IntroductionDemand for rice in West and Central Africa (WCA) hasbeen growing at the rate of 6% per annum since 1973,and now amounts to over 8 million tonnes per annum.Increased consumption is due both to population growth(average of 2.9% per annum) and to the increased shareof rice in the diet of the African population. Over thesame period, production of rice has been expanding atthe rate of 5.1% per annum, with 70% of the growthdue to an increased area of cultivated rice, but only 30%due to higher yields. Much of the expansion has been inthe rainfed systems, particularly the two major eco-systems that make up 78% of rice land in WCA: namely,the upland and rainfed lowland systems.

To cover the increasing gap between rice demand andsupply, rice is being imported at an increasing rate.Shipments of rice to Africa rose by 11% to slightly morethan 5 million tonnes in 1999/2000, Nigeria being themajor importer at 0.9 million t, closely followed by Côted’Ivoire. The cost of this importation is approxi-matelyUS$1 billion of foreign exchange each year. Withpopulation growth rate exceeding that of regional food

Development of New Rice for Africa (NERICA) andparticipatory varietal selection

HE GRIDLEY, MP JONES AND M WOPEREIS-PURA

West Africa Rice Development Association (WARDA), 01 BP 2551, Bouaké, Côte d’Ivoire

One of the major research thrusts of the West African Rice Development Association (WARDA), to meet the ever-increasingdemand for rice in West and Central Africa (WCA), has been the development of higher-tillering varieties. By 1996 WARDAhad developed a range of new interspecific varieties derived from crosses between the Asian rice, Oryza sativa and theAfrican rice, O. glaberrina. These varieties, later termed NERICAs (New Rice for Africa), have shown stable and high yieldsand tolerance to major biophysical production constraints in a range of upland environments. Their rapid disseminationto small-scale rice farmers in WCA has been achieved through participatory varietal selection (PVS), an applied and adaptiveresearch mechanism in which farmers play an active role in varietal selection, development and spread. PVS lasts for 3 yearsduring which farmers select and evaluate varieties on their own farms. PVS was initiated at Côte d’Ivoire in 1996 and by2000 all 17 WARDA member countries had initiated PVS, involving 4000 farmers at 105 sites. Particular progress has beenmade in Guinea where from 1997 to 2000 the number of farmers participating in PVS rose from 116 to 20,000 and the areasown to NERICAs from 50 to 8,000 hectares.

Abstract

production, and limited foreign exchange for increasedlevels of imports, the future for Africa’s poor appearsgrim. For food security, WCA needs to increase pro-duction capacity for when cheap rice can no longer beimported.

The birth of a new rice for AfricaThe Asian rice Oryza sativa was brought to Africa over500 years ago. Within a short period, because of its highyield potential, it replaced most of the indigenous Africanrice, Oryza glaberrima, which was pushed to inhospitable,scraggy and infertile upland soils or flood-prone inlandvalleys and deltas. In Africa, however, O. sativa sufferedattack by pests and diseases known only on the Africancontinent. Toxic soils with limited nutrients also reducedits yield potential. The hardy African rice, O. glaberrima,with a history of 3,500 years of cultivation, withstandsthese problems better than O. sativa. To combine the besttraits of the Asian and African rice species, the challengeis to transfer into the high-yielding but stress-proneO. sativa the desirable genes in O. glaberrima whileshedding its undesirable traits of lodging, fewer grainsand grain shattering.

24 Development of New Rice for Africa (NERICA)

Conventional breeding efforts to develop interspecifichybrids had failed due to a high level of sterility in thehybrids. In 1991, WARDA launched a new effort tocombine the genetic potential of the two rice species,using conventional and anther culture techniques toovercome sterility and to hasten the breeding process.Crosses were made and embryo rescue was used toremove fertilized embryos and grow them in artificialmedia. Anther culture allowed rapid fixation and helpedto retain interspecific lines combining desirable featuresof the two rice species. The generous support of donorsin Japan and the Rockefeller Foundation in the USA tothe Joint Interspecific Hybridization Project allowedconsiderable progress to be made, and by the late1990sinterspecific lines of a radically different plant type,tailored for dryland rice farmers of WCA, had beendeveloped and were being tested and evaluated in a rangeof environments.

Agronomic characteristics of interspecificlinesSince the mid-1990s many interspecific lines have beengenerated, evaluated and characterised for a range ofagronomic traits and reaction to important diseases andpests (WARDA 1999, 2000a, 2000b, and 2001a, and2000b). In 2000, these interspecific lines were dubbedNERICA (New Rice for Africa).

NERICAs have been identified with one or more ofthe following characteristics:

● Wide and droopy leaves that help to smother weedsin early growth stages

● Strong stems that can support heavy heads of grain● More tillers with longer grain-bearing panicles than

either parent and non-shattering grains● Stems with secondary branches on their panicles that

can carry up to 400 grains● Early maturity, 30–50 days earlier than currently

grown cultivars● A good height that allows easy harvest of panicles● Good tolerance to drought● Resistance or tolerance to Africa’s most serious pests

and diseases – African rice gall midge, rice yellowmottle virus and blast (Magnaporthe grisea)

● Tolerance to acidic soils but responsive to limitedorganic and inorganic fertilizers.NERICAs have shown stable yields under both low-

and high-input conditions and are expected to reducerisk and increase productivity in farmers’ fields, therebyreducing the need to clear new land. Reduced risk willalso give farmers incentives to use more inputs, intensifyland use, and gradually abandon shifting cultivation, thusimproving system sustainability. The rapid introductionof NERICAs to small-scale farmers in WAC was, there-fore, considered a vitally important first step towards thesustainable intensification of Africa’s fragile uplands. TheNERICAs were not intended as a total replacement forlocal varieties but for integration into the existing varietalportfolio of farmers.

Figure 1. Diagrammatic representation of relative time scales for conventional and PVS to deliver new varieties to farmers

Development of New Rice for Africa (NERICA) 25

Participatory varietal selection (PVS) metho-dologyIn conventional breeding schemes selection and testingprocedures involve a series of multilocational trials over 8to 12 years of a diminishing number of varieties (Figure 1).This favours the selection of a few, widely adapted linesfor release, often with little regard to farmer andconsumer preferences, or to addressing the myriad farm-level microenvironments.

At WARDA, this conventional approach to transferof new varieties has given way to an applied and adaptiveresearch mechanism, termed PVS, that favours farmersplaying an active role in varietal selection, developmentand spread (Figure 1).

The goal of PVS is to efficiently transfer improvedrice varieties to farmers in order to:● Reduce the time required to move varieties onto

farmers’ fields● Determine the varieties that farmers want to grow● Learn the traits that farmers value in varieties to assist

breeding and selection● Determine if there are gender differences in varietal

selection criteriaWARDA’s approach to PVS research (PVS-R) is a

3-year programme (Figure 1). In the first year, breedersidentify centralised fields near villages and plant a ‘ricegarden’ trial of up to 60 upland varieties. The varietiesrange from traditional and popular O. sativas toNERICAs, African O. glaberrimas and local checks. Menand women farmers are invited to visit the plotinformally as often as possible, but the farmers arebrought as groups for formal evaluation of the varietiesat three key stages: maximum tillering, maturity andpost-harvest. For the first two, farmers compareagronomic traits, including weed competitiveness,growth rate, height, panicle type and growth cycle, whilstthe third focuses on grain quality attributes such as size,shape, shattering, ease of threshing and husking andpalatability. Each farmer’s varietal selection and thecriteria for selection are recorded and later analyzed.

In the second year, each farmer receives as many assix of the varieties he or she has selected in the first yearto grow on his or her farm. Thus genetic diversity entersthe communities. PVS observers, who may compromisebreeders and/or technicians from NGOs and extension

Table 1. Evolution of PVS-R activities by year in West Africa, 1996–2000

NumberYear Countries with PVS-R activities Countries Sites Farmers

1996 Côte d’Ivoire 1 1 551997 Benin, Ghana, Guinea, Togo 5 17 5701998 Burkina Faso, The Gambia, Guinea-Bissau, Nigeria, Sierra Leone 7 38 12931999 Cameroon, Chad, Liberia, Mali, Mauritania, Niger, Senegal 17 64 24912000 17 WARDA member countries 17 105 >4000

services, visit participating farmers’ fields to recordperformance and farmer appreciation of the selectedvarieties. At the end of the year, farmers evaluatethreshability and palatability to provide an overall viewof the strengths and weaknesses of the selected varieties.For the third year, farmers are asked to pay for seeds ofthe varieties they select, thereby providing evidence ofthe value they place on them. Thus, in 3 years, PVS-Rallows the farmers to select varieties with specificadaptation and preferred plant type and grain qualitycharacters. These, in turn, can be integrated into thebreeding programmes to tailor varieties for farmers. Inthe conventional scheme at least 12 years are necessaryto reach this point and, even then, farmers andconsumers may not appreciate the varieties selected.

A PVS extension (PVS-E) phase has recently beenintroduced to compliment PVS-R and acceleratedissemination and official release. Four to six of the morecommonly selected varieties in the second year of PVS-Rin an ecoregion are disseminated widely to farmerswithin the region for evaluation in the third year. After2 years of PVS-E, the more-preferred of these varietiesare enrolled in multilocational trials to generate data forofficial release. Simultaneously, these varieties entercommunity-based seed systems (CBSS) for multiplica-tion to ensure adequate seed supplies for rapid dis-semination of the varieties officially approved for release.

PVS activities in West AfricaPVS-R was initiated in 1996 through a small project inBoundiali in Côte d’Ivoire, where farmers were providedwith a rice garden of 57 varieties amongst which theywere able to select those that met their own needs. Theresults were so encouraging that WARDA decided toextend the participatory approach. By 1999 all 17WARDA member countries had instigated PVS-R at 64sites with 2,491 farmers participating; by 2000 over 4,000farmers at 105 sites were involved (Table 1) (WARDA1999).

As early as 1997, significant gender differences invarietal selection were detected at Danane in Côted’Ivoire, but further analysis of the large body of datacollected from other countries is needed to determine ifthis is common. Marked specificity in varietal selectionbetween countries, however, was evident (Table 2),

26 Development of New Rice for Africa (NERICA)

S Bouaké 189 ✔ 1S BW 293-2 ✔ 1S BW 348-1 ✔ 1S Chinese ✔ 1S CICA 8 IR 2042-178-1 ✔ 1S FKR 21 ✔ 1S FKR 243 ✔ 1S FKR 33 ✔ 1S FKR 41 ✔ 1S IDSA 10 ✔ 1S IDSA 85 ✔ 1S IR 12979-24-1 ✔ 1S IR 28228-45-3-3-2042-178-1 ✔ 1S IR 46 ✔ 1S IR 47701-6-3-1 ✔ 1S IRAT 144 ✔ 1S IRAT 216 ✔ 1S ITA 150 ✔ 1S ITA 222 ✔ 1S ITA 398 ✔ 1S Niono 2 ✔ 1S RP 1045-25-2-1 ✔ 1S RP 17846-111 ✔ 1S SK 51-5-2 ✔ 1S Suphanbouri ✔ 1S TGR 75 ✔ 1S TOX 3093-35-2-3-3-1 ✔ 1S TOX 3098 ✔ 1S TOX 3100 ✔ 1S WAB 181-18 ✔ ✔ ✔ 3S WAB 365-B-1-H1-HB ✔ 1S WAB 50-HB ✔ 1S WAB 515-B-16A2.2 ✔ ✔ 2S WAB 515-B-16H2-2 ✔ 1S WAB 56-104 ✔ 1S WAB 56-50 ✔ ✔ ✔ 3S WAB 570-10-B-1A2.6 ✔ ✔ 2I WAB 450-11-1-1-P31-HB ✔ ✔ ✔ 3I WAB 450-11-11-P50-HB ✔ 1I WAB 450-11-1-2-P41-HB ✔ 1I WAB 450-24-2-3-P33-HB ✔ 1I WAB 450-24-2-3-P38-HB ✔ 1I WAB 450-4-1-A16 ✔ 1I WAB 450-I-B-P-105-HB ✔ ✔ ✔ 3I WAB 450-I-B-P-135-HB ✔ ✔ 2I WAB 450-I-B-P-147-HB ✔ 1I WAB 450-I-B-P-153-HB ✔ 1I WAB 450-I-B-P-157-1-1 ✔ 1I WAB 450-I-B-P-157-2-1 ✔ 1I WAB 450-I-B-P-160-HB ✔ ✔ 2I WAB 450-I-B-P-163-4-1 ✔ ✔ ✔ ✔ 4I WAB 450-I-B-P-163-HB ✔ 1I WAB 450-I-B-P-20-HB ✔ 1I WAB 450-I-B-P-33-HB ✔ 1I WAB 450-I-B-P-38-HB ✔ ✔ 2I WAB 450-I-B-P-82-HB ✔ 1I WAB 450-I-B-P-91-HB ✔ ✔ 2I WAB 450-I-B-P-28-HB ✔ ✔ ✔ 3

Total 6 4 6 6 3 4 7 5 3 5 6 5 7 5 5 77

1. S = O. sativa; I = interspecific (O. sativa x O. glaberrima)

Table 2. The 58 varieties in advanced PVS testing by West African countries, 2000

Ben

in

Bu

rkin

a Fa

so

Cam

eroo

n

Côt

e d’

Ivoi

re

Gh

ana

Gu

imea

Bis

sau

Gu

inea

Mal

i

Mau

rita

nia

Nig

er

Sen

egal

Sier

ra L

eon

e

Cha

d

The

Gam

bia

Togo

Tota

l

Type1 Variety

Development of New Rice for Africa (NERICA) 27

reflecting a combination of differing varietal adaptationto the wide range of on-farm (micro-) environments anddiverging consumer preference. Of the 58 varieties inadvanced PVS-R testing in 15 countries in 2000, 46 wereselected in only one country. These accounted for 79%of the varieties, with the rest accounting for 1% or less— namely, six in two countries, five in three countriesand one in four countries (Table 2).

The most frequently selected varieties were derivedfrom WARDA crosses designated by the code ‘WAB’, andamongst these the NERICAs were as popular as theO. sativa varieties. The NERICAs were particularlypopular in Côte d’Ivoire, accounting for all varietiesselected there, and in Guinea, where NERICAS were allbut one of the varieties selected. Five NERICAs have nowbeen released in Côte d’Ivoire and two in Guinea. Onaverage, five varieties were selected in each country,ranging from a minimum of three for Mauritania to amaximum of seven for Guinea and Chad (Table 2).

Selection criteria applied by farmers amongst 17countries were more consistent (Table 3). The mostfrequently applied criteria across countries were:● Yield (14)● Height (13)● Short growth cycle (11)● High tillering (11)

Togo ✔ ✔ ✔ ✔The Gambia ✔ ✔ ✔ ✔ ✔Chad ✔ ✔ ✔ ✔ ✔ ✔Sierra Leone ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔Senegal ✔ ✔ ✔ ✔ ✔Niger ✔ ✔ ✔ ✔Nigeria ✔ ✔ ✔ ✔Mauritania ✔ ✔ ✔ ✔ ✔ ✔Mali ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔Liberia ✔ ✔ ✔Guinea Bissau ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔Guinea ✔ ✔ ✔ ✔ ✔ ✔ ✔Ghana ✔ ✔ ✔ ✔ ✔Côte d’Ivoire ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔Cameroon ✔ ✔ ✔ ✔ ✔Burkina Faso ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔Benin ✔ ✔ ✔ ✔

Total 14 13 11 11 9 9 4 4 4 4 4 3 3 3 3 2 2 1 1 1 1

Table 3. Farmers’ selection criteria applied in PVS-R in 17 West African countries, 1999

Yie

ld

Hei

ght

Shor

t gro

wth

cyc

le

Hig

h ti

ller

ing

Gra

in s

ize

Gra

in-l

arge

Wee

d c

omp

etit

iven

ess

Gra

in c

olou

r

Gra

in-b

old

Goo

d r

esp

onse

to fe

rtil

izer

Lod

gin

g re

sist

ance

Pan

icle

siz

e

Tast

e

Non

-sti

cky

grai

n

Dro

ugh

t tol

eran

ce

Med

ium

gro

wth

cyc

le

Bri

d d

amag

e re

sist

ance

Ad

apta

bil

ity

Em

erge

nce

rat

e

Aro

ma

Dis

ease

res

ista

nce

Selection criteria

● Grain size (9)● Large grain (9).

Thus across countries, high-yielding, early varietieswith profuse tillering and large grain are the most desired,but breeders must be alert to variation between countries,especially taking account of the less-frequently notedcharacters.

PVS in GuineaPVS-R and NERICAs have made a major impact inGuinea. The first phase of PVS-R began in 1997 withsupport from the World Bank that enabled The Institutde recherche agronomique de Guinea (IRAG) andService National de la Promotion Rurale et de la Vulgari-sation (SNPRV) to collaborate with WARDA to identifyvarieties that could be rapidly transferred to farmers. Asecond phase was started in 1999.

The number of farmers participating rose from 116in 1997 to 20,000 in 2000, with the area of NERICAssown over this period increasing from 50 to 8,000hectares (Table 4) (WARDA 2001). In 2000, the gain inproduction and in income from NERICAs was estimatedat 15,000 tonnes of rice worth US$2.5 million withadditional income estimated to rise to US$300,000 in2002 (Table 4).

Country

28 Development of New Rice for Africa (NERICA)

ReferencesWARDA (West Africa Rice Development Association) (1999) Rice

Interspecific Hybridization Project: Research Highlights 1999.WARDA, Bouaké, Côte d’Ivoire. 34 pp

WARDA (West Africa Rice Development Association) (2000a) RiceInterspecific Hybridization Project: Research Highlights 2000.WARDA, Bouaké, Côte d’Ivoire. 34 pp

WARDA (West Africa Rice Development Association) (2000b)Annual Report 1999. WARDA, Bouaké, Côte d’Ivoire. 74 pp

WARDA (West Africa Rice Development Association) (2001a) Bintuand her new rice for Africa. WARDA (West Africa RiceDevelopment Association), Bouaké, Côte d’Ivoire. 32 pp

WARDA (West Africa Rice Development Association) (2001b)Annual Report 2000. WARDA, Bouaké, Côte d’Ivoire. 84 pp

Table 4. PVS in Guinea: Number of farmersparticipating, area sown to NERICAs, and estimatedgain in production and income from NERICAs,1997–2002

Area (ha) Estimated gainNumber of sown to in production

Year farmers NERICAs and income

1997 116 50

1998 380 150

1999 1,680 250

2000 20,000 8,000 15,000 tonnesUS$2.5 million

2002 300,000 tonnes

IntroductionThe study was initiated in 1998 and is part of the activitiesof the International Plant Genetic Resources Institutes(IPGRI) global project, Strengthening the Scientific Basisof in situ Conservation of Agricultural Biodiversity inNepal. It was jointly implemented by the Nepal Agri-cultural Research Council (NARC), a non-governmentalorganisation (NGO) Local Initiatives for BiodiversityResearch and Development (LI-BIRD) and localinstitutions including community-based organisations(CBOs), a network of nodal farmers, and farmingcommunities.

Enhancing on-farm conservation of traditional rice varietiesin situ through participatory plant breeding at threecontrasting sites in Nepal

BR STHAPIT1, J BAJRACHARYA2, A SUBEDI3, KD JOSHI4, R RANA3, SP KHATIWADA2, S GYAWALI3,P CHAUDHARY3, PR TIWARI4, DP RIJAL4, K SHRESTHA3, BK BANIYA2, A MUDWARI2, MP UPADHAYA2,D GAUCHAN3 AND D JARVIS1

1. International Plant Genetic Resources Institute (IPGRI), 10 Dharmasahila Buddha Vihar Marg, Nadipur Patan, Ward 3, Kaski,Pokhara 3, Nepal

2. Nepal Agricultural Research Council (NARC), Singhadurbar, Plaza, Kathmandu, Nepal3. Local Initiatives for Biodiversity Research and Development (LI-BIRD), PO Box 324 Bastolathar, Mahendrapul, Pokhara, Nepal4. Department for International Development (DFID) Plant Sciences Research Programme (PSP), Centro Internacional de

Mejoramiento de Maiz y Trigo (CIMMYT), South Asia Office, Singh Dubar Plaza Marg, Bhadrakali, Kathmandu, Nepal

Understanding rice diversityWe found that communities maintained a rich varietaldiversity ranging from 21 cultivars (100% of which werelandraces) at the high-mountain Jumla site to 69 cultivars(91% local, farmer-named cultivars) at the middle-hillKaski site. Farmers at the highly accessible Bara village,where extension services had made many interventions,and who are close to the national rice research centreand the Indian border maintained 33 cultivars (62% ofthem landraces). More landraces were maintained atboth the household and community level rather thanmodern cultivars at all three sites. The area covered by

Participatory plant breeding (PPB) is one of the strategies employed for on-farm management of traditional rice varietiesin three contrasting in situ conservation villages of Nepal. PPB programmes in Nepal were designed to investigate whether:1. Traditional farmers’ cultivars can be conserved per se2. PPB contributes to the enhancement of biodiversity by broadening the genetic base that provides benefits to community3. The PPB process encourages farmers to maintain the evolutionary processes that allow the crop to evolve over time,

while remaining adapted to diverse, local production niches4. Landraces can be conserved by improving them for traits that farmers consider important to make them more

competitive against alternative varieties.This paper describes the processes of PPB that encourages farmers to locate rice diversity and their custodians; understand

the amount of rice diversity (the number of local and modern varieties at both the household and community level) and itsdistribution (the frequency of households growing each variety); analyse the preferred traits of rare, endangered and locallycommon farmers’ varieties; and develop options for adding benefits to cultivating local landraces. It describes preliminaryresults on understanding genetic diversity in terms of its value to local communities and the participatory methods used toselect landrace parents for PPB. Participatory methods, such as biodiversity fairs, diversity blocks, diversity kits and communitybiodiversity registers, were integrated to sensitise the local community to varietal diversity issues, to understand the valueof local crop diversity, and to strengthen the roles of farmers and the informal sector in the local crop development process.

Abstract

30 Enhancing on-farm conservation in Nepal

rice landraces in Bara was 17%, followed by 73% in Kaskiand 100% in Jumla.

Landrace enhancementWe found that resource-poor farmers, compared toresource-endowed households, were more dependent onlandraces for their food security. This was particularlytrue in marginal rather than in better-off environments(Rana et al., 2002). We also found that some landraceswere very competitive to modern cultivars in certainniches and that such landraces could be promoted morewidely for similar agroecological niches without furtherimprovement. Results also indicated that a few culturallyimportant cultivars, such as Anadi, could be conservedthrough market links and consumer awareness.

Genetic diversity of landracesWe also assessed the genetic diversity of a few locallycommon and well-known landrace populations oflandrace Jetho Budho in Kaski district, and landracesKariya Kamod and Lalka Basmati in Bara district. Thelandraces from Jumla had a low level of diversity (Jaccardsimilarity index of 0.87 and 12.8% of polymorphic loci)(Bajracharya et al., 2001).

This high degree of similarity indicates a possibilitythat these landraces, that have unique traits for toleranceto cold water chilling injury but are susceptible to neckblast, could have originally come from the same source.Compared to the low diversity among Jumla landraces,landraces from Kaski and Bara had a considerable levelof genetic diversity (Jaccard similarity index = 0.48–0.50), reflecting the high ethno-botanical variabilityindicated by many different farmer-given names anddiverse seed-related phenotypes (Bajracharya et al.,2001). Although the intra-population genetic variation

of Jetho Budho and Basmati was generally low, variationfor quality traits has been found and is a potentialopportunity for landrace enhancement. Throughconsumer surveys we have established the quality traitsof Jetho Budho. These include (in order): softness ofcooked rice, aroma, elongation ability, and taste. TheProject plans to help farmers market locally produced,quality-assured rice seed and grain using local brandnames as a way of improving rural livelihoods andconserving genetic diversity.

Social seed networkWe found that 96–100% of farming households aredependent upon an informal seed source (Baniya et al.,2002). The seed flow occurs basically through farmers’social networks. The community managed a rich ricediversity through this social system, that involved:exchange on a barter basis (60–63%), gift (20–25%),borrowing of either seed or seedlings (10–12%), andpurchase (<5%) (Subedi et al., 2002). Nodal farmers whosearch, select and maintain rice diversity have been foundto play an important role in seed flow in the informalseed system and were selected as collaborators in the PPBprogramme.

Performance of PPB bulksThe 13 segregating materials from crosses between locallandraces, and local landraces with exotic varieties, weredistributed to interested and nodal farmers for growingan F

2 or F

3 bulk of their choice, and a farm walk was

carried out by researchers and farmers to assess the fieldperformance of different populations.

In Begnas village F3 segregants of Mansara (locally

adapted to marginal, drought-prone rainfed and low-input conditions) and Khumal 4 (a good-quality moderncultivar having the local variety Pokhereli Masino as oneof its parents) showed promising results in upper Begnasareas (1000–1300 m) (Figure 1). The breeding goal is toincorporate the good ‘eating’ quality and yield potentialof Khumal 4 into Mansara without losing the adaptivetraits of Mansara. The cross between Pusa Basmati 1 andJetho Budho (a local high-quality rice in high marketdemand) was also doing well in Kholakochewenvironments (650–690 m). In Bara district, six largesegregating populations were evaluated in farmers’ fields,among which farmers preferred the populations ofthe cross Lajhi x IR62161–22–1–2–1–1 and selectedamong them for lodging resistance and post-harvesttraits. The performance of these segregants was verygood.

Future plansSelected populations of Jetho Budho landraces will beevaluated for quality traits and taste with diverseconsumers — hoteliers, millers, housewives, farmers, andresearchers.

Figure 1. Promising plants in the bulk of the crossKumal 4 x Mansara

Enhancing on-farm conservation in Nepal 31

The PPB bulks of the Mansara x Khumal 4 have beendivided into three bulks based upon grain colour andphenotypes. These bulks have been distributed to allnodal farmers and interested farmers for on-farm testingand selection. Researchers will monitor the spread ofthese PPB bulks in 2003 through farm walks (Figure 2)and household interviews.

Figure 2. Participants in a farm walk, Begnas, 2002

References

Bajracharya J, Steele KA, Witcombe JR, Sthapit BR, Jarvis DI (2001)A study of genetic relationships in rice landraces of Jumla, Kaskiand Bara ecosites using microsatellite DNA markers. Paperpresented at the National Workshop of in situ Conservation ofAgrobiodiversity On-farm, held at Lumle, Pokhara, Nepal, 23–25 April, 2001

Baniya BK, Subedi A, Rana RB, Tiwari RK, Chaudhary P, Shrestha S,Tiwari P, Yadav RB, Gauchan D, Sthapit BR (2002) What are theprocesses used to maintain the genetic diversity on-farm? In:Gauchan D, Sthapit BR (eds.) Nepal’s contribution to agro-biodiversity conservation in-situ: A scientific basis for policyrecommendation. IPGRI/NARC/LI-BIRD

Rana RB, Gauchan D, Rijal DK, Subedi A, Upadhaya MP, Sthapit BR,Jarvis DI (2002) Factors influencing farmers decision onmanagement of local diversity on-farm and their policyimplications. In: Gauchan D, Sthapit BR (eds.) Nepal’s contri-bution to agrobiodiversity conservation in-situ: A scientific basisfor policy recommendation. IPGRI/NARC/LI-BIRD

Subedi A, Chaudhary P, Baniya B, Rana RB, Tiwari RK, Rijal DK,Jarvis D, Sthapit BR (2002) Who maintains genetic diversity andhow? Policy implications for agrobiodiversity management. In:Gauchan D, and Sthapit BR (eds.) Nepal’s contribution to agro-biodiversity conservation in-situ: A scientific basis for policyrecommendation. IPGRI/NARC/LI-BIRD

32 Farmer PPB and PVS in eastern India

Farmer participatory breeding and participatory varietalselection in eastern India: Lessons learned

TR PARIS1, RK SINGH2, GN ATLIN1, S SARKARUNG3, G MCLAREN1, B COURTOIS4, K MCALLISTER5, C PIGGIN6,S PANDEY1, A SINGH7, BN SINGH11, ON SINGH8, S SINGH8, RK SINGH8, NP MANDAL8, K PRASAD7, RK SAHU9,VN SAHU9, ML SHARMA9, RKP SINGH10, R THAKUR10, NK SINGH10, D CHAUDHARY10, S RAM11

1. International Rice Research Institute (IRRI), DAPO Box 7777, Metro Manila, Philippines2. International Rice Research Institute (IRRI), Liaison Office, 1st Floor, CG Block, NASC Complex, Dev Prakash, Shanstri Marg,

Pusa, New Delhi 110 012, India3. Rice Research Institute, Department of Agriculture, Chatuchak, Bangkok 10900, Thailand4. Centre de coopération inernationale en recherche agronomique pour le développement (CIRAD), 34398 Montpeiller, Cedex 5,

France5. 4 Armshore Drive Halifax, Nova Scotia, Canada6. Austrlian Council for International Agricultural Research (ACIAR), GPO Box 1571, Canberra, New South Wales, Australia7. Narendra Deva University of Agriculture and Technology (NDAUT), Kumarganj, Faizabad 224 229 Uttar Pradesh, India8. Central Rainfed Upland Rice Research Station (CRURRS), PO Box 48, Hazaribagh 825 301, Jharkhand, India9. Indira Gandhi Agricultural University F/2 Krishak Nagar, Rapir, 492 012, Madhya Pradesh, India10. Rajendra Agricultural University, Pusa, Samastipur 848 125, Bihar, India11. Central Rice Research Institute (CRRI), Indian Council of Agricultural Research (ICAR), Cuttack, 753 006, Orissa, India

This paper describes the objectives and methods used in a participatory breeding (PPB) project in several rice ecosystemsin eastern India. Farmers were interested in a large range and combinations of traits. The selection criteria of farmers weredetermined by hydrological conditions (water depth) and land types (toposequence) followed by the adaptation of thevariety to different user needs (food, livestock fodder, thatching, cash). Other factors determining preference were: thecompatibility of the variety in the cropping systems, socioeconomic status, gender and culture. Several lessons learnedfrom the three years’ experience led to the refinements in the participatory methods in the next phase of the project.

Abstract

IntroductionPoverty in Asia is most severe in rainfed areas wherepeople depend on rice for subsistence. To these ruralpoor, rice is not only the staple food but also has manylivelihood uses. Sustaining livelihoods and alleviatingpoverty require a major increase in agricultural prod-uctivity. This means, among other factors, developinghigher-yielding well adapted rice varieties.

Classical breeding approaches have been successfulin developing improved varieties of rice for favourablerice environments. However, these approaches have beenless successful in rainfed environments because they failto account for the high levels of social and agro-ecologicaldiversity in these areas. Witcombe et al. (1998) reportweaknesses in the formal testing system in India that havereduced the chances that varieties released for marginalareas would meet farmers’ needs. The failure of the

system is evidenced by, for example, the rejection of manyvarieties by farmers, and the rapid and high adoption byfarmers of such non-released varieties such as Mashuririce that had been rejected in the formal testing.

The goal of this project was to increase food securityby providing varieties capable of producing for high andstable yields. The specific objectives were to:● Test the hypothesis that farmer participation in

rainfed rice breeding can help develop suitablevarieties more efficiently

● Identify the stages along a breeding programme wherefarmer participation is most necessary.

● Enhance the capacities of the national agriculturalresearch systems (NARS) in farmer participatoryresearch and gender analysis for rice plant breedingand varietal selection (Courtois et al., 2000).

Farmer PPB and PVS in eastern India 33

Methodologies and major findingsSelection of the research sites

This project was conducted in eastern India, (easternUttar Pradesh, eastern Madhya Pradesh, Assam, Bihar,West Bengal, and Orissa). Eastern India is the largest rice-growing region in the country and accounts for about63% of India’s rice area (26.8 million ha). However, about80% of the rice farming in this region is rainfed.

The project involves two rice ecosystems and six sites:one in the upland ecosystem and five in the rainfedlowland ecosystem representing different agro-ecologies(drought-prone, submergence-prone, or both). At eachsite, two to three villages were chosen on the basis oftheir access to market, agro-ecological diversity andextent of adoption of improved varieties. A team of socialscientists (mostly agricultural economists) and plantbreeders from each centre conducted this research.

Finding out farmers’ selection criteria based on theiragro-ecological and socioeconomic environment

Farm household surveys and participatory ruralappraisals (PRA) were conducted to characterise farmers’agro-ecological and socioeconomic environments,farming systems, rice diversity, farmers’ crop manage-ment practices, gender roles, and seed managementpractices. Farmers’ were consulted about their preferredtraits and the positive and negative traits of local andimproved varieties. Through graphic illustration of traits,male and female farmers valued the importance ofvarietal traits according to land types (lowland, midlandand upland), by socioeconomic groups and by gender.The criteria for varietal choice were:● Adaptation of the variety to the hydrological

conditions (water depth) and land type (topo-sequence location) of their fields. For the lowlandsand submergence-prone ecosystems, long-duration,photoperiod-sensitive, semi-tall and tall varieties werewell adapted. For the uplands and shallow-depthmedium lands, short to medium duration varietieswere preferred to escape the terminal drought(Courtois et al., 2001; IRRI, 2001)

● Adaptation to different user needs such as food,livestock fodder, thatching and cash. Farmers grewmore than one variety not only because they farmheterogeneous rice fields but also for different end-uses. Different varieties fulfilled different livelihoodfunctions (food, livestock fodder, thatching, and cash).For example, farmers like varieties with long finearomatic grain, because these are used as gifts forspecial occasions (marriage) and for religious cere-monies.

Poor farmers were more interested in the qualityof leftover rice that should remain tender and soft –characteristics found in traditional varieties. Similarly,traditional varieties were perceived to be better forpreparing puffed rice and other rice products (Paris

et al., 2001a). In the uplands, farmers preferred suchtall varieties as Brown Gora, Vandana, RR 151-3 andKalinga III because they need the straw for animalfodder (IRRI, 2001). Other farmers grew traditionalvarieties with a purple-pigmented base in drought-prone areas in Bihar, Madhya Pradesh, and Cuttack.This trait helped farmers distinguish weeds from riceespecially in direct-seeded fields where weeds are amajor problem (Sahu et al., 2001)

● Compatibility with existing cropping systems. Formedium lands farmers preferred medium-heightvarieties in the medium lands of short duration (90–110 days) that allowed the timely sowing of a followingcrop of pulses, wheat, or vegetables. In Bhubaneswar,Orissa, farmers rejected certain rice varieties that haddense root growth because these inhibited theestablishment of a relay crop of pulses (black gram)broadcast into the standing rice crop (IRRI, 2001).

● Socioeconomic status of farmers. Farmers preferreddifferent grain types according to their socioeconomicstatus or degree of market integration. For example,in Faizabad, eastern Uttar Pradesh, farmers and fieldworkers of lower castes with small landholdingspreferred varieties with coarse grains, that give thema feeling of fullness due to their slow digestibility. Thehigher-caste farmers with large landholdings who sellrice to the market preferred fine slender grains, thatcommand a higher price. In general, smallholderfarmers from the lower castes used rice mainly forconsumption, while farmers from the upper caste withmore land sold their surplus (Paris et al., 2001b).Farmers in the uplands of Hazaribagh preferredvarieties that do not require high inputs. Farmers whodepend on family labour preferred varieties with arange of maturity dates so that harvests can bestaggered (IRRI, 2001).

● Gender-specific roles in rice production, postharvest, consumption, and livestock care. At all thesites, there were gender-specific tasks in rice pro-duction. Women from poor farming householdsprovided most of the labour in rice production(pulling seedlings, transplanting, weeding), post-harvest (winnowing, hand threshing, seed drying),and seed management (selection, storage). Men weremainly responsible for land preparation, applicationof chemicals, and transporting inputs and products.Male and female farmers in Faizabad agreed that grainyield and duration were the most important traitswhen choosing varieties for upland and lowland areas.However, women gave more importance to such traitsas competitiveness to weeds and post-harvest qualitiessuch as ease of dehusking, or threshing, and highmilling recovery or suitability for different foodpreparations (e.g., puffed rice). In Raipur, MadhyaPradesh, women consistently identified straw quantityand quality as important, whereas the men never

34 Farmer PPB and PVS in eastern India

mentioned these criteria. Women’s criteria for varietalselection were likely to be related to their roles andresponsibilities (Paris et al., 2001b; Sahu et al., 2001).

Selecting new varieties suitable to farmers’preferences and agro-ecological conditions

Participatory plant breeding (PPB). Farmers andbreeders selected individual plants from segregatingpopulations (F

5) from different crosses, using the

pedigree selection method. Trials were held on-farm andon-station. Plants selected by breeders and farmers wereadvanced separately through several generations untilfixed. Farmers and breeders evaluated these genotypesat maturity on the basis of panicle and grain characters,and susceptibility to stem borers. Breeder-selected andfarmer-selected materials were then compared.

Promising lines were selected and these genotypeswere multiplied and supplied to farmers for evaluation.

Participatory varietal selection (PVS) In PVS, similarsets of fixed varieties (13–25 advanced lines and a localcheck) suited for the specific hydrological conditions inthe area were tested on-station and in farmers’ fields.The advanced lines were from the IRRI Shuttle BreedingProgram and from breeding programmes of projectpartners in eastern India. Two or three farmers per villageconducted the on-farm trials under their own level ofmanagement. At vegetative (pre-flowering) andreproductive stages (maturity) farmers and breedersvisually ranked the rice lines grown in both on-stationand on-farm trials. The Kendall coefficient of concor-dance and the Spearman rank correlation coefficientwere used to analyse the agreement of ranking of thegenotypes among farmers, among breeders and betweenfarmers and breeders (Courtois et al., 2001). The resultswere as follows:● There was strong agreement among farmers’ visual

rankings, but not always among those of plantbreeders

● Agreement of breeders with farmers was fairly goodon plant traits but many quality traits were over-looked by breeders

● Agreement between visual ranking of traits by menand women during the pre-harvest stage was good.However, they differed during post-harvest assess-ment

● Yield and duration were important traits consideredby farmers. However, ranking preferences were notalways correlated, indicating that these traits are notthe only factors taken into consideration by farmerswhen selecting rice varieties

● Sensory evaluation showed that the mode of ricepreparation (parboiled or not parboiled) influencesfarmer selection. Sensory ranking did not correlatewith results of classical physico-chemical analysis(Singh et al., 2001).

Lessons learnedSeveral lessons were learned from the 3 years ofexperience in developing and testing the methodologiesfor farmer participation. These lessons are related to theconcerns:● Choice of representative sites. During the first year,

a few of the sites selected for on-farm trials were notrepresentative of the environments targeted in thebreeding work. They were chosen for conveniencereasons (e.g., close to the research station). Thus newtrial sites, which better represented farmers’ environ-ments, replaced some of the sites.

● Number of villages to represent a specific agro-ecology. Due to limited resources and staff, only twoor three trials and farmers were included in eachvillage. Thus the risks of losing information due tosevere drought, poor management of trials, etc. werehigh. The Mother and Baby trial model may providean alternative way of reliably testing a large numberof cultivars under farmer management (Atlin et al.,2001; Appendix).

● Choice of the varieties to be included in theexperiments. The material chosen for the PVS trialswas not always ideal. The lack of clear-cut differencesbetween some of the varieties included in the PVSmade it more difficult for the farmers to rank them.The number of varieties tested should be balancedbetween what is useful for the breeders (many lines),acceptable to the farmers (fewer lines) and what ispossible on a reasonable plot size. Farmers haddifficulty in visually ranking too many (13–25) ricelines from 1 (best-liked) up to n (least-liked). Farmerswere willing to test a maximum of five varieties ontheir own fields. A rating system, for example, 1–3(bad, average, good) or 1–5 numerical scale, for traitsis a simpler method (Atlin et al., 2001).

● Constraints in post-harvest operations of too manylines. The need to harvest, thresh, and weigh thedifferent varieties in small quantities was toocumbersome. Moreover, this work increased theburden of the women cooperators who did the post-harvest work. The number of lines for testing andselection should be reduced. A local field technicianis required to assist in post-harvest work and to ensurethat the varieties are not mixed.

● Number of varieties to include in sensoryevaluation. Men and women found it difficult toevaluate the cooking and eating quality of too manylines. Orgonoleptic tests should be modified so thatfewer varieties are tested at any one time.

● Access to new seeds. During the first 2 years of theproject, the availability of seeds was quite limited andnot all farmers who wanted to participate in theselection of new rice lines on their own farms couldbe included.

Farmer PPB and PVS in eastern India 35

● Institutional constraints. There was a lack of plantbreeders at centres and NARS with experience onparticipatory approaches. In some centres, it wasdifficult for the breeders to change their practices andincorporate the participatory approach into theirformal breeding programme. There were fears thatfarmer participatory breeding will replace, rather thancomplement, conventional breeding. Moreover, theskills in doing this kind of work, that involves multi-institutional participation, diverse socio-culturalsettings and many stakeholders, were not welldeveloped. In many of the centres, there were nofemale social scientists or female plant breedersincluded in the team. Thus, it was difficult at thebeginning for the male plant breeders to increase thenumber of women cooperators in the PVS on-farmtrials. Thus, there is a need to develop partnershipswith NGOs and extension research institutions.

ConclusionsPVS is essential in unfavourable rainfed environmentsand diverse socioeconomic groups that use rice for theirlivelihoods. Farmer participation improved the selectionof suitable varieties for complex rainfed environmentsin eastern India because: farmers screened new varietieson their own farms under their own levels of manage-ment, and breeders better understood farmers’ qualityrequirements.

The close collaboration between scientists andfarmers helped ensure that farmers had access to usefulgermplasm that was adapted to their circumstances andmet the requirements of the community for quality andlivelihood use. Providing farmers with diverse materialsalso helped to enhance genetic diversity. Continuousfeedback mechanisms between breeders and farmersshould be established to ensure appropriate materialsare disseminated to farmers.

There was little evidence of major differences in theway farmers and breeders visually rated varieties in thefield. Hence, simply having farmers take over the earlygeneration visual selection is unlikely to result insignificant improvements in the acceptability of cultivars.On the other hand, extensive farmer-managed trialsappear to be a promising approach to reducing the effectsof random variability and increasing gains fromselection.

Although it is too early to show the impact of PVSthrough the spread of the materials generated from thatPVS, there are promising signs that farmer-selectedvarieties are spreading easily through farmer-to-farmerexchange. Plant breeders are beginning to change theirmindsets and have realised the necessity of interactingwith both farmers, and socio-economists. However, thechallenge lies in institutionalising farmer participationas integral in the formal breeding programmes and inscaling up PVS and disseminating high-quality new seeds.

ReferencesAtlin G, Paris T, Courtois B (2001) Sources of variation in rainfed

rice PVS trials: implications for the design of “Mother-Baby” Trialnetworks. Paper presented at the workshop Quantitative Analysisof Data from Participatory Methods in Plant Breeding, 23–25August 2001 in Giessen, Germany

Courtois B, Bartolome B, Chaudhary D, McLaren G, Misra CH,Mandal NP, Pandey S, Paris T, Piggin C, Prasad K, Roy AT,Sahu RK, Sahu VN, Sarkarung S, Sharma SK, Singh A, Singh HN,Singh ON, Singh NK, Singh RK, Singh RK, Singh S, Sinha PK,Sisodia BVS, Takhur R (2001) Comparing farmers and breedersrankings in varietal selection for low-input environments: A casestudy of rainfed rice in eastern India. Euphytica 122: 537–550

Courtois B, Singh RK, Pandey S, Paris T, Sarkarung S, Baghel SS,Sahu RK, Sahu VN, Sharma SK, Singh S, Singh HN, Singh A,Singh ON, Sisodia BVS, Misra CH, Roy JK, Chaudary D, PrasadK, Singh RK, Sinha PK, Mandal NP (2000) Breeding better rainfedrice varieties through farmer participation: some early lessonsfrom eastern India. Pages 208–223 in: Lilja N, Ashby J, Sperling L,(eds.), Assessing the impact of participatory research and genderanalysis. Participatory Research and Gender Analysis Program,International Center for Tropical Agriculture, Cali, Columbia

IRRI (2001) Farmers and scientists: building a partnership forimproving rainfed rice in eastern India. Final narrative reportand technical report submitted to the International DevelopmentResearch Centre (IDRC), Canada, September 2001. InternationalRice Research Institute (IRRI), Metro Manila, Philippines

Paris TR, Singh A, Luis JS, Hossain M, Singh HN, Singh SS, SinghON (2001a) Incorporating gender concerns in participatory riceplant breeding and varietal selection: preliminary results fromeastern India. Pages 109–121 in: Lilja N, Ashby J, Sperling L (eds.),Assessing the impact of participatory research and genderanalysis. Participatory Research and Gender Analysis Program,International Center for Tropical Agriculture (CIAT) Cali,Columbia

Paris TR, Singh A, Luis J (2001b) Listening to male and female farmersin rice varietal selection: a case in eastern India. In: CIAT, AnExchange of Experiences from South and Southeast Asia.Proceedings of the International Symposium on ParticipatoryPlant Breeding and Participatory Plant Genetic ResourceEnhancement, Pokhara, Nepal, 1–5 May 2000. CentroInternacional de Agricultura Tropical, Cali, Columbia

Sahu RK, Sahu VN, Sharma ML, Paris TR, McAllister K, Singh RK,and Sarkarung S (2001) Understanding farmers’ selection criteriafor rice varieties: a case in Madhya Pradesh, eastern India: a casestudy in eastern Uttar Pradesh, India. In: CIAT, An Exchange ofExperiences from South and Southeast Asia. Proceedings of theInternational Symposium on Participatory Plant Breeding andParticipatory Plant Genetic Resource Enhancement, Pokhara,Nepal, 1–5 May 2000. Centro Internacional de AgriculturaTropical, Cali, Columbia

Singh RK, Prasad K, Mandal NP, Singh RK, Courtouis B, Singh VP(2001) Sensory evaluation of rice varieties with farmers: anexperience in eastern India. In: CIAT, An Exchange of Experiencesfrom South and Southeast Asia. Proceedings of the InternationalSymposium on Participatory Plant Breeding and ParticipatoryPlant Genetic Resource Enhancement, Pokhara, Nepal, 1–5 May2000. Centro Internacional de Agricultura Tropical, Cali,Columbia

Witcombe JR, Virk DS, and Farrington J (eds.) (1998) Seeds of Choice.Making the Most of New Varieties for Small Farmers. Oxfordand IBH, New Delhi, and Intermediate Technology Publications,London

36 Integrating conventional and participatory crop improvement in rainfed rice

IntroductionIRRI’s participatory crop improvement research isfocussed on increasing the impact of rainfed ricebreeding programmes. Particularly in upland rice, thesebreeding programmes have had only limited success in:● Identifying the traits that are important to farmers● Identifying lines that outperform traditional varieties

in the most stressful environments● Giving farmers access to new varieties.

IRRI is collaborating with national programmes inIndia, Laos, and Indonesia in participatory varietialselection (PVS) research that addresses the problem ofthe limited power of on-station experimentation topredict farmer adoption of new lines, and on reducingthe high costs of introducing effective participatoryresearch into formal breeding programmes. This researchrecognises that random genotype x environment inter-action (G x EI), or variation in ranking among experi-mental lines across trials, even within narrowly definedtarget environments, is large in rainfed rice (Cooper

Integrating conventional and participatory cropimprovement in rainfed rice

GN ATLIN1, TR PARIS1, B LINQUIST2, S PHENGCHANG3, K CHONGYIKANGUTOR3, A SINGH4, VN SINGH4,JL DWIVEDI4, S PANDEY1, P CENAS1, M LAZA1, PK SINHA5, NP MANDAL5, SUWARNO6

1. International Rice Research Institute (IRRI), DAPO Box 7777, Metro Manila, Philippines2. International Rice Research Institute (IRRI), Lao Project, PO Box 600, Luang Prabang , Lao PDR3. Northern Agricultural and Forestry Research Institute (NAFRI), Luang Prabang , Lao PDR4. Narendra Deva University of Agriculture and Technology, (NDUAT), Kumarjang, Faizabad 224 229, Uttar Pradesh, India5. Central Rainfed Upland Rice Research Station (CRURRS), Central Rice Research Institute (CRRI), PO Box 48, Hazaribag, 825 301,

Jharkand, India6. Balai Penelitian Padi (BALITPA), Jalan Raya No. 9, Sukamandi, West Java, Indonesia

et al., 1999). The only reliable way to minimise the effectsof random G x EI is to evaluate new cultivars over manyfarms and several seasons (Atlin et al., 2001).

The experience of IRRI and its national partnersindicates that linking small rainfed rice breedingprogrammes into collaborative breeding and varietytesting networks is an effective way to achieve theenvironmental replication needed to make genetic gainsin stressful environments. This paper summarises theresearch results of IRRI and its collaborators onapproaches to integrating and scaling up PVS withinnational rainfed rice programmes.

Key points for introducing participationExperience has identified three key points whereparticipation is needed in breeding programmes:

Identifying traits valued by farmers throughparticipatory assessment

It is important to identify at the outset quality andproduction traits preferred by farmers, in order to allow

To increase the impact of rainfed rice breeding programmes, participatory varietal selection (PVS) must be institutionalisedin the varietal screening process. PVS can be conveniently integrated in two steps: 1. preliminary PVS, in which a relativelylarge number of lines is screened for adaptation and acceptability on a few farms, with a high level of researcher involvementper farm; 2. advanced PVS, in which a small set of varieties is screened by many farmers, with minimal researcher involvementper farm.

Use of farmer ratings for yield and other agronomic traits, collected via surveys during and after the growing seasoninstead of through crop cuts and researcher measurements, permits even small breeding programmes to scale up advancedPVS, sampling the large numbers of farmers and environments needed to reliably detect differences among varieties. Researchin rainfed rice systems in India, Laos, Indonesia, and the Philippines has shown that farmer pre-harvest yield ratings arehighly correlated with researcher-measured yields and with farmer preference ratings taken at the same stage. Little cultivarx farm interaction is observed in rainfed rice PVS trials, indicating that broadly adapted cultivars can be identified.

Abstract

Integrating conventional and participatory crop improvement in rainfed rice 37

the selection of genotypes with preferred traits early inthe on-station evaluation process. Failure to determineand screen for traits critical to farmer preference meansthat resources are expended on developing and in-creasing lines that have no hope of being adopted. Thisis particularly important in upland systems where farmerquality and plant-type preferences are often not wellunderstood by breeders, but it is also a necessaryprerequisite to the scaling up of PVS. This is becauselarge-scale PVS, in which dozens or even hundreds offarmers are asked to assess the performance of new lines,must rely on quantitative survey instruments involvingfixed questions and numerical rating scales that permitstatistical summary and analysis, rather than on open-ended questionnaires. The questions included inquantitative surveys must therefore be designed tocapture information on the traits most valued by farmers.

We have found that several participatory techniquescan be used to elicit farmer priorities and preferences.Among the most effective of these is preference analysisundertaken with individuals or small farmer groups inthe field. In these exercises, farmers view a variety trial(usually just before maturity), and discuss the strengthsand weaknesses of particular varieties by explaining whythey like, or dislike, a particular variety.

Preliminary farmer-managed screening of fixed lines:PVS with many genotypes

A second point at which farmer participation is neededis the preliminary evaluation of fixed lines that have beendeveloped and evaluated agronomically on-station.Typically, rainfed rice breeding programmes in Asiapromote about 20 such lines annually from a preliminaryon-station yield trial to advanced yield testing in multi-environment trials (MET) or in replicated on-stationtrials. Information on farmer acceptability andperformance under farmer management is needed onthis relatively large set of lines. Without it, there is a highprobability that the lines will fail in wide-scale PVS.

PVS strategies permitting the evaluation of manygenotypes on few farms are useful for this purpose. Wehave found that the Mother and Baby PVS trial designcan permit a set of up to about 20 varieties to be evaluatedunder farmer management. An incomplete-block designis used, with three to eight varieties tested per farm, andeach variety tested on three to five farms. This designrequires considerable researcher involvement, but isuseful in getting a preliminary farmer assessment of alarge number of varieties, allowing varieties that performwell on-station but fail under farmer management to bequickly eliminated.

The importance of this step was confirmed in a PVSprogramme conducted by the Indonesian upland riceprogramme in the wet season of 2002, in which 11varieties were evaluated on six farms (Suwarno et al.,2002). Although all the lines performed well on-station,

several failed on-farm due to blast or brown-spot diseasethat had not appeared with severity in screening trialson-station, probably because of differences in fertilitymanagement. Differences in plant density and weedpressure between the station and farmers’ fields may alsohave contributed to the poor performance of some lineson-farm.

It is important for the credibility and success ofadvanced PVS trials that lines with a high probability ofnon-acceptance by farmers be excluded. This PVS stepis at least as important as replicated on-station or METtesting. Most formal breeding programmes can directlymanage such trials on a small number of farms, but mayneed to shift resources into them and out of other on-station activities.

Advanced PVS and the use of farmer ratings

The products of the preliminary PVS trials describedabove must be evaluated in advanced PVS trials by largenumbers of farmers in all the conditions that occurwithin the target environment. Such trials, which mayrequire the participation of hundreds of farmers, mustbe designed to require limited researcher involvementin each trial. The informal research and development(IRD) design described by Witcombe et al. (2001), inwhich each participating farmer evaluates only one testvariety, is suitable for this purpose.

Farmer assessments are collected through surveys inwhich farmers use a simple numerical scale to rate testcultivars as better than, the same as, or worse than theirown cultivar for agronomic and quality traits. Surveyquestions must be carefully designed to target thecharacteristics that are important to farmers, andadministered in a way that facilitates quantitativeanalysis. Questions can be developed using informationobtained in the steps above.

Some breeding programmes will be unable to managedirect farmer contacts on the scale needed for advancedPVS, but will need to link with extension or developmentorganisations to recruit farmers, distribute seed, andadminister surveys. However, advanced PVS can beefficiently undertaken directly by researchers linked inmulti-station breeding networks, if each station managesthe programme in several nearby villages.

The use of farmer ratings to replace agro-economic measurements in advanced PVS

The relationship between farmer yield ratings andresearcher-measured yield

Successful implementation of large-scale PVS pro-grammes in rainfed rice will require the use of methodsthat minimise researcher involvement per trial.In general, this will involve the use of farmer ratingsrather than quantitative measurements of yield and othertraits.

38 Integrating conventional and participatory crop improvement in rainfed rice

It is important, therefore, to estimate the correlationbetween farmer ratings and quantitative measurements.This has been done for grain yield in PVS trials in severalcountries and ecosystems (Table 1). In both upland PVStrials in the Philippines and a rainfed lowland trial ineastern India, farmer ratings of cultivar yield were highlycorrelated with researcher-measured yields. Farmerperception of yield captured by a simple rating systemtherefore appears to be strongly associated with grainyield measurements taken by researchers. It is likely thatyield ratings collected in surveys may substitute for crop-cut measurements in large-scale PVS trials, greatlyreducing their cost.

Genotype x environment interaction in PVStrialsOne of the proposed effects of increased farmerparticipation in plant breeding is increased local orspecific adaptation of cultivars. This would imply thatcultivars emerging from PPB and PVS programmesexhibit increased G x E interaction. This hypothesis wastested in farmer-managed, multilocational PVS trials inJharkhand, India and in northern Laos. Little cultivar xvillage interaction was observed for grain yield in theIndian trial (Table 2). No significant cultivar x provinceinteraction was observed for grain yield in Laos, but astrong interaction was observed for farmer preferenceamong early cultivars (Table 3). The reasons for thisprovince-specific preference are unknown.

Table 2. Variance components for upland ricecultivars evaluated under farmer managementin four villages over 3 years near Hazaribag,Jharkhand, India

Source of variation Variance component

Cultivar 0.077Cultivar x location 0Cultivar x year 0Cultivar x location x year 0.246

Table 3. Significance level of F-tests of cultivar andcultivar x province interaction for eight early-maturing and eight late-maturing cultivarsevaluated in Luang Prabang and Sayaboulyprovinces, Laos, 2001

Source of variationCultivar x

Trait Cultivar province

Yield of early cultivars * ns

Yield of late cultivars ns ns

Preference rating forearly cultivars ns *

Preference ratingfor late cultivars * ns

Overall, these results indicate that superior cultivarsidentified in PVS trials may be broadly adapted andperform well at a regional level, a finding that has alsobeen reported by Joshi et al. (2001), and is consistentwith the broad farmer-to-farmer diffusion of preferredrainfed rice cultivars such as Mahsuri (Mackill et al.,1996).

However, local preferences for cultivars withparticular agronomic traits may also be detected in PVStrials with sufficient replication within environmentalsubclasses such as regions or management regimes topermit cultivar x class interactions to be tested againstthe pooled cultivar x farm variance within classes (Atlinet al., 2000).

Table 1. Correlations among researcher-measuredyields, farmer yield ratings, and farmer preferenceratings in on-farm PVS trials in India, Laos, and thePhilippines

Trial location Correlation

Farmer yield ratings versusresearcher-measured yieldsBatangas, Luzon, Philippines 0.72Faizabad, Uttar Pradesh, India 0.90

Researcher-measured yields versusfarmer preference ratingsTamanbogo, Sumatra, Indonesia 0.66Faizabad, Uttar Pradesh, India 0.67Luang Prabang, Laos (medium-maturing) 0.82Luang Prabang, Laos (early-maturing) 0.54

Farmer yield ratings versus farmerpreference ratingsFaizabad, Uttar Pradesh, India 0.81Batangas, Luzon, Philippines 0.80

The relationship between grain yield and farmerpreference

Questions have been raised about the extent to whichrainfed rice producers, who grow rice mainly for foodsecurity, are concerned with grain yield in comparisonwith other production parameters. This question wasaddressed by examining the correlation between farmerpreference ratings and grain yield measurements orratings in PVS trials conducted in the Philippines, India,Indonesia, and Laos. In all cases, although there werehigh-yielding cultivars that were not preferred, there wasa strong correlation between measured grain yield andpreference ratings (Table 1). The relationship betweenfarmer yield ratings and farmer preference at harvest wasat least as strong as the relationship between researcher-measured yield and preference. Grain yield, either asmeasured by researchers or as visually perceived byfarmers, appears to be an important criteriondetermining farmer preference.

Integrating conventional and participatory crop improvement in rainfed rice 39

In general, our results show that participatoryapproaches can be usefully integrated at several pointsin rainfed rice breeding programmes, including identi-fication of farmer priorities, initial elimination ofunsuitable genotypes, and the final stage of cultivarevaluation. Farmer ratings for grain yield are closelyassociated with both preference and measured yield, andcan serve as the basis for selection decisions in large-scale PVS trials, permitting low-cost scaling-up ofparticipatory selection. PVS products appear to bebroadly adapted across farms within the targetenvironments sampled in these studies.

Integrating PVS in rainfed rice breedingnetworksBecause of the sporadic nature of drought and floodingstresses in rainfed rice production environments,resulting in a large random G x EI variance, large-scale,multilocational testing systems that link small breedingprogrammes together have proved to be an effectiveapproach for increasing breeding progress. In the EasternIndian Rainfed Lowland Shuttle Breeding Network, IRRIcollaborates with nine Indian breeding programmes atIndian Council of Agricultural Research (ICAR) researchstations and state agricultural universities to jointlydevelop and evaluate lines across the spectrum of rainfedrice production environments.

A similar interstation network, involving eight sites,has been operating for 2 years in northeast Thailand.These breeding networks annually evaluate over 130 linesin multilocational trials at high levels of precision, andare ideally structured to implement large-scale PVS asthe final step in the varietal evaluation process. Networksin which each individual centre manages a preliminaryMother and Baby PVS trial requiring researcher layoutin three or four communities, with the participation ofthree to five farmers per community, could, in colla-boration, obtain information on varietal performanceand farmer preference from 70–150 farmers in a singleseason at a manageable cost in time and resources.

Large-scale advanced PVS networks relying on theuse of farmer ratings could sample many hundreds offarmers in the target region, permitting local adaptationto be reliably detected if it exists. Lines selected fromsuch networked PVS trials are likely to be broadlyadapted, stress tolerant, and acceptable to farmers.

ReferencesAtlin GN, Cooper M, Bjørnstad Å (2001) A comparison of formal

and participatory breeding approaches using selection theory.Euphytica 122: 463–475

Atlin GN, Baker RJ, McRae KB, Lu X (2000) The effect of subdividinga target region on selection response. Crop Sci 40: 1–6

Cooper M, Rajatasereekul S, Immark S, Fukai S, Basnayake J (1999)Rainfed lowland rice breeding strategies for northeast Thailand.I. Genotypic variation and genotype x environment interactionsfor grain yield. Field Crops Res 64: 131–151

Joshi KD, Sthapit BR, Witcombe JR (2001) How narrowly adaptedare the products of decentralized breeding? The spread of ricevarieties from a participatory plant breeding programme inNepal. Euphytica 122:589–597

Mackill DJ, Coffman WR, Garrity DT (1996) Rainfed lowland riceimprovement. International Rice Research Institute, Los Baños,Philippines. pp 242

Suwarno BK, Arjasa WS, Atlin GN (2002) Participatory selection onupland rice in Sumatra. Pages 61–63 in Breeding rainfed rice fordrought-prone environments integrating conventional andparticipatory plant breeding in south and southeast Asia:proceedings of a DFID Plant Sciences Research Programme/IRRIConference, 12–15 March 2002, IRRI, Los Baños, Laguna,Philippines (Witcombe JR, Parr LB, Atlin GN eds.). London andManila, Department for International Development (DFID)Plant Sciences Research Programme, Centre for Arid Zone Studies(CAZS) and International Rice Research Institute (IRRI)

Witcombe JR, Joshi KD, Rana RB, Virk DS (2001) Increasing geneticdiversity by participatory varietal selection in high potentialproduction systems in Nepal and India. Euphytica 122: 575–588

40 Participatory scaling up of participatory varietal selection

BackgroundFor over 5 years in Nepal, Local Initiatives for Bio-diversity Research and Development (LI-BIRD), incollaboration with the Centre for Arid Zone Studies(CAZS), has been developing and promoting partici-patory crop improvement (PCI) and participatory plantbreeding (PPB) approaches. These approaches have beenfound to be simple, rapid and effective in identifying anddisseminating farmer-preferred technologies (Joshi et al.,2001; Joshi and Witcombe, 2002).

The next important step was to find better ways ofscaling up to improve cost effectiveness. The firstassumption was that the active involvement of farmersin this process – ‘participatory scaling up’ – would greatlyimprove cost effectiveness. Participatory scaling upinvolves farmers as seed producers and disseminators.When there is a high degree of farmer involvement it iseasier to benefit resource-poor farmers in remote regions,but the technique also works well for more favourableenvironments. Another important advantage is thattechnology verification and adoption take placesimultaneously. We begin by briefly reviewing the roleof farmers before discussing the different stakeholdersin this process and the partnerships that are evolving inNepal.

Participatory scaling up of participatory varietal selection

KD JOSHI1, S GYAWALI2 AND JR WITCOMBE3

1. Department for International Development (DFID) Plant Sciences Research Programme (PSP), Centro Internacional deMejoramiento de Maiz y Trigo (CIMMYT), South Asia Office, Singh Durbar Plaza Marg, Bhadrakali, Kathmandu, Nepal

2. Local Initiatives for Biodiversity Research and Development (LI-BIRD), PO Box 324, Bastolathar, Mahendrapul, Pokhara, Nepal3. Department for International Development (DFID) Plant Sciences Research Programme (PSP), Centre for Arid Zone Studies

(CAZS), University of Wales, Bangor, Gwynedd LL57 2UW, UK

Farmers are the ultimate beneficiaries of new agricultural technologies and the best indicator for the success of a technologyis its acceptance and uptake by them. We discuss the benefits of the active involvement of farmers in scaling up the resultsof participatory varietal selection (PVS) and participatory plant breeding (PPB). We discuss the stakeholders and partnershipsfor participatory scaling up of varieties. In Nepal, we describe the evolving partnerships between government organisations(GOs) and non-government organisations (NGOs), and community-based organisations (CBOs) that is designed to combinethe strengths of the stakeholders whilst sharing responsibilities and resources.

The role of farmers’ participationParticipatory scaling up is designed to maximise thecontribution of farmers in the process. For example,farmers who are key individuals in the informal seedsupply system in the village are identified. Such farmersare contracted to produce seed for the project fordissemination by informal research and developmentmethods (see Appendix), or they are given good-qualitysource seed and asked to grow it and disseminate theharvest through their usual informal network. Duringthe participatory scaling up process individual initiatives,the actions of farmers’ groups (FGs), and the results ofinformal farmers networks have been documented toillustrate the role of farmers in the process.

One example of the role of an individual is DayaRam Khandka, Chairman of the Dhadwar VillageDevelopment Community (VDC), Bardiya District. Heis also a farmer, and a member of the Kulanath KrishakSamuh (an FG). Khandka took on the responsibility ofdistributing seeds, periodically monitoring activities, andmaintaining records of the yield and yield componentsof all the varieties on offer from a LI-BIRD and DistrictAgricultural Development Office (DADO), Bardiyacollaboration.

The FG saved all the seeds of all the varieties forsowing in the following year. They have their own normsfor exchanging seeds; any farmer wishing to obtain seed

Abstract

Participatory scaling up of participatory varietal selection 41

of new a variety shortly after harvest within the monthof Mangsir (16 Nov–15 Dec) can do so just by exchangingseed for rice grains. However latecomers will have to buythe seed paying cash at a rate of two rupees less than thestandard price for seed fixed by the Agricultural InputsCorporation (AIC). This community initiative for vari-etal evaluation and seed management indicates a sustain-able way of achieving impact from PVS.

One of the greatest displays of interest in testing andadopting new varieties from a single-season’s exposurewas at Beladevipur village, Kailali District. Most of thepeople in the village whom the monitoring team wereable to see could tell the names of all the varieties testedat that site, the relative performance of all the varieties,and which particular variety they would grow next year.Such interest is only possible when an external agencyhas been the catalyst in providing seeds and informationso that farmers have interacted intensively with theirfellow farmers on the merits of the new varieties.

There was also good evidence of farmer-to-farmerseed spread in Kailali District. Nearly 1.5 t rice seed ofthe new varieties will be sown in the 2002 season at allthe sites in Kailali from an initial 300 kg that wasdistributed (Ram Ayoudhaya Mahato, Chief DADO,Kailali, pers. comm.)

The spread of variety BG 1442 in Chitwan andMakwanpur Districts provides one of the most convin-cing examples of the effectiveness of farmer networking.The variety was introduced in farmers’ field trials duringthe early 1990s as a Chaite (early season) rice variety butit is yet to be formally released by the national researchsystem. The variety is promising for the Chaite season.It is adapted to partially irrigated well-drained conditionsas a main-season rice variety and is equally suitable forthe Bhadiaya system (rainfed land in the Terai whereearly-maturing rice varieties are grown). The crop issown in May and is called Bhadaiya because it is harves-ted in the Nepali month of Bhadra (late Aug–Sept).

The variety is spreading very quickly from farmer tofarmer since LI-BIRD introduced a few kilos of seed ofthis variety in participatory varietal selection (PVS) trialsin 1998 and 1999, and subsequently in informal researchand development (IRD) from 2000 to 2002. The varietyincreased rapidly in area grown from 2001 to 2002 tocover nearly 1200 ha, about one third of the Chaite areain Chitwan District. The spread was encouraged by thebreakdown of the predominant variety CH 45 to blastdiseases. BG 1442 is also spreading in villages in theneighbouring district of Makwanpur where it coverednearly 50 ha in the main season of 2002. Most of this isfrom farm-saved seed spreading through the farmer tofarmer network, facilitated by distribution of source seedby DADO, Chitwan, LI-BIRD and some farmercooperatives in eastern Chitwan.

Although farmer-to-farmer spread of seed is veryimportant and has been the major source of seed in mostdeveloping countries, including Nepal, it is not the fastestsystem. Farmers consume seed on-farm and sell it asgrain in the markets, so only a proportion can be usedas seed for sowing. Our experience with most of thefarmer-to-farmer spread has been that a three-to four-fold increase is often found, but a sixty-fold increase caneasily be achieved with an organised seed increase anddistribution programme. This efficient source ofofficially multiplied seed can be used most effectivelywhen it is distributed widely and to many farmers.

Stakeholders and modes of partnershipOnce the central role of farmers has been established, afirst step in scaling up is the identification of stakeholdersand devising the most appropriate mode of partnership(Table 1). The lack of lasting effects of some of theresearch and development projects designed and runthrough a top-down approach in the past was due to thelack of farmers’ participation, and ignoring the centralrole that farmers can play. There was also a lack ofidentification of all stakeholders and their roles andresponsibilities.

Scaling up is more complex than simple technologygeneration or verification because it involves morestakeholders. Technology generation often ignoresfarmers and multi-partner approaches. A multi-partnership in which farmers play a central role is themost suitable one for scaling up (Table 1).

The benefits of this approach are particularlyimportant when, as is common, linkages betweenresearch and extension are poor and governmentextension services are restricted by limited resources toserving only the more accessible areas.

Table 1. Stakeholders1 for participatory scaling up

Stakeholders Mode of partnership

Donor CGIAR NARS Ignores farmer, extensionand NGO sector

Donor CGIAR NAES Ignores farmer, researchand NGO sector

Donor CGIAR NGOs Ignores farmer and GOs

Donor NGOs Ignores farmers, GOs andCGIAR

Farmer NARS NAES Multi-partner collaborationof all stakeholders. Farmersfirst

NGOs CGIAR Donor –

1. CGIAR = Consultative Group on International Agricultural Research;NARS = National Agricultural Research System; NAES = NationalAgricultural Extension System; NGO = non-governmental organisation;GO = governmental organisation

42 Participatory scaling up of participatory varietal selection

Partnership modes for participatory scaling up

NGO and community partnerships

An important form of collaboration that is emerging forparticipatory scaling up is collaboration between NGOsand community-based organisation (CBOs). The CBOsare usually FGs that were already established forcooperative enterprises or for specific users’ interest –e.g., drinking water.Three CBOs were as follows:1. Women’s group (Narayani Mahila Milan Club,

Amaltari, Kumarti-9)2. Mixed FG (Gramin Swablamban Bikas Kendra,

Bishaltar, Devchuli-1) and3. Water-users’ group (Naya Belhani Khanipani

Upabhokta Samiti, Naya Belhani-2)LI-BIRD signed a Letter of Agreement (LoA) with

three CBOs in Nawalparasi. One of which was formedby the DADO office while the other two were self-organised. In this type of partnership, both organisationshave a high degree of institutional flexibility. Its successwill depend on both types of partners respecting eachother’s roles and responsibilities (Box 1) and, in the caseof LI-BIRD and the three CBOs there were three separateLoAs to make these facts clear. However, the scale ofoperation of NGO and CBO partnerships will usuallybe limited as most FGs are local organisations.

NGO and GO partnerships

LI-BIRD and DADO Chitwan. One mode of working iscollaboration between an NGO and a GO. An examplein Nepal is the partnership between LI-BIRD and thegovernment line agency DADO in Chitwan District.One unique feature of this partnership was the signingof an LoA for 3 years, which defined the role and

responsibilities of the two partners (Box 2). However,the sustainability of this approach could be improved iffunding were from the recurrent budget of His Majesty’sGovernment of Nepal rather than a special projectfunded by the Plant Sciences Research Programme.

This agreement between an NGO and a GO may beone of the first examples of its kind in Nepal (LoA signedby the District-level Authority of the Department ofAgriculture (DoA) rather than at the Department level)although there are other examples of GO–NGOcollaboration for specific development objectives at theDepartment level, such as the collaboration betweenCARE Nepal and the Departments of Soil and WaterConservation and the Department of Health Services.

The initial interest in formal collaboration came fromthe GO. However, informal collaboration started beforethat. When LI-BIRD initially set up the research projectit consulted DADO Chitwan on the identification ofproject villages and DADO staff regularly participatedin the monitoring of the participatory varietal trials.DADO then started scaling up unreleased farmer-preferred crop varieties using their own funds for over ayear.

Expanding the collaboration to more district offices.A stakeholders meeting was held in Kathmandu inFebruary 2001 to discuss building on the existingpartnership between LI-BIRD and DADO, Chitwan. Themeeting agreed on the advantages of participatoryscaling up and set up a joint Working Group of the DoA,Nepal Agricultural Research Council (NARC) and LI-BIRD. The Working Group developed a proposal to scaleup the activities in Chitwan to four more districts:Dhanusha, Sarlahi, Bardiya and Kailali. This wassubsequently approved by the authorities from DoA,

Scaling up through farmers’ groups in Nawalparasi District, Nepal

The process● Participatory rural appraisals (PRAs) to identify community-based organisations (CBOs) suitable for participatory

scaling up, and the analysis of secondary information through agricultural service centres● Documentation of profile of short-listed farmers’ groups (FGs) by PRA● Selection of established FGs● Letter of Agreement (LoA) signed with three FGs, two mixed and one of women farmers● Implementation of programme through FGs.

Role of FGs● Responsible for carrying out all field activities: farmer identification, (focus on poor and medium-poor farmers),

distribution of seeds, record keeping, including the monitoring of varietal adoption● Seed production and farmer-to-farmer distribution● Group decision making.

Role of LI-BIRD● Impart necessary skills for scaling up through orientation, training and exposure visits● Provide technical backstopping● Assist in monitoring and evaluation● Provide financial and institutional support for scaling up.

Box 1

Participatory scaling up of participatory varietal selection 43

Letter of Agreement between DADO Chitwan and LI-BIRD for participatory scaling up

Roles of DADO● Full responsibility for implementing scaling-up activities through their networks● Preparing a joint annual plan and obtaining the approval of the Ministry through the regular planning process● Periodical monitoring – collecting field information and preparing progress and technical reports● Providing financial and other institutional support to seed production of the new, unreleased varieties, as applicable to

released varieties● Farmer and staff skill development on participatory approaches through orientation, training and exposure visits● Organising periodical coordination and review meetings with LI-BIRD.

Roles of LI-BIRD● Arranging source seeds of farmer-preferred varieties for scaling up● Assisting in preparing the annual plan● Imparting skills to DADO staff through joint activities and through orientation and training programmes● Sharing in data analysis and report writing● Organising seasonal monitoring tours involving DADO, DoA, NARC and IAAS staff● Providing agreed financial and institutional support.

NARC and LI-BIRD. The World Bank-fundedAgricultural Research and Extension Project (AREP)agreed to provide a small grant for this work.

Farmer-preferred rice varieties were produced anddisseminated in 2001 in the major rice-growing domainsin each of four districts. Farmers tested the new ricevarieties in long-standing water, irrigated, partiallyirrigated, and rainfed conditions, and for the Bhadiayasystem.

Consolidation and further expansion tomore organisationsA workshop was jointly organised in January 2002 byDoA and LI-BIRD to share the findings from the fivecollaborating district offices. Representatives from sevenDADOs participated and developed seven projectconcept notes (PCNs) for participatory scaling up ofPVS-identified and PPB-produced varieties.

The possibility of incorporating the work plans ofthe PCNs into the regular planning process of the DoAto attract regular funding was discussed. Some of theDADOs have already done so, indicating a greateracceptance and uptake of the approach by the DoA.

Although the extent of activities differs greatly,currently activities have been extended to 17 Terai andtwo hill districts. In these initiatives, LI-BIRD is alsocollaborating with such other NGOs as FORWARD,REGARDS, CARE and PLAN International, in additionto working closely with the DoA.

OpportunitiesCurrently there is a helpful environment for participatoryresearch and scaling up in Nepal. The Ministry ofAgriculture and Cooperatives has adopted partnershipas one of the main approaches to agricultural researchand development. His Majesty’s Government of Nepalhas set up a competitive funding system for actionresearch and scaling up called the National AgriculturalResearch and Development Fund (NARDF). Agriculturalextension will very soon be devolved to the localgovernment, which necessarily involves decentralisationand a greater emphasis on people’s participation. DADOsare also considering contracting out of some of theirservice delivery functions to improve their efficiency.

AcknowledgementThis publication is an output from projects R7281 andR7542 funded by the UK Department for InternationalDevelopment (DFID) Plant Sciences Research Progra-mme (PSP) for the benefit of developing countries. Theviews expressed are not necessarily those of DFID.

ReferencesJoshi KD, Sthapit BR, Witcombe JR (2001) How narrowly adapted

are the products of decentralised breeding? The spread of ricevarieties from participatory plant breeding programme in Nepal.Euphytica 122 – 589–597

Joshi KD, Witcombe JR (2002) Participatory varietal selection in ricein Nepal – a comparison of two methods by farmers’ selectionand varietal adoption. Euphytica 127: 445–458

Box 2

44 Farmers’ participatory breeding for upland rice in eastern India

IntroductionMost upland rice farmers, particularly in eastern India,are resource-poor and are not able to use purchasedinputs. Improved varieties developed for rainfed uplandshave not been fully adopted by farmers. The localcultivars have the desirable characters to withstandunfavourable growing conditions but they yield less thana tonne per hectare. Traditional breeding approaches fordeveloping improved rice varieties have not been veryeffective. By the time on-farm trials are conducted, themajority of the genotypes that might have appearedpromising from the farmers’ point of view have usuallybeen discarded. Farmers prefer varieties that will performwell under multiple cropping and integrated farmingsituations. They are also looking for varieties that mightprovide products and by-products for household needsother than food.

Past experience clearly shows how breeders have failedto adequately address the problems of upland eco-systems. Variety adoption has been meagre and spreadhas been slow and restricted. The impact of modernvarieties was poor mainly due to their shorter stature,lack of early vigour and poor responsiveness to low-inputmanagement. Farmers are not in favour of taking risks

Farmers’ participatory breeding for upland rice in easternIndia

NP MANDAL1, PK SINHA1, RK SINGH1, M VARIAR1, RK SINGH2 AND GN ATLIN2

1. Central Rainfed Upland Rice Research Station (CRURRS), Central Rice Research Institute (CRRI), PO Box 48, Hazaribag 825 301,Jharkhand, India

2. International Rice Research Institute (IRRI), DAPO Box 7777, Metro Manila, Philippines

in unpredictable environments. On-farm research withfarmers’ participation aims at involving farmers inselecting and testing new varieties along with existingtraditional types under their own management practices.A participatory breeding project was started in 1997 withthis objective, under the Indian Council of AgriculturalResearch (ICAR) – International Rice Research Institute(IRRI) collaborative programme. The project has twomajor components – plant breeding and social science.The breeding component consisted of two distinct parts:● Participatory varietal selection (PVS), where existing

cultivars or fixed breeding lines were exposed toselection by farmers (Witcombe et al., 1996)

● Participatory plant breeding (PPB), in which farmerswere involved in selecting lines from segregatingpopulations.

Materials and methods

Site selection

Three villages in Hazaribag District were selected for thisproject on the basis of their representation of theenvironments targeted in the breeding work, diversityand range of ecological conditions, involvement ofwomen in farm activities, availability of previous survey

Lack of access to new, high-yielding rice varieties, and to information about new varieties, are major constraints for resource-poor farmers in eastern India. There is no effective local seed system, and new varieties have not been adopted because oftheir lack of early vigour, short stature and poor response to low-impact management. Although local cultivars can survivethe poor environmental conditions, they do not produce high yields. Traditional breeding approaches have been ineffectivein producing modern, high-yielding varieties that can withstand growing conditions under local management practices.A participatory breeding project begun in 1997 aimed to get farmers involved in testing new varieties on their own fieldsalong with traditional cultivars. Results were outstanding. Farmers consistently selected more successful varieties thanbreeders, and were able to select superior progenies from the segregating generation. The project showed that participatoryvarietal selection (PVS) and participatory plant breeding (PPB) are excellent ways of helping breeders to understand theproblems facing farmers, and allow the farmers to take an active part in driving their farming practices forward.

Abstract

Farmers’ participatory breeding for upland rice in eastern India 45

data and easy access to the site. Upland rice is grown onapproximately 20% of the total cultivated area. Farmersgrow Brown Gora rice in upland systems though its yieldis very low (usually less than 1 t ha-1).

Farmer selection

Ten farmers from each village were selected, based ontheir primary occupation being rice farming, especiallyupland rice, ownership of the operational land holding,and some degree of literacy. With a few exceptions mostof the farmers selected in the first year continued toparticipate until the third year.

Participatory varietal selection (PVS)

The PVS trials were conducted for 3 consecutive yearsduring 1997–99, with 15–16 upland rice genotypes(including the local check, Brown Gora). In the first year,the trial was conducted in farmers’ fields in three villages,one set managed by farmers and another set managedby breeders in adjacent blocks. Breeders also conductedthe same trial at the research station. From the secondyear the trial was conducted in all three villages underfarmer management, except for the seeding, and at theresearch station by breeders. Most of the test entries ofthe PVS trial were continued for 3 years. A few entriesranked as very poor by both farmers and breeders werereplaced with new ones in the second and third years oftesting. The trials were unreplicated complete-blockdesigns (CBD) in farmers’ fields and replicatedrandomised complete-block designs (RCBD) at stations.The participating farmers of a given village and thebreeders evaluated all the trials conducted in that villageand also at the research station. The genotypes wereranked at two phenological stages (vegetative andreproductive) as ‘most-liked’ to ‘least-liked’, on the basisof their selection criteria. In addition, breeders alsorecorded duration, height, yield and yield components,and reaction to diseases and pests in all the trials. Tocompare the ranks given by farmers and breeders, theKendall coefficient of concordance (W) was used asdescribed in Siegel (1956). Farmers’ and breeders’rankings were compared following the Spearman rankcoefficient of correlation. Details about the methodologyused can be found in Courtois et al. (2001).

Participatory plant breeding (PPB)

PPB started one year after PVS, after farmers were givensome basic training on the objectives and methodologyof single plant/line selection in a segregating population.One hundred segregating lines from 12 crosses in the F

5

generation were grown at one on-farm site (Khorahar)and at the research station. Lines were scored by farmersand breeders at both sites at different crop stages. Atmaturity, single-plant selections were made separatelyby farmers and breeders at both sites. The plants selectedby farmers and breeders on-station and at Khorahar were

grown in separate blocks in the same field in theirrespective sites in the 1999 wet season. The farmers andbreeders continued their selection on the respectivematerials and sites and, during the 2000 wet season, foursets of final (bulk) selections were made, i.e., breeders’and farmers’ selections on-station, and breeders’ andfarmers’ selections on-farm. These four sets were pooled,along with three checks (Brown Gora, Kalinga III, andVandana) and evaluated both on-farm and on-stationduring the 2001 wet season in an alpha-lattice designwith two replications. Analysis of the PPB trial data wasconducted using a mixed model, with selector andselection environments taken as fixed, and lines withinselection environment x selector combinations consi-dered random. The analysis was conducted with theREML algorithm of SAS PROC MIXED.

Results

PVS trials

The coefficient of concordance among farmers washighly significant in all the trials conducted in all 3 years.This indicated that farmers’ rankings were not randomlyattributed and that there was a good agreement amongthem. The concordance among breeders’ rankings wasalso high but often not significant because of the smallernumber of breeders. The rank correlation betweenfarmers’ and breeders’ average rankings was highlysignificant except in a few cases (1997 on-farm resultsand on-station in 1999). This indicated a good agreementbetween farmers and breeders in varietal choice. Thefarmers’ or breeders’ rankings at maturity were notalways highly correlated with yield, especially thefarmers’ ranking in low-yielding trials. This may bebecause farmers consider criteria other than grain yieldwhile selecting varieties for their farms. It may also bebecause ranking was undertaken before harvest.

There was no correlation between farmers’ andbreeders’ rankings and duration or plant height, with afew exceptions. The variance component analysis for theon-farm trials showed that there were large andsignificant differences among cultivars in grain yieldunder farmer management. The grain yield data averagedover the four on-farm sites and 3 years showed that oneelite line (RR 348-5) significantly outyielded Vandana,and produced about 100% more grain than Brown Gora(Table 1).

No variety x location or variety x year interaction wasdetected (Table 2). The three-way interaction, variety xlocation x year, was large, but cannot be separated fromthe within-trial error in this analysis, as the trials wereunreplicated. This result does not support the hypothesisthat varieties exhibit specific adaptation to particularsub-environments within the target region of theresearch station. The repeatability or broad-senseheritability of different on-farm and on-station trials

46 Farmers’ participatory breeding for upland rice in eastern India

were estimated to judge the precision of the trial. It wasfound that the repeatability of on-station trials was poor,compared to on-farm trials. This may be because of thelow yield of the on-station trials, where continuouslygrown upland rice results in poor soil fertility.

The regression of on-farm performance on on-stationperformance was non-significant. This means that on-station testing is not able to predict the on-farmperformance of the genotype, and there is a need tointegrate on-farm testing at an early stage of the breedingprogramme. At the end of 3 years of PVS trials farmersselected five varieties: RR 354-1, RR 347-166, RR 151-3,RR 166-645 and RR 51-1 using their own selectioncriteria. The highest-yielding line, RR 348-5, was notselected by the farmers. Farmers do not consider yieldthe only important trait.

RR 166-645 was preferred for its long, slender grains.RR 51-1 appealed to the farmers for its dark green leavesand high tillering ability. Farmers preferred tall varietiesas they need fodder for their cattle. Farmers’ perceptionon cooking quality was that RR 354-1 and RR 347-166were better than their local variety or even Vandana.

During the 2000 wet season, seeds of these varietieswere given to farmers in two villages to grow in theirfields along with the check (Brown Gora). In spite ofsevere drought, RR 347-166, RR 151-3 and RR 354-1

performed better than the farmers’ variety in most ofthe farmers’ fields. However, RR 51-1 failed as it wasfound to be susceptible to drought. Drought tolerancewas not taken into account by farmers, as the last 3 yearswere almost favourable. Now we need to improve thisotherwise good genotype for drought tolerance.

Mother and Baby trials (see Appendix) were started

Table 2. Variance components for cultivars evaluatedunder farmer management in four villages nearHazaribag, Jharkand, over four wet seasons

Source Variance component

Cultivar 0.077Cultivar x location 0Cultivar x year 0Cultivar x location x year 0.246

Table 3. Mean grain yield (t ha-1) of upland cultivarsand breeding lines evaluated under farmermanagement in three villages and on-station inHazaribag, Jharkhand, wet season 2001

Cultivar Chichi Khorahar Peto Mean1 CRURRS2

RR 354-1 1.32 1.47 1.70 1.49 2.67RR 347-1 1.46 1.53 1.30 1.43 2.92RR 347-166 1.61 1.55 1.10 1.42 3.25RR 348-5 0.74 1.65 1.50 1.30 2.78RR 345-2 1.47 1.38 1.00 1.29 2.30Kalinga III 1.13 1.30 0.80 1.08 1.83RR 361-1 1.03 1.28 0.90 1.07 2.35CR 876-6 1.29 1.20 0.70 1.06 1.95RR 51-1 1.44 0.87 0.80 1.04 2.67Brown Gora 1.26 1.22 0.60 1.03 1.85RR 151-3 1.09 1.10 0.50 0.90 2.10Vandana 1.01 0.95 0.70 0.89 2.43RR 166-645 1.24 0.87 0.40 0.84 2.25RR 363-737 0.90 0.92 0.60 0.81 2.38RR 139-1 0.55 1.32 0.35 0.74 1.87RR 361-783 0.51 0.90 0.33 0.58 1.15LSD

.050.70 0.70 0.70 0.42 0.76

1. Over villages. 2. On-station

Table 1. Mean grain yield (t ha-1) of upland cultivarsand breeding lines evaluated under farmermanagement in four villages near Hazaribag,Jharkand, over four wet seasons

Cultivar Number of trials Mean1

Brown Gora 12 1.12Vandana 11 1.74RR 347-166 12 1.92RR 348-5 12 2.25RR 151-4 9 1.11RR 203-16 9 1.47RR354-1 12 2.12RR 51-1 12 1.78RR 151-3 12 1.68RR 166-645 12 1.49RR 50-5 9 1.63RR 139-1 12 1.25LSD

.05 for means over nine trials 0.47

1. Over years and locations

Table 4. Combined analysis of variance for cultivarsevaluated under farmer management in threevillages near Hazaribag, Jharkand, wet season 2001

VarianceSource MS F Pr>F component

Cultivar 0.64 3.43 0.002 0.05

Cultivar x location 0.19 1.49 0.077 0.02

Residual 0.13 0.13

during the 2001 wet season. The analysis of Mother trialdata at three on-farm sites showed that there were largeand significant differences among cultivars in grain yieldunder farmer management (Table 3). The on-stationdata are not included in the ANOVA because there werelarge differences between on-station and on-farmperformance. There was no significant cultivar x locationinteraction (Table 4). This implies that there was no rankchange in varietal performance within this sub-environment.

Averaged over the three sites, RR 347-1, RR 354-1,and RR 347-166 significantly outyielded Vandana(Table 3). These lines also performed well during 1999–

Farmers’ participatory breeding for upland rice in eastern India 47

2000. They yielded about 40% more than Brown Goraand about 30% more than Kalinga III in 2001, andconsistently outperformed these lines in previous years.They should be evaluated in large-scale PVS trials.Variance component analysis was conducted to estimatethe precision of on-farm trials through estimation ofrepeatability or broad-sense heritability. The broad-senseheritability of grain yield measured in the on-farm trialswas relatively high, as it was in previous years. The resultsshowed that evaluation of genotypes at three sites, withthree replicates per site, would give adequate precisionfor the detection of cultivar differences (Table 5).

In Baby trials, three varieties were given to each

significantly outperformed breeder’s selections, and thatselection on-station was superior to selection on-farm(Table 6). Farmers’ selections yielded almost twice asmuch as breeders’ selections (Table 7). Of the fivehighest-yielding lines, four were selected by farmers onthe research station (Table 8). They were also foundsuperior to check varieties.

Table 5. Predicted broad-sense heritability (H) ofgrain yield for cultivars evaluated under farmermanagement in three villages near Hazaribag,Jharkand, wet season 2001

Sites Replicates H

1 1 0.231 3 0.443 1 0.503 3 0.70

Table 6. F-tests for effects of selection by farmersversus breeders and selection on-farm versus on-station on the grain yield of selected linesevaluated under farmer management at Khorahar,Jharkand, 2001

Source F Pr. >F

Selector (farmer vs breeder) 16.6 0.0002Selection environment 20.4 0.0001 (station vs farm)Selector x selection 5.5 0.0246 environment

farmer to grow along with the local variety. Farmers’perceptions for different varieties were recorded. AtChichi, two farmers preferred RR 354-1, and one farmereach preferred RR 347-1, RR 363-737, Vandana andKalinga III. In most of the trials severe weed infestationwas the reason given by farmers for failure of the trials.In a few cases drought damaged the crop at key pheno-logical stages.

PPB

Farmers concentrated on fewer crosses than breeders.They rejected crosses that did not produce plants thatfit their criteria. In general, farmers preferred breedinglines with tall stature, high tillering and long panicles.Plants generated from certain crosses (VHC 1253 x Sathi34–36, N 22 x RR 20–5, Annada x RR 151–3 and RR 139–1 x IR 57893–08) were preferred by both farmers andbreeders.

The results indicate that on average, farmer selections

Table 8. Means and selection history of the checksand five highest-yielding lines

Grain yieldLine Selection history (g plot-1)

RR 356-77 Farmer-selected on-station 831RR 356-72 Farmer-selected on-station 825RR 356-74 Farmer-selected on-station 769RR 356-71 Farmer-selected on-station 569RR 356-51 Breeder-selected on-station 519Kalinga III 413Vandana 413Brown Gora 256LSD

.05412

DiscussionFarmer agreement on ranking varieties was highlysignificant, although there were differences in opinion.The relatively high agreement contradicted our initialassumption that farmers’ preference would vary becauseof the diversity in their socioeconomic backgrounds. Thismay be due to the limited number of farmers involvedin the project, or to the low diversity in wealth, caste,ethnicity etc. The high agreement among breedersindicated similarity in selection. The agreement betweenfarmers and breeders was good in most of the cases.Participation brings little improvement when there isclose agreement between farmers and breeders.

The degree of agreement is highly influenced by thematerials used for the selection. For example, in theuplands of eastern India, the variety has to be tall andmature within 100 days. Less variability within the testedgenotype perhaps influenced the farmers’ and breeders’preference for variety or traits. The rankings of farmers

Table 7. Effect1 of selection by farmers versusbreeders and selection on-farm versus on-stationon the grain yield (g plot-1) of selected lines eval-uated under farmer management at Khorahar,Jharkand, 2001

Selection Selectorenvironment Farmer Breeder Mean

On-farm 243 162 202On-station 566 264 415Mean 405 213

1. F-tests of main effects of selector and selection environment were significant

(Pr>F 0.0002 and 0.0001), respectively

48 Farmers’ participatory breeding for upland rice in eastern India

and breeders were correlated with yield in only a fewcases, as farmers also consider other factors in theirdecision. Though duration and stature had littleassociation with yield, these are important characters invarietal choice. Again, because of the low variability inthe tested material, these were not identified asimportant. Therefore, ranking of varieties must becombined with survey data about farmers’ selectioncriteria to get the actual information (Courtois et al., 2001).

The results of 3 years of on-farm testing indicatedthat the varieties tended to rank similarly across farms.Therefore, the G x EI was not a significant factorinfluencing varietal performance. This indicated that anon-station breeding programme can serve a purpose inthe targeted region. But the precision of on-farm trialswas as high as that of on-station trials, and 3 years ofon-station trials failed to predict the on-farmperformance. This clearly indicates that PVS should beintegrated at an early stage of testing for better predictionof varietal performance.

Farmers’ intentions of voluntarily testing the newvarieties in their own fields indicated that access to thenew varieties or information is the major constraint, asthe local seed system is not very effective. Out of fivevarieties they tested during the 2000 wet season two didnot perform well and were rejected by farmers. At anearly stage farmers may adopt more varieties, but somewill later be dropped, as the farmers test them over anumber of years. Once the variety is adopted by farmersit can be spread from farmer to farmer, because thefarmers from the neighbouring villages also took seedsfrom the adopting village (Joshi et al., 2001).

In PPB, where both farmers and breeders madeselections from the segregating population, farmersselected progenies from fewer crosses. Farmers imme-diately rejected the crosses that did not produce pro-genies according to their criteria. Knowing the back-ground of the crosses, perhaps breeders were expectingdesirable segregants in the later generation.

The significant superiority of farmers’ selection overbreeders’ selection was quite unexpected. It has establi-shed that farmers are able to identify high-yieldingentries (Ceccarelli et al., 2001). The better performanceof on-station selection over on-farm selection indicatedthat selection at the research station can be useful in thetarget region. This may be due to the low, but consistent,yield of the research station. Of course, we need toconfirm this result on a larger scale in the next season.

ConclusionThe PVS programme has given the breeders a systematicway to approach the farmers. The interaction withfarmers and social scientists involved in the projecthelped breeders and farmers develop a better under-standing of the complexity of the problem. As the localseed system is non-functioning, access to new varietiesor information about new technology seemed to bemajor constraints. Proper allocation of resources forvarietal testing is needed, as on-station evaluation alonecannot predict the performance of the cultivar in thetarget region.

The PPB work has given very interesting results.Although it is very early in the testing stage, it appearsthat farmers can manage a segregating population andwere able to select superior progenies from the segre-gating generation. This indicates that it is possible toexpose breeding materials to farmers at an early stage.

ReferencesCeccarelli S, Grando S, Bailey E, Amri A, El-Felah M, Nassif F, Rezgui

S, Yahyaoul A (2001) Farmer participation in barley breeding inSyria, Morocco and Tunisia. Euphytica 122: 3, 521–536

Courtois, B, Bartholome B, Chaudhary D, McLaren G, Misra CH,Mandal NP, Pandey S, Paris T, Piggin C, Prasad K (2001)Comparing farmers’ and breeders’ rankings in varietal selectionfor low-input environments: A case study of rainfed rice in easternIndia. Euphytica 122: 3, 537–550

Joshi KD, Sthapit BR, Witcombe JR (2001) How narrowly adaptedare the products of decentralized breeding? The spread of ricevarieties from a participatory plant breeding programme inNepal. Euphytica 122: 3, 589–597

Siegel S (1956) Non-parametric Statistics for the Behavioral Sciences.Macgraw Hill, New York, USA, 312 pp

Witcombe JR, Joshi A, Sthapit BR (1996) Farmer participatory cropimprovement. I. Varietal selection and breeding methods andtheir impact on biodiversity. Expl Ag 32: 4, 445–460

IntroductionBetween 1992 and 2000 more than 13,000 traditionalrice samples (representing over 3000 different varieties)were collected from Laos. More than 7,300 of thesesamples were upland rice varieties. Using this collection,a system was set up to identify superior varieties for widerelease. While germplasm from breeding programmesis also evaluated, the focus is on the traditional varieties.The rationale behind this is that the Lao prefer glutinousrice and there are relatively few improved glutinousvarieties.

The process of evaluation involves four steps. TheHuay Khot Research Station in Luang Prabang provinceannually receives a new set of 300 accessions from theLao genebank, that are evaluated on-station in anobservational nursery (OBN 1). From this, 20–30% ofthe varieties are selected for a second observation nursery

Participatory varietal selection: Lessons learned from theLao upland programme

K SONGYIKHANGSUTHOR1, GN ATLIN2, S PHENGCHANH1 AND B LINQUIST3

1. Northern Agriculture and Forestry Research Station, Luang Prabang, Lao PDR2. International Rice Research Institute (IRRI), DAPO Box 7777, Makati, Philippines3. International Rice Research Institute (IRRI) Lao Project, PO Box 600, Luang Prabang, Lao PDR

Several thousand traditional upland rice varieties have been collected in Laos. Between 1992 and 2000, approximately 2000of these were evaluated for grain yield and other agronomic characteristics in researcher-managed trials in northern Laos,in an effort to identify high-yielding traditional varieties for dissemination to upland farmers. A few of the tested varietieswere given to farmers to evaluate in collaboration with district extension officers before 2001, but data were returned fromonly 6% of the on-farm trials. In 2001, a new participatory varietal selection (PVS) programme was designed andimplemented to obtain information about farmer preference and on-farm performance of the upland varieties, and toincrease the efficiency of the screening effort. In the first stage, on-farm PVS was conducted with 32 farmers in five provinces.Farmers evaluated eight short- or eight medium-duration varieties selected by researchers in the earlier agronomic testingprogramme. Preference data were successfully obtained from 84% of the farms and yield data from 63%, a 10-fold increasein the on-farm trial success rate over the previous programme. Farmers strongly preferred early-maturing, large-seededvarieties with large panicles and strong stems, and disliked varieties with few tillers or tillers that ripened non-uniformly.These criteria will be incorporated into the initial stages of varietal screening in the future. The correlation between varietymeans over farms for grain yield and preference rating was 0.82 in the medium-maturity trial and 0.54 in the early trial.Occasionally, high-yielding cultivars were not preferred, but the lowest-yielding cultivars were never preferred, indicatingthat agronomic selection for grain yield helps to select varieties farmers prefer. A farmer-preferred, early-maturing variety(Nok) that significantly outyielded the local checks was identified. This and five other lines are being evaluated in a scaled-up PVS programme in 2002.

(OBN 2). A limited number of varieties (8–16) areselected from the OBN 2 for multilocational yield trials(MLYT). Variety evaluation in these first three stages isbased primarily on yield and duration (preference forearly-maturing varieties). Varieties selected from theseMLYTs are then evaluated on-farm. Between 1992 and2000, over 2000 varieties were evaluated through thissystem and 27 were selected for on-farm evaluation. Foron-farm evaluation, varieties (typically 60 g of eachvariety per farmer) were sent to district officers, whoidentified cooperating farmers. The success rate in thisfinal stage was poor. Between 1994 and 2000, 235 farmerswere involved in on-farm evaluation. However, data wereonly collected from 15 farmers (6%). As of 2001, therewere no varieties that could be recommended for theuplands, due mainly to limited data on farmer preferenceand performance under farmer management.

Abstract

50 Participatory varietal selection: Lessons from Lao

In initiating a participatory varietal selection (PVS)programme there were two broad objectives. Firstly, therewas a need to accelerate the evaluation of uplandmaterial. Continuing at the current rate (about 300varieties entering the evaluation process each year), itwill take over 20 years to evaluate all the upland varieties.Identifying and using farmer-preferred characteristics,it is possible to rapidly screen varieties using passportdata (passport data are available for all rice samples heldin the genebank) and in the OBN 1. Secondly, it wasrecognised that changes needed to take place in the finalstage of evaluation (on-farm testing). While the pro-gramme was able to collect agronomic data from the firstthree stages of the evaluation process, the final stage hasproduced no information on farmer preference.

To achieve these objectives, activities in the first yearof PVS involved:● Training researchers and district officers in parti-

cipatory research methods● Evaluating varieties on-farm using participatory

research methods● Identifying characteristics farmers prefer in their rice

varieties.

Materials and methodsOn-farm PVS was conducted with 32 farmers in fiveprovinces: Luang Prabang (10 farmers), Oudomsay(6 farmers), Luang Namtha (4 farmers), Sayaboury(6 farmers), and Xieng Khouang (6 farmers). Seedavailability limited the number of farmers able toevaluate the 16 glutinous upland rice varieties includedin the trials. Twelve varieties were selected from the 2001MLYTs and 4 from previous trials (Vieng, Hom, Damand Makthoua). These varieties were divided into twogroups depending on their duration (Table 1).

For the 2001 PVS trials, participating farmers weregiven one set (early or medium) of eight varieties (110 gper variety). This was enough to sow a 15-m2 area ofeach variety. Farmers were given assistance with plotlayout and sowing but managed the plots according totheir normal practices.

Between flowering and maturity, researchers visitedeach farmer and conducted a preference analysis (PA).In the PA the farmers were asked to rank each varietyusing seeds, placing more seeds on boards for varietiesthey liked. They were then asked to indicate the positiveand negative characteristics of the varieties. Finally, theywere asked which varieties they would continue toevaluate next year on their own. At harvest, the resear-chers and farmers harvested each plot and measuredgrain yield.

AnalysisOf the 32 farmers evaluating varieties, 27 participatedin the PA and yield data were collected from 20. Yielddata from four of the sites where the data was collectedwere not used in the analysis because of poor plot layout.

Individual farmer variety rankings were standardisedacross farmers by expressing the ratings (seeds placedby each farmer on the board) as a proportion of the totalnumber of seeds placed on the board by each farmer.The ratings were further standardised to a mean of 0and SD of +1 for each farmer. Least-squares means wereestimated using PROC MIXED METHOD=REML inSAS, with varieties considered fixed, and farms randomeffects.

Correlations between measured grain yield andfarmer preference ratings were estimated for those farmsfrom which both types of data were collected.

Results and discussion

General

Compared to previous years, the 2001 on-farm trials werevery successful. PA data were collected from 84% of thefarmers and yield data from 63%. This good success canbe at least partially attributed to timely training(particularly in participatory research methodology) andcontinuous follow-up during the course of the season.

In some cases, the PA could not be conducted due toearly crop damage by livestock. Yield data could not becollected from some farmers due to damage by livestock,wild pigs or rats. These are factors that the researcherscan do little about. However, at four sites, PA and yielddata were collected, but due to poor plot layout, the datacould not be used in the analysis. Related to this, someof the trials were in fields where the local variety (usedas the check) had longer growth duration than thosebeing evaluated. This resulted in worse than normal pestdamage. In future, there will need to be training in theseareas, and simpler designs involving fewer varieties willbe used

Preference analysis: preferred plant characteristics

During the PA, farmers were asked to identify reasonswhy they liked or did not like particular varieties. Thesereasons were summarised from all the PAs and are givenin Figure 1. Large panicles, big seed and strong stems

Table 1. Varieties and accession numbers evaluatedduring the 2001 PVS trials

Early-maturing Medium-maturingAccession Accessionnumber Variety name number Variety name

LG 7347 Taa roon LG 7051 Do homLG 6911 Deng said LG 6432 Makhine soungLG 7771 Kou yongke LG 7023 Da cheungLG 6499 Nok LG 7084 PaneurLG Vieng LG 6655 LebmeuLG Hom LG 1724 DamLG 6905 Sang LG 6593 SangonLG 6501 Deng LG 2387 Makthoua

Participatory varietal selection: Lessons from Lao 51

were the most commonly referred to positivecharacteristics. In contrast, farmers did not like varietiesthat produced few tillers, had short panicles or hadpanicles that matured non-uniformly. All the varietiesevaluated were of short and medium duration, therefore,preference for short duration did not stand out promi-nently in this PA. However, early maturation was men-tioned as a positive character and late maturation as anegative character. Earlier research had identified farmerpreference for early- and medium-maturing varieties,and therefore, only such varieties were included in theon-farm evaluation.

Using these criteria, varieties can now be eliminatedearly in the evaluation process, saving time and limitedresources. For example, passport data (including datafor some of these characteristics) exist on most varietiesin the collection. Using this database some varieties couldeasily be eliminated from the evaluation process. Also,in the OBN 1, variety scoring should include thesecharacteristics, to ensure that varieties selected for furthertesting will not be rejected by farmers for reasons thatare already obvious.

Grain yields and farmer variety ranking

In the early-maturing trial the highest grain yield (GY)was 2.14 t ha-1 (Nok). This was significantly higher thanthe average of the local early-maturing varieties whichyielded 1.5 t ha-1. In the medium-maturing trial thehighest yield was 1.94 t ha-1 (Makthoua), similar to themean yields of the local varieties (1.82 t ha-1).

An analysis of data from only those farmers fromwhom both yield and preference data were obtainedshowed significant differences in yield and preference inthe medium-maturity group, but not in the early-maturing group (yield, P=0.06). Yields and farmerpreference ratings (PR) (standardised) were associated.The correlation between variety means over farms forGY and PR was 0.82 in the medium-maturity trial and0.54 in the early trial. Occasionally, high-yieldingcultivars were not preferred, but the lowest-yieldingcultivars were never preferred. This indicates thatagronomic selection for grain yield helps to selectvarieties farmers prefer.

A test of cultivar x province interaction was con-ducted, using the within-province cultivar x farmerinteraction as an error term. There was no detectablecultivar x province interaction for GY. A significantinteraction was observed for PR in the early cultivar set;the cultivar Kouyonke was strongly preferred in LuangPrabang, but had a very low preference score inSayabouli. The reason for this rank reversal is not clear.Certainly, one may expect interactions between cultivarsand different farmers, ethnic groups, or environments.However a larger number of farmers must be surveyedfor such an analysis to be reliable.

What next?Based on farmers’ preferences and GY, Taa roon, Nok,Do hom, Makhine soung, Dam and Makthoua wereselected for further evaluation in 2002 (Table 2). Seed of

Figure 1. Results from the preference analysis (PA) summarizing positive and negative plant characteristics most mentionedby farmers. Twenty-seven farmers participated in the PA. The number of times a characteristic is mentioned can be morethan 27 because farmers were evaluating eight varieties

52 Participatory varietal selection: Lessons from Lao

these varieties are being multiplied during the 2002 dryseason. A Mother and Baby trial design (see Appendix)will be used for the 2002 analysis.

One Mother trial will be placed in each village, whichwill permit many farmers to evaluate varieties side byside. Baby trials will consist of each farmer receiving asingle variety to compare with their own variety. In eachvillage, we anticipate three or more farmers evaluatingeach variety (a total of 18 or more farmers per village).The number of villages depends on the success of theseed multiplication. Yield data will be collected from theMother trial. In the Baby trials, farmers will be asked tocompare the two varieties and to provide informationcollected on growing conditions, soil type, etc. Thisinformation can be gathered during and after the growingseason to provide some comparison of grain quality.

Farmers need to be involved earlier in the selectionprocess. We have already indicated that results from thePA could be used to help screen out varieties early in theevaluation process. However, farmers could also beinvolved in the MLYT. This was done in 2001 but resultsare not yet available. Farmer involvement at this stagerequires a different methodology.

Lessons learned● Compared to earlier on-farm work the success of on-

farm evaluation using participatory methods was verygood.

● In the future, more training emphasis needs to beplaced on good plot layout if a number of varietiesare to be evaluated at a single location.

● Better care needs to be taken in working with farmers,so that the duration of the varieties being testedmatches the duration of the local check variety.

● Too many varieties were evaluated by each farmer. Thisresulted in several problems: 1. farmers could notremember variety names, 2. the PA took too long, and3. it was easy for farmers and researchers to getvarieties mixed up. Based on our experiences, no more

than five varieties should be evaluated by any onefarmer. When evaluating many varieties a simplequestion could be, ‘Which varieties do you plan tofurther evaluate next year?’ Farmers were always quickto identify such varieties. These would naturally bethe ones in which the farmer is most interested

● In future evaluations it would be good to gatheradditional information from the farmer about theenvironment such as, soil type, soil colour, droughtoccurrence (if so, early, middle or late-season). Thiswould allow us to test interactions between cultivarsand environments.

● In future evaluations we need to ask farmers toevaluate post-harvest characteristics of the rice, suchas threshability, milling quality and taste. This was notdone this year due to the limited amount of seed eachfarmer received.

● More seed (at least 0.5 kg and preferably 1 kg) needsto be provided to each farmer for evaluation. This willrequire that a system of seed multiplication is in place.Seed multiplication during the dry season is currentlybeing evaluated.

ConclusionBased on 2002 results, we identified how two broadobjectives of the variety improvement programmes canbe achieved. Firstly, the evaluation process could beaccelerated by taking into account farmer preferencesearly in the evaluation process (either by using passportdata or in screening the OBN 1). Secondly, a system foron-farm evaluation has been identified, that hasproduced encouraging results

Table 2. Average grain yields and ranking in 2001. Varieties in bold are those selected for evaluation in 2002

Early varieties Medium varietiesGrain yield Standardised Grain yield Standardised

Variety (t ha-1) farmer preference1 Variety (t ha-1) farmer preference1

Nok 2.14 0.53 Makthoua 1.94 a 0.46

Hom 2.08 -0.44 Makhine soung 1.75 ab 0.62

Taa roon 1.98 0.60 Do hom 1.72 ab 0.49

Kouyongke 1.88 0.02 Dam 1.55 bc 0.39

Vieng 1.86 -0.16 Peek 1.48 bc -0.70

Deng said 1.67 0.04 Da cheung 1.33 c -0.11

Deng 1.62 0.06 Sangon 1.31 c -0.41

Sang 1.34 -0.65 Lebmeu 1.19 c -0.76

LSD ns ns 0.39 0.83

1. Farmer rank is standardised so that the mean is 0.0

IntroductionThe Rice Varietal ATM Project is a research/extensionstrategy that involves seed testing, evaluation, selectionand multiplication by farmers of upland rice varietiesfrom farmer-managed trials.

The trials were carried out with upland rice farmersin Arakan, Cotabato. It may be noted that prior to theconduct of trials, most of the farmers apply fertilizersand chemical sprays to control pests but yields aregenerally low at about 1 t ha-1. The landholdings weresmall as over 80% of the farmers only had 0.1 – 1.0 ha ofland.

MethodologyThe Rice Varietal ATM project (Figure 1) followed thePublic Education Campaign (PEC) model (Figure 2) thatwas guided by the following concepts: facilitate access toinformation and technology through training, farmerstest (farmers try the technologies on their fields), andmultiply (produce in larger scale or disseminate to otherfarmers for wide use of technology).

The Public Education Campaign (PEC) model

Poor or low utilization of technologies could be due tofarmers’ lack of information or access to seeds and otherinputs. By conducting short courses (e.g., 2- or 3-day

Upland rice varietal access, test and multiplication (ATM)

RFD HONDRADE AND EG HONDRADE

University of Southern Mindanao, Kabacan, North Cotabato, Philippines

Upland rice farmers usually conduct their own on-farm evaluations when presented with a new technology, such as seedsof a new upland rice variety. This strategy was conceived as the Rice Varietal ATM (Access, Test and Multiplication) Project.The project was designed to provide both the hardware and software of selected upland rice production technologies toresource-poor rice farmers. Farmers were provided with small quantities of new, high-yielding upland rice varieties, andother researcher-generated rice production technologies. If the farmers found the material promising, it was multiplied forwider use. All cooperators underwent a one-day training course prior to the start of a farmer-managed trial.

Results revealed that all farmer-cooperators who grew the new rice lines in upland areas succeeded in their seed-testingphase. Usually, successful Rice Varietal ATM trials were conducted by farmers who were prepared to plant upland riceduring the cropping season, who had the required land available, and who were willing to fund and manage the trial. Mostimportantly, the new seed materials usually yielded more than the farmers’ existing or local varieties.

Abstract

Figure 1. Rice varietal ATM methodology

54 Upland rice varietal access, test, and multiplication (ATM)

training) and providing start-up materials (e.g., 1 kg seedkits), farmers get access to information and the hardwareof technology so that they can proceed with their owninformed evaluation. If satisfied with the performanceof the technology on their own farms, it is very likelythat they will multiply the seeds (in case it is a rice variety)for wider use. Chances are that close relatives and friendswho have observed the performance of the technologywill try to access via seed exchange or purchase, therebyfacilitating farmer-to-farmer sharing of technology. Thetransfer of capability or skill from researcher to farmerjustifies the ‘education’ dimension of the model.The model works best under the following conditions:● Committed manpower to analyze farmers needs and

identify appropriate teaching methodologies● Availability of and timely provision of seed kits or the

hardware of the technology● Effective linkage with other partner institutions for a

sustained and effective provision of technical adviceafter PEC

● Continuing liaison with such information/technologysources as research institutions.

1. Breeding network trial

Promising breeding lines of upland rice, collected fromseveral countries, were tested in the breeding networktrial. From this trial, promising lines were identifiedbased on the following criteria:

1. Adaptation to local conditions2. High yield3. Resistance to drought and blast.The presence of the breeding network in their local

municipality allowed farmers to observe the varieties orlines. Some of these achieved yields, that were 50% higherthan their own varieties, and had resistance to pests,diseases and drought.

2. Advanced yield trial

The advanced yield trial (AYT) was composed ofpromising, high-yielding lines from the breedingnetwork trial. The AYT further tested the performanceof these lines. Yield, productive tillers per linear metre,

days to maturity, plant height, and pest and diseaseoccurrence were all measured in the AYT.

3. Farmers’ evaluations of rice lines in the AYT

Farmers were usually asked to evaluate the upland ricelines in the AYT at the trial site during field days. Theholding of field days by the Upland Rice ResearchConsortium (URRC)1 in Arakan became a necessarycomponent of the trials and allowed farmers to closelyobserve and evaluate entries in the AYT using their owncriteria. Farmers were asked to write down the rating oftheir most-preferred varieties and their least-preferredones.They were also asked to identify the criteria theycommonly used in judging a preferred variety.

Results of farmers’ evaluations were usually con-sidered in succeeding trials or activities. Entries thatperformed well in the AYT that were highly favoured byfarmers were given to them for further testing in theirown fields.

4. Farmer-managed varietal assessment

a. Seed Testing Phase: small-area varietal evaluation.The varietal trials were carried out and managed byfarmers. Each participating farmer was assigned atleast two promising varieties (5 kg of seed each) andasked to grow them alongside their local variety inthe same field under the same management. Thefarmers were trained to use a 2-page monitoringinstrument with a calendar, with which they recordedtheir farm practices and general observations on theperformance of the varieties. The monitoringinstrument was based on an earlier survey conductedon the criteria used by upland rice farmers inevaluating a variety. Farmer-to-farmer sharing ofinformation on promising lines occurred in this phase.

Eight cultivars were tested in 22 farmers’ fields.Among the promising and recommended varietiestested were: B 6144, PR 23813-2-53, IR 55423-01, PR23706-26, C-22, UPL Ri-5, PSB Rc-1 and PSB Rc-5.

Only 6 of the 22 farmer-cooperators were able toplant the seed kits given them on their upland farms(Table 1). Each farmer tested at least two upland ricevarieties/promising selections. The rest reported thatthey sowed them in lowland areas.

b. Seed Multiplication Phase: wide area-varietal evalu-ation. If farmer-cooperators are satisfied with theresults in the seed-testing phase, they will likelymultiply the seeds for wider use. Usually, the producefrom the seed-testing phase is used as material for seedmultiplication in a wider area. It is hoped that farmer-to-farmer information dissemination will be strongestduring this phase.

1. The URRC is a collaborative programme of the International Rice Research

Institute (IRRI) and the member country’s research agency involved in

upland rice – in this case the Philippine Rice Research Institute – and the

university, which co-implements the rice trials.

Figure 2. The PEC Model: training/seed kit distribution

Upland rice varietal access, test, and multiplication (ATM) 55

Results and discussion

Selected socio-demographic characteristics of ricefarmer-cooperators

Twenty-two upland rice farmers from Arakan, Cotabato,served as the farmer-cooperators of the project. Most(77%) of them were male members of the Ilongo ethnicgroup (82%), with ages ranging from up to 40 (41%) to41–60 (46%). Generally, the upland rice farmer-cooperators were literate; about 91% of them hadreached or finished elementary and high school. Almosthalf of them (41%) had been growing rice for 21 yearsor more. It should be noted that most (82%) of thesefarmer-cooperators owned a small portion of upland ricefarm, ranging from 0.1 to 1.0 ha. Usually, the farmers’sources of seeds were their neighbours or fellow farmers(77%), USM (36%) and cooperatives (55%).

Rice farming practices during the pre-ATM phase

Fertilizer application. Most (68%) of the upland ricefarmer-cooperators used compound fertilizers on theirfarms. The rate and time of application ranged from 1to 3 bags ha-1 at the tillering stage (47%). Other farmersused urea (36%) and ammonium sulphate (23%) at therate of 1 to 3 bags at tillering and panicle initiation stages.Pest management. Most (91%) of the farmers usedpesticides to control rice pests and diseases. Very few

Table 1. Number of successful trials per variety perlocation

Variety Farmers1 Upland2 Lowlands3

B 6144 10 5 5UPL Ri-5 2 2 –PR 23706-26 1 1 –PR 23813-2-53 7 2 5IR 55423-01 6 2 4C 22 2 2 –PSB Rc-5 1 1 –PSB Rc-1 1 1 –

1. Each farmer received at least two varieties2. Successful under upland conditions3. Varieties sown in lowland areas

(9%) claimed to use such biological control agents asplants with pesticidal properties. The upland rice farmersmanually control weeds usually with the help of familymembers.Yield from their own varieties. Before the interventionof the project to test promising upland varieties, thecooperating farmers grew traditional varieties andlowland rice varieties such as MASIPAG lines and IR36.When these farmer-adopted varieties were compared toeach other, the highest average yield obtained by ricefarmers was from IR36. The average yield of IR36 was1.25 t ha-1 followed by Dinorado (1.00 t ha-1). TheMASIPAG lines yielded an average of 0.78 t ha-1.

Varieties/promising rice lines identified by the projectand performance of recommended cultivars

Four varieties (C 22, UPL Ri-5, PSB Rc-1 and PSB Rc-5)and four promising rice varieties (B 6144, PR 23813-2-53, IR 55423-01, and PR 23706-26) were identified bythe project (Table 2).

Yield performance of some upland rice varieties/selections in farmer-cooperators’ fields.

Based on the data, UPL Ri-5 gave the highest averageyield (Table 3). This was followed by PR 23813-2-53 andIR 55423-01. B 6144 PSB Rc-1, PSB Rc-5 and C 22,yielded less than other test entries.

Table 3. Yield (t ha-1) of upland rice varieties/promising rice lines and farmers’ preferences

Farmers’ preferenceAverage (f = number of farmers reporting)

Number of yield Yield range Variety tested Farmers’ varietyVariety cooperators (t ha-1) (t ha-1) f % f %

C 22 2 2.28 2.14–2.42 1 50 1 50PSB Rc-1 1 1.78 – – – 1 100PSB Rc-5 1 1.63 – 1 100 – –UPL Ri-5 2 3.22 2.85–3.60 2 100 – –B 6144 10 2.44 1.80–3.38 6 60 4 40IR 55423-01 6 2.86 2.63–3.10 5 83.3 1 16.7PR 23813-2-53 7 2.91 2.60–3.23 7 100 – –PR 23706-26 1 2.23 – 1 100 – –

Table 2. Varieties/promising rice lines identified andtested by the Rice Varietal ATM Project

Cultivars tested ReleaseCultivars Released Pre-released year

B 6144 ✔ ✔ –PR 23813-2-53 ✔ ✔ –IR 55423-01 ✔ ✔ –PR 23706-26 ✔ ✔ –C-22 1972UPL Ri-5 1980PSB Rc-1 1990PSB Rc-5 1997

56 Upland rice varietal access, test, and multiplication (ATM)

Farmers’ preferences. Generally, most farmer-cooperators preferred the new seed materials to theirexisting variety. Judging by the limited number ofrespondents for the other rice varieties, a good numberof farmers preferred PR 23813-2-53.Positive and negative qualities of new upland ricevarieties as observed by farmers. The released varietyUPL Ri-5 is high-yielding but is susceptible to bacterialleaf blight, leaf folder and stem borer (Table 4). Two pre-released varieties, B 6144 and IR 55423-01, were observedto be susceptible to lodging. B 6144 lodged at the milkgrain stage. All the other entries showed a high degreeof resistance to lodging. The farmer-cooperators whotested B 6144 claimed that this line was high-tilleringbut was susceptible to narrow brown spot, leaf folderand bacterial leaf blight.

IR 55423-01 and PR 23813-2-53 were observed to besusceptible to bacterial leaf blight and narrow brown spotdiseases.

Table 4. Positive and negative qualities of upland rice varieties as observed by farmers

Variety Positive qualities Negative qualities

UPL Ri-5 High-yielding Susceptible to bacterial leaf blight, leaf folder and stem borer

PSB Rc-1 – Susceptible to narrow brown spot

PSB Rc-5 – –

C-22 – Susceptible to sheath blight

B6144 High-tillering Lodges at milk grain stageSusceptible to narrow brown spot, leaf folder andbacterial leaf blight

IR 55423-01 High-tillering Susceptible to lodgingSusceptible to bacterial leaf blight

PR 23813-2-53 Susceptible to narrow brown spot and bacterial leaf blight

PR 23706-26 Early-maturing –

ConclusionsThe primary objective of the project was to find outwhether selected researcher-generated technologies (e.g.,promising new upland rice varieties) perform well infarmers’ fields under their own management. The RiceVarietal ATM project has demonstrated its value as a wayby which farmers can evaluate a conventional technologypackage (e.g., variety) under their own field conditions.This project also facilitates the provision of feedbackfrom the farm to the experiment station. Such infor-mation can be used in setting research priorities andgoals to serve the varietal needs of the farmer-clientele.

Results of the farmer-managed trials indicated thatfarmers can make better objective evaluations whenexperiments are done by them, since the project providesa forum for rice farmers to integrate their indigenoustechnologies with introduced research-generatedtechnologies.

Although one of the released varieties (UPL Ri-5)performed better than any of the pre-released varieties,this result is not conclusive since has only been testedfor one cropping season.

IntroductionIn Uttar Pradesh rice is grown in 5.5 million ha, of which2.2. million ha (40%) of rice lands are rainfed lowlandand 0.7 million ha are upland (13%). The impact of theGreen Revolution is hardly visible in such areas due toseveral factors:● Poor and inconsistent performance of on-station

developed varieties when grown under farmers’management

● An inability to meet farmers’ preferences for quality● A lack of information about new varieties or poor

access to seeds of them, or lack of both informationand seed.The major production constraints in rainfed rice

include an unpredictable combination of droughts andfloods, build-up of pests and diseases, poor physical soilconditions and socioeconomic limitations. Farmers areresource-poor, which limits their risk-managementstrategies in rice cultivation. Therefore, the developmentof rice varieties suitable for these fragile ecologies willmake a positive impact on the millions of poor peoplewhose livelihoods depend on them.

Addressing genetic improvement and on-farm diversitythrough farmer participatory breeding: A case study ofrainfed rice in Faizabad and Siddharthnagar districts,eastern Uttar Pradesh, India

ON SINGH1, S SINGH1, VN SINGH1, JL DWIVEDI1, A SINGH1, HN SINGH1, RK SINGH2, TR PARIS3, GN ATLIN3

AND VP SINGH3

1. Narendra Deva University of Agriculture and Technology (NDUAT), Kamarganj Faizabad 224 229, Uttar Pradesh, India2. International Rice Research Institute (IRRI) Liaison Office, 1st Floor, CG Block, NASC Complex, Dev Prakash Shastri Marg, Pusa,

New Delhi 110 012, India3. International Rice Research Institute (IRRI), DAPO, Box 7777, Metro Manila, Philippines.a

Narendra Deva University of Agriculture and Techno-logy (NDUAT) in Faizabad, eastern India, in collabora-tion with International Rice Research Institute (IRRI),initiated a farmer participatory breeding (FPB) progra-mme in 1997. In this approach, the needs and prioritiesof farmers were taken into account during the early evalua-tion of rice genotypes, both on-station and in farmers’ fields.

This paper discusses the FPB approach and its impacton on-farm diversity conservation by the participatingfarmers.

Materials and methodsThree typical rainfed sites were chosen – Sariyawan andMungeshpur in Faizabad, and Basalatpur in Siddharth-nagar districts. Basalatpur represented a favourablelowland environment, while Mungeshpur and Sariyawanrepresented shallow but submergence- and drought-prone areas. Three approaches were followed:

Participatory plant breeding (PPB)

This included F4 generations of six different crosses

involving diverse parents. These segregating populationswere jointly assessed by farmers and breeders on-station

Resource-poor farmers in Uttar Pradesh, India, face a number of production constraints to rainfed rice growing, rangingfrom poor performance of traditional varieties and a lack of access to new ones, to unpredictable weather. A farmerparticipatory breeding (FPB) programme that began in 1997 sought to address those problems by developing new ricegenotypes. During the 4 years of the study, several important points were established: Farmers could visually rate for yieldwith high accuracy, and their visual yield ratings were the best predictor of farmer preference. Some cultivars producedvery high yields under low-input management and were highly preferred by farmers. However, some high-yielding varietieswere not preferred, while some farmers preferred varieties that did not give a high yield.

Abstract

58 Addressing genetic improvement and on-farm diversity in Uttar Pradesh

and in farmers’ fields. Two breeders and five farmersselected separately from F

4 to F

6 during the 1997, 1998

and 1999 cropping seasons. Altogether, 50 populationswere developed and tested in three replications on-station and in two replications on-farm. From these, sixlines for Faizabad and five lines for Siddharthnagarconditions – identified by breeders, farmers or both –were selected and compared in field trials. Observationswere recorded on such traits as yield.

Participatory varietal selection (PVS)

This was initiated during the 1997 rainy season andcontinued in the 1998 rainy season. Fifteen to 20 entrieswere tested on-station and in farmers’ fields. The entriesincluded advanced breeding lines from the IRRI-ledShuttle Breeding project, released varieties, and the mostcommon local varieties. Two farmers each from thevillages of Mungeshpur (drought-prone), Sariyawan(drought- and submergence-prone), Faizabad andBasalatpur (submergence-prone), and Siddharthnagarwere chosen for PVS experiments. Final PVS trials wereconducted during 2001, both on-station and on-farmin Faizabad district.

Mother and Baby PVS programme

This was initiated during the wet season of 2001, andconsisted of replicated researcher-managed Mother trialswith 15 genotypes at Masodha Rice Research Station,Faizabad. In these experiments recommended fertilizerdoses were used. Nine farmers in Siddharthanagardistrict conducted the Baby trials in three villages. EachBaby trial included five genotypes and was entirelymanaged by farmers. The alpha-lattice incomplete-blockdesign was used for the layout and analysis of the Babytrials. Simple rating scales (1–5) for different characteri-stics were used for statistical analysis. A simple form wasused to elicit additional information from the farmerson the traits and to obtain suggestions on how to improvethe trials.

Important findings of PPB and PVS

PPB

With few exceptions, the genotypes flowered earlier infarmers’ fields than on-station. However, yields infarmers’ fields were generally lower, as the genotypessuffered prolonged waterlogging, particularly atSiddharthnagar. While a coarse-grained type (IR 700803-43-NDR-2-9335) selected by the farmers performed bestboth on-farm and on-station at Faizabad, the farmersin Siddharthnagar preferred an early type (123-dayduration) that gave a maximum yield of 5.1 t ha-1 on-station and 3.2 t ha-1 in farmers’ fields.

PVS

Ten farmers (five women and five men) visited individualplots and ranked the genotypes grown in different

farmers’ fields. Farmers’ selection criteria were based onecological needs, livelihood uses, gender and social andeconomic background. Mungeshpur and Sariyawanfarmers preferred genotypes with drought tolerance andearly to medium maturity. In contrast, Basalatpurfarmers selected slightly later-maturing genotypes withsubmergence tolerance and long, slender grains.

In general, rice grain yields observed on-farm werelower than on-station. However, some of the cultivarsproduced very high yields (> 4 t ha-1 under low-inputmanagement) and were highly preferred by farmers.

Farmers did select more than one variety (eight inthis case), indicating that PVS/PPB can help maintainvarietal diversity on-farm. Such traits as threshability,submergence tolerance, and grain shape were alsopreferred in this set of cultivars. Straw traits, and pestand disease resistance ratings were not selected byfarmers. Data for grain yield collected from the on-farmtrials were highly repeatable.

Mother and Baby PVS trials

Significant differences were observed for grain yieldamong the genotypes tested in the Mother trials. Farmersperceived significant differences between the cultivarsfor grain yield, submergence tolerance, threshability(particularly women farmers), grain shape and overallacceptability.

Differences in measured grain yield were highlysignificant. The reliability of yield measurements wasmuch higher than expected for on-farm trials. Two ofthe highest-yielding genotypes gave over 4 t ha-1 on-farm,and were highly preferred by farmers over their ownvarieties. However, one high-yielding genotype was notpreferred, but no clear reason for this anomaly could bedetermined from the questionnaire results.

On a single-plot basis, farmers’ visual yield ratingswere highly correlated with measured yield (r = -0.84)(Table 1). Farmers’ yield rating and farmers’ preferencerating were also highly correlated (0.75). Breederspreference rating was highly correlated with measuredyield (-0.81) and farmers’ yield rating (0.94).

On a genotype mean basis, farmers’ yield rating andmeasured yields were highly correlated (0.90). Farmershad a strong preference for cultivars they rated highlyfor yield (r = 0.81) (Table 2).

On-farm diffusion and varietal diversity

While PVS trials continued, farmers picked up a few lineswhich they found promising and grew them on largerplots. For example in Faizabad villages, 23 farmers grewthree genotypes namely NDR 40032, NDRSB 9830102and NDRSB 9730020 in 3.83 hectares during the 2000crop year. The following year, 36 farmers grew the samegenotypes in 4.84 hectares. In Siddharthnagar, thenumber of farmers who used the genotypes such asNDRSB 96005, NDR 40032, NDRSB 9730020, and

Addressing genetic improvement and on-farm diversity in Uttar Pradesh 59

Table 1. Correlations among plot values for measurements and ratings taken by farmers and breeders in Babytrials in three villages in Siddarthnagar, rainy season 2001

Farmer Breeder Farmer Farmer Farmer Farmerpreference preference yield submergence threshability grain shape

Particulars rating rating rating tolerance rating rating rating

Measured grain yield 0.59 0.81 0.84 0.53 0.69 -0.46Farmer preference rating 0.75 0.74 0.48 0.61 0.59Breeder preference rating 0.94 0.49 0.69 0.6Farmer yield rating 0.49 0.74 0.62Farmer submergence

tolerance rating 0.34 0.45Farmer threshability rating 0.33

Table 2. Correlations among cultivar means for measurements and ratings taken by farmers and breeders inBaby trials in three villages in Siddharthnagar, rainy season 2001

Farmer Breeder Farmer Farmer Farmer Farmerpreference preference yield submergence threshability grain shape

Particulars rating rating rating tolerance rating rating rating

Measured grain yield 0.67 0.87 0.90 0.61 0.89 0.57Farmer preference rating 0.80 0.81 0.69 0.77 0.75Breeder preference rating 0.96 0.74 0.83 0.74Farmer yield rating 0.70 0.84 0.75Farmer submergence

tolerance rating 0.70 0.75Farmer threshability rating 0.53

Table 3. On-farm diffusion of genotypes selected through PVS, 2000-2001

Sites (Year) Genotypes Number of farmers Total area (ha) Yield range (t ha-1)

Faizabad (2000)NDR 40032 2 0.16 3.5–5.5NDRSB 9830102 18 2.17 2.8–3.7NDRSB 9730020 3 1.50 3.5–4.8

23 3.83Faizabad (2001)

NDR 40032 4 0.56 3.2–4.1NDRSB 9830102 23 3.10 2.5–3.6NDRSB 9730020 9 1.18 4.2–5.8

36 4.84Siddharthnagar (2000)

NDRSB 96005 6 1.15 2.4–4.2NDRSB 40032 8 0.80 3.5–5.2NDRSB 97300201 12 2.85 3.7–5.6Kamini 6 0.85 1.6–2.5

32 5.65Siddharthanagar (2001)

NDRSB 96005 5 0.90 3.4–4.7NDR 40032 8 1.20 3.6–4.8NDRSB 9730020 19 4.60 3.9–5.6Kamini 11 1.70 1.5–2.5

43 8.40

1. These seeds were disseminated in Nepal by relatives of farmer cooperators. In Siddarthnagar, other genotypes such as NDRSB 9730004, NDRSB 9830102 and

Sugandha were also adopted by farmers.

60 Addressing genetic improvement and on-farm diversity in Uttar Pradesh

Kamini increased from 32 to 43 from 2000 to 2001. Thearea grown to these genotypes also increased from 5.65to 8.40 hectares during the same period. (Table 3). Thisclearly indicates that farmer participatory breeding willnot only increase the existing diversity of rice varietiesgrown under rainfed conditions but also allows farmersto grow improved varieties which were chosen by them.

Conclusions● Farmers could visually rate for yield with a high degree

of accuracy● Farmer visual yield ratings were the best predictor of

farmer preference● Some preferred varieties were not high-yielding (e.g.,

NDRSB 9830102, that produced 2.5–3.6 t ha-1), andsome high-yielders, such as NDR 40032 (3.2–4.1 t ha-1)and NDRSB 9730020 (4.2–5.8 t ha-1) were notpreferred

● Some cultivars produced very high yields (> 4 t ha-1)under low-input management, and were highlypreferred by farmers

● Over the years of the study, farmers did not continueto grow all the lines that they had selected in previousyears. This was because they took into accountadditional information from various trials and laterobservations (such as sensory evaluation, millingrecovery, and drought incidence), in order to finallydecide whether to select or reject a variety.