietal diversity deployed among farmers' fields as well as within them is the key to reducing vulnerabilíty to diseases and pests (Zhou et aL 2000). The process also ensures farmers better access and control of acceptable germplasm (McGuire, Manicad, and Sperling 1999). Both PPB 1 and particípatory varietal selection (PVS2) are likely to have a negative ímpact on the diversity oflandraces. Both methods are íntended to change the structure ofloeal crop populations to make them competitive witb other available options that could totally replace the existing diversity ofloeal landraces (Witcombe et aL 1996; Brown and Young 2000). Nevertheless, PPB is Iikely to be more beneficial for conservation goals because it works with variable, segregating material s that are derived trom or similar to materials already in the local farming system (Brush 1999: 288). In contrast, PVS is likely to be negative for conservation goals because it is based on replacement of local populations with new and less variable ones trom breeding programs (Witcombe et aL 1996). It is therefore important to distinguish between the processes ofPVS and PPB if one ofthe purposes of PPB is to enhance biodiversity and production. There are concems that PPB products may replace díverse local crop populations, and new a1leles or combinations are expected to increase at the expense of other alleles, whích that may well disappear. Brush (1999: 288) advocates that as better alle1es or combinations arise and enjoy selective advantages, other less useful alleles thereby become less competitive and decline. This is the cost of evolutionary substítution and the price paid for a110wing evolution to continue. Therefore, ít is assumed that in situ conservation and PPB strategies faíl 10 preserve aH biodiversíty at the gene leveL Although it may not prevent a reduction in genetic variabilíty, it certainly limits the amount and rate of genetic erosiono There is limited published literature tbat has monitored tbe products ofPPB over the long termo Sthapít (1998) monitored the spread ofPPB products and theír impact on landraces fOI seven years. The impact of PPB was reported as positive in terms of biodiversity enhancement and production in hígh-a1titude areas ofNepal where rice varietal diversity was limited because of chi)ling temperatures and associated disease complexes (Sthapit, Joshí, and Witcombe 1996; Sthapit el al. 1998). PPB tends to enhance biodiversity in potential hígh-yield areas as well, where diversity has already been reduced by the rapid spread of modem cultivars (Witcombe 1999). However, the impact ofPPB in biodiverse areas is nol yet well documented in the literature. The factors that most influence how genetic improvement affects biodiversity are the production system (marginal, heterogeneous, or hígh potential), the degree offarmer participation employed in plant breeding (centraIized or decentraIized participatory methods), and tbe breeding methods employed (narrow or broad geneti~ base), The Global In Situ Projecf ofthe International Plant Genetic Resources Institute (IPGRl) has initiated PPB programs in selected countries and crops to examine such concerns. The goals ofPPB can l. PPB is defined as tne "selection hy farmers Qf genotypes froro genetically variable, segregaring materials" (Witcombe et .L 1996), 2. PVS i5 the selection offixed tines (released, advanced lines, or landraces) by farmers in their target environments using theirown seleclion criteria (Josili and Witcombe 1996). 3, The project, UStrengthenlng the Scientific Basis of itl Situ Conservation of Agrobiodiversity OnMFarm;~ is globaUy coordinated by IPGR!, Rome, witn three objectives: (a) understanding the scientifíc framework of farmers' decision-maktng processes in 30 D, Sthapit et al, ínc1ude improving and conserving germplasm, broadening the base of the population, enhancing germplasm (pre-breeding), increasing optíons and access to a wider range of germplasm to provide greater opportunities lo low-income furmers, empowering local farming communities, and increasing self-reliance. Not a1l goals will be addressed by the Global In Situ Project, but enhancing biodiversity and production is the central concern, with strengthening the capacity of farming communities in in situ conservation and utilization. Research questions Before PPB is accepted as a strategy for on-farm conservation, the following research questions need to be answered: • Can farmer cultivars, per se, be conserved in situ? • Can PPB contribute to the enhancement and conservation of landrace diversity in situ and provide benefits to the community? • Can genetic improvement be acbieved without loss of genetic diversity? This paper presents preliminary results from the PPB program in tbree in situ sites of Nepal ~Talium, lumIa (2000 m), Begnas, Kaski (600-1400 m), and Kochorwa, Bara (80-100 m}based on the objectives of enhancing biodiversity and production. The methodological constraints of participatory approaches in setting breedíng goals in the context of biodiversity enhancement and production objectives in biodiverse areas have been discussed. The paper also documents (1) how the needs of farmers and can be íntegrated with the objectives of biodiversity enhancement when breeding goals are set and (2) how useful genetic diversity can be developed by bringing in new, restoríng old, and generating new genetic diversity (local x exotic) in the agroecosystems of the tbree eco-sites in N epa!. Materials and methods Study sites and priority crops Three contrasting physiograpbic regions were selected to represent the bigh-mountain, middle-bill, and low altitudes of crop production in Nepal (table 1). The lumIa valley is remote, wíth basically subsistence-oriented traditional farming systems. lt has a unique range of crop varieties finely adapted to local conditions. The area is a transitional zone between valley bottoms (where a wínter cereal ís followed by a summer crop) and bigher elevations (where only one crop can be obtaíned). Most crop varieties are landraces except for few introduced vegetable and fruit crops. The valley is known for ¡ts cold-tolerant Jumlí marsbi rice, which ís the on1y variety grown at the bighest altitude of the world (Shahi and Heu 1979). Finger millet, barley, buckwheat, and cucumbers are other important crops in the valley. The second site, Begnas village is sítuated in the Pokhara valley of the Kaski district, a middlemountain ecosystem in Nepal. The Pokhara valley is known for bigh-quality rice in the western bilis ofNepal. It is characterized by a number oflakes, broad alluvial valleys, ísolated bilis, terraced on-fann management of crop diversity. (b) strengthening national capacity to plan and impiement in situ conservation, and (e) broadening the use of agrobiodlversity. The nirte countries involved in the project are Sumina Faso; Ethiopia. Nepai, Vietnam. Pem, Mexieo. Moroeeo, Turkey, and Hungary (Jarvis and Hodgkin 1991), 31 V.> , N' Table l. Contrasting Slte Characteristlcs ofIPGRI's in Situ Projeet, Nepal - ------ ----- JI.&';&I 01 lechnical and markat Agroecological Socloaconomlc Sil. characteristlc$ charaderlstics Human managed Talium, Jumla Hlgh mOtJntaln wilh valJey botlom Mlxed ethnícity: 55% Chetto Fertítizer for MV Populatlon size=4570 Fertilizor (2200-3000 m) Out~migration 866 mm/annum Total ctoPplng area =258h. ArcUc to cooI temporale Average farm·size~ 0.33 ha Bertey=5 Calcarle cambiaol, utric cambisol. utric fluvlsol 8 indigenous land-use systems Total households=759 Buckwheat=6 Begnas, Kaski Middle mountain with valley .. Rice = 21 . Impar/eel market Poor access to Compost =19.2-30.1 !!ha Taro:=1 Area under rice MV::;:O Cucumber-13 technologies and ¡nputs No road networl<: kg 87±25 kg (14) - partner~ __ NARC, LI-BIRO, NGOand OoA No 01 parcel. =18.9%0.9 Food'su!ficiency (montlhs) =7,5;;0.3 (600-14oom) 3979 mml.nnum Subtroplcal lo temperale Dystrict luvisol, systrlct eampisot, lutlhle reglsol, fturisol Tolal hOUS-_________.-_ __ Table 3. Steps Used in the PPB Process under Ihe in Situ Project in Nepal (Continued) Moda 01 particlpation PPB proCjlss PartlcipaUon Tools Farmer Breeder Institullon E. Impact of PPB' Baseline information on genetíc diversity to measure number of variatal choices available 10 farmers and their geographic pattem (domaln) Collaborative Inventory of fanner..named and exotic cultlvars and lis! of theit uses Time-series ínformation to measure the change 01 GD over time at village and community levels PhD research C8R Sample survey Effective number of oo~farm ... C80s LI-8IRO NARC - LI-SIRO ... LI-SIRD NARC landraces Pedigree allalysis to detect the level of alJelic richness and the level of polymor phism for genetic markers PhD research Need for monitorlng diversity of use-values, e.g., Jacal adaptatlon. uses fer local cuistne, refigious & cultural rituals. fodder value, etc. Collaborative Assessment of productivity and sustainability of the system Collaborative Llnking with national breedíng and ex situ conservation strategy Consultatrve w Impact on cropping systemSl1andraces that are not chosen for improvement by Molecular marker Average, weighted and temporal diversity estimates Focus-group díscussíen Usl el !armers' select10n criteria U-SIRO NARC Partícípatory • project evaluation CoIlaboraüve Number al landraces used as parents in national breedíng Collectlon 01 lecally rare landraces CBR NARC ... ." .. U-BIRD NARC U-BIRO NARC U-SIRD NARC Souree: Adapted and modified fram Joshi el al. (1999). Note: CBR==community biodíversity register. Passive role. ** Subordinate role. • *uLead role. 1. Imp.c! of PPB is no! the producl oflhe participalory planning process, The firsl step was lo locate ecosystems and communities Ihat harbor good biodiversity wilh local knowledge and interest. Multidisciplinary research teams assessed local situations using participatoI)' tools lo locale diversity, Ihe community's interesl, and the local capacity of cornrnunity-based organizations for participating in particípatory crop improvement. In 1997, a diversity fair5 was organized in all three locations in Nepal in order lo sensitize communities aboul crop genetic resources, to locate diversity and custodians of diversity, and to understand Ihe value of landrace diversity in Ihe context 01' the local 1'ood culture, market forces, and socioeconomic and agroecologícal settings. The materials collected from Ihe diversity fairs were displayed in farmers' fields as a diversity block6 to measure the rnorphological diversity structure within landrace popu5. A diversíty fair ís a dIsplay úf local crop diversity through competilion al víllage level (Rijal et al. 2000). 6. A ruversity block 15 a participatory research techniquetocharacterize local1andraces under condítions oftypical fanner manage~ ment Germplasm 10 be grown ín the divers1ty block may be selected from materiaJs displayed in diversity fairs or from cormnunÍty members' seed stocks. Fanners using traditional practices manage the crops, while fanners and scientists monitor the plants to observe and record agromorphologícal characteristks, Diversity blocks can be used ro select parent plants and sources ofseed for the crossing program (Sthapit. Sajise, and Jarvis 2000). 36 B. Slhapit el al. lations and also to analyze preferred and undesirable traits. Male and female farmers representing aH socioeconomic strata participated in this activity. Participatory methods, such as diversity fairs, diversity blocks, and community biodiversity registers, were used to understand the use value of local crop populations and to assess the richness in the use value of existing diversity. Amount 01 genetic diversity The genetic diversity in farmers' fields was measured by the number offarmer-named and exotic varieties, the number of farming households growing each variety, and the area covered by each variety. A baseline survey and participatory tools such as the diversity fair and CBR7 were used to collect the data and validate the inforrnation. Participatory rural appraisals (PRAs) and diversity fairs are quick methods for assess biodiversity in situ. CBRs provide a complete census ofbiodiversity at the community level. Distribution 01 genetic diversity Baseline household (HH) surveys were carried out on a sample of 180 to 200 households per site, depending upon the variability of agroecological and socioeconomic factors (Rana et al. 2000). Data were coHected for each cultivar on HH features, farm characteristics, the status of cultivars (growing environment, area, and productivity), preferred and undesirable traits, along with use value, including medicinal and religious/cultural values, if any. An SPSS Data Entry module was used for data entry and SPSSIPC and MlNIT AB were used for statistical analysis. Understanding the value 01 local crop diversity This is a key step before participatory goal setting is initiated. The CBR provides a list of farmer -named cultivars and their use value at the community level. Based on the baseline survey data and CBR, farrners' varieties can be placed into four broad categories in terrns of area cultivated and the number of farrners maintaining them in order to understand the relative importance of specific landraces: 1. 2. 3. 4. landraces grown in large areas 8 (wide) by manY farrning households (common) landraces grown in large areas (wide) by a few farming households (rare) landraces grown in small areas (local) by manY farming households (common) landraces grown in small areas (local) by a few farming households (rare) This kind of broad distribution analysis helps to understand why sorne landraces are grown in a small area by many farmers whereas sorne landraces occupy a large area but are grown by few farmers. It is very important to understand the pattem of such distribution and the reasons for such decision making by farmers. Setting breeding goals The dilemma project members faced before initiating the process ofPPB was where to start? On aH varieties? Those in high demand in the market or those maintained by only a few farmers in small areas? Or those grown by many farmers in larger areas? 7. Afterthediversity fair in Nepal, local cornmunities were motivated to keep an inventory offanners' varieties, including rare and endangered cultivars. This record is being maintained in the cornmunity biodiversity register. 8. Far example, at Begnas, 63 rice landraces were reponed in the baseline study (n= 206 households). The average area under rice varieties was 1.22 ha, maintained on average by 11 households. These figures were used te categorize area (> 1.2 ha = a large areaand < 1.2 ha=a small area) and frequency (> 11 =many HH and < 11 =few HH). The size ofareaand HH numberdiffered in the lumIa and Bara sites. 37 Enhancing Biodiversitv and Productíon Ihrough Participatory Plan! Breed... in"bg_ _ _ _ _ _ _ _ _ _ __ It was no! known which particular group of farrner-named varieties was important from a conservatíon angle. However, it was not possible to inc1ude all the landraces for PPB, therefore it became necessary to categorize them on sorne obvious eriteria. Sinee there was no previous literature available to guide this process, the team decided to categorize the rice landraces into four cells, based on the average area covered and the average number ofhouseholds growing them. This was planned in such a way that at least one representative variety from each cell would be included in the crossing programo A consultative participatory mode was used to assess the needs of farrners and the project goal. Focus-group discussions (FGDs), attended both by men and women farmers, were organized at the Kachorwa, Bara, and Begnas, Kaski, sites with the objective of identifying landrace parents for PPB. Farmers from across socioeconomic strata and gender were consulted in the FGDs to select Ihe landraces for participatory plant breeding. Participatory approaches were used to select al least one landrace per cell for parents. Listed farrner-named cultivars were analyzed using preference matrix ranking (Guerrero, Ashby and Gracia 1993) for preferred and undesirable traits in order lo identifY traits that needed improvement. The third step was lo identifY the best landraces from the four cells using preferenee matrix ranking. During discussion, the preferred traits of the landrace parenls were documented, while the traits Iha! needed improvement were thoroughly analyzed using ordinary ranking or paired matrix-ranking methods. Finally, landrace parents preferred by farrners for the PPB program were short-listed to represent each celL A relatively large number of landraces were selected from the cell with both large diversity and arca. The exotic parent was then identified looking at the traits to be improved in an individuallandrace, as well as adaptabílity of the exotic variety in the area and other farmer-preferred traits. F inally a cross-combinatíon for each of the sites (Bara and Kaski) was finalized by the team. Results and Discussion Amount 01 díversity Farrner-named varieties are a practical indicator of genetic diversity in farmers' fields. However, ít is assumed that there is sorne degree of inconsistency in farmers' naming andJor distínguishing traits. Table 4 surnmarizes the number of farrner-named cultivars of rice, barley, finger millet, buckwheat, taro, sponge gourds, cucumbers, and pigeon peas across three study sites in Nepal. Qnly the data on rice are consídered in this paper for further analysis. The míddle-mountain ecosystem at the Kaski ecosite harbored the highest number of rice cultivars (69), followed by Bara with 53, and Jumla with 21. The farrner-named cultivars at these sites were characterized to find out their genetic distinctiveness. The diversíty of micro-satellíte markers in random subset samples oflandraces showed uniqueness in Ihe cullivars. Therefore, the number of farmer-named cultivars can be considered as a measure of genetíc diversity on-farm. Sorne landraces particularly adapted lo heterogeneous areas tended to vary between subpopulatíons, as evident from molecular characterizations of Jetho budho, Basmati, and Gurdi (Bajracharya el al. 2000). We found Iha! severallandrace populations share a common name bul farmers distinguished Ihem by their specific morphotypes and uses (table 5). Most farmer-named cultivars of self- and clonal breeding crops have a high degree of consistency in their names and farmers' distinguishing traits (Bajracharya et al. 2000). We also found varialion 38 B. Sthapit el al. Table 4. Tbe Amount oC Genetic Diversity and Its Distribution in Three Eco-Sites of Nepal, 1999 # AVérage area Farmer.. # Consisteney of namad modem varlely vari8~ farmer..named variety and distinctiveness Area under crop (ha) undar fandrace (ha) Talium. Jumla (2240-30QOm) 21 O Medium 80.6 0.13 O Begnas, Kaski (600-1400m) 63 6 High 363.4 0.36:tO.02 O.2:tO.02 Kachorwa, Bara (80-9Om) 33 20 High 718.4 0.3tO.03 0.7±0.05 TaHum, Jumla 12 O High 34.7 0.04±0.01 O Begnas, Kaski 24 O High 133.4 2.2±003 O Kachorwa. aara 6 O High small 0.04±0.03 O HG NA O 24 O O High 6egnas, Kaski High HG 0.001-0.03 O Kachorwa, Bara 1 O High HG NA O 13 O Medium HG NA Begnas. Kaski 14 Medium NA O O Kachorwa. Bara 4 NA NA HG Medlum HG NA O Ecosite Rice Average afea under modarn v.riely (ha) Finger mllle! Taro Talíum, Jumla Cucumbw Talium, Jumla Sponge gourd Tallum. Juml. O O NA HG O O Begnas,Kasló 13 O Medium HG O Kachorwa. Sara 16 O Medíum HG NA NA 5 O High 119.7 0.07±0.01 O 6 O High 9.8 0.04 ±0.01 O 5 O High Bund planting 0.21±0.02 O O Barley Talium, Jumla Bu~kwheat Tslium, Jumla Plgeon pea Kachorwa, Sara Source: Rana e! al. (2000). HG= horne garden (few plan!s). NA= No! available. within fanner-named cultivars (rabIe 5). Work on molecular and agromorphological characterization is in progress to address problems of consistency in names; however, !he list of farmer-named cultivars can be considered a basic unit of diversity on-fann, used by fanners as a management tool. Dístríbution 01 rice diversity and use value Figure 1 ilIustrates the extent and distribution of rice diversíty from !he Begnas eco-site. It helps to understand !he importance of each farmer-named cultivar and the value of genetic diversity. 39 J¡;nhancing Biodivers.irv and Production throughParticipatorv Plant Breeding Table 5. Examples ofFanner-Named Landrace Populations and Farmer-Named Cultivars witbin the Landracc, Kaski Site. Nepal Landrace poputation Dlstln;uishing tralts Gurdi Gurdj type 01 grain io distingulshed by Fatmer-named eultivars Translation 1. Seto gurdi White gurdí 2. Naulo gurdi New gurdi small grain with b!ack apiculus color, niedium-quality grain, adaptad to hills Jhinuwa Jhínuwa group of rice is distinguished by fine black grain and known for cooking quality-such as aroma and softness of cooked rice 3. Kathe gurdi Foolhills guroi 4. Lahare gurd¡ Oustered gurdi 5. Ganjale gurdi Maskara eye shaped 6. Kalo guro! Black gurdl 7. Sano gurdi Small guroi 8. Thulo gurdi Blg guroi 1. KaJo jhinuwa Black fice grain 2. Chobo jhinuwa 3. Tarkaya jhinuwa 4. Pahenle jhinuwa Yellow fine graln 5. Tunde Jhinuwa 6. Kalo tunde jhinuwa Awned fine gralo 7. Lamo jhinuwa long fine grajn 8 Seto jhinuwa 9. Masino jhinuwa 10. Jnlnuwa basmati Jemeli JemeJi group of rice Is distinguished by medium-coarse and long type rice Ghalya Ghaíya group 01 rice is distinguished 1. Pakhe ¡emeli 2. Dhabe jem_~.!~_~_ 1. Seto gtiaiya for its ecological adaptation to dlred-seeded upland rice 2. Rato ghaiya Black~awned fine grain Whíte fine grajn Fine grain Seented fine ~raln Foothills. medium-coarse grain Adapted to swampy area White upland direct-saeded rice Red upland dlrect-seeded rice 3. Bicharo ghaiya 4. Gurdí ghaíya 5. Jjre ghaiya 6. Chaba 7. Jhayali rato ghaiya 8. Kunchhall ghaiya 9 Upland direct-seeded rice frort1 Kunchha Katuse ghaiya 10. Lahare ghaiya Clustered upland rice 11. Masino ghaiya Fine grain upland 12. Kanajire ghaiya (Continued on "ext page) 40 B. Sthapit et al. Table 5. Examples ofFarmer-Named Landrace Populations and Farmer-Named Cultivars within the Landrace, Kaski Site, Nepal (Continued) Landrace population Olstinguishlng traits Anadi Anadi group of rice is distinguished for its glutinous rice, coarse grain, and Farmer-named cultivars Translation 1. Seto anadi White sticky rice 2. Rato anadi Red sticky rice 3. Dudhe anadi Milky sticky rice bread leal Madhise Bayemi Bayemi rice Is distinguished for its high-quality, aromatic. and black-seeded fine grain 1. Thulo madhise Big rice trom terai 2. Sano madhise Small rice 3. Naulo madhise New rice trom terai 1. Junge bayerni Awned bayemi 2. ~alo bayerni 3. Bayemi jhimuwa trom terai Blael< bayemi Bayemi fine grain Landraces grown in large areas by many households. It is interesting to note that only 13% of the landraces were grown in large areas (> 1.2 ha) by many households (> 11). Only eight farmer-named cultivars were found in this cel!. They are used for four major purposes: subsistence, commercial, local adaptation, and quality. Table 6 iUustrates the comparative value oflocallandraces and their bad traits as perceived by farmers. We found sorne contradictions in farmers' perceptions of good and bad traits and, therefore, only frequently reported information is included in order to avoid such bias. Ekle, Kathe gurdi, Thulo gurdi, and Madhise are relatively high-yielding landraces adapted to high-yield-potential areas ofthe study site. Mansuli and Radha-7 are the only two modem cultivars cornmon in the area. Farmers use grains of these varieties for subsistence purpose, whereas landraces such as Jetho budho and Pahenle are grown to meet market demando These varieties are basically sold at the farm gate because oftheir high market price, based on their good cooking qualities. Resource-poor farmers in Begnas have allocated a large portion oftheir rice fields to Jetho budho (0.24 ha ±O.II) and Pahenle (0.20±0.14 cf. 0.14±0.14 ha). Through interviews, it was found that these farmers seU highly priced rice in the market and buy modem varieties for their own consumption. In contrast, resource-rich farmers and farmers in the medium wealth category grow modem rice (Masuli) in larger areas than do the resource-poor farmers. It is assumed that these farmers have good access to land with a high yield potential and have the capacity to purchase the inputs required for modem cultivars. In this category, 43 farming households also grow Mansara, which covers 14.3% ofthe area for this category. Mansara is specificaUy adapted to poor soils and low-input management. The faith in Mansara is safe as long as farmers do not find competitive options. This variety tends to be grown by resource-poor farmers; however, its adaptive traits may be useful for future crop improvement for marginal lands. No farmer wants to improve this variety because ofits low quality and poor straw and grain yield. From the conservation perspective, the question is whether these types of cultivars should be considered for PPB since many resource-poor farmers in large areas grow them. Or should we concentrate on locally cornmon landraces such as Pahenle, Ekle, and Jetho budho to address the needs of many farming households? Farmers proposed Pahenle, Ekle, and Thulo gurdi for participatory improve- 41 Enhancíng Biodíversíty and Production through Particip..atory Plant JJ..c;re"e",d:::in",g~_______________ Rice Jandraees grown in !arte area by rnany households Rice landracn grown in largl;J area by few households Thulo Gurdi (24) 6A% Paheníe (23) 7.<)% Jetho Budho (43) Elde (67) 32.0% Gurdi (9) 49,6% 8.6% Sano Madhise (B) 50.4% Kathe Gurdj ,0.5% Madl'lise (48) 17,8% ~atoAnad¡(71) 34,9% Pakhe,.".!~"" '12) _.J r-1 Saoo Gurdl (4) 10.1% Jamell (16) 12.0% Rice lal"ldraces grown in amalJ area by many hOusaholds Tunde ruce landraces grown In smalt atea by few households Figure 1. Categorization of rice landraces of Begnas, Kaski, based on area and number of Carmers growing them (figures in parenthesis represent number ofhouseholds) ment, whereas Mansara was selected for adaptive traits. letho budho was selected for landrace enhancement because it is more competitive to other available options. Landraces grown in large areas by a few farmers. There are very few cases in this category (figure 1). In Begnas, Gurdi and Sano madhise are the only two landraces found, whereas in lumIa and Bara, none was identified. These cultivars are grown for home consumption and are more tolerant to storage pests. Sano madhise is perceíved as good for better rice yield and soft straw for animals (table 7). Farmers perceived that Gurdi is valued for multiple positive traits, such as good taste, long straw, low input response, and better milling recovery. Farmers selected Gurdi for a PPB paren!. Landraces grown in small areas by many farmers. Figure l shows that few rice varíeties are found in this category. Most landraces in this cell have a special value in local food culture. Landraces with religious and cultural significance are likely to be maintained in small patches for local 42 B. Slhapil el al. Table 6. Comparative Use-Value of Rice Landraces Grown in Large Areas by Many Households Fa.....r- HH namad cuUivars , # Ekle Ecosystem Major use Irrigated Subsistence Us...value and constralnts perceived by farme", ' fanners : Good taste (97) HYP 87 i Undes1red traits perceived by High water demandíng (43) Long slraw (87) ! 800-1000m Prone to inseets in storage (35) Good mil~ng recovery (60) High nument demanding (36) Lale malUrít)! (27) Madhise Irrigated I 48 ! ! Subsistence HYP 600-90Om 47 Good laste (48) : Poor straw yield (21) Good miDíng recovery (19) I Requíres more water (12) & numanls (11) Good yield potantial (17) , I Tari Ka!he gurdi i Subsístence : LVP , 800-1400m Adapted to low-Input minfed condltions (45) : Poor straw yíeld (23) Low yíeld (6) Good tasle (36) Poor milling recovery (5) Early melUrit)! (21 ) : Adapted lo peor Rainfed tarI Mansara 43 early malUring (23) Poor tasle (22) land and low- Adapted to very poor soil and Iow-ínpul condltions (17) Poor stmw yield (22) Commercial HIgh-inpul demandlng (9) HYP Good-qualit)! aroma/sofiness (74) 600-00Qm Hlgh price (22) LYP ¡ input condltions 900-1400m Jetho budho : Irrigated teri 43 Poor mllling recovery (17) Low yíeld (6) Oualit)! straw for mato (8) : Irrigated tari Thulo gurdí 24 • HYP Subslstence . : 800-1100m , Good taste (36) Demanding for water & nutnents I Milllng recovery (17) (18) Low yietd (3) Good straw yíeld (22) , Pahenle : 23 Commercial Irrigated Iowland HYP 600-80Om Jhinuwa ghaiya 23 59 Good .Iraw yield and quality (12) Water demanding (20) Adapted to Good taste among ghalya (15) • Nutrí""t demandlng (6) I LYP upland poor soíl Market price (5) ¡ Straw value (4) ! Poor straw yíeld (3) Poor mllling recovery (4) Good taste (SO) Poor slraw yield and quallty (24) HYP : commercial High yíeld (40) More input demandlng (36) 100-700m • Modem variety Milllng recovery (25) I Adapled lo wann water (14) Irrigaled ! Low yíeld (4) Insect di ...a ... (2) ~ Tari 800-120Om Mansuli Good quality wilh aroma (30) High marl 1.2 ha) by Few Households « 6 HH) in Begnas Village, Kaski Ecosite, Nepal cultivars I HH# Sano madhise • 6 Farmer-named Gurdj : Ecosystem TalililTigated Subslslence Use~yalue and constraints perceived by fanners I Undesired traits : perceíved by farmers • High yield (3) ! Poor straw yield (3) HYP I 700-1100m '9 Irrlgated i HYP I Majoruse • 800-1200m ! Subsístence • Adapted to low~nput conditions (5) Relatívely good taste (5) Good milling recovery (4) Good straw value in tenns 01 yield (3) Poor yield (2) I Source: Baseline survey. 1999. Note: Figures in parentheses indicare number of respondents in survey. Only top three frequencies of positíve and negalive traíts were considered as perceived value ofeach variety. Tari = indigenous classífication of land types, upland rainfOO ríce ecosystem, for specific domes tic uses. Their small population size may ¡ead to genetic drift. Does PPB have the scope to merease their produetivity so that useful alleles from the Bayerni, Jerneli, and Anga populations are maíntamed? lfthe crop-improvement program is successful in incorporatmg good quality with yield advancements, will PPB products replace the díversíty of other landraces that are not chosen for improvement? Fal1I1ers value Anga for its multiple traits and it has been crossed with NR 10291-6-1 for better yield. Landrace enhancements for Beyerni and Jemeli have also been suggested for improving yield, Landraces gr()wn in small areas by few farmers. In all sites, the majority of farmer-named cultivars fall into this category. In Begnas alone, out of 63 landraces grown, 48 landraces were mamtamed by onIy a few farmers in small patches ofabout 0,5 ha (figure 1). We need to understand why farmers grow so many landraees in small patehes, as well as when and where they grow them and how they maintain and use them at the locallevel (table 9), Except for a few, the majority of landraces are maintained m small areas scattered in fragmented plots. This group of cultivars falls into locally rare materials, which should receíve priority for ex sito conservation. Of 48 cultivars, 24 farmer-named cultivars were mamtained by virtually a single household and can be defined as endangered, Should these be improved by PPB? Or are they candidates for a genebank before they disappear from the cormnunity? Do these varieties have specific genetic value? Or are farmers maintaining them because they do not have any better options? Or are they selected from locally cormnon landrace populations? If so, should they be candídates for PPB? 44 B. Sthapit et al. Table 8. Comparative Use-Value ofRice Landraces Grown in Small Areas « 0.5 Ha) by Many Households (> 11 HH) in Begnas Village, Kaski Ecosite, Nepal Farmer..named HH cultlvars # Ratoanadi 71 Ecosystem Major use Irrigated; dhab Food ""llure HYP 700-00Qm Setoanadi 54 lnigated; dhab : Food cuttura HYP I ! Jeme!! : 700-900m 116 Rainled I Bayami I 11 ! I LJ Poor mílling recovery (20) High input requirament (15) Low yi.ld (10) Good lar sticky latte nce (47) : Coa".. grain (6) , Good for many local reclpe. sueh . I (42) khatte (22) .oe t t :. asslrua. High-qualíly rice , Adapted lo low-input rainled condltions (45) for home consumption Good tasle (35) 6Q0-90Om i Good lor latte recípe (56) Medicinal value (59) Good for many local recipes such as sirula (35). khalte (17). puw. 16 ,tote 12), chiura 5 UndesJred traits perceived by fa"".ro I (7), cmur. (5). puwa h tarildhab I US8*value and constralnts percaived b tarmers I Low yield (3) I Poor millíng recovery (2) . Input demanding (2) Earty maturily (21) Tari Qualíly arornatie HYP rice forhome 7OQ..1000m Uninigaled tati . i LYP : l000-140Om I consumption Medicinal use Good qua/ily rice: aroma, .oftn ••s (22) Medicinal value (5) : Good lor mal. (7) . Adapted to very poor soil and rainled plots ( 9) Medicinal valu. ( 5 ) Good fodder Low yield (6) High input demanding (6) Peor tasle (6) Low yield (3) Red nee (6) ! Traits similar to wild rice (3) So"ree: Baseline survey, 1999. Note: Figures in parentheses indicate number ofrespondents in survey" Only t-op three frequencies oC positive and negative traits were considered as perceived vaIue of each variety. Tari=upland rainfed rite ecosystem. Dha/J= permanent waterlogged ecosystem. Table 9 illustrates the use value ofminor vaneties from the Begnas site. For example, Sano gurdi is valued for its moth tolerance in on-farm storage, whereas Biramphoo¡lO and Ramani are kept for their excellent cookíng quality. These vaneties are, however, low yielding and special skílls are needed for cookíng them in the traditíonal kasaudi (a thíck, round, nickel pot for slow cookíng on the fire). The introduction of rice and pressure cookers has replaced old cookíng practices and skílls and has also sIowly reduced the demand for these vaneties. Naltume is a niche-specific vanety adapted to shaded areas. Tunde is concentrated in droughtprone plots. Many Jerneli and Bayerní types are maintained for multiple quality traíts despite their low yields. More case studies may be needed for varieties that are conserved without special value. The challenge is ro identify the special genetic value of these rare Iandraces and find ways to assist the continued selection oflocal Iandraces that conserves the evolutionary process oflandrace diversity. TraditionaI knowledge about such cultivars is limited, as few mrrners maintain them. In the PPB program, farmers decided to select Biramphool for its high-quality traits, whereas Naulo madishe was selected for its local adaptatíon to rainfed condítíons. Biramphool wíll be crossed wíth another modern aromatic rice with better plan! stature. Naulo madhíse is crossed with IR 36 to incorporate its good yield potentíal (table 9). lO, The survey showed that two resource-rich farmers conserved this variety in an area of 0.30 ha. This variety is higbly valued for its aromatic quality, which is controlted by a single gene. 45 Enhancing Bio. Maniead, and Sperling (1999). Details ofinternatlonal cQUaooration in this area are also given in Sthapit. Joshi, and \Vitcombe (1996); Gauchan, Subedi, and Shrestha (1999); Jarvis, Sthapit, and Sears (2000); Witcombe, Vírk, and Farrington (1998). 7. In the late 19705 in Uttar Pradesh, there was an u on4arm" program involving several universities te develop composite varieties and relevant agronomy practkes for fanners with high fanner involvement. At the same time, a report by a very experienced maize breeder invo)ved in a Centro Internacional de Mejoramiento de Maíz y Trigo (CIMMYT) regional network was stopped because it outlined the altemative approach to maize crop improvement that was being tried, Deep controversy i5 not new to plant-breeding practitioners, 64 S. Biggs and D. Gauchan Farmer-Ied or formal-Ied research. Anolher dicholomy in Ihe literature is Ihe difference between "farmer-led research" as opposed lo "format-Ied research" (Weltzien et al. 2000; McGuire, Manicad, and Sperling 1999). This distinction worries us on a number of seores. First, Ihis is a sort of contradiction ofterms. If Ihere is respect ror Ihe different knowledge Ihat farmer and researchers bring to a discussion, then each party recognizes that there are things the other party does nol know. Each party has lo trust tbat the judgments being made by the other party are in good faith. In farmer-led research, farmers have to trust the "relevance and safeness" ofthe exotic materials being suggested by the researcher. Yes, the farmer may decide to use Ihal new variety; however, ifthal variety has come, say, from research processes using modern biotechnology methods, the farmer will have to trust the information abeut those processes that the scientists have provided. A different, equally qualified researcher ftom the formal system, but with different scientific judgments on genetically modified organísms, might have given the farmers different advice and farmers might have made different choices. Clearly, the information scientists give affects the path taken by farmers even in the farmer-led approach. Terms such asfarmer-managed orformal-managed research might better represen! the difference that is being searched for here. However, even ifthal change is useful, we may have the situation (ifPPB is like olher types ofparticipatory technology-development processes) where there is always fluidity in the process. What started off as farmer managed might become formal managed and vice versa. A second concern centers on the notion tha! it is useful to separate farmer-Ied from formal-Ied research. There is considerable evidence to show that new technology is often developed and promoted by formal and informal allianceslcoalitions/fartnerships. g Those involved in participatory plant breeding projects in Nepal confirm this view. This is also the pattern of cooperation that is emerging in a sustainable soil-management program in Nepal (Subedi and Bajracharya 2000). The problem, then, with farmer- or formal-Ied difference is that it places emphasis on a simple difference, rather than helping us to look at and understand the way alliances/partnerships were formed with different actors playing complementary roles. To place the emphasis on who "Ied" the process might be misguided if one were interested in a broader-based institutional analysis of what happened and why. As to the formation of allianceslpartnerships in the future, farmer-led or formal-led classifications might well hinder rather than help in the promotion of collaborative activities. A third problem centers around the notion that farmer-led PPB is in any sense a new phenomena and Ihat ways forward might be based mainly on the experience from recent projects Ihat have promoted farmer-led PPB. Our reaétion comes partly because Nepal and India have a very long open boarder along the Tarai, which for all practical purposes cannot be regulated regarding the flow of germplasm. Farmers cross from one side to the other and sélect materials in a truly participative manner. This has been going on for generations and continues very actively toclay. Sorne farmers (perhaps with and without formal Iraining in science) even go to sorne of the big agricultural uilÍversities in India to seek advice and obtain new varieties. We would call these activities "farmer-Ied inítiatives." They are very cornmon and have resulted in the spread of improved materials in Nepal from India and vice versa. We would sugges! that this source of information on innovative farmer-led PPB should be used as much for understanding the past and for guiding future technology and institutional policy as more recent "prnject-inspired" ínítiatives in this area. 8. See Tendler (1993) and Biggs and Smith (1998). 9. See Sth.pit and Joshi (1998). This was .Iso. strong view expre,sed to us by Dr. Aní! Subedi ••he direo.orofLl·BIRD, ane ofthe major NGOs involved in PPB in Nepal (Subedi 2000). 65 Plant Breeding in [he Changing Agricultural and Natural Participatory crop improvement versus conventional crop improvement. A recen! new term to be promoted by sorne of the dominant actors in this area is participatory crop improvement (pe!), which by implication is differentiated from conventional crop improvement (CCI). The term enables nongenetic technological options, such as integrated soil fertility management, integrated pest management, irrigation, mechanization, and post-harvest technology, to be taken on board. For those who havc worked in India, this tcrm has resonance with the AH India Crop Improvement Programs. Experiences with the practice ofthose programs as planned approaches to scienee have been very mixed and their periods ofhigher or lower usefulness can only be understood if one looks at the broader macroeconomíc poliey context and the institutional context of seience and teehnology at the time (Rajeswari 1995). A c10ser examínation of the activities of those programs would show different types of participation of farmers, NGOs (both local and intemational, sueh as the Rockefeller and Ford Fouodations, CIMMYT, etc.), public-sector researchers and development agencies, and other actors in differen! arenas at different moments in time. A detailed institutional analysis of those networks/partnerships/eoalitions would be interesting for informíng future policy in the area of integrated crop improvement policy. However, the point that is being made here is !hat a new simple dichotomy is being introduced: peI programs on the one hand and CCI programs on the other. We feel that this simple approach will neither help us understand the past nor provide a good framework for addressing the complexíties of developing partnerships/linkages, etc., where sorne parts ofthe "old" conventional approaches might be relevant lo "new" participatory approaches. From a participatory perspective, one of the most difficultproblems of getting different groups ofresearchers (both social and natural scientists) to participate and work together is as likely to be as difficult for the new PCI programs as it was for the CeI programs. Narrow definition 01participation A further characteristic of the current debate is lo define "participatíon" in a very narrow way.l0 Participation appears to concentrate mainly on only two sets of actors: "researchers" (plant breeders) and "users" (farmers).There appears lo be little analysís of relationships within these two groups, e.g., gender analysis as regards the significance of the relationships between men and women of different ages within the user or researcher groups is little analyzed in sorne PPB programs. We mentioned earlier the difficulties of getting different groups of researchers lo participate together in R&D, and there is'plenty ofliterature on trus subject. ll However, for contemporary analysis of participatory issues, we have to take into account tbe relationships between different actors in the private and public sectors. This has special importance if one is concemed with the actual practice ofR&D as well as "implementatíon" and "institutionalization" issues. Whether a plant breeder trained in formal scientific methods is in the private sector (e.g., a furmer or a seed merchant in a private breeding company), in the public sector, in the NGO sector, or par! of a donor project, etc., will in most situations have great significance for any debate conceming the usefulness and spread ofPPB. In most situations, tbe history of specific actors and the history of tbe institutions will also be of significance. A narrow discussion about relationships between researchers and farmers, with little or no serious analysis of these broader institutional 10. In this paper we do not address the issue ofhow different communjty/vlllage levei actors perceive the enterpriselproject being undertaken by researchers. For a salutary aniele which demonstrates that the "outsider" will always be víewed in ways that tha! outsider may no! want, see Burghart, (1994). 11 See for example Memll-Sands, et al. (1991); Byerlee. Triomphe and Sebillotte. 66 S. Bíggs and D, Gauchan issues may well be diverting us away from sorne ofthe more important topics of contemporary research poliey. 11 is interesting to refleet thal there is never any situation in R&D where there is no participation. There is always participation by definition, as R&D is a social proeess. It is the nature of the participation that de serves analysis. We would argue that the use of simple dichotomies and the narrowness of the PPB debate have been sorne of the reasons why major policy issues in thís arca have no! been addressed l2 : issues such as the sharing of inforrnation and the mobilization of national R&D human resources (in whatever institutionallocatíon) to address national R&D problems. In addition, there are legítimate concems about CPB prograrns tha! need addressing by the promoters ofPPB. Legitimate concerns of CPB programs There are five importan! concems about conventional breeding prograrns in theír resístance to accept and adopt PPB principies and praetices in N epa!. They are (1) fear of vulnerability of genetic material s to disease/pest epidemics when they are promoted in the early stages ofbreeding, (2) concerns for wide agroecological adaptation of the genetic material s, which prolongs ínvolvement of farrners in the early stages, (3) requirements for a regulatory framework for variety release and seed productíon and promotion, such as distinctive, uniforrn, and stabilíty (DUS) characteristics, (4) inadequate knowledge on the part of participating farrners about the future genetic performance of materiais, (5) the nature of partnerships and ownership of the participatory technology-development process. Reasons for the narrowness of the debate and the fragmentation of activities Before continuing, ít is worth exploring sorne of the reasons for the apparent narrowness of the plant-breeding debate and the lack of concems wíth broader natíonal research polícy íssues. Demands ofthe project cycle One of the reasons is the demand of current funders to see the "impact" of "their" projects. The project cycle, with íts assocíated methods and techníques of log frames, reviews, and monítoring and evaluation (M&E) activities sometimes leads lo the creation of differences, for project purposes, In competitive situations, differences may also be created in order to secure funds. Differences are created categories, which in sorne ways are created for reasons related to the project cyc1e rather than to helping apply science lo problems. One of the reasons for the idea that PPB ís "new" comes from the need for funders to fund "new" ideas that can be tried and experimented wíth before being multíplied and promoted more general1y. Thís makes a far more exciting project for a funder to support than a project that suggests ít is trying to change the balance between different altematíve approaches and methods in a broad-based crop-improvement strategy. The fbrrner makes claims of 12. The type of debates mentloned aboye have led to sttuations in Nepal for example where PPB projects have been critical ofCPB, but faHed tú point out that roan)' of the useful v3rieties used in participatory program carne fmm ePB programs. On one occa~ sion, an eminent plant breeder connected with a major CGIAR center, visited the country tú work wíth an NGO on PPB. However, even though the center had had for many years worked in partnership with the relevant pub líe sector plant breeding programs, the visitor did not contact the public sector program .. This was in part due to the separation between PPB and CPB, This approach in our yiew, does not encollrage new creative partnerships, 67 PI~n¡ Breeding in ¡he Changíng Agricultural and Natural Resolirce. R&D Sys¡em in Nepal c1ean-cut separations and dichotomies, which many reviewing such projects know are based on many unsustainable assumptíons. In the present context, sorne ofthe ways PPB projects are beíng drawn up and described as separable from other parts of the R&D system are not necessarily leading to the better use of scarce resources in Nepal. 13 Social status and institutionallocation 01 researchers The second set ofreasons for sorne the activities conceming PPB in NepaJ concem the history, experiences, seniority, and status of people in different parts of the overall research-and-technology promotion system. While many scientists subscribe to an ideal model of scíence where there is a free flow of information and ideas, the way science aetually takes place in any situation is affected a great deal by the social attributes ofthe actors involved. In its most obvious form, these things are reflected by who gets invited to meetings, what makes up the formal and informal agenda, and the authority ofthe meetings to have "recommendations" "implemented." The reasons for the demise of the strong NARC farming systems division and the current locations of the old staff of the two British-funded stations in Nepal are important institutional determinants for the nature ofthe curren! discourse conceming the pros and cons ofPPB. Path dependence A third importan! determinant affecting tbe nature of the present activities and discourse concems the "path dependence" of scientific debates (Hogg 2000). It Ís very hard for organízations to change, especíally ifthey are part of, or are linked to, established bureaucracíes or traditions in the way they do things. Thís ís not only for technical research priorities but also the institutíonal mechanisms for managing research and establishing línkages with other actors. In the current context of PPB in Nepal, it ís very challenging for the Nepal Agricultural Research Council (NARC) to consider the possibility of developing new partnerships wí!h the private sectorlNGOs for plant breeding and the on-farm testing of a wide range ofmaterials where farmers and NGOs can have a major say in decísion making (Sthapit, Gauchan and Rana 1996). Issues ofpath dependence are a1so key features for explainíng sorne ofthe behavior of different parts ofthe CGIAR system and other international actors. Sorne of the consequences of the battles within the plant-breeding fraterníty about wha! ís or is no! PPB, PCI, or CPB and who should or should not be doing what, are that important policy issues are not being addressed and scaree human reSOUTees in Nepal are no! being well used. Opportunities to have polícy debates and actíons on such things as a strategy for future rice research in Nepal, using all known professíonal expertise whether in the public or private/farmerlNGO sector, are being missed. Sorne of this is due to the narrowness of tbe PPB debate and the associated behavíor of the actors involved. Participatory technology development activities in Nepal We have suggested that sorne ofthe intemational and local plant-breeding debates have been narrow and parochial in their orientation. We now briefly describe sorne ofthe history ofbroad-based 13. There has been a Jot written about the way dIfferent actors use concepts oflogframes, cost!benefit anaiysls, the proJectcyde, etc., for differentpurposes in general developmeut practice. For ex.mple. see G.sper (1997), Woód(1998), Bíggs (1997), Apthorpe and Gasper (1996), Clay and S:e::::m",e",n,--t_ _~ _ _ _ _ _ _ _ __ Table 3. VarietaI Knowledge and Preference Ranking of Maize for Men and Women Women Men 1. bhote' pahelí 2. pahelí 3. dudhe' selí 4. bhole' selí 2. dhude' selí 3. paheli 1. manakamana-l (MV) 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 "iunovative" 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 employed 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 involved 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 cenlers 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 production 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 between 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 situated 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 contribute 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 empowerment. 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 technícal 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 language 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 hungerpangs 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 confusion, [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 appeasement. 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 territory 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 explanations and thereby becorne an ernpowering elernent for farrner participation. Wíthín such a context, 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 ofagricultural research and techn%gy promotion. Agricultural 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. Freire, 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 Programo 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, Maryland: University Press of America Khanal, N.R. 1992. Study ofgeo-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"technology 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 acceptability of new erop varieties generated through the breeding programo C,n Tho University, as the leading researeh institution for adapting participatory approaches to rice improvemenl, 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 Resourees 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 selections 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' involvemenl in the whole process or, al least, in the evaluation process. This has resulted in better understanding 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 genetically variable, segregating materials and PVS as involving the selection by farmers ofnonsegregating 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, therefore, 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 materials. 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 selecting 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 usedlor 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 ,-----~ Breeders Technicai asslstance lit : r" ' - - - - - - - - ' MooitOfing Project Sta" Materíals !Crop.improvement sedor ; , PGR conservation & development ¡ Supporting sectors ~ Experience shafing • Traíning - Farmers' field days - Field visrt ·IPM1. EPM2 .1 i------------to>I¡ Group 1 (2·3 rafmOn;) .1 (5·10 Group 2 '------------+1. rarme,.) I I PartlclpatQry plant breeding ·Broeding . Segregaq materials selection On·farm trials ~ PVS • Varietal yeUd triai - Seed selecOOn 1 ' - - - - - - - - - - - - + 1J ~ Group 3 (3~Or~m) I 1 ! • Seed muttipfication • Seed distribution ~ FarmenJ In the eommunity I • Planting material$ & saed production - Olssemination I Farmet'$ In oth$l' communltfes I I~Advanced fanners with good knowledge and skili in breeding. Group 2~Fanners with good knowledge and .kiU in seed seleclion. Group 3~F'llflers with good knowledge and in seed production. 1. lPM~Integrated pest management. 2. EPM~Ecological pest management. Note: Group ,km 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 F4) to the farmers ror selection 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 assistance from breeders and agricultural extensionists. These farmers also keep records on field conditions 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 separate 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 speeialized 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 eonununity, 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 finished 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 multiply 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 monitoring and evaluation ofPVS activities. Breeders, field staff, extension workers, and farmers participate 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 13240108-2-2-3), 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 selection to itnprove tbe formally released varieties for grain quality and adaptation to specific conrutions 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 Conservation (CBDC) project, PPB and PVS have been íncluded in their current form in the crop-improvement 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 , III.Y Thanh Nhut Ninh Testad Selected TR 252 8 DWR 20 6 MR 18 4 HYV 5 1 TR 23 3 Vea, 1994 1995 HYV 1 Tested Selected Ke Sach Testad Salected TR i 4 5 5 1 22 2 5 3 169 16 9 1 1 MR HYV 1997 9 ,TR MR 9 222 2 7 Lost' ! MR 11 HYV 12 22 1 34 9 32 29 20 9 89 _, i HYV 1998 Selected Tested I -~----- 1996 long Thanh 25 16 8 20 1 3 12 6 I 8 18 19 9 24 5 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, Table 2. Common Varieties Selected from PVS Activities at Four Communities in the Mekong Delta Rice varieties TR i Nhut Ninh My Thanh Nep Thom, Tal Nguyen, Me Huong long lhanh TaiNguyen MR MTL83, MTL 124 HYV IR54883, 8976B, MTL138,205 IR49517, IR64, MTl156. 157, MTL159,199 KeSaeh Souree: CBOC (1998), Note: TR~Traditional rice; DWR= Deep-water rice; Y!R~ MTLa3 MTL99.101. MTL142, 157, MTL164, 190, MTL199,201, MTL202 IR64. MTL 138, MTL142,147. MTL149, 150. MTL 156, 157. MTL159,199 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 __________________________________________________________~N~gu~~~n.~gocDe 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,) F, Community : F, ... F4 F, 13 MyThanh 20 8 3 1 10 4 2 1 Long Thanh Total : i Farmers' selection 13 Nhul Ninh Ke Sach - 20 11 63 36 L246-10-1-B L246-7-3-B (SiC-1) L247-1·5-B (SiC-2) i 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 multiplying 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 ~ducationallevel 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 materials 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, additional funds, and facilitation is very important. Cooperation with groups and eornmunities on PPB and PVS gives better results than working 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 Vietnam 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[ 1997. 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 Conservation 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 Educarion. Xuan, V.T. and N.N. De. 1993. Present status of agricultural extension in Vietnam. Paper presented at the first Southeast 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 stapIe 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 consumption: 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 during 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 always 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 potatoes, 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 February 1998. The experimental design ls a randomized complete bloek wifu three replíeations. The size of individualplotsís 1.6 mx 3.0m. Spacingis 80 cm betweenrowsand 15 cm to 18 cm betweenbills. Harvesting 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 Criteria of selection Resulls (in arder) Production Kinta, Toweko, Lemekuara, Umakmbi, Pipombi 11 Ski n color Toweko, Pipombi, Lemekuara, Umakmbi, Kinta 111 Root shape Umakmbi,Toweko, Kinta, Pipombi, Lemekuara IV Flesh color Toweko, Umakmbi, Lemekuara, Kinta, Pipombi V Unilonmily (shape and size) Umakmbi, Toweko, Pipombi, Lemekuara, Kinta VI Skin smoothness Toweko, Pipombi, Lemekuara, Kinta, Umakmbi VII General acceptance 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 appears to be the most preferred cultivar in this area. Farmers in Desakolot plant sweet potatoes for cornrnercial purposes. They have several requirements, 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 ACCQS- No sion No. Local name W0139 Toweko 2 W0331 Kinta 3 WOlll Umakmbi 4 WOl13 Lemekuwara 5 WOl09B Produc- Ski n tion color ... ~ Roo! shape Uniformity Skin Flesh (sh.pe smooth~ Generat color and slle) neS$ acceptance ..... 1> H·U' • ..... "'~ .. .. 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 exhíbíts 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 season 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 multiplyíng 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 potato as a source of planting material. 167 Using Former Knowledgefo.~.f.ar(icipa/ory Swee/-Po/ato Variety Selee/ion in Garu/, West Java, indonesia Pipombi (WO 1098) 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. Indonesia: 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. Schneider. 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 No 1 Accession No W0131 I Cultlv3r ,Son 11 0.56 . 0.14 Not marketable X I 11 111 X 1.94 0.88 1.81 1.67 0.83 1.44 8.47 8.93 2.36 1.39 2.92 • 2.22 111 2 WOI94 Yaronambiri 5.83 12.5 3 W0116 Helalekue 7.08 7.08 7.22 7.13 2.92 • 1.11 1.94 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 2.06 1.39 1.67 1.71 12 W0158 Musanaken baru 15.14 1028 9.31 5.42 3.19 3.19 3.93 13 W0220B Helalekue lama B 0.37 - - 0.69 0.23 14 W0220A Helalekue lama A 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 10.28 10.23 2.22 2.22 2.50 2.31 ¡ ! 0.26 • 2.06 2.50 -~,.'1 1.25 , ¡ 1.99 :;, 22 W0139 Toweko 12.08 8.33 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 4.26 0.14 - 31 W0331 Kinta 1.67 2.22 1.81 1.90 32 W0339 1.53 : V8 0.97 : 1.76 33 W0253 34 .W0041 ¡ 0.56 12.22 - 13.19 12.22 8.61 11.34 Kuning 10.97 5.69 9.17 8.61 Yoban 4.58 • 4.72 5.28 , 4.86 1.39 Pusemangken 0.42 1.53 0.65 0.83 I - I ! I 2.22 0.05 1.67 1.76 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 Ma~ketability (Continued) Yield (Ton/Ha) Marke!able No Accession No I Cultivar 11 No! marketable 111 X 1 11 111 X 35 WOO10 IMusan 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 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 4.17 4.31 10.28 6.25 3.61 '.~~ --- I 48 W0156 Soepak Saru 49 W0206S Andelan B 4,72 0,56 0.42 1.90 153 • 097 1. 50 W0206C Andelan C 1.67 1.25 1.25 1.39 1.11 • 0.42 0.69 51 W0167 Anewun 0.83 - - 0.28 0.56 0.28 0,42 • 0.42 Tabimbi 4.03 5.69 5.28 5.00 0.83 0.14 1.11 0.83 1.25 • 1.02 52 I O 0.74 0.69 53 WOO05 Hoboak 8,19 2.22 6,53 5,65 0,97 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 3.33 - 1.81 1.71 0.83 0.56 1,39 0.93 Bis 183 12,36 13.06 13,61 13,01 403 028 4,44 2.92 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 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 4,58 3,33 4.44 59 WOO39 60 61 170 ! Tinta Kuning - I 4.95~ i I - 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. individual 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 distribulion within domains can a program on diversity deployment and biodiversity conservation be effectively 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 ecological 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 important cereal crop in the regíon, so farrners have an in-depth knowledge ofrice-growing environrnents 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 strategic 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 methodologies 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 representatives 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, technical 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 characterization. Participanls were also asked to estímate the size of each domain and place different landraces/varieties in their right domains. Estimating tbe relative size of each domain was straíghtforward 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 complicated 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 disagreements 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 5011 type Production potentlal Cultlvated landraceslvarletles Ucha (bhadalya rice cullivated on availability 01 water, good wlnter crops) Balaute " sandy low (111) Mutmur, Sotwa, Sokan, Sara,., Samlal (Good erop 01 bhadaiya rice and winter erops, aaghanl rice can be grown) Domat" Loam No modem varleties grawn, (ujar " whitlsh) HIgh (1) Balaute domat " sandy lcam (whltish and brown) 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 Nicha/Khalar (Good erap of aaghani rice and medlum winter crops) Matiyar " Clay1 PokharllMan (can only graw aaghanl rice) Matiyar " Clay? Hlgh (11) (Piyar" Yellowi5h) Ba5matl, Lajhl, Mansara, Karma, Balsar, Rat ranl, Faram, Kamod, Madhumala (Mansula, Sabetri, Pankaj, Nat masula, Jaya, K, Mansuli ... )-Modern varieties (kalolkariya Low (IV) =black) Bhati, Megraj, SilahouL No modern varieties grown. Source: Chaudhary (2000). Table 2. Agroecological Domaln! at Begnas Eco-Site Domains Slzeof doma!n Mula khallBhale khetlKhule !2 ha) 8% 4% 1.00 ha 0.70 ha Male head 72% 51% Female head 40% 14% 7 7 Share-out Owner-cultivated Labor inputs in rice (dayslha) Male farmers Female famners Categories ollarmers (%) Ave. operational size Literacy rates (%) Average family size 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 application is done by women in Basalatpur). In Mungeshpur, women from the lower social status dominate 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 contributions 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 adoption rates in Mungeshpur, these varieties ofien suffer from submergenee, drought, and stress al reproduetive 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 MVs 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 levels of fertilízer, and more irrigation, but more farmers prefer to grow MVs 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 Mashurí, Oríswa, Malwa Ari. Bagrí, Balbagra, Chaini Improved NDR-97, Sa~u-52. PNR-381 Saket-4, NDR-80, 91,118 NDR-359, Pant-4. Pant-10, Pant-12, Sarju-52 TradHional Kalamanak, Motibaddam, Malwa, Malasia Bilaspurí, lodrasan Improved Mashuri, Rajshree, Sambha Mashuri Mashuri, Madhu, BKP-246, Dwarf Mashun Shaliow lowland~owland Topographical adaptations Farrners generally match varieties wíth their environment. For rainfed rice, this means an adaptation to the hydrological conditions of their fields, Each field position in the topo-sequence corresponds 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 progressively 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, Motibaddam, 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 variety, 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 ----- lowland Upland female Male Fomale 36.67 39.50 48.67 25.83 34.5Q 0,67 Grain príce 0.00 0,00 Resistance ID abiotic stress 8.33 Biomass quality Traits Male Grain yi.ld Ouration Upland lowland Mal. Female Male 49.67 41,67 35,96 42,06 40.45 1.00 20.56 25,84 20.56 15,QQ 15.67 16.00 1.67 2.81 2.97 1.82 6.10 0.67 0.33 6.10 6.18 5.10 5.00 3.33 2.50 5.33 4,61 5.00 2.25 5.52 8.64 ! Female 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 Postllarves! quaHIy 3.33 3.00 2.33 0,67 5.00 4.49 5.52 6.36 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 0.00 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100,0 TOTAL 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 recovery. 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 upper 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 difter in agrohydrological char~cteristics 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 adoption, 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 varietíes 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 postharvest 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 varieties. 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 lowland 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 responsibilities 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 Listening to Farmers' Perceptions through Participatory Rice Varietal Selection which flowers in 136 days, and Sugandha, which flowers in 124 days). Scientists distributed the seeds through the FPB project. In this approach, breeders select the most promising lines with farmers, and farmers are given a "basket of choices," growing several genotypes in their specific environments. Ten farmers (five women and five men) visited the individual plots and ranked the rice genotypes grown on farmers' fields past the maturity stage. Farmers were asked to rank the rice lines from I (exceIlent) to 14 or 16 (worst) on the basis ofvisual assessment. The rankings ofthe new cultivars by the farmers generated an n x k matrix, where n equals the lines being evaluated and k equals the farmers evaluating the crop performance. KendaIl's Coefficient ofConcordance (W) was used to measure the agreement in rankings arnong male farmers and among female farmers, and the correlation between male and female farmers' rankings. High and significant correlation values indicate cIose agreement on the ranking of the rice genotypes by men and women in the sample. Tables 5a to 5d show that in the two villages, male and female evaluators were in cIose agreement in the ranking ofthe lines. The Ws were highly significant, revealing that farmers' and breeders' rankings are ofien acceptable. Table 6 shows the surnmary of the ranking of male farmers, female farmers, and plant breeders indicating their choices. Ofthe 14 and 16 varieties ranked in Basalatpur and Mungeshpur, PVS 1, PVS3, PVS7, PVS9, PVSlO, and PVSI5 carne out as the farmers' and breeders' choices in 1999. The traits of these lines are shown on table 7. During the crop season in 2000, several ofthese lines were compared with local check through PVS. Twenty-three farmers in two villages in Faizabad grew three rice lines, while 50 farmers in six villages in Siddathnagar grew six rice lines obtained from PVS trials. Table 5a. Summary Ranking of Rice Genotypes in Basalatpur, Siddathnagar District, 1999 M.tes(5) Field 1 Femates(5) Breeders (3) No. Unes Ave. Score R.nk Ave.score R.nk Ave. score Rank PVSl NDR-40032 2.4 3 2.6 2 3.0 2 PVS2 Kamini 8.4 8 8.8 6 11.3 12 PVS3 NDR-9730004 5.8 5 7.0 5 4.0 3 PVS4 Bindili 6.4 6 8.8 6 10.3 11 PVSS NDR-9830103 10.6 10 13.2 11 9.3 10 PVS6 Sugandha 10.8 10 7.0 5 12.0 13 PVS7 NDR-96005 6.8 7 7.6 7 6.3 5 PVS8 4113 14.0 11 12.4 10 14.0 14 PVS9 NDR-9730015 3.0 2 1.8 1 5.3 4 2.0 1 4.0 3 2.0 1 PVS10 NDR-9730020 PVSll Malasia 9.6 9 5.2 4 8.7 9 PVS12 RAU-1308-10-11-3·1·2-4-3 8.6 11 7.4 5 6.7 6 PVS13 CN-l03S·61 4.8 4 10.0 9 8.0 8 PVS14 RAU-1411·10 10.4 10 9.2 8 7.0 7 W=.73·· ··Significant at 0.5 and .10 per cent leve!. 186 W=.63** w=.70'" T.R. París el al. Table 5b. Summary Ranking of Rice Genotypes in Basalatpur, Siddatbnagar District, 1999 Males (5) Fleld2 Br.eder. (3) F.mal •• (5) Ave, score Ave. Rank score Rank seor. Rank 2.2 2 3.8 3 3.3 4 10 Ave. No. Lines PVSl NDR·40032 PVS2 Kamini 7.2 6 7.8 7 10.7 PVS3 NDR·9730004 8.2 1 5.4 5 2.7 2 4 11.7 11 PVS4 Blndili 5.6 2.6 2 PVS5 NDR·9830103 8.0 7 9.2 8 9.7 9 PVS6 Sugandha 604 5 6.2 5 9.3 8 PVS7 NOR·96005 4.6 3 6.4 5 5.3 5 pvsa 4113 11.0 9 12.2 10 13.3 12 PVS9 NDR·9730015 1.8 1 1.8 1 1.3 PVS10 NOR·9730020 2.4 2 5.0 4 3.0 3 PVS11 Malasia 12.6 10 7.2 6 9.3 8 PVS12 RAU·1308-1Q.ll·3-1·2....3 13.6 11 12.2 10 7.7 7 PVS13 CN·1035-61 8.6 8 12.2 10 6.0 6 12.8 10 11.0 w=.72!t'A 9 11.7 11 PVS14 RAU·1411·10 W=.90 u w=31*" **Significantat 0.5 and .10 percent level. Table Se. Summary Ranking ofRice Genotypes in Mungesbpur, Faizabad District, 1999 Mal•• (S) Field 1 No. Avescores Unes Breadars (3) Fem,le. (5\ Rank Ave scores Rank Ave scores Rank PVSl NDR40032 3.2 3 2.6 2 1.7 1 PVS2 Kamíni 15.8 16 15.2 14 15.3 16 PVS3 NDR·9730004 6.6 6 6.0 4 3.0 2 PVS4 NDR·9730003 10.4 13 7.2 7 3.7 3 PVS5 RAU-1308·9-3-1·10-3-4-3 8.4 8 9.0 8 13.0 13 13 PVS6 PSRM·1-1648-1 13.8 15 14.8 13 14.0 PVS7 NDR·9830102 2.9 1 1.8 1 5.7 5 PVS8 NDR-9730002 9.2 10 12.6 . 10 7.0 PVS9 NDR-!l730015 8.0 7 6.6 5 5.0 8 4 PVS10 NDR-!l730020 5.4 4 7.0 6 6.0 6 10.6 10 12 PVSll Mashuri 6.6 5 9 9.7 PVS12 RAlJ..1308-10·11-3-1 .... 3 10.2 11 13.0 11 12.0 PVS13 NDR-96012 9.0 9 8.8 8 8.0 9 10.0 11 , PVS14 RAU-1411-10 lOA 12 6.0 4 PVS15 NDR·9830103 3.0 2 3.4 3 6.7 7 PVS16 RAU·1400-13-20Q.4-6 14.0 14 13.2 12 13.3 140 w"71- w=.81** w=O,79*" **Sjgnificant at 0.5 and .10 per cent level. 187 Listening lo Farmers' PercePlions Ihrf)ugh Participalory Rice Varietal Seleclíon Table 5d. Summary Ranking of Rice Genolypes in Mungeshpur, Faizabad District, 1999 Mal•• (S) Fieid 2 Ave scores B;~d~;;'(4) Female. (5) Ave seores Ave scores No Unes PVS1 NDR-40032 4.2 3 3.4 3 2.3 1 PVS2 Kamini 11.4 12 14,4 14 14,7 11 Rank Rank= Rank PVS3 NDR-973004 8.0 7 6.2 4 4,7 2 PVS4 NDR-973003 8.6 9 8.0 8 8,0 6 PVS5 RAU-1308-9-3-1-10-3-4-3 14 12.0 12 14.3 10 10 PVS6 PSRM-I-16-48-1 12.8 13 11.8 11 12.3 8 PVS7 NDR-9830102 3.6 2 2.4 2 7.0 5 PVS8 NDR-9730002 8.0 7 10.0 9 8,7 7 PVS9 NDR-9730015 5.6 5 6.4 5 5.0 2 PVS10 NDR-9730020 5.2 4 7,0 6 6.0 4 PVS11 Mashun 10.6 10 13.6 13 7,0 4 PVS12 RAU-1308-10-11-3-1-4-3 8 10.2 la 12.7 9 9 PVS13 NDR-96012 10,8 11 72 7 9.3 7 PVS14 RAU-141HO 7,0 6 10,0 9 9,0 7 PVS15 NDR-9830103 1.6 1 1.4 1 5.3 3 PVS16 RAU-1400-13-20 15,0 15 10,0 9 9.7 w-z.6S ...• 6 w=.60 u W=.65"''' "Significanl al 0.5 and .10 per cenl leve!. Table 6. Summary Ranking ofPreferred Lines by Male and Female Farmers and Plant Breeders, 1999 Mala farmer. Female farmers Plan! breeders Field 1 Fleld 2 Field 1 Field 2 Field 1 Field 2 PVSl 3 2 2 3 2 4 PVS3 5 7 5 5 3 3 PVS7 7 3 7 5 5 5 PVS9 2 1 1 4 PVS10 1 2 3 4 1 PVSl 3 3 2 3 PVS3 6 7 4 4 PVS7 1 2 2 5 5 5 4 2 Basalatpur 3 Mungeshpur 1 2 2 PVS9 8 5 5 PVS10 4 4 6 6 6 4 3 1 7 6 188 T.R. París el al. Table 7. Farmers' Assessment ofNew Rice Lines during tbe 1999 KharifSeason Lines (LocationL. _.::N"'a"m"'e_ _ _ _-'P'-'o"s"'lüv"'·"'"..:t:.:'a"'II:::5_ _ _ _ _ _ _ _ _ _......:.N"'."'9=at"'ív:.:e'-'lra=ít"'s'-_ _ _ __ PVSl Good yíeld Medlum planl height Good straw (quantity and qualityj Has regeneration capaclty (faster recovery after submergence) Short. bold. h ••v:¡ grains Best for puffed rice, has good PYS-3 NOR-973004 Me 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 farmers 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. Annual rajnfall in RaJasthan \ , I chUIIJ I aika~ " J I 350-550 mm Sikar / NaQaut , , ~ • Sawai~ //~Tri;' jI a~ I 1/ ,Jodhput Pia" / ~riee of lndle dep!cted in this l1Ul9 a/'é l'I#li1tJer (;(ItNid l\Of a~ TM ~mal Figure 1. The state of Rajastban in the northwest ofIndia 194 The extemaI boutldario!$ ol~ ~ in ~ mep are neMher t.tlI'f'tId: nor~. 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 "varieties," 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 eharaeteristies 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 environmental 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 interactions 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 ofprofessional 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 background; 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, including 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 developed. 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 separated from the straw, although sorne farrners prefer to select for panicles in the field before harvesting, 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, selecting for desired plant types among outcrosses: One or more new plant types will become dominant, 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 characteristics, 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 conditions. 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 traditional 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 diversíty of socíoeconomic conditíons: farmers who grow traditionallandraces with or without selection; 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-management 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 c1imatic factors. Other important factors seem to be the size ofthe landholding (market-oriented or subsistence-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 betwcen 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 assocíated with grain yield in the stress envíronments and negatively associated in the non-stress environments. 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 Environments Mild terminal drought Favorable MAN97 J0091 Graio weighl 0.69*· 0.75" 0.42** . Panicle girth 0.70** 0.83" 0.42" Lcafwidlh 0.38- Slcm diameler 0.62" 0.69- 0.33" 0.41·· No. of panicles -0.54" Tíllers -O.54*" -0.46" -0.58- Nodal lillering -0.65" Plan! type diversity -0.57" Traits PAL97 Early drought MAN98 J0098 0.08 -0.60- -0.25' -0.62- -0.24' -0.65" 0.90- -0.14 0.01 -0.41- 0.67- 0.36- 0.56" 0.27* -0,36'" 0.32'- 0.11 -0.41 -0.24' 0.48" 'p < .OS. "p<.OL 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 different 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-management 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 supported 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 material 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 ofpotential 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" landraces (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 improvement. 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). However, direct observation and interview data from this study have revealed that this view does not apply 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 locallandraee 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 control over the negatíve effects of introgression. Fanners 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 exotic 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 rejected, 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 potential 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 farmers' 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 lmprovement. 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 offarmers 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 01choice: Makíng the most ofnew 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 fiDdiDgS 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 referenee 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 particular ecosystems. The role of farmers in creating agrobiodiversity is also evident from their involvement 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 enhaneement 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! important factors that motivated fanners to diversifY crop and Iivestock in the past were probably ensuring Iivelihood and meeting qualitative preferences and requirements (Roder 1995; Sperling and Berkowitz 1994; Sthapit and Joshi 1998). Roder (1995), has reiterated the faetors motivating farmers 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 strengthening 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 conservation, 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 mechanism 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 mainlaíning theír genelic diversity through Ihe partnership of farmers wíth formal science institutes 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 lntemational 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 publíe and private institutíons, has played an important role in bringing erops and varieties to their present 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 farmers to make the story successful. Stages of deveJopmant Process Farmers' role Wlld relativas Ni ~ rl u! RI OomesticatKm H Farmers' varAR A: LI sEL E ~ Exploratlon, coUection, testing Experimentation Oif!usion ~-t{~~~~=~ ' Experimentatlon Conv, breeding :--- - - - - - - - - - - , , H)lbridiZation. Improvement, : for local adaptation and vbhty Expeñmentation. testing for adaptation, verification, "tedion, nomendahJre ~ exchang$, purchase, 9ift, _ing, theft mass, and Pl setectlon Pareru selection. need and praference, Sék!ctlon, regeneration Extension (E il1$t.fdevl NGO} , Adaptation. diffusíon e I-~~P:::C:'(~P~PBIP~VS/P::.":~G=E)~ T~ I ~ 1-l-L~N"".w:v~a~",,~ty,-_~ 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 ¡dentifYing 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) 2 Bhadaíya basmati 3 BaSInatJ' 4 Lal tengar · Aghani Saro=bhadaiya type Slightly 8hadalya= .arly seasoned, rango~líke color; fine grained 8as;:aroma Like B, basmati: long panicle length; fine graio; 8as=aroma aroma; awn on a few grajns · Reddish UP color; boid graín INith long stout awn; grown rn shallow water Lal=red, Tenga,. type 01 fish with stout spur 5 Amaghauj 6 Dudhraj 7 La/ka fararo 8 Harinker 9 Parewa pankh 10 Kariya keroadh · Ama=guava, Ghauj=cluster · Oudh=milk Whitish UP color; m¡¡ky-white grain Lal=red, Yellowish UP color with minute reddlsh stripes · · · · Yeflowish grain; two to fOUf grains originating from a single point giving c'uster~like look; long and strong stalk Faram= research ¡nstitutian Harin=spotted deer Par.we~ plgeon, Pankh=feather Karlya=black i IJP during milking and dough stage lookslike .pot- I ted deer; small round grajo. ~terile lemme l. long, ooverlng tIle graio Irom both sides fine grein; bladaIQ. nakhí sara, mansa,., Bloticlablotle stress ReSOOice l'lCh SocloeconomlesJ H -,.,.. No basmati cuJtivated Bhltthi, amagMuj, sakhat, faJ te,,~r (1)rought fTóm n.lghboring dlstrtctS) : aJl modm1 v,atietias I I I I I I 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 organizations 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 farmers' 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 conserved 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 Farmer-Ied PPB External agents support larmers' own system 01 crop development Forrnal-Ied PPB Farmers joio in lonmally initialed process 01 crop development McGuire, Maoiead, and Sperling (1999) 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 landraces, 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 perfonned 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 selection 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 different 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 russemination (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 requirements 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 p, Chaudhary. S,P. Khatiwada. and K.D, Joshí !ll!!J&llm> I1I>o Kothi .. Gunny bag • Woodoo structore • Metal structure .. Gtain .. Panicle ~ Cob .. Fruit .. Pod ~ .. aome • Mud structure .. Purchase • Malerial exchaoge • B:siá ~~- lO ' Philippines Nauru . ....'. .0 .- .. .. '" Sotomon ...... \ ~ • Islands ' Tuvaiu .. Tokelau Samo,a.", V.anuatu·~ , ''! t ".. .. Fiii '1" T!.>. rtlt~. C::hhotanagpur Platea" of Eastern India • long cobs with high placement on the stem • prolificacy (two to tmee ears per plant) • resistance to lodging, disease, and insect pests • nonhybrid varieties because ofthe cost and difficulties ofpurchasing hybrid seed every year Breeding strategies To develop new varieties of malze, two strategies were adopted: • breeding open-pollinated composite varicties • breeding intervarietal hybrids Composite breeding To breed new open-pollinated composite varieties a base population was initiated in the main season of 1997 by making nine crosses between three yellow-endospermed flint varieties (BM 1, Suwan, and Chandan 3) and three white-endospermed flint varieties (GDRM 187 from Gujarat AgriculturaJ University and KRlBP west, Gujarat [see Goyal, Joshi, and Witcombe, this volume]; Shweta, from Uttar Pradesh; and Rudarpur local, from Uttar Pradesh). The parental varieties were either farmer-preferred varieties or had complementary traits. The three yellow varieties are medium- to late-rnaturing and have a higher yield potential when water 1S not limiting than the three earlier-maturing white varieties. By 1999, the population had been randomly rnated for five cycles by using a pseudo-random hill planting plan. In each cycle, 50% ofthe plants were detasseled, and paJe yellow grains were harvested from lhe detasseled plants (Goyal, Joshi, and Witcombe, this volurne). At lhe C 3 and C4 cyc1es, two open-pollinated (C3/98-99 and C4/99) varieties were extracted frorn the base population. The yellow-grained variety C3/98-99 was formed from deep yellow seed harvested from about 200 early-maturing, detasseled plants of lhe C3 cycle of lhe base population in the post-rainy season of 1998-99. In the rainy season of 1999, lhe random-mating population was grown frorn these seeds. In the post-rainy season of I 999--{)0, farmers were invited lo visit the research station at Ranchi lo select desirable plants. Farmers graded them inlo three categories and the third preferences were rejected. In the rainy season of2000, lhe seIection will be repeated and the population will be tested in research station triaIs. Similar procedures were followed for variety C4/99. A white-endospermed population was also developed by bulking white grains frorn three sources: frorn selected p1ants ofthe C 4 cycle ofthe base population, frorn the C3/98-99 population grown in the rainy season of 1999, and from plants selected for rnaking the C4/99 population. The first random rnating will be carried out in the rainy season of2000, and farmers will be involved in selection before and after flowering. Intervarietal hybrid and composite breeding Intervarietal hybrids offer a faster approach to creating new varieties for farmers than generating new cornposites, but they require more complex seed rnultiplication than open-pollinated varieties. However, the seed of intervarietal hybrids is cheaper and sornewhat easier to produce than that of 230 A. Kumar el. al. single-cross hybrids based on inbred lines. If intervarietal hybrids were greatly preferred, then KRIBP would attempt to produce seed within project villages, In singIe-cross hybrids, advanced generations from farrner-saved seed are considerably lower yielding tban tbe original F I generation. However, tbe advanced generations of intervarietal hybrids may still yield welL How much hybrid vigor is los! if farmers retain the seed ofhybrids is being evaluated in tbe rainy season of 2000 by using advanced open-pollinated generations from tbe F I intervarietal hybrid, Sorne farrners have preferred tbe open-pollinated varieties Suwan, BM 1, and Chandan 3 to their local varieties. These varieties, along witb Megha, a drought-tolerant and early-maturing variety from Punjab, were used as parents to produce three intervarietal hybrids in the rainy season of 1997: BM I x Suwan, BM I x Chandan 3, and Megha x Suwan. Evaluation The new open-pollínated varieties developed have not yet been tested for yield on farmers' fields. However, intervarietal hybrids were tested in farmer-managed participatory-research (F AMPAR) trials in !he rainy season of 1998 as well as in research-station trials in tbe pre-rainy season of 1998-99 and the rainy season of 1999. Ofthe three hybrids tested, BM 1 x Suwan yielded tbe mos! in trials conducted in !he pre-rainy season of 1998-99 (rabIe 1). The advantage of!he intervarietal-hybrid approach is clear: the hybrid yields more !han either parent and is earlier tban tbe later, highest-yielding parent (Suwan). F AMPAR trials in tbe rainy season of 1998 showed tbe following: Table 1. Performance of Three Intervarietal Hybrids of Maize on the BAU-KRIBP Research Farm, Rancbi, Bihar, during the Post-Raiuy Season of 1997-98 (Summer 1998) 50% silking (d) 50% tasseling (d) Maturity (d) Plant height (cm) Earlength (cm) Yieldl plant(g) Suwsn x Megha 102 110 148 151 17 125 8M 1 x Suwan 99 108 147 136 17 145 BM 1 x Chandan 3 94 98 139 135 16 115 Suwan 104 111 159 146 16 105 8M 1 95 99 139 127 14 100 Megha 93 96 135 137 13 93 Chandan 3 96 101 144 129 125 GDRM 187 86 93 132 116 17 13 Hybrid 88 • Farmers preferred BMI x Suwan and Chandan x Suwan because oftbeir yellow flint grains, higher yield, medium maturity, and higher fodder yield. • Hybrid Megha x Suwan was rejected because of a high proportion ofpoorly developed and diseased plants. 231 Participat()ry Plant Breeding in}{aize [or (he Chhotanagpur Platear¡ oC Eas(ern India The two farmer-preferred hybrids were further tesled in the raíny season of 1999 al the BAuKRlBP farm (figure 1). Hybrid BM l x Suwan, the highesl-yielding entry, yielded more Ihan both parents and was earlier Ihan the highest-yielding parent. Chandan 3 x Suwan was earlier than eilher parent, and although it yielded less than Suwan, the advantage in earliness of 19 days would mean Iha! farmers would prefer il to Suwan. Both intervarietal hybrids had cob placement equivalent lo ¡he best parenl, Suwan (high cob placement protecls from jackal damage). 205 105 5 t;' 188 83 209 109 211 107 146 49 128 47 Planl height (cm) Cob height (cm) 4 '§ '- ~ 3 CI> 's. .!: 2 [! ~ 1 O BM 1 xSuwan BM 1 Suwan, Ch. 3 x $uw. Chandán 3 Local Figure 1. Performance of two intervarietal hybrids in the rainy season oC 1999 at Ihe BAU-KRIBP Research Farm, Ranchi, Bihar (Days lo maturity are indicated on !he bars.) The intervarietal hybrids themselves are certainly promising. What now needs to be done is to look at the feasibility of their seed production-.either with village-based organizations in the development project area or, on a more commerciallevel, wíth the public- or prívate-sector organizations. How important this will be will depend, in par!, on how well the íntervarietal hybríds compete with open-pollinated varieties derived from the composite. References Goyal, S.N., A. losbi, and l.R. Witcombe. Participatory crop improvement in ma¡ze in Gujarat, India. This volume. 232 Participatory Plant Breeding in Rice in Eastern India se Raví Kumar, D.N. Síngh, Prasad, J.s. Gangwar, D.5. Virk, and JR. Wilcol1lbe Abstract Tbis paper describes a partícipatory plant-breedíng (PPB) program for rice in castem India, targeted at a range of ecosystems. Only a few crosses with large populatíons were used. Mas! resources were used on selcetion by farmers among F. progeny bulks grown on a research station, but farmers also gelected within F. bulk populations on lheir fields. Farmers (both male and female) from villages served by lhe KRIBHCO Indo-British Rainfed Farming Prejec! made selections among bulk progenies derived from a cross between Kalinga III and IR64 that were grown at Ranchi, Bih.r, in the main (rainy) season of 1998. The tbree mosl selected progeny bulks were promoted to an AH India Co-ordinated Rice Improvement Proje.1 varielal trial in Ihe main season of 1999. Oue of Ihese was Ashoka 200F, me result of selection withín on F4 bulk in a farmer'. fi.ld. The performance of all tbree varieties in the innial varietal trial al Birs. Agricultural University (BAU), Ranchi, in lhe main season of 1999 was good. Ashoka 200F yielded 2.90 1 ha' compared 10 1.95 t ha" for KaJinga m. AII Ihree cutries were as .arly and slender-grained as Kalinga IlI, and all were more resist.nlto lodging.ln lhe main season of2000, mese varietÍes will be tested in a participatory varietal seleclÍon program in si" project villages. The approach of using a low-cross-number, large-population breeding strategy with both consultative .nd coll.boralÍve partícipation has rapidly ímproved K.ling. 1Il, lhe mas! wide!y adopted upland rice in India. Among other strategies, we have used modified-bulk populatíon breeding lo provide heterogeneous .nd homozygous bulks to farmers for solectíon. Introduction Bíhar, a typical eastem Indian state, has 5,4 million ha planted to rice, with yields of, on average, onIy 1.2 t ha- 1 of grain. More than half ofthe rice area is rainfed, inc1uding the drought-prone, upland ecosystem. In tbis ecosystem, most farmers grow traditional varieties and productivity is very low. Most farmers prefer to grow traditional varieties. Many of the varieties bred and released by the formal system, both nationalIy and at the state level, have no! been adopted by farmers because they lack traits important to farmers (Virk and Bhasker Raj 1996). However, variety Kalinga I1I, which was promoted by the project in its target area covering mne districts ofBihar, West Bengal, and Orissa, has severa! advantages-excellent grain quality and extreme earliness, which allows it to escape end-of-season droughts. However, because it has weak straw, a major objective of the participatory plant-breeding (PPB) program was to breed varieties to replace Kalinga I1I that díd not have this wealrness. Breeding strategies The breeding strategy was to cross a popular, locally adapted cultivar (in this case, Kalínga ill) with exotic, high-yielding cultivars from a centralized breeding program (Witcombe et al. 1996). Varieties IR64 and IR36 were chosen as the high-yielding cultivars since both are grown in large areas in eastem India. A strategy of a few crosses with large populations was used (Witcombe and Virk, in press). Raví Kumar and D.N. Singh are wíth the Department ofPlant Breedíng and Genetics, Birsa Agricultural University (BAU), Ranchi. S.C. ?rasad and J.S. Gangwar are with the KRIBHCO Indo-British Raínfed Farming ?rojee! (KRlBP), Kanke Road, Ranchí. D.S. Virk and J.R. Witl:ombe are wlth the Centre for Arid Zone Studies (CAZS), Gniversity of Wales, Bangor, UK. 233 At the request ofthe Centre for Arid Zone Studies (CAZS), crosses were made at the Intemational Rice Research Institute (IRRI), in the Philippines. Because only a few crosses were used, more resources could be devoted to each cross, so large population sizes and many progeny rows were employed in lhe breeding programo A large F 2 population was raised al Ranchi in lhe main season of 1997 and the F3 was grown in the off-season at the Central Rice Research Instítute (CRRI), Cuttack, in 1997-98. In the main season of 1998, the crop was grown at lhe collaboratíve research farm of Birsa Agricultural University (BAU) and the KRIBHCO Indo-British Raínfed Farmíng Projeet (KRIBP), Ranehi. Each year since then, two crops have been grown. We describe lhe breeding strategy for the Kalinga 1II x IR64 cross. Two participatory methods were used; they varied according to the main type of particípation employed, Le., consultative or collaboratíve (Joshi and Witcombe 1998). Consultative breedíng In the consultatíve approach, breeders grew all of the tríals on a research farm wíth moderate applícatíon ofpurchIehru Krushí Vishva Vidhyalaya (JNKVV), Madhy. Pr.desh. Selection was done in an appropriate envrronment: low-fertility fields under management typical of local farmers. The traits selected for were those identified by farmers. In sorne of the later generations, furmers were invited to carry out mass selection in Ihe populations. Early in the breeding program, farmers were given Ihe composite to evaluate in Iheir fields, and as soon as varietíes were produced' from the composite, they were included in PVS trials. Several white-endosperm and yellow-endosperm varieties were produced from the composite by selection for grain color afier randorn mating was completed. Three white-endosperm varieties, GDRM 185, GDRM 186,andGDRM 187, were tested in formal trialsandon farmers' fields in participatory trials. GDRM 187 was bred as an extra-early variety ofrnaize. Extra-early varieties, such as Chandan Safed 2, can play an important role in the farming system, particularly for growing in rows Ihat alternate with olher crops. Chandan Safed 2 had been appreciated by farmers in Ihe participatory trials, particularly for intercropping with pigeonpea, as Ihe maize could be harvested before it had a significant competitive effect upon the pigeonpea crop. Moreover, early varieties can escape end-of-season drought and produce a harvest al Ihe time when grain is se arce, thereby fetching a high price. It was assumed Ihat Chandan Safed 2 could be improved because it was a direct introduction from South America. GDRM 187 was bred from al) six parents (table 1), but in Ihe third generation of random mating, selection was made for plants that had Chandan Safed 2 as a maternal grandparent, and these lines were backcrossed to Chandan Safed 2. 238 SN. Goya/. A. Joshi. and JR. Witcombe Results The white-grained maize varieties were tested by Gujarat Agricultural University (GAU) in the system of state trials that is used to identify varieties for release. The mean performance of these entries was superior for grain yield by 7%-29% (depending on the variety) in a multi-year, multilocational trial to lhat ofGujarat Makka I (GM 1), a variety that out-yields the most widely grown locallandrace by 10% for grain. They also silked two to six days earlier than GM l. In lhe rainy season (kharif) 1997, GDRM 185 and GDRM 187 were tested in farmer-managed participatory-research (F AMP AR) trials in Itawa and Bihar (Madhya Pradesh); Smjumi, Bar, and Katarani Palli (Gujarat); and Khundini Rupa, Mathura Khali, and Kunda (Rajasthan). Focus-group discussions showed that both varieties were much preferred by farmers over the local varieties. Both were perceived to be earlier than the local varieties (GDRM 187 particularly so) and to be higher yielding. GDRM 187 was reported lO have much better grain quality than the most widely grown local variety, and GDRM 185 was reported to have somewhat better grain. Both varieties were reported to have fewer plants lhat failed to produce cobs, more plants with two cobs, larger cobs, and, unlike the local varieties, cobs that were filled to the tipo In kharif1998 GDRM 187 was tested in three villages in Gujarat and one in Rajasthan (figure 1). Yield increases in farmers' fields were higher in percentage terms lhan those found in higher yielding research station trials. Overall, GDRM 187 was the variety most liked by farmers. Like GDRM 185 and GDRM 186, it yielded more than the local varieties, but it had the added advantage ofbeing significantly earlier to mature. In addition, farmers cornmented that its cobs were tightly and completely enclosed by the husk, reducing insect attacks, and they al so cornmented on lhe superior quality of its grain. In crease (%) over local varieties 2.0 íii' 31 12 18 32 1.5 § -~ :! 1.0 0.5 0.0 ·,i cJ''0. 0'1> <,<:;.'<'1> <¿l' Village Figure 1. The performance ofGDRM 187 in FAMPAR trials compared to the local varieties in four villages: Sarjumi, Gujarat (20 trials); Umariya, Gujarat (5 trials); Bansripada, Rajasthan (6 trials); and Bar, Gujarat (15 trials) (The advantage of GDRM 187, averaged over 46 trials, was 28%.) 239 Parlicipatory CrofJ Improvemenl in Maize in Gujaral, I"dlo Discussion PVS does nol a/ways work hUi it builds (he foundation of a sound PPB program The PVS in maize was nol very successful. Perhaps Ihis is nol surprising since mosl oflhe varieties that were tesled were not bred in Ihe target environment. Gujaral Makka 1, the only cultivar tha! was bred in the area, was selected from the locallandrace "Farm Sameri." Although this selection was successful in producing a statislically higher-yielding variety in tríals, the difference was insufficient. On farmers' fields ¡Is 10% yield advantage was not noticed by farmers. Even a PVS program that does not identífY highly successful varieties is of use. In Ihis case, it enabled target traits to be idenlified-for example, the preference for whíte grsín and extreme earliness. Mosl important, it a1Iowed the ídentification of parental genotypes. Was this a participatory plant-breeding program? Biggs (1989) classified participation inlo four types, two of which are collaborative and consultative participation. Collaborative participation by farmers mass seleeting in the populalions in their own fields was attempted, bul failed. It was difficult for farmers to preven! cross-pollination of the composite with locally grown material by, for example, planting the erop in an isolated plor. Farmers were reluctant lo mass seleet by uprooting undesirable plants because ofthe IOS5 ofyield this would entaiL The allemative of delasseling undesirable plants and rejecting them al harvest lime would be possible but difficult However, the breeding program did involve consultative participation-farmers were involved in the identification ofparental material and target traits, and in the evaluations of the breeding generations on the research farm. Of majar importance was the . decentralization ofthe breeding programo Although selection was not in farmers' fields,.it was in the largel geogrsphícal area. The composite was grown under lower input levels Ihan normaIly found in a research station and cIoser to the levels used by farmers. The breeding program also had innovative aspects Ihat were not related directly lo farmer participation. Wide crosses were made between yellow- and white-endosperm maize with reselection for white. Quite elaborate designs during Ihe random mating of the composile were employed: hill planting and detasseling was done to increase the pollination between progeny of the original nine crosses and reduce sibbing within them. In the random-mating generations, grains with pale yellow were selecled to advance Ihe riext generalion. Thís color is the mos! probable phenotype ofheterozygotes and selecting for it maximized the possibility of recombination around the locus controlling grain color. It is not possible to know which component was most important in the success oflhe program--<:ollaborative selection of parents by PVS, consultative PPB, decentralization lO the target environment, or innovative breeding techníque5. However, part ofthe breeding philosophy in PPB, argued by Witcombe et aL (1996), is the need lo concentrate on ouly a few crosses or populations, which allowed the required resources for the novel leehníques used in Ihe breeding programo Was PPB cost effective? Conventional breeding had never produced a cultivar thal was preferred by even a significant minority oflhe farmers in the projecl area, In about five years, PPB had produced al least one cultivar, GDRM 187, that was liked by most farmers for most oftheir fields. It yielded significantly more grain (about 15%-30% more) even though it was significantly earlier to flower (about one week earlier). Ihis combination ofhigher yield and earlier flowering is extremely valuable for farmers 240 _ _ _ _ _ _ _ _ _ _ _.::.:sly Goval. A. Joshí, ond .IR. Witcombe and ís nonnally a difficult combínatíon to achíeve in any maize-breedíng programo GDRM 187 al so had other advantages, including improved grain qualíty, that should íncrease íts speed of adoption and íts adoption ceiling. AH this was achieved with modest resourres, since only a single composite was created and only a few varíeties were derived from ir. References Biggs, S.D. 1989. Resource-poor farmer participation in research: A synthesis ofe:xper;encesfrom nine natíonal agricultural research systems. OFeOR Comparativo Study Paper No. 3. The Hague: International Service for National Agricultural Researeh. loshi, A. and 1.R. Witeombe. 1996. Farmer participatory crop improvement. srudy in India. Experimental Agriculture 32:461-477. n: Participatory varietal seleetion, a case loshi, A. and 1.R. Witcombe. 1998. Farmer participatory approaches lor varietal improvement. In Seeds of choice: Moking the mast ofnew variet;es for smallfarmers, edited by J.R. Witcombe, D.S. Virk, and J. Farrington. New Delhi: Oxford IBH; London: Intermediate Technology Publications. Witcombe, J.R., A. Joshi, K.O. loshi, andB.R. Sthapit. 1996. Farmerparticipatory crop improvement.1: Varietal selection and breeding methods and lheir impact on biodiversiry. Experimental Agnculture 32:445-460. 241 Towards a Practical Participatory Plant-Breeding Strategy in Predominantly Self-Pollinated Crops J.R. Witcombe, M. Subedi, and K.D. Joshi Abstract There is a limil to the eapacity of any breeding program, and lhe more crosses tha! are made. lhe ,maller the ,izo of each eross. The theory of the optimum numbcr of erosscs in inbrceding crops is brictly reviewed. The !heory is unsatisfactory in detennining the optimum number of crosses, but models lhat take linkage inlo aecounl show Ihat very large populations are needed to recover specified genotypes. Henee, one possible strategy is lo seleel a small number of erosseS lhal are .ansidered favorable and produce large populations from them. This stralegy is idealIy suiled lo Ihe particular ccnstrainls and advantages ofparticipatory plant breeding (PPB). When a breeding program is based on few erosse., lhe choice of parents is crucial ;md [aoner participalory melhods are highly effeclive in narrowing !he choice, Modified bulk-population breeding methods are desirable slralegies in lhe participalory planl breeding of self-pollinaling eraps when combined wilh • low-eross-number .pproach, and a p.rtieipatory breeding program for rice in Nepal ís described. Introduction In most, perhaps all, conventional breeding programs for inbred crops on research stations, breeders deal with many crosses each season. Even with fairly Iimited resources many hundreds, or even thousands, ofF4 or Fslines can be tested. Unless there is considerable researcher input into the layout oftrials in farmers' fields, participatory plant breeding (PPB) has to employ many fewer crosses and entries than conventional or classical breeding. In farmer-designed, farmer- managed trials, each farmer usually grows only one entry (e.g., Joshi and Wítcombe 1996) and the number of particípating farmers thus límits the number of entries. However, a very large population of any entry can be grown, with líttle or no cost, or even wíth a benefit. In PPB, a farmer replaces hís or her cultivar with a population for PPB on land that would normally have been devoted to the crop. The cost ofthis replacement is any decrease in value ofthe harvest caused by the replacement and the benefit ís any increase in harvest value, In contrast, in classícal breeding all the costs of any increase in the area of the cultivated crop are borne by the breeding program. We briefly review the theoretical evidence on the number oY crosses that are required in a breeding progral11- We describe a rice breeding program in Nepal that is using a low-cross-number, hígh-population-size strategy. Theory on the number of crosses in a breeding program The optimum number of crosses required in an inbreeding crop was reviewed by Witcombe and Virk (forthcoming) and only a summary is presented here. To calculate the optimum number of crosses, crucial assumptious are required on the rate ofthe inevitable decline in the potential value of each cross as more and more crosses are made. If the decline is very significant (e.g., a few J. R. Wilcambe 1$ al tlte Centre for Arid Zoo. Studies, University ofWales, Bangar, Gwynedd, UK. M. Subedi and K. D. Jo,hi are witlt Local Initialives for Biodiversity Resea:rch and Develapment (LI-BIRD), Pokhara, Nepal. Thi, document is an OUtpUI from project R7 I22, funded by tlte Plan, Scieru::es Research Prograrnme nf Ihe UK Department for International Development for the benefit of developing counnies:. The views expressed are not necessarily those of tite DFID and L1-BIRD. The .utltors are graleful for ,he help ofal! Ihe participating farmers in Chitwan and thank tlte LI-BIRD staffwho helped in conducting the research described in this papeL 243 Towards a PracÜcal Participa/O/y Plam-Breeding Slrategy in Predominantly Sel(-Pollinated eraps crosses can be identified as having a higher probability of giving desirable segregants than others), then only a few crosses are needed. Ifthe decline cannot be predicled, then many are required. The lack of quantitative data 10 support assumplions on the rate of decline limits Ihe role oftheory in deciding the optimum number of crosses. However, 10 recover specified genotypes, large population sizes are needed that, given a limit to ¡he overall sizc of any breeding program, willlimit the number of crosses. Whether a high-cross-number or a low-cross-number approach should be used depends greatly on the judgment of thc rescarcher as to whether the value of crosses can be predicted with any certainty. In a decentralized breeding program, the target envíronment and the requíred traíts in a finíshed variety are known, and the knowledge of exísting adapted germplasm is considerable. This allows such predietions to be made, so a low-cross-number strategy appears sensible. Many fewer eros ses than are cornmon in most breeding programs will be used, and for all of them there will be logical reasons as to why ¡he cross should have a hígh probabílíty of producing favorable segregants. There will be many fewer crosses than cornmonly suggested from theoretícal calculations that ínvariably assume there is no prior ínformatíon on the value of illIy cross, Le., that al! cr.osses are considered to have an equal chance of success (Yonezawa and Yamagata 1978; Wricke and Weber 1986; Hueho 1996). In a large-cross-number strategy, population sizes are likely to be limited to a few hundred rather than several thousand. In a low-cross-number strategy, population sizes ean be larger and increase the probability that desirable segregants tha! are an improvement over the best parent are recovered. AH that is needed is that the two parents differ significantly for an important trait (a practical certainty) at sorne point in the genome. A segregant tha! has a genome substitution from the other parent al this point will be superior, providing the sum ofthe rest ofthe genome is equal to the best parent The existence of a eros s that canno! give rise to superior segregants is theoretically impossible, although the population size required to recover desirable segregants may be impracticably large. However, ehoosing complementary parents increases the likelihood Ihat there will be a sufficiently high frequency of desirable segregants for thern to be selected. Towards a practical participatory breeding strategy PPB is ideally suited to the strategy ofrigorously selecting parents, using a small number of crosses and employing large populatiQns. Participatory varietal selection (PVS) is the first step in selecting desirable parents. It allows local and introduced germplasm to be evaIuated using participatory approaches; it identifies candidate varieties having suitable traits and determines their acceptability to farmers. A PPB program in an inbreeding crop can start on the basis of one cross or very few crosses. Even with a low-cross-number strategy, the number of crosses eovered will gradually increase over time if one, or a few, new crosses are made each year. This will help to maintain the farmers' interests by a supply of novel germplasm and allows a continuing incorporation of new genetic material from more centraIized breeding programs. Pedigree breeding generates a large number oflines (the selection units) tha! can only be accornmodated with difficulty in a PPB program. The mosl effective rnethods keep the number of selection units to a mínimum, thus allowing one, or an acceptably low number, of selection units per farmer. However, large population sizes can be used because the marginal costs to the program of increasing population size are very low (figure 1). Hence, bulk-population breeding is ideal for PPB, in either its pure form or modified by dividing the population into sub-populations according to 244 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _~JR lVüevmbf. M. Subed;. and K.D. .fusili Figure 1. A very large population grown by a farmer, Cbitwan, October 1998. Tbe only possible cost to tbe farmer is tbat tbere might be a reduction in tbe yield of the F. bulk of Kalinga III X IR64 (right) compared to Masuli (left). farmer-important traíts. Bulk-population approaches have been used with success in classicaJ breeding, e.g., Carver and Bruns (1993) report that 30% ofwheat releas es from a breeding program resulted from bu1k population breeding that took less than 8% of the resources. We are conducting a PPB program in rice, targeted at a [ange of environments in Nepal. These vary from the Terai (alluvíal, low-altítude, fiat land in the southern part ofNepal at about 150 m altitude) in both the main season (sown in June) and the chaite season (sown in February). The breedingprogram is also targeting a range of irrigaled environments up to 1500 m altitude. Only a few crosses have been made during lhe course of this breeding program, which commenced in 1996 wíth two crosses made by lhe International Rice Research Institute (lRRI) at the request of the project and one cross made al the Center for Arid Zone Sturues (CAZS), Bangor, by Dr. D.S. Vírk. AH three crosses ínvolved the upland rice variety Kahnga III as one of the parents. Kalinga II1 was identified in western India in a PVS program (Joshi and Witcombe, 1996). Fanners like it for Íls very short duration and, an unusual traíl for an upland rice variety, íts slender grains. Although it is an upland rice variety adapted to marginal conditions, it is widely adapted even though it was rejected from AH-India Co-ordinated Crop Improvement Program multilocational tríals. It was released for raínfed, drought-prone, cold-susceptible environments only in Orissa, on the basis of trials in that state, but is now widely grown in Bihar, West Bengal, Madhya Pradesh, Rajasthan, and Gujarat. In PVS trials, it performs extremely well as a chaíle rice in the Nepal Temí under partially irrigated condítions and can be grown as a main-season rice in the low hills ofNepal up to 1000 m under rainfed conditions. 245 T())vards a Practical Participa ton! Plant-Breeding Strate[V in Predominantlv Selt-Pollillated Crops One of the erosses made at IRRI was Kalinga III x IR64. IR64 is a longer duration, high-yielding variety adapted to irrigated conditions. At one time it was the most widely grown rice genotype in the world and has occupied the majority ofthe rice-growing area in the Philippines and Indonesia. It has also been released in India forTamil Nadu but is widely grown in other states as wel!, e.g., West Bengal and Haryana. It has wide adaptability, multiple pest and disease resistance, and slender, translucent grains. It is clear that in this low-cross-number strategy, an enormous amount of information is available on the parents. Kalinga III has weak straw and a low yield potential. IR64 has complementary traits: a very high yield potential and it is highly resistant to lodging. It can also contribute pest and disease resistance to Kalinga III, even though this variety has few susceptibilities. In targeting highyield-potential environments, Kalinga III can eontribute earliness to IR64, and beca use of its extreme genetie divergenee from IR64, it is reasonable to expect transgressive segregation for yield in these environments. For the chaite season and somewhat higher-altitude, high-yield environments, Kalinga III contributes cold tolerance. In PPB, an essential part ofthe strategy of selecting appropriate parents is that one ofthem is local!y adapted. Kalinga III has been adopted by farmers for partial!y irrigated conditions in the chaite season. However, it is a niche variety and other eros ses have been made involving lhe most popular chaite rice, CH45, and lhe most popular main-season variety, Masuli. In the early stages of creating the bulk populations, the breeding program for the eros s Kalinga III x IR64 was entirely researcher managed on land rented from a farmer. Initial!y 290 F3 lines ofthe cross were grown in the chaite season of 1998. The progeny rows were highly diverse and they were grouped into six bulks based on their height (tal! or dwarf) and maturity classes (early <110 days seed to seed; medium 110-125 days; and late >125 days). The bulks were named as fol!ows: ED = early dwarf; ET = early tal!; MD = medium dwarf; MT = medium tal!; LD = late dwarf and LT = late tal!. Of these, the performance of the early dwarf proved to be unsatisfactory and it was dropped. Ihe performance of the MI bulk was good but highly variable, so it was further divided into four: MIl = earlier shorter; MI2 = earlier taller; MI3 = later shorter; and MI4 = later tal!er (figure 2). After dropping lhe ED bulk and dividing the MI bulk into four, there were eight bulks. Ihree were then advanced without further division (the two dwarfbulks and the late bulk). However, in five of the bulks, further division was made among the Fs seed into grain type, i.e., long, intermediate or short in length. In the F 6 generation, the resultant bulks were grown by researchers and evaluated by farmers (consultative participation). Combinations of maturity and grain types were selected and rejected. For example, in later-maturing bulks that more or less matched the maturity ofCH45, only nonslender types were acceptable. For rice ofthis maturity, the harvest ofwhich coincides with the rains, it is only economic to produce roasted, flattened rice, for which only less-slender-grained varieties are suitable. In contrast, in the earlier groups, al! grain types were acceptable. By the F6 generation the bulks were recognizable by farmers, because most of the plants shared cornmon traits, but the bulks still had significant genetic heterogeneity within them for farmers to be able to make selections. In the chaite season of2000, farmers were given the bulks at the F7 stage and lhe results ofthis farmers' selection wil! be evaluated. As wel! as the modified bulk populations breeding approach, we are al so trying variants of single-seed descent (SSD) such as equal-seed deseen!. In classical breeding programs, SSD is increasingly employed to rapidly and cost-effectively produce homozygous lines. It concentrates 246 IR. Wircombe, M. Subedi, and K.D. Joshi Figure 2. An example of rice suh-hulks at the F s stage in the cross Kalinga III x IR64, Chitwan, Octoher 1999. Note the large population sizes (the people in the background are working in the same bulk as the one seen in the foreground) and the two bulks-earlier-maturing bulk MT2 (right) and a later-maturing bulk MT3 (Ieft). seleclion in advaneed generations that are highly homozygous and where selection is more efrcetive thanin earlier, more heterozygous generations (e,g., Delzer, Busch, and Hare1and [1995] and Van Oeveren [1992] in wheat; Fahim et al, [1998J in rice), We have modified SSD to retain even more variation by using equal- rather than single-seed deseent in the earlier selfing generations, It also aUows multiplieation so lhat by the F 5 or F 6 generation, large quantities of seed of each bulk (or sub-bulk) can be supplied lo many fanners, The probability of se1ecling desirable segregants is increased when the entire selection process is replicated across fanners, Participatory techniques mus! complement and caunot replace classical breeding. Somc low-heritabilíty traÍts can only be selected under controlled environments, and modero techniques that facilitate wide crossing, such as embryo reseue, are confined to the laboratory. No single participatory plant breedíng program can hope to screen more than an ínsignificant proportíon ofthe germplasm available in collections of genetic reSOllrces or, for example, attempt to create poplllations with novel resistance traits. Classical breeding is a strategic approach that creates improved parents for Ihe cost-effectíve, adaptive approach of participatory plant breeding. References Carver, B.F. and R.F. Bruns. 1993. Emergence of altemoti ve breeding methods for allogamous crops. In Proceeding$: Vo/I. Tenth Australian Planl Breeding Conferenee, Gold Coasl, 18-23 Apri11993, edited by B.e. ¡mne and J.B. Hacker. Canberr.: Australian Conventíon and Travel Servíces. 247 TOrvards a Practica! Parlicipatorv Plant-Breeding Slrare!{V in Predominuntlr Sdf~Púll¡naled Crop:i. ______ Delzer, B.W., R.H. Busch, .nd G.A. H.rel.nd. 1995. Recurrent seleetion Crop Sciel1ce 35:730·735. fOf grain protein in hard red spring wheat. Fahim, M., M.P. Dhanapala, D. Senadhira, and M.J. Lawrence. 1998. Quantitative genetics ofrice. H: A eompallson of the effieieney of four breeding methods. Field Crops Researe/¡ 55:275·266. Huchn, M. 1996. Optimum number of eros,es .nd progeny per eross in breeding self·fertilising erops. \: General approaeh and fiest numerieal results. Euphytica 91:365·374. Joshi, A. and J.R. Wi!eombe. 1996. Farmer participatory erap improvement. ll: partieipatory varietal seleetíon, a case study in India. Experimental Agrieulture 32:461-477. Van Oeveren, AJ. 1992. A comparison between single seed deseen! .nd early eross seleetion in wheatbreeding.lI: The evaluative generations. Euplzytica 64:91·97. Witeombe, J.R. and D.S, Virk. (Forthcoming). Number of erosses and population SIze for participatory and elassical plant breeding. Euphytica. Wricke, G.•nd W.E. Weber. 1986. Quantitative genetics and se/eclinn il! plan! breedíng. Berlin·New York: W. de Gruyter. Yonezawa, K. and H. Yamagata. 1978. On the number and size of eross combinations in a breeding program of self·fertilizing crops. Euphytíca 27: 113-116. 248 Participatory Crop Improvement for Intercropped Maize on Bar; Land Terraces with Trees T. P. Tíwari. Daljit S. Virk, and F ergus L. Sinclair Abstract Maize (Zea mays L.) is the most importan! crop io the middle hills ofNepal and is mostly grown in association with finger millot (Eleusino coracana Gaertn) and fodder trees. Seven maize varieties have been recornmended forthe hills but few farmers have adopted them. As a prerequisite to designing a participatory maize-improvement program for the middle hills Ihat could reverse declining yields, local knowledge .nd practice were investigated and combined with micro-meteorologieal measurements at three sites, eaeh with 20 participatíog farmers. In a participatory maize-improvement program, both participatory varietal selection (PVS) and partieipatory plant breeding (PPB) were camed out side-by-side with vari.ties selected on the basis of etitena derived from farmers' knowledge. Four different vari.ties were tested (Manakamana-I, Arun-l, BA-93-2126#2, Population-22) with local vaneties al e.eh site. Particípatory mals, where eaeh farmor grew a new variety alongside local varietíes, were eombíned with display trials of all tbe v.rieties al five loeations. Questionnaires and focus-group discussions were used lo assess farmers' ev.luatÍon ofvanetíes. Population-22, despite its late maturity, w.s Iiked by f.rmers for disease toleranee, higher yield potential, white and large grains, .nd its slay-green charactenstics. StatistÍcal analysis of grain yield confirmed farmers' preference for Population-22, since this out-yielded the ather new vaneties (p<.OS), which were similar in yield to local vaneties. A seed-multiplication program ofthis preferred vancty has becn initiated by participating farmers. As part of the PPB program, the best four local v.neties (Marga local, Muga local, Madi local, and F.kch.tnara loc.l) were colleeted from vatious parts of the middle hilIs and crossed with adapted exolics (Maoakamana-I, Arun-l, Population-22, and Pool-21). Five compasites have becn created by random mating so as to offer choices to farmers in Ibe coming seasons. thus increasing the genetie diversity they are able to evaluate and utilize, Introduction Maíze (Zea mays L.) is the most important crop grown in association with finger núllet (Eleusine coracana Gaertn) and fodder trees in Nepal. About 80% of maize is grown in the hills, which constitutes 20% ofthe total cereal production of the country with producíÍvity of slightly more than 1.5 t ha- l (CBS 1997). There hasbeen a decline of 20% in maize productivity in the hills since the mid-1970s (palikhe 1996; Adhikari 1998; NMRP 1997). This is proof ofthe inefficiency ofthe traditional approach to maíze improvement. The problem with the present approach is that it has assumed that biophysical and socioeconomic factors are common1y shared. The nature and importance of farmers' knowledge is poorly understood, and farmers' involvement in the research process has not been realized. The complex system of growing maize/millet with trees has been overlooked and fanners have nol been recognlzed as research partners in lhe process of maize T. P. Tlwari is with the Agricultural Research Station, p.khribas, Ohankuta, Nepal; O.ljit S. Virk is with the Centre for Arid lon. Studies, Univérsity ofWales, Bangor, Gwynedd, UK; and Fergus L. Sinel.ir is wíth!he School of Agricultural and Forest Sciences, University ofWales, Bangor, Gwynedd, UK. The authors are grateful to all the participating fanners who contributed to various aspects of this project This researoh was funded by OPIO, PSRP, UK and Nopal Ihrough HARP. Pi.ld research was cartied out al the Agricultural Research Station, Pakhribas. one of the hiH research stations under!he Nepal Agriculture Research CounciL The support oftnese institutions is híghly acknowledged. CommenlS made on me manuscript ofthis paper by Oc K.R. Regmi and Mr. O,B. Subba are also acknowledged. 249 Parlicipatorv Crop lmprovement (oy ¡ntercropped .Maize on Bari Land Terraces with Trees technology generation. Consequently, the ímpact of seven newly released maíze varietíes has been ínsígnificant. It is not surprising lhat most ofthe maize-growing areas in the middle hills ofNepal are covered by the traditional varieties. Either new varielies are no! reachíng a majority offarmers or those farmers with access lO them are not continuing lo use Ihem (Khadka et al. 1993). Pham, Waddington, and Crossa (1989), in lheu review report on the ímpact of germplasm from the Intematíonal Inslitule for Wheat and Maize Improvement (CIMMYT), mentíoned that in most developíng countries, maíze farmers are, by and large, growing the old established landraces. It was Iherefore realízed that it is necessary 10 develop a more efficient and effective approach where researchers, maize breeders, and farmers can work together for a common goal. The need lO bridge Ihe gap between local and scientific knowledge is a necessary prerequisite lo developing an effective maize-improvement strategy. A farmer-participatory approach would involve developing a community-basedadaptive research capacity, achieved by working with groups of farmers, maximizing Ihe use ofrural resources, and utilizing farmers' knowledge in parallel. This approach can bríng farmers' knowledge -(farmers' perspectives) and scienlific ideas (researchers' perspectives) logether (Walker et al. 1997; Sinclair and Walker 1998; Wagner 1993; Joshi 1997). Besides lhe acquisitioo oflocal knowledge, a fuller understandiog oflhe maize-growing environment and how farmers manage scarce resources are major prerequisites for investigation. Farmer participatory maize improvement During the early 1970s, wheo research 00 maize started, Ihere were high expectalions that the development of maize in Nepal would offer better varieties lo farmer •. However it has been realized that Ihe adoption of new varieties by farmers was not as simple as lhe researchers and developmenl workers had thought. The farmer-participatory approach began in response lo lhe inefficient, traditional, top-down approach, where more focus was paid 10 a few researchers' traits of interest ralher Ihan to lhe needs of farmers managing complex and heterogeneous systems. There are many good reasons lo encourage farmers' participation in the process of agricultural research and development (Farrington and Martín 1988; Farrington 1998; Witcombe et al. 1996; Joshi and Witcombe 1996; Witcombe and Joshi 1996; Witcombe and Virk 1997; Sperling and Scheidegger 1995; Sthapit, Joshi, and Witcombe 1996; Subedi, Rana, and Joshí 1997). The complexity of the system i8 only understood by lhe farmers. The traditional approach is deficient both in understanding such systems and in using farmers' talents. The participatory approach will help empower local groups of farmers by enhancing production (through the acceptance of preferred varieties), genetic diversity, and "togethemess" (Sperling and Scheidegger 1995; Eyzaguirre and Iwanaga 1996; Chambers and Mascarenhas 1990). As part offarmer-participatory maize improvement, bolh participatory varietal selection (PVS) and participatory planl breeding (PPB) were carried out side-by-side, although the latter is usually initíated when PVS faíls lo identify farmers' preferred genotypes (Witcombe et al. 1996; Joshi and Witcombe 1996). The two activities were carried out al lhe same time in order lO create broad, genelic-based populations simultaneous1y with PVS aclivities so as to offer choices to Ihe farmers as quickly as possible. 250 _ _ _ _ _~ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _'_'T'_P'_.ccTt=\vari. D.S Virk. and FL Sinclair Materials and methods Participatory varíetal selection Fanners' criteria fOf selecting maize genotypes were based on local knowledge. Suítable varietíes were sough! to meet the importan! rraits that were identified as preferred by farmers, particularly those rela!ing to grain size, color and type, plan! height, suitability for agroecologícal niches, and compatibility with the system~ Varie!ies were selec!ed as suggested by Witcombe et al. (1996) from the releases fOf the same domain (Manakamana-I), for one o!her domain (Arun-l fOf lower hills bu! forthe middle hills a new introduction), and from pre-releases (Population-22 and BA-93-2126#2). Sites were chosen where maize is the important crop fOf household income, in farming systems Ihat were representative in terms of agroenvíronrnental and socioeconomíc conditions, and where ¡here were no political or social obstacles 10 effective researcher-farmer interactions. Marga, Patle and Fakchamara were selected for the farmer-managed, participatory-research (F AMPAR) trials. A total of60 packets (15 of each variety) containing 500 g ofseed were distribuled randomly to 20 farrners at each site to compare with theír local varieties. Therefore, F AMP AR trials of one variety were replicated over five particípating farmers al each site. Farmers were asked lo grow the new varieties alongside their local variety in the same field and under the same management conditions. However, fields for the F AMPAR trials were to be selected mutually by farmers and researchers for their representativeness (not too sloping, not too marginal or too fertile, and with sorne degree of tree shade, if possíble). Periodic farm visits and interactions wíth farmers were made so as to observe performance of varieties at different stages. Assessments of the pre-harvest traits of test varieties were made by joint visits between researchers and farmers to each participating farmer's fields. A wide range of issues, covering field management and performance ofvarieties in the complex and heterogeneous environrnent, were discussed. F arrners' observations of experimental varietíes and their own local varieties were discussed at greater length and were recorded using household·level questionnaires (HLQs). Farmers were asked to harvest both new and local varieties separately and to measure grain yield using their local measurement uruts. They were also requested to store the harvests separately usíng existing practices and to assessievaluate other postharvest characteristics, such as grit-to-flour ratio, graín type and color, cooking quality, taste, and market value. They were also asked to assess fodder quality. AIso as part of the PVS program, demonstration trials were conducted in five different sites (Marga, Patle, Fakchamara, Murtidhunga, and Parewadin). The same four FAMPAR varieties were given to one farmer at each site to grow together with hislher local variety for comparison; 500 g of seed of each variety was given to mrmers to grow on their own farms. Group visits were organized to see FAMP AR trials in the field grown by indívidual farmers in various growing conditíons. The performance ofthe FAMPAR varieties was assessedjointly, and final!y, farmers were brought to see varietal demonstrations to compare al! varieties at one site. At the end of the session, focus-group discussions were organized and views were collected as per questionnaires developed for the discussions. Male and female farrners were grouped separately and discussions were initiated accordingly. Based on the performance ofF AMP AR varieties, farmers were asked to rank the varieties. 251 Participatorv Crop Improvement [or lntercropped Maize on Sari Land Terraces with Trees Participatory plant breeding The germplasm that farmers felt was best adapted to the eastem middle hills of Nepal was co~ lected. 1 Before collection, the fields where these varieties were grown were visited and their performance was assessed. Individual growers and local farmers were consulted in order to identify the best-adapted local germplasm. The varieties Muga local, Madi local, Fakchamara local, and Marga local from local varieties and Manakamana-l, Population-22, and Arun-l (white) and Pool-2l (yellow) from the. improved varieties were used in the PPB programo A total of five composites were prepared with different crossing combinations offarmer-preferred varieties (table 1). Sowing time was staggered according to the maturity class ofthe variety so as to synchronize flowering. A purposive randomization was followed to equalize the chances of random mating. Three seeds per hill were sown and later thinned to one. At the vegetative stage, individual plants were tagged to detassel later. Diseased and other abnormal plants were removed as soon as the tassel appeared. These composites were sown on-station with irrigation, because the previous year there had been asevere drought (the longest in 35 years). Seed priming was practiced for early establishment. Final selection was done by farmers from tagged and detasseled plants. Laboratory selection was done for flinted and white grains, rejecting yellow, dented, and diseased grains. Farmers prefer white-grained maize varieties; however, sorne yellow-grained types possess desirable traits. A novel PPB prograrn was followed to exploit yellow-grained types in composite breeding, where pale yellow grains are used for further cycles of random mating (Goyal, Joshi, and Witcombe, this volume). Pale yellow grains represent a cross between white- and yellow-grained varieties. All other grain types, being more likely to be parental types, are rejected. Results and discussion Unlike PPB, which requires a long phase ofbreeding before its products can be tested in FAMPAR trials, PVS provides a means for immediate identification offarmer-preferred varieties. The products ofPPB are not yet ready for such an evaluation but the results of the PVS program are available and are presented in this paper. Participatory varietal selection Group interviews were conducted to compare all experimental varieties with local varieties at the end ofthe growing season. The objective and expected outcome ofthe project was reviewed once again as a reminder to the group members, since they are the ultimate users. Groups of farmers visited each other' s fields to see all the FAMPAR varieties grown by different farmers, which may be under different management but were grown under similar growing conditions (with respect to altitude and system). Each trait they mentioned was recorded. Most of the traits were compared against local varieties; however, lhe overall ranking ofthe acceptability of the varieties was made among test entries, including local varieties. Farmers' perceptions of major pre- and postharvest traits are summarized in figures la and lb. It was noted lhat most ofthe farmers could not make confident contributions regarding cooking quality and taste. Assessment of problems with pest in the stored grain is continuing because it has not yet been taken out of the thangkro (maize crib). 1. 252 Descriptions based on farmers' knowledge about and experience with varieties were prepared but not included in tbis report. Table 1. Farmer-Participatory Maize Breeding Plan Parent selection: WeU~adapted landraces wfth good phenotypic trans were salee'eo by farmers. Locally adapted new vedalles identified from Ihe sean:h program as 5uggesled by Witcombe el al. (1996), A.White composite Composite 1 PopulaUon-22 Muga local F/mara local Second year Thlrd ye., Fourth Sowing time adjusted according lo malulity class, Upgrading continued. FAMPAR lrials, Co-coordinated multilocatíonal trials (CVTs). Formal on-farm trials, 50% plants detasseled ano ssleellan to allow aqual chanca for random concentrated only fmm those. maUng at Rampur during \Yintar. Other techniques not changao, Upgrad¡ng continued at station. Good malze growers ¡nvited for fi&ld selectíon. FAMPAR tríals, Seed ¡n crease in farmers' fíelds. dítto ditto dillo ditto Best ears from the respective farrn-. ers' fie!ds selecled by farmers and collected from three different sitas. Seed of different V3rieties :m1xed and sown as composita. Good maize growers invitad to select in tield. Bulk sowlng, Random mating as other composites. FAMPAR trials, Upgr.ding continued, FAMPAR trials, Co-coordlnated multilocational tlials. Seed ¡ncrease in farmers' Broad-based com09site Purposlve randomization done lo al low equal chanca for random mating. w Good malze growers invitad durlng First 3 composites grown together to prepare broad-based composite, Sown by purposive randomization the fleld selection, Lab selection for heatthy, whUe-ffinted grains. Composite 2 DIsease nurseries. Seed increase in farmers' fialds. ditto Manakamana-1 Marga local. Flmara local Composite 3 ditto Arun~l, Madi local Marga local Composjte 4¡, dltto all abova tour locals Composite 5 FAMPAR varietíes yoa, Fi'st yaar. Season·1 (Moren sowing. hill) Season·2 (Septamber. toral) 50% oí plants detasseled for random mating. ditto Table 1. Farmer-Participatory Maize Breeding Plan (continued) B. Composile breeding using yellow Iypes Pool-21 yellow (Female) Female parent sown in alternate with other varieties. Arun-1. Madi local AII female and undesirable and diseased plants from male rows detasseled. Manakamana-1 Population-22 rQWS Good maize growers invited for tield selection. Pale yellow seeds sown at Rampur. 50% plants detasseled. Fleld selection. Lab selection; only white seeds selected to continue. Other colors discarded. Upgrading continued. White seed continued by random mating. FAMPAR trials. Good maize growers invited for plant selection. Co-eoordinated multiloeational trials. FAMPAR trials. Seed inerease in farmers' fields. I ditto Deep yellow and white seeds and other diseased grains discarded. - Only pale yellow seeds (being hybrids) from yellow female parent continued. Note: Two years on-fann testing to satisfY yariety release cornmittee is to be condueted befare proposing yariety to be released. * Adapted local gennplasm from yarious parts ofthe middle hills should be continuously collected, eyaluated, and eombined in local composite so as to make broad genetic base which could be used for future crossing prograrn,s. Recurrent selection of these composite3 (randomly mated) should be continued _________________.___________.:.r"-p~...I1_'_i,::.v::.ar:.:i, D.S. Virk, and FL Sine/a ir . .-c=. . Dlsease, Man-1 --~ . .---.. Arun-1 BA·S3 Pop-22 Lodgíng, Man·l Arun-1 BA-93 Pop-22 Drought Iol"",nl, Man·l Arun-l BA·93 Pop-22 Maturity. Man-1 Arun-1 BA·93 Pop-22 Stay green, Man-.1 Arun-1 BA·ro Pop-22 Man~ 1 Arun·l BA·93 Pop-22 lST. liking, Ear size. Man-1 Aruo-l BA-ro Pop-22 Ear Compact, Man-1 Arun-l BA·93 Pop-22 Shade tolerant, Man-1 Arun·l BA·93 Pop-22 Stoveryield, Man-1 Arun·l BA-93 Pop-22 Productlon, M.o·l Arun·l BA·93 Pop-22 J ~____:=_d§:::__~~ ·100 ·so Worse than local o 50 100 8etter than local Figure la. Farmen' perceptions of pre-harvest traits of four new maize varieties (Farmers' perceptions as to whetber tbe test varieties were better or worse tban tbe local varieties are indicated by lines. Tbe sborter the line, tbe more similar tbe variety is to the local varieties.) 255 Participatory Croplmprovement for Inte!crC!pped Ñfaize ~ Grajo size, Mana-1 011 B..ªE! LOlld Terraces with Trees .. Arun-1 BA-93 Pop-22 - Grain color, Mana-1 Arun-1 BA·93 Po".22 *Taste, Mana. 1 Aruo-1 BA-93 POD~22 "1: ·Cookíng quaUty, Mana-1 Arun-1 BA·93 Pop-22 ~Market value. Mana-1 Arun-1 BA-93 Pop-22 -SO Worse than local Betler than local "'Results based on single site. Figure lb. Fanners' perceptions on postharvest traits of four new maize varieties (Farmers' perceptíons as to whether the test varieties were better or worse tban tbe local varieties are indicated by lines. The sborter tba line, tbe more similar the variety is lO Ibe local varietíes.) Preharvest traits. Farmers observed that the germinative ability of the new varieties was better than thal ofthe local varieties since better quality seed was given lo thero. During the group discussion, one of the participating farmers said that if the quality of seed of the new varieties was as poor as the usual inferior (ínsect attacked) quality of the the local varieties, Ihen the germiuation percent of the local varieties would be higher under stressed sítuations (drought and soil capping). Additionally, the farmers said that when the maize was sown, Ihere was sufficient moisture in the soil, and as a result, there were no germination problems this year. Farmers a1so perceived that the new varieties had stronger stems and shorter pIant height than the local varieties, resulting in reduced lodging. The test entries were better with respect to foliar diseases, particulatly turcicum blight, but they had problems with ear rol. Within the new varietíes, Population-22 was preferred. This was mainly because it had larger ears and lower rates of infection with turcicum blighl. Farmers thoughl this was because it was less affected by tree shade. New varieties roatured earlier than local varieties except for Population-22 (figure la). The new varietíes had similar requirements for fertilizer and water as the local varieties; however, theír drought tolerance was less. There were míxed responses from fanners on ear size, production estimates, shelling percent, and graín size, Despite the desirable thinner stems of the local varietíes for livestock 256 TP. Tiwari, D,S. Virk, and FL Sine/air stover, farrners preferred the new varieties for this purpose because Ihey had improved stay-green characteristics. Except for Population-22, the new varieties were not shade-tolerant. Postharvest traits. Assessment of postharvest trai!s revealed tha! the local varieties were better with respect lo grain color and type, taste, grit-to-flour ratio, stored-grain pest infestations, and cooking quality. The farrners who were able to comment on taste reported that Manakamana-l was good but still inferior to the local varieties. The taste ofPopulation-22 was inferior to local varieties and to Manakamana-I. However, these varieties all fetch good market prices compared with yellow types (figure lb). The overall ranking of the tested varieties trom different sites with different groups of farrners reveaIed that despite its Iateness, farmers liked Population-22 in field conditions (table 2). The traits farmers Iiked were higher yielding potentiaI, taller plants with multiple ears, stay-green characteristics, freedom trom foliar diseases, and tolerance to lodging, Because of the taller plant height, there was less shading of millet when tbe Iower leaves are stripped by farrners to harvest fodder and reduce competition with the millet. However, at tbe Murtidhunga and Parewadin sites, farmers saíd it affects mílIet because of its larger leaves and late maturity. Table 2. Overall Rank ofVarieties froID Different Sites with Different Groups ofFarmers (1999) Marga CVs\sites Mana~1 Fakchamara. Palie M/Dhum~. Tankhuwa Mar.. Female Male Female Mixed MI...d Mlxed Mlxed 2 5 3 2 5 4 3 3 1 1 Pop-22 1 BA-93 4 3 5 5 4 3 5 2 Aron-l 2 3 2 2 4 2 2 4 4 4 4 3 3 5 2 5 Locals There was little distinction between tbe preferences of maJe and female farmers. It was rather surprising lbat the late variety Population-22 scored the highest (40), followed by Arun-l (26). The seore ofMana-l, local varieties, and BA-93 was similar (21). This was supported by the observed grain yield trom FAMPAR trials, where Population-22 was found to be significantly superior (p<.05) to local varieties (figure 2). Other entries were on par with local varieties fOf grain yield. As in the F AMP AR trials, Population-22 was found to be the highest yielder in tbe multilocational varietal display trials, although this result was not statistically significant (V=.38). The mean graín yield, irrespective of site, ranged trom 2294 kg ha- I to 2949 kg ha· l . Arun-I was tbe lowest yielder. Most of the farmers who grew Arun-l commented lbat because of its early maturity, birds and rodents were attracted to it. A further problem was tbe theft of ears. Thus, there was no seed to keep for the folIowing year or 10 assess for postharvest traíts. However, because of its earliness and otber desirable traíts, farmers were willing to contÍnue to use it. Sorne farrners also expressed the opinion lbat it provided early food and that demand for it would inerease in the future when green ears were marketed locally for roasting. Impact 01 F AMPAR varienes The impaet of any variety is assessed by looking al the area covered by that variety in a particular Ioeation and how confidently farmers have taken to that variety. Although it is too earIy lo assess 257 Parlicipatory Crop fmprovement [or Interc.,:opped Maite on Bar; Lalld Terraces withJf.e-::es'----_ _ _ _ __ 3000 lSD (5%) : 502.3 2500 ~ al C. 80 households, good irrigation facilities, land suítable for double rice cropping, and good access to agricultural markets, There are more than 53,000 ha of cultivated land in Chitwan and over 64,000 ha in Nawalparasi. Both districts have more than 72,000 ha of main-season rice, About 22% of the land is irrigated in Chitwan and about 36% in Nawalparasi. Fanners grow two or three crops per year. Main-season rice is lhe major crop in June lo October and covers about 1600 ha in the study villages. Rice is followed by lentíls or wheat in lhe winter, followed by maize and chaíte rice in the spring. There is diversity in soil type, irrigation facilities, and production potential. Productivity is generally higher in East and West Chitwan than in Nawalparasí. There are also variations in the fanning systems within clusters, e,g" sorne farmers in Chitwan grow maize and vegetables in the winter instead of wheat and lentils, A survey of 1487 households in Chitwan and Nawalparasi conducted in 1997 showed high diversity in physical and socioeconomic conditions, In the study arca, 23% of farmers were resource rich, 34% were c!assed as having average resources, and 43% were resource poor (Ram et al. 2000). There was wide variatíon in lhe size ofland holdings, access to irrigation, and lhe use of production inputs, which has resu!ted in different cropping pattems: rice-vegetables-maize or ricemaize-vegetables and rice-wheat-maize in East Chitwan and mostly rice-wheat-rice, rice-fallowrice, or rice-lentil-maize in West Chitwan and Nawalparasi. There is wide variation in lhe rice ecosystem, from perennially irrigated land with varying durations of retained standing water, to seasonally irrigated land, to rainfed lowland conditions. The productíon potential is high. Yields of themost cornmonly grown main-season rice variety were measured in fanner-managed particípatory research (FAMPAR) trials in 1997. The average yield ofthe predominant main-season rice cultivar, Masuli, was 4.2 t ha- t (Joshi et al. 1999), Participatory approaches Two approaches-participatory varietal selection (PVS) and informal research and development (IRD}-were used lo provide a choice ofvarieties to farmers in Nepal. In PVS, introduced varieties were tested in intensively evaluated FAMPAR trials using lhe melhods described in Joshi and Witcombe (1996), lRD uses less intensive evaluation and has been proven to be effective for popularizing new varieties by lhe LurnJe Agricultura! Research Centre, Nepal (Joshi and Sthapit 1990). In each cluster of six villages, F AMPAR tria!s were conducted in three and lRD trials in three. In the lRD trials, lhe same ranges of varieties were used but there was no monitoring or participatory evaluatíon during lhe growing season. Instead, fanner's perceptions were evaluated after harvest by informal interviews wilh a sample of farmers, Data were collected on subsequent adoption and fanner-to-fanner seed dissemination. There were 536 F AMPAR and 5461RD trials from 1997 to 1998. Twelve new varieties of main-season rice were fust offered to farmers to experiment with in lhe main season of 1997 and five more varieties were given out in 1998 (table 1). In each village, for each variety a l-kg bag of seed was given to two fanners ineach ofthree weallh categories (see below). Plot sízes varied because of differences in nursery raising practices. Planting melhods. use of manure and fertilizers, and intercultural operations were unchanged. The fanners grew lhe new variety alongside their existing variety, usualIy MasuJi, as a control. CaTe Was taken to avoid any 268 -~-~ .. _------------------- K.o. Joshi ami J R.1JI.íl"pmbe Table 1. Rice Varieties Includad in the Participatory Varíetal Selectíon Program, 1997-1998 (The first 12 varieties were first offered in 1997, the last 4 in 1998.) Q.ountry and year 01 ralease .!'Iame of variety Entry name Parenlage India Nepal IR51672 IR51672 Narendra 80 NOR80 N22/IR36 1986 NR Radha 11 (Indla:Rajshree) TCA8Q..4 Local selection In India 1989 1995 PR Rampur Masull AS781-1 Lalnakanda/IR30 Pant Ohan 4 BG 90-2 IR262/Ramadja 1984 NR Pant Ohan 10 IR9763 IR32/MahshurillR28 1993 NR Talnan 3 mutanVBasmatl 370 1992 NR PNR 381 1997 PR 103 IRS61 IR8/IR127-2-2 1976 NR PR 106 IR665-79 IR8/Peta/Bella Patna 1978 NR IR54/PR106 1993 NR Pusa basmatl-1 Pusa 615 Pusa 1501Karnallocal 1989 NR Swama MTU7029 Vaslsta/Mahshuri 1982 NR Improved Sabarmatl/Ratna 1983 NR IARI 5901-2/IR 8 1993 NR IR 50lPusa 33fIlR 50lPusa 33 1995 PR 111 Pusa 33 Pusa 44 Pusa 834 Pusa 44-33 Sarwatl NR NR Note: NR = not released; PR:: pre-release; (-) infonnation not avaiJable. chance of mixing the new variety wilh Ihe existing farmer' S variety from seed sowing through to post-harvest assessment. The area of Ihe trial plots was measured by researchers, while farmers measured yield in local volumetric units, which were later converted to metric units. A paired !-test was used to test the significance of the difference for yield between Ihe test entry and the existing rice variety, To conduct Ihe trials, each trial site was joint1y identified and demarcated by Ihe participating farIIlers and researchers. There were regular visits by researchers to the trial plots with Ihe participating farmers to see Ihe performance of the variety at different growth stages, A farm walk was orgaruzed in which researchers, participating farmers, and olher interested farmers saw Ihe standing crop in al! or most of Ihe plots when Ihe crop was near to rnaturity. Immediately after each farm walk, a focus-group discussion was held, which included preparing a narrative summary of each rice variety, describing al! its positive and negative traits, and preparing an overall preference ranking of al! Ihe varieties. A post-harvest evaluation ofthe rice varieties was done on Ihe basis offarmers' perceptions two to three monlhs afier the harvest of Ihe crop, This gave the farmers enough time to assess post-harvest traits. A structured questionnaire was used, which included questions on grain quality, market preference, and the farmers' intentions on whether to adopt or reject Ihe variety, Questions were also asked on Ihe distribution of Ihe seed of the variety by fanners to monitor the adoption and spread of Ihe new rice varieties through 1997 to 1999. In 1999, households that receíved seeds in 1997 and 1998 were visited first (purposive sampling) and then new adopters were interviewed based on the distribution list provided by each farmer. 269 Participatory Varíe/al Se/.elion. Food Security and V.a.rietal Diversitv in a High-Potential Productíon System The project mobilized existing farmers' groups in the project villages. These groups had been formed for different purposes, including agriculture, livestockldairy, and water use. Dístribution of the seed of the new varieties was done following discussions with the groups. Participatory well-being ranking was done to identiIY farmers from dífferent resource categories, Through group consensus, an equal numberoffarmers from all tbree well-being categories were selected to participate in the trials. A brief overview of all the varieties included in the trials was gíven to farmers. Varietal diversity in the project area before peI The baseline study showed that varietal díversity was low in chaite rice, wheat, and maize (Rana et al. 2000). In chaile rice, CH 45 covered over 97% ofthe chaite rice area in the project villages. In maíze, varieties Arun 2 and AIUn 4 occupied ahout 70% ofthe area, and Rampur Composite about 30%. In wheat, two varieties, UP 262 (50%) and RR 21 (20%), occupied most ofthe area. For main-season rice, the greatest varietal diversity was in the East Chitwan cluster (ECC) of víllages where 11 different rice varieties were grown by the farmers, ofwhich Masuli and Ekhattar'(a sister line ofSabitri) together occupied two-thirds ofthe rice area (figure 1). 27% Masuli 111 Radha 17 O Janaki • Blhari 8] • Ekhattar Radha4 • Kanchhi Masuli .TW • Sabitrl Figure 1. Area under main-season rice varieties in three study villages of East Chitwan cluster, 1997 (Himali and Chaite 6 occupied an insignificant area and are not shown.) Six differem rice varieties were grown by the furmers in the West Chitwan cluster (WCC) but Masuli alone covered 98% ofthe total rice area (figure 2). The narrow varietal diversity in this cluster could be attributed to a more uniform physical environment as the majority of the area is low lying and retains standing water during most of the rice-growing season. Another reason is that in WCC, in contrast to ECC, few vegetables are grown. Vegetable growing promotes diversity because farmers grow rice varieties of shorter duration than Masuli to allow timely sowing of the vegetable crops. 270 KD. Joshi and J.R. Witcombe 0% 1% 98% liI Radha 7 Masuli • Radha 9 Figure 2. Area under main-season rice varleties in three villages of West Chitwan Cluster, 1997 (Sabitri, Kanchhi Masuli, and Radha 4 occupied an insignlficant area and are not shown.) The varietal díversíty at the Nawalparasí cluster (NPC) is closer to WCC than to ECC. The main differences are that in N awalparasi there is more Masuli and Sabítri and no Ekhattar at al! (figure 3). 63% 1% 12% o Masul; 1% O Radha 4 O Radha 9 !! Radha 17 • • Sabitri Janak! Figure 3. Area under main-season rice varieties in tbree villages oC Nawalparasi cluster, 1997 (Kanchbi Masuli and Radha 7 occupled an insigníficant area and are not shown.) 271 Participatory Varíe/al Selection, Foad Securíty ond Varíetal Diversity in a High-PotentlOl Production System Varietal dynamics The distribution ofvarieties over time is dynamic, as new varieties are adopted and old and obsolete varieties are dropped, How dynamic the system is can be quantified by measuring temporal diversíty. A dynarnic sítuatíon ís found not only in high-potential systems with modem varieties, it also occurs in marginal areas and even for Iandraces (Joshi and Witcombe, this volume). As a result of the introduction of new varieties by PVS, most farmers indicated that the new varieties they were adopting would replace Masuli. Other varieties also likely lo be replaced were Kanchhi Masuli, Radha 4 (also known as Chaurasi or Bammorcha) and Sabitri. Twenty varieties were listed as likely to be replaced, but 16 ofthem accounted for only 18% ofthe total varietal replacement indicated by farmers. On-farm varietal diversity The introduction of new modern varieties contributed lo an increase in on-farm varietal diversity when diversity is measured simply as the number ofvarieties grown in each village (figure 4). o BElfore PCI in 1997 • Afiar PCI jn 1999 8 4 o Study vitlages Figure 4. Varietal diversity io rice before aod after a participatory crop improvement program across all nioe FAMPAR villages, 1997 lo 1999 Grain yield Four varieties (Swarna, PNR 381, Pant 10, and PR103) had a statistically significant yield advantage over the farmers' existing varieties (table 2). From a few kilograms of seeds in 1997, these four variel:tes covered 22 ha by 1999 in the FAMPAR víllages, which contributed 65 t of additional yield. A further 25 ha were occupied by four other new varieties: Rampur Masulí, Sarwati, IR51672, and Pusa 44. On average, these yielded 7% more than existing varieties (p <.05 in a pooled analysis). The added yield from these varietíes was about 9 t. A similar or higher amount can 272 K.D. Joshí and iR. Witcombe Table 2. Yields ofNew Main-Season Rice Varieties Compared to Existing Varieties in Participatory Varietal Trials in Eight FAMPAR Villages, Main Season, 1999 Graln Yleld of riCé varieties (t ha-1 J Difference relative lo Masuli Variety name New Exisling Yield ('Yo) Swama 4.40 3.35 31.0 m PNR 381 4.04 3.45 Pan! 10 4.37 3.95 PR 103 4.45 4.17 Other new varieties I Maturity (da)!s¡ Area covered (%) by 1999 +5 5.2 17.0' -30 2.1 13.5' -25 0.7 3.86 15.3" -18 0.8 3.80 7.0' 3.0 • p<.05 . .. p<.Ol. u. p<.OOl. : Mean ofRampur Masull.IR51672, Sarwati .nd Pusa44. be expected for the IRD víllages that were found to have higher fanner-to-fanner spread of new varieties than the FAMPAR ones. The monitoring of varietal adoption and spread done in 1999 confirmed that Swama, Rampur Masuli, PNR 381, Pant 10, PR 103, and Sarwati covered significant areas, although other varieties, such as IR 51672, Radha 11, PR 106, and NDR 80, were also adopted to some extent. Discussion The existing varietal diversity in main-season rice was low in general and very low in the West Chitwan cluster. The differences between clusters reflected their physical and agronomic diversity. Because the dominant crop varieties grown by the fanners in the villages of the study area were 30 to 35 years old, fanners were not benefitting from several decades of progress in plant breeding, and because ofnarrow varietal diversity, these systems may be more vulnerable to pests and disease attacks, which contribute to instability in food production. The participatory varietal selection program was successful in thÍs high-potential production systemo Fanners identified and adopted seven new rice varieties from the 16 given in PVS, Some of these, such as Swarna, PNR 381, PR 103, and Pant 10, had a distinct yield advantage over the varieties fanners were currently growing. Others were preferred for therr early maturity, lower water and nutrient requirements, or berter grain quality. New varieties were adapted to specific niches. For example Swarna is suitable for fields where the water stands for nearly all ofthe growing season; Pant 10, PNR 381, and Sarwati are suited to conditions ofpartial irrigation and medium fertility; and PR 103 and PR 106 were adopted for more fertile, higher yíelding environments. Radha 11 was found to be suitable for late planting conditions and for transplanting when the seedlings are more than one and one-half months old. This is an important trait for areas where rice transplanting is dependent on unpredictable monsoon rains. Varietal diversÍty can be quantified but such quantification is scale sensitive. Diversity estimated overall the FAMPAR villages as one unit gíves differentresults Ihan ifit's estimated on the basis of clusters. The varietal diversity in the WCC increased far more than in the other two clusters, whích 273 ParticipatoD' Varietal Seleclion. Food Security and Varietal Díversily in a High-Potential Production System both had higher inirial varietal diversity. From the viewpoinl of diversity deployment to enhance food security, increasing diversity in the most vulnerable arcas is not only important for the communities in those areas, but it also reduces the vulnerabílity of the system as a whole. The PVS approach in main-season rice has helped enhance varietal diversity on-farm in the same way that it has for other crops and areas (11alhi et aL, this volume; Virk el aL, this volume; Witcombe 1999a, 1999b; Joshi et aL 1997). Participatory varietal selection was effective in increasing production in HPPSs by matching agroecological niches 10 the most appropriate varieties. Such increases in production are essential if the deve10ping world is 10 feed its rapidly growing populations. References AMDD. 1995. Food Balance Sheet olNepal (l994//995). Lalitpur, Nepal: Agricultural Marketing Develópment Division. loshi, A. and J.R. Witcombe. 1996. Fanner participatory .rop improvement:Il: Fanner participatory varietal selection in India. Experimental Agricalture 32:461-477. losm, K.D. and B.R. Sthapit. 1990.lnlonnal research and development (IRD): A new approach lO research and extensían. LARC Discussion Paper 1990/4. Pokhara, Nepal: Lumle Agricultural Research Centre. loshi, K.D. and J.R. Witcombe. Impact of participatory plant breeding on ¡andrace diversity: A case study for high-a1titude rice in NepaL TMs va/ame losm, K.D., R.B. Rana, D.s. Virk,.nd J.R. Witcombe. 1999. Evaluatíng the relevanee 01particípalory variety selection approaches lar offering varietal choice on main-season rice in high poten/ial productian systems 01 Chitwan and Nawalparasi district 01 Nepal (1997-/998). LI-BIRO Technical Paper No 111999. POkhara, Nopal: Local Inítiatives for Bíodiversity Research and Development. loshi, K.D., M. Subedi, R.B. Rana, K.B. Kaday.t, .nd B.R. Sthapi!. 1997, Enhancing on-fann varietal diversity through participatory varietal seleclÍon: A case study for chaite rice in Nepal. Experimental Agriculture 33:335--344. Malhi, S.S" J,R. Witcombe, D,S. Virk, and K.B, 5ingh. Participatoty varietal selection in rice in the Punjab. This volume. Rana, R.B., K.D. Joshi, D.8. Virk, D. Harris, andJ.R. Witcombe, 2000. Baseli"e stady report on participatorycrop improvemenl project in Chitwan and Nawalparasi dístricts olNepal. CAZS Discussion Paper 4. Bangor, Wales: Centre for Arid Zone Studies, University ofWales. Virk, D.S., D. Harris, B.S. Raghuwanshí, A.O,B, Raj, P.S. Sodhi, and J.R, Witcombe. A holistic approach to participatory crop improvement in wheat. This volume. Witcombe, J ,R. 1999•. Do fanner-participatory methods apply more to high potential areas than marginal ones? Outlook on Agriculture 28:43-49, Wilcombe,l.R. 1999b. Does planl breeding lead lo a 1055 of genetic diversity? In Agrobiodiversity: Characterization, utilization and management, edited by D. Wond and J.M. Lenné. Oxon, UK: CAB Internationa!. 274 A Holistic Approach to Participatory Crop Improvement in Wheat D.S. Virk, D. Harris, B.S. Raghuwanshi, A.O.B. Raj, P.s. Sodhi, and JR. Witcombe Abstract The term "participatory erap improvement" is used to cover all aspects of erep improvement where farmers are involved in a participatery role. In this paper, we describe the approach aud results for participalory crep improvemenl in wheat, in Lunawada subdistrict, Gujaral, India. Nine villages were selected for study, and within eaeh village, farmers were ranked into categories by wealth. An initial baseline survey on farming praetices was conducted by semistruetured inlerviews on a sample of rarmers. Selected farmers from eaeh weallh calegory kepl weeldy f3rm calendars of all operations in their wheal fields. Selecled fields were termed "intensive data plols." They provided a basis for analyzing Ihe farming syslem and profitability by weallh calegory, as well as for identifying conslrainls. The baseline surveys revealed lha! upper-calegory farmers benefitted most from the sale of wheal produce. The lower-category farmers consumed a large par! oftheir produce. Intensive dala plots showed Ibal upper-category farmers accrued higher nel gains from wheat cultivation Ihan Ihe lower-calegory farmers. Participatory varietal selection (PVS) offered new varieties to farmefs for selection. PVS resulted in significanl replaceroent ofthe old variety Lok 1, grown in abou! 90% oflhe area, by many varietíes Iha! increased yield levels and on-farm biodiversity. Resource-poor farmers benefitted as much as Ibe better-offfarmers from PVS activities. Participating farmers experimented on a simple, cheap agronomic inlervenlian: seed priming. Mosl farmers intended to adopt it beeause of ils multiple beneficia! effeels, incJuding íncreased yields. This holistic approach to participatory methods was effeclive in analyzing poverty issues, idenlilYing constraints and new opportunities, and moniloring impact. Introduction Fanners in high-potential production systems (HPPS) of the Indian subcontínent adopted modero Green-Revolution cultivars in the 1960s and 1970s. Indigenous cultivars were quickly replaced with CIMMYT wheat and IRRI rice varieties. Rates of adoption of modero varieties have since slowed. For example, in India the average age of cultivated varieties is between lOto 27 years for most cultivated crops (Virk, Packwood, and Witcombe 1997). SIow tumover mtes of cultívars mean that fanners are growing older, and therefore inferior, genetic material. The extent of adoption of new varieties by farmers depends on multiple factors, including agronomic and socioeconomic constraints. We used a holistic approach to partícipatory crop improvement in wheat in the Lunawada subdistrict in Gujarat, India, to anaIyze constraints, provide new opportunities, and monitor the adoption of new cultivars chosen by farmers. D.S. Virk, D. Harris, and J.R. Witcómbe are al the Centre for Arid Zone Studies, University ofWales Bangor, Owyoedd, UK. s.s. Raghuwanshi .nd P.s. Sodhi are with lbe Grarnm Vikas Trust, Dahod, Gujarat, India. A.O.B. Raí loe.red in Hyderabad, AP, India. This d .. 700 600 500 o!! ~ "<.,: - .!l! 40 30 Q. . lO -., <: ~ 20 lO ~ < 10 O 1-15 May 16-31 May 1-15 Ju"" 16-30 June 1-15 July Figure L Rice area transplanted from 1 May to 15 July in Punjab and the project area ofthe Patiala district (The averages are over four years (1996-1999), based, for Punjab, on a sample of 1076 farmers in 11 districts in 1996 and 1997, and 855 farmers in 1998 and 1999; in Patiala, based on a sample oC 105 farmers in 1996 and 1997, and 100 farmers in 1998 and 1999.) 284 SS ,Va/hí, J.R. Witcombe, D,S, Vírk, and KB. Singk Why farmers praclice early transplanting contrary lo extension recornrnendations is an interesting question, Participatory rural appraisals (PRAs) done with farmers reveal sorne ofthe reasons farmers transplant late: • the availability of tube-well irrigation and a cheap, fia! rate for electricity • the continued employrnent oflabor afier the wheat harvest • the limited choice of early-rnaturing varieties, since high-yielding cultivars tend to have longer maturation periods and need earlier transplanting Early transplanting ofrice has led to multiple problems such as the foIlowing: • a loweríng of the water table from greater exploitation of ground-water resources (During May and June, the water requirements for crops are at their peak. The early transplanted crop requíres 20% to 30% more water [PAU 1996].) • the loss of nutríents frem evaporation in the extremely hot months, resulting in increased use of chemicals aríd degradation of the environment • an inerease in diseases and inseet pests • less opportunity for green manuring Specific varietal adoption patterns Old varietíes are cultívated on a large area. P AU has reeornmended a number of varieties of rice; however, farmers still prefer to grow old varieties, The varietal surveys conducted by P AU's sernor extension specialist (farm management) showed that 36% orthe area in the stale duríng 1999 was occupied by varieties released 15 years ago, e,g" PR 106, IR8, Jaya, PR 103, and Govind (Singh 1999), Weighted average age of varíeties is high. The average age of varieties, weighted by the area grown lO them in the Punjab, was 12 years in 1996, 11 years in 1997, and lO years in 1998 and 1999. This average is veryclose to the 12 years reported by Witcombe el al. (1988) forthe whole ofIndia, More recently, farmers have replaced their varieties more rapid1y, but the average age remains higher lhan what could be expected of an agriculturally advanced state, Varíeties of wheat and barley grown in the UK in 1999 had an average age of only five years (analysis of data from the Nationallnstitute of Agricultura! Botany by A.G, Bhasker Raj,personal communicatíon), Nonrecommended varieties occupY large areas. Despití! many reeornmendations by PAU, there is sigrúfieant adoption of nonrecornmended varieties in the state. In faet, the area under non-PAU varieties increased in 1998 and 1999 (table 1), In Patiala, the adoption of nonrecornmended varieties was higher than in the Punjab as a whole (average of 53% over rour years), Among nonrecornmended varieties, Pusa 44 has the highest adoption, Itoecupiednearly 50% ofthe areain the Patiala distriet in 1996 to 1999, Pusa 44 is highly susceptible to bacterialleafblight (BLB), and the large-scale cultivation ofPusa 44 has helped to build up the BLB pathogen, which causes losses in other varieties. However, farmers prefer Fusa 44 for its high yield and resistance lo lodging, 285 Participatory Varíe/al Sel,clion in Rice in /he Table l. Pun~ Area ofNonrecommended Varieties in the Punjab and Patiala District from 1996 lo 1999 Area 01 nonrecommended varieties ('lo 01 total rice area) Area of Pus. 44 ('lo of total rice area) Patiala Year Punjab _______' -'-""'-_____ ~== ==_________'P_'u:::n"ja=b~ 1996 31 1997 33 1998 1999 35 38 43 47 60 60 Mean 34 53 Patiala 24 43 28 30 28 28 47 56 54 50 Note.' See figure I for infonnation on sample size5. Methods and materials Participatory approaches Three participatory approaches were used in thís study: l. farmer-managed particípatory-research (F AMP AR) varietal trials, in which farmers grow new varieties alongside their local variety under farmcr management, with evaluation of many cultivar traits by hoth scienlists and furmers 2. informal research and development (IRD), in which furmers evaluate new varietíes with little intervention from scientists; evalualÍon is mainly !Tom the examination of adoption trends 3. single-replicate design (mother trials), with aH varieties grown togelher as demonstration plots to assess the relative performance of varieties (researcher-designed but farmer-managed trials) Selection ollarmers and villages Eleven villages (Kalifewala, Chalaila, Kalwa, Barsat, Bhedpura, Gajjumajra, Kaidopur, Dhengera, Partapgarh, Kartarpur, and Jauramajra) were selected to represent agroclímatic situations in the Patiala district. Three villages(Gajjumajra, Bhedpura, and Barsat) represented salt-affected arcas with soils having a pH hetween 9.0 and 9.5. Ofthese II villages, F AMP AR trials were conducted in six and IRD. in the rest. AH viHages have either metaled or good earthen approach roads. AH of the agriculturalland is irrigated !Tom canals or tube wells. Farmers were selected lo represent small, medíum, and large landholdings. Willingness to experiment with new varieties was tbe key factor in selecting farmers. A total of 497 farmers were involved in participatory research in the kharif(monsoon season) of 1999. Farmer-managed trials Twelve varieties were tested in participatory trials: IR36, IR64, HKR 120, HKR 126, Pant Dhan 4, Pant Dhan 10, Gurjari, Kalinga TIl, Govind, Pusa 834, PR 111, and PR 114. Of these, varieties, PR 111 and PR 114 are recornmended for the Punjab. All other varieties are out-of-state released varieties. Small bags (2-5 kg) of seed (varying according lO the demand of farmers) were given to farmers with the understanding tbat they would grow the new variety alongside their local variety under the same management and that they would participate in the evaluation. 286 S.S Malhi . .IR. Wílcombe. D.S, Virk. and K.B, Sing! The plol area for F AMPAR trials varied from 40-5000 m2 Mos! trials had an area of more Ihan 1000 m 2 under any variety, Sorne farmers, particularly in IRD víllages, pooled ¡he seed 10 grow a larger area. Researchers and farmers jointly evaluated the trials. Frequent farm walks, focus-group discussions, and household-Ievel questionnaires were used for recording farmera' perceptions. Graín yield data were reeorded jointly; researchers measured the plot size and farmers weighed Ihe plol yield. Demonstration plots of all varieties grown in the same field in a single-replicate trial were grown in aH villages as mother trials. Results and discussion Ofthe 12 varieties tested with farmers, Ihree (IR64, IR36, and PR 114) were preferred but IR64 was ¡he most preferred. We shall restrie! the description of trials to IR64 only. Variety IR64 was tested with 43 farmera (26 in FAMP AR villages and 17 in IRD villages) and compared to Pusa 44. F armer trials over several dates 01 sowing The greatest power of participatory trials was experienced in Ihis study when IR64 was lested over a span oftime representing the whole ofthe transplanting period in Ihe Punjab. This was no! delibera!ely designed bu! was a result of the reasonably large sample size Ihat represented the normal practices of farmers. This was nol possible in earlíer on-station trials that were invariably sown over a restricted, usually late, periodo These on-station trials, done in 1985, 1986, and 1987, did nol identifY IR64 because il yielded less than the check varieties in trials that were transplanted in July. PerlormanceolIR64 IR64 had a significant yield superiority of 5% over Pusa 44 in 43 trials, givíng an extra 300 kg of grain ha'¡ over a base of 6550 kg (figure 2). IR64 showed the besl performance (a 12% yíeld increase over Pusa 44) when transplanted from 21-24 June. The yield advantage decreased when IR64 was transplanted earlier or laler in June, which fits very well with the exlension recornmendation to spread transplanting equally around 20 June. An important feature ofIR64 ,is Ihat it matures 26 days earlier tban Pusa 44, This trait, along with high yield, favora its adoption in various situations (figure 3). Farmers' perceptions for traits other than grain yield (figure 4) identified IR64 to be superior to Pusa 44 for number of tillera per plant and resistance to BLB, slem borer, and leaf folder. IR64 is shorter so il is resistanl to lodging, whích allows it to be responsive to inputs, Advantages 0lIR64 over Pusa 44 IR64 had the following advantages over Pusa 44: • superior grain quality and higher yields • earlíer rnaturity, leading to a saving of iITÍgation water • resislance lo BLB and tolerance to whíte-backed plant hoppers • resistance to lodging 287 Figure 2. Yield (kg haO') of IR64 and Pusa 44 in 43 farmers' field tríaIs (26 FAMPAR and 17 IRD) in the Patíala district during the monsoon season of 1999 (Tbe overalI mean yield of 6860 kg ba- l of IR64 was significantly bigber [at tbe 1% level] tban tbe 6550 kg ba- l yield of Pusa 44 witb a t-value of 4.1 over 43 sites.) • allowing a green-manure crop Of summer mung (Vígna radiata [L] Wi\czek) to be grown between the wheat harvest and rice transplanting Adoption and further tesnng of IR64 AH participating farmers saved IR64 seed in 1999 for growing in kharíf2000. There was considerable seed exchange from farmer-to-farmer. Seed demand in kharif2000, from farmers who had seen the trials was considerable, but only five tones of seed could be procured and supplied lo farmers. Sorne entrepreneuriaI farmers and farmers' groups in the state have already become active in producing and procuring IR64 seed. As a consequence of the participatory trials in Patiala, PAU ís retesting IR64 at a number of research statíons under appropriate management. The Krishi Vígyan Kendra (KVK), Patiala, has undertaken large-scale testing on farmers' fields in PatiaIa and other districts ofthe Punjab in kharif 2000. To exploit the advantage ofIR64's early maturity, new agronomic practices and cropping patterns are being tested by the KVK Patíala in more than 40 triaIs with farmers. These are on growing summer mung and green manuring with sesbania in kharif2000. 288 SS Mal/¡¡, J.K Wítcombe, D,S, Virk, and x'f}, Singh 17S -* rr 170 --.~---. ¡ I M.turiq;(d) IR64 , -11- I M.tU'i~US.44 165 .... .c o. ~ ":;::"i 160 r 155 r 150 f- 140 ... .- 145 ¡ I I 6·10 11·15 16-20 21·25 June Juno June June i 26·30 June Time of transplanting No, of farrners 7 8 14 9 5 Figure 3. Days to maturíty ofIR64 and Pusa 44 in 43 farmers' field trials (26 FAMPAR and 17 IRD) in the Patiala district during the monsoon season of 1999 (Over 43 trials, IR64 matured signíficantly earlier [144 daysJ compared to Pusa 44 [170 days).) 80 60 40 Plant helght Tiller number ala Sto.. 00"" leaf folder Traits scored Figure 4. Farmers' perceptions (%) for IR64 in comparison to Pusa 44 for plant height, tiller number per plant, and resistance to bacterialleaf bligbt (BLB), stem borer, and leaf folder over 48 farmen (Like Pusa 44, IR64 was found to be 100% lodging resistant.) 289 Participatory Variela/ Se/eelion in Rice in ¡he Pun",ja",b,--_ Conclusions The PVS approach has been shown to be a potent tool: • to identify fanner-preferred varieties • to identify the correet recommendation domain of a variety (IR64 was previousJy tested in formal trials but was rejected for the Punjab because formal testing díd not represent Ihe temporal variability Ihat exists in high-potential production systems) • to correctly determine the best time oftransplanting of a variety • to identify varíeties Ihat give farmers new agronomie options • to promote the rapid adoption and dissemination of a variety References PAU. 1996. Package ofpractices for kJtarif crops ofPunjab. Volume XIII, No. 1 (March 1996):5. Ludhiana, India: Punjab Agricultural University. Singh, J. 1998. Paddy crop in Punjab. /997-98. A reportoflhe Directorate ofExtension Educativn, .nd Department of Economics and Sociology. Ludhiana, India: Punjab AgriculturaJ University. Singh. J. ¡ 999. Paddy erop in Punjab: A survey repaTI of1999. A repar! ofthe Directorate ofExtension Education, and Department of Economics .nd Sociology. Ludhiana, India: Punjab Agricultural University. Witcombe, J.R., A.J. Packwood, A.G.B. Raj, and D.S. Virk. 1998. The extent and rate ofadoption ofmodem cultivars in India. In Seeds ofchoice.' Makíng the mast ofnew varieties for smallfarmers. edíted by J.R. Witcombe, D.S. Virk, aud J. Farrington. New Delbi: Oxford lBH; London: Intermediate Technology Publications. - - 290 ...... _ .... _--- -------------------- Equity Issues in Varietal Dissemination through Farmers' Fairs (Kisan Melas) in Punjab, India J. Singh, s.s. Malhi, J.R. Witcombe, and D.S. Virk Abstraet In Ihe Punjab state of India, grain produclion has rapidly increased. One factor in this increase has beeo the fasl adoption of oew varieties. Punjab Agricultural Uníversíty (PAU) has played a major role in dislributing certified seed of new vaneties to lhe farmers ofthe stale, Mosl of Ihe seed is distributed by sales al farmers' fairs (kisan melas) held al PAU and ils regional research slalions. In this study, equity issu.s in Ihe sale ofwheal seed were examined in farmers' fairs held in Seplember 1999, In Ihe PAU kísan mela, smallholder farmers were found lO be considerably underrepresenled and large farmers considerably overrepresenled. The geographical distribution oflhe farmers who purchased seed was also studied. As might be expected, farmers lended to come lO where Ihe kisan mela was held from nearby administrative areas (termed blocks). This resulted in certain blocks being poorly represenled, PAU needs 10 .ddress equity issues, both socioeconomically and geographically, by incrcasing lhe outlels for seed sales in remote distriet, and areas oflbe state, and by encouraging smal! farmers to attend the kisan melas and purchase seed. Introduction Ihe Punjab State ofIndia has witnessed a rapid increase in the proouction offood grains, particularly wheat. Wheat production was only 1.74 million tonnes in 1960-61, but it rapidly increased to 14.46 million tonnes in 1998-99 as a result of increases in both yield and L!-¡e area under the crop. Wheat yields averaged only 1.2 t ha- t in 1960-61, but this increased to reach 4.3 t ha- t in 1998-99, Ihis very large increase in productivity was due to several factors, including the breeding and popularization ofhigh-yíelding varieties (HYV s), increased irrigation and fertilizer use, and the meehanization of fann operations. The fast adoption of quality seed was a major-perhaps the most important-factor. A survey of the wheat crop in the Indian Punjab (Singh 2000) showed that 79% of fanners kept seed from the previous crop, 12% purchased from private seed dealers, and 6% kept part ofthe seed and purchased part from seed traders, Only 3% offanners practiced fanner-to-fanner seed purchase. About 4% ofthe purchased seed was bought from institutional SOUTces such as the Punjab Agricultural University (PAU), the Punjab State Seeds Corporation, or the Natíonal Seeds Corporation. However, for new varieties, fanners tended, in the beginníng at least, to purchase seed from PAU. PAU produces and disseminates seed. Its primary responsíbility for production is breeder and foun· dation seed. However, ít also produces certified seed of recornmended varieties and, for wíder disseminatíon, recently released varieties. Most of this certified seed i5 distributed during fanners' fairs (kisan melas) that are held at the main campus at Ludhiana (PAU mela) and at fOUT regional research stations (RRSs) situated at Rauni, Bathínda, Ballowal Saunkhari, and Gurdaspur. In this J. Síngh ís wilh lhe Departmenl ofEconomies and Sociology. Punj.b Agrieultural University, Ludhiana, Punj.b, India. S.s. M.lhi i. witn Punjab Agricultuml University, Krishi Vigyan Kendra, Patiala, Punjab, India, J.R, Wifeombe and D,S. Vir\{ are with the Centre for Arid lone Studies, University of Wales, Bangor, Gwynedd, UK. Thi. document is an output ITom prajeel R7323, funded by Ihe Plant Seiencos Research Program .nd!he Natural Resources Systems Program ofthe UK Departrnent for Intemalional Develapment (DFID) for tIle benefit af developing countries, The views «pressed are not necessarily those of DFtD. 291 Equity lssues in Varietal Dissemination thrf?ugh Farmers' Fairs sludy, we examine the equity issues in PAU's wheal-seed distribulion system at the time of the fannerg' fairs. Methods PAU holds fanners' fairs twíce a year al Ihe maín campus and al four RRSs. Al Ihe fairs, certífied or truthfully labeled seed is sold for Ihe kharif(monsoon season) and rabi (winter season) crops. The seed is soJd on a "firsl-come-first-served" basis-fanners queue for their turn lo buy seed for cash. In September 1999, wheat seed sales at Ihe five kisan melas were surveyed by distributing a simple questionnaíre lo the fartners in the queues. There was a random sample of 359 fanners who purchased wheat seed at the PAU campus mela and a random sample of285 fanners at the RRS melas. Fanners were asked about their fann size, Ihe location o f their farm and Ihe amount of seed they had purchased. Results and discussion Station-wise and variety-wise seed sales Nearly 28 t of wheal seed was sold in all kisan melas. A major share of Ihe seed was sold al Ludhiana (70%) because it is centrally placed and is the main campus ofthe university (figure 1). When fanners visit Ludhíana for seed purchases, they also have the opportunity to ¡earn about other technologies. AIso, this mela is widely advertised and is a more significant event than the regional melas. Afier the PAU campus, Rauni (l 0%) and Bathinda (12%) accounted for mos! ofthe remaining seed sales (figure 1). Ludhiana 70% B, SaunkM4% Others 9% Rauni 10% By stallon By varlety Source: Director of Seeds, PAC and Ludhiana, personal cornmunÍCation, Note: "Others'" include vanotios PDW 233 (1.8%). PBW 138 (2.1%), PBW 175 (0.3%), PBW 299 (0.4%), PBW 373 (2.5%), and PBW 396 (1.7%), aH of which individuaHy .ccoun! for less than 5% of ,eed sales. Figure 1. Wbeat-seed sales ofPAU at the main campus and regional research stations Melas at Gurdaspur (4%) and Ballowal Saunkhari (4%) do not aecount for major sales of wheat seed. Gurdaspur is located on the northem comer of the state and is not well conneeted. Ballowal Saunkhari represents the main1y rainfed kandí belt of the state-a 10 km traet adjoining Ihe hiUy state of Himachal Pradesh, where irrigation facilities are very poor. F anners in this area largely belong to the low-resource category. 292 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ .. J. Singh. S.s. Malhi. J.R. Witcombe. and D.S. Virk Variety PBW 343 was in the greatest demand and accounted for 84% ofthe total seed sales (figure 1). The only other varíety to account ter an appreciable proportion of seed sales was WH 542 al 7%. The remaining five varieties accounted, in total, for only 9% of the sales. Patterns o/ seed distribution in addressing equity issues Overall seed distribution in the state in aIl kisan melas. A large proportion (45%) ofthe farmers in the Punjab have small landholdings of fewer than 5 acres. These farmers own only 12% of the cultivable land (table 1). In contrast, 29% of farmers who have more than 10 acres own 67% ofthe cultivable land (table 1). Table 1. Patterns ofWheat-Seed Sales at Fumers' Fair at PAU, Ludhiana, September 1999 Proportion offarmers in the state by Farmers attending the mela Quantity of seed sold (acres) Number 0/0 Tonnes % Number{%) Area (%) <5 20 6 0.9 5 45 12 5 to 10 49 14 2.0 11 26 21 10 lo 20 125 35 6.1 35 23 40 > 20 165 46 8.7 50 6 27 Farm size When farmers attending all the melas were categorized by the size oflandholding, it was found tbat smallholder farmers with fewer than five acres were extremely underrepresented (7% of purchasers versus 45% ofthe farmers as a whole). The 7% ofthe farmers from this category purchased 6% of the seed sold (figure 2). In contrast, farmers with large landholdings were hugely over-represented (46% of purchasers bu! only 6% of the farmers in the state). Less marked, but nonetheless quite large, underrepresentation occurred for farmers in the five- to lO-acre landholding category, and there was overrepresentation for farmers in the 10- to 20-acre category (figure 2). A similar, bu! less marked, bias was found for seed quantities purchased relative to the area of land held by each category of farmer (figure 2). Seed sales as a percentage oftotal sales varied ¡ittIe from the data for farmers purchasing seed, Le., once farmers decíded to purchase seed, there was IittIe difference in the quantity purchased, whatever the category of fiumer. The same analysis was done, disaggregated into the PAU mela (table 1) and the regional melas (table 2). Although, in both cases, there was underrepresentation of smallholder farmers and over-representation oflarger Iandholding farmers, the situation was better in the regional melas. The biggest difference between the regional melas and the Ludhiana mela that there were fewer large landholding farmers purchasing seed (46% in the Ludhiana mela compared to 28% in the regional melas). Spatial coverage The geographical distríbution ofthe farmers who purchased seed was also studied. The Punjab state is divided into 136 administrative units, caBed development blocks, tbat represent clusters of contiguous villages. As expected, farmers tended to come from nearby administrative areas or blocks 293 !:"quity lssues in Varietal Dissemi!!.~lion t~qugh Farmef 50 ~------------------------------------------~ ~ ~ ;:;.... 40 e -.. 20 a: 10 o :¡:¡ 30 c "'f! o. . O < SACRES 5-10 ACRES 10-20 ACReS > 20 ACRES Landholding size category of farmer • ~ ot all buye,.. El % of aU seed bought O % of farme:rs in Punjab • % o, landholding af'eS in PunJab Figure 2. Wheat-seed sales by PAU at its main campus and four regional research stations, categorized by landholding size (Sales by number of purchases and quantity of seed purchased are compared to the number of farmers and the area of land in the state by the landholding categories. The data prescnted are from a random sample of 644 farmcrs: 359 at the main campus and 285 at regional research stations.) Table 2. Parteros of Wheat-Seed Sales at Farmers' Fair at Four PAU Regional Rcsearch Stations, September 1999 Farm size Farmers attendlng !he mela Quantity of seed sold (acres) Number % Tonnes % <5 25 9 0.9 7 5 to 10 65 23 2.4 19 10 lo 20 114 40 4.5 36 >20 81 28 4.7 37 where the kisan mela was held. Fanners who visited kisan melas at the main campus and RRSs belonged lo 95 blocks oul of 136 blocks in the Punjab. lO In the PAU campus fair, the farmers sampled came from 65 development blocks oflhe Punjab state. Farmers also came from nine development blocks of the surrounding states of Haryana and Rajasthan. In the regional fairs, farmers came frem 59 developmenl blocks lo buy seed. The geographical distribution al block level shows the following: • Seed is only disseminated to 70% ofthe blocks in the Punjab despile the five kisan melas in the state. Forty-one blocks showed no representation among the fanners who were sampled. --------------~---------~_ 294 .....~- 1. Sillgh. ss. Malhi, J.R. Witcombe. and D.S. Virk • The majority of underrepresented blocks were in the Arnritsar and F erozepur districts where no fairs are presently held. PAU developed its seed-dissemination system in the post-Green-Revolution period to improve the equity of seed distribution in the state. In this system, small kíts of seed are sold to many farmers rather than larger quantities being sold to a few better-off farmers. When it was felt tha! farmers from remote areas were unable lo travel to the main campus in Ludhiana, regional kisan melas were started in order to make seed available in the regions. However, the seed-dissemination system of PAU at present does not address these issues satísfactorily. It is not known if these equity issues have always been present or if they have worsened over time. It is possible that over years, small farmers and those in remote geographical areas have become less enthusiastíc about traveling lo kisan melas, and small farmers have become dependent on larger farmers for their seed supply. Another factor may be that farmers with smaller landholdings are less prepared to take the risk of trying new varieties immediately afier their release and wait until Ihey can judge their performance on ¡he fields ofbetter-off farmers in theÍr village. Why small farmers have lower representation in melas and why ¡hey buy less seed are important issues that need to be addressed. Large farmers, who generally employ labor for farm operations, can afford to be away from their farms. They have the means and the time to travellong distances lo purchase seed to inerease farm revenues. On the other hand, small farmers • lack the resources to travellong distances • lack time because of theír involvement in farm and off-farrn activities, partíeularly in September when lhey are busy attending to the maturing rice crop • lack sufficienl funds to purchase seed at the time when they have incurred heavy expendítures on the standing rice erop, and have yet to gain a return from it • perhaps lack enthusiasm to try new varieties because their possible failure represents for them a g¡eater risk to their livelihoods than it does for larger farmers Although not ideal, the representation of small farmers is slightly better at the regional fairs because, on average, seed purchasers have traveled less faro Even there, they buy seed in smaller quantities than their representatíon. Small farmers require smaller quantitíes of seed because of their smalllandholdings, but this may also indicate that they lack money to buy more and that they have greater aversion lo risk than large landholders. Despite the sale of seed at regional stations, there are 41 blocks that were not served by the system in the sample. Most of these are in the border dístricts of Amritsar and Ferozepur where there are no RRSs. Ferozepur borders on Haryana and Rajasthan. Lack of availabilíty of seed from sources in the Punjab probably leads to a hígher adoption of varieties from adjoining states. Conclusions The P AV system needs to open more outletS for seed sales to address both equíty issues. If new regional stations catmot be opened in the Amritsar and Ferozepur districts, Idsan melas can be held in these districts in collaboratíon with the Department of Agriculture. More kisan melas, especially in poorly served blocks, may also help address the needs of small farmers in the state. Policies al the state level, involving Punjab State Seeds Co!poratíon and the Department of Agriculture, tha! are 295 Equitv lssIles in VarietaJ Dissemination through Farmer more smallholder-farmer friendly need to be formulated and adopted. Extension workers could create greater awareness among small farmers ofthe benefits of replacing seed more frequently and adopting new varieties earlier. One way of doing this is to encourage farmer experimentation by recommending that farmers try new varieties on a small area lo compare them to the exisling variety (see Malhi et al., this volume). References Malhi, s.s., J.R. Witcombe, D.S. Virk, and K.B. Síngh. Partícípatory varietal seleetion in rice in the Punjab. This volume. Singh.1. 2000. Survey reporl 01wheat crop ill PUlljab 1999-2000. Mimeographed reporto Ludhiana. India: Department ofEconomics & Sociology, Punjab Agricultural University. 296 Participatory Varietal Selection in Rabi Sorghum in India B.S. Rana, SL Kaul, Chari Appají, Parbhakar, NS. Kalyanshetty, B. Vs. Reddy, JR. Witcombe, and D.S. Virk Abstract Sorghum ís the thírd most important cereal crop in India "nd, over both the rainy (kharij) and the post-rainy (rabi) seasons, totals a combined area of more than I1 mi Ilion ha. Rabí sorghum is important for both food and fodder in the drought-prone areas of the states of Maharashtra (3.3 mili ion ha), Karnataka (1.5 million ha), and Andhra Pradesh (0.45 million ha). Genetic enhaneemenl .nd technology development have doubled the productívity ofkharif sorghum. Progress in rabí sorghum has becn slower because of several factors, such as more prevalent drought, shoot-fly infestations affeoliog lhe ioitial planl stand, low response oflandraces to applied nulrienls, and a limited choice of cultivars Ihal have Ihe traits required for adaptation to the rabi season. As a consequence, farmers eontinue to grow Ihe cultivar M 35-1, developed in 1935, lhal was a selection from Ihe Maldandi landrace. A pa'rticipatory varietal selection program for rabi sorghum, lO overcome Ibe laek of cultivar choice, is descríbed in Ihis paper. Introduction Partícipatory varietal selection (PVS) provides an opportuníty for farmers lo select one or more varieties from a basket of recently developed genotypes from plant breeding programs. Witcombe el al. (1996) reported that if a suitable choice of cultivar exists, PVS is a more rapíd and cost effective way ofidentifyíng farmer-preferred cúltivars than conventional, transfer-of-technology, extension methods. In India, Maurya, Bottrall, and F arrington (1988) tested advanced lines of rice witb villagers in Uttar Pradesh and successfully identífied superior material tbat was preferred by farmers. Also in India, Joshi and Witcombe (1996) identified farmer-acceptable cultivars of rice and chickpea from a range of released and nonreleased cultivars tested in farmer-managed participatory trials. Farmer-acceptable cultivars were found among released varietíes but not among those recommended for tbe area. Relevance of PVS in rabi sorghum The participatory approach to varietal selection is considered valuable when formal breeding and seed-supply systems have been unable lO fulfill tbe needs of users. This ofien occurs where the agroecological or socíoeconomíc environment differs significantly from those anticipated and tested foc in the formal system of variety testing. In rabi sorghum, several factors mean that PVS could he a useful approach: low adoption of improved cultivars, variable growing conditions and multiple productíon constraints in farmers' fields that are difficult to simulate on the research station, and local preferences for grain quality. a.s. Rana, S.L. Kaul, and C. App'íi are with lbe National Reseafoll Centre for Sorghum (NRCS), Rajendranagar, Hyderabad, Indía. Dr, Parbhakar ís atlbe Centre for Rabí Sorghum, Solapur, .nd N.S. Kalyanshetty ¡sal Kri.hi Vigyan Kendra, Shrí Sidde,hwar Kri.hi Vidnyam PTa••rak Samstha, Solapur, Indía, B. V.S. Reddy is with the International Crop, Rese.rch Instítute for !he Semí-Arid Tropícs (ICRlSAT), patancheru, India. j,R. Witcombe and D.s. Virk are wi!h the Centre for Arid Zone Studies, University ofWales, Bangor, UK, This document is an QUlput fiom proje.1 R7409 funded by lhe United Kingdom Department fOf International Development (DFID) Plant Sciences Research Programme for me benefit of developing countries. The views expressed are not necessarily those ofDFID. 297 Particip.atory Varietal Se/ce/ion)" Rabi Sorghum in India Activities Six nongovernrnental organízations (NGOs), six centers of the All-India Co-ordinated Sorghum Improvement Project (AICSIP) located in state agricultural universities, the Nationa! Research Centre for Sorghum (NRCS), and !he Intemational Crops Research Institute for the Semi-Arid Tropics (ICRISAT) collaborated in the research. The activities involved the identificatíon of villages, NGO user groups and farmers in those villages, and the conducting of rapid rural appraisals (RRA) to identify which varieties farmers cultivated and how they cultivated !hem, as well as to assess constraints to productivity. Farmer-managed trials of lOor more identified elite varieties, hybrids, and selected local control varieties were conducted by farmers on !he ir fields. Joint monitoring by researchers and farmers was done at a mínimum of three crop stages, and data were collected on !he performance of the entries. before the rabi sowing, the NGOs selected !he participating farmers by organizing group discussions !hat inc!uded both farmers and officials. Materials AICSIP has continuously developed new dual-purpose cultivars wi!h rabi-adapted traits, such as resistance to droughl, shoot-fly, and charcoal roí. 1t has involved direct selection from landraces, as well as hybridization and progeny selection. Genotypes in the trials inc1uded several !hat were developed and tested in initial and advanced varietal trials of AICSIP in recent years, and three selections from an ICRISAT populatíon based on M 3S-2. There were 11 genotypes in the rabi 1998 trials and 22 in !he rabi 1999 trials. These included the following: • five recently released cultivars: variety CSV 14R, GRS I (DVS 4), 9-13 (DVS S), SeU, and a hybrid (CSH ¡SR) • three elite genotypes from AICSIP advanced varietal trials: SPV 1155, SPV 1359, and SPV 1380 • six from AICSIP initia! varietal trials: RSLG 262, SPV 1360, SPV 1375, SPV 1411, SPV 1428, and SPV 1429 • four genotypes tested earlier: SPV 655, SPV 1215, SPV 12!7, and GSS 2 • three population bu!ks derived from M 35-1: BLK 1, BLK 2, and BLK 3 • the popularly grown cultivar, M 35-1 Five genotypes-CSV 14R, CSH 1SR, SPV 1359, SPV 1375 and M 35-I-were uniformlytested by al! six NGO groups, but others were tested selectively by from one to five NGO groups, dependíng on previous experience. Varietíes for which farmers could maintain the seed themselves were preferred over hybrids. Trial design Each of the six NGOs selected three villages, each with six participating farmers. The number of varietíes tested by each NGO ranged from 10 to 12. The NGOs, in consultation with farmers, had decided to give each farmer 2 kg seed of each entry for advanced varietal trials and 1 kg seed of each entry fOf initial variela! tria!s. However, involving more farmcrs by providing each of !hem with less seed was considered a more appropriate designo Each genotype was tested by three farmers lO 298 represent three replications. A trial consisted of growing the new cultivar alongside the local cultivar in a similar-sized plol without any plant protection and under farmer management. Observations on grain yield, dry fodder yield, grain appearance, and farmer-preferred traits (for male and female farmers) were recorded by skilled helpers. Farm walks, focus-group discussions, and house-level questionnaires were employed. Results Studies undertaken during rabi 1988 revealed Ihat farmers' practices varied greatly. In mostplaces, varieties were grown Iffider rainfed conditions, bul sorne farmers provided a single irrigarion, and nitro gen applications varied from 0-100 kg urea per acre. At Dhulia center, farmers planted deep behind lhe plough with no fertilizer and no pesticides. At Parbhani, farmers used four to five cart loads of farmyard manure and two 50-kg bags of 20:20:0 compound fertilizer per hectare. Al Solapur center, the crop was planted in shallow soil, and 100 kg urea per acre was applied under irrigation, but other farmers did no! apply fertilizer under rainfed conditions. At Bijapur, farmers applied 25 kg urea + 25 kg di-arnrnoniurn phosphate (DAP) per acre. In 1998 in Maharashtra, SPV 1359 and SPV 1155 were ofien preferred by farmers over lhe local cultivar M35-1. At sorne locations, olher varieties, such as SPV 1380 and the ICRISAT bulk derived from M35-1, were also preferred over M35- L Local germplasm seleetions, such as RSLG 2623, were preferred at locations outside of lheir locatíon of origino This led liS to test lhe local germplasm in aH participating centers in 1999. For 1999, allhough the genotypes were tested by all the NGO groups, only the data from Solapur are presented in detail (table 1). In six trials, SPV 1359 was found mostproductive with 3.7 t ha'! grain yield, compared to 1.7 t ha'! grain yield oflhe local cultivar. Thus, lhe grain yield ofSPV 1359 in farmer-managed tiials was more Ihan double that of lhe local cultivar. There were more trials of SPV 1380 and CSH 15R; both gave almost double the grain and fodder yields oflhe local cultivar (table 1). M 35-1 was also tested against lhe locally grown landraces. In 16 such comparisons, its grain yield was 2.4 t ha'¡ (compared to 1.5 t ha· l for ¡he local cheeks), an inerease of 66%. The inereases over locally grollm cultivars are sumrnarized in table 2. Genotypes tested in lhe initial varietaI trial also performed well (table 3). The cultivar SPV 655, earlier dropped from coordinated trials, gave the highest grain yield, 3.2 t ha'¡ against only 1.3 t ha'¡ of lhe farmer-grown local cultivar, an ínerease of 146%, and its fodder yield was double lhat of the local variety. The grain and fodder yield of SPV 1413 was also double Ihat of lhe local variety grown by farmers. Two other genotypes, RSLG 262 and SPV 1411, gave more lhan 1.5 times the grain and fodder yields of the local varieties grown by tarmers. These genotypes will be tested in 2000-2001 in more trials. Farmers' perceptions ortbe improved genotypes During 1998, farmers in general were satisfied wilh Ihe grajn yield of the new varieties, compared to their local cultivar, and demanded more seed from the new varieties. The popularly grown variety M 35-1 was not liked at certain plaees because ofits side tillers. Women preferred bold and pearly seed, medium plant height (since this was convenient for harvesting the heads), higher flour 299 Understan.~lit1g Table 1. Agroeco.!.ogical Domains ._ __ Grain and Fodder Yields of Improved Genotypes in Farmers' Fields in Advanced Varietal Testing, Rabi Season, 1999, Solapur, India Gra;n yield (t ha-1) Fodder yield (t ha-1) No_ oftrials Improved Local Improved Local SPV 1359 6 3.7 1.7 4.5 3.0 SPV 1380 Entry 40 2.8 1.4 6.0 3.0 SPV 1155 2 2.4 1.8 5.2 4.1 M 35-1 16 2.4 1.5 5.4 2.9 CSH 15R 25 2.3 1.1 6.0 2.9 Table 2_ Percent !nerease of Improved Genotypes over Farmer-Grown Local Cultivar in Farmers' Fields in Advanced Vanetal Testing, Rabi Season, 1999, Solapur, India Grain yield Entry Percen!age of Irlals Fodder yield (%) ofloc..1 Range with >20% increase (%) of local Range SPV 1359 116 7-195 32 (-)20-87 SPV 1380 96 4-194 67 88 113 20-244 M 35-1 66 (-)36-287 56 89 (-)14-382 CSH15R 101 (-)7-228 88 105 (-)17-276 table 3_ Cultivar Performance in Farmers' Fields in Initial Varietal Testing, Rabi Season, 1999, Solapur, India Grain yield (1 ha") Improved Local local (%) 2.5 1.5 69 5.6 3.8 47 5 1.6 1.4 14 6.3 5.3 20 3 2.1 1.6 29 4.3 3.1 40 RSLG262 3 SPV1462 BRJ 356 SPV 1413 ~PV655 SPV1411 M 35-1 (B-3) Increase over Local Improved En!1}' Fodder yield (t ha") I"crease overlocal !%) NO.of trlal5 5 2.5 1.2 109 6.0 2.8 117 15 32 1.3 146 6.6 3.2 109 15 3.0 1.7 72 6.4 4.0 57 1.6 3.8 recovery (9: 1), better cooking quality (good dough), soft and good tastíng chapatti, and a longer storage life of the flOUL During rabi 1999 al Solapur, farmers reported on the high grain yield and good fodder quality ofthe improvedcultivar SPV 1155 compared to M 35-1. Farmers sald that SPV 1359 was excellentfor lts higher grainyield and bold grain but that it had no sweetness in the stemand thus its fodderwas not preferred. In the case ofSPV 1380, funners' reactions were that it had excellent grain yield, bold grain, and loose panicles that helped stop birds sitting on thero to eat the grain. However, they reported that it had poor-quality fodder because of a longer intemodallength and leaf falL 300 For the hybrid CSH 15R, farmers reported that it was good for high grain yield under irrigation, tha! it was earlier in maturity than the local cultivar, and that its fodder was moderately preferred. They were unhappy with the 60% to 70% grain filhng that reduced its yield. Conclusions Participatory varíetal selection appears to be an effective approach to supplement plant-breeding efforts in marginal areas, where progress with varíetal adoption has been slow. It enables farmerpreferred varíeties lo be idenlified and tests the rigor of Ihe varíetal-tesling program in multi-environment coorrunated trials. In contrast to fue general belief that M 35-1 is a popularly grown variety, access to NGOs and farmers revealed that various landraces are stíll grown in the Solapur area in addition to M 35-1. Improved varieties such as SPV 1359, SPV 1380, and SPV 1155 from !he AICSIP advanced varietal trials, and SPV 655, a rejected genotype from the ACISIP trials, performed excellent1y. The first two have already been identified for re1ease. Thus, the varíetal tesling al !he research slation is usualIy, bu! not always, satisfactory to determine adaptability under realistic farmer management. Further PVS success will depend on newly evolved varieties, based on fue farmers' perceptions learned in fuese studies. References Joshi, A. and J.R. Witcombe. 1996. FanneT particlpatory CTOp improvement. 11: Partícipatory varietal selec!ion, a case study in India. Experimental Agriculture 32:461-477. Maurya, D.M., A. Bortrall, and J. Farrington. 1988. Improved hveliboods, genetic diversity and fanners' participatíon: A strategy fOT rice-breeding in rainfed aTeas ofIndia. Experimental Agriculture 24:311-320. Witcombe, J.R., A. Joshí, K.O. Joshi, .nd B.R. Sthapit. 1996. Fannerparticipatory crop impTovement. 1: Varietal selec!ion and breeding methods and tb.iT impact on biodiversity. Experimental Agriculture 32:445-460. Witcombe, J.R. and D.S. Vírk. Forthcoming. The success ofparticipatory varieta! selectían in a hígh-potential production system in India. How important ís the participatíon? In Particípatory Plan! Breeding ond Rural Development, edited by J.A. Ashby and L. Sperlíng. Cah, Columbia: CGIAR Systemwide Program on Participatory Research and Gender Analysis, Centro Internacional de Agricultura TTopical. 301 The Impact of Participatory Plant Breeding (PPB) on Landrace Diversity: A Case Study for High-Altitude Rice in Nepal K.D. Joshi, B.R. Sthapit, and JR. Witcombe Abstract Partícipatory plant-breeding (PPB) melhods were used to develop Iwo .ceeplable, cold-to\erant nce varieties in Nepal: Maehhapuchhre-3 (M-3) and Machhapuchhre-9 (M-9). Both were derived from Ihe eross Fuji I02/Chhomrong Dh.n. Following the introduclion oflhese varieties from 1993 to 1998_ ¡he changes in the rice landraees and vaneties that fanners grew were studied in 10 villages. In seven of the villages, forwhich dala were analyzed for bOlh 1996 and 1999, fanners grew 191andraces and fourmodem varieties, ofwhich three (M-3, M-9, and Lumle 2) were !he produc!s ofPPB. These three varieties eovered 11 % of the total surveyed area in 1999. The introduction of the PPB varierÍes had tne greatest impact on the more commonly grown I.ndraees. During the years studied, because the new vaneties had exotic gennplasm in their parentage, there was an overall inerease in vanetal diversity. However, in the future, ¡ncreasing adaption ofM-3 and M-9 could result in significant reductions in variet.l diversity. Introduction Participatory plant breeding (PPB) is increasingly being used for decentralized crop improvement (Weltzein et aL 2000; Eyzaguirre and Iwanaga 1996; Sthapit, 1oshi, and Witcombe 1996; Witcombe et aL 1996). Important elements ofPPB commonly include the use in the breedíng program of a locallandrace or locally adopted variety as a parent, the sereeníng of segregating matel'Íals in the target environment, and the participation of farmers in goal selting, selection, and evaluation. Farmers in the bilis and mountains ofNepal continue to grow landraces because centralized plant breeding has had limited success in producing varieties that farmers wish lo adopt. The use of decentralized, participatory methods could remove this constraint lo the adoption of new varieties. However, the products ofPPB, ifbighly preferred by farmen;, could have a considerable impact on local agrobiodiversity, In recent years, there has becn a growing awareness ofthe value and utility ofagrobiodiversity, and local nongovernmental organizations (NGOs) and intemational organizations are concemed about the conservation and utilization ofbiodiversity. For example, during the third global meeting ofthe Intematíonal Plan! Genetic Resources Institute (IPGRl), in July 1999, Pokhara, Nepal, the in situ crop conservation project of DI. Ramnath Rao of IPGRl presented one possible impact that PPB products could have on landrace diversity (figure 1). K. D. Jo,hl ¡, with local lnitiative, for Síodiversity Researeh and DeveJopment (U-SIRO), Pokhara, Nepal. B. R. Sthapit i. wíth the Intematíonal Plant Genetie Resources Instítute (IPGRl), Asía, the Pacitic and Oceania, Pokhara, Nepal. 1. R. WítcomOO ís with the Centre for Arid Zone Studies, University ofWales, Bangor. UK. Most of,his work was carried out with core funds ftom U-SIRO. The ¡nitia! monitoring of 'he adoption and varietal spread of PPB products "'as joíot1y funded by the UK Dopartment for Intemational Dovelopment (DFIO). project R6636, for tite OOnetit of developing countries and by the Intemational Development Research Centre (IDRe). The views expressed are not necessarily those of DFIO, U-SIRO, or IORC. Machhapuehhre-3 and Machhapochhre-9 are the products ofthe Agricultural Resean::h Station, Lurnle, Nepal Agricultural Researeh CounciL Wo acknowledge the contribution ofall the farmers who eollaborated in the ioítia1 PPB aod in the ,preadofM-3 and M-9. Thetieldwork by B. B. Paude!, community organizer of U-SIRO is apprecíated. 303 The Itrlpact o(Partícipatory Plan/ Breedí/1g 0/1 Landrace Dlversitv L1 L2 L3 L1 L2 L4 LS Le L4 L5 L7 L8 19 L9 V1 X V2 V3 L4 L5XV3 l7 Selected Initiat system landraces L5 LB L7XV1 LB L9 New parents Possible result Figure 1. Tbe ¡mpact ofPPB producís on local agrobiodiversity (In tbis scenario biodiversity is increased because it is assumed that non e of the landrace are entirely replaced by new variedes produced by PPB.) Materials and methods Participatory plant breeding of high-altitude rice was initiated in 1993 by the Lum!e Agricultura! Research Centre (LARC) in the villages ofChhomrong and Ghandruk, both at an altitude of2000 m, in !he Kaski district of Nepal. Eighteen farmers collaborated in selecting between, and sometimes within, 10 Fs bulk lines derived from three different crosses made by !he Agricultural Botany Dívisíon ofthe Nepal Agricultural Research Council (Stbapit,.Joshí, and Witcombe L996). As a result of this program, in June 1996, the Variety Release and Registration Cornmittee (VRRC) of N epal made the first release of a variety produced with the extensive use of particípatory methods: Machhapuchhre-3 (M-3) (Joshi et al., 1996). In a participatory varietal selection (PVS) program, farmers at Chhomrong also identified Machhapuchhre-9 (M-9), a sister !ine to M-3, as an acceptable variety. Starting in 1996, M-3 and M-9 were introduced into víllages situated between 1200 ID and 2300 m altitude by NGOs such as the Local Initiatives for Biodiversity Research and Development (LI-BIRD), CARE Nepal, the Annapurna Conservation Area Project (ACAP), and LARC. The adoptíon and spread ofM-3 and M-9 were monitored from 1996 to 1999. Five víllages were surveyed in both 1996 and 1997, and 10 in both 1998 and 1999. Only the surveys in 1999 are reported here (table 1). Information was collected froq¡ the surveyed households using semi-structured ínterviews. Samplíng was purposi ve (on1y from househo1ds known to have been given seed of M-3 or M-9). In 1998, farmers were asked about their adoption intentions to assess the possible impact ofPPB products on the diversity of rice landraces. The 1999 survey, which covered about 18% of the households that had adopted and grown PPB products within the last three years (table 1), also collected information on the 1andraces farmers grew in 1996. For each household, the total area of khet land (irrigated and bounded terraees of land where rice is grown) was determined from the land-ownership certificates, and a total inventory of rice varieties, with the area that each variety occupied, was compiled. The rice varieties and landraces were analyzed by the area in which they were grown and the number of households that grew !hem. Changes between 1996 and 1999 were assessed for area and household number for the more cornmon landraces. The statistical significance of changes in area was determined by a two-tailed paired f test between the areas reported for 1996 and 1999. 304 KD, Jos},i, S,K Sthapit,{jnd JR, Witcombe Tabla 1. Tbe Study Villáges, 1999 Households: Altitude and aspect in the village in 1999 survey with data for 1996 Village Oistrict Chhomrong -Ghandruk Kaski 1800-2000. NE 55 12 12 Chane-Kimche, Landruk-Tolka Kaski 1500-1900, SW 142 14 14 Marangche Kaski 1400-1600. SE 34 30 30 11 Kande Kaski 1400-1600. N 55 11 Khanigaun, lwang Kaski 1600-1900. SE 68 11 O Patlikhel Myagdi 1400-1700. SW 50 6 6 Bhakimle Myagdi 1600-2300. N 181 16 16 Chipleti Myagdi 1400-1800. S 78 11 11 Bangsing Deuralí Syangja 1300-1500, S 64 6 O Bangephadke_ Syangja 1400-1600, S Total 28 10 O 754 127 100 Results and discussion Adoption 01 M-3 and M-9 in 1999 M-3 was introduced to alllO study villages and was adopted in al! ofthern, whíle M-9 was introduced to seven ofthe víllages but was adopted in three (figure 2). The rnost important factors in determining adüption were the altitudes of the villages and the year in which they first received seed. Apart frorn the low-altitude víllage of Bangephadke, adoption of either M-3 or M-9 was at leasl 10% ofthe rice area in víllages that had received seed before 1998. Impact 01 M-3 and M-9 on varietal richness Since the ancestors of the landraces were not kIlown, no analysis of diversity could be done that required a kIlowledge ofthe relatedness ofthe cultivars with each other. However, richness can be assessed by the number oflandraces and varíeties grown (figure 3) for the seven villages for which Ihere were data for both 1999 and 1996. The total number ofrice cultivars decreased líttle in the study víllages. Thís was despite the adoptíon of varíeties produced by PPB that might have been expected to have replaced several of the landraces. The number of rice cultívars grown in 1999 increased in two ofthe study víllages and decreased in two, while in three ofthe villages there was no change (figure 3). The decrease in díversity in Chhornrong and Ghandruk i5 not 5urprísing since the inítial PPB prograrn was conducted in these villages. In the early stages, as manyas nine lines were grown in 1996 at Chhornrong alone, but by 1999, the undesirable Unes had been dropped. Another case of decrease was in Chane and Kimche, where adopting households dropped the Tairige and Takmare landraces to grow M-3 even though M-3 covered less than 15% ofthe total rice aTea. In a11 of the seven study víllages, sorne of the rice area that was under landraces in 1996 was occupied in 1999 by M-3 and M-9. This increased genetíc diversity, since M-3 and M-9 have exotic germplasm in their ancestry. M-3, M-9, and Lumle-2 all have a locallandrace, Chhornrong Dhan, as a parent. Fuji 102, an exotic varíety from Japan, ís a parent of M-3 and M-9, and IR36, an 305 The lmpaCl o[ ParticipalOly PIOlll Breeding 011 LandraCl 50 Allilude _lg~ High Mid '"g¡ 40 "'" 30 '" ¡,; '" Study villages Figure 2. Adoption in 1999 of Machhapuchhre-3 (solid ban) and Machhapuchhre-9 (open bars) in 104 sampled housebolds in the 10 study villages (The year of flTst introduction of M-3 or M-9 is indicated aboye tbe ban.) Interoatíonal Rice Research Institute (IRRl) varíety, is a parent ofLumle-2. Chhomrong Dhan was grown in only tbree oflhe seven villages, so in four oftbem, there was no cultivar Ihal was genetically related lo tbe PPB products. Classification 01 Jandraces by their relative abundance In 1999, farmers grew 19 landraces and five modero varíeties in tbe seven study villages for which both 1996 and 1999 data were available. Of the five modem varietíes, !bree were tbe products of PPB (M-3, M-9, and Lumle-2). The average area devoted to any landrace by tbe households in tbe study villages was quitesmall «0.3 ha) (figure 4). Oflhe 191andraces in these seven villages, 12 were reasonably common (figure 4). Oftbe seven less common, five were grown by only one oftbe sampled households and two had a combination oflow household number and a small average area. While studying the occurrence and diversity ofIocallandraces in Kaski Ca low to mid-hill site, 750 m to 1300 m) and Bara (100 m to 150 m), Joshi el al. (1999) found tbat only a few landraces were widely grown. The great majority oflandraces or varieties were less common and had eitber a small area or few households growing them, or botb. A similar result was found for ghaiya (upland rice) landraces (Joshi et aL, forthcoming). This was also found for modem varíeties in the Nepal Teraí (Joshi and Witcombe, tbis volume). 306 KD. Joslri. S.R. Stlrapit, and JR. WÍlcombe 16 14 ~ ... 12 '5 10 <1> C.l 8 'l5 ...<1> 6 E. 4 Z 2 1/1 u '¡: .Q ::1 Study vi llagas Nore: Open bar, 1996; solid bar, ~ 1999. Impact 01 M-J and M-9 on landrace diversity Farroers' pereeptions in 1998. In 1998, fanners' perceptíon8 of the iropact that PPB products would have on local landrace diversity were recorded. Most of the respondents reported that they would increase the area under M-3 or M-9. About 24% ofthe respondents reported that the adoption ofM-3 or M-9 would either reduce the area under landrace Kathe or entírely replace it A similar situation was perceived for landraces Kalopatle (8% of respondents), Maisara (6%), Raksali (3%) and Darmalí (3%). A fllrther 10% ofthe surveyed households al80 mentíoned the possíble partial replacement of I Oother landraces and one modern varíety, No households reported that they would entírely replace the landrace Chhornrong Dhan or Ihe modern variety Khumal-4, even though at leas! one household mentioned the complete-replacement of at leasl one ofthe remaining 19 landraces. Results olthe 1999 survey The 1999 survey confinned mosl of the 1998 perceptions, The area and number of adoplíng households ¡ncreased significantly for M-3 and M-9 (figure 5). The ¡ncreasing adoptíon ofM-3 and M-9 is líkely lo have far greater impact on landrace diversíty in the future than what had already taken place by 1999. In 1999, the area under 12 out of the 19 landraces had decreased, whíle for eight of them, the number of adopting households decreased. Area was more dynamic than the number of households probably because a decision to change the area under a landrace is more common than to entirely drop a landrace or adopt a new one. 307 The lmpact o( Partícípatorv Plafll Br~eding on Landrace Diver,;ty 0.7 O~6 : • I Marsi 0.5 8 0.4 • Juwari 0.3 .. • O 0.2 e 0.1 O eOarmall Sl_ • O .- Gharkole • .S'lI'lJali Raksale Rakae ""..... • Chhomrong Dhan ....... o o 5 15 10 20 25 30 Numb<>r of househoh:l. Note: The more eornmon landraces are marked with solid círcles and labeled with names. Tthe less cornrnon landraces are marked witb grey circles and are Phake Dhan, Gunta, Pahenle, Rakse, Maísare, Takmare, Galaiya, Tarkaya and Anga. Figure 4. Landrace frequency by number of adopting bousebolds and average an~a grown byeach housebold in 1999 in the seven villages for which 1996 and 1999 data were available (see table 1) .="J: .. 'O ~ -.. .e o J: Q. 0.18 0.16 0.14 012 0.1 ••• 0.08 ~ 0.06 " 0.04 ~ 0.02 '" 1: M-9 •• • M-3 o o 10 20 30 50 HOUSéhold number Note: Lines are labeled next to the 1996 ongin cf the lineo Both éultivars have slgníficant changes in area ("'.* = P :s .0(1). The significance of <::hanges in adopting households was not tested. Figure 5 Change in area and household adopters from 1996 to 1999 for M-3 and M-9 in seven villages (see table 1) 308 K.D. Joshi, B.R. Sthapit, and J.R. Wi/combe As an example, the changes in adoption of rice cultivars from 1996-1999 were analyzed for the seven villages shown in figure 3. The decrease in area was statistically significant for eight oflhe 10 most common landraces, i,e., Chhomrong Dhan, K.hate, Kalopatle, and Sinjali (p < .001), Raksalí and Rakse (p < ,O 1), and Darmali and Maisare (p :::; .05). In al! cases, this decrease was largely accounted for by a compensating íncreasc in M-3 and M-9. Ofthese six landraces, four ofthem had becn mentioned by farmers for possible replacement in the 1998 survey. Most sígnificantly, tbree ofthe eight landraces where the number of adoptíng households declined were those that were grown by the most households. Hence, it was mainly Ihe most common landraces Ihat had fewer adopters in 1999 than in 1996, and the less common landraces were the most buffered against change, AH of Ihe five landraces with only a single household in 1996 were also grown by a single household in 1999 (figure 6). . 0.7 ¡ 0.6 ~ ~ o ",•• o ".. 0.5 0.4 1 ~ .... Q I! & .. I! ¡¡ --+-- ChlllH'llfOnQ 0.3 0.1 ../ ---- -'Ó--SiuA1ll -<:>--- .....- ........ --..-Gilrkhol. .....__M>Bi ·•• ·•• ./ • ;.:Jt 0.2 -o--- tnmI.1i b 11 /~ --J~ri --......,. ·• o .0 ____ PWMt Ch,,, • : o o --........ ! """"-0.... Po_ ...." _. -'-Malnrl 5 10 15 20 25 30 -- •..• TlkI'n.... :._-lt- Household number • • - Note; Cultivars with significant changes in area have been indicated by asterisks (*.* = p .$ .001; ** =P S .01; *' "'" P ~ .05). Nonsignificant changes are indicated with 'n,' elose 10 the 1996 ori¡¡in of!he lineo Tho significance of changes in adopting households was not tested. Figure 6. Change in area and household adopten from 1996 to 1999 for higb altitude rice landraces alter ¡be introduction of M-3 and M-9 in seven villages (see table 1) Varietal change is a common and continuous process in most subsíslence farming where farrners allocate different proportions oftheir land lo a cultivar from one season to anolher. Landraces !hat most c10sely match Ihe new varieties, but have a lower yield or other undesirable traits, are replaced first. The landraces wilh Ihe greatest reduction in area and adopting households were Chhornrong Dhan, Kalhe, and Kalopatle. The niches ofthese varieties closely match Ihose ofM-3 and M-9. By 1999, six years after Ihe commencement of the PPB program, Ihe products of PPB occupied about 11% of Ihe total rice area and about 14% of Ihe surveyed households. There is a continuing trend ofincreasíng adoption ofM-3 and M-9 in both area and household number. In Ihe past, in spite of concerted efforts by government extension agencies lo promote modem rice varieties, Iheir adoption was very poor. For example, only 100/.,-11 % of farming households were growing improved rice cultivars in a survey of 1688 households in 11 districts of eastem and westem Nepal 309 The Impact o(Participatary Plant Breedmg 011 Landro< nearly three decades after the intervention of improved varieties (Chemjong e! al. 1995; LARC 1995). Targeting specific niches that were not addressed by conventional breeding programs is one ofthe objectives ofPPB. The increasing acceptance ofPPB products in the study villages provides evidence for its success. To conserve landraces, maintaining diversity at the community leve! should be sufficient. Although there was an oyeran 105s in landrace richness in the sample, it was no! severe and M-3 and M-9 added to the diversity. Landraces found to be mos! al risk can be utilized in particípatory plant breeding programs so tha! their useful genes are incorporated in more productive genotypes and hence conserved. In terms ofutility and food security, diversity at the household level may be more important, and the addition of either or both M-3 and M-9 to the varietal portfolios of about 14% of the farmers would contribute to this diversity. An important finding was tha! the adoption of landraces was highly dynamic, wi!h losses and gains at the village level and cornmon changes in areas. Ex situ conservation is simply a "snapshot" of a situation in the year in whích the coHechon was marle. PPB, in producing varieties that farmers Iike, contributes to the dynamism. It accelerates cbange by introducing genes and genotypes but may not fundarnentally cbange the age-old process ofvarietal adoption. Indeed, as argued by Witcomhe et al. (1996), PPB in ils collaborative form in farmers' fields is a dynamic form of in situ genetic conservation. References Chemjong, P.B., B.H. Baral, KC. Thakuri, P.R. Neupane, R.K. Neupane, and M.P. Upadhaya. 1995. The irnpaclof Pakhribas Agricultural Centre research in the eas/em Hills of Nepal: Farmer adoption of nine agricultural technalagies. Dhankuta, Nopal: Pakhribas Agricultural Centre. Eyzaguirre, P. and M. Iwanaga (Eds). 1996. Proceedings of the lDRCíIPGRl Workshop on Participatory Plant Breeding. 28-30July. Internacional Agricultural Research Centre, Wageningen. Rome: International Plant GenetÍcs Resources Institute. Josm, K.D., B.R. Sthapit, R.B. Gurung, M.P. Gurung, "ud J.R. Wilcombe. 1996. Propasal for release of rice variety Machhopuchre-3. Kaski, Nepal: Lumle Agricultural Research Center. Josm, K.D., D.K. Rija!, R.B. Rana, S.P. Khaliwada, P. Chaudb.ry, K.P. Shrestha. A. Mudawari, A. Subedi, and B.R. Sth.pi!. 1999. Adding benefits l. Through partícípatory plant breeding (PPS), seed nelWorks and grassroots strengtheníng. In Conserving agricultural biodiversity in situ: A scíentific basi. for susta/noble agricu/ture. Proceedings ofthird global meeting ofin situ crap conserva/ion project, 5-12 July /999, Pokhara. Nepal, edited by D. Jarvis, B. 8th.pit, .nd L. 8ears. Joshi, K.O. andJ.R. Wítcombe. Partícípatory valÍetal selection, food security, and varietal díversíty ín. high-potentia1 production system ín Nepal. This volurne. LARC. 1995. The adoptivn ond diffusion and incremental benefits offifíeen technologies for craps, horticulture, /ivestock and forestry in the western hills ofNepal. LARC Occasional Paper 9511. Pokhara, Nepal: Lurnle Agricultural Research Centre. Sth.pil, B.R., K.D. Joshi, .nd J.R. Witcombe. 1996. Fanner participatory crop improvement. 1lI: Particípatory plant breeding, a case ofhigh altitude rice from Nep.l. Experimental Agriculture 32:479-496. Weltzein, E., M. Smith, L.S. Meítzner, .nd L. Sperling. 2000. Technical iss"es in participotory plom breedingfrom the perspective offonnal plant breeding: A global analysis of issues. results and cu"ent experiences. Working Document no. 3. Washington, DC: CG!AR Systemwide Program on Participatory Research and Gender Analysis for Technology Development and Institutionallnnov.tíon. Wilcombe, J.R., A. Joshi, K.D. Joshi, .nd S.R. Sthapí!. 1996. Fanner panicipatory crop improvement. 1: Methods for varietal seJection and breeding and their ímpact on biodiversiry. Experimental Agriculture 32:453-468. 310 Role of Farmers in Setting Breeding Goals M. Subedi, P.K. Shrestha, S. Sunwar. and A. Subedi Abstraet Maize (Zea mays L.) is Ihe mas! important crop in Ihe hill farming system in Nepal. It plays an important role in the livelíhood of the people living in lbe Mis. The hilIy area of the Palpa, Gulmi, and Arghakhanchi districts eXlending lowards Pyuthan and further weSI has a unique geophysical environmenl, which is different from other maize-growing areas in Nepal. Farroers in Ihis area not only have poor aecess lo agricultural inputs, including improved genetic materials, bUI the improved varieties tested so far do not exactly match Ihe unique growing condilions .nd the needs of farmers in the area. Therefore, Ihe major proportion of maize in the Palpa, Gulmí, and Arghakhanchi districts is domínated by local variolies. Several f.cton; are responsible for low productívity and for other associated problems of maíz. production in Ihe area. lnitia!ly researchers perceived low yields assocíated with inferior local vanetíes as lhe main constrainl in maize produclion for lbe area. Based on pasl expenence and success in upgrading the productivity of local landraces through lbe introductíon of high-yielding varieties and subsequenl seed selechon, a breedíng program was formulaled in order lO address Ihe problem. The inít.al objective of the program was to increase fanners' access to new, improved genetic materials and provide them with trainíng on mass seleelion. However, a dífferenl seenario emerged during lhe site-selecrian survey and lhe process of settiog research goals. Farroers reported lhal maíze productíon in lhe area was aff.cled maínly by lodging problerns, Farmers in Ihe area have deveJoped and maínlaíned a variety called Thulo pínyalo Ihal produces good yields and a1so meets theirfodder requiremenls. However, the variety is prone lO severe lodgíng, resulting in yield losses of 15% lo 85%. Farrners lherefore strongly suggesled tha! ralher lban introducing new varieties, their local varieties be improved 10 address Ihe problem. In Ihis way, Ihe breeding program changed trom increasing grain yield to reducing lodging in Ihe target envíronrnenl. Thís paper discusses how farmers set their own breedíng goals and Ihe implícalions for methodological approaches ro participatory plan! breeding. Introduction Maize (Zea mays L.) is the second most importan! crop after rice in Nepal. It is grown largely on ban land (rainfed upland cornmonly associated with farm forestry) during summer and usually rotated with millet or beans. Maize is also grown as the sole crop at lower altitudes (below 1000 m) and at higher altitudes (above 1600 m). It is also grown in khet land (bunded land where at least one crop of puddled rice is cultivated) at altitudes below 1000 m during the spring season. Maize cultivation occupies nearly 0.8 million hectares (almost 30% of the total cultivated area), and 80% of this is under terraced hill fanning, producing over 1.3 millíon tones/annum (MoA 1995). The productivíty ofmaize is quite low (about 1.7 tonneslhectare), which is reflected by a high incidence offood-deficit households in the hills ofNepal. A number of factors appear to be involved in M, Subedi is a prograrnme officer (plant breeder), P.K. Shrestha is a programme officer (socioeconomist), S. Sunwar is an asst, plant breeder, and A. Subedi is executive direetor (lf Local Initiatives for Biodiversity Research and Development (LI-BIRD), Pok.hara, Nepal. . Thi, work was carried out in collaboration with fue CGIAR Program on Participalory Research and Gender Analysis (PRGA) for Teclmology Development and Insrirutional Innovation, convened by CIAT, Cali, Columbia. Farmers of fue Gulmí, Palpa, and Arghakhanchi districts wno parti<:ipated in the process in one way or another and provided information are duly acknowJedged. The contributions ofMr. D. Shanna, Dr. K.a. Koirala, Mr. c.a. Kunwar, and Mr, T. Rijal frem NMRP, Mr. T Shrestha from DADO Gulmi, and Mr. M. Chaudhary, Mr. B,B. Paudel, and Ms. N.K. Khatri frem Ll-B1RD in the site-selection process are híghly appreci.ted. Sincere acknowledgement ís also due lo Dr. a.R Stbapit and Mr, K.B. K.dayal for sharing the information, whích was very helpful in planníng and defining the processes. 311 Role o(Farmers ill Sellíng Bre!?... d,,,,·"g,,-G,,,{J~a,,,ls________ the low productivity ofmaize in the middle hílls ofNepal. These ínclude raínfed farming with uncertain rainfall, poor access to chemical fertilizers and declining application of organic manure, and lack ofvarietal options and access to improved genetic malerials suitable lo local conditions.ln areas where improved maize varieties have been introduced, farmers tend lo grow the same seed for a number of years without replacing it or without practicing standard seed-selection procedures. As a result, these varieties generally deteriorate rapidly due to genetic contamination with poorer heterogeneous landraces aml/or due lo unconscious selection for negative traits, as farmers generally use either grain for seed or seleet harvested cob for the seed. Practice ofselecting standing plants for the seed is rarely seen among the farmers. From the point ofview ofvarietal improvement, the problem ofmaize production in the hilly areas ofNepal is therefore threefold. First, farmers' access to new, improved germplasm is highly limited; second, the recommended varieties do no! mee! the multiple varielal needs of local farmers; and third, varietal deterioration occurs over time in the farmers' fields. To address these problems, Local Initiatives for Biodiversity Research and Development (LI-BlRO) is currently researching a farmer-Ied participatory plant-breeding (PPB) exercise in maize in the Gulmi district ofthe westem hills ofNepal. The maize-growing envirornnent of Gulmi has a unique geophysical envirornnent and represents the large hi1ly areas of the Palpa, GuImi, and Arghakhanchi districts extending towards Pyuthan and further west. The maize is grown in outward sloped terraces of bari land under raínfed conditions, with minimal external inputs (seeds, fertilizers, and plant-protection measures). Farmers in the area have poor access lO agricultural inpuIs, including improved genetic materials (Kadayat et al. 1998; Sthapit el al. 1997). Moreover, access 10 new sources of maize germplasm-thal closely matches farmer-preferred traits-in the traditional seed-supply system is limited. A survey ofpreferred trails carried out in 16 villages in the Gulmi district revealed that grain and fodder yield, aato (grit) recovery, taste in various cuisines, graín color, resistance to lodging, and time ofmaturity are the most cornmonly cited preferred traíts (Subedi and Shrestha, Unpublished; Kadayat et al. 1998). As a result, the major proportion of the maize area in the Palpa, Gulmi, and Arghakhanchi districts is planted to local varieties. The local varieties are the products of continuous seed selection carried out by farmers, consciously or unconsciously, over many generations and are well adapted to the local envirornnents and meet furmers' multiple needs. However, these varieties have a number of undesirable traits that require urgent attention in order lo ensure food security in the regíon. Researchers' perceptions of the problem LI-BIRO carried out a study to analyze the situation in the Gulmi and Arghakhanchi districts lo develop a future strategy for agriculture. Maíze was the most important crop; however, average productivity was reported to be low: below 1.5 tfha in both districts (Kadayat et al. 1998; Sthapit et al. 1997). This may be partly due lO a low supply of inputs in these districts, as the improved seed sold by Ale during 1996/97 was 1.22 mt in Gulmi and 0.91 mI in Arghakhanchi (Kadayat et al. 1998; Sthapit et al. 1997). Researchers concluded that the low maíze 'yields were due to poor access to new, improved genetic materials and deterioration offarmers' maintained variety because ofpoor seed-management practices (figure 1). In such a situation, providing farmers with improved maize varieties and seed-selection skills appeared to be a practical and sustainable solution. As a resuIl, helping farmers improve local maize varieties for yield-related traits became the goal of the programo 312 M Subedi, P.K. Shreslha. S Sunwar. and A. Subed. I Poor performance of maíze crop I Low yíeld potentíal of local varíety : I Selectíon practíces leadíng lo degeneralíon Figure 1. Researchers' perceptions ofthe causal relationships contributing to low productivity of maize in the mid-hill region of Gulmi and Arghakhanchi districts, Nepal However, A different scenario emerged duríng the selectíon survey for lhe research site and in research-planning discussions wilh farmers at lhe research sites. Farmers felt lhat poor production performañce was associated wilh the lodging of maize plants ralher than yield traits, lhemselves, in most commonly grown local maize varíeties. Farmers' perceptions of the problem An extensive reconnaissance survey was conducted in large areas of the Palpa, Gulmí, and Arghakhanchi distrícts during the process of selecting research sites for the project A rapid survey of28 villages was done, and farmers were consulted to verify the research problems in maize production and determine lhe suítability of these villages for implementation ofthe research programo Potential sites were screened and narrowed down to síx villages. Particípatory rural appraisal (PRA) and field observations were done by a multidisciplinary team in lhese villages. Discussions were held in the farming cornmunities during the site-selection process in order to colleet information about lhe geophysical condition ofthe area, socioeconomíc situation ofthe farming cornmunities, and farmers' interest ín ~ lhe proposed programo Problems were discussed with farmers in greater length during lhe survey. Preferred-trait analysis was done during the PRA to verify the researchable problems. Major traits of interest and problems associated with the preferred traits were identified in the process. Varietal traits o/ interest Varietal performance for the trait of interest was díscussed wilh farmers duríng lhe site-selectíon survey in order to understand farmers' needs and varíetal strengths and weaknesses in relation to a particular trait. Thís exercise was important in order to develop a breeding program based on needs and problems. In this process, ínformatíon on lhe desirable and undesirable characteristics ofbolh local and reeornmended ímproved varietíes was colleeted. Farmers were found to grow a number ofvarietíes (viz. Thulo pinyalo, Thulo seto, Sano pinyalo, Sano seto, Amrikane, Kaude, Rato dhanthe, Thorgeli pinyalo) to suit their growing environment 313 Role.. o! Farmers in Setting Breedíng Goals and to meet theír household needs. Thulo pinyalo is the mosl popular variety ofthe region and occupies as rnuch as 80% ofthe maize area in sorne villages. Farmers liked rnost ofits traits. This variety has good taste in all recipes, good grain and fodder yield, the biomass (both green and dried) is very rnuch líked by the livestock, and it is easy to sel! and barter because it has bold, fiint grain with an attractive grain color. However, farmers had lodging problems with this variety, leading to as much as 85% production 1055 in the worst season (table 1). Lodging problerns are equally high in other local varieties (viz. Thulo seto and Amrikane); however, the arca under these varieties is very low. It was reported that the low production of Thulo pinyalo has more significant implications for the food security ofthe region than any other variety. So, the lodging in Thulo pinyalo was considered a major problem. Resistance to lodging frorn thick stalks and strong, stout plants has been perceived by the farmers of the surveyed villages as the rnost desired characteristic in a recommended improved variety (table Table 1. Desirable and Undesirable Traits ofLocal Varieties ofMaize Grown in Surveyed ViIlages, 1999 Surve ed villages Parameters DIgam. Gulmi D/Devislhan. Gulmi Si Chaun arl, Palpa ¡ Banjha. Palpa Kaule. Arghakhanchl Desirabie traits High yleld potentlal High fodder yleld • • • • • • * • • * • Hígh ftour recovery Good taste Undesirable traits Lodging Head smut • 2). The least desired characteristics were a relatively low grain and fodder yield compared to that of large local varieties, followed by inferior taste. Low fodder yields have been found to be associated with the low height of improved maize varieties, compared to local varieties. Farmers of Banjha reported lhat al! fue improved varieties under cultivation in the village were introduced nearly six years before, and now there is no difference between local and ímproved, due to heavy and récurren! cross-fertilization with local varieties. F armers of the surveyed villages reported that high-yield potential and resistant to lodging were the most preferred traits for maíze, followed by good taste and high stover yield (table 3). Farmers perceived that graín yield is closely associated wíth the extent of lodging; they felt that these two parameters are highly interrelated and essentially synonymous. Farmers ofDarbar-Devisthan reported that lodging problems are due to tall plant height, and therefore, they perceived relatively shorter plant height as one of the mos! preferred traits to be considered in the maize improvement programo Revisiting farmers 10 discuss maize-production problems in the targeted area and to verify research hypotheses with farmers revealed that causal relationships in poor maize performance were no! properly established. Earlíer, a new research hypothesis surfaced, which explained Ihat the poor 314 M. Subedi. P.K Shrestha. S Sunwar. and A. Subedi Table 2. Desirable and Undesirable Traits oflmproved Varieties ofMaize Grown in Surveyed Villages, 1999 Surveyed vlllages Parameters Dígam, Gulmi D/Oevisthan, Gulml Simichaur, Gulmi Chaun bari, Palpa • • • • • Banjha, Palpa I Kaule. . : Arghakhanchl Desirable traits Nonlodginglthick stalks Eal1y ma!urity · S!rong Istout plan!s Undesirable Iral!s • • lodging Lowyield Low fodder yield • • • • . Inferior taste More iosee! problems : • • • · · ·• • I : Table 3. Ranking of Preferred Traits ol Maize in Surveyed ViIlages, 1999 Surveyed vlllages Traits Higher grain yield I D/Devisthan, . Simichaur, Chaun Gulml : bari, Palpa Gulml DIgam, Gulml 1 2 2 1 2 3 1 1 2 More stover yield 5 3 3 3 Need lor less soil lertility 6 5 5 4 4 dctlparGood tasle 2 : 4 1 Short plan! heigh! Good husk cover 3 4 White grain color Early maturity Kaule, Arghakhanchl 1 Nonlodging More grit recovery BanJha, Palpa i 5 i performance of maize in the area is not due to yield traits but to lodging tendencies, and this, in turno leads to poor production (figure 2). Redefining breeding goals In light ofthe new research hypothesis that emerged during the site-selection survey, a one-day village workshop was organized with the farmers at each research site selected for the implementatíon of the program. Farmers at the research siles opined that the local variety Thulo pinyalo has good yield and meets their requirements. They strongly suggested improvíng Thulo pinyalo for lodging resístance rather than just introducing new varieties. The underlying causes of lodging in Thulo 315 Role o[Farmers in Setlillg Breedillg Goals Poor performance of maize crop Loss of maize yíeld due lo lodging of local variety Selectíon practices leadíng to tall plants prone to lodgíng Figure 2. Farmers' perceptions of causal relationships for low productivity of maize in the mid-hiU region of Gulmi and Arghakhanchi districts, Nepal Pinyalo were explored with the farmers' group in order lo understand and tackle the problem. Farmers perceived the following as Ihe causes of the problem: • The very tall plant stature ofthis variety is the main reason for lodging. Farmers reported it having as high as 27 ¡eaves in one plant. In field observations, the plant height of Thulo pinyalo was found to be as high as 5.1 meters. Ear height has been found lo be more than two meters under good growth conditions. The weight of the tassel and eob al such a height contributes to the extensive lodging of the thin-stalked Thulo pinyalo, even under mild wind pressure. • Thulo pinyalo artains luxurious grov.ih in fertile land, which is one ofthe reasons for lodging. • Disease and insects attack the stem. • The lodging is greater after prolonged rainfall foIlowed by winds. Aceording lo farmers, they faee substantíal yield reductíons even with mild winds, as very weak plants lodge under such conditions and fall on other, nonlodging plants. This phenomenon oecurs in cycles and can affeet large areas. • The plants are more prone lO lodging during the lasselíng stage because of Ihe increased weighl al the top of the plant. • Yield is inversely related lo lodging. Yield los ses due to lodgíng in this variety are as high as 85% in the worst season. Thulo pinyalo produces more grain than high-yielding varieties (HYVs) in a normal season and less if there is a Jot of rain and wind. • Lodging is greater in wet areas al lower e1evations than in flat areas at the lop of the hills. • Lodging does not QCcur every year. However, there is no distinct partero. High winds during tasseling contribute to severity of the problem. Several possible options were discussed with the farmers lO achieve the goal. The options that could be implemented within Ihe project framework and which farmers considered possible lo imple- 316 M. Subedí, P.K. Silrestha, S. Sunwar, and A Subedí ment, considering their resources (time), knowlcdge, and skills, were chosen by the farmers' group. There were mainly three types of activities: a mass-selection program, a crossing program, and a participatory variety selection (PVS) programo Refining tbe research process The involvement of farmers in analysis of researchable problems helped change the researchers' perceptions ofthe problem (table 4) and redefine the goal oflhe maize-improvement programo The redefinirion ofthe breeding goals ofthe maize-improvement program provided guidelines for refining the research process !hat had been proposed initially. A multiple approach (mass selection, crossing, screening of improved/pipeline varieties, and PVS) was taken to address the problems, some of which had not been considered before, Farmers liked the mass-selection technique because they perceived it as a simple method and as a possible option to improve specific traits, keeping the desirable traits ofthe variety intact. The crossing program was chosen in consideration ofthe slow genetíc gain in fue mass-selection method, partícularly in farmers' fields, Considering the long gestation period ofthe variety-improvement program, which may delay the delivery ofbenefits to the farmers, the variety-selection program was planned. This would provide farmers with access to new, improved genetic materials to test in ruverse farming situations, A farmers' research committee was formed at each site in order lo empower farmers and to ensure farmers' leadership in the project. It was decided that the committee would be equally responsible for the planning, implementation, and mO!litoring ofproject activities. The committee works as an interface between farmers and researchers. It is expected that involving farmers in the planning and Table 4. Changes in Researchers' Perceptions after Involvement of Farmers in Prohlem Analysis Researchers' perceptlons Parameters Before farmers' Involvement After farmers' involvement Varietal intervenlion LQW Lowand limitad Landraces Low ylelding and inferior Despite good yield potenliaHow production due to lodglng . Problem Lodging Lowyiald Not known Yield loss: 15%-85% depending upon severity Extent 01 problem ----~------------------------~ : Tall plant and ear helght Contributing laclors Notknown 01 the problam • Thin stalk i Ethno-perception Nol known Varietal requirement HYV Objectlve Increase aecess to genetlc materíals Provide mass-selectlon tralnlng to farmers Wind pressure Locallandraces are well filled in different niehes : Widely grown Thulo pínyalo has ell goOO trelts : but prona to lodging Lodging-reslstant varlety . Improve Thulo pínyalo for lodglng reslstance Provide mess-selection tmlnlng lo farmers 317 implementatíon process will help in capacity building and increase ¡he farmers' sense of ownership in the programo Farmers are very supportive and cooperative in the project area. However, in some technical matters farmers' had different perceptions and altitudes, which changed along with the time. For exampie, farmers perceived that plants with short height could not produce good yields, that detasseling leads to total sterility in maíze, etc. In the beginning, Ihis made it difficult for researchers to facilitate some oflhe field activities, such as crossíng, demonstrating short-statured varietíes, etc. Later, the farmers found thal their perceptions were not correct, and their faith in the researchers increased, leading to better understanding, cooperation, and collaboratíon. Some farmers who were no! positive about the program in the begirming are the strongest members ofthe team now. Conclusions Involvement offarmers in the plarming process resulted in the development of more specific breeding objectives, which were more focused on the farmers' perceíved needs. It has helped to refine the context and process of the participatory plant-breedíng program and has gíven farmers a leading role in the decision-makíng process. Bibliography Kadayat, K.B., R.B. Rana, K.O. Joshi, A. Subedi, and B.R. Sthapit. 1998. Report on panieipatory variety seleerían un load crops in Gulmi district, Nepa/: Need assessment and future strategy Gu/mi Arghakhanchi Rural Deve/opment Projeet/LI-BIRD, Dee 1998. Pokhar., Nepal: Local Initiatives for Biodiversity Research and Developmento MoA. 1995. Agricultural statísties o/Nepal. Kathm.ndu, Nepal: Ministry of Agriculrure, HMGN. Sth.pit, B.R., K.B. Kaday.t, A. Vaidya., A. Subedi, K.O. Joshi, and R.B. Rana. 1997. Report on need assessment /or partieipatory variety se/eelioo 00/000 crops in Arghakhanchi district. Nepal. Gulmi Arghakhanchi Rural Development Project/LI-B1RD. May 1997. Pokhara, Nepal: Local Initiatives forBiodiversity Research .nd Developmen!. Subedi, M.•nd P,K. Shrestha. (Forthcoming). Site seleetion repon o/farmer-led partiCípalory moize breeding program/or the Middle HilIs o/Nepal, 1998. Pokhara, Nepal: Local lnillatlves fOT Biodiversity Research and Development. Jl8 Sensory Evaluation of Upland Rice Varieties witb Farmers: A Case Study R.K. Singh. K. Prasad. NP. Mandal, R.K. Singh. B, Courtois, vp, Singh. and T. Paris Abstract As part of a participalory planl·breeding project wilh melhodological objeclives lo improve rainfed rice in eastem India. an evalualion of sensory characteristics was conducled with farmers in a village of Bibar. Twenly·four farmers (12 women .nd 12 men) evalualed 15 upland rice varieties as raw rice and parboiled rice for milIed and eooked rice appearance, coloT, odor, texlUre, stickine.s, taste, .nd overall acceptabilily. The rice samples were milled and cooked by the women farmers following their ordinary practices, One variely recorded good resulls with bOlh raw and parboiled modes of preparation, The preferences ofwomen and men farmers díd nol differ significantly. The rankings based on preferences were compared with the rankings ofthe varietie. for varíous physico-chemical characteristics measured in Ihe laboratory, Most correl.tions were not significant, indícating Ihat, for Ibe sel of lesled varieties, lhese pararnelers were poor prediclors of farmers' preferences. Tho rankings based on preferenees were compared with fanners' ficld rankings, and lhecorrelation was positive for raw rice andnegative forparboiled rice. Fanners' trade-offbetween field performance and grain qualily is Iherefore importanl lO .5ses. for al leas! parboíled rice. The resul!s of Ihis first sensory evalualion experiment will be used to simplity Ihe methodology and to improve varietal evaluation in Ibe formal breeding process, Introduction In eastem India, rainfed rice represents a major component in thediet and income,of.millions of resource-poor people. In these harsh environments, the rate of adoption of modem rice varieties is Iow. Subsístence agriculture ís stilI quite important, although market integration is slowly progressing (Pingali 1997). In these transition systems, grain quality and taste strongly ínfluence the adoplion of modem vaneties. The maín source of vanation in grain qualíty ís the vanety, although envíronment and genotype-x-environment interactíons also affect grain quality. Different grain types, and therefore dífferent vaneties, are needed for self-consumptíon, market sale, and vanous preparatiollS or to pay wages in kínd. For plain rice, precooking practíces influence the vanetal choíces. Among the most common is parboíling, which is an age-old practíce in sorne regions of eastem India, where rice ís partly cooked before being air-dried and then sun-dried to improve íts nutritíonal, cookíng, and storage attributes, Preferences may vary across income levels, various social groups requiring vanous vaneties. Qualíty tests for breeding lines are routinely conducted by scientists in the laboratory. In the frame of a partícipatory plant-breeding project with methodologícal objectives started in ¡ 997 under the collaborative program wíth the Indian Council of Agricultural Research (ICAR) and the Internatíonal Rice Research Institute (IRRI) (Courtois et al. 1999), we developed a methodology to evaluate the grain quality ofrice vaneties in collaboration with fanners. To test the methodology, the R,K, Singh, K, Pra'ad, and N"P" Mandal ate al lhe Central Rainfed Up!and Rice Reseatch Station, Hazaríbagh, Jharkhand. India, RK. Singh, B, CourlOls, V,P, Singh, and T, París are wilh lhelmemationa! Rice Research Institute, Philippines (lRRI), B. Courtoi, was seconded te IRRl from the Centre de coopération intemationale en recherché agronomique pour le développement, France. The authors thank the Korahar fanners who actively and enthusiasticaHy panicipated in this study foc their contribution. Tbis projee! ¡. partly funded by lhe Inlemational Developmenl Research Centre (lDRC), Canada, and partly by me collaboratjng research institutions, lt is a component afthe System~W¡de lnitiative ofthe COlAR on Partlcipatory Research and Gender Analysis. coordinated by CIAT, and it benefitted of the overalI organization of the inittative 319 ~f?n§(}ry Eva/tia/ion o{ Uplalld Rice Varieties-"w"'it!!.".!-F:-"'arcem"'e;e:rs'--_ _ _ _ _ _ _ _ _ _ _ _ _ _ __ sensory evaluation of a set of upland rice varieties was organized in a village of eastem India. The objectives of this study were (l) lo document the process of rice preparation at the farm level for raw and parboiled rice, (2) to estimate the influence of the two modes of preparation on rice quality and identify the best varÍeties in each case, (3) to colleet informa1Íon about quality characteristics that determine varÍetal acceptability by female and male farmers, and (4) 10 relate the preferences with the physico-chemical properties of the varieties determined in laboratory. Materials and methods Fifieen modem upland rice varieties and a local check (Brown Gora, widely grown by upland farmers) were tested. The test was conducted in 1998 in the village ofthe Korahar dÍstrict ofHazaribagh, Bihar, India. These varieties had been prevÍously tested for their agronomic values in a participatory varietal tria! conducted in the same víllage (Courtois et al., submitted). Rawrice F or each variety, two kilos of sun-dried paddy of good quality were used. The paddy was dehulled and mílled using a dhenld, a big wooden bar moving up and down around an axis. The dhenld was operated by two women, one of them moving Ihe dhenld wÍIh her leg, the other shuffiíng the paddy grain afier every stroke of the dhenki. Al! Ihe varieties were dehulled and milled by Ihe same two persons under the same condítions. The times necessary for completion of dehulling and milling, and Ihe milling recovery (percentage of milled rice weight on rough rice weight) were recorded. The head rice recovery (unbroken grains) was not quantified but estimated visually (milled rice appearance ). Before cooking, one kilo of c1eaned rice was washed with water. Aluminum vessels called bhude/i were used to cook each variety separately. All bhude/i were ofthe same capacity. The women sug· gested using 3 liters of water to cook I kg of raw rice. The bhude/i wilh water was kept on the fire up to Ihe boiling point, when the washed rice was added. The cooking test was done by pressing the cooked rice between Ihumb and index finger. The same woman did the eooking test for all varie1Íes. The cooking time of each variety was recorded. The excess water was drained and Ihe cooked rice was displayed on a pattal (leaf mat) for sensory evaluation. ParbQiled rice As decided by the women, 2.5 kg of paddy were soaked in 3 liters of water in a tin container for 18 houis. A common belief is that the soaking of paddy should be done in the evening rather than during daytime, wilh the excess water drained in the moming, to avoid Ihe heat oflhe day. A temperature Ihat is too high would induce Ihe soaked paddy to ferment, leading to poor rice quality, high breakage, and bad odor (Bhattacharya 1985). The soaking ofpaddy in water startedat 4:00 p.m. and the water was drained al 10:00 am the next day. Afier decanting Ihe water, the soaked paddy was steamed on Ihe fire. During Ihe steamÍng process, the tin containing the soaked paddy was covered with a gunny bag to avoid loss ofheat. When Ihe husks of the paddy started cracking, Ihe container was taken off the fire. The steamed paddy was spread in the shade on a mud floor for drying. The paddy was dried in the shade for 48 hours wilh intermittent mixing. It was then exposed lo Ihe sun for complete drying. An indigenous technique was used to test the proper drying ofpaddy. Twenty lo 30 grains ofpaddy were dropped on a hard floor. The graíns were crushed underfoot by rotating Ihe heet If this removed Ihe grain husk, Ihe rice was considered to be well dried and ready for 320 RK Singh el al. dehulling. For dehullíng and milling, 2 kg of c1eaned paddy were used and the same process as for raw rice was followed. More water is needed to cook parboiled rice than lo cook raw rice. The women suggested adding 7 liters ofwater to cook 1 kg ofparboiled rice. For the subsequent operations, the same process was followed as for raw rice. Sensorial evaluation A protoco! for lhe practica! organization ofthe sensory evaluation was desígned following the recommendations of Arnerine, Pangborn, and Roessler (1965) and Del Mundo (1991) and adapting them to the realities of an eastern lndian village. Twenty-four farmers (12 women and 12 men) particípated in Ihe sensory evaluation, A hedonic scale was used. The farmers were asked lo indicate whetherthey líked (score 1) ordisliked(score O) the varieties for mílled grain appearance, cooked rice appearance, odor, color, texture (softlhard), stickiness, laste, and overall acceptability. The samples were numbered and randomized to límit the "first-sample bias." The raw rice and parboiled rice were evaluated on different days to limil the teslers' fatigue. Physico-chemical characterization o/ the samples under laboratory conditions The tests were perforrned at the technology laboratory Gf the Central Rice Research Institute, Cuttack, India, for raw rice and in N.D. University of Agriculture and Technology, Masodha, Faizabad, India, for parboiled rice, The parameters measured for raw rice were milling recovery, head rice recovery, grain length and width, alkali value, volurne-expansion ratio, kernel-elongation ratio, and amylase content. For parboiled rice, hulJing and milling recovery and grain shape were measured. StatisticaJ analysis Forrank comparison, Spearrnan's coefficient of correlatíon was used when only two rankings were compared. A Kendall coefficient of concordance was used, as described in Siegel (1956), when more than two rankers were involved. The mean comparisons were perforrned using a Student's t-test. Results and discussion Mil/ing No difference between the two modes ofpreparation was observed for mílling time (table 1). Raw rice took significantly less time to cook as compared to parboiled rice. Milling recoverywas significantly higher for parboíled rice in comparison to raw rice. There was no significant difference between farmers' practices and laboratory method for raw rice but recovery was higher with farrners' practices for parboiled rice. The lower coefficients of variation in the case of parboiled rice índicated a buffering effect ofparboiling across varieties for recovery, which explains why parboiling is considered an excellent means to recover poor-qualíty samples. Sensory evaluation The method of rice preparation had a great impact on the ranking ofthe rice varieties for aH traits, as shown by the nonsignificant and sometimes negative rank correlations between the two seis of 321 Sensory Evaluation o[Upland Rice Varieties with Farmers Table 1. Comparison of the Milling Properties and Cooking Time of Raw and Parboiled U pland Varieties Prepared by Farmers, Korahar, Bihar, India, 1998 Recovery farmers' practicas ('lo) Milling time (minutes) Varlely R_ ParboiJed Brown Gara 19 RR139-1 16 Recovery laboratory ('lo) Raw Parboiled R.w 30 70 71 17 63 77 Cooking time farmers' practicas (minute.) Parbolled Raw 58.5 75.0 11.0 23.0 62.3 80.0 8.5 33.5 Parboiled RR151-3 18 19 69 75.0 10.0 17.0 22 19 57 72 74 67.3 RR1514 67.5 80.0 8.0 20.5 RR166-645 15 23 65 74 59.5 76.3 11.0 23.0 RR203-16 15 17 63 73 56.0 76.3 11.0 22.0 RR2-6 27 18 70 72 60.5 76.3 13.0 33.0 22.0 RR265-1 20 15 70 72 76.5 77.5 8.5 RR347-166 20 17 66 74 73.5 7.38 15.5 23.0 RR348-5 30 17 71 72 66.3 78.8 9.0 230 RR348-7 13 16 69 74 51.0 77.8 13.0 32.0 RR352-1 16 24 66 72 64.0 76.3 14.0 27.0 RR354-1 20 23 59 75 69.8 77.5 16.0 29.0 RR50-5 18 20 67 71 67.8 80.0 13.0 34.0 RR5H 19 18 66 72 58.8 75.Q 10.0 26.0 Vandana 17 19 74 70 72.0 76.3 13.5 25.0 Mean 19.1 19.5 66.6 12.8 64.4 77.0 11.4 25.8 SD 4.4 3.8 4.5 1.8 8.9 1.9 2.5 5.1 t raw/parboiled Note: ** = 7.11** 0.28ns signifieant at (be 1% leve1; ns ~ 4.29 H 12.04" flot significant seores (table 2). The preferred varieties in tenns of aeceptabilíty were RRI51-3, RR352-1, and RR354-1 for raw rice, and RR50-5, RR352-1, and RR354-1 for parboiled rice. For breeding purposes, it was interesting to identifY varieties that could perfonn well under both preparations. RR352-1 and RR354-1 scored quite well in this respect. The farmers were also asked to indicate the fOUT varieties they liked the mosl (high seore indiealed high preferenee) and the fOUT varieties they liked the least (this time high seores indicated high dislíke). By this means, only one variety, RR354-1 recorded a good seore for both raw and parboiled rice (table 3), being liked by 67% ofthe farmers as parboiled rice and 58% ofthe fanners as raw rice. RR151-3 and RR352-1 were apprecíated by the farmers as raw rice but not as parboíled rice. Inversely, RR2-6, RR I 66-645 , and RR265-1 were líked by the farmers as parboíled rice but not as raw rice. For raw riee as well as parboiled rice, the rank correlatíons among characteristícs scored by funners were very strong and posítive (table 4) except for stickiness, for which they were also positive but more seldom significant. This means that there is probably no need ID ask the fanners to seore aH these traits. The aeceptability or the choice of the three or four most preferred varieties should be enough to represent the group of traits. A simplification of the testing procedure an important in order to facilítate the integratíon of partícípatory approaches ín the fonual breeding system and tD sustaín fanners' participatíon. 322 Table 2. Sum of Seo res Given by 24 Farmers for Cooking Varlety MiUed rica appearance Raw BrowoGora RR139-1 RR151-3 RR1514 RR166-645 RR203-16 RR2-6 RR265-1 RR347-166 RR348-5 RR348-7 RR352-1 RR354·1 RR50-5 1 4 17 17 4 8 8 19 21 1 22 12 21 RR51-' 9 Vandana 12 Rank oorref, apeéaranCa Ra", 5 18 20 16 11 13 9 13 20 19 11 10 1 16 11 6 13 18 6 22 13 10 23 16 7 2 7 20 15 14 11 15 -0.12 0.10 := Par. 13 19 9 19 18 14 21 18 17 20 16 21 24 24 14 9 T.stet Textura Cooked rice Par. Notes: Par. == ParboHed rice; Accept. Characteristics of Upland Rice Varietles, Korahar, Bihar, India, 1998 Odar Raw Par. 2 11 13 18 18 10 12 23 11 17 9 16 8 23 10 16 11 17 17 13 6 16 20 20 14 23 21 15 12 15 0.12 12 11 Color Raw Par. 2 7 21 17 9 20 17 19 6 8 14 15 12 23 13 21 17 21 20 16 19 4 18 20 23 19 22 15 10 16 19 5 ¡soltlhard} Raw Par. 10 2 13 17 18 8 13 15 12 10 5 12 6 17 12 18 14 14 15 16 5 14 17 17 16 20 10 22 7 13 16 8 Stlckln ••• Raw Par. 8 12 11 20 16 10 10 22 9 13 15 8 8 19 9 19 12 16 15 15 11 13 16 21 7 22 11 19 13 15 11 8 0.20 0.06 -0.19 flavor Accept. Raw 10 15 Par. Ra.. Par. 13 20 4 9 9 18 11 13 9 13 14 18 16 15 20 16 13 14 14 20 22 Z2 11 12 18 16 9 13 16 17 17 e 21 18 14 13 12 0.26 9 8 9 12 15 6 17 17 14 11 12 6 19 18 11 20 18 12 16 13 22 24 21 13 6 0.23 Acceptabílíty; Varieties with high scores a.re the prefcrred ones. :.:. >: ~ w N w '" ",... ~ Sensory Evalu"tion oI Up/and Rice Varieties \Vil/¡ Farmers Table 3. Preferences of Farmers for the Various Varieties in the Sensory Evaluation Conducted in Korahar, Bihar, India, 1998 Variety Most liked* Raw Brown Gora o RR139-1 Least fiked'" Parbolled Raw Parboiled O 13 5 4 O 2 RR151-3 16 RR151-4 8 O 6 2 RR166-645 1 10 10 RRZ03-16 3 9 4 o RRZ-6 1 11 14 O RR265-1 2 10 3 O RR347-166 8 2 RR348-5 7 6 2 3 RR348-7 3 2 15 11 RR352-1 14 O RR354-1 14 16 2 1 RR50-5 4 4 4 4 8 15 2 7 7 RR51-1 3 3 3 Vandana o o 13 *Farmers were asked tú give the codes ofthe sorne of them gave only 1 oc 2 s\:ores. Table 4. fOUT vaneties they ljked most and the four variettes they hked lease, Howevef, Correlations between Farmers' Ranks for Quality Traits of Raw and Parboiled Upland Rice Varieties (Women's and Men's Rankings Pooled Together), Korahar, Bihar, India, 1998 Trait Millad rice app. Milled rice app. I Color Texture Stickiness Taste Accoplability Odar Cojor Textura Taste Slickine"" ! Raw PatboIled Cooked rice app. : Raw Odar , Cooked rice app. 0.59' I Partdled OA55+ Raw Parboiled O.58*" O.as" Raw Parboiled 0.60'" 0.60'" 0,84- Raw Parboí1ed 0.46 0.76" 0.50'" 0.87""* 0,12.-0 I O.SS"" 0,87·· ! O.asO,BS·'" I 0.80·· 0.87"" 0,85" 0.45 0.66"" 0.29 0.48 0.S7"" O.7'f·... 0.72"* 0.83" 0.72** 0.82" 0.71" 0.72*0.74" 0.79"* 0.75"" 0.750.71"" Raw Parboíled 0.18 0.47 0.29 0.62' Raw oas" Parboíled 0.53" Raw Parboíled 0.67* 0.S1·· 0.52 I 0.81- ! I o.ar 0"83""* 0.20 ¡ ! 0.52" 0.28 0,39 0.23 0,39 0.90"' 0.91 Opinions of women and men fanners were similar, with significant to highly significant correlations between their rankings for milled rice appearance, cooked rice appearance, texture, color, and taste (table 5). The on1y traite for which their agreement was weaker was stickiness ami, to lower 324 RK Singh Table 5. el al. Correlations between Women and Men Farmers' Mean Ranks for Cooking Characteristies of Raw Rice, Korahar, Bihar, India, 1998 Trail Spearman rank coefficient of correlalian Milled rice appearance 0.97" Cooked rice appearance 0.57" Odar 0.45 Color 0.75" Texture 0.55* Stickiness 0.22 T aste/flavor 0.54' Acceptability 0.83** Most liked 0.88" Least liked 0.95" Note; Sample size was 12 women and 12 meno .. =; Significant at 5% leve!. *'" Significant at 1%. extent, odor. In terrns of overall acceptability, there was no difference in women and men farmers' opinions on the tested varieties nor in their final choices of the varieties they liked most and leasl. Laboratory analysis versus sensory evaluation The ranks given by farmers for the various quality traits were compared with the ranks ofthe same varieties for the main chemical properties of raw rice measured in the laboratory: alkali value, volume expansion, amylase content, and eIongation ratio. Elongation ability was negatively correlated with stickiness r -0.55, significant at the 5% leve!) but that was the only significant case. In the samples tested, amylase conten! did not seem to have any link to farmers preferences for texture r = -0.14) or stiekiness r = 0.04). It is unexpected to see so few relationships between consumer preferences and measurable chemical properties, since these are standard parameters used by all chemistry laboratories. However, for the varieties inc1uded in the evaluation, the variability for sorne traits was limited and therefore consumers had difficulty assessing differences. Field performance versus grain quality There was little relationship between farmers' field ranking and grain quality for parboiled rice, as shown by the very low coefficients of correlation for rank and a negative one for the ranking based on yieJd (tabIe 6). The relationship was stronger and positive for raw rice. Thete was no particular reason why the rankings should be correlated, but a strong negative correlation would complicate the breeding work. These results confirm tha! participatory varietal selection should not stop afier harvest. Sinee a compromise might be necessary, at least for parboiled rice, the trade-offbetween eritena for agronomic performance and cooking quality applied by farmers has to be assessed. 325 Table 6. Correlation Between Field Ranking and Yield, and Farmers Preferences based nn Grain Quality, Korahar, Bihar, Indía, 1998 rankeng Ranks based on obselVad yietd (1) (2) Raw Parboiled Raw Parboiled 10,5 15,5 14,0 16,0 14,0 7,5 14,0 9,0 8,5 15,5 Farmers fleld Variety Most liked • Acceptabilily - Brown Gora 15,0 RR139-1 12.0 RR151-3 4,0 16 2,0 1,0 14,0 1,0 RR151-4 2,0 10,5 4,5 6,5 4,0 5,0 RRI66-hrestha. M. Suhedi. D. POI/del. and S Sunwar Table 3. Gender Roles in Maize Production and Utilization (Percentage Time Contribution) Activities Male Female Children A. Maíze productlon activities l. Seed preparation (shemng cobs, dryíng and storage) 24.8 61.1 14.1 2. Canry composVFYM lo the field 17.4 63.5 19.1 3. Land preparanon 54.6 36.5 8.7 4. Seed sowing 11.7 42.8 45.5 5. Fíeld supervision for seed germination 43.4 52.1 4.5 6. Weeding and earthing up maize crop (ftrst) 34.1 49.4 165 7. Weedíng and earthing up ot maize crop (second) 41.0 54.5 4.5 8. Intercrop sowíng of bellOS, cowpeas, pumpkin etc. 11.9 74.3 13.7 9. Relay transplanting 01 fingermíllet in maiza fiald 308 56.0 13.2 lO. Fiald supervision 01 IOOging 01 maiza plants 41.3 52.1 6.6 11, Harvastíng and transporting 35,9 SO.3 13.8 12. Making bundles 01 maiza stover and transportlng 53,9 39.5 6.6 33,8 45.5 20.7 33.5 52.1 14.4 13. Processing (khostyaneljhuto pame) and starage 01 cobs Total B. Consumption and marketing actlvlties l. Shelling cobs 24.2 57.1 18,7 2. Processing (cieaning and drylng) grains lor milling 15.3 76.6 8.1 3. Canrying gralns to processing milis 27,2 52,0 20.7 4, Canrying grains to market Ior sellíog 49.7 50,3 5. Purchase 55.1 443 0.6 34.1 56.2 9.7 Total C. seed management activltles 1, Selection 01 cobs for seed 37.3 57.1 5.6 2, Shelling gralns trom Ihe selecled cobs 31.1 52.4 16.5 3, Seed processing and Irealment (cleaning, drying and lrealment) and seed storage' 21.7 74.4 3.9 4. Preparing storage potlstructure for seed slorage 26.3 72.5 1.2 29.1 64.1 6.8 Total marketing and seed management The gender analysis thus suggests that women have important roles and a stake in the varietal-improvement programs designed to develop farmers' preferred varieties. Their particípation in the whole process of variety development should be ensured and properly utilized. Distribution of breeding knowledge Particípatoryplant breeding seeks to use the knowledge and experiences farmers have accwnulated over generations. It al so creates an environment for mutual learning and sharing, which closes the knowledge gap and sets the stage for a working partnership between the farmers and researchers. 337 {'2~Qrporation Table 4. al Users' and Gender Perspective in Farm Gender Differences in Decision Making in Maize Production and Utilization (Percentage Contribution in Decision Making) Actlvitles Male Female A. Maize productlon actlvities 1. Seleclion of maize variety for nex! .eason planting 49.2 50.8 2. Seleclion 01 land selection according lo Ihe variety 46.1 53.9 3. Daleltime 01 sowing 51.5 48.5 4. Selec!ion 01 erops lor in!ercroppíng with maize 27.0 73.0 5. Dale/time 01 weedíng and earthing up 01 maize 36.2 63.8 6. Dale/lime 01 maize harvest 44.6 55.4 42.4 57.6 Tolal B. Consumption and marketing aclivities 1. When and how much graios lo shell 30.6 69.4 2. Quantity 01 grilslflour lo be milled al a lime 23.2 76.8 3. When lo carry maíze grains lo the mili (tor milling) 27.6 72.4 4. Food ítems lo be cooked daily 33.0 67.0 5. Whelher lo sale maize or nol 44.8 55.2 6. Quantity 01 maize grains to sold 37.7 62.3 7. Whether lO purchase maize or no! 41.5 58.5 8. Quanlity 01 maize grains lo purchased 36.1 63.9 Total C. Seed management actlvllies 1. Selection of maize varleties lor nex! season 46.2 53.8 2. Quan!ity of seeds of dlfferenl varie!les for nex! season 39.9 60.1 3. Wayslmelhods 01 storing seed 35.3 64.7 4. Number 01 sun-drying 01 stored saeds and using other !realmenls 30.7 69.3 5. Whe!her lo change old seeds or nol 48.0 52.0 6. Type and quantity 01 seeds 01 new variety to be planled 48.8 51.2 7. Giving sell-produced saeds tó other larmers 36.1 63.9 36.3 Facilitating and supporting farmers in their plant-hreeding activities then becomes easy and smooth. Based on this understanding, farmers' breeding knowledge was assessed by surveying a sample of 113 households selected randomly. An analysis of the influence of gender, wealth, and ethnicity on the distribution of such knowledge was also done and ís presented in table 5. Ibe majority ofthe households (more than 90%) separate seed and graín in advance, but the seed selection is almost entírely done from the cobs, and generally righ! after the barvest Farmers virtually do no! practice seed selection on standing crops. Ibe majority of the households select big, good-Iooking cobs with big, bold grains for seed. Similarly, almos! all farmers follow tbe practice of discarding grains on the tips ofthe cob when the cobs are shelled for seed. Only about a quarter of the farmers are knowledgeable about the role ofseed replacement in maintaining varietal purity and vigor. Farmers' knowledge on the more technical side ofbreeding, such as identification ofmale 338 Table 5. Distribution oC Breeding Knowledge by Gender, WeaIth and Ethnicity (% Households) - - - - - ----------- Characteristics Separate seed and graio in advance ., On standing Q, Immediately after harvest AlI 96,2 1---- . Male Famale - - - - - ------------ ------------------ ------Elhnl. cat.lIorio. Weallh .alellori •• Gender cateaori&s Rich Average Poor BCJ GMN KDS 94,1 97,2 90,0 91,7 0,0 1,0 0,0 91,0 939 97.7 94,1 0,1 10,0 0,0 0,0 1,0 0,0 100.0 0,8 96,0 1,0 97,0 44,0 32,0 3,8 10,8 8,0 12,0 0,0 0,0 1,0 0,0 0.0 1,0 0,0 63.6 75,1 67,6 30,0 26.0 32,0 7,0 7,0 32,0 32.0 31,0 72,0 79,4 5,0 8,0 9,0 30,3 18,0 17_0 10,0 12,0 8,0 43,0 1,0 2,0 14.0 37,0 4,0 4,0 Stage 01 .eed sele<:tion ClOp e, From stored cobs Basis of cob selection far seed a, Cobs with big and bold grains 67,2 b, Big and good looklng cab. e, Matured cobs 83,6 36,0 d, Heallhy robs wllhoulln.ect and dísease damage 35,2 32,3 47,0 38,2 d, Cobs no! damaged by bíros and rodents 1,6 1,0 2,9 0,0 2,0 0,0 2.0 - t Uniform grajo colour 0,0 4,0 2,9 1,0 0.0 0,0 3,0 2,0 Practice of discard grains on either tips of the cob while selecting seeds 97,7 98,0 97,0 95,3 97.1 100.0 98.1 100 91,7 Knowledge about the need for seed re~ placement to maintain varietal purity and vigour 24,2 24,0 25,0 27,S 23,5 13,5 28,0 0,0 8,3 Knowledge about mate and female maize f10wer a, Male flower 6,0 8,0 0,0 8,7 3,0 0,0 7,2 0,0 b, Female f10wer 6,0 8,0 0,0 8.7 3,0 0,0 7,2 0,0 0.0 0,0 12,0 13.1 9,0 17,1 5,7 12,6 0,0 16,7 0,0 9,0 0,0 16.7 2,8 10,0 ·1.0 16.7 Knowledge about lhe use of tassel and silk a, b, Tassel SlIk Knowledge abeut the reason of variatal mixture No/e: Ethnicíty is represeoled as BeJ 9,0 11,1 3,0 11.4 6.7 6,7 10,5 13,1 3,0 14,3 6,7 = BrahminlChhetri/Jogi; GMN = Guruog/Magar/Newar; KDS = Kami/DamaiISarki, - - - - ------------ lncorparation al Users' and Gender_FerspecthJe in Fan and femate plants and theír functions, was found te be very pOOL Similarly, a majority ofthe farmers also do not know the actual mechanism tha! causes new maize varieties to rapidly deteriorate, compared to other cereal crops like rice and wheat. The survey thus revealed that there is good scope and a need for sharing scientífic breedíng knowledge prior to the inception of a partícipatory plant breedíng program in order to enhance farmers' confidenee and thereby inerease theír ínterest and participation, Incorporation of the users' perspective in the research process Considerations made in the research process The project on fanner-Ied participatory plant breeding of maize has just completed one season of work. A number of consíderations have be en made, as suggested by the analysis of the users' and genderperspeetive ofmaize produetion and utilization. These are briefly discussed below. Breeding objective and selection olbreedíng maleríals The breeding objective has been redefined to ímprove the production performance of a widely grown maíze variety, Thulo pyanlo, rather than creating a large díversity of maize varieties in order to improve the productivity ofthe niche envíronment. This variety has all the traits preferred by the farmers except one, i.e., lodging resistance, Reducing lodging in this variety is now the maín objective of the breeding program. In addition, the selection of improved maize varieties to be used as one ofthe parents for crossing with Thulo pyanlo was done in a way that ensured that they met most ofthe farmers' preferences for different traits, These included relatively taller, stout plant varieties like Ganesh I and 2, Rampur composit, Rampur 1, Khumal yellow, and Pop 22. This would help to combine good traits from a large number of varieties into a few fanners' preferred maize varieties. At the same time, attention has also been gíven to meeting the specific needs of the niche environment through a participatory variety-selection program, which provides farmers with a choice from a large number of maize varieties. Selection 01research larmas Farmers have fonned their own research cornmíttee at both the research sites to ensure their partieipation in and influence on the ~esearch process, These research cornmittees are well represented by different categories offarmers and 41% ofits members are women, The Farmers' Research Committee. in consultation with the farmers at large, decide the breeding objectives and the research process. They also select research farmers to participate in the farmer-led maize breeding prograrns implemented at the research sites, Since farmers themselves select research farmers, it is envísaged that this wílllead to the development of maize varieties preferred by a large number of fanners. Similarly, under participalory variety-selection program, care is taken to distribute the seed of new maize varieties to different categories of frmners. Selection 01 trainees and contents Based on the findings ofthe survey on the distribution of maize-breeding knowledge among farmers, field-based training was provided lo the research farmers in order lo supplement farmers' knowledge with practical scientific breeding knowledge, Attentíon was given to representation of different categories of farmers, inc\uding women. Forty-five percent of the total trainees were women. This consideration will also be made in future farmers' training programs. 340 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _-'-P-".K"'.c;;S"'hc.;re::;s=tha, M. Subedi, D. Paudel, and S Sunwar Collection and analysis of users' and gender-differentiated data The initial survey indicated that farmers use multiple eritería for the selection of a particular maize variety. Farmers may give different weights to these eritería to suite their individual needs and resources. Wíth this in mind, the colleetion and analysis of users' and gender-differentiated data have been built into the research process to ensure Ihat users' and gender perspectives are incorporated into the partícipatory breedíng programo Data are collected in a form that allows users' and gender-differentiated data to be anaIyzed, which will facilitate the drawing of inferences about whether users' and gender differences make a significant difference in the process and product of participatory plant breeding in open-pollinated crops like maize. Conclusion The users' and gender analysis indicates tha! the differences among maize-growing households in regard to wealth, ethnicity, and gender do not have any significant influence on their choices for dífferent maize varieties. Similarly, farmers across aH wealth, ethnic, and gender categories grow only one 10 two maize varieties per household; therefore, their varietal needs are not very diverse. This is contradictory to what has been found in the case of self-pollínated crops. This appears to be largely because a large number of varieties is díffieult to maintain and manage in open-pollinated crops like maize. Farmers, however, use multiple cnterí,a in selecting the maize varieties they grow and prefer to have as many traits oftheÍr preference as possible in one to two varieties. It is, therefore, important for the particípatory breeding program to focus on developing fewer maize varieties with tbe multiple traits that farmers prefer. Women farmers have strong preferences about the quantity and qualíty of the fodder by-products of maíze and the suitability of new maize varíeties for intercropping with legumes. The research process should allow farmers of different categoríes to use their eritería in developing and selecting new maize vaneties, Farmers of a!l categories generally lack adequate practical breeding knowledge, and they are specifieally poor in scientific reasoníng, regardless of whatever breedíng knowledge they have. Supplementing farmers' knowledge with practica! scientific breeding knowledge is, therefore, necessary to empower farmers to sustain theÍr breedíng ínítiatives. References me Acharya, M. and L. Bennet!, L. 1981. The rural women ofNepal: An aggregate analysís of8 village ,lUdíes. In status ofwomen in Nepal, Vol 9. PartlL Katbmandu, Nepal: Centre for Economíc Development and Administration. Bajracharya, R 1994. Gender issues in Nepali agriculture: A review. Research Report No. 25. Katbmandu, Nepal: Minístry of Agriculture (HMG)lWínrock lntemational. CBS. 1999. Statistical yearbook of Nepal. Katbmandu, Nepal: Central Bureau of Statistics, National Planning Commissíon Secretariat. Cromwell, E. and S. Wiggins, wíth S, WentzeL 1993, Sowing beyond the sta/e: NGOs ami seed supply in developing countries. London: Overseas Developmen! Institute. J3rvis, D., T. Hodgkin, P. Eyzaguirre, G. Ayad, B. Sth.pit, .nd L. Guarino. 1997. Fanoer selechon, natural seleerion and crop genetic diversity: The need for a basie dataset. In Strengthening the scientific bosis ofin situ conserva- /ion ofagricultura! biodiversity on-farm. Proceedíng oJa workshop to develop tools and procedures for in situ conservolion on-Jorm. 25-29 August, 1997, edited by D. Jarvis and T. Hodgkin. Rome: Inlemational Plan! Genetic Resources Institute. 341 Incorpora/ion o(Users' and Gell~~!.:..r~~~pective in Farmer-Led Particípatorv pralll Breeding in lvfaize Joshi, K.O. 1995. Seed regulatory framewor/(s in Nepal: Partícipa/ory aad o/her al/erna/ive approaches lo seed productioo aod distribution in Nepal. London: Overseas Development Institute and Centre for Arid Zone Studies. Rana, K. and K. B. Kadayat. 1999. Report on haseline slady ofP VS sites, Mahottari. Pokhara, Nepal: Local Initiatives for Biodiversity, Researcb and Developmen!. Shrestba, P.K. 1998. Gene, gender and generation: Ro!e of tradiliona! seed supp!y systems in the majotenance of agrobiodtversity in Nepal. In Managing agrobiodiversity: Farmers' changing perspectives aad Instllu/Ional responses in Ihe Hindu Kush-Himalayan Region, edited by Pratap and B. Sthapit. Kathmandu, Nepal: IntemationaI Centre for Integrated Mountain Development and lntem.tíonal Plant Genetíc Resources Institute. 342 Understanding Farmers' Selection Criteria for Rice Varieties: A Case in Madhya Pradesh, Eastern India R.K Sahll, v.N. Sahll, M.L. Sharrna, T. Paris, K McAllister, R.K Singh, and S. Sarkarung Abstract This paper presents infonnation from a participatory breeding project ¡niti.red in 1997 at the Internatíonal Rice Researeh Institute (IRRI) ín collaboratíon wíth plant breedees and social scientists from six national agricultural research institut;ons loeated in eastem India. The Indita Gandhí Agricultural Uníveesíty (IGAU) ín Raípur, Madhya Pradesh, ís one of the collaborating eentees. The informalion gíven here is based on a sample survey of75 riee-farmíng households in Ihree villages oflhe Raipurdistrict, Madhya Pradesh. Surveys were conducted lO charaelerize fanners' croppínglfanning syslems, rice varietal diversity, dogre. of market orientation, gender roles, as well as socioeconom;c differences, and lo relate these to farmers' rice varietal preferences. The foeus is on methodologíes for improving understanding of fanners' (including worneo farmer's) eriteria for seleeting specific rice v.rieties and how !hese criteria are considered io particípatory breeding strategies for raíofed lowland conditions in Madhya Pradesh, eastem India. Introduction Rice is the principal crop grown during the wet season (June-October) and i5 the staple food in Madhya Pradesh, eastero India. In this regíon, rice is cultivated on 5.35 million hectares, wíth an annual production of 6.46 millíon tons. This state contributes 9% to the national production from 12.8% of ¡he national acreage. Eastem Madhya Pradesh, k:nown as Chhattisgarh 18 considered the rice bowl ofthe state. Ofthe total rice area, 80% is rainfed, and drought, which occurs every two years, i5 a major constraint ín íncreasing rice productívity in the regíon. The rice yield in the regíon ís low (abon! 2.3 tons per hectare) and ís below the national average. Because of the frequent droughts, the majority offarmers are not willing to risk investíng in farm inputs to inerease productívity. Sustainabílity and yield stabilíty are the most important considerations of farmers in the management of their farming systems. Rural poverty still persísts in this regíon, and about one-thírd of the total poor in Madhya Pradesh depend on rice production as the basic source of Iivelihood. Thel'efore, improving rice pl'oduction and productivity could directIy lead to a substantial reduction in the rural poverty in the regíon (Janiah et al. 20(0). F or the last four decades, a total of 512 modero rice varieties have been released in Indía. Howevel', hardly 10 to 20 ofthe released varieties are in the seed-productíon channel. For example, the average age of cultivars for which there i5 a demand fol' breeder seed is 11 years. The average age of cultivars in certified seed production ranges fiom 12 to 17 years in the states of Gujarat, Madhya Pradesh, and Rajasthan (Virk, Packwood, and Witcombe 1996). Only a few modero varieties have been successfu11y adopted in the irrigated ecosystem. RX. Sahu, V.N. Sahu, M.L. Sharma, T. Pati" K. MeAlIister, R.K. Singh, and S. S.rkaruog are scientists from lndira Gandhi Agricultural University (IGAU), Raipur, Madhya Pradesh; !he IntemaMnal Rice Research InstItute (IRRl). Los Baños, Philippines; and Ihe IRRl-Delhi offic., New Delhi, lndi •. The aUlhor, wish 10 acknowledge Ihe cornments ofDr. A.S.RAS. Sastri, Associate Director cfRescarch, IGAU .•nd Dr. Y.P. Singh, Coordinatorofthe Rainfed Lowland Rice Research Project in Eastem India, IRRJ. 343 Understanding Farmers) Selection Criteria [or Rice Varieties One of the main reasons for low adoption of released varieties in the rainfed environments is lhat farmers have inadequate exposure to new cultivars. If adoption rates are to be improved, farmers need to try a wide range of novel cultivars in their fields in partícipatory varietal-selection (PVS) programs. The cultívars should include prereleased cultivars, advanced hnes, and already released cultivars from other regíons or countries (Whitcombe et al. 1996). This would give farmers a 'basket of choices' of varied genetic material (Chambers 1989). Another reason for low adoption of modern varietíes is that the breeding process does not meet fanners' diverse needs. Released rice varieties are not suited to the complex and heterogeneous rainfed agroecologícal environment or to the diverse uses and needs of dífferent socioeconomíc groups of fanners. In Uttar Pradesh, India, Maurya et al. (1988) tested advanced Hnes of rice in villages and successfully identified superior material that was preferred by fanners. Understanding farmers' preferences and needs is crucial for successful adoption and dissemination of improved rice cultivars. In 1997, a fanner participatory breeding projecl was initiated at the Intemational Rice Research Institute (IRRI) and conducted in castem India (Courtois et al. 2000). This is a collaborative project among plant breeders and social scientists from IRRI and six national agricultural research institutions located in eastern India. The Indira Oandhi Agricultural University (IOAU) in Raípur, Madhya Pradesh, is one of Ihe collaborating cenlers. The main objeclives for pursuing fanner participation in plant breeding are as follows: • lo test the hypothesis that farmer participation in raínfed rice breeding can help develop suílable varieties more efficiently • to identify stages along the breeding process where faImers' participation has the most impact and to develop and test a methodology for effectively involving fanners in the breedíng program • 10 improve understanding of male and female criteria for selecting specific rice varieties • to differentiate between the influence of fanner participation and decentralizatíon of the breedíng program • to develop rice varietíes suítable for heterogeneous rainfed environments and which meet fanners' preferences Thís paper focuses on methúdologies for improvíng our understanding of fanners' (including women farmers') criteria for seleeting specífic rice varieties and how these eriteria were considered in participatory breeding strategies for rainfed lowland conditions in Madhya Pradesh, eastem India. Methodology This study ís based on a sample survey of75 rice-fanning households in Ihree villages oflhe Raipur district, Madhya Pradesh. Surveys were conducted to characterize fanners' cropping/fanning systeros, rice varietal diversity, degree of market orientation, gender roles, as well as soeioeconomic differences, and lo relate these to farmers' rice varietal preferences. Farmers were interviewed in regard lo the positive and negative attributes of the traditional and improved varieties they grow and other seed-related information. A method of particípatory weighted ranking was uscd to elicit male and female farmers' eritena for selecting rice varielies accordíng to specific land elevations and information on how they trade offbetween traits. Basic informatíon (name, age, sex, caste, size 344 RK Sahu el al. oflandholding, elevation ofrice plots, etc.) was colIected from male and female heads ofseparate households who are actively involved in rice farming. Twenty cards that iIIustrate traits of rice cultivars were shown and explained to lhe farmers. Referring to a particular land elevation (upland, for example), each farmer was asked what traits he/she considered when selecting rice varieties for lha! elevatíon. The traits that the farmer did not consider important were discarded. Wilh lhe remaining cards representing the chosen traits, the farmer was lhen asked how much weight he/she gave to each trait out of 16 ana (16 ana= 100 paise, 100 paise = 1 Rs). For this process, a total of 16 pieces of stone were provided to the respondent to assign the weights according to hislher choice. An average weight was then computed by getting the sum of all lhe values assigned per trait, divided by lhe number of respondents, afier which lhe proportion of each trait to all traits was calculated. This melhodology in eliciting farmers' perceptíons also provides room for trading off between traits (Sharma el aL 1998; Paris et aL 1999) Farmer participatory approaches for lhe identification or breeding of improved crop cultivars can be usefully categorized into participatory varietal selection (PVS) and participatory plant breeding (PPB). PVS is a more rapid and cost-effective way ofidentifying farmer-preferred cultivars, if a suitable choice of cultivars exists. A successful PVS program has four phases: (1) a means ofidentifying farmers' needs in a cultivar, (2) a search for suitable material to test with farmers, (3) experimentatíon on its acceptability in farmers' fields, and (4) wider dissemination of farmer-preferred cultivars (Whitcombe et al. 1996). In all ofthese phases, understanding farmers' local knowledge, perceptions, and criteria for varietal selection ís important in ímprovíng rice varieties for rainfed ecosystems. Two approaches were used to strengthen farmers' involvement in the project: (1) farmers were invited to lhe research statiDn to view a broad range of genetic materials, and (2) farmers were asked to grow a set of diverse materials in their own fields using their own level of management and inputs. Two farmers in each village volunteered to evaluate 16 rice genotypes on lheir fields using lheir own labor and level of management. Two sets of medium-duration rice genotypes were planted in two farmers' fields in Tarpongi, which has comparatively lighter soils. One set each of late-duratíon varieties was planted in Saguni and Khairknt villages, which have heavy-textured soíls. The set of rice genotypes include prereleased genotypes (F7-F8), advanced lines from lhe Shuttle Breeding Project, and a local check. During specific phenotypic stages of rice production, farmers and plant breeders, using a visual melhod, evaluated and ranked the same set ofrice genotypes on lhe station and on farmers' fields. Kendall' s coefficient of agreement was used to measure the agreement among farmers, among plant breeders, and between farmers and breeders. Farmers recorded lhe reasons for their ranking in lheir diaries. This was done for consecutive years from 1997 to 1999. In 2000, lhe number ofrice genotypes was reduced to five choices (plant breeder, farmer, one common, and a local check). These genotypes will be evaluated before harvesting, bolh at lhe station and on farmers' fields by pIant breeders and farmers. Results and discussion Characteristics 01 the research sites and the larm households This research is being conducted in three villages in lhe Raipur district located on lhe Chhattisgarh plains ofMadhya Pradesh. On lhe Chhatisgarh plains, rice is grown mostly in the lowlands in a drought-prone ecosystem. Drought is a major climatic constraint for rice crops in lhis region. The general c1ímate of the region is dry sub-humíd, where annual potential evapotranspirationallosses 345 Underslanding Farltl~~s..:..s.~leclion Crileria fo,. Rice Varielies are higher than the annual raínfall, whích is about \300 mm. Over 90% ofthe rainfall is reeeived during the period from June to October. The monsoon sets in by 15 June and withdraws around 15 September. Winter conditions set in by mid-November, when the average minimum temperature reaches around ¡5°C. Hence, the rice erop should mature before this time. Sometimes winter conditíons set in early-by the thírd week ofOetober--and thís results in íncreased sterilíty and, thereby, low productivity. Under such fragíle eondítions, the identificatíon of suitable genotypes should be based both on climatic and edaphíc eharacteristies (IRRI-IGAU 2000). The research sites are located in tbree villages: Tarpongi, Saguní, and Khairkut in the Raipur distriet. Tarpongi is 29 km in the north of Raipur; Saguni and Kharkut are 5 km to the west of Tarpongí. These villages are located within 50 km ofIGAU. There are 200 to 250 households in each village. More than 90% ofthe farming households in these villages belong to the other backward caste with small and margínallandholdings (owning less than a hectare), ofwhich the majority are Hindus. Male heads of households have an average of four years in school, while the majority of the women have lower levels of education and did not go to schooL AH ofthe farmers interviewed owned their own land. In eaeh village, 25 farmers were interviewed with regards to their fanníng and eropping systems, rice díversíty, and their eriteria for varietal selection. The survey was conducted in 1997 and 1998. The areas for rice production in these representatíve villages are heterogeneous. Farmers in these villages classify their land according to the topography/slope, such as upland, midland, and lowland. The light so¡ls in the uplands are cIassified by farmers as bhata (entisols), while the sandy loam in the midlands are referred to matasi (ínceptisols). The heavy-textured soils in the lowlands are referred to as kanhar (vertisols). Most of the drought-prone areas have light-textured soíls, whereas the more favorable arcas have heavy-structured soils. Tarpongí has líght-textured soils while the other two villages have heavy-structured soils. The length of the rice-growing season is primarily dependen! on moisture availabilíty, whích ís dependent on slope and soiJ type. Rice ís grown mainly in the rainy season (kharif) in a biasi system. Land preparatíon is done by bulJocks and rice is dry-seeded at the beginning ofthe rainy season in June. When enough rain has accumulated in the field, 25- to 30-day-old seedlin[s are wet-plowed, laddered, and redistributed. This traditional practice, ca1led beushening or biasi, is common in many rainfed areas of eastem India, particularly in Madhya Pradesh. Farmers continue tms practice with the beliefthat ít helps to control weeds and stímulate root growth (Fujísaka el aL 1993; Singh, Singh, and Singh 1994). Farmers grow purple-colored rice varieties as a strategy to identífy and eradicate wild rice (which is prevalent in this region) at an early stage of crop growth. F amily members provide the major source oflabor for rice cultivation. While maJe family members do most of the land preparatíon, rice broadeasting, and applícation of chemicals, females are predominantly responsible for weeding, applying farmyard manure, harvesting, threshing by band, winnowing, and managing seeds for storage. Seed selectíon ís done by both husband and wífe. Other post-harvest activities, such sun drying, dehusking, and parboiling are exclusively done by women. Caring for livestock and, consequently, daiJy collection of green fodder for the livestock is done mostly by women (Sharma et aL 1997). Thus, women's criteria for rice varietal choices may be influenced by their roles and responsibilíties in farming and their social and relígíous obligations, and may differ from those ofmen. The majority ofthe farmers obtain new seeds from their neighbors and from extension workers. Only 24% obtain new seeds from IGAU. This indicates a lack of awareness among farmers about the new technologies developed at the university. Weeds are prevalen! in farmers' fields, and roguing the rice fields to protect the purity of seeds is not 346 cornmonly practiced ín these villages. Rice mixtures and weed seeds are commonly found in the seed stocked for the next season. The cropping intensity in these villages is low because ofthe lack of supplementary irrigation water during the rabi season. The cropping systems in the villages are rice-fallow, rice-lathyrus, or rice-chíckpea (table 1). The chickpea and lathyrus crops are grown as relay crops (locally called utera in rice). Table 1. Characteristics ofthe Rice Land in the Research Sites in Raipur, Madhya Pradesh, Eastern India ! Upland S¡ope . (ndulating) LOIYland (Ieveled and genlly undulating and terr.ced fields) I Mídland (gently , undulaling) Lowland Lowland (¡oveJed) (Iow Iying) Naja Matas; Dorna Kanhar (inceptisols) (alfisols) (vertisols) (verlisoJs) Texture Gravely course , Sandyloam loamy to sandy Silty c!ay Clayey Clayey Depth (cm) Very shallow (5-30) (3~0) Moderate to deep (80-150) Deep (>150) Rapid Moderate Bhata (entisols) Soil. Internal drainage Mechanical composition (%) I Moderate I I Moderate to slow : ! : Deep • (>150) Slow , I Slow (surface "oOOing) i a. Sand ~O 30-50 25-35 20-30 20-30 b. Silt 15-22 30-40 25-30 20-30 20-30 c. Cla 9-20 20-33 33--45 >45 Cropping patt.ms i Rice-Fallow Ri....FaWow i Ríce·Lathyrus or : Chickpea I : Ric.-Lathyrus; >45 Rice-Lathyrus Oikes are plantad wilh pigeon pea Duration of rice varleties suited to these land ShOrl (90-110 days) Intermediate (110-130 day.) Long (130-145days) Long : (> 145 days) Long (> 145 da s) Adoption of rice varieties A high díversity ofrice varíeties exists in these villages. The names of the varíeties grown by farmers in these villages are shown in table 2. Ofthe total area grown to rice in the lowlands ofTarpongi, 73% is grown with traditional varieties, while the rest (27%) has modem varíeties. Twenty years ago, there were about 20 traditional varíeties; however, this number has declined. In contrast, in the uplands of Saguni and Kharkut, the adoption of modem varíeties is slightly higher than thethe adoption of traditional ones. Traditional varíeties such as Safri-17 and Chepti gurmatia are popular in the lowlands. The main reason for adoption of traditional varieties in the lowlands with heavy soíls is because aH the traditional varíeties are tall and can sustain even late biasi operations. According to the rainfall pattem and soíl types of Chhattisgarh, farmers grow varieties according to the land elevation, hydrology, and soils. Rice varieties with a growth duration ofless than 110 days are grown on the upper (undulating) portion ofuplands with loamy to sandy Boíl bhata (entisols). Rice varíeties with a growth duratíon of 110 to 130 days are allocated mainly to the midland (gently undulating) sandy loam matasi (inceptisols). Varieties with a growth duration ofup to 140 days are best suited for light soils, such as those found in Tarpongi village. Late-maturing varíeties (140 to 155 days) are ideal for low-lying, heavy-textured dorasa andkanhar soil types, such as those found 347 Underslanding Farmers' Selection Critería for Rice Varíetíes .. Table 2. Area (Rectares) Planled lo Modern and Traditional Rice Varieties by Sample Farming Households, Elevatlon of Rice Land, and ViIlage, Raipur, Madhya Pradesh Varieties Tarpongi (n = 25) Sagunl (n=50) Khalrkut ("=50) Modem Upland Lowland Upland Lowland Upland lowland Swama 0.8 7.82 27.64 9.86 38.66 5.0 late (150) 2.6 2.22 1.4 6.6 1.0 Medium (130) 8.8 1.8 4.9 Medium (130) 0.1 0.8 Medium, (125) 0.4 Medium (130) Mahamaya Kranti 6.8 6.9 262 7.5 2.1 H.M.T. Purnima 2.4 0.4 late (145) : IR36 Culture Duratlon (days) 0.8 1.86 0.6 Early (120) 1.2 Others Medium (130) 0.7 ·1 Total MVs 20.62 40.52 Tradltlonal 2.9 7~04 1.2 10.7 12 0.44 , 1o.a 7.0 .I'l.anlkajar i 1.8 1.4 ! Bhala safri Aojan safri .. 4A4 0.5 Ganga 6alri 0.3 Nankershar 0.2 Dubraj 7.8 0.1 , 3.8 6.3 , i : ! : 5.2 Late (150) Late (155) 064 5.0 Medium (130) 1.84 5.68 04 Med!um (130) 0.4 212 1.6 ._M Medlum (130) Late (145) , Late (145) I , Late (135) 1.6 Cheptl Total Tradltional 64 3.2 . 6.6 i 40.62 Safri-17 Cheptl gurmat!a 51.36 28.4 ! Safri-BD 15.06 4.7 Medium (130) 20.14 57.0 29.5 18.82 49,50 122 Total of all vanetles 37.64 77.62 70.02 33.88 100.86 18.8 %MV 46.49 26.57 57.87 44A5 50.92 35,11 %Traditlonal 53.51 73.43 42.13 55.55 49.08 64.89 Note: Modem = semi·dwarf, high-yielding vaneties, Traditional = taH in slature wnelher Improved or not improved by seleclÍon. Upland no bunds between plots. in Saguni and Khairkut. Crops are grown chronologically wífu fue lowland fields planted first and the upland helds planted last. Lowland fields are submergence-prone and need to be sown early so fuat seedlings are already establíshed before fue fields are flooded. F armers' perceptions of traditional and modem rice varieties Afier identifyíng the modern and traditional varieties fanners grew, questions were asked about positive and negative attributes. These questions were open-ended and no attempt was made to ¡mpose a priori categories of answers, Table 3 shows the list of positive traits of popular traditional varieties such as Safrí-17 (late duratíon) and Chepti gurmatia (medium duration). Alfuough fuese traditíonal varieties have !ower yields, fanners prefer fuem because of fueir combined positíve 348 R. K Sahu Table 3. el al. Farmers' Assessment of Popular Traditional Varieties Variely Sam-17 (late maturing) Positive Iraits _ _ _ _-+...:N.:.:e:;,.9:ative Iraits stable yield every year resistanl to pests and diseases drought toleran! good for heavy-textured 5011 9000 for beusheníng method 01 land preparation tall (157 cm) and submergence tolerant has lower ylelds (2-3 t/ha) than Swarna and Krantl susceptible to IOOglng due to height (157-168 cm) can'l be used lo distinguish wild rice (karaga) too much slraw and less grain competes with weeds raquires less water and fertilizer photosensitive good lasle and eating quality gOoo grain quality (slender, fine, shinlng) commands high market príce high milling recavery good quantity and quality of straw for making rope matures near religious festival (Diwall) Chepti gurmatia (medium duratlon) good grain yield (3 t/ha) yields lesser !hao Swarna competes with weeds susceptible to lodging because it is lall (137-142cm) tolerant to drought ideal fO( líght soil or Matasi darse medium duration and cán be haIVested early, allowing rabi crop susceptible to bacterial blight and stamborar has more straw than grain purple pigrnests (brown plant hopper) susceptible to shealh blight I • dark green color helps dlstinguish wíld rice . • purple lea! sheath and purple auncle help identíly wild rice not good eating quallty poor mllling recovery-has more broken gralos after milllng early to medlum duration-can harvesl sconer and grow rabi crops Qommands hlgh market prlce has bold, heavy gralns . • good quantity and quality 01 s!raw more fertíle splkelets resistant lo lodging-interrnediate helght responsive lo fertilizar prelerred by millers and !raders far bealen rice (unbroken poja) and lor puffed rice (murmura) because it expands easily prelerred by.peor farmers and agricultural laborers because It remalns soft after cookIn9 and makes !hem feel fuI! even when consumad In sflla:;.lI"'q;::ua:::n.::I"'IIy<-_ _ _ _ _-'-_ Mahamaya was only released in 1997. Both Swarna and Mahamaya were released for irrigated rice ecosystems, but because oftheir perceived ability to tolerate drought and theÍr high market demand by traders, these two varieties have become ver)' popular, Millers and traders prefer Mahamaya for making beaten rice and puffed rice. Poor farmers and agriculturallabarers who are paid in terms af 350 R. K Sahu el al. rice prefer Mahamaya because they feel that it satisfies their hunger. Mahamaya has bold, coarse grains that they believe last longer in the stomach. F armers also prefer Swama for basi (Ieftover rice from dínner, dipped in water with a little salt and eaten the following day for breakfast or lunch). Male andfemalefarmers' eriteria in seleeting riee varieties Despite the active involvement of women in rice production, post-harvest, and seed-management activities, scientists, who are mastly men, aften talk with male farmers only. Ignoring women's knowledge and preferences for rice varieties may be an obstacle lo adoption of improved varieties, particularly in areas with gender-specific tasks and in farm aetivities where women have considerable influence. Far example, a released variety such as Pant-4 is high yielding but is rejected by wamen farmers because it is difficult to thresh by hand. In contrast, traditional varieties that are low yielders are still grown because of their desirable taste and their eating and cooking qualities that make them well-suited forrice produets that women prepare. Knowing men's and women's eriteria in rice varietal selection and access to and control of new seeds, information, etc., willlead lo more efficient dissemination ofimproved rice varieties for rainfed conditions and their subsequent adoption. Thus, in 1998, a team of scientists from the Directorate of Extension, ¡GAU, conducted focused research in the same villages. Our objeetive was to test and develop a methodology for eliciting male and fernale farmers' eriteria and to determine whether there are gender differences in these criteria in rice varietal choice. The majority of the women farmers are illiterate and are less exposed to household surveys; therefore, we used a simple participatory method of elicíting their perceptions regarding the useful traits they consider when selecting rice varieties. Men and women were separately involved in this activity. This method, which is like a game of cards (see methodology section), gave the farmers more time to think as weil as to elljoy the process. Tables 5 to 7 show the important traits thatmale and female farmers eonsider when selectíng rice varieties according lo land elevation and size of landholdíng. The results show that grain yield was the most important eriterion for both men and women farmers in selecting rice varieties for allland types and sizes oflandholding. Both men and women gave more value to eating quality (laste) and durationlmarurity for rice varieties grown on upland fields. However, women were more concemed with market price, drought tolerance, pes! and insect resistance, and competítiveness to weeds. On the other hand, men gave more importance to graín size and shape than women did. For midland conditions, women gave higher values lo eating quality and market price, while men gave more importance lo duration and marurity. For lowlands, eating quality and market price were considerations for both men and women. Women consistently gave higher values to the multiple use of straw for varieties grown in allland types. We also assessed whether there were differences in eriteria between men and women from marginal and large farms. Table 6 shows that there is not much difference between the eriteria across size oflandholdíng. Both men and women wíth large farms gave the highest value lo grain yield. Aside from grain yield, both men and women from the same economic category gave more importance to eatíng quality and market price. Duratíonlmarurity was more importan! to male farmers from large farrns than to women ofthe same category, similar to marginal farmers. Women from both large and small farms gave a higher value lo the multiple use of straw than men did. In summary, the most importan! traits tha! both men and women value in selecting rice varieties are grain yield, eatíng quality (taste), marke! price, durationlmarurity, drought tolerance, and resistance to pests and diseases. Women placed higher weights on multiple uses of straw aeross allland types and for both large and small landholdings. Men did not consider this as important, obvíously 351 Understanding Farmers' Selection Criteria [or Rice Varieties Table 5. Men's and Women's Perceptions of Useful Traits of Rice Varieties by Land Elevation, Raipur, Madhya Pradesh Midlands Uplands Traits Lowlands Men Women Men Women Men Women Grain yield 19 19 27 25 30 27 Eating quality (taste) 16 11 6 17 11 19 Market price 3 10 8 13 9 13 Duration/maturity 13 10 13 6 7 3 Drought tolerance 6 11 5 3 3 1 PesUinsect resistance 6 10 8 6 6 4 Multiple use 01 straw O 8 5 11 6 11 Grain size and shape 16 O 2 2 4 3 Milling recovery 9 O 2 2 4 4 Lodging resistan ce 3 O 3 4 2 3 Fertilizer responsiveness 6 3 5 3 4 2 Weed competitiveness 7 7 3 1 2 2 Submergence tolerance 5 5 1 2 2 2 Good lar rice products O O 2 2 1 0.5 . Disease resistance O O 3 <0.5 3 0.5 Adaptation to soils 3 0.5 2 1 2 1 Adaptation to land level O 0.5 2 1 0.5 1 Storage quality O 2 1 <0.5 2 1 Ful1ness in stomach O 1 <0.5 1 1 Cooking time O 3 1 1 0.5 100 100 100 100 100 100 Note: Values have been rounded off. Values were computed by weighted-ranking method. because women are more responsible than men in caring for the livestock. Rice straw is used as feed for the livestock and also mixed with cowdung to make a cake for household fue!. Thus, women consider both grain yield and rice biomass in selecting rice varieties according to their specific environments. A rice variety that has high grain yields but low quantity and quality of rice straw has a lower chance of adoption by women farmers. Men gave more importance to grain size and shape for varieties grown on the uplands. Men owning smaIl farms considered adaptation ofthe variety to specific soil conditions as being extremely important (second to yield) but were the only group to rank this highly. This may be because poorer farmers cultivate more marginalland (explaining the need for adaptation ofthe variety to soil type). Women did not rank this characteristic highly, probably because oftheir role in production (men tend to choose the varieties and cIear the land). LogicaIly, drought tolerance was more important for upland and midland areas than for lowland areas. Women weighted this more highly than meno While the participatory ranking method was use fui in assessing the trade-offs between traits valued by farmers, this method could be improved by incIuding traits mentioned in the open-ended 352 Table 6. Perceptions of Useful Traits of Rice Varieties, by Size of Landholding and Gender, Raipur, Madhya Pradesh Traits Marginal tarmers larga farmers Women Men Women 36 13 34 19 21 12 18 12 Duratíon/maturi1y 8 10 9 6 3 7 Mulliple use 01 straw 4 7 3 8 10 Drought tolerance 4 6 4 4 PesVinsecl reslstance 7 5 6 7 Grain size and shape 5 2 Milling recovery 8 1 <0.5 2 9 6 Lodging resistance 3 2 4 2 Fertilizer responsiveness 3 2 7 3 Weed competitiveness 1 2 2 1 Men Grain yield Eating quality (taste) Market price i 13 Submergence tolerance 1 5 1 1 Good lor rice products 1 <0.5 1 1 Disease resistance 1 2 <0.5 Adaptation to sails 1 12 <0.5 Adaptation lO land level 1 1 <0.5 Storage quality 1 1 Fullness in stomach O 1 Cooking lime 2 1 2 100 100 100 I 1 <0.5 questionnaires. The cards shown by the researcher limited the choice of desired traits--other traits based on specific cultural practíces, such as a preference for purple-colored rice varieties or for varieties suited to the beushening method of land preparation, were not mentioned at all. Moreover, other social considerations, such as a preference for late and medium varieties to coincide with a religious festival such as Diwali were not captured. Farmers usually harvest rice onIy afterthe Diwali festival. During this festival, families give special rice as gifts to relatives. Participatory varietaI selection Although scientists accept that farmers are careful managers and possess a wealth of knowledge about theÍr production systems, this knowledge is not sufficiently used in the formal breeding process (Kshirsager et al. 1998) Several strategies were used to mvolve farmers in PVS. Farmers volunteered to grow 16 early- to medium-duration group varieties and late-duration varieties on their own fields for three consecutive years. The early/mediuru-duration group varieties were tested at Tarpongi village on two farmers' fields that have light soíls. The late-duration varieties were tested on two farmers' fields at Saguni village under heavy soíls. The new varieties had sorne of the preferred eriteria mentioned by farmers obtamed in the intervíew and partícipatory-ranking activities. Farmen and breeders ranked the rice Hnes on the station and on farmers' fields in the research siles. 353 Understanding Farmers ' Selection Criteria (ar Rice Varieties Table 7. Comparison between Ranks Attributed by Farmers and Breeders at Different Growth Stages in the PVS Trials, Raipur Villages, Eastern India, and IGAU Station, 1997-99 Triallocation Station Tarpongi Saguni Station Tarpongi Saguni Agreement among farmers Agreement among breeders Correlation between farmers' & breeders' rankings No varo No No F. B. W W r F 16 8 1 0.34*" - -0.20 1 M 16 8 1 0.51*" - 0.11 1 F 16 5 - 0.51*" - - 1 M 16 4 2 0.55·' 0.47 013 Year Trial code 2 Stage 97 1 97 97 97 1 97 2 F 16 5 - 0.50" - - 97 2 M 16 7 2 0.34** 0.53 -0.03 97 1 F 16 7 - 0.30** - - 97 1 M 16 6 2 0.44*" 0.30 -0.18 97 2 F 16 5 - 0.79** - - 97 2 M 16 5 2 0.54** 0.56 -0.06 98 1(M) F 16 8 2 0.32** 0.77 0.16 98 1(M) M 16 6 2 0.26 0.60 0.50' 98 2 (L) F 16 8 2 0.31** 0.54 -0.04 98 2 (L) M 16 6 2 0.67** 0.70 0.28 98 1(M) F 16 5 1 0.55** - 0.46 98 1(M) M 16 4 1 0.30*** - 0.20 98 1(M) 98 2 (L) F 16 4 1 0.56" - 0.07 98 2 (L) M 16 4 1 0.59** 0.51' -0.01 CROP FAILURE 98 2 (L) F 16 6 1 98 2 (L) M 16 4 1 0.44' - Station 99 1 M) M 16 7 3 0.49** 0.91"'* 0.33 Station 99 2 M) M 16 7 3 0.65** 0.89" 0.62* Tarpongi 1 99 1 M) M 16 6 3 0.65" 0.94" 0.61' 0.84** 0.46 0.15 Khairkhutl 0.38** 0.02 2 M) M 16 5 3 0.62** 99 1 (L) M 16 7 3 0.53** 0.81" 99 2 (L) M 16 7 3 0.34** 0.76** 0.11 0.66** Tarpongi 2 99 Station Station Saguni 1 99 1 (L) M 16 7 3 0.50** 0.93" Saguni 2 99 2 (L) M 16 6 3 0.66** 0.91 ** 0.64- Station il9 1 V 20 5 3 0.98** 0.94** 0.90" Station 99 1 F 20 5 3 0.98** 0.98" 0.91 ** Station 99 1 M 20 5 3 0.96" 0.97- 0.89" Khairkhut 99 2 V 20 5 3 0.98** 0.95" 0.87" Khairkhut 99 2 F 20 5 3 0.94** 0.99- 0.92" Khairkhut 99 2 M 20 5 3 0.90** 0.97" 0.41** Note: - = not tested. W = Kendall's coefficient of concordance. r = Spearman_'s coefficient ofcorrelation. F = farmers. B = breeders. 1. Stage: V = vegetative stage, F = flowering, M = maturity. 2. Trial code: L = late, M = medium. 354 Farmers' rankings were compared with breeders' rankings during different stages of crop growth (vegetative, flowering, and maturity) as shown in table 7. Correlation between breeders and farmers al all siles and in al! the years was consistently low. Very few of the trials showed significant or highly significant agreement between farmers and breeders (trials that showed any significant agreement were mainly in 1999). In general, agreement was insignificant or even negalive (although not strongly so). It was impossible lo make an assessment of agreement between farmers and breeders in 1997 and 1998. However, in 1999, although there was high agreement in varietal ranking among farmers and among breeders, there was generally low agreement between farmers and breeders, which may indicate that farmers and breeders consider different criteria. Farmers' rankings are not correlated with yield, indícating Ihat farmers consider other criteria in their rankings. Assessment uf late-duratiun varieties included in PVS in Saguni, Raipur The breeders' top five favorite late-duratíon varíeties ín the 1999 trials included Swama, BKP-232, R650-18l7, R304-34, and R738-1-64-2-2 (aH modern varieties). These varieties also ranked in the top five in yield. The farmers' top five favorite varieties included Swarna, Safii-17, R 738-1-64-2-2, Mahsurí, and R650-1817. These were not always the highest yielding varieties-in fact, Mashuri gave one of the lowest yíelds and Safii-17 (a tradítional variety) was somewhere in the middle. These varieties were likely selected for other reasons than yield. Varieties preferred by both groups (ranking on average in the top 5) included Swarna (first choice ofboth farmers and breeders, and also high yieldíng), R650-1817, and R738-1-64-2-2. These are aH modero varieties, and are also the three varieties that had the híghest yíelds in the trials (table 8). Table 8. Assessment ofLate-Maturing Varieties Included in PVS, Saguni, Raipur, Madbya Pradesh, Eastern India Varíety . Ranking Swama (check) : Favorite 01 both farmers and breeders Consistently ranked highly in the tep 5 by beth groups in the field sites and en-station Safri-17 (check) R73&-64 R304-34 Always ranked in the !op 5 by farmers, bu! no! so well ranked by breeders , Thls 15 ranked in !he top 5 by farmers and breeder5 in !he farmers' flelds, but les5 , well ranked In on-station trials. Ranked flrst by breeders, but not IIked by farmers, even though yleld 15 quite good ! (5 tJha) . Ranked low by both groups in fleld sites ! Bold rains, not susceptible to disease, oommands hl h market prlee Mahsun On-stalien. ranked wíthin top 5 by farmers, en station and in one farm site, although yield is consistently iow Ranked consistentiy low by breeders On-slation. ranked highly by breeders Yield is gaod, bu! larmers don't like it (one of their least favarttes) _ _ _ _ _ _-"-.. Ra"'r1.ked Jow by all in larm trials IR54896 Assessment of medium-duration varieties in Tarpongi, Raipur, Madhya Pradesh In Tarpongi, the top ranking medíum-duration varieties for breeders were R574-11, IR42342, Chepti gurmatia, BG380-2, R703-1-52-1, and ORl158-261. AlI of these were also the top six 355 Understanding Farmers' Seleclíon C~ite",ri"Oa.t..:fo"-r-,R",jc",e,-,~,-,a,,-r¡,,,·e:::tie::;s,--_ _ _ _ _ _ _ _ _ _ _ _ __ __ yielding varietíes. Al! are modem varieties except for Cheptí gurmatia. For farmers, the top ranking varieties included BG380-2, ORI158-261, R714-2-9-3-3, IR63429, and R574-11. These are al! modem varietíes, bu! no! always top yielding. R714-2-9-3-3 gave medium yields, while IR63429 gave relatively low yields when compared with the other varieties. Farmers and breeders agreed onIy on R574-11, BG380-2, and ORl158-261 as their favorite varieties (table 9). Table 9. Assessment ofMedium-Duration Varieties Included in PVS, Raipnr, Madhya Pradesh Varlety I Ranking R714--2-9-3-3 , Ranked highly by farmers on farmers' fields and in 2nd on-slatioo replicatíon, and is among , the farmers' favorites i Consistently marked low by breedars R574·11 Top ranked by tarmers and by breaders in station tríals. Also, highest yield On-farm, la still in top 1-2 for breeder. but drops to 3-10'" rank for fanners . Yleld on farm ls less (4th and 6th rank) ! ! OR1158-26 , Ranked abou! 5-6 (on average) in all siles excep! in one field, where it was4t1 among farmers , Yield ranges from 3-6 tlha Among !he top varieties lor !anners and IR63429 Ranked well by farmers in all sites bol conslstently ranked low by breaders Lower-yieldlng varíety compared to others, but fanners seam to IIke it In any case Early, long grain, intennediate hei ht IR42324 Consis!ently highly ranked by breeders, bu! given low rank by fanners in all sites excapt station repllcation #1 Cons¡stently high yield, ,but even with highes! yield on larm, larmers don't I¡ke it Chepti gurmatia (local check) Consistently ranked well by breeders, also one 01 the top 5 yielding varieties However, ij ranks in the middle with !armers BG380-2 Ranked hlghly by breader. and fanners In field and on-statlon : Generally has goOO yleld During the kharif season 2000, Ihe medium-duratíon' varielies that were further evaluated on-stalion and on farmers' fields were IR4234 (breeders' choice), R574-1I (farmers' choice), BG380-2 (common choice), and Chepti gurmatía (best local choice). The late-duration varieties were BKP-232 (farmers' choice), R304-34 (breeder's choice), R650-1817 (common choice), and Swarna (local check). The challenge facing plant breeders in IGAU and IRRI ís lo develop new cultívars that are better Ihan Swama and Maharnaya, while a1so meeting the other requirements and criteria thal furmers have for their given rice environments. While it is impossíble 10 combine all the requirements in one single variety, giving farmers (both men and women) an opportunity 10 test the performance of different rice genotypes on their own fields and 10 evaluate their cooking and eating qualíties can ¡ead 10 more efficient rice varietal improvement in the Chhattisgarh region in Madhya Pradesh. Conclusions This paper focused on methodologies for improving our understanding ofthe eriteria used by farmers (both men and women) in selecling specific rice varieties and ofhow these criteria are considered in partícipatory breeding strategies in the rainfed lowland environments of the Chhatisgarh region in Madhya Pradesh, eastem India, Different methods for understanding farmers' eriteria in 356 Sahu et al. selecting rice varíeties were used. These melhods were (1) a questionnaire with open-ended queslions eliciting positive and negative attribules ofthe most popular modem and traditional varíeties, (2) a participatory weighted-ranking method, disaggregating the perceptions of men and women by land types and size oflandholdings, and (3) participatory varietal selection, where farmers evaluated severa] prereleased and local varíeties on their fields as well as on-station. The results of the study highlíght the importanee farmers attach to characteristies other than grain yield: eating quality (taste), rnarket price, durationlrnaturity, drought tolerance, and pest and inseet resistance. Both men and women have similar eriteria in choosing rice varíeties. However, straw quality for multiple uses is an important consideration for women farrners but not for meno F armers, particularIy women who do most ofthe weeding, prefer rice varíeties that are inherently dark green orpurpie to distinguish them from wild rice and enable the farmer to eradicate the wíld rice at an early stage of crop growth. Wild rice is a prevalent pesl and a constraint to high rice productivity in the Chhattisgarh regíon. The attributes considered by men and women farmers, however, are not generalIy used as screenlng eriteria in most formal breeding prograrns, where the emphasis is mainly on grain yield. Qualíty attributes should be' emphasized more Ihan they have been in the past in breeding prograrns for rainfed areas. Because of the proximity of the villages lo !he markel, farmers prefer lo grow varieties Ibat no! only mee! their own eonsumption needs bul also those of consumers, including millers and traders. Therefore, farmers maintain their rice diversity and grow both traditional and modero varíeties that meet their varied interests and needs. Using approaches like farmer participatory breeding and varíetal selection from many rice lines provides an opportunity to fanners to choose varieties suitable lo their environment and needs as well as access to new seeds. Breeding lines R574-1l, BG308-2, and IR42342 performed well over the tbree years ofthe project in the medillm-duratíon group and showed tolerance to drought. Breeding lines R304-34 and JET-14444 (R 738-1-64) also proved promising. A large quantity of seeds have been multiplied by one ofthe farmers of Saguni village where blight is a problem. References Courtois B., Bartolome, D. Chaudhary, G. MeLaren, C.H. Mishra, N.P. Mandal, S. Pandey, T. Paris, C. Piggin, K. Prasad, A.T. Roy, R.K. S.hu, S. Sarkarung, S.K. Shanna, A. Síngh, H.N. Singh, O.N. Singh, R.K. Síngh, S. Singh, and B.V.S Sisiodia. 2000. Partíeipatory varietal seleelion for low input environments: A case study of rainfed rice in eastem India. Euphytica (forthcoming). Fujisaka,1.S. 1990. What does buildresearch onfarmer practice mean? Rice crop establishment (beushening) in eastem India as an ilIustration (/990). Social Sciences Division Paper No. 90-40. Los Baños; Philippines: Iotemaliona! Rice Research Institute. IRRI-IGAU. 2000. Collaborative prajeet on farmers' participatory breeding program: Agrometeorological analysis ofrice genolypes ofdifferenr durations grown under rainfed lowland rice ecosystem. Raipur, India, Indira GanAI -llíDP fiigure 1. The rainfall zone. of Jordan Barley is mainly cultivated in tbe dryland areas tbat cover part of east Jordan. These areas are characterized by low rainfal!, irregular1y distributed, with most of the rain falling during the winter. Temperatures vary widely, witb frequent feost in early spring and in late spring, resulting in head sterilíty, low yields (table 2) and often in crop failure. The unpredictable envirorunental conditions, along with poor soils and crop management, have made it difficult to introduce new cultivars and obtain yield increases. 384 Table 1. Planted Ares, Production, and Productivity ofBarley in Jordan (1990-1997) Year Area thousand dunum Productien thousand ton Productivity (kgldunum) 90 34.44 36A 1060 91 25.58 26.8 1190 92 83.63 103.2 1230 93 69.19 44.2 640 94 62.56 34.3 550 95 83.83 57.7 690 96 52.22 44.9 860 97 50.29 42.8 850 Average 57.34 48.8 883 Table 2. Mean Graln Yield, Biological Yield, Straw Yield, Plant Height, Harvest Index, Days to Heading, Days to Maturity, and GraÍl~ FilIing Period at Four Locations In Jordan Locations Tralts Rabba Khanasrl Ghweer Ramtha Mean Grain yield (g/pIOI) 289.58 95.90 154.02 129.48 167.3 Bielogical yield (g/piel) 99490 423.0 807.1 639.2 716.0 Straw yield (g/piel) 705.32 327.1 65308 509.72 548.7 Plant height (cm) 59.50 33.1 47.9 56.6 49.3 Hatvest index (%) 29.30 22.5 19.3 21.1 23.1 Days to heading (days) 118.50 110.8 85.6 128.7 110.9 Days to matunty (days) 152.0 137.8 113.2 155.9 139.7 Filling perlad (days) 31.50 24.0 26.5 24.2 26.5 Nore: Data are !he means of84 barley lines during the 199611997 growing soaSoo. Project objectives and expected OUtputs The long-tenn goal oflhe project is the improvement oflhe welfare ofsmall, resource-poor fanners by increasing and stabilizing barley and livestock productíon. The imrnediate objectives of the project are • to develop a participatory approach to breedíng barley for stress conditions • to improve barley varieties Ihat fulfill Ihe needs of poor fanners in Ihe marginal rainfed environrnents of Jordan • to enhance the rate of adoption of new varietíes through fanners' participation in selection and testing 385 lncreasing ¡he Relevance ol Breedin.""~-"to:..:S""m,,,a!ClI:...!l'.!·a::.r:::m"'er:.::s_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ • to identify dífferences in seleetion eriteria used by different types of farmers (aceording to gender, enterprise mix, and other farm characteristics) At the end of the project we expect the following outputs: • documented and validated information on farmers' objeetives, know ledge, and field condilions • the performance and quality, under both farmers' and station conditions, of barley fines selected by farmers in their fields, compared with the performance and quality oflines selected on the experiment station using breeders' setection erítería • doeumentation of the selection eritería used by different types of farmers andlor different members offarm households • a number oflines selected and developed through this participatory breeding program multiplied by farmers and tested by neighboríng farmers • the importance of the interactions between selection erítería and selection envíronment assessed • incorporation of participatory approaches by the two national breeding prograrns Methodology Orientation and targeting Al each ofthe locations included in the project area, cooperating farmers ("host farmers"), who will hos! breeding plots and make individual selections, will be recruited from the pool of participants in previous on-farm research and cooperatíve research programs in ongoing research-and-devetopment projects. A rapid-appraisal exercise will be conducted wíthin the agricultural cornmunity assocíated with each ofthe selected agroecologicallocations, and a group oflocal "expert farmers" wiU be identified and recruíted on the basis of reputatíon, key farming contacts, past performance, gender representation, producer and consumer categories, and se1f-selectíon. The expert farmer groups, together with the host farmers, wíll participate as key informants in the indigenous-knowledge study and will perform group selections from their respective host furmers' germplasm collectíons. lndigenous knowledge Ibis component has several crucial outputs for developíng the partícipatory-breeding approach. First, there will be an enquíry into farmers' objectíves, reasons for producing barley, and different end-uses ofthe crop. This will include theírperceptíons ofthe dífficulties they experíence in reaching these objectíves. Household economic securíty and rísk considerations will also be considered in the context of production objectives and genotype evaluation. The índigenous-knowledge study will pro vide the information needed for the analysis of concepts 5uch as how farmers, both men and women, value various characteristics ofthe barley erop and how much they understand adaptatíon for specific environments and uses. The methodology for data collectíon and analysis will rely prímarily on formal ethnographíc techniques used in sociocultural anthropology, including participant observation, structured interviews, and taxonomic and componential analysis of labeled traits. As much as possible, badey characterístícs recognized by 386 fanners will be classified hierarchically to enable selection procedures to be applied one afier the other aeeordíng to prioritíes reported by fanners. Indígenous methods for reeognízing desirable characteristies within populatíons ofbarley eultívars will be documented, and activities of fanners applying these methods will be recorded in detail. An important aspect ofthis component i5 the ídentíficatíon ofwomen's seleclíon eriteria, partícularly, but not only, al those locatíons where barley is u5ed for human consumption. Specific outputs for tbis component inelude the following: • evaluatíon of the innovatíve capacity of fanners and insíght into theír potentíal for direct participation ín formal breeding programs • lí518 of desirable characteristícs, prioritized and cross-referenced to environment and utilization • indigenous knowledge and perceptions of environment-genotype ínteractions in barley landraces • the theory, objectíves, and implementation ofthe participatory-breeding prograrn will be discussed thoroughly with the host fanners and expert farmer groups in order to obtain their input into the design of the breeding scheme, including selection procedures, such as the proper time for selection, how ofien selection is done, etc. From centralized nonparticipatory io decentralized participatory barley breeding This componen! represents the major empirical thrust ofthe project and will quantify the effects of the selection environment (experiment station vs. fanner's field), of who does the selection (breeder vs. farmer), and whether these effeets interact or vary from year to year. The traits that fanners select for, and the eriteria they use in their selection, will be recorded by the breeders and social scientists, and compared with objective measures of trai18 used by barley breeders, including the yíeld and quality of graín and straw. A common set oflines and populations (including the farmers' cultivars) will be grown on a typically well-managed experiment station field and on one fanner's field at each of six !ocations in Jordan under fanners' management practices (fertilízer use, rotations, date and method of sowing, land preparation, etc.). The locations will be as follows: Al-Mohay Al-Muaqure Ramtha Khanasri RaMa Ghwer 60 km southeast ofK.arak and about 130 km south of Aroman, with an annual rainfull of about 13Q....1 50 mm 55 km east of Amrnan, in the arid areas, with an annual rainfaIl of 150 mm 160 km north of Amrnan, with an annual rainfall of 250 mm 135 km north of Amrnan, with an annual rainfall of200 mm 140 km south of Amman, with an annual rainfall of 340 mm 160 km south of Aroman, with an annual rainfall of 280 mm By ineluding locations with less than 200 mm average annual raínfall, there will be opportunities to investigate the performance of breeding material in environments where barley is a forage crop rather than a grain erop. In addition, small graín-producing areas occur wíthin the < 200 mm zone. These are seasonally f100ded wadi f100rs (marrabs), where high graín yields are normal wíthín a generally arid environment. Because of theír ímportanee locally, and the uniqueness of the agroecosystem, these !ocations are included in the project. 387 [n the project area, the majority of farmers still grow barley landraces ¡hat are heterogeneous populations composed of a large number of individual genotypes. Although the population buffering of such heterogeneous populations-and, hence, their role in reducing ¡he risk of crop failures-is well documented, we do no! know whether farmers perceive this type of diversity as importan! and jf this is the reason for the popularity of landraces. To gain information on this specific point, the genetic material will include high-yielding fixed or nearly fixed lines, segregation populations, and farmers' cultivars. The use ofbolh pure lines and heterogeneous populations will provide a means for testing the attitude of farmers towards heterogeneity, as opposed to the conventional breeders' propensity for homogeneity. The farmers' cultivars, which are likely to be different at each location, wíll be collected from each farmer during the harvest of the previous year, and all farmers' cultivars will be grown at each site. Selection will be conducted on the experiment station by breeders, and in each host farmer' s field, selection will be conducted by both the breeders and the host farmers, their spouses, andlor olher household members. Whenever possible, neighboring farmers wiJI also participa te in lhe selecnon process. The collaborating farm householders will make selections from theír fields. Following a group selection procedure similar to thatused by ICRISAT in Rajasthan (ICRISAT 1996:98-100), the expert farmer groups wíll be asked to select material from lhat grown by lheir host farmers, material lhat they think would be use fui for- them and olher farmers in lheír area. The selectíon wiIl be conducted in such a way as to reveal the criteria being used by lhe farmers and olhers when they make their choices. There will be detailed discussions regarding lhe cultivars selected and the eriteria used in selection. Farmers' observatíons, expected performance, and crop-management practices wíll be recorded. At lhe end ofthe frrst year, in addition to the breeders' se1ections from lhe experirnent stations, for each particípating farmer, the following groups of selected Iines wíll be available: 1. 2. 3. 4. lines selected by lhe breeder Iines selected by the farmer Iines selected by other household members lines selected by lhe farmer's neighbors In lhe second year, each host farmer wilI growall lhe Iines selected in hislher ficld in the fust year, regardless of who made lhe selection, Le., groups 1 to 4 aboye, as well as the lines selected by the breeder in lhe experiment station. The selections will be grown as one populatíon oflines without obvíous distinctions between the groups to avoid any possible bias in the second cycle of selection. AH lhe lines selected in lhe first year wíl1 also be grown on the experiment station in the sccond year to provide enough seed for lhe third year. Data on grain and straw yield will be collected at each host farmer' s field and at lhe experiment station. Response to selection will be evaluated using lhe farmer's cultivar as reference. In lhe second and trurd year, selection will be done, as in the first year-on the lines resulting from the first and second cycle of selection. Thus, during lhe second and third cycle (year) of selection, lhe farmers and lhe breeders wí11 be exposed to the material selected by each olher. By lhe third year, the project will have involved a total of36 households in lhe target area and will have simulated tbree cycles of selection of the sarue type of cyclical processes !hat take place in conventional breeding prograrns (figure 2). 388 S. Ceccarelli el al. Formal breeding program generates diversity Neighbor Farmer Nefghbor Farmer Neighbor Farrner Neighbor Farmer - Formalor informal Netghbor Farmer NeighbOr Farmer s ystems Nelghbor Farmer Neighbor Farmer NeighOOl' Farmer Figure 2. Scheme of the decentralized participatory barley-breeding program for one location (The number of farmers is arbitrary. The same scheme Is repeated in six locations.) During the selection proeess, the eriteria ofboth fanners and breeders will be monitored and eompared. Of particular inlerest will be the frequeney with wmeh me fatrners,in me seeand and thírd year, seleet from the material they selected themselves in the first year and from among the material selected in me first year by the breeders. This will give not only an indícation ofthe eonsistency of fanners' seleetion eriteria, but also an indieation of the possible effects of fluetuations in environmen! over years on genotype performance and fanners' perceptions ofthese effects. Tm3 component is designed to quantify the following effects: • the effeet of the selection environment (experiment station vs. farmer' s field) by comparing, both on the experiment station and on the fanner' s field, the superiority over the fanner' s cultivar ofthe Iines selected by the breeder on-station with the superiority ofthose seleeted by the breeder in the fanner' s field • the effect of seleetion eriteria (breeder vs. fanner) by comparing, in the fanner's field, the superioríty over the fanner's cultivar ofthe lines seleeted by the breeder with the superiority of those seleeted by the fanner (this comparison wíll be extended to cover seleetions done by others, Le., farm household members and/or neighbors.) At the end of the first three years, it is expected tha! the number of selected lines wíll be small enough 10 stimulate the interest of the participating fanners, and possibly of sorne neíghboring fanners, to grow one or more of them as commercial erops. The experimental material will be assembled and distributed by the barley breeders to ensure a UIÚform seed souree. 389 References Bínswanger, H.P. and B.e Bamh. 1980. Yield risk. risk aversion. and genotype se/eclíon: Conceptual issues and ap· proaches. Research Bulletín No. ). Hyderabad, India: ¡nt.malional Crops Research Inslitute for Ihe Semi·Arid Trapics. Byerlee, D. and T. Husain. 1993. Agrieultural researeh strategies for favored and marginal areas: The experience of fanning system research in Pakístan. Experimental Agrieulture 29:155-171. Cecearelli, S. 1994. Specific adaptalion and breeding for marginal conditions. Euphytica 77(3):205-219. Ceccarelli, S. 1996. Positive interpretation of genotype by environmenl interactlOns in relation to sustainability and biodiversity. In Plant adaptation ond erop improvement, edited by M. Cooper and G.L Hammers. Wallingford, UK: CAB Intemational. Ceccarelli, S., S. Grando, R. Tutwiler, J. Baha, A.M Martini, H. Salahieh, A. Goodchild, and M. Michae\. 2000. A methodological study on participatory barley breeding. 1: Selection phase. Euphytica 111: 91-104. CeccareHi, S., S. Grando, A. Arnri, F.A. Asaad, A. Benbelkacem, M. Harrabi, M. Maatouguí, M.S. Melmi, H. Mimoun, R.A. EI-Emen, M. EI-Felah, A.F. EI-S.yed, A.S. Shreídi, and A. Yahyaoui. 2001. Decentralized and participatory plant breeding for marginal environments, In Broadening the genetic base of crap praduction, edite'd by H.D. Cooper, C. Spillane, and T. Hodgink. WaHingford, UK: CAB Intern.tinna!. G.lt, D. 1989. Joining FSR lO cornrnadity programrne breeding efJarts earlier: Inereasing plan! breeding efficiency in Nepal. Network Paper 8. London: Overseas Develapment Institute. Grisley, W. 1993. Seed for be.n production in sulrSaharan Africa: ¡ssues, problems, and possible solutíons. Agricultural Systems 43: 19-33. Hardon, J.J .•nd W.S. de Boef. 1993. Linking farmers and breeders in local crap development. In Cullívating knowl· edge, genetic diversity. farmer experimen/ation and crap research, edited by W. de Boef, K. Amnor, K. Wel· lard, .nd A, Bebbington. Londnn: Inlennediate Technology Publications, Ltd. ICRISAT. 1996. ICRISAT Asia region annual report 1995. Semi·formal publication. Andhra Pradesh, India: Interna· tÍonal Crops Research Institute for tbe Semi·Arid Tropics. Saade, M., F. Nassif, A. Arnri, and H. El Bagbati. 1993. Constraints to the adoprion ofbarley varle/ies in Morocco. ¡CARDA Expert Report (Semi·formal publicalion). Aleppo, Syria: ¡ntemational Center for AgricuItural Researeh in the Dry Areas. Sirnmonds, N. W. 1991. Selectíon for local adaptalion in a plant breeding prograrnme. Theoretical and Applied Geneties 82:363-367. Sperling, L., M.E. Loevinsohn, and B. Ntabomura. 1993. Rethinkíng the farmer's role in plant breeding: Local beanexperts and on·stalion seleelion in R wand•. Experimental Agriculture 29:509-519. 390 Present Status of Participatory Plant-Breeding Research on Wheat at the National Wheat Research Program ofNepal M.R. Bhatta, G.O Ferrara, B. Gurung, T-P- Pokhrel, NR. Gautam, P. Gurung, and R.B. Neupane Abstract Participatory varieta! seleclion work io the form of coordinated farmers' fi.ld trials iovolving farmers, extension agents, and researehers has becn a regular component of the wheat-improvement program of Nepal sinee Ihe 1970s. In this system, variet.1 tesling is camed oul in farmers' tields, bul the participation of farmers in actual breeding work has varied greatly. Recently, more farmor-collaborative plantbreeding work has beco initiated in Ihe Bankatti village ofthe Rupandehi district A pre-breedingparticipalory assessment involving 20 male snd female farmers was conducled lo delermine Ihe preferenee crileria tba! farmers emp!oy in seleeling for wheal varieties. There was some degree ofvari.lion ín preferenee criteria listOO by women and men farmers. Male and female farmers were allowed lo selectlevaluate 12 wheal varielies grown in farmers' fields al near rnaturity, based on preference crileria sel oul during fue pre-breeding exereíse. This paper summarizes the results oflhe pre-breeding survey conducled al Bankatti village and Ihe varietal evalualion done by male and female farmeTS. Comparisons are made belween farmers' preferences and tbe !rails sel out by the Nalional Wheat Researeh Program in developing wheal varieties for different domains. Introduction Wheat ís lhe third most ímportant cereal crop in Nepal, afier rice and maíze. Until lhe mide 19605, wheat was consídered a minor cereal and its cultivatíon was limíted only to the far westem hill regíon. Coordínated research and extension efforts in wheat during the last 30 years have sígnificantly contributed towards an increase in the area planted to wheat and in produetion and productivíty. At present, the wheat crop covers more than 640,000 hectares wíth a total production of 1,086,000 metric tonnes. The natíonal average productivity of wheat is 1700 kglha (CBS 1999). Wheat occupies 22% ofthe country's total eultivated area and 20.2% ofthe total planted lo cereals. It contributes 16.2% ofthe total cereal produetion in the country. Although there has been a tremendous ínerease in area and production, national wheat productivity is stilllow, whích is attributed to many factors. Some ofthese are a poor rate ofseed replacement, slow varietal replacement, use of poor-quality seed by fanners, suboptimal fertilizer use, ínsufficient irrigation facilities, and a low fann-gate price. More than 90% ofthe wheat seed moves from farmer to farmer. The present seed-replacement rate is 4% to 5%, and a newly released variety takes five to 10 years to cover a large area, depending on the acceptance ofthe variety. Ninetypercent of the country's wheat area is covered by modem wheat varieties; however, many farmers still grow old, dísease-susceptible, low-yieldíng varieties, resulting poor yield. M.R. Bh.na, T.P. Pokllrel, N.R. Gautam, and R.B. Neupane are with Ihe N.nonal Wheat Research Program, Bhairahawa, Nepal. G.O. Ferrara is with the Intematl0nal Maize and Wheat lmprovement Center, Soutb Asia Regional Office. Kathmandu, NepaL R Gurung and P. Gurung are with the System-wide Program on Parti~lpatory Research and Gender AnaJysisl Intemational Center for Tropical Agriculture (PRGA·CIAT). 391 Presen! Status ofl'artícípatol]J Plant-Breeding Research on Wheat The national wheat research program was establíshed in 1972 at the Botany Division in Kathmandu Valley. In 1975, the headquarters moved down to Bhaírahawa, westem lera!. with the following natíonal mandate: l. develop, ímplement, coordinate, and monitor multilocational and multidisciplinary adaptÍve crop-írnprovernent research for developing superior varieties resistantltolerant lo biotic and abiolic stresses for different agrochmatic conditions 2. colleet, evaluate, identífY, maintain, and use suitable donors for different biotie and abiotic stresses 3. develop appropriate crop-productíon technologies for optimal use of resources in a sustainable rnanner 4. produce a nuc!eus and breeder seeds of popular varieties in required quantities 5. establísh national and intemational linkages for strengthening wheat-improvement researeh in the country The National Wheat Research Program's wheat-breeding objectives are to deve[op wheat varieties with the following major traits: . 1. high yield polential 2. resistance to multiple diseases 3. widely adaptive 4. medium in height 5, with bold, white grains 6. early in maturity 7, tolerant to late heat stress 8, tolerant to drought 9. tolerant to lodging 10, with high protein content ( aboye 10%) 11. tolerant to sterility Wheat-production tones In Nepal there are two major wheat-production zones, One líes in terai, and terai-like arcas in river basins and lower valleys up to 500 meters and the other is in the mid- and high hills, aboye 500 meterso The former zone represerits 60% ofthe total wheat area and contributes 63% ofthe total wheat productíon in the country. Tnis zone is further subdivided into three production environments: (1) rainfed, which represents 28% of tbe terai wheat area, (2) irrigated, which represents 72% of the terai wheat area, and (3) Late-sown, which represents 250/0-30% of the wheat in the terai. Rice-wheat is tbe dominant croppíng pattern in this zone. The mid-hil1s represent about 40% ofthe total wheat area and production, and the high hills, about 37%. In these areas, wheat is grown after ricc-wheat in irrigated khet land and after maize and millet in rainfed barí land. The wheat-research program has released 27 improved wheat varietíes since 1960, and if we look critieally at their adoption rate, five out of27 have very high adoption, 10 have high, six have low, and five out of the 27 have a very low adoption rate (table 1). Present yield gap Figure 1 cJearly indicates the research-generated technologies practiced on experiment stations, revealing a 4.8 tiha grain yield, compared to 3.2 tlha obtained in farmers' field trials. Howeyer, the 392 .. Table 1. -'-.~~ M.R. Shalla el al. .. Improved Bread Wheat Varieties Released Sínce 1960 and Tbeir Adoption in Nepal Pedigree Year released Area of adllptation Lerma-52 MentanalKenya 324 1960 HiIIs High LR-64 Y50IN10BIIL521312'lR 1967 HíIIs Verylow Pilie-62 YT54IN10B 126.16 1967 Hills Very low Kalyansona Pj"S"IGabo 55 1968 Tarai Low RR-21 1154-388IAN/31YT54IN 10BIILR64 1971 Hills and terai Very high NL-30 HD832-5-5-0YIBB 1975 Westam lerai Low HD 1982 E5557IHD845 1975 Wastern terai low UP262 S 308/BAJIO 66 1978 Tara; Very high Lumbini E48711PJ62 1981 Terai Very low Triveni HD19631HD1931 1982 Tera; low Vinayak LC 55 1983 Teraí High Siddhartha HD20921HD196211 E4870 /3/K65 1983 Tera! Hlgh Vaskar TZPP/PUI7C 1983 Mld-wesl terai Very low Tera! Very high Teraí Hlgh '1988 HiIIs High Variety Adoption Nepal297 HD2137/HD2186/1 HO 2160 1985 Nepal251 WH147/HD216011 WH147 1988 Annapuma 1 KVZlBUHOIlKALl8B Annapurna 2 NPOITOBl/8156/3/ KAUBB 1988 HiIIs Very low Bl1022 PVNIBUC 1991 Western teraí High Annapurna 3 KVZlBUHOIIKAL/BB 1991 HiIIs Hlgh Bhrikuti CMT/COC7513/PLOII FURY/ANA75 1994 Terai Very high Bl1135 QTZ/TAN 1994 Teraí Hlgh Annapurna 4 KVZl3ICCIINIAI/CNO/ ELGAUISN64 1994 HiIIs Hlgh Aehyut CPAN188IHD2204 1997 Tera; High Rohini PRUTONillCHIL 1997 Terai Very high pasang PGOISERI 1997 HiIIs low Kanti LlRAlFUFANI7/NEE#5 1997 HiIIs Low Bl1473 Nepal 297 INL 352 1999 Tera! ? national average is 1.7 tJha, clearly indicating that a threefold yield increase-compared to experiment-station yields-and a twofold increase-compared to fanners' field trials-is possible, To achieve this, new technological advances are to be made in the area of crop improvement through participatory plant breedíng and partícipatory varietal selection. This will facilitate fue rapid adoption of newly released varieties, along with a faster seed-replacement rate. And ultimately, fanners will get varieties wifu fueir preferred traits. This will further help in the identification and release oflocation-specific wheat varieties for different agroecological niches. 393 Present§tatus Di Participato.rv Plam-Breeding Research O" Wheat 6000 4800 5000 íiI r. en ::- 3000 Qi 2000 4000 '"O ;;: 1000 O Nat" ave. On-lann Exp. yield res. yield Sourc•.· Mudwarí, Bh.lta, and Pkharel (1998). Figure 1. Present yield gap in experimental plots and on-farm research plots compared with national average yields 'Participatory crop improvement The National Wheat Research Program has been involved in participatory crop-improvement activities since the 19705, in tbe form of coordinated farmers' field trials (CFFTs), farmers' acceptance tests (FATs),frontline on-farm research, district seed self-sufficiency programs, supplying experimental germplasm to differentnongovermnental organizations (NGOs), and participating in joint monitoring trips to evaluate wheat genotypes witb farmers. Figure 2 illustrates the presen! and past wheat varietal development process of tbe wheat variety release system. Farmers' informal participation starts from tbe initial evaluation trial onwards, where farmer-visitor groups are allowed to evaluate wheat genotypes planted at different research ,tations. Coordinated farmers' field trials are carried out in tbe farmers' fields in multilocational test sites. The genotypes in on-farm trials are evaluated jointly by farmers and researchers. As in farmers' acceptance tests, small seed kits of pre-released and released varieties are distributed to a number of farm families tbroughout tbe country, along with a questionnaire to be retumed by them. The questionnaire used in evaluating coordinated farmers' field trials and tbe farmers' acceptance test are given in annex l. Frontline, on-farm research is a triangular activity where farmers, extension personne1, and research scientists are involved right from síte selection to variety evaluation. In frontline research, a complete technology package is introduced, along with an improved variety, in a large area where many farmers are allowed to join in. In the district seed self-sufficiency program, tbe multiplication of seed from recently released varieties is carried out jointly by farrners groups, extension personnel, and research scientists. The seed produced is procured and marketed by tbe farmers groups tbems<;lves. A part of tbis wheat-research program is continuously supplying genetic materials to several NGOs and actively participating in variety-evaluation work. Thus, with tbe involvement of farmers and various developmental agencies, tbe participatory crop-improvement program could help in (1) the ídentificatíon and release of farmers' preferred varieties, (2) the release of location-specific varieties for diverse agroclimatic ruches, (3) faster 394 M.R. Bhatta el al. Hybridization O F, generation O lo O F. generalion O Advanced lines O Inniai evaluatíon trial (Farmer. part1cipation starts) O Advanced varietat triaj [1 Coordinated !armers' field trial O Farmers' acceptance test O Variety ralease Figure 2. Variety release system followed by National Wbeat Researcb Program adoption of newly released varieties, (4) higher seed replacement and varietal replacement within short periods oftime, and (5) the ineorporation of a gender perspective on farroers' preferences in formal plant-breeding research. Materials and methods A participatory assessment to determine the preferenee eriteria that farmers employ in selecting wheat genotypes was carried out at Bankatti village in the Rupandehi distriet. Twenty male and female farroers were invited, divided into two groups by gender, and asked te list the traits tha! they preferred in selecting wheat genotypes. After Iisting their preference eriteria, individual members were alIowed to rank the preference eriteria. Based on their preference eriteria, 12 wheat varieties with two replieations were planted in farroers' fields in Bankatti vilIage. These differed in maturity, height, tillering, grain size, disease response, grain yield, and other traits. At physiologieal rnaturity, women and men farroers were allowed to seleet wheat genotypes based on the preference traits listed and ranked previously. A format-containing list of 12 wheat genotypes and preference traies listed by the two groups was distributed to both the groups, and they were asked to rank the genotypes from one te 12, Afier harvesting the crop, farmers were asked to rank the grain size and color for their preference. Grain-yield samples frOID the harvest were weighed, adjusted to 12% moisture, and analyzed statistically. 395 Results and discussions The preference eriteria ser by the two groups is presented in table 2. The women's group listed 12 traits, while men listed only níne traits. There were eight traits that were common to both groups. If we compare the eriteria set by tbe two gender groups for selecting wheat genotypes and tbose set by the National Wheat Research Program, we can see many similarities, except in sorne traits related to quality, drought, and sterility. This is because Bankatti is an irrigated and sterility-free area and farmers have never experienced drought and sterility. Table 3 reveals the preference eriteria set by the two gender groups afler ranking in order. It clearly shows the differential ranking of traits by gender. Men gave top prioríty to tolerance to late heat stress, followed by large, white graillS and tolerance to shattering, while women ranked resistance to diseases first, followed by high tillering and high yield. Table 2. Prefcrence eriteria Set by Women and Men Farmer Groups Separately for Selecting Wheat Gcnotypes (pre-Breeding Survey) Wome~n~ ___________________M~en~_________________ 1. Shattering tolerant Shattering tolerant 2. Lodging tolerant Lodging tolerant 3. Good chapati (soft) Good chapati (soltness) 4. High yielding High yielding 5. Disease resistant Disease resistant 6. Medium height Medium height 7. Early maturity Early maturity 8. BoId and white grain Bold and while grain 9. High tillering late heat stress tolerant 10. Reslstant to pests 11. Large spike 12. Short awns Tabla 3. Preference Criteria for Women and Men Listed by Ranking in Order of Priority Men 1. Resistance lo diseases Late heat stress toleranca 2. Early In maturity and pesl resistan! White large grains 3. HIgh yielding Shattering tolerance 4. High tillering Disease reslstance 5. Medium height Lodglng tolerance 6. White bold gra;ns Early in maturity 7. Lodging toleranee High yielding 8. Larga spikes Medium haight 9. Good tasle for ehapati-making (softness) Good taste far chapati-making (softness) 10. Resislance to shattering 11. Short awns 396 UR. Bhatta el al. Tables 4 and 5 show the preference ranking of wheat genotypes by women and men farmers, respectively. Both lhe groups selected BL 14 73--a recently released variety-as lhe number one choice, followed by Nepal 297. The differential ranking appeared when selecting lhe lhird genotype, where women group ranked NL 731 as third and lhe men group selected Bhrikuti (table 6). Table 7 represents the genotypes evaluated, along with lheÍr main characteristícs. It can be seen that the farmers' number one choice is not grain yield but other important traits like bold grain, earliness, and disease resistance. Tahle4. Preference Ranking of Wheat Genotypes by Women Farmers at Physiological Matunty Wheat genotypes Tralts Nl Bl Nl Nl Bhrikuti 753 1473 181 Bl 1124 5 5 4 3 4 872 12 8 6 7 6 3 5 2 5 4 Nl Nl Nl 750 5 Early maturily 297 2 2 11 731 1 4 5 7 Hígh yielding 5 8 4 6 8 High tillering 7 4 6 4 Medlum helght 4 3 2 White and bold grains 1 7 2 Disease resistance Nl Bl Bl 183 10 1692 1810 9 3 12 4 7 4 6 3 2 10 5 5 10 1 2 2 8 5 12 3 4 5 5 2 5 11 9 5 7 6 Lodging toJerance 3 4 4 3 4 1 2 2 4 5 2 6 Large spike 8 9 3 4 4 4 4 10 4 2 8 6 8 5 4 6 4 7 6 6 6 6 5 5 5 2 41 45 56 31 62 60 3 42 5 48 9 6 68 4 47 111 V VIII X IX IV VII XI VI JI 111 Goodchapatí Shattering tolerance Shortawns 4 Total 37 3 62 Ranking order JI X ....OveraJl ranking V IV 8 Table 5. Preference Ranking of Wheat Genotypes by Men Farmers at Physiological Maturity Traits Wheat genotyp•• Nl Nl Nl 297 750 10 11 12 5 1 10 12 731 5 Bhrikutl 2 7 3 6 6 4 4 6 1 7 5 12 1 9 9 Heat-stress tolerante 3 White and bold grain 2 Shattering tolerance 3 Disease resistance 7 Lodging toIerance 1 Early maturity 3 Hlgh yielding 4 4 Nl Bl Nl Nl 753 1473 1 872 9 1 7 781 12 10 4 1 5 4 7 7 2 2 4 10 5 2 1 1 Medlum plan! helght 11 Nl Bl Bl 1692 6 783 8 6 6 3 8 1810 6 4 10 2 7 11 8 1 1 Bl 1724 11 1 1 4 12 6 8 2 5 7 2 11 1 6 7 2 1 54 55 36 48 36 42 VIII IX IV V VII IV VI 1 Good chapati 24 62 24 32 Ranldng order 11 X 11 111 OYerall rankínQ IV Total 111 40 V 20 11 397 Present Status o{ Partlclpatory Plant-Breeding Research on Wheat Table 6. Table 7. Differential Ranking of Wheat Genotypes by Gender Gronp Women Men 1. Bl 1473 Bl1473 2. Nepal297 Nl 731 and Nepal297 3. Nl 731 Bhrikuti 4. BL 1724 Bl 172 4 and BL 1692 5. Bhrikuti NL 753 Wheat Genotypes Evaluated in On-Farm Site Bankattí along with Main Characteristics Olseases Genotype OH OM PHT TGW Graln yield kg ha" HLS lR BL 1692 75 123 44 4000 84 O Bll724 80 123 41 4000 73 O 41 Bl1810 79 125 85 93 91 3575 83 O Nl750 85 126 91 41 3950 73 O Nl731 82 123 96 44 4400 84 O Nl753 79 123 93 44 3750 84 O Nl781 85 127 89 42 3700 73 O Nl783 82 125 95 41 3875 83 O 44 Nl872 84 126 89 3900 86 60S Nepal297 77 123 89 47 3250 85 20S Bhrikuti 80 124 41 3750 83 O Bl1473 77 123 89 99 45 3500 73 O Mean 3804 F test NS CV% 13.24 Conclusions The results of this exercise revealed lhe following: 1. Farmers have vas! knowledge and skills regarding varietal preferences and evaluation techniques (i.c., they can truly idenlifY suitable genotypes and can look very critically al lheir farming system). 2. Women farmers seem to have an inherent ability 10 look at the required traits faster tban male farmers do. 398 References Bhatt., M,R, 2000, Wheat research strategy, NepaL A discussion paper presented during the 22'" n.tional surruner crops workshop held at Lurnle Agricultur. Researeh Stalinn from March 27 to 30, 2000, CBS, 1999. Final eslimales. Kathmandu, Nepal: Central Bureau ofStatistics, Mudwari, A, M,R, Bhatt., and D.R, PokhareL 1998. Increasing wheat production in NepaL In Wheal research needs beyond 2000 AD, edited by S, Nagarajan, G, Singh, and 8,S. Tyagi, K.mal, India: Indian Cnuncil Agriculture Rese.reh, Direetorale of Wheat Researeh, Annex. Existing format for evaluating genotypes in on-farm varietal trials A, Coordinated fanners' field Iríais Questions to fue larmers: 1. Which vanety do you like the bes!? 2, Reasons lor líking the variety a) b) e) d) 3 Which variety do you plant nex! year? _ _ __ 4. Researcher's comment about the vanefies tested, B, Fanners' acceptance test Answer the lollowing questions or tick mark in the eorreet places 1. Wheat variety planted _ _ 2, Date 01 planting _ _ 3. Wheat planted on (a) Khet 4. Fertilízers usad (kgs) (a) Urea _ _ (b) DAP (b) Ban _ _ (e) Potash _ _ (d) FYM _ _ 5. Number 01 irrigatíon applied _ _ 6, Area plantad _ _ 7. Date 01 wheat harvest 8. Grain yield (kgs) _ _ 9. Did you like fuis variety _ _ Did no! lika lO. Reasons of lika and dislike 11. Did haya saved the seeds 01 this variety? Yes No 12, Araa are you 90in9 to plant this variety nex! yaar? Yes No 13, 1I yes, how mueh area? _ _ 14. How did you like !his program _ _ 15. Commen!s and suggestions 01 agricultura tachnieians abou! fuis variety 399 Conserving Agricultural Biodiversity Sunderam Verme Abstract Research on chilis and agroforestry that received the 'Using Diversíty Award' from the Intemational Development Research Centre (IDRC), Canada, is blief!y presented. Introduction The Society for Research and Initiatives for Sustainable Technologies and Institutions (SRISTI) has extended administrative support lo the International Oevelopment Research Centre (IORC), Canada, for monitoring the Using Diversity Awards-a small granl of approximately Rs 3 lacs each, given lo individuals and institutions for conservation of agricultural biodiversity in Soulh Asia. This research award has provided support lo my research ideas and enabled me to continue with experiments, successfully proving that locally developed valieties can be superior lo the commercially released high-yielding varieties. Under the Using Diversity Award Program, the different chili varieties were exhibited at lhe Golden Jubilee celebrations of the Agricultural University, Jobner. Rajasthan. India. The Governor of Rajasthan. along with a large nurnber of scientists, farmers. teachers. and students of schools and colleges (a total of 10,000 people) were among the audience al the exhibi!. Table 1. Salient Features of My Results S. No. 1. i 2. 3. Numberof Unique characteristics as comparad lo varieties developed by varieties , agricultural scientists Neme ofcrop Garlic 12 Hlgher produclivity ln 1 variety Onion 10 6 varielles with hlgher productivlty 22 3 venelles with hlgher produclivity , Cluster bean 4. Sasame 5 1 variety highly produclive and resistant lo red rol disease 5. Green gram 5 1 variety performing balter if sown out 01 season 6, Cabbagel Cauliflower 7. I Fenugreek ! I 1 variety suitable for sowing Agell, Madhyam. and Pichhcholi '8 2 varieties highly productive and 3 vaneties resistant to chachia disease 16 4 vanelies resistanl lo root rot disease 8. Check peal 8eogBI gram 9. Míllet 36 2 varíeties drought resistan!, 1 variaty highly productiva Coriander 31 11 varietíes highly productive Cumin 101 lO. 11. 20 , 12. Total , Chili 2 varieties resistant to jhuJsa disease , 48 , 2 varieties high yielding, 1 variety with higher color value as compared to Rams hybrid sead company 314 Sundaram Verma is a farmer in Rajasthan, India. 401 My experiments in dry farming and agroforestry This ís another area of my experímentation and innovation. I had leamed Ihe basíe lechniques of dry farming in Ihe young farmers' learning eourse al the Indían Agricultural Research Institute (IARI) In 1982. The water needed for irrlgatlon was very searee. The only persisting problem was to stop evaporation and water 1055 by caplllary aetlon. Thi5 could be done by breaklng capíllary movemen! and eontrollíng weed growth by plowíng or díggíng ¡nto the 5011 up lo 20 cm in depth. This should be done twlce-1 O days after the firsl monsoon and jusi after Ihe lasl monsoon. Table 2. Comparative Analysis of the Two Varleties of Chili , Danta variety ' (Sundaram Verma's) Features RS1 variejy (Rajasthan Select) Sowing time Late sowing (April) Sowing usually in February-March Pes! atlack Less More Pes! and disease resistance More Less Cold resislance More Less Flowering AII al Ihe same lime (in flushes) Bul 5-6 limes Not a! !he same time 10-12 limes Time between pickings More Less Produclion 200-225 quinlalslheclare 150-170 quintalslhectare Colorvalue Three limes more Ihan usual Average Dry weighl More Less Markel value ..... _---- Two times mOre ¡han usual Note: Danta selection is compared lO the existing two varielies of chili. namely RS1, Table 3. My Experiment in Dry-Land Forestry Descríplion General procedure Sowing season for tree seedlings Before mensoon Preparatory procedures Pits are dug Inrigation Each plan! requires 15 Ifters of water al least 4 times -------------------4----: Proves lo be more expansiva COS! Sundaram's method I After monsoon Pits nol required : A single plan! requires only one liter once and no extra walering Provas lo be less expensive Type of trees : Masl of!he timas only thomy : shrubs survive Any kind of tree can be planted (fuel wood. fodder. or olher wood-yielding varielies) Survival rale of planted saplings i 50% lo 65% 80% lo 90% 402 Experience, Research, and Facts Related to Local Species of Paddy Chapel Div Bhagal Mali Assistedby Pashupati Chaudhari and Parmananda Chaudhari Bara Abstract This paper presenls a discussion of the qualities 01 local species 01 paddy rice, along wi!h !he advantages 01 growing it-marketability, flavor, use in religious ceremonies~from a larmer's perspective. It also compares produclíon technologies tor local paddy and improved varieties, with information on traditional techníques lor seed management. Use and irnportance of the local species I am Kapil Dev Bhagat Mali, an ordinary farmer owning about one and a half bigha of land in Bara and supporting a family of five members. I have been cultivating the same local specíes of paddy in the traditional method as had been practiced by my ancestors in the pas!. I have both types of land-bhadaiya and aghahani land. (In bhadaiya land there is a winter crop and in aghahani there is the summer crop.) In both types of rice fields, I cultivate local species of paddy. The advantage of the local varieties is their power to resíst díseases and pests, and less labor ís needed in weeding. This paddy gives good tumover despite less fertilizer and less water. Besides beller production, in our religious and social rituals, only the local paddy is accepted and we cannot make use ofthe improved variety. In such sacred rituals, all types of aghahani species and in bhadaiya, the sathhi species are of great importance in the local community. The paddy species locally called Sokan and Sotawa, are even given to patients suffering from high fever, as the rice texture is soft and readily digested. In fact, some of !he local types of paddy are no less in production than the improved varieties. Another advantage of the local paddy is tha! even in less fertile soil, it gives an excellen! harves!. Some specíes 01 !he local paddy have good production in deep-water fields and some other local species are equally good in ralher dry fields too. However, in both these types of land-wet and dry-the improved species of paddy cannot be cultivated. Special qualities of sorne of the local species of paddy I have seen many local species of paddy and have also personally experienced their cultivation. Each of Ihe local species has.its own specific characteristics and qualíties. I am giving a brief description of the specific qualities of the local species of paddy: • Among !he aghahani, the Basmati has a special fragrance, hence il is called Basmati. It is very tasty and very costly. • In Aanga there is tunda; so these local types are no! easily allacked by diseases and pests. • The Dudharaj variety is thick textured, and it is white in color; hence, il is called Dudharaj (king of milk). This species grows equally well in all types of land, wet, deep, or dry land. • The species called Bhathi is grown in deep water. II can grow equally well in fields tha! have water as deep as a man's heighl. • Budhi dayan is another species of paddy that can be cultivated in high or low places. It is good for making beaten rice (chiura) and the rice is also excellent. • Anandi is rather sticky in teXture and is very good for rice dishes. • The Kanna species can be cultivated in both types of land, in high and low areas. The rice in yellow in color and has a very good taste. 403 Experience. Research. and Facts Related 10 Local Species of Paddv • Kamodh is fragrant and delicious and grows equally well with less fertilizer. • Lajhi is fairly good, grows well in les s fertilized fields, is good for daily consumption, and lhe production is high-about two and a half maund of rice per each katha of land, almost equal to the improved variety. • Mansara and Mansari have good straw and are very good for daily use. • Bachi basmati (Chanchur) is fragrant, good for eating and makes good beaten rice (chiura) Production technology and production status of the local and the improved varieties of paddy • The local variety, in comparison to Ihe Bikashe, is much less in production, but in orderlo get a good retum on the production of the Bikashe species, there must be proper provísion of fertilizers, good irrigation facilities, and as they are easíly atlacked by pests, it is necessary to make the field pest free with pesticides. Having provided all these facilities in cultivation, then lhere is only the possibility of getting 3 to 4 maunds of paddy per katha of land. Bu! lhe local species, depending on !he monsoon rain, with less manure and fertilizers and without the use of pesticides, can produce paddy in the same ratio as the Bikashe varíety. • In the case of paddy saplings to be replanted in the fíeld, the timing is very importanl for the Bikashe, butforthe local species, Ihere is no such limitation, as Ihe saplings can be either 15 to 20 days old or 35 to 40 days old. • The straw (para/) of Bikashe paddy from one katha of land is equivalent to the straw from local species coveríng an area of 10 kathas. • If the family members are few in number, il is nol wíse to cultivate Bíkashe species because il requires greal labor, as in sprinkling pesticides, spreading fertilizers, weeding, etc. • Those farmers who cultivate Bikashe paddy make use of tyanki but I have never used any such thing. • I experimented by cultivating two species of paddy in one field: in half the field, the Bíkashe species callad Masula, and in the other half, local varieties callad Dudhraj and Chhataraj. In both, I used equal measures of irrígation and fertilizers, and I dídn't find much difference in paddy production. I usad 2 kg of urea fertilizer in each katha of land. • Though I have the facliity of irrigation in my fíeld, I will no! discontinue cultivating the local paddy because, with good irrigation, lhe rate of production of local species will also be equally high. Cost differences between local and Bikashe (improved) paddy The cosl of Ihe Bikashe paddy called Mansuli is Rs. 100 more per quintal lhan the local species of paddy. Bul rice varieties like Basmatí and Kamodh are one and a half times more lhan Mansuli. During festivals like the Tíhar (Dewail), Ihe cosl of local species of paddy like Sathí and Khera is higher than al olher limes. Besides the produclion of rice straw (para/) being more, if the sale of rice and straw can be made together, there is greater profit in Ihe cultivation of Ihe local breed of paddy. Role of farmers in improving paddy harvest In improving lhe paddy harvest, farmers have very importan! roles lo play. If it had not beeo for the farmers who carefully kept the seeds of different species of paddy, il wouldn't have been possible 404 Chapel Div Bhagal Mali forthe scientists to bring about the improved variety ofseeds now.ln facl. Ihey would nol be able to see the presenl species of paddy Ihat we have now carefully conserved among uso Sesides. if we had nol broughl seeds from our relatives. neighbors, and olher people. all Ihe different species found now would nol be available. It is because we have conlinued to cullivale with Ihe old, lradilional methods, selected Ihe seeds, dried them, and stored Ihem locally Ihal these local species are still found now. Therefore, lhe extensive practical knowledge Ihat we have of Ihe local species of paddy. now, is also nol wilh Ihe researeh scientisls. Non-formal seed-management techniques We are eontinuing lo apply Ihe same old lechnology in seed selection. When we keep the seeds for planting, we eilher select Ihe seeds after we bring Ihe paddy into Ihe huge circular space (khalihan), or after the paddy has been de-husked, or from lhe slorage room itself we select lhe seeds for Ihe next planling. Since Ihe different speeies of paddy are repeatedly Ihreshed in lhe same spot (khalihan), Ihere is Ihe possibility Ihat every year Ihere is some degree of mixing up of seeds. Then there is again the possibility of mixing up seeds in slorage. or by planting seedlings logelher in Ihe same nursery. 50metimes lhe flow of water carries away the seeds and mixes up the seedlings. or Ihey are mixed up in Ihe process of keeping henga, or due to birds and animals. So Ihere are many chances of mixing Ihe different breeds and species of Ihe paddy. Some farmers deliberalely mix the differenl species of paddy in order lo secure a good profit in lhe market. For this reason. these farmers mix togelher varieties of lhe same type of color and síze so lhat nobody can see the discrepancy and find out lhe farmers' profit-oriented molives. For example. il is found Ihal farmers mix Mansara in Basmati. Mansara looks ralher like 8asmati bul is Ihicker in texture. After Ihe paddy has been separated from Ihe stalks by threshing. il are kepl in Ihe kha/ihan (Ihe huge circular floorfor threshing Ihe paddy) for a day or two lo dry, and aftar Ihal , it is weighed and stored in bhakari (huge cane baskets painled wilh a pasle of mud and cow dung). If Ihis storage is . full. Ihen Ihe paddy is kept in sacks. and Ihe nexl year. Ihe seeds are laken from lhis stored container for paddy plantalion. If il is necessary lo bring seeds from neighbors, Ihen lhe seeds are exchanged at lhe ratio of one saek of paddy seed lo one and a half of de-husked rice. If Ihe seeds are brought directly from the khalihan, Ihen Ihe exchange rate is Ihe same in ratio. Likewise, sometimes the seeds are borrowed from relatives, somelimes boughl from the markel or at times exchanged or even boughl from olher farmers. Training and tours I have taken two training courses. aboul one or two days in polato cultivation. In lerms of lours, this year I have visited lhe PaddylRice Harvesl Research Center in Hardinath, organized by NAARC, U-SIRO, and IPGRI, where I gol lo see Ihe excellent species of paddy cultivation as well as our local varieties of paddy, which were grown there for study and observation. So now, I feel lhe necessity for us to be given training and skills in proper melhods of keeping seedlings lo maximize production from our local paddy species as well as technology in irrigating and fertilizing. Conclusions I am a small-scale farmer, eultivaling a small plol of land and sustaining my small family, trusling the maxim lhal 'contenlmen! is bliss," and I have continued wi!h the tradilional lechnology of local paddy cultivation practiced by my aneestors sinee Ihe remole past. I know lha! !he local variety of seeds have many good qualitíes which are no! available in the improvedlBikashe seeds. The local varie!ies have less chance of being altacked by diseases or pests, they are able lo withstand climalie condítions, and give good produclion with les s fertilizer and less care Ihan Ihe improved varieties. Besides Ihese, Ihe local varieties are delicious for eating, are acceplable in religious ritual s, 405 Experience, Research. and Facts Re/oled lO I,o~c,,=a~1S2p"'ec"'ie"'s'-'o',L[P'--a...d=dz..y_ _ _ _ _ _ _ _ _ _ _ _ __ and have medicinal properties, Some varieties al so have a higher market value than the improved variety of rice. II is nol only difficult and labor consuming to cultivate the improved variety of seeds, bul in lerms of rice production, there is not much difference if the method of cultivating both types of rice is carried the same way, I believe that farmers who are commitled lo developing improved varieties of paddy seeds must give equal importance to protect and promote the local varieties of seeds because these seeds are the basis for future improved varieties of species that can give beller produclion al harvesl. So if the government ilself took the iniliative to give the required training and make arrangements for educational tours, then Ihe farmers could contribule beller in farming and cultivation. We farmers are much behind in farming skills. We do nol have the knowledge of!he latest technology in farming, So we follow the same old technology that our ancestors have followed from past times. About selecting and keeping the seeds, storing Ihem. we have the same traditional techniques, When we don'! have seeds of our own, then we borrow from our relatives and friends as we don't trust the seeds that we buy out in the market. Neither have there been any effective measures taken by Ihe government to ensure good storage syslems for seeds lo have beller produclion. Suggestions and recommendations I request our farming community brolhers and sislers lo continue lo cultivate our local paddy. Though less in quantity, the local species have less of a chance of pesl and disease infestation, require less fertilizer. and grow equally well, Even in insufficiently irrigated fields, the production is good, and Ihese local varielies are acceptable in our sacred and social rituals. As cooked rice also, il is delicious and it is good tor our health, In poor soil also, il gives good production. If we lel !he local varieties and species disappear, than there is less chance for developing improved seeds for the future use, Today !he need Is for an initiative tha! must be started from !he govemment level lo protect the local paddy seeds, which are fast dísappearing and unaccounted tor: lo conserve, lo protect, and to highlight them through local development clubs, district-Ievel agriculture development offices, the Agriculture Ministry, departments, etc. The scientists must not only work towards making !he rice grains trom long to short or bringing grealer production, but they must also make provísions for Ihe availability of proper irrigalion syslems. In a similar manner, the government musí also make provision lo supply fertllizer in time, good seeds and Ihe means lo control and destroy the diseases and pesls that attack the improved varieties. - - - - - - - - _ __ . _ - - - - - - - - - - - - - - .. 406 Problems of Maize Cultivation and the Role and Approach of Local Farmers in Solving this Problem Srí Harí Prasad Aryat Secretary Maize Cultivaüon Research Committee Darbar, Gulmí Abstract Maize is the maln crop in Ihis village in Ihe Gulmi dístrict ofNepal. Farmers from lhis village discuss Iheir experiences participatíng in maize-ímprovement research wilh Ihe Local Initíatives for Siodiversity Research and Development (U-SIRO). The farmers preferred Iheir traditional variety for its lIavar and fadder, bu! it tended to fall. This papar describes how Ihey warked with U-SI RO on maintaining Ihe characteristics they preferred in Iheir traditianal maize variety, while worklng lo overcome the problem af íls weak stems. The majority of the people in the Durbar Devisthan VOC in the district of Gulmi, which falls under Ihe Western Development Region, are dependen! upon agriculture. In this area, the prominent crop is maize. Olher species of maize are also planted, but in this hilly region, the most common species is the large-sized yellow-colored variety. This type of maize is remarkable tor grealer production and the higher amount of fodder for animals, and Ihe corn Is delicious to eal. So the majn crop here is maize, bul farmers are facing a great problem as Ihe malze plants have the tendency lo fall. Olher problems are relaled lo Ihe lack of irrigation facilities, the problem of pests and diseases, and also Ihe scarcity of better technology lo improve our crops. . We are making our own efforts to solve the problem of maize plants falling down easily. What we do is when the plant is about 20-30 days old, we dig and weed out the grass around the saplings. Then when the plant is about 45-55 days, then there is a special type of digging done just to loosen the roots, and we cut one side of Ihe root, loosen il, and lay the plants on the soi!. This stops Ihe growlh of Ihe height of the maize plants for so me days and the rool is strengthened. Some farmers also plan! maize in straighl line. Sometimes, aflerweeding out the grass around the maize saplings, Ihe soil is pressed around il. In Ihis way the farmers make great efforts lo prevent Ihe maize plants from falling down easily. Initially, when the Locallnitiatives for Siodiversity Research and Development (U-SIRO) came visiting our village, we were most happy. AII of us villagers gathered together. At that time they were in the procesa of choosing the appropriate place to carry out their research. We requested them to carry out Ihe program in our village only, so Ihal we too could participate in the programo Afler that, il was settled tha! the program would be conducted in our village. Under Ihis program, they told us their plan-that they would give us new improved species of maize seeds, and we could choose those species that would be most suitable under our soil and climatic conditions. About this, we were not in agreement as we knew Ihat the species of maize tha! we locally grew, the big yellow variety called Piya/o, was good for uso We were not willing lo change from the local variety because this species gave a good harvest, the quantity of com-liour after grinding was more, il was delicious eilher dry roasled or grilled over the lire, and the maize plants also provided excellent fodder for our animals. The U-SIRO teachers had asked us whether we would grow a species of maize Ihat was very similar lo the Piyato variety in taste, production, and as fOdder, bu! the maize planta would be shorter in height. We suggested tha! we didn't want any of Ihe new variety bul we would be happy to be taught Ihe techniques to prevenl the Piya/o maize planls from falling down easily. Then we asked if there could be improvemenls in this variety. So as we wanled, LI-SIRO is helping us lo improve the big variety of Piyato. We have started a cross-pollina!ing program in which we have crossed the big 407 Problems ofMaize Cultivation and the Role ond Approach o{Local Farmers in Solving this Problem variety of Piyalo with improved varieties. Another program is mass selection. In this, the main worl< is lo improve the big-grained Piyalo itself, by separating and selecting seeds from tal!, weak maize stalks, identifying and marking poor seeds by removing the straws from such ears. We have done other programs too, and ourfarmer brothers have experimented by growing other improved varieties. Some of the species have also been appreciated by our folk. Likewise, we have also considered Ihe 36 varieties of maize species brought by CIMMYT, and from among them we have selected some species for cultivation. In this way, U-BIRO has helped us to solve our problem and to bring improvement to the cultivatíon of malze. We have been given training in how to do cross-fertílizatlon and mass selection. We farmers are also commltted lo improving the species of big Piyalo and we are very optímistic about this improvement. It is true that we have leamed about crossing the maize species and plucking out Ihe straws from weak and tal! maize stalks in order lo selee! good seeds, but sliII, we feel thal if we are Irained in other melhods of improving our crop production, we would be capable and successful farmers. Therefore, Ihrough mutual efforts, underslanding, and cooperation among our farming community and NGOs, we would surely be able to bring out improved maize specíes líke Ihe yellow Piyalo variety according lo the suitabilíty of our soll and weather conditions. 408 There Is the Possibility of Simichaur Becoming Makai Chaur: My Experiences in Crossing Maize Species Mrs. Lal Kumari Basnel Daha Simichaur, Gulmi Abstracl A maíze farmer describes her experiences leaming and using new technologies tor improving her maize crop. From the earliest times, Simichaur was known for cultivation of maize. Every year we suffered losses through Ihe falling down of maize plants. Now we have an office Ihal helps with maize cullivation, so we are happy. The species we grow is the big yellow variety, but Ihe main problem wilh Ihis species is, Ihe planls fall down. But we are still conlinuing wilh Ihis species, only beca use il is delicious lo eat, eilher roasled dry or as corn meal; il gives more eorn flour, grows more, and gives much fodder for our domestic animals. In arder to proteel Ihe planls from falling, we have a proeedure by whieh we weed out Ihe grasses and cul Ihe rools on one side. In Ihis way, Ihe growth is controlled and the plant is strenglhened. Bul Ihen Ihe faHing process conlinues. Then we slash oul Ihe leaves. This makes Ihe leaves less dense and Ihere is plenty of spaee for Ihe movemenl of winds. Even Ihen Ihe plants fall. So whalever means we have laken unlil now lo proleet our plants from falling, we still have no solution and Ihis problem continues to be among uso The teachers carne. AII of us-Ihe village farmers-galhered. They asked us whal speeies of maize we cultivaled. We replied Ihe species we grewwere Ihe big yellow variety Khumaltarand the small yellow variety. We lold them Ihat we liked Ihe big yellow variety bul !hese planls always fall. We asked them if Ihere was Ihe possibility of making Ihis big yellow variety nol fall. The leachers said Ihat Ihey had not come lo make Ihe maize plants nol fall bul Ihey came lo visi! Ihe villages in order lo introduce new species of maize thal we eould grow in Simalchaur. If Ihe local people didn't want lo change Ihe species of maize they have been cullivating and if !hey wanted to find the means to preven! Ihe planls from falling, they could all work logelher to find the solution lo their problems. Then we exchanged ideas as lo how lo prevent Ihe plants from falling, and we farmers became participanls in Ihe discussion. We discussed Ihe ways and means of solving Ihe problem and Ihen we reached our unanimous decision lo creale a new species by making Ihe yellow variety of maize smaller in height. In my field, we crossed belween Ganesh-1 males and big yellow variety females. We planled three lines of male seeds and six lines of female seeds. Likewise, we advised some five olher farmers lo cross-fertilize Ihe big yellow variety of maize with olher species. While cross-fertilizing, we planted Ihe male and female seeds separately, with a difference offive days. But all the time we were worried Ihal Ihe seeds mighl nol grow. The seeds grew, bul maybe due lo Ihe growth of Ihe grass or something else., the saplings became yellow and Ihey didn't grow well. Then I was very much frightened. I was worried Ihat if we didn't have a goOO maize harvest, how are we going to survive? Then I visited the fields of Ihose farmers who were asked lo cross-fertilize maize seeds Jike me. At thal time my husband came home. He was also surprised atwhat I had in the fields and asked whal we were going lo eat if we didn'! have a good maize harvest. I assured him Iha! !here was no! mueh to worry about as we had in our village the experts working in maize cultivation, so we would be happy if we could find out the means lo stop the plants falling and preven! Ihe yeariy losses. So even if we didn'l have a good maize harvest this year, there was still hope for a better return in Ihe next year. 409 There Is the Possibilitv o[ Simichaur Becoming Makai Cltaur In the hope that the teachers would come and Instruct us, we had not prepared Ihe field for cross-fertlllzlng. Bu! they carne late. Here the seeds we had planted could not grow properly due to weeds. Then I took the rlsk, whatever the consequences, and we began lo dlg and lo loosen the 5011 around !he malze saplíngs. Then we pul In a líttle urea and the malze plants showed good results. Then we put one teaspoon of urea Into each planl and the resul! was excellent. Then !he teachers came and gave us tralnlng. In tha! !ralnlng, ¡ ha ve lo know thal even malze has male and female f1owers. I unders!ood tha! even maize has a cross-fertilization process. When we cross the malze plants, we have lo take out the straws from the ears of corn, and we were asked to take out the straws from Ihe female maize plants. I was amazed Iha! taking out the straws from !he maize could result in !he better growth of malze kernels. Even my father used to say If we remove the straws from the malze, there wouldn't be good corncobs. Then it was time lo take out Ihe straws from Ihe malze. I was slill uncertain abou! the outcome of Ihis bul I reassured myself Ihat since we were belng told by people who are experts in maize cultivalíon, I shouldn't doubt. AII the maize didn't grow the straws al Ihe same time, so every day I went and pulled out the straws. After this work was completed, !hen it became my habit lo go inlo !he field at least !WO times to sae the result. I was all the time worried about whether the maize cullivated in one ropani of land would yield as much crop as we expected. Every day I saw corn cobs growing on the maize plants. Then I wanted lo find out if the cobs had kernels inslde. I was nol without worrles. Then one day I pulled out the eorn eob, took out the outside layers and looked inside. There were beautiful kernel s and now I was at peaee. Now I was confident and happy that we eould bnng out a new species of maize. Now !hal we have learned !he teehnology to cross species, we want to learn how and when the new species will come lo be, how lo grow many species, and if the maize is nol well filled wlth kerneis, what is to be done? If the seeds are of mixed variety, what ís to be done? What measures can be taken lo ensure the qualíty of one particular variety only? Ifwe can get answers to such quenes in our minds, 1am sure our village called Simichaurwould be changed lo Makaichaur. 410 How Did the Farmers of Chhomrong Improve the Local Paddy Species? Participant Representatives: Mr. Om Bahadur Gurung Mr. Najarman Gurung Mrs. Min Kumarl Gurung Mrs. Nauli Gurung Ghandruk VDC., Ward no. 9, Chhomrong Abstract Farmers frem the village ef Chhemrong in Nepal describe their experiences with experimental varieties ef paddy rice, grewing experimental varieties and selecting for specific qualities, in an initiative tha! started in 1993. Introduction Chhomrong ís a 5mall village situated at the foot of the Machhapuchhare and Annapurna Range of mountaíns. It lies al Ihe heighl of 1800-2000 meters aboye sea level, on the Pokhara-Baglung Highway, six to eighl hours' trek lo !he north from Nayapul. This village has a majority of Gurung residents. II is believed that in the year 1962/1963 pakhe red rice paddy species was introduced. This species was originally brought and cultivaled as a specimen by aman who lived in Lumle VDC and had worked in Shillong, India. Gradually, this paddy was found to be planted in other villages too, and in Ihis way il gained enlry into Chhomrung village. Here people began lo cullivate Ihis variety of paddy loo. To improve upon this species, the Lumle Agriculture Center took the initiative and this variety was recammended as the Chhomrong local. From the yaar 1993, this villaga was selected for research under !he Lumle Agricultura Centar for the Participatory Plant Breeding Programo The main objective of this program was lo improve !he species oflhe local red paddy.11 is rather hard textured and took long time in husking, and as there was no alternative species Ihat could tolerate the cold of the local place, so research was begun to improve upon the local variety so Iha! il could wilhstand Ihe cold and make!he grain white ín color. In the initial stage of the program, 250 specíes of paddy variety brought from !he national and international paddy research program were planted in the nursery on an experimental basis. These different species were carefully selected and kepl under the joínt care of the technicíans of the Lumle Agriculture Center and the Chhomrong VDC participant farmers. The selection was as follows: • Chhomrong Local ripened in its usual time periodo • The straws were !he same as that of Chhomrong. • Good grains of rice. • Capable of tolerating diseases and pests. • Capable of tolerating the cald. These species were carefully selected and those that gave white grains of rice were particularly taken care of. Along with !he experimental nursery planting, the participant farmers had also planted lines of paddy species tor the experiment. The species used for the experiment were as follows: 41l How Díd ¡he Farmers o[Chhomrong lmprove ¡he Local Paddy Species? 1. Machhapuchhre -2 2. Machhapuchhre -3 3. Machhapuchhre -4 4. Himchuli -1 5. Himchuli -2 6. Nilgiri -1 After planting the selected species, the technicians, the scientists, and the participant farmers visited !he fields and carefully selected the species on the spot. The process of selecting the species started from Ihe year 1993 to 1995. Then those thal was selected were carefully studied by using different methods. • paddy abundantlnot abundant • giving goOO rice grains/not giving rice grains • rice breaking into pieces/not breaking • rice swelling/not swelling • good rice straws/not good rice straws • delicious rice/not delicious • not quickly digested/quickly digested • lime taken to Ihrash paddy The process of keeping and selecting these qualities in Ihe species was initiated asa joint venture' of the Lumle Agricultural Research Center and participant farmers. It was formally handed over to the Chhomrong Agriculture Oevelopment Committee. Then this Chhomrong Agriculture Oevelopmenl Committee requesled !he U-BIRD Organizalion for assistance lo give conlinuity lo Ihe research programo Accordingly, the agreement was reached belween U-BIRO and the Chhomrong Agriculture Research Commfttee lo carry ahead the joint participatory programo In 1997 ano!her improved species of paddy called Machhapuchhare-9 was being experimentally cultivated and developed. We request that U-BIRO will continue lo assist us in giving conlinuity lo our paddy research programo 412 Local Species: Methods of Cultivation, Sorne Successes and Sorne Problems Ram Ashraya Saha Kalewar Bara Abstract Farmer's description of paddy species. melhods of cultivation. and use. Introductlon My house is in the district of Bara at Karchowa VDC. For the las! few years, I have !aken land on lease, and I am cul!ivating the local species of plants. These species are quite rare species, which I am certain can'! be available in any par! ofthe Bara district. Why I am interested in cultivating these species is that the condition of lhis land Ihat I have laken on lease is mosl suitable. In other words, Ihe type of land Ihat I have leased is such that it is always water logged, where neither any of the improved variety of seeds grow nor any of Ihe usual local varieties. The positive quality of this land is lhat during the monsoon, Ihe water collects. It is so deep that sometimes men can be drowned. I brought the seeds of Ihese species from Ihe village al Rautahat distric!. The following are some of the details of Ihe species I have been cultivating. Paddy species: 8hathi The paddy species called Bhathi was bought from the Uchidiha VDC from Ihe village called ltawal at the price of RS.15 per kilo. I have been planting this species of paddy for the las! 10 years. Every year, I have planted this variety of paddy in abou! one bigha of land. The return has been 18 maunds (720 kilos of rice l. Bas and Basthan These species of paddy can be cultivated in areas where Ihe water is deep. Therefore, this variety needs a place where there is always water available. This type offarming can be done up to nine months only. The reason is that at the time when we sow the seeds, the land has to be dry and in the month of Chitra ( March/Aprill, the sowing of seeds is done. The harvesting season is always in the month of Mangsir (NovemberIDecember) regardless of Ihe time paddy may have been planted. The roots extend from the rice stalks on Ihe water's surface so plants seem to balance on lhe surface of the water. Therefore, there is no possibility of the paddy plants falling, however deep lhe water level may be. The other good point about Ihis paddy species is no malter how much the water level may in crease during the night, Ihe rice plants seem lo grow in equal propor!ion. External appearance The length of the rice stalk is almost Ihree meters high. Since the stalk is pretty thick, there is less possibility of the rice plants falling over. The leaves are broad and light green in color. The roots are quite strong so Ihat Ihe normal sor! of f100d can't drag them out and sweep away lhe plants. The rice grains are shapely and big. The rice grain is red in color and there isn't any tunda. Method of cultivation In the month of Chaítra ( March/April), the tractor is used to plough the land and Ihe seeds are sown. At this time of the year, the land is dry. A few days after lhe sowing of the seeds. the rain star!s and the land starts getting filled with water. As the water level rises, the length of the rice stalks grow. As the species grow in water, there is less possibility of the plants being atlacked by 413 Local Species: Methods ofC"lti¡lation, Some Successes and Some Problems pests and diseases, Bul Ihe rice plants are infested wilh ínsects called gawaro. This varíety of paddy neither requires weeding out grasses nor Ihe need offertilízers. The harvesting time for Ihis species ís the month of Mangsir (November/December). In this monlh, the level of Ihe water comes up to Ihe knee and the paddy can easily be cut down wilh Ihe help of a curved knife called hasiya or kachiya. After the paddy is cut down and carried home, il is dried, beaten, threshed, and sto red wel/, ready for use. Uses and its importance The rice is softand delícious to eal. It is readily digested. From this rice beaten rice (chíura) can be made, and varielies of local delicades líke bread, thakuwa, fried bread, etc., are prepared The paddy of Ihis species is accepted in our sacred rituals and practices. This variety of rice is general/y used in feasts and parties. Possibi/ity of improvíng the qua/mes of this specíes of paddy I bough! Ihe seeds oflhis species of paddy at RS.15 per kilo. Therefore, Ihe market price ofthis variety of rice can'! be less Ihan RS.18/- to RS.20/-. Duríng festival s, Ihe price of this species of rice goes up to RS.30/-. So there is no loss in cultivating this varíety of paddy in our fields. This is very advantageous, especially if Ihe land is being left unused due lo water-Iogging conditions; thus, farmers can maximize their profit. Sin ce the labor is less, any one can farm Ihis variety of paddy. Need for conservation of the specíes This endangered variety ot paddy, having such odd qualities, is in Ihe process of being losl. If we do nol direct our attention in time, its extinction is certain. Now there is an increasing tendency to fil/ up the water-Iogged land and convert il into residential areas. This is a really dangerous situation because il ¡ncreases Ihe possibility Ihat Ihis species will be completely wiped oul. For Ihis reason, both the govemment and Ihe NGO must take the iniliative lo conserve this varíety of paddy and work towards helping Ihe farmers wilh Ihe technology lo improve their living conditions. Name of paddy: Amadhouj and Sakhar Both Ihe stalks and Ihe ríce grains of these two varieties look Ihe same. The two species are cultivated in fields that require less depth of waler than for Bhafhí, bullhe Sakhar paddy is shorter than Ihe Amadhouj species. Both are cultivaled in a similar manner. They are sown by spreading the seeds on the prepared fields. They require less fertilizer but need to be weeded. If Ihe fertilizer is loo much, Ihen there is the possibility of the paddy plants falling. If the cultivation is done wel/, the return harvest can be as much as four to six maunds (160-240 kilos) per each ka/ha of land, The rice crop is long and the grains are fal and heavy. This year I boughl bolh Ihese species of paddy trom the village cal/ed Sonarniya in the Raulahal distríct. I cullivated Amadhouj variety on Ihree katha of land and Sakhar on one bígha of land. But in the field there was too much watercol/ected, and Ihe paddy was good only in two ka/has of land. These two varíeties of paddy are acceptable in our sacred rituals. From Ihe rice straws, we can make floor mats. The cooked rice is good and any person who is sick can ea! this rice without any problem. In the Amadhouj stalks, three or four grains of paddy grow together, so Iha! at a glance, Ihe paddy crop looks almosl like wheat. But there is a superstilion attached lo the cullivalion of Ihis paddy. II is believed thal iflhe paddy grows In equal measures in thefourcorners ofthe field and the produc- 414 _ _ ~ ... _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _-,R~a!!:m!..;A:!:s",h:..tra!Ly~a",Sa",h",a-.:'lG",a",le",w",-ar líon is equal in all the fourcorners, then the farmerwho has such a harvest will suffer so me evil, like some one wiU be sick or some one in the famíly may die. Species of paddy: Khera In the month of Jestha (May/June), the seeds are sown, and in the month of Asadh ( June/July). the rice saplings are planted, and in lhe month of Mangsír (November/December), the paddy is harvested. As the level ofwater is less. the paddy needs to be weeded. It can also be given fertilizers. I had personally used three kilos of urea and 2.5 of DAP in Ihe paddy fields. As the paddy grains have tunda there is less possibility of disease and pest problems. But !his year, the paddy was infested wilh bolh disease and pests. and I had lo use the pesticides called che/amín and metacid. After taking all these measures, the harvest was three maund (120 kilos) per each katha of land. This variety of rice is importan! from a religious point of view. In the worship of our family God, Gobín Maharaj. this variety of rice is absolutely needed and no other variety will do. The cooked rice from this species is equally delicious. From the rice straws, the f100r mats are made and even the cattle have a special preference for lhis type of rice straw. This variety needs less fertilizer. 415 Role of Farmers in Selecting Crop Species Jhapad Bahadur Bhandari Gitanagar VDC, Ward no. 8 Indrapuri, Chitwan Abstract Farmers describe their activities investing in and developing their agricultural output, with details of their production and Suwarna paddy rice (an improved variety). Introduction We have been following the traditional method of cultivating crops according to the methods used by our ancestors. In those early days it was not possible to know the kinds of soil, the type of crop that was suitable to the kind of soil, the types of pests and diseases, the methods of controlling them and how much fertilizer is required for a particular type of crop, and hence, the harvest was not satisfactory. I feel the govemment is not much interested in the agricultural sector, but without improvement in fhe agricultural sector, there can't be improvement in industries and commerce. In every country the role of agriculture is prominent. Because of the three basic human needs-food, shelter, and clothing-food occupies the prominent place in man's life. The basic need of all sorts of living beings, from the wealthy to the poor beggars and birds and animals, is food. Without food, nobody can live; it is a universal fact. This point is most significant for us to understand. Nepal is an agricultural country, so the people here would be most happy if they were given knowledge about the formation of land, the types of soil, means of irrigation, and given the priority to develop improved seeds for local use. In 1929 we migrated from Lamjung to Geetanagar VDC, Ward no.8. At that time there was no irrigation system in the village. The field had mixed cultivation of different species of paddy in the same plot, like Dudharaj, Aap jhuthe, Battisara, Gola, Mansara, Thapachini, Jetho bure, Ghaiya, etc. Development In terms of harvest, there used to be 25 to 35 muri of paddy per bigha of land. At Tandi, we grew maize and mustard. There was no system of cultivation by rotation. Later on there carne new species of paddy called Achhami masino, Mansuli radha-4, and Radha-17. Among them, the Mansuli was the best, so there was extensive cultivation of this species of paddy everywhere. Vegetable cultivation was limited to small kitchen gardens for vegeta bies like green vegetables and radishes, but later, from the year 1937, with the assistance of the Agricultural Development Branch Office, we began to grow vegetables on a larger, commercial scale: improved species like cauliflower snowball, Kathmandu local, snowcom cabbages, radishes, carrots, mustard, etc. Now, these improved varieties require good irrigation systems, so we took a loan from the Agriculture Development Bank, dug out deep wells, kept motor engines, drew out water, and kept sprinklers to irrigate the vegetable fields. Now we have at our rescue the LI-BIRD Organization, who, by keeping in contact with the different research centers, have made available for us new and beller improved seeds, like Rajma, Panta-11, BC 1442, PNR 381, and Sarvati. These crop species are suitable for the soil at Tandi, but in water-Iogged fields where the paddy plants tend to fall easily, the Sawarna species of paddy seems to be appropriate. 417 Role of Farmers in Selecting Crop Species Experiences regarding the sawarna species This variety of paddy is excellent in churned fields. The plant stalk is strong and does not fall easily. The rice is tasty, good in texture, has weight, has salid grains, is easy at milling and at threshing prior to milling. Seventy-six percent of the rice grains remain during the processing period and do not break easily into pieces in the de-husking process, and unlike the Mansuli species, even when the plants are shorn of green leaves, there is no difference in the usual production of rice kernels. Last year for the research study, the U-BIRD Organization and the Agricultural Development Branch office at Bharatpur made improved seeds available, and so we were able to cultivate the sawarna variety of paddy in 10 ka/ha of land. In terms of production, it gives us four muri of harvest per each ka/ha of land. This year we harvested 80 muri of paddy. In the village of Indrapuri, for Tandi there are varieties of paddy species that can be cultivated, like BG 1442 and Panta-10, IR13155. We farmers are very happy now because a mini-kit has been prepared for the farmers with the different species of paddy, exhibition of research results, trial checks, specimens, etc. From such programs, we have highly benefitted and learned that now we can choose by ourselves which of the species would be most suitable for our land, soil, and climatic conditions so we can be selective in cultivating crops ourselves. 418 Role of Farmers in the Improvement of Crops: New Species of Paddy in Maramche Mr. Chandra Kanta Poudel Assisted by Mr. Dil Bahadur K. C. and Mrs. Sita Poudel, Maramche, Kaski Abstract This paper describes the agricultura! siluatíon in Maramche. Village larrners have beeo partidpating with Ihe Local !nitialives for Siodiversily Research and Development (LI-SIRD), testiog improved varieties 01 paddy rice. Details of the qualities 01 lhe new varietíes. qua!ilies preferred by Ihe farmers, and experiences wilh !he participalory plant-breeding program are included. Introduction Maramche is a small hilly village in the Dhikur VDC near the Lumle Agriculture Research Center. It is aboul 30 kilometers from Pokhara, the tourist area in the Kaski district, lo !he northern side of!he western Baglung Highway. II has a cold climate and a diversity of ethnic tribes. II is 1600 melers aboye sea leveL In Ihis village there are about 36 households only. Allhough the village contains a diversity of ethnic tribes, the majority of the people belong to the Brahmin caste. Just as Cheerapunji is the world-renowned place for rain, likewise Maramche is also the village in Nepal that has the highest rainfall. . Due to poverty in !he country, many of the people have left their homeland in search of work, while the rest are dependen! on agriculture. Looking into Ihe picture of farming in this village, we find much diversity. The following table helps to illustrates the important crops and vegetables grown in this area. Farmers Land ATea • Muen Mueh ! less Paddy. maize. wheal. millal . Less Patatoes. mustard, radishes, beans (bodi, simi, bakulla). gourds • (ghiraula), pindalu. sOybeans, iskush (squash), karela (bitter : gaurd), garlic, onions : Caulifiower, carro!s, salgam (!urnips), kerau (chick peas), sugari cane, green vegatables (chamsur, palungo). jau (barley). uwa, i gaha!, lentils (arhar) tarul, turmeric. gingar Among the crops mentioned in the table, paddy is most extensively cultivated. But unlike!he other crops, paddy has the maximum number of local varieties. The following are the names of the local species: 1. Kanthe 2. Kalo patle 3. ReksaJi 4. Mansara 5. Juwari 6. Si/ayam 7. Tarkange 8. Chhomrong local 419 Role o[Farmers in (he Improvement o[Crops Introduction of M-3 and M-9 As Ihe above-menlioned names of local paddy species show, Ihey are many and Ihey possess Ihe qualities lo wilhsland maximum rain and Ihe cold climalic condilions of Ihe place. Therefore, as Ihe people had no olher means lo find oul aboul species Ihal have Ihese qualilies, Ihe local people conlinued lo cullivale Ihe species available lo Ihem locally. In Ihis conlexl, in Ihe year 1996 wilh Ihe cooperalion of U-SIRO and Ihe local farmers, Ihe Participalory Planl-Sreeding Program was inilialed.lmmedialely afterlhe program was slarted, U-SIRO dislribuled improved seeds M-3 and M-9 lo five farmers in Ihe village. After Ihe paddy was planled, il was found Ihal Ihese species of paddy were capable of wilhslanding Ihe climalic condilions of Ihis village. So Ihe people of U-SIRO and Ihe Crealive Molhers' Group wenl lo Ihe field area lo inspecl Ihese species of paddy cullivaled on an experimenlal basis. The inspeclion and sludy of Ihe paddy was made on various faclors, like produclion, Ihe heighl of Ihe paddy slraws, and Ihe shape and size of Ihe rice grains. After Ihorough mulual discussions and giving priority lo Ihe farmers' inleresls, Ihe species M-3 and M-9 were accepled as Ihe righl choice for cullivalion. The reasons for their acceptance 1. Maximum produclion 2. Tolerance of fertilizer 3. Tolerance of Ihe cold climalic condilions 4. Abilily lo lolerale wel and moisl condilions 5. Suitable in less waler 6. Quick ripening period 7. Less waslage Ihrough falling ( nol Ihal Ihey will nol fall wilh Ihreshing) Nature of the popularity of M-3 and M-9 species 1. After Ihe inspeclion, Ihe requesl for Ihe improved seeds as per Ihe needs of farmers 2. Sased on Ihe advice of Ihe farmers who inilially cullivaled Ihe improved varielies 3. Inleresl in cullivaling new species of improved paddy, and by Ihe dislribulion of seeds 4. Sased on Ihe requesl of local organizalions to do Ihe experimenl again and mulually exchange and share ideas, and Ihe syslem of laking away Ihe improved seeds for cullivalion So Ihe sludy of Ihe popularity of M-3 and M-9 reveals Ihal Ihe local species of paddy do have some weaknesses, which are as follows: 1. Nol able lo lolerale fertilizer 2. Thin flowerings and smaller harvesls 3. Less produclion and grains nol full and solid 4. Nol able lo lolerale Ihe cold 5. More lime needed in cooking The qualities that we farmers would want in our paddy The paddy musí have long slraws, full and so Id rice grains. II musí be delicious lo lasle, fragranl, able lo lolerale heavy rains, able lo lolerale Ihe cold, able lo lolerale Ihe fertilizer. II will nol f1ower, nol easily fa 11 , give heavy paddy crops, lake less lime lo ripen, and increase in volume in cooking. 420 Chandra Kanta Poudel As these qualities were no! available in the paddy, we continued to cultivate the local variety. But in the year 1996, LI-SIRO not only brought the M-3 and M-9 varielies but also 144 other varieties of paddy. In the land that belonged to Indra Prasad Poudel, the nursery was made and all the differenl varieties of seed were planted in that plot of land. At that time, due to hailstones, the experiment suffered setbacks, but even then, among them, 30 species were saved and selected for our purposes. In this experiment, three groups were actively involved: U-BIRO, the Mothers' Group, and the Progressive Youth Club. In this way, in the year 1998, the 30 species-selected on the basis of discussions among the local organizations-were distributed among the 30 households of the village, so that each household got to cultivate one particular variety. On the other hand, in the year 1996, U-SIRO had sent us 25 species of paddy and we had also cultivated them. Among the 25 species, our farming community had selected one particular variety on the basis of the harves! produclion, the height of the straw, the taste of the rice, the shape of the rice grain, etc. Summary Among the 144 varieties from 1996, only 30 species were selected, and from the 30, three (the process of selection is continuing). In 1999, from among the 25, one variety was selected. This species was cultivated by Maheswor Poudel in 1999 in two separate fields, and in 2000, he is planning lo cultivate il in Ihree fields. Moreover, al the instigation of the local club, we are planning lo cullrvate that single variety at the rale of 5.5 and 9 by buying the seeds ourselves. We do no! know the name of this paddy. Now In the year 2000, under the joint auspices of U-BIRO. the Mothers' Group, and the Youth Club, we are golng to give it a name. The basls of selection made by the Youth Club and Mothers' Group 1. The height of the straws and the productivity of the paddy crop 2. Falling/not falling 3. Less chance of disease and size of the rice grain 4. CapabJe of tolerating the cold and wet conditions and quick ripening period The above-mentioned basis tor selection was made after the field inspection and discussions in the group meelings among our participant farmers and club members. In this. the local organization relays the information and also tea ches us how to do Ihe work. In the end, we review the whole matter and wilh the participation of the entire farming community, we selec! the paddy species. Participatory plant-breeding program We appreciate this program highly, for it respects our experiences and the traditional technology that we have been following in farming. When looking into the statistics available. we found that without the participation of the farming communily at the nationallevel. there had been recommendations made for more than 42 species of crops, allhough this sort of selection had nol much affecled Ihe people living in Ihe hlgh hilly areas. Therefore, il is most necessary that we have a participalory planl-breeding program among uso For example, we can take Ihe case of paddy species M-3 and M-9 thal we have been cullivating in our own village. Necessity of el plant-breeding program 1. A partlclpalory program means the collective presence of the farming communily: they can selee! tor themselves the paddy species that suit their soll and cllmatic conditions. 421 Role o[Farmers in ¡he Improvemenl o[erops 2. The farmers themselves are more aware oftheir own needs and requlremants. 3, In the seleetlon of the paddy species, the farmers themselves are participants. 4. The farmers leam the technology about how to breed between two speeies to ereate several varieties, Reasons 1. Climatie conditions differ according to altitude and have different affects on farming. 2. Land and climatle conditions differ in the hilly region. 3. When improved species are selected according lo the suitability ofthe particular place and climate, they have a high degree of tolerance and survivability. The drawbacks of the participatory breeding program The disappearanee of the local species. For example, after the introduction of M-3 and M-9, Ihe local species called Kalo patla and Reksali have gradually disappeared. Important suggestions In the process of developing a plant-breeding program, we must remember to include the local species so that the genes of Ihe local wilt not disappear completely. Role of the local organizations Important and active organizations in our village are the Progressive Youth Club and Creative Mothers' Group-two local organization Ihat haya worked closely with the Participatory PlantBreeding Program since 1996. The most important work has besn to develop M-3 and M-9 improved seeds and so we have sent to LI-BIRD and other agricultural organizations about 5 muri of improved seeds in 1998. The local organizations continue to develop the improved seeds and distribute them for cultivation. For example, in areas like Lumle, Paudur, and Salyan VDC, Ihe improved varieties like M-3 and M-9 have already been sent. Plan o, the local organizations • to give continuity to the work carried out by the Participatory Plant-Breeding Program • to develop and distribute the improved Machhapuchhare-3 and -9 paddy species • to conserve, develop, and distribute the selected species • to distribute to Ihe farmers the newly developed improved seeds sent by olher research centers • to increase the village's agricultural production under the leadership of local organizations • to develop systematic and sustainable methods of paddy conservation from the nursery to storage • to inerease our own leaming skilts and technology among ourselves • to develop new improved paddy species 422 Leaders among the farming community Priority is given lo persons who have been successful in order to encourage and bring maximum participa tia n of farmers. Or Ihe farmers have themselves selected one among them or have won the confidence of Iheir farming community. Among the 25 species, one variety was selected and it is cultivated in the field of Maheswor Poudel. This species of paddy was cultivaled al his inltiation. He has said that he took this iniliative because this species has all the good qualities to be found in paddy. As he says, this variety has good tasle lo ea!, long slraws, salid grains of rice kernel and Ihe grains are nol Ilkely lo fall off easily from the plants. Therefore, Ihis species of paddy without name as ye!, has the grea! possibility of becoming popular in this village. The official and scientific research Bahadur Poudel. trom LI-BIRO was carried out under !he leadership of Indra 423 The Importance of Crop Improvement in Conservation of Diversity K. C. Adhikary, S.P. Adhikary, and K.R. Adhikary Development and Environment Club Bagnas, Kaski Abstract In lhis paper, fanners describe lhe agroecology of Bagnas and the diversíty of crops grown lhere. Brief description of the Bagnas Village Bagnas is a small hilly village to the northeast and 16 kilometers away from the well-known tourist center Pokhara. Facing the north-south direction, the village lies about 600 lo 1400 meters aboye sea level. II lies in lhe mid-hill region, and like Ihe rest of the other villages in Nepal, it is equally remarkable for agricultural diversity. A variety of agricultural production and one single crop with a diversity of species is the greatest wealth of this village. Sorne time ago, farming was carried out in Ihis village on the basis of absolutely personal Interest and experience. But since the lasl 1010 15 years, we farmers have been attracted to collective and partlcipatory farming systems. Caste-wise, the village has a majorlty of Brahmins, and 97% of the population is dependent upon agricultura. The land is su eh that in the north are Ihe steep hilly areas, and there are plenty of gullies and hil!ocks wllh narrow strips of land between them. So comparatively speaking, the upper areas of Ihe village are cold and dry/parched, and the lower bell is molst and hot. On Ihe basis of our experience and experiments, in the north belt we grow millet, maize, and Ghaiya (a type of paddy), and in the lower part of the land, we cultivate paddy. There is no irrigation system, so we are entirely dependent upon the monsoon rain. Although during the monsoon months of Ashadhl Srawan (June-September), we collee! water in small man-made reservoirs and pools forfarming. Slnce the main erop in this village is paddy, we are fully involved in cultivating paddy. We are nol engaged commercially in agriculture but in terms of sustaining our lives. The following is a proposal aboul the Participalory Plant-Improvement Program and our experiences and suggestions about paddy cultivation in our Begnas Village. Common species of crops We, the villagers of Begnas, have been eultivating crops aecording lo our long traditional methods of trial and experienee so tha! we have been able lo study wha! sort of erops are suitable in one place and what crop at which time. Having observed !hese details, we have been eullivating our fields. The common crops tha! we have been cultivating are as follows: . • Pakho (hillside, slanted, uncultivable slope): Millet, maize, Ghaiya (paddy), Pidalu, sugarcane, buckwheat, and vegetables (potatoes, radishes, gourds) • Field (for paddy plantatíon): Paddy (and also minimum cultívation of whea! and mustard) Status of paddy species system The proeedures for cultivating the paddy specíes have been based on the traditional technology followed by our forefathers. Actually, the majority of the farmers in the village follow the same methods offarming. Even now, the local species ofpaddy have a grealer influenee in the choice of - _.... _----------~ ------------------425 The Importance o( Crop Improvemenl in Conserva/íon of Díversíty cullivation, although in some places, improved specles of paddy are also being used by some farmers. The common varletíes of paddy belng cultivated are Mansuli, Taichung, and'Radha-7, while some years back, we (one-third of the farmers) culllvaled the varielies called Mansule, Madhise, Chhote, and Radha-7 in order lo make a beller profit. Bul Ihese improved varieties needed chemical fertilizers, good irrigalion facilities, and pesticides to control pesls and diseases. When Ihese requiremenls were not fulfilled, we losl half the harvest in some years. So !hen we re-verted lo cullivaling our own local varieties, and among Ihe most popular ones are almosl 35 lo 40 species. That is why we cullivate Ihe dlfferenl varietíes of the local species. We do no! only work on an individuallevel bul also collecl ourselves inlo differenl groups and work towards preserving and prolecting Ihe local species. In our village, on the basis of our yearly experiences with Ihese different varietíes ot local paddy, we continue to grow Ihem and alternate the seeds in Irying lo create continuity lo our work and efforts al conserving Ihe local paddy species. Diversity In the local paddy species In order lo fulfill different usage requirements and needs, we have been preserving differenl local species. Al presenl in Begnas Village, we have 16 collective development groups. and among !he participaling farmers, Ihere are from nine lo 43 species of paddy !hal are being locally cultivated. (Iable 1l. Now we are planting the following local paddy species (Iable 2). These local species have many qualities Ihat are suilable tor our soil and climatic condilions, so we are giving continuity to our local species. Yel there are some qualilies Ihat we think il would be greal lo improve upon, so we have asked tor assislance from the INSITU Conservation Programo With Ihe help of this program, Ihe local species like Aanga, Thulo gurdi, Sano gurdi, Biramful, Pahe/e, Ekle, and Mansara are being crossed with improved varieties like Savitri, HimaN, and Khumal-4. In Ihis way, if we could cross local with improved varieties, the local species would no! disappear. In fact, many of !he local varieties like Marsl, Pahele, Thulo Marsi, Seto, and Jadan are in the process of being lost, and some other species like Anga, Rate, Chobo, and Jhinuwa have completely disappeared. Table 1. Collective Development Groups and Participating Farmers, Begnas Village Group names No. of paddy species received in exhibition 1 Daralhar Coilec!ive Developmenl Committee (CDC) 43 2 Bisaunaihar CDC 30 3 4 Majhihar CDC 30 Chaur CDC 21 5 Ko1bari CDC 6 Aduwabari CDC 24 20 7 Poudeilhar CDC 2 8 Archalthar CDC 9 Paurakhe CDC 19 18 18 15 15 14 12 9 10 Adhikarithar CDC 11 12 13 14 Sundaridada CDC 15 426 TalbesiCDC Kalirnati CDC Kholbesi CDC Sirnalpata CDC K. e Adhikary, SP. Adhikary, and K.R. Adhikary Species of paddy under cultivation Groups. Names of farmers Archallhar Tara Tiwari 16 Adhikarithar Goma Adhikari 13 Poudellhar Padam Raj Poudel 11 Bisaunathar Padma Kumari Adhikari 11 Table 2. Common Paddy Specles Found In Begnas Village Reásons for being Main species .elected by farmers Drawbacks Interest of the farmers Eil