irri limited proceedings progress in crop improvement research

The International Rice Research Institute (IRRI) was established in 1960 by the Ford and Rock-efeller Foundations with the help and approval of the Government of the Philippines. Today, IRRI is one of the 15 nonprofit international research centers supported in part by more than 40 donors: members of the Consultative Group on International Agricultural Research (CGIAR – www.cgiar.org), other government funding agencies, foundations, the private sector, and nongovernment organizations.

The responsibility for this publication rests with the International Rice Research Institute.

Copyright International Rice Research Institute 2009

This publication is copyrighted by the International Rice Research Institute (IRRI) and is licensed for use under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License (Unported). Unless otherwise noted, users are free to copy, duplicate, or reproduce, and distribute, display, or transmit any of the articles or portions of the articles, and to make translations, adaptations, or other derivative works under the following conditions:

Attribution: The work must be attributed, but not in any way that suggests endorsement by IRRI or the author(s).

NonCommercial: This work may not be used for commercial purposes.

ShareAlike: If this work is altered, transformed, or built upon, the resulting work must be distributed only under the same or similar license to this one.

l For any reuse or distribution, the license terms of this work must be made clear to others.l Any of the above conditions can be waived if permission is obtained from the copyright

holder.l Nothing in this license impairs or restricts the author’s moral rights.l Fair dealing and other rights are in no way affected by the above.l To view the full text of this license, visit http://creativecommons.org/licenses/by-nc-sa/3.0/

Mailingaddress: IRRI, DAPO Box 7777, Metro Manila, PhilippinesPhone: +63 (2) 580-5600Fax:+63 (2) 580-5699Email:[email protected]:www.irri.org.RiceKnowledgeBank: www.knowledgebank.irri.orgCourieraddress: Suite 1009, Security Bank Center 6776 Ayala Avenue, Makati City, Philippines Tel. +63 (2) 891-1236, 891-1174, 891-1258, 891-1303

Suggestedcitation: Hossain M, Bennett J, Mackill D, Hardy B, editors. 2009. Progress in crop improvement research. Los Baños (Philippines): International Rice Research Institute. 112 p.

ISSN 1607-7776

IRRI Limited Proceedings SeriesThe series allows IRRI scientists and partners to quickly share information with specialized institu-tions and individuals. It consists of proceedings from conferences, meetings, and workshops. To permit rapid publications, the review and editing may not be as rigorous as with formal proceed-ings.

Progress in Crop Improvement Research

M. Hossain, J. Bennett, D. Mackill, and B. Hardy, Editors

L P2009 No. 14

Breeding rice for drought-prone areas of eastern India: accomplishments in the recent past and current scenario iii

Contents

Preface. .. . . .. . . .. . . .. . . .. . . .. . . .. . . .. . . .. . . .. . . .. . . .. . . .. . . .. . . .. v

Drought-prone ecosystems . . .. . . .. . . .. . . .. . . .. . . .. . . .. . . .. . . .. . . .. 1Breeding rice for drought-prone areas of eastern India: accomplishments in. . .. 3 the recent past and current scenario M.N. Shrivastava and S.B. Verulkar

Deepwater/boro ecosystmes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13Breeding rice for flood-prone, deepwater, and boro areas . . . . . . . . . . . . . . . . . 15 of eastern Uttar Pradesh J.L. DwivediBreeding for rainfed lowland, deepwater, and boro land in Bihar, India: . . . . . . . 21 achievements and challenges R. Thakur, N.K.Singh, and J.N. RaiBreeding rice for deepwater and flood-prone areas of Thailand . . . . . . . . . . . . . 28 Wilailak Sommut, Kalaya Kupkanchanakul, Prayote Charoendham, Udompan Promnart, and Suthep Nuchsawasdi

Shallow-flooded (submergence-prone) aman areas of South Asia . . . . . . 43Breeding rice for submergence-prone and aman areas of India . . . . . . . . . . . . . 45 S. Mallik, J. Ahmed, S.K. Bardhan Roy, J.N. Reddy, and G. AtlinBreeding rice for submergence-prone and aman areas of Bangladesh . . . . . . . . 57 M.A. Salam

Tidal wetlands/problem-soil ecosystems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63Rice breeding for the tidal wetlands of Indonesia . . . . . . . . . . . . . . . . . . . . . . . . 65 S. Sulaiman, I. Khairullah, and T. AlihamsyahRice breeding for acid-sulfate soils in Vietnam . . . . . . . . . . . . . . . . . . . . . . . . . . 75 Bui Chi Buu and Nguyen Thi LangBreeding rice for salt-affected areas of India . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 R.K. Singh, B. Mishra, A.M. Ismail, and G.B. Gregorio

Upland ecosystems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91Breeding rice for the sloping uplands of Yunnan . . . . . . . . . . . . . . . . . . . . . . . . . 93 D. Tao, F. Hu, G.N. Atlin, S. Pandey, P. Xu, J. Zhou, J. Li, and X. DengProgress of upland rice breeding in Indonesia since 1991 . . . . . . . . . . . . . . . . . 98 Suwarno, E. Lubis, and B. KustianoBreeding rice for the Indian plateau uplands . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 P.K. Sinha, M.Variar, and N.P. Mandal

iv Shrivastava and Verulkar

Breeding rice for drought-prone areas of eastern India: accomplishments in the recent past and current scenario �

Preface

This limited proceedings on the progress of crop improvement research highlights rice breeding research in various ecosystems in South and Southeast Asia. Specifically, these ecosystems cover drought, deepwater/boro, shallow-flooded (submergence-prone), tidal wetlands/problem soils, and upland areas. The drought-prone area involves rice breeding in eastern India, whereas deepwa-ter/boro highlights rice breeding in eastern Uttar Pradesh and Bihar in India and some areas in Thailand. India and Bangladesh are the cases presented for submergence-prone areas, while Indonesia, Vietnam, and some salt-affected areas of India fall under the tidal wetlands/problem-soils ecosystems. As for upland ecosystems, the focus is on Yunnan’s sloping uplands, rice breeding in the uplands of Indonesia since 1991, and the Indian plateau uplands. We hope that this limited proceedings will contribute significantly to the further-ance of rice breeding in these different ecosystems, providing research updates that will enhance ongoing research and encourage future undertakings in these ecosystems.

�i Shrivastava and Verulkar


Drought-prone ecosystems

� Shrivastava and Verulkar


Rice is the staple food of millions of people and it plays a predominant role in determining their livelihoods. Eastern India, which comprises the states of Assam, Bihar, Jharkhand, Chhattisgarh, Madhya Pradesh, Orissa, eastern Uttar Pradesh, and West Bengal, and smaller states of the northeast region, accounts for 63.3% of the total rice area in the country. About 80% of this area is rainfed and is often exposed to abiotic stresses such as drought, low soil fertility, flood, and stagnant water. Although eastern India accounts for about three-fifths of the total rice area, it produces only 48% of the total rice produced in India. The rainfed rice ecology in eastern India can be broadly classified into upland and lowland. Weed competition, frequent drought, low input application, high incidence of blast and brown spot diseases, and the lack of suitable varieties are the major problems faced in the upland ecosystem. In the lowland ecosystem, especially in the shallow lowland areas, high incidence of insect pests and diseases, weeds, terminal-stage drought, and unavailability of good-quality cultivars are the major constraints. Seedling-stage drought and submergence are major issues under semideep- and deepwater situations. The coastal saline soils are often affected by iron and zinc deficiency, which causes chlorotic and low-tillering plants. The varietal improvement programs in the different eastern Indian states, which started in the first or second decade of the 20th century after the establishment of the Imperial Council of Agricultural Research in 1892 (and reoriented periodically), have led to the release of many improved varieties. Strategies for developing drought-tolerant, high-yielding varieties using conventional and molecular approaches are discussed in this paper.

Breeding rice for drought-prone areas of eastern India: accomplishments in the recent past and current scenarioM.N. Shrivastava and S.B. Verulkar

Rice is the staple food of millions of people, playing a pre-dominant role in determining their economy, survival, migra-tion, social structure, and religious beliefs and rituals. Besides being cultivated in 122 countries, the crop is the lifeline of the three most populous countries of the world—China, In-dia, and Indonesia. Of all the rice commercially produced, 90% is directly consumed as food. A majority of the poor in several Asian countries spend as much as half of their income on rice. They depend on rice for around two-thirds of their calories and some 60% of their protein. Hence, any technology that improves the productivity and profitability of rice and reduces risk under adverse growing conditions can make a big difference in the lives of millions of poor people all over Asia. Water is the basic requirement of life. In agriculture, availability of water is a major concern for farmers around the world and drought stress has been identified as one of the most pervasive threats posed to agriculture by the environment. Drought is a major abiotic stress that limits plant growth and productivity and is a major cause of yield instability. With limited irrigation, eastern India is highly exposed to the vagaries of the monsoon. Eastern India comprises the states of Assam, Bihar, Jharkhand, Chhattisgarh, Madhya Pradesh, Orissa, eastern Uttar Pradesh, and West Bengal, and smaller states of the northeast region. Here, rice is grown in the basins of the Ganga, Yamuna, and Mahanadi rivers (and their tribu-taries); this area has the highest intensity of rice cultivation in the country. Eastern India accounts for 63.3% of the total

rice cropped area in the country but produces only 48% of the total rice. About 80% of the rice area of eastern India is rain-fed and exposed to abiotic stresses such as drought, low soil fertility, flood, and stagnant water (Singh and Singh 2000). Even though the region receives good rainfall, yield losses caused by drought every year reach 2.9 million t annually (Widawsky and O’Toole 1990). The yield losses caused by drought at various stages combined are highest among all the biotic and abiotic stresses and about double those caused by weeds (Widawsky and O’Toole 1990). Some of the poorest rice farmers live in this region. However, this area has great potential—productivity can go up significantly. The yield potential of rice cultivars under favorable environments appears to have reached its respective plateaus for more than 20 years, regardless of tremendous breeding efforts. Over the next 50 years, the world population is pro-jected to increase by some 3 billion, mainly in developing countries. The projected rice demand and targets of rice yield in the eastern region are shown in Figure 1. The percentage of area under rice in India has remained static: in 1950-51, it was 23.6 million ha; it was 23.3 million ha in 2000-01. Overall, rice productivity is showing signs of decline, rice area expansion is limited, investments in irriga-tion have virtually ceased, increased fertilizer use threatens the environment, and good rice land is being diverted to other purposes. The only option left is to increase yields of existing unfavorable lands in the years to come. This challenge, associated with the existence and well-being of


millions of poor people in Asia, requires an integrated and systematic approach; a better understanding of type of rice culture, hydrology, climate (rainfall and temperature), soil constraints, biological constraints (weeds, diseases, and insect pests), and socioeconomic factors (Mackill et al 1996); and an appreciation of the importance of the sustainable use and management of natural resources.

Rainfed rice production ecosystems

The rice crop is grown in a complex set of socio-physical and biological environments that determine the performance and adaptation of a variety. It necessitates enhanced efforts toward research prioritization and technology development, delivery, and impact assessment activities. Before defining any strategy aimed at changing a genotype for a particular environment, breeders must be acquainted with and should understand fully the target environment (Fischer et al 2003) in terms of type of rice culture, soil hydrology, rainfall pattern and temperature, soil type, predominant pest complexes, as well as social structure. A broad classification of rice envi-ronments in different eastern Indian states is given in Table 1 (see also Fig. 2).

Table �. Rice area (million ha) under different ecosystems in eastern India.

State Irrigated UplandRainfed lowland

0−30 cm

30−50 cm

Deep water (50−100

cm)

Very deep water

(>100 cm)Total

Assam 0.2 0.2 0.9 0.5 0.4 0.1 2.3

Bihar 1.5 0.5 1.7 0.5 0.4 0.7 5.3

Orissa 1.1 0.7 1.7 0.5 0.4 0.1 4.5

Chhattisgarh 0.6 0.7 2.2 0.1 – – 3.6

Uttar Pradesh 1.0 0.7 1.9 0.3 0.2 0.5 4.6

West Bengal 1.3 0.9 1.7 0.5 0.4 0.7 5.5

Total 5.7 3.7 10.1 2.4 1.8 2.1 25.8

Source: Singh and Singh (2000).

0

10

20

30

40

50

60

1996 2000 2005 2010 2015

Year

Demand (million tons )Y ield targets (q ha–1)

Fig. �. Projected rice demand and targets of rice yield in eastern India. Source: Singh and Singh (�000).

The rainfed rice ecosystem is usually classified into upland and lowland. Various constraints contribute to lower the yield. These include the following: (1) moisture stress due to erratic and often inadequate rainfall, high runoff, poor soil, and lack of facilities for rainwater and soil moisture conservation; (2) flash floods and submergence due to poor drainage, low-lying physiography, and high rainfall in sub-mergence-prone lowlands (as in Assam, West Bengal, and Bihar); (3) accumulation of toxic decomposition and iron toxicity; (4) continuous use of traditional varieties due to nonavailability of improved genotypes and lack of farmers’ awareness; (5) low soil fertility due to soil erosion leading to nutrient stress and imbalance in the use of fertilizers; and (6) heavy infestation by weeds and insect pests and lack of knowledge of control measures. Poor crop establishment (due to several factors) and poor adoption of improved crop production technologies (due to inappropriate technology and economic condition of farmers) (Singh and Singh 2000) further aggravate the situation.

UplandRainfed upland rice is characterized by the absence of stand-ing water in fields a few hours after the rain stops. The total area under rainfed upland in the country is about 6 million ha, which accounts for 13.5% of the total area under rice cultivation. Most rainfed upland areas are found in eastern India. The productivity of upland rice is very poor (0.9 t ha–1). Weed competition, frequent drought, low input application, high incidence of diseases (such as blast and brown spot), and lack of suitable varieties are the major constraints in this ecosystem.

LowlandThe lowland areas in the eastern region of the country are located in Assam, West Bengal, Bihar, Orissa, Chhattisgarh, and eastern Uttar Pradesh. The total area under this ecosystem is estimated to be around 14.4 million ha, which accounts for 32.4% of the country’s total area under rice. The aver-age productivity is low and ranges from 1.0 to 1.2 t ha–1.


The lowland can be further classified into shallow water, semideep water, and deep water. The major constraints to higher yield under shallow water include high incidence of insect pests and diseases, weeds, terminal-stage drought, and unavailability of a wider choice of cultivars with good quality. Seedling drought and submergence are major issues under semideepwater and deepwater situations.

Coastal saline areaIn eastern India, the coastal saline area is situated in West Bengal and Orissa. Yield is about 1 t ha–1. The coastal saline soils are often deficient in iron and zinc, which results in chlorosis and reduced tillering.

HillsThese areas lie in the hilly regions of Chhattisgarh. The total area is around 5% of the total cropped area. The productivity of this region is very low (<1.0 t ha–1). The major problems encountered are cold injury, blast, frequent drought spells, and a very short cropping season.

A varietal improvement program

l Firstphase.The varietal improvement program in different eastern Indian states started in the first or second decade of the 20th century after the estab-lishment of the Imperial Council of Agricultural Research in 1892. Its focus has been reoriented periodically. The approach during this period was to identify better purelines among similar types with the same name within each major group under cul-tivation. The pureline selection method was mostly followed, which resulted in the release of cultivars such as N22, HB10, T141, T1242, Safri 17, Jhona 349, Basmati 370, and others in different states. Also, salt-tolerant lines such as SR26B and flood-tolerant varieties such as FR13A were developed. During the second quarter of the 20th century, the

0

0.5

1.0

1.5

2.0

2.5

A ssam Bihar Oris sa Chhattis garh UttarPradesh

Wes tBengal

States

Area (million ha)

Drought-proneDrought- and submergence-proneFavorableSubmergence-prone

Fig. �. Rice areas (million ha) under different ecosystems in eastern India.

emphasis shifted from selecting within types to selecting between types, as well as to combining characters through hybridization. Attention was also diverted toward developing varieties with better grain quality, including aroma. Improving locally cultivated landraces through pureline selection, im-proving grain yield potential through hybridization, and evolving quality rice continued to be the major themes during the early period of the third quarter of the past century (1951-66) as well. The breeding objectives then were (1) earliness, (2) deepwater and flood tolerance, (3) lodging resistance, (4) drought tolerance, (5) nonshattering grains, and (6) higher response to heavy manuring. Pureline selection was mainly used for genetic improvement, which resulted in the development of 394 varieties. Breeders also tried to create new variability through hybridization, followed by the pedigree method of selection. This resulted in 51 varieties. All these im-proved traditional varieties were high-yielding under low inputs, were resistant to prevailing insects and diseases, and were bred before the exploitation of the dwarfing gene, which resulted in very high-yielding, fertilizer-responsive cultivars, thereby changing the rice breeding scenario.

l Secondphase.A new phase in rice breeding started with the introduction of the semidwarf plant type in 1965. The establishment of the International Rice Research Institute (IRRI) in the Philippines and the coming into existence of the All India Coordinated Rice Improvement Project ushered in a new era in varietal improvement. Between 1966 and 1975, several dwarf varieties came from IRRI and central institutes such as the Central Rice Research Institute, Cuttack; Indian Agricultural Research Institute, New Delhi; Directorate of Rice Research, Hyderabad; and various agricultural universities made recom-mendations at different stages. The more important


varieties were Taichung Native 1, IR8, Jaya, Bala, Cauvery, Sona, Ratna, IR20, IR28, IR36, Phalguna, Surekha, and Sarjoo 49. A number of donors were identified and varieties were bred for resistance to diseases such as blast, bacterial leaf blight, and tungro virus and insects such as gall midge, brown planthopper (BPH), and whitebacked planthop-per. Resistance breeding received rice breeders’ considerable attention. It also became a priority as various biotic factors were seriously limiting yield, particularly where no other control measures were available. It is important to understand the impact of these varieties. As an example, in Chhattisgarh, much emphasis was given to incorporating the genes for gall midge resistance into improved varieties. The extensive cultivation of these varieties has substantially reduced losses due to this pest. The other pest, BPH, was not important initially and not much effort was made to control this menace. The result was that this insect gained prominence as the most damaging pest in rice fields. A list of varieties that evolved prior to 1991 and that remain popu-lar is given in Table 2. Adoption of high-yielding dwarf varieties was 96.5% and 96%, respectively, in Tamil Nadu and Punjab, followed by Karnataka and Andhra Pradesh (Table 3).

Most of the varieties, which became popular under the drought-prone rainfed ecology in the eastern states, particularly Madhya Pradesh, Chhattisgarh, Uttar Pradesh, and Orissa, were spillovers from work done for the irrigated

Table �. Area covered by dwarf variet-ies in some states of India, �000-0�.

Zone State Percentage

Eastern states Assam 56.0Bihar 62.5CG 69.1Orissa 72.6

Southern states AP 77.9Karnataka 87.0TN 96.5

Northern states Haryana 64.7Punjab 96.0

Table �. Varieties developed before �99� that are still popular with farmers in eastern Indian states.

State Varieties

Chhattisgarh Kranti, Aditya, Tulsi, K3, IR36Madhya Pradesh Poorva, Anupama, Jawahar 75Uttar Pradesh Narendra-1,2,97,118Orissa Rudra, Subahdra, Annanda, Kalinga 3Jharkhand Vandana, Brisa Dhan 103, 201Bihar Rajshree, Jaishree

ecology as the local parents used in hybridization quite often possessed a good amount of tolerance for drought (Table 4). Breeding materials, however, were rarely subjected to planned drought situations. The mechanisms and traits as-sociated with drought tolerance and recovery were poorly understood. Varieties such as IR36 played an important role in bringing drought-prone area under HYVs. Nevertheless, semidwarf cultivars spread considerably in rainfed drought-prone environments. An excellent example is Chhattisgarh, where semidwarf varieties occupied as high as 90% of the cultivated area in plain regions, although more than 80% of the area was rainfed (Fig. 3). Most of the changes came only in the late 1980s or early 1990s, when separate trials were constituted for rainfed conditions at the all-India level by directorates of rice research under the All India Coordinated Rice Improvement Project.

Current efforts

Mackill (1986) expressed the view that rice breeders were well aware of the importance of drought as a major constraint to yield and its stability. However, only a few crosses or selections were made because of the lack of knowledge and confidence in drought screening and selection protocols. Breeding for drought tolerance was simply not a part of their mainstream rice improvement program. The irrigated environment usually requires fewer genotypes, resulting in maximum phenotypic expression. In contrast, the rainfed lowlands have a range of environments and involve so much diversity (moisture regime, fertility, flooding incidence, biotic factors, daylength, interaction among each other, interaction with genotypes, etc.), resulting in complex phe-notypic expression and requiring a range of genotypes for maximum phenotypic expression (Fischer 1996). It is critical that the factors contributing to low yield under the rainfed ecosystem be understood, particularly in the eastern Indian context. These include low crop yield at original sites, co-

Table �. Number of rice varieties released for different ecosystems in India, �9��-�000.

EcosystemVarieties released (no.)

Up to 2000 Until 2005

Irrigated 314 422

Upland 84 283

Lowland 123

Semideep water 30

Deep water 14

Saline alkaline 15

High altitude 33

Scented fine grained 19

Total 632 705

Source: DRR (2000), AICRIP Annual Workshop (2005).


evolution of pests leading to a larger number of pests, wide topographical/ecological variation within a small geographic area, tropical/subtropical areas with lower yield potential, a narrow genetic base, inefficiency of selection (more so in unfavorable conditions), abiotic stress undefined in time and space, and a few but vital biotic stresses. The other problems are inadequate funding for breeding programs and training of breeders for specific environments.

Occurrence of drought and a breeding strategy

Availability of appropriate moisture is necessary to obtain optimum yield from any crop plant. Under the rainfed rice ecosystem of eastern India, inadequate moisture, which is linked to erratic rainfall during crop growth, is a crucial fac-tor that affects overall yield. Although mean rainfall is quite good, seasonal distribution and year-to-year variation are high, as the rains may cease for any number of days at any growth stage from germination to maturity. However, the rainfall pattern in the past 12 years at Raipur indicates that terminal-stage drought is very frequent and medium-duration varieties are prone to terminal drought almost every second year. Developing very early duration varieties is not a good option as these are prone to damage by rain at maturity and a yield penalty is incurred due to early maturity. Therefore, developing drought-tolerant, 110–115-d-maturing, high-yielding, and stable genotypes is one of the major objectives of rice breeding. Terminal-stage drought can be managed to a good extent by recommending varieties of appropriate duration in the context of the water-holding capacity of the soil. Based on this, breeders have developed and recommended many varieties in different states (Appendix 1). These varieties have played a big role in improving yield in the eastern states and bringing self-sufficiency on the food front. The situa-tion, however, is likely to worsen in view of ever-increasing population, current rice and food crises, and a slowdown in productivity increases. Concentrated efforts will have to

0

10

20

30

40

50

60

70

80

1965 1970 1975 1980 1985 1990 1995 2000 2004

Year

Area (% of total rice area)

Fig. �. Approximate area under modern (dwarf) varieties in Chhattisgarh.

be made to get higher yields from the same or reduced area planted to the crop and perhaps with a lesser quantity of avail-able water. This requires evolving a well-planned strategy to meet the challenges ahead.

Components of the strategy

The following represent key components of a breeding strategy: l Recommending and popularizing early-duration

varieties to enable farmers to cope with terminal-stage drought.

l Using more local germplasm: The exploitation of a narrow genetic base in breeding programs has resulted in reduced gain in improvement. Plants have evolved several mechanisms to combat drought damage. It is believed that local landraces and wild species still have a good number of untapped genes. Thus, it is important to explore new donors that exhibit stable performance under water stress and have good yield potential. The different mechanisms for drought resistance can bring in potential new genes/QTLs into the breeding pool.

l Identifying morphological and physiological traits associated with drought tolerance: Considering the complexity of the rainfed ecosystem, it is imperative to understand the myriad physiological interactions and reduce that enormous complexity to a mana-gable level. There is a need to give information on how it might be manipulated genetically to achieve practical results (Fischer 1996). Work on this aspect is in progress at different places. The work at IGAU has resulted in understanding interactions—for ex-ample, delay in flowering under drought conditions was related to low water status and was an indicator of drought susceptibility; delay in flowering was also associated with higher spikelet sterility. Some geno-types increased root growth at reproductive-stage


drought and yield stability was attained by maintain-ing higher plant water status. This maintenance of higher water status under water-limiting conditions was a key to achieving drought tolerance.

l Standardization of drought-screening protocols and development of rapid screening protocols: The prob-lem of within-field heterogeneity is a major one as it reduces selection efficiency and heritability of traits under stress. The well-managed screening protocol in the field needs to be standardized to increase the heritability of drought-related traits and repeatability of results. Agronomic adjustments such as sowing and transplanting may be made to increase the prob-ability of exposing test material to drought.

l Proper selection of field: When water is drained from a field, the water level becomes uneven; this imposes heterogeneity in the field and genotype performance becomes unpredictable. Physical (recording the depth of freely available water in soil), biological (resistant and susceptible checks), and statistical tools can be used to overcome the problem of within-field heterogeneity. There is also an urgent need to develop rapid screening techniques for large-scale screening of segregating populations.

l Incorporating resistance to major diseases: The drought-prone environments are quite often as-sociated with incidence of particular diseases and insects. Blast and brown spot are two such diseases. It is necessary to incorporate at least a good level of resistance to blast as this is controlled by major genes.

l Under rainfed conditions, weeds are a major yield constraint. The genotypes need to have early vigor to compete with weak growth.

l Water-use efficiency and stay-green traits: These two traits have not yet received adequate attention. Their role in raising the yield potential of genotypes under drought situations needs to be understood and used to the extent possible.

l Exploiting molecular tools: Current advances in molecular biology have raised the hopes of making a breakthrough in improving tolerance of plants for biotic and abiotic stresses. Identifying major as well as minor genes (quantitative trait loci [QTLs]) and/or candidate genes has become possible. These technologies may lead to some basic understanding as well as an increase in selection efficiency, opening up possibilities of developing transgenic lines with increased stress tolerance.

l Developing local mapping populations can be an important approach to identifying genes that may impart adaptability to local adverse conditions re-lated to the environment or the soil.

l Following appropriate breeding methodologies: Adoption of appropriate breeding methods is es-sential for combining genes of interest. Therefore,

the pedigree method is most widely used in most situations, and the backcross method for incorporat-ing QTLs is now increasingly used. Since drought situations are quite varied and location-specific, test-ing of material in farmers’ fields is now being done. Involvement of farmers at various stages of breeding would ensure better adoption of lines when these are finally released. Multilocation and multiyear testing (though a regular component of all breeding programs) is much more pertinent in the case of developing varieties for rainfed environments than for irrigated areas.

l The use of hybrids in less favorable rainfed environ-ments: Because of their genetic plasticity and faster growth, hybrids can yield better under adverse con-ditions (Table 5). Their adaptation under such con-ditions is to be encouraged. Considerable progress has been noted recently in some areas, particularly in Chhattisgarh.

Most of the abovementioned approaches are known and are being used by breeders, but, now that productivity in favorable irrigated areas has already reached a plateau, no stone is left unturned to improve yields in rainfed environ-ments. It is heartening to note that most breeders are adopting more and more systematic approaches and modern tools to achieve their targets.

References

DRR (Directorate of Rice Research). 2000. Annual progress report.Fischer KS. 1996. Improving cereals for the variable rainfed system:

from understanding to manipulation. In: Singh VP et al, editors. Physiology of stress tolerance in rice. Proceedings of the Inter-national Conference on Stress Physiology of Rice, 28 Feb.–5 March 1994, Lucknow, Uttar Pradesh, India. p 1-10.

Fischer KS, Fukai S, Lafitte R, McLaren G. 2003. Know your target environment. In: Fischer KS, Fukai S, Lafitte R, Atlin G, Hardy B, editors. Breeding rice for drought-prone environments. Los Baños (Philippines): International Rice Research Institute. p 5-11.

Mackill D. 1986. Varietal improvement for rainfed lowland rice in South and Southeast Asia: results of a survey. In:Progress in lowland rice. Los Baños (Philippines): International Rice Re-search Institute. p 115-144.

Table �. Promising hybrids identified for the rainfed ecology.

Hybrid Yield (t ha–1)

Yield advantage (%) over check

KRH-2 3.7 31PSD-1 3.6 28PHB-71 3.6 27CORH-2 3.6 26PA 6201 3.6 25Best check 2.8 –

Breeding rice for drought-prone areas of eastern India: accomplishments in the recent past and current scenario 9

Mackill DJ, Coffman WR, Garrity DP. 1996. Rainfed lowland rice improvement. Los Baños (Philippines): International Rice Research Institute. 242 p.

Singh,VP, Singh RK, editors. 2000. Rainfed rice: a source book for best practices and strategies in eastern India. Los Baños (Philippines): International Rice Research Institute. 292 p.

Widawsky DA, O’Toole JC. 1990. Prioritizing the rice biotechnology research agenda for eastern India. New York: The Rockefeller Foundation.

Notes

Authors’ address: Indira Gandhi Agricultural University, Raipur 492006, India.

Appendix 1

Rice varieties developed/released in Chhattisgarh (�99� onward)

Variety Parentage Suitabilitya Duration (d) Yield (t ha–1) Grain typebFeaturesc

(resistance to or tolerance of)

Year released

IR64 IR5657-32-2-1/IR2061-465-1-5-5

RF, Irr 112 4.0−4.5 LS Bl, BLB 1992

Shyamala R60-2712/R 2389 RF, Sh. LL 130 4.5 LS – 1993

Mahamaya Asha/Kranti RF, LL 125 5.5−6.0 LB GM, BLB, drought 1995

Poornima Poorva/IR8608 RF, DS 110 3.5−4.0 LS GM, BLB, drought 1996

Danteshwari Samridhi/IR8608 RF UpL & Sh. LL 105 3.5−4.0 LS GM, drought 2001

Rice varieties developed/released in Madhya Pradesh (�99� onward)

Variety Parentage Suitability Duration (d) Yield (t ha–1) Grain type Features Year released

JR3-45 Lodhi/IR36 RF, DS 100 2.5–3.0 SB, red ker

Drought 1997

JR353 IR36/JR75 RF, DS 110 3.0–3.5 MS LS, Dr 1998

JR201 – RF, DS 105 2.5–3.0 LS GM, BLB 2001

Richa – RF, DS 100–105 LS MT 2004

Rice varieties developed/released in Jharkhand (�99� onward)


Birsadhan 103

Fine Gora/IET 2832

RF Sh. LL 95–110 2.2–3.0 LB BM, SB, Bl, BLB, BS 1992

Vandana C 22/Kalakeri Upland 90–95 3.0–3.5 LB SB, termite, Bl, BS 1992

Hazaridhan IR42/IR5853-118-5

RF Sh. LL 115–120 4.0–4.5 LS SB,WBPH, Gundhi, Bl, BLB, BS, drought

2003

Sadabahar BRRI SAIL/ IR10181-

58-3-1

RF Sh.LL 105–110 3.5–4.0 LB SB, Bl, BS, BLB, drought

2003

aRF = rainfed, Irr. = irrigated, LL = lowland, Sh. LL = shallow lowland, UpL = upland, DS = dry season. bLS = long slender, LB = long bold, SB = short bold, MS = medium slender, LM = long medium, MB = medium bold. cBl = blast, BLB = bacterial blight, GM = gall midge, SB = sheath blight, WBPH = whitebacked planthopper, RTV = rice tungro virus, LF = leaf folder, SR = sheath rot.

�0 Shrivastava and Verulkar

Varieties released in Bihar (�99� onward)


Kamini Katarni RF LL – – MS BPH. BLB, drought 1992Shakuntala Pankaj/BR 8 RF UpL 145–150 3.0–4.5 LM BPH, BLB, Bl 1994Turant Dhan Rasi/Sattari Timely or

late sown 70–75 3.0–4.0 SB SB, Bl, drought 1994

Vandehi Beldar Drought-prone LL

– 3.0–4.5 LB Bl, BLB, RTV, drought

1995

Prabhat IR2033-521-1/IR261-264-2//IR3

Early and late sow-ing

95–100 3.5–4.5 MS SB, other major insect pests, BLB, BS

1997

Satyam RD19/Desaria 8

RF LL 145–150 3.0–6.0 LS BPH, WBPH, LF, BLB, Bl, BS, BLB, Bl

1997

Kishori IR8/Barogar RF LL 140–145 6.0–7.0 LB LF, BPH BLB, BS 1998Richharia Pusa 33/

IET7464Upland 100–105 3.5–6.0 Long fine SB,BLB, BS 2000

Santosh Pankaj/BR34 Intermedi-ate land

145–150 3.5–5.0 LS WM, Gundhi bug, BLB, BS, SR

2001

Saroj Gautam/ Type 3

Fav. upland, medium land

115–120 4.5–5.0 LS SGP 2001

Varieties released/developed in West Bengal (�99� onward)


Bhupen C22/IR26 and C22/054

Drought- prone areas

110–114 2.0–2.4 LS SB, drought 1993

Jamini BG 280-12/PTB 33

Drought-prone laterite

95–110 3.2 LB Bl, BLB, ShR, ShB 1996

Khanika Jaya/CR 237-1 Drought-prone DS

75–85 3–4 LS BPH, BS, ShB 1996

Shantabdi CR 10-114/CR 10-115

Drought-prone DS

112–145 3.5–4 LS ShB 2000

Varieties release/developed in Eastern UP (�99� onward)


Narendra 7 N 22/Ratna Drought-prone

85–90 4.5–5.0 MS Bl 1992

Barani Deep

(-1064-5?IR 9129-320-3-3-3/IR54)

Drought-prone

–

95–100

–

4.5–5.0 MB Drought 2001

Breeding rice for drought-prone areas of eastern India: accomplishments in the recent past and current scenario ��

Varieties release/developed in Assam (�99� onward)


Moniram Pankaj/Mah-suri

Sh. LL 155 5–5.5 MS BLB, Bl 1992

Ketekijoha (Badsabhog /Savitri)/ Badsabhog

Sh. LL 160 3–3.5 SB GM, SB, BLB, ShB 1994

Satyarajan IET9711 /IET11162

Medium lands

130 – MB GM (biotypes 1&2), BPH, WBPH, SB, LF, Bl.

1996

Rice varieties developed/released in Orissa (�99� onward)


Badami Suphala/An-napurna

Uplands/bialy lands

90–95 3.5 MB GLH, SB, BPH, BLB, BS

1992

Ghan-teswari

IR2061-628/N 22

Irr, RF UpL 90–95 3.3 MB SB, GM, BLB, SR, ShB, RTV

1992

Khandagiri Parijat/IR13429-94-3-2-2

Irr, RF UpL 90–95 3.2 MS BPH, GM, Bl, SR, ShB, BLB, RTV

1992

Nilagiri Suphala/DZ 192

Medium lands, LS

90–95 3.0 MB SB, GM, BPH BLB, SR

1992

Gajapati OR 136-3/IR13429-196-1-20

Medium lands

125 4.4 MS GM, SB, BPH, WBPH, Bl, BS, ShR, BLB, RTV

1999

Kharavela Daya/IR13240-108-2-2-3

Medium lands

125 4.4 MS SB, GM, BPH, WBPH, Bl, RTV, ShR, ShB, BLB

1999

Konark Lalat/OR 135-3-4

Medium lands

125 4.5 LS GM, SB, BPH, Bl, BS, ShR, BLB, RTV

1999

Lalitagiri Badami/IR19661-364

Upland 95 3.2 MB SB, GM, WBPH, BS, ShR, Bl, Dr

1999

Udayagiri Savitri/IRAT 138/IR 13543-66

Upland 95 3.5 MB SB, BPH, WBPH, Bl, RTV, ShB, ShR, BLB, Dr

1999

Bhanja IR36//Hema/Vikram

RF, Irr me-dium lands, Sh. LL

135–140 4.0 MB GM,SB, BLB, BS, Bl, ShR, lodging, shattering

1992

Birupa ADT 27/IR 8//Annapurna

RF, Irr me-dium lands

130–135 4.0 Coar. SB, GM, cutworm, WBPH, ShB, SR, BLB, Bl, RTV

1992

Mahalaxmi Pankaj/Mah-suri

Irr, RF Sh. LL 145–155 4.0 MB BPH, ShB, FSm, LF, ShB, BLB

1992

Manika CR 1010/OBS 677

RF, Irr LL, LS 155–160 4.5 MB SB, GM, BPH, ShB, RTV,BLB

1992

Meher (OBS-677/IR2071)/(Vikram/W 1263)


135–140 4.0–6.5 MB GM, SB, BPH, WBPH, BLB, SR

1992

Samanta T90/IR8//Vi-kram///Siam 29/Mahsuri


140 4.5 MB GM (biotypes 1&4), SB, WBPH, BLB, BS, Bl, SR, ShB

1992

�� Shrivastava and Verulkar

Rice varieties developed/released in Orissa (�99� onward)


Santep Heap 3

Pankaj/Sigadis RF, Irr Sh. LL

150–155 4.5 MB GM, BLB, SR, ShB 1992

Urbashi Rajeswari/Ja-jati

RF and Irr Sh. LL, LS

145 4.0 MB GM, SB, BLB, SR, ShB, RTV

1992

Bhoi Gouri/RP 825-45-1-3

120 3.9 MB GM,WBPH, BPH, SB, Bl, BS, BLB, ShB, RTV

1999

Indravati IR56/OR 142-99

Sh. LL 150 4.1 MS BPH, WBPH, SB, Bl, BS, ShR, RTV

1999

Mahanadi IR19661-131-1-3/Savitri

Sh. LL 150 4.4 MB SB, WBPH, LF, Bl, BS, ShR, BLB, RTV

1999

Prachi IR9764-45-2-2/CR 149-3-2

Sh. LL 155 4.3 MB BPH, WBPH, BS, BLB, RTV

1999

Ramchandi IR17494-45-2-2/jagan-nath

– 155 4.0–6.5 MB Shallow water and flash floods

1999

Sebati Daya/IR36 Medium lands

– 4.1 MS GM, Bl, BS, BLB, RTV, ShB, ShR

1999

Surendra OR158-5/Rasi Medium lands

135 4.7 GB SB, GM, LF, BPH, WBPH, Bl, BS, SB, ShR, RTV, BLB

1999

Breeding rice for flood-prone, deepwater, and boro areas of eastern Uttar Pradesh 13

Deepwater/boro ecosystems

14 Dwivedi


Flood-prone rice areas in India are mainly located in various eastern states such as Assam, Bihar, Orissa, Uttar Pradesh, and West Bengal. Of the total of 2.3 million hectares of these lands, eastern Uttar Pradesh has 0.39 million ha. These areas are more vulner-able than the other rainfed ecosystems as their yields are low and quite unstable. In eastern Uttar Pradesh, four different flood-prone subecosystems exist: deep, semideep, flash-flood-prone, and off-season deep stagnant. Many types of rice varieties have been selected by farmers over the centuries that meet the varied and harsh conditions of this ecology. These traditional types give stable but low yields. The IRRI-ICAR-NDUAT collaborative project on flood-prone rice established in 1991 resulted in a characterization of the environment followed by strong work on genetic improvement of rice grown in various conditions of this fragile ecology. This paper deals with the accomplishments of this collaboration with IRRI, with special reference to breeding rice for flood-prone, deepwater, and boro areas of eastern Uttar Pradesh. The technologies have significantly helped to increase on-farm productivity, meet household needs, and generate income for resource-poor farmers of the flood-prone areas of eastern Uttar Pradesh.

Breeding rice for flood-prone, deepwater, and boro areas of eastern Uttar PradeshJ.L. Dwivedi

India has approximately 2.4 million hectares of flood-prone rice (FPR), of which 80% are now in eastern India. The extent and distribution of rice areas under different ecosystems in India as well as eastern India are provided in Table 1. Eastern Uttar Pradesh shares about 390,000 ha of FPR area. This area suffers from uncontrolled flooding and drought. Low rice yields in this area, as a rule, have yet to meet the projected needs for the future, and an increase in rice productivity will likely have to come from this fragile ecosystem. Rice farming in eastern Uttar Pradesh is complex, diverse, and risk-prone. Increased production in this area depends on submergence-tolerant varieties and other appropriate technologies. Many types of rice have been selected by farmers over the centuries that meet the varied harsh conditions of this ecosystem. Combining their adaptive mechanisms with high-

yielding traits is the basis for developing improved varieties for this ecosystem. The IRRI-ICAR-NDUAT collaborative project on FPR since 1991 has resulted in the characterization of rice environ-ments, development of nondestructive screening techniques for assessing elongation ability, identification of suitable donors, and development of improved varieties for different subecological FP conditions, besides limited validation of improved germplasm in the target environment.

Characterization of the rice environment

A good understanding of the socioeconomic and biophysical aspects of the environment is crucial to breeding research. The IRRI-ICAR collaborative project helped in the systemat-ic mapping of the extent and distribution of rice under excess water conditions, supported by field observations. In eastern Uttar Pradesh, four different flood-prone subecosystems ex-ist: deep, semideep, flash-flood-prone, and off-season deep stagnant flooding. In the deep subecology, maximum water depth is usually >100 cm and could increase up to 300 cm. The semideepwater situation includes areas with maximum water depth up to 100 cm. The rise of water is slow and the water is not turbid. Flash-flood-prone is the largest area among the four subecosystems mentioned. Flooding can occur up to 9 times per season, with depth ranging from 30 to 120 cm depending on rainfall (Table 2). Damage caused by flash flood can be quite severe. Off-season deep stagnant rice grown in flood-prone depressed areas during the dry season is called boro.

Table 1. Extent and distribution of rice area under different ecosystems in India.

EcosystemTotal area (million ha)

All India Eastern India

Upland 7.0 5.2Rainfed lowland 12.1 9.4Shallow water (0–30 cm) (8.1) (5.9)Intermediate (30–50 cm) (4.0) (3.5)Deep water 5.4 4.6Semideep (50–100 cm) (3.0) (2.5)Deep (> 100 cm) (2.4) (2.1)Irrigated 17.8 7.1

Source: CRRI, Cuttack, India.

16 Dwivedi

Genetic improvement

Collaboration strengthened rice breeding research by devel-oping nondestructive screening techniques for elongation ability and refining the submergence screening test, identi-fying donors for better elongation ability and submergence tolerance, and studying the inheritance of these flood-prone traits during the early stage of collaboration. In addition, substantial progress has been made in this subecology in developing improved germplasm and varieties. Brief accom-plishments related to these aspects are described below. l Screeningtechniques: Efforts were made to develop

nondestructive screening techniques for assess-ing elongation ability in flood-prone varieties as the screening test used for elongation ability was destructive. Elongation is essential for survival in floating rice. Floating rice elongates rapidly under submergence. Accordingly, three nondestructive techniques, GA3 application at the 3–5-leaf stage, as-sessing elongation without flooding, and elongation under shallow water, have been developed. These techniques could be used in assessing segregating populations without losing nonelongating types. They were successfully used to determine the ge-netics of elongation ability (Dwivedi and Senadhira 1996).

l Genetic studies: Dwivedi and Senadhira (1996) found elongation to be controlled by two dominant complementary genes and heritability was also high. Submergence tolerance is necessary for rice grown in medium-deepwater areas. Complete submergence could occur at any time during crop growth. Dwivedi

and Senadhira (1996) studied the genetics of these traits and reported a single major gene with additive effects involved in this trait. IR31142-14-1-1-3-1-1-2, IR31406-333-1, and IR40931-33-1-3-2 were identified as good donors for submergence tolerance. Rice yields are severely lowered by flash flood in addition to stagnant flooding.

Development of varieties

Flood-prone deepwater ecologyTheRice Research Institute of Thailand took the responsibil-ity for developing and distributing improved deepwater (DW) rice in Southeast Asia in 1992. Because of the difference in latitude, breeding materials developed in Thailand were not suitable for most DW areas of India as Thai genotypes were late in India. To solve this problem, selections were made in Thailand and at IRRI. With this improvement in selection, the adaptability and acceptability of germplasm provided by IRRI and Thailand were enhanced. It is believed that this collaboration resulted in improving grain quality traits. To further streamline FPR research in northeastern Thailand, special research and training workshops were held in Thai-land in 1992, 1994, and 1995, in which FPR breeders from this region also participated. Substantial progress was made in this fragile ecology. Since 1993, four improved varieties—Jalnidhi for floating ecology (1993); Jalpriya and Jal Lahari for semideep and shallow deep water, respectively; and Barh Avarodhi (1995) for flash-flood areas—were released. These varieties are spreading in their respective areas.

Table 2. Flood spells at Ghagharaghat during the wet seasons.

Year Water inception date Flood duration (d) Peak water depth (cm)

2000 (9)a 8 June 4 35

22 June 3 40

2 July 3 11

8 July 3 21

12 July 14 36

29 July 10 70

8 Aug 2 30

11 Aug 12 72

30 Aug 15 36

2001 (3) 15 July 6 55

27 July 10 52

21 Aug 3 16

2002 (3) 12 Aug 3 12

22 Aug 5 23

6 Sept 6 36aNumbers in parentheses indicate number of flood spells.


Two of the most promising floating lines, NDGR 426 (IET 11873) and NDGR 427 (IET 11871), have been recom-mended for multiplication of seeds and minikits for the areas where water depth surpasses 100 cm for at least 30 days, in addition to NDGR 433, 444, 445, 448, and 449. Promising lines possess fast elongation ability besides kneeing ability and nodal tillering. NDGR 231, 272, 279, 260, 268, 283, 290, and 269 for semideep water have been identified. These promising lines have moderate elongation. In eastern Uttar Pradesh, damage due to flash flood is quite high. Yields are highly unstable and depend on flooding depth, duration of crop submergence, and stage of the crop when submerged. With increased submergence tolerance, yields of this subecology could be increased. To identify genotypes possessing submergence tolerance, 25 genotypes along with tolerant variety FR13A and susceptible variety IR42 were evaluated following the Standard Evaluation System developed by IRRI. NDGR 83, 84, 85, 86, and 109 showed submergence tolerance on a par with FR13A and Sabita. High survival of 75–80% was recorded in these promising lines. These breeding lines were also assessed for days to 50% flowering, plant height, number of tillers per hill, and panicle length. Some of these breeding lines showed little elongation coupled with high survival, indicating they are true submergence-tolerant types (Table 3).

Off-season deep stagnant/boro riceBecause of submergence coupled with waterlogging in the rice season, the flood-prone environment is subject to a high risk of crop loss. Off-season deep stagnant boro rice grown in the dry season is one promising option for such deeply flooded areas. About 50,000 ha of potential boro area exist in the state. However, at present, barely 4,000–5,000 ha of land are under boro rice. Most of the improved germplasm grown in the boro season is cold-sensitive. No systematic breeding program was undertaken in the past to improve boro varieties although there is immense potential to increase the rice productivity of such lands. Accordingly, efforts were made in 2001 and 2003 to identify the potential boro rice-growing areas in eastern Uttar Pradesh with the aim of evaluating the available technology and improved rice germplasm developed in other states through a scientist-farmer participatory approach (Dwivedi 1996). High-yielding cold-tolerant varieties developed at RAU, Pusa, Bihar, and Vivekanand Parvatiya Krishi Anu-sandhan Sansthan, Almora, Uttaranchal, were evaluated at the Crop Research Station, Ghagharaghat, Bahraich, U.P., during 2000. Three boro rice varieties, Richharia, Prabhat, and Dhan Laxmi, from RAU, Pusa, and 20 cold-tolerant lines from Almora were planted. Though all the entries had

Table 3. Performance of promising submergence-tolerant breeding lines/varieties for yield and other traits.

Genotype Date to 50% flowering

Plant height (cm)

Panicle length (cm)

No. of tillers hill–1

Yield (kg ha–1)

Madhukar 15 Oct. 180.1 24.4 8.2 3,400Barh Avarodhi 27 Oct. 175.7 26.2 10.2 3,900NDGR 24 30 Oct. 161.3 25.7 6.7 3,100FRG 10 27 Oct. 179.7 26.0 8.4 3,636FRG 13 27 Oct. 181.4 27.2 10.2 4,040NDGR 60 21 Oct. 181.5 24.3 8.6 3,800NDGR 63 20 Oct. 171.6 25.4 6.8 3,750NDGR 65 19 Oct. 181.7 26.4 8.2 4,200NDGR 80 27 Oct. 188.6 26.3 9.0 4,350NDGR 81 4 Nov. 156.8 25.8 9.2 4,667NDGR 82 27 Oct. 163.8 27.5 8.2 3,437NDGR 83 28 Oct. 182.4 26.4 8.6 3,850NDGR 84 29 Oct. 176.7 27.2 8.4 3,700NDGR 85 28 Oct. 181.8 26.8 8.6 3,850NDGR 86 30 Oct. 181.3 27.4 8.8 4,440NDGR 87 30 Oct. 188.6 26.4 8.2 3,840NDGR 88 23 Oct. 186.4 26.8 8.4 3,900NDGR 37 25 Oct. 188.7 24.4 8.2 3,750NDGR 104 16 Oct. 190.4 26.0 7.4 3,867NDGR 105 28 Oct. 163.7 24.4 8.0 3,667NDGR 106 15 Oct. 184.7 26.2 8.4 3,637NDGR 107 25 Oct. 183.7 24.4 8.2 3,737NDGR 108 26 Oct. 191.4 23.8 7.8 3,850NDGR 109 14 Oct. 186.7 24.6 8.0 3,950NDGR 110 22 Oct. 175.7 25.8 8.6 4,267NDGR 111 28 Oct. 184.7 24.6 8.6 3,900NDGR 103 21 Oct. 176.7 26.8 8.2 3,037NDGR 70 22 Oct. 181.4 27.0 8.4 3,900FR13(A) 20 Oct. 156.4 24.8 8.2 3,850Sabita 30 Oct. 176.9 25.2 7.4 4,040

18 Dwivedi

good germination, seedlings could not tolerate the chilling cold prevailing at this location. As a result, high mortality was recorded at the seedling stage. However, VL 93-288, VL Dhan 81, and VL 88-1011 were promising. During 2002, eight improved boro rice varieties were evaluated in farmers’ fields in the traditional boro rice-growing areas of Gorakhpur and Sant Kabir Nagar districts. Seedling vigor, days to maturity, phenotypic acceptability score (range of 1 = excellent to 9 = poor), and grain yield of these varieties are given in Table 4. Prabhat (5.21 t ha–1) was the outstanding variety, followed by Dhan Laxmi (5.02 t ha–1) and Richharia (4.20 t ha–1). Other varieties showed moderate performance. Gautam was slightly late in maturity. These varieties were also assessed on larger plots in farmers’ fields involving 13 farmers from Ramnagar Kajraha Village in Gorakhpur District and 4 farmers from Jhumia Village in Sant Kabir Nagar District. Similar patterns in yield and maturity duration were observed. This indicates that there is immense scope to increase the rice productivity of boro areas, where the main rice crop fails due to excess water, by providing integrated crop management practices such as cold-tolerant, short-duration varieties and nursery management to reduce the effects of cold injury.

Validation of improved germplasm in target environments

Rice yields are severely reduced by flash floods in addition to stagnant flooding. Tolerant rice varieties can make a sig-nificant difference in rice yield. Improved varieties released in the past have been found limiting in adaptation to varying on-farm conditions probably because of the very wide variability in this ecology, besides farmers’ needs and preference. It is therefore necessary to consider the farmers’ participation in varietal evaluation. The farmers’ participatory varietal selection approach was followed for selecting appropriate varieties by taking into account farmers’ preferences. To stabilize rice yield in

flood-prone ecologies, on-farm trials were conducted using lines/varieties suitable for flash-flood cultivated areas in eastern Uttar Pradesh using participatory selection during the 2000-02 WS. On-farm trials included seven varieties/lines (Table 5). These activities were implemented in two districts of eastern Uttar Pradesh, Bahraich and Barabanki, involving three villages and 5–7 farmers from each district, where most of the farmers were small to marginal. Farmers’ field evaluations were done in researcher-de-signed and -managed trials. Farmers were provided with only seed as an input. The rest of the inputs had to be supplied by the farmers. Farmers were asked to grow the lines/varieties using their own methods. Seeds were given to the farmers along with a questionnaire with varietal characteristics to be scored in comparison with the local popular variety. These characteristics were vigor, lodging, reaction to diseases and pests, yield, grain quality, threshability, and overall farmers’ opinions and reasons. Scores given to characteristics were 1 = very good, 2 = good, 3 = medium, equal to farmers’ variety, 4 = bad, and 5 = very bad.

Table 4. Performance of some improved varieties during boro season, 2002.

Variety Seedling vigor50% flowering from the date

of transplanting (days)

50% flowering from the date of nursery (days)

Days to maturity

Yield (kg ha–1)

NDR-97 Moderate 63 130 158 3,823Barani Deep Moderate 64 130 160 3,676Gautam Good 69 137 175 3,241Prabhat Very good 69 137 160 5,210Richharia Good 54 131 170 4,203Saroj Good 72 145 171 4,071Dhan Laxmi Good 75 147 170 5,021Joymati Average 63 140 167 3,648Boro Dhan Good 60 130 180 4,040

Date of nursery: 23 November 2001 CD: 0.682Date of transplanting: 28 January 2002

Table 5. Overall scores for farmer opinion and average yield of breeding lines/varieties evaluated under flash-flood conditions during 2000-02 across locations in Barabanki and Bahraich of eastern Uttar Pradesh.

VarietyFarmers

(no.)Average score

for opinionAverage yielda (q ha–1)

2000 2001 2002

NDGR 70 9 2.4 1,930 2,310 –NDGR 24 12 2.9 1,700 2,040 –FRG 13 9 3.0 1,730 2,030 1,040NDGR 63 9 2.9 1,550 1,930 –Madhukar 9 4.0 1,380 1,760 870Barh Avarodhi 15 2.3 1,630 2,240 1,090Sugapankhi 9 3.1 1,280 1,650 1,510NDGR 88 5 3.0 – – 1,070NDGR 82 7 3.0 – – 820

a2000 WS = excess rain coupled with flooding, 2001 WS = normal year, 2002 WS = drought year.


The yields of flood-prone rice are highly unstable and depend on flooding depth, duration of crop submergence, and stage of the crop when submerged. Yield of tested entries was variable during 2000-02 because of changes in rainfall distribution and flooding pattern (Table 2).The lowest yield was recorded during the exceptional drought year (2002) and entries yielded higher in 2000 and 2001. However, yield trends were similar. Most of the farmers scored variety Barh Avarodhi as good and it was rated first in both districts for its bold and shining grain (Tables 5, 6). Even other varieties such as NDGR 70 and NDGR 24 yielded more, along with NDGR 63, which flowered 20 days earlier without sacrificing yield. Rice varieties performed well in Assam and Bihar under the multilocational farmers’ participatory varietal selection program in five states of eastern India with flood-prone conditions. Since the farmers’ perceptions on adoption and acceptability of Barh Avarodhi were positive within and be-

tween states, it can be disseminated to other states/countries having similar ecological conditions. There is not much choice in the released varieties of Uttar Pradesh for stagnant deep flooding. Hence, to expedite the interstate flow of promising rice varieties developed elsewhere in the country for deepwater conditions, rice varie-ties Rajshree from Bihar, Ranjit and Bahadur from Assam, along with Barh Avarodhi and Jalpriya were evaluated fol-lowing the farmers’ participatory approach during the 2002 WS in four districts—Bahraich, Barabanki, Faizabad, and Basti—involving three farmers from each district. Varietal performance differed among the test sites. Rajshree yielded the highest (3,583 q ha–1), followed by Barh Avarodhi (2,468 q ha–1), Ranjit, and Bahadur (2,324 q ha–1) considering overall performance across locations (Table 7). Most of the farmers scored the varieties good in com-parison with their own variety because of their higher yield. Among these, Rajshree was rated first because of desired

Table 6. Number of scores of farmer opinion given to each entry and corresponding remarks made by farmers.

Variety

Number of farmers giving

scorea of Mean Farmers’ remarks

2 3 4

NDGR 70 7 – 2 2.4 Moderate submergence toleranceNDGR 24 4 3 3 2.9 Bold shining grainFRG 13 4 1 4 3.0 Stiff strawNDGR 63 4 2 3 2.9 Moderate survivalMadhukar – – 9 4.0 Susceptible to neck blast, red

pericarpBarh Avarodhi 13 – 2 2.3 Bold shining grain, better survivalSugapankhi 4 – 5 3.1 Drought-tolerant and stableNDGR 88 – 5 – 3.0 Moderate performanceNDGR 82 – 7 – 3.0 Good survival and medium slender

grain

aScores for farmers’ opinions: 1 = very good, 2 = good, 3 = medium (equal to farmers’ variety), 4 = bad, 5 = very bad.

Table 7. Average yield of varieties/lines under stagnant flooding and over-all score of farmers in participatory varietal trials conducted in different districts of eastern Uttar Pradesh during WS 2002 and 2003.

VarietyAverage yield (q ha–1) Average score

for opiniona

Barabanki Bahraich Basti Faizabad

2002 2003Ranjit 2,360 1,870 2,746 2,493 2,429 2.4Bahadur 1,700 2,610 2,510 2,490 3,088 3.0Rajshree 2,800 2,436 2,200 2,776 3,583 2.3B. Avarodhi 2,573 2,785 1,956 2,558 2,498 2.4Jalpriya 1,766 – 2,137 – 3,031 3.1Jal Lahari – – – – 4,141 2.3

aScores for farmers’ opinions: 1 = very good, 2 = good, 3 = medium (equal to farmers’ variety), 4 = bad, 5 = very bad.

20 Dwivedi

maturity and lowest infestation of diseases and pests (Ta-bles 7, 8). The majority of the farmers preferred Jal Lahari because of its high tillelring ability and good yield, followed by Rajshree. The most promising lines rated highly by most farmers and that produced relatively higher yield were Jal Lahari and Rajshree. Bahadur, although it produced a higher yield (3,088 q ha–1), was not rated good because of its lodging tendency and moderate infestation of stem borer. It was also moderately susceptible to brown spot. Though modest, the varieties have significantly helped to increase on-farm productivity. It is not envisaged that the varieties of the future will alone enhance productivity but they should have stable performance. Therefore, efforts should address yield stability and greater cropping diversity in varietal development besides yield alone for this complex fragile ecosystem.

Table 8. Number for scores for farmer opinion given to each variety and corresponding remarks made by farmers.

VarietyNumber of scores

Mean Remarks1 2 3 4 5

Ranjit – 7 – 2 – 2.4 Brown spot infestationBahadur – 4 1 4 – 3.0 Lodging tendency, brown spot

and stem borerRajshree – 13 – 2 – 2.3 Timely maturity, good yield

under shallow waterB. Avrodhi – 7 – 2 – 2.4 Slightly late, good survival, bold

shining grainJalpriya – 4 1 4 – 3.0 Fine grain, neck blastJal Lahari – 13 – 2 – 2.3 Good tillering and high yield,

trace of bacterial blight

References

Dwivedi JL. 1996. Problem and prospects of boro rice in Eastern U.P. Paper presented at International Flood-prone Rice Workshop held at Rajendra University, Pusa, Bihar, India, 28-31 October 1996.

Dwivedi JL, Senadhira D. 1996. Genetic control of elongation ability in flood-prone rice. Paper presented at the 2nd International Crop Science Congress held 7-24 November 1996, New Delhi, India. (Abstract.)

Notes

Author’s address: Senior rice breeder, Crop Research Station, N.D. University of Agriculture and Technology, Masodha, Faizabad 224 133, Uttar Pradesh, India.

Breeding for rainfed lowland, deepwater, and boro land in Bihar, India: achievements and challenges 21

Rice in Bihar covers about 3.8−3.9 million ha of land under diverse ecological conditions. More than 70% of this area is rainfed. The rainfed lowland is the most important ecological system, followed by deep water. The state has two geographi-cal divisions, north and south, and in both divisions rice is the main crop. Because of the wide diversity in riceland in Bihar, productivity is poor. However, in recent times and in collaboration with IRRI, there has been significant progress in developing improved varieties for different situations, with the breeding program being reoriented on the basis of existing challenges and needs. Rainfed lowlands in Bihar fall within two broad categories: unfavorable and favorable. The favorable ecosystem virtually resembles the irrigated system, whereas the unfavorable ecosystem is highly variable. Varietal requirements for these two ecosystems are different; consequently, selection and evaluation criteria are modified. New entries are now evaluated jointly by farmers and scientists using farmers’ management practices. Through this process, a pureline selection, Vaidehi, was developed, while two other high-yielding varieties, Satyam and Kishori, were bred under the shuttle breeding program. Both have multiple resistance to biotic and abiotic stresses and are adapted to late planting. Santosh, a variety with good grain quality, was developed through the farmers’ participatory breeding project. In deepwater rice land, excess water causes serious damage and more tolerant and higher yielding varieties are needed. Boro (dry-season) rice, grown from Oct.-Nov. to Apr.-May, is highly productive, but germplasm with tolerance for cold stress at the seedling stage is required to expand the boro area. Recently, an EMS mutant genotype, Rasi, was developed with high cold tolerance and high yield potential (more than 8 t ha–1). It was released as Gautam. In subsequent years, Richharia, Dhanlaxmi, and Saroj (all with different maturity durations) were developed for the boro season. These varieties are becoming more popular and cultivated area is progressively expanding in traditional and nontraditional boro regions. There is now a subtle change in the attitude of farmers: yield is no longer the sole criterion for adopting new varieties; they also assess other traits such as cooking quality and an attractive look. This shift in preference could help popularize new varieties with good grain quality, even in the rainfed ecosystem.

Keywords: rainfed lowland, deepwater rice, boro rice, elongation, submergence, chaur, cold tolerance

Breeding for rainfed lowland, deepwater, and boro land in Bihar, India: achievements and challengesR. Thakur, N.K.Singh, and J.N. Rai

Bihar is a typical eastern state of India where rice is cultivated on 3.8–3.9 million ha of land under diverse agroecologi-cal conditions. Rice is cultivated on more than 60% of the cropped area. Agriculture is the most dominant sector of the state economy and 82% of the population is directly depend-ent on it. Consequently, rice production determines food security and stability in the state. Rice as a food crop is well known for its unique adaptation. It is grown as an upland crop on precarious moisture conditions and in water as high as 2–3 m, just like in deepwater areas of Bihar. There are several intermediate conditions between these two extremes. Apart from these, rice is also grown under diverse climatological and edaphic situations, which are often not conducive to proper growth and productivity. More than 70% of the rice area in Bihar is rainfed, with productivity determined mainly by the pattern of the monsoon.

Physiographic features

The Ganga River divides Bihar into two distinct broad regions: the north Bihar alluvial plains and the south Bihar alluvial plains. North Bihar alluvial plains. This region is practi-cally plain land sloping toward the southeast. Many natural riversGandak Ghaghra, Burhi Gandak, Kamal Balan, Bag-mati, Koshi Balan, Mahananda, and their tributariesflow through this area. The region has two distinct agroclimatic zones: northwest and northeast. The northwest (zone 1) cov-ers Chapra and Tirhut divisions. The zone slopes toward the southeast, as seen in the direction of the river flow. Vast wa-terlogged areas in the districts of Saran, Vaishali, Samastipur, Muzaffarpur, and Begusarai are found. Due to the near-flat landscape and saucer-shaped depressions (chaurs) between the rivers, vast areas become flooded during the rainy sea-son in practically all districts in this zone. When floodwater recedes from most areas, the saucer-shaped depressions and the abandoned channels of the rivers and lakes remain flooded

22 Thakur et al

for various durations. Due to the construction of irrigation infrastructure and canal systems under the Gandak project, most of the natural drainage systems had been disturbed. This caused more flooding and waterlogging in the central and western parts of the zone. Embankments and roads also worsened the waterlogging problem. The Ganga River serves as the major drainage channel in which all riverflow from various points meets. Total rice area in the region is about 1.5 million ha. The northeast plain (zone II), covering Kosi Division, generally slopes toward the southeast; Kosi, Mahananda, and Ganga are the major rivers. This zone is full of abandoned beds and dead channels of the Kosi River and its tributaries. In addition, small lakes, oxbow lakes, and patches or marshy grounds are frequently encountered. The total area of rice in this zone is about 0.79 million ha. SouthBiharalluvialplains. This region has two divi-sions and slopes toward the northeast. Important rivers other than the Ganga are the Sone, Punpun, Falgu, Badua, and Chir. Except for the Sone River, all others are seasonal. The region has comparatively higher irrigation resources and more than 60% of the land is normal, not like the situation in the north Bihar plains. Total rice area is about 1.5 million ha.

Challenges and constraints

When high-yielding varieties (HYVs) with features such as dwarf habit; short, erect leaves; and better response to N fer-tilizer were developed in the late sixtieswhich triggered the Green Revolution in India and elsewherebreeding methods were modified everywhere to accommodate these HYVs. The general belief was that the Green Revolution technology could provide solutions to all rice production problems. Rice breeders were consequently made to develop short-stature, N-responsive varieties for all rice ecosystems, including deepwater areas (Morshima-Okinow 1964). The success of the efforts was not tangible. It was soon realized that the Green Revolution technology was best suited for the more favorable ecosystems such as irrigated lands but had little relevance to the more heterogeneous and unfavorable rainfed production systems. A few varieties were developed through this approach for the rainfed lowland and deepwater ecosys-tems. Their average yields were very high in managed trials and they were therefore recommended for release. However, none of them were widely adopted. Mansarowar was released in the late eighties for the rainfed lowlands based on yield data from the all India coordinated trials. When this variety was included in frontline demonstrations in farmers’ fields in typical lowland areas and was planted a little late, it did not flower. Such varieties can better express their yield potential in a favorable environment. Because the rainfed system is highly heterogeneous, varieties for this ecosystem must be able to withstand local environmental fluctuations. Rainfed lowland and deepwater ecologies are parts of a continuum of rice ecosystems, but they differ with respect to water depth and other abiotic stresses. Consequently, the varietal

requirements are different. The constraints encountered in these ecologies are dealt with separately in this paper. Rainfed lowland. This ecosystem constitutes more than 50% of the rice area in Bihar and is the most important, However, it faces the following constraints: l The hydrology is irregular. When monsoon rain is

heavy, there is excess water in the field; when rain is scanty, drought is experienced at any growth period.

l Flash flood could inundate the fields for at least a week at any time during the season.

l The monsoon pattern determines sowing and plant-ing time. If rain is timely, sowing is normal. Planting is delayed when rain is delayed. This happens in canal-irrigated systems also.

l Photoperiod-sensitive tall cultivars are predomi-nantly cultivated; these have weak culms and are prone to lodging. The yield potential of these culti-vars is low, but they are adapted to prevailing abiotic stresses. They seem to tolerate both excess water and drought and are adapted to delayed sowing and transplanting, which occur because of variation in the onset of the monsoon season.

Deepwater rice. Deepwater rice land, locally called chaur, is found all over north Bihar, covering about 0.83 million ha (Catling 1992). This is the main source of liveli-hood of millions of resource-poor farmers. Water on the peripheral parts of the chaur remains shallow, medium in the middle, and very deep in the center. Varieties are used on the basis of this water depth gradient. Tall, photoperiod-sensitive cultivars with no elongation ability are grown on the rela-tively shallow periphery, and taller elongating cultivars are grown in the middle. Cultivars with high elongation ability are grown in the deeper central parts. However, almost all current cultivars are traditional. The transplanting method is used in the shallow areas and direct seeding is practiced in the deeper areas. The major constraint in this ecosystem is the unpredictable rainfall during different growth stages and reasonable yield is thus not assured. If a crop survives and matures in the central portion, harvesting is usually done using boats, as the water remains in the field for a very long period, even up to March or April. Bororice. Boro rice is grown during winter in low-ly-ing areas, especially in river deltas in both eastern India and Bangladesh. It is commonly grown in deeply flooded areas of northeastern Bihar, West Bengal, Assam, eastern Uttar Pradesh in eastern India, and in many areas in Bangladesh. Groundwater in these areas is shallow and waterlogging/water stagnation prevails in some places during winter. The winter rice crop grown in Myanmar also resembles this boro rice. It suffers from damage due to low temperature of vary-ing degrees at the seedling stage. Where temperature during winter is not too low, boro rice covers large areas as in As-sam (Pathak 1995), West Bengal (Chatterjee 1995), and in small areas in Bihar (Thakur 1995). In Bihar, cultivation is mainly confined to favorable areas adjoining West Bengal,


the traditional boro region, and has not spread beyond this area mainly because of very low temperatures in so-called nontraditional boro areas. Boro rice produces more grain yield than wet-season rice in the same ecology (Singh et al 2003). As previously mentioned, some varieties such as Gautam recorded yields between 8 and 10 t ha–1, even in farmers’ fields (Thakur et al 1994). The productivity of boro rice is high mainly because of higher solar radiation, lower night temperature throughout the crop growth period in winter, and favorable temperature during ripening (Singh et al 2003). Major constraints: Low temperature during winter prevents the expansion of boro rice in nontraditional low-lying areas. This is because low temperature leads to l Poor and slow germination of seeds; l Stunted seedling growth and death in some cases

when temperature becomes very low; l Nonsynchronized tillering and flowering; l Delayed flowering and maturity, which increases

irrigation cost and exposes the crop to higher tem-peratures during summer.

Among the biotic stresses commonly encountered dur-ing the boro season are insect pests such as brown planthop-pers (BPH), stem borers, and leaffolders and diseases such as brown spot and sheath blight. Drought and western rains during maturity are also encountered. Currently, boro rice receives greater attention and higher priority in the state’s research agenda because of its very high yield potential and less vulnerability. It is an excellent option to increase and sustain the productivity of deepwater rice land.

Collaboration with IRRI

The Rajendra Agricultural University has a very long partner-ship with IRRI through shuttle breeding, farmers’ participa-tory breeding, and deepwater rice breeding projects. Bihar also participated in the IFAD-funded project on validation and delivery of new technologies for flood-prone rice lands. Through these projects, Pusa was selected as the main center for testing and distribution of improved deepwater rice in eastern India and Bangladesh.

Progress in breeding (1991-2000)

Significant achievements in breeding varieties for rainfed and boro rice production systems were noted during this period. The hybridization program was reorganized, and evaluation of new entries was modified to suit growing conditions. These efforts and adjustments resulted in the development of several popular varieties, replacing most of the traditional cultivars. Earlier, selection of segregating populations was made using nurseries grown under good management condi-tions on research stations. Cultivars identified through this approach were released as varieties with record yields, but these were not adopted widely in the ecosystems for which they were intended. This aspect was carefully considered

while restructuring the new breeding projects designed for the rainfed ecosystems, where different breeding programs were developed and executed for each of the major ecosystems. A clear understanding of the major challenges in these ecosys-tems became the basis for revising the breeding objectives. Rainfedlowland. Broadly, rainfed lowlands have two distinct subecosystems: favorable and unfavorable. Unlike the situation in the unfavorable ecosystem, several HYVs were available for the favorable ecosystem. Consequently, local cultivars were predominant in the unfavorable areas. As a breeding strategy, pureline selections that were sufficiently tested were promoted in areas with extreme adverse condi-tions based on practical experience and the poor performance of the released modern varieties (Thakur 1995). Local germ-plasm was thus included in the multilocation trials conducted for the rainfed lowland, along with improved breeding lines. The following are examples of varieties common in these areas: l Mahshuri. This cultivar was once very popular in

the lowlands. However, it virtually failed when planted late (more than two consecutive years) due to delayed monsoon rains. It also became highly susceptible to BPH and drought. The breeding pro-gram then focused on developing a variety that is more suited to these growing conditions. As stated earlier, experimental trials were usually conducted under normal sowing and planting conditions. However, through the shuttle breeding program, the performance of new entries under delayed planting was assessed for the first time, along with normal planting, and this has been very successful. Three varieties were subsequently developed based on the work of the shuttle breeding network (Thakur et al 1994,1998).

l Vaidehi. This pureline selection from a local cultivar is tall and has a thick and sturdy stem and dark green foliage. It is highly resistant to drought and tolerates water depths of up to 100 cm. Photoperiod-sensitive and suitable for delayed planting, it is recommended for growing in relatively more adverse conditions.

l Satyam. This variety was developed from a cross between floating rice cultivar Desaria and an im-proved deepwater variety, RD19. It is semitall, with 140-d duration, and is highly adapted to delayed transplanting. It tolerates water depths of up to 90 cm and is resistant to BPH, whitebacked planthop-pers (WBPH), and bacterial leaf blight (BLB). The entry is exclusively tested in eastern India under the shuttle breeding program for 3 years.

l Kishori. Developed from IR8 and Barogar (a deep-water local cultivar), this variety underwent exten-sive testing through the shuttle breeding network in eastern India. It is semitall with 145-d maturity. It is highly adapted to delayed planting conditions and is resistant to BPH, WBPH, and BLB.

24 Thakur et al

Bihar is also an active partner in the farmers’ partici-patory breeding network. The project was mandated to seek farmers’ views and opinions when selecting varieties and fixed breeding lines that would be grown along with check varieties popular in the target region. In the past, farmers’ views were seldom taken. This project practically docu-mented the criteria used by farmers in accepting or rejecting a cultivar. Many times, a cultivar was rejected by breeders on the basis of such criteria as yield, duration, and reaction to diseases and pests. However, farmers have their own (and even different) opinions. For example, in spite of the release of several improved varieties for a specific ecology, farmers continue to rely on their own traditional varieties. A few years earlier, they participated in testing new varieties. Brown Gora, a traditional cultivar for the uplands, is predominantly grown in Chotanagpur Plateau. Several improved varieties have been released for this system, but this cultivar remained popular because some inherent qualities in it are lacking in the new varieties. We came across an interesting situation in which several improved cultivars were included in a demonstration trial both in farmers’ fields and on the research station. Those with poor ratings were rejected and new cultures were added. A survey was then conducted to find out the composition of varieties under rainfed lowland in the village where the demonstration trial was held. Cultivar RAU1306 was found to be grown on a large scale, replacing local landraces. This line was selected from the demonstration trial in farmers’ fields; ironically, it was dropped from the research station trials because of relatively poor yield. Subsequently, this line was formally included in multilocation trials and was released as Santosh in 2001 (Thakur et al 2003). Santosh was developed from a cross between Pankaj and Br8. It is tall, with 135−140-d duration, and phenotypically resembles Mahshuri. It has very good grain and cooking quality. It performs well under low-input conditions and is adapted to late sowing and planting, traits not found in Mahshuri.

Deep water

To date, no significant achievements in breeding for this ecosystem have been made, in Bihar or elsewhere. Still, tra-ditional varieties predominate (Thakur 2004). Excess water is a major impediment, where it sometimes remains in the deepwater chaur lands up to March and April and sowing of dry-season wheat becomes impossible. An alternate technol-ogy is therefore critical. Cultivation of boro rice in such lands is feasible from October-November to April-May. Currently, Vaidehi, which is released for the rainfed lowland, is also grown in these deepwater areas. Boro rice in West Bengal and Assam is grown on a large scale during this period. In fact, it has long been grown in some districts of Bihar near West Bengal. It first started in chaurlands, then moved on to irrigated medium lands. The climatic conditions (temperature, soil type) in these Bihari districts are similar to those in West Bengal and so this area is considered a traditional boro region. Boro rice cultivation,

however, did not extend beyond this region, and the most important constraint was the extremely low temperature dur-ing winter. Varieties grown in the traditional boro region did not perform well elsewhere because of the low temperature. If cold-tolerant varieties were made available, then it would be possible to expand boro rice in chaurland to make this deepwater ecosystem more productive. Research to develop cold-tolerant varieties for the boro season began in the early 1990s at Pusa (hot spot for cold screening), where a large number of improved cultivars, including those grown in traditional boro regions, were evalu-ated. One of the EMS mutants of Rasi yielded as high as 8.8 t ha–1. Named Gautam, it had an average yield of more than 7.5 t ha–1 in multilocation yield trials (Thakur et al 1994). Gautam has dwarf habit, profuse tillering, dark green leaves, and high tolerance for low temperature at the seedling stage and for high temperature at anthesis. Its grains are medium fine. The success of Gautam in deepwater chaur land in-tensified research efforts to develop varieties with different durations and superior grain quality. Richharia, Dhanlaxmi, and Saroj were developed in subsequent years (Thakur et al 2002a,b). l Richharia: Developed from IET7464/Pusa 33. It

is a semidwarf and an early-maturing variety with long grain. Released in 2000, it has excellent syn-chronized flowering in the boro season.

l Dhanlaxmi: Developed from ES1-2-3 (very early)/IR36. It has medium duration during the boro season and good grain quality. It is tolerant of Zn deficiency and resistant to bacterial leaf blight, brown plan-thopper, and bacterial smut. It was also released in 2000.

l Saroj: Developed from a cross between Gautam and type 3 (Basmati rice). It was basically developed for the kharif (wet) season but has now been widely adopted for the boro season. Its grains are long and slender. It was released in 2001.

These varieties became popular in selected nontra-ditional and traditional boro areas in low-lying deepwater chaurlands. However, there is still a demand for aromatic boro varieties. A program started and several cultivars and breeding lines were evaluated during the boro season. One of the breeding lines, RAU1397-19-3-7-9-4-2 derived from IR36/type 3, has been identified. It has a yield potential of 7 t ha–1 with long cylindrical grains and fine aroma. This will be the first aromatic variety for the boro season.

Identification of cold-tolerant breeding lines

Cold-tolerant varieties are needed to expand boro rice area in the vast waterlogged chaur lands in Bihar and Uttar Pradesh. Normally, this is not required in traditional boro regions, but in 2000-03, a severe cold during winter damaged most of the boro seedlings, especially in Assam. This triggered a large-scale screening of breeding lines developed locally through hybridization and of those received from the Directorate of


Rice Research, Hyderabad. About 80 entries were tested during 2001-02 and 210 entries were evaluated in 2002-03. Thirty entries were identified as highly cold-tolerant. The 15 best entries are listed in Table 1.

Evaluation of screening methods

The evaluation of new improved germplasm was done through an on-farm research trial. It comprised 8-10 entries that differ in height, duration, and grain quality, along with locally adopted checks. The trials were conducted jointly by farmers and scientists at several sites following farmers’ prac-tices. The same set of lines was also grown at the research farm. The good performers across all trials were then included in the on-farm trials using two to three entries per farmer and a large number of sites (10−15). The data from all sites were compiled and analyzed; the entries that spread across the whole village were selected and considered accepted by farmers. This method was developed during the evaluation of Santosh; the same protocols were applied for testing and release of Saroj and Prabhat for the boro season and Satyam and Kishori for the wet season in the rainfed lowland.

Progress in deploying new varieties

The Agriculture Department in Bihar is well established to disseminate new technologies developed by the university to farmers. Joint meetings of both agencies are periodically held and the seed production program was set on the basis of target demand. Seed production of varieties favored by farmers is done on a limited scale as these varieties normally spread relatively faster among farmers. Some difficulties are encountered in disseminating the new varieties recommended for adverse conditions; it takes a long time for these varie-ties to become accepted. In a situation like this, university and Agriculture Department staff discuss ways to enhance delivery mechanisms to end users.

Vaidehi was initially recommended for the rainfed low-land, but it was also found suitable for the mixed-cropping scheme with mungbean/sesame in medium deepwater areas. This variety has an erect growth habit and is not affected when grown with nonrice crops. Mixed cropping, practiced during March-April, currently covers areas in Samastipur, Darbhanga, and Muzaffarpur districts. Satyam has become popular in BPH-endemic regions, especially in south Bihar. It covers more areas in Madhubani and Darbhanga districts. Kishori, because of its bold grains, is getting popular in areas where special rice is in demand. Santosh has spread in all zones and has a wide coverage in Zone I in north Bihar. This variety performs very well under low fertilizer input and adverse conditions, particularly in the irrigated Zone IIIB. During 2002-03, the monsoon season was delayed, and the majority of rice varieties were planted late. This resulted in nonsynchronized flowering of most varieties such as Nata Mahshuri and Sita but not Santosh, which made this variety more popular in these areas. The most visible impact of new varieties was seen with the new boro lines. When Gautam was released, it spread so fast that, in the first year of its release, it covered about 100 ha of chaurland in Zone II in Madhubani District. It also became popular in the traditional boro region of Saharsa District. Within just a few years, all new boro areas were covered by this variety alone. However, with large-scale cultivation, this variety started to show high susceptibility to BPH, suggest-ing the need for a new resistant variety. This variety is also late maturing; the water stored in chaurland was sometimes not sufficient to irrigate larger areas. Richharia, Dhanlaxmi, and Prabhat (varieties recommended for the upland) started to cover wider areas in traditional as well as nontraditional boro areas. At present, Prabhat is most popular in Madhubani District and has replaced Gautam. The government of Bihar launched a project to popular-ize boro cultivation in the nontraditional boro regions, with the hope that a wide coverage of these areas will lead to

Table 1. Cold-tolerant breeding lines for the boro season identified in 2002-03 at Pusa-Samastipur, Bihar.

Designation Pedigree Survival (%)

RAU1406-1-3-2-2-5 Turant Dhan/Rasi 90.4RAU1406-2-1-4-2-3 Turant Dhan/Rasi 95.2RAU1401-28 Rajshree/Gautam 92.0RAU1400-5-7-CT Gautam/Type 3 96.2RAU1395-2-7-3 Vaidehi/Sita 97.0RAU1429-3-9-10-11 IR68888A/Saroj 97.0IET17257 HKR 239/Basmati 370 98.0IET16958 IR38699/IR 19431 92.5IET17296 Indrajani/SYE 3-43-57 98.5IET17243 – 99.0IET17256 HKR239/Basmati 370 96.0IET17457 CO 1 064-5/IR9129-320-3-3-3 96.5IET17269 IR54 95.6IET17519 IR38699/IR19431 98.3IET16935 JLD4/ADT39 98.4

26 Thakur et al

enhanced productivity of deepwater land. Selected varieties from Bihar, West Bengal, Assam, and Uttar Pradesh were jointly introduced in on-farm trials over the past several years. Gautam in West Bengal, Prabhat in Uttar Pradesh, and Saroj in Bihar and Assam were recognized as the top yield-ers (Chaudary et al 2003). Saroj performed exceedingly well in the traditional regions of Bihar and was highly preferred for its grain quality. This variety also became popular in the irrigated midland plains of south Bihar.

Breeding objectives and challengesin the next 5–10 years

There is now a subtle change in farmers’ attitude toward new varieties and yield is no longer the sole criterion that ensures adoption of a particular variety. Besides yield, grain quality receives prime consideration. An example can be cited to elu-cidate the point. We developed a very early-maturing variety (75 d total duration), Turanta Dhan, for upland conditions and also for double rice cropping. It has a yield potential of 4 t ha–1. When we asked a resource-poor farmer who had successfully grown this variety, he said he did not retain its seeds (he sold all of them in the market) in spite of its high yield. His reason is that it had very poor eating quality. Even a farmer of this category now needs better quality rice. Though good grain quality is preferred, greater attention should be given to the cooking and eating qualities. Disease-resistant varieties are needed to replace cur-rent susceptible ones. The university recently released a BPH-resistant, high-yielding variety, Kanak, for the irrigated midland and favorable lowland areas. Its yield in frontline demonstration trials in a 8-ha field was 7.5 t ha–1. A serious extension campaign was carried out to disseminate it; the aim was to replace susceptible variety Sita. These efforts failed, however, and Sita remained popular. The main reason again was Kanak’s poor grain and cooking quality. While deciding on the objectives of a breeding program, one emerging aspect is the focus on phenotypic appearance of the genotype, which seems to help in the spread of a variety. Sita was released in Bihar in the early 1970s and became very popular in the irrigated areas. However, it was later found to be highly susceptible to BLB. A BLB-resistant, high-yield-ing variety, Rajendra Dhan, was subsequently released to replace Sita, and it outyielded Sita in yield trials. Again, the variety was not accepted mainly because its look was not that appealing. Leaves were erect, virtually covering the panicles. Besides, its grains were not as long as Sita’s. Sita, in contrast, has an attractive look; erect, dark green leaves; and long drooping panicles. When this variety is grown by a farmer in a village, other farmers are attracted to grow it solely because of its attractive look. This variety has been cultivated in the last two decades. Another example is upland variety Prabhat. At the dough stage, all leaves of this variety are droopy, except for the flag leaf, and only the grains are visible. This look has helped it gain popularity in the boro

season as well as in upland areas during the wet season, not only in Bihar but also in other states. Thus, high yield as the primary target in breeding may not be sufficient; other features such as good grain quality and attractiveness should be considered. For yield to stabilize at higher levels, new varieties must be tolerant of common stresses associated with the ecosystem—pests, diseases, drought, and submergence. Several stresses are as-sociated with specific unfavorable ecosystems and regions as evidenced by the popularity of specific cultivars in a par-ticular region. Since local cultivars are usually grown under adverse conditions, the new varieties must have such distinct advantages as higher grain yield, but they at least must retain similar grain and cooking qualities. Several varieties need to be developed to account for the wide range of variability in these ecosystems. This great challenge for breeders could be overcome if conventional breeding approaches were supplemented by biotechnology tools. Yield and quality do not seem to be well related, and there might be some barriers for their combination. This problem could probably be resolved through molecular breed-ing. That this could be achievable is reflected in the efforts to develop Basmati rice for these ecosystems. Eventually, there is a need to devise special and novel breeding strategies for the adverse rainfed lowlands and deepwater ecologies as breeding endeavors have so far not been very successful. As such, it is better to use local cultivars in hybridization and improve their deficiencies. For deepwater areas, boro rice is probably an excellent choice for Bihar and Uttar Pradesh, but varieties that are highly tolerant of cold stress are also needed.

References

Catling HD. 1992. Rice in deepwater. Manila (Philippines): Interna-tional Rice Research Institute. 542 p.

Chatterjee SD. 1995. Status of boro rice in West Bengal. In: Report of an INGER boro rice monitoring visit to India, Bangladesh and Myanmar. Manila (Philippines): International Rice Research Institute.

Chaudhary VK, Thakur R, Dwivedi J, Maity S, Deo Chaudhary PK, Singh NK, Nillanjay. 2003. On-farm evaluation of improved boro rice germplasm in eastern India. In: Singh RK, Hossain M, Thakur R, editors.. Boro rice. IRRI India Office, NASC, Pusa New Delhi. p 221-232.

Morshimo-Okino H. 1964. Report of observations on floating rice in Thailand. 9 p. (unpubl.)

Pathak AK. 1995. Problems and prospects of boro rice in Assam. In: Report of an INGER boro rice monitoring visit to India, Bang-ladesh, and Myanmar. Manila (Philippines): International Rice Research Institute.

Singh RK, Thakur R, Chatterjee SD. 2003. Harnessing boro rice poten-tial for increasing rice production in deepwater areas of Eastern India. In: Singh RK, Hossain M, Thakur R, editors. Boro rice. IRRI India Office, NASC, Pusa, New Delhi. p 51-72.


Thakur R. 2004. Genetic improvement of flood-prone rice: Where are we today and what are the future prospects? In: Bhuiyan SI, Abedin MZ, Singh VP, Hardy B, editors. Manila (Philippines): International Rice Research Institute.

Thakur R. 1995. Status of boro rice in Bihar, India. In: Report of an INGER boro rice monitoring visit to India, Bangladesh, and Myanmar. Manila (Philippines): International Rice Research Institute.

Thakur R. Sahi SP, Mishra SB, Singh UK, Mishra M, Rai JN. 1994. Gautam: an improved rice variety for winter (boro) season in Bihar, India. Int. Rice Res. Newsl. 19(2):19.

Thakur R, Singh AK, Singh RS, Mishra SB, Singh NK, Rai JN. 1998. Satyam and Kishori, two high-yielding varieties developed fro the rainfed lowlands of Bihar, India. Int. Rice Res. Notes 23(3):20-21.

Thakur R, Singh AK, Singh RS, Singh NK, Mishra SB, Mishra M, Singh UK, Rai JN. 2002a. Dhanlaxmi and Richharia, very early varieties released in Bihar, India. Int. Rice Res. Newsl. 27(1):31-32.

Thakur R, Singh AK, Singh RS, Singh NK, Mishra SB, Singh UK, Chaudhary VK. 2002b. Saroj an early maturing variety released in Bihar, India. Int. Rice Res. Newsl. 27(2):18.

Thakur R, Singh NK, Chaudhary VK. 2003. Recent advances of boro rice research in Bihar. In: Singh RK, Hossain M, Thakur R, edi-tors. Boro rice. IRRI India Office, NASC, Pusa New Delhi.

Thakur R, Singh NK, Mishra SB,. Singh AK, Singh KK, Singh RK. 2003. Santosh: a high-yielding variety for rainfed lowland developed through participatory breeding for Bihar, India. Int. Rice Res. Notes 28(2):25-26.

Notes

Authors’address: Department of Plant Breeding, Rajendra Agricultural University, Pusa, Samastipur 848125, India.

28 Sommut et al

Rainfed rice in Thailand accounted for approximately 88% of the total cultivated rice area and 79% of total production in 2004. Flood-prone rice (FPR), one of the rainfed rice ecosystems, accounts for 4.7% of total cultivated area, 5.4% of rainfed rice area, and 1.4 million tons of rice production annually. The cultivation system has changed somewhat from monocropping of FPR varieties in the wet season to double cropping, primarily due to the expansion of irrigation infrastructure. However, monocropping of FPR still predominates. Breeding for FPR in Thailand was strengthened with the establishment of collaboration with the International Rice Re-search Institute (IRRI) during 1991 to 2004. With the Prachinburi Rice Research Center (PRRC) and Huntra Rice Experiment Station (HTA) as key sites, collaborative activities involved exchange of germplasm and capacity building through degree and nondegree training and workshops. Breeding strategies followed included enhancing genetic variation by effective collection, evaluation, and use of rice germplasm, hybridization and selection, screening for biotic and abiotic stresses, the use of rapid generation advance (RGA) to shorten progeny selection time, screening of F2-F3 populations under natural flooding in farm-ers’ fields prior to replicated yield testing, and farmer participation in multilocation yield trials. More than 1,000 crosses were made and hundreds of breeding lines were tested by collaborators. Some lines were elevated to replicated yield trials and a number have been tested under acid-sulfate soil conditions, with a few lines in the process of being released. Three Thai varieties have also been released in Cambodia and five varieties, including two new plant type FPR, were released in Thailand. Training activities and workshops were held at PRRC and three scientists received doctoral degrees from the University of the Philippines Los Baños under IRRI’s sponsorship.

Breeding rice for deepwater and flood-prone areas of ThailandWilailak Sommut, Kalaya Kupkanchanakul, Prayote Charoendham, Udompan Promnart, and Suthep Nuchsawasdi

Thailand lies between 5o and 21oN latitude and between 97o and 106oE longitude. This peninsular Southeast Asian country shares boundaries with Myanmar in the west, Lao PDR and Cambodia in the northeast, and Malaysia in the south. The South China Sea touches the east coast, while the Indian Ocean and Andaman Sea border the west coast. Thailand has 51 million ha of land area, of which one-third is cultivated with annual crops and about 7% is under cultivation with perennial crops. Thailand belongs to the warm subhumid tropics. Four seasons are recognized: southwest monsoon from May through September, a transition period from the southwest to the northeast monsoon during October, the northeast mon-soon from November through February, and a premonsoon hot season during March and April. Temperatures in the Central Plain during the rainy season (May to November) average 27 ºC, with only an 8–10 ºC difference between the daily minimum and maximum temperatures. There is a brief cool period (December and January) when temperatures are as low as 2–3 ºC in the northern highlands.

Rice environments and production

Rice is not only a traditional staple food but is also linked with life and culture in Thailand, aside from being the coun-

try’s most important export crop. Administratively as well as geographically, Thailand is divided into four regions: central, north, northeast, and south. Each region has different rice-growing environments and yield potential (Table 1). Almost one-third of the land area of Thailand is located in the northeast region, which has more than one-half of the rice land but accounts for only 37% of total rice production. The average rice farm size is smaller than in other regions and soil erosion and drought during the dry season are acute problems. The central region, an intensively cultivated alluvial area, accounts for about one-fifth of the total cultivated rice land. Average farm size is large, and most farms have irriga-tion facilities, thus allowing two rice crops per year. Almost 75% of the dry-season rice grown under irrigated conditions is located in this region, which produces up to 31% of total production. Farm operations are almost entirely mechanized and farmers use the direct-seeding method of crop establish-ment to save labor. The northern region has almost one-third of the land area of Thailand. Upland rice is grown in the lower altitudes of high hills and in upland areas. Lowland rice is grown mainly in lower valleys and on some terraced fields where water is available. This region has about 24% of the total rice land and accounts for 28% of total production.

Breeding rice for deepwater and flood-prone areas of Thailand 29

The southern region constitutes about 14% of the to-tal area of the country. The region has only 5% of the total rice area and accounts for only 4% of total rice production, mainly for local consumption. Among the four regions, the north and central are more productive in terms of rice yield during both the major and second rice crops. Thailand has two main rice ecosystems—rainfed and irrigated. The rainfed ecosystem can be further divided into four subecosystems: rainfed lowland accounting for 80% and distributed mainly in the northeast and north, deepwater, floating (also called flood-prone), and upland. This paper covers the flood-prone rice (FPR) ecosystem, which includes floating rice (FR) and deepwater rice (DWR).

Importance of flood-prone rice

The large proportion of FPR is mostly grown along the Chao Phraya River in the Central Plain of Thailand, the most pro-ductive rice area (Table 2). The FPR area declined from 0.85 million hectares in 1982 to 0.50 million hectares in 1992, approximately 5.8% of the cultivated rice area (Charoendham et al 1995). However, it has also been reported that 21.9% of Thailand’s total rice area is devoted to FPR (Huke and Huke 1997). Nevertheless, FPR provides around one million tons of rice each year. Two types of FPR rice are grown: the deep-water rice (DWR) type that grows at average water depths of 50–100 cm for up to one month or more and the floating rice (FR) type that is grown at a water depth of 1–5 m with a flooding period of 3–4 months. Submergence tolerance is the most essential characteristic for DWR while good elongating and kneeing ability are desirable for FR. FPR has been grown as a monocrop in the past using traditional varieties that, while popular, yield only 1.2–1.8 t ha–1. Since 2000, however, double cropping has been prac-ticed using modern varieties (Table 3; Sommut et al 2004). Short-duration pre- or postflood rice crops have also replaced FPR in some areas of the Central Plain of Thailand (Table 3). Classification of areas by time of arrival of floodwater, rate of rise of water level, range of maximum water depth, and

time of recession of the flood is essential for the selection of appropriate FPR varieties.

Special characteristics of FPR

Many types of FPR can be found in a region and different maximum water depths require distinct FPR types. Farmers may use several FPR varieties where water depths change with topography over a short distance: slowly elongating (2–3 cm day–1) varieties for 50–80-cm water depth and fast-elon-

Table 1. Rice area, production, and yield by region in Thailand, 2003-04.

Region

Planted area

(million ha)

Production

Total (million tons)

Yielda

(t ha–1)

Major rice Second rice



North 2.30 0.46 4.96 1.85 2.64 4.09Northeast 6.04 0.10 8.54 0.30 1.72 3.14Central 1.67 0.66 4.70 2.89 3.04 4.60South 0.50 0.04 0.85 0.11 2.13 2.90Total 10.51 1.26 19.05 5.15 2.38 4.24

aBased on harvested areaSource: Center for Agricultural Statistics (2004).

Table 2. Flood-prone rice areas in Thailand based on water depth, crop year 1992-93.a

ProvinceCultivated area (ha)

TotalWater depth<100 cm

Water depth>100 cm

Ayutthaya 33,328 64,123 97,451Nakorn Sawan 59,527 11,593 71,120Pichit 52,136 13,859 65,995Nakorn Nayok 17,686 25,875 43,561Ang Thong 12,326 29,037 41,363Lop Buri 12,793 21,325 34,118Pitsanulok 21,049 8,598 29,647Prachin Buri 5,796 16,248 22,039Singh Buri 9,856 8,917 18,773Ratcha Buri 9,989 3,943 13,932Suphan Buri 7,778 2,456 10,234Sara Buri 4,336 5,278 9,614Chainat 4,114 1,917 6,031Nakorn Pathom 3,403 1,233 4,636Uthai Thani 3,369 383 3,752Petcha Buri 1,661 1,001 2,662Nongkai 492 3,072 3,564Khon Kaen 327 2,846 3,173Chaiyaphoom 2,887 18,986 21,873Udorn Thani 10 622 632Total 262,863 241,306 504,169

aEstimated area = no. of families × average cultivated area/family.Source: Charoendham et al (1995).

30 Sommut et al

gating (15–20 cm day–1) varieties for water deeper than 150 cm. The rice in the first category has the greatest potential to increase yield. FPR has three special adaptation mechanisms: (1) elongation of stems and leaves; (2) kneeing ability, which is the upward bending of the terminal parts of the plant; and (3) the ability to develop nodal tillers and roots from upper nodes in the water. Kneeing keeps the reproductive parts above water as the floods subside. The first two traits are accentuated in floating rice such as Leb Mue Nahng 111, Pin Gaew 56, Plai Ngahm, Prachinburi, Leuang Yai, Khao Luang, and Tewada. Nodal tillers arising during the flood period can sometimes compensate for sparse stands.

Elongation abilityThe ability to elongate is an escape mechanism for survival from partial or total submergence. Total plant elongation, including an increase in lengths of leaf blades, leaf sheaths, and stems, may be as much as 20–25 cm in 24 h during initial flooding (Choudhary and Zaman 1970). The stems produce new nodes where leaves are attached and the internode can increase in length when submerged. Elongation of leaf blades and leaf sheaths is important for seedling survival but internode elongation is the most important mechanism for increasing plant length in very deep water (Vergara et al 1975). Internode elongation commenced between 4 and 6 weeks of age for most of 100 lines of FPR growing in rising water (Puckridge et al 1990). The elongation induced after panicle initiation affects the final 4–5 internodes and is not influenced by water depth (Morishima 1975, Bekhasut et al 1990). Elongation stops after flowering in all varieties.

Kneeing abilityKneeing is the bending upward of the upper parts of the culms (stems) as water levels fall and rice plants lodge during the recession of floodwater. When culms lie on the water surface and the upper leaves are held vertical above the water, the plants appear to float even though their base is still attached

to the soil; hence, the common name “floating rice.” Knee-ing keeps the canopy and panicles erect and above the water level. It maintains grain quality by preventing submergence of panicles in water and protects grain from damage by aquatic fauna (Vergara et al 1977, Vergara 1985). Traditional FPR cultivars have good kneeing ability but some modern varie-ties lack this trait.

Submergence toleranceSubmergence tolerance and elongation ability are distinct plant responses that represent opposite mechanisms for flood adaptation. Attempts to combine these two mechanisms in a single rice variety have failed. DWR may be completely submerged if floodwaters rise rapidly, but elongation is usu-ally sufficient to raise part of the foliage above the water level. Photosynthesis can then continue and starch and sugars in emergent leaves and in plant parts under the water are maintained (Setter et al 1987) and further elongation can take place. Submergence tolerance is most useful for nonelongat-ing rice that may be submerged for a few days by rapidly increasing and then decreasing water levels. Puckridge et al (1994) surveyed 87 farmers’ fields in the Central Plain in 1988-89 and found 37 DWR varieties grown. Yield ranged from 0.65 to 4.87 t ha–1, with a mean of 2.13 t ha–1. This compared well with the estimated 2.0 t ha–1 average yield for DWR (Catling et al 1982). In another survey conducted during the 1992-93 cropping season involving 889 farmers from 184 villages in 20 provinces with major DWR areas in north, central, and northeast Thailand, 86 varieties were recorded (Charoendham et al 1995). Only 18% of the farmers grew the five recommended varieties. Local varie-ties were grown by 75% of the farmers, who considered the recommended varieties to be unsuitable to their areas and to have uncertain or low yields of 2.2 t ha–1. Farmers sold 80% of their produce, kept 8% for seed, and used 8% and 4% for family consumption and land rent, respectively. The most desired FPR characteristics were high yield, good elongation, and drought tolerance. The varieties grown matured from early December to mid-January and were 150–200 cm tall, had high tillering, long panicles with good exsertion, good kneeing ability, medium to nonshattering panicles, drought and submergence tolerance, long-slender grain with light hull color, and aroma and good cooking quality, and were nonlodging. Some 98% of the farmers kept their own seeds, with only 7% doing seed multiplication. More recently, Som-mut et al (2004) found 120 rice varieties grown in the wet season, but only 18 in the dry season.

Major constraints of FPR

The major constraints for FPR cultivation are drought, flood-ing, problem soils, and diseases and insects. Drought is the major problem in areas where rainfall is bimodal, with a dry spell occurring between the two peaks (Figs. 1, 2). Damage from drought occurs mainly at the seedling stage. Although FPR can tolerate flooding, either through the mechanism

Table 3. Current cropping patterns (% of land) in the flood-prone ecosystem of Thailand, 2000-01.

Cropping patterna Lowland Lowland Deepwater Deepwaterrainfed irrigated rainfed irrigated

Permanent fallow 1.0 3.0 0.7 0.3Upland crops-fallow 0.0 2.7 1.5 0.8TV rice-fallow 55.4 13.8 58.2 42.8ITV rice-fallow 25.3 7.0 22.9 8.4MV rice-fallow 13.2 46.3 8.5 16.1MV-TV rice 0.0 4.2 4.9 12.8MV rice-MV rice 5.1 20.8 0.7 15.2Other crop-rice 0.0 0.2 8.3 16.4Total 100.0 100.0 100.0 100.0Cropping intensity 105.1 127.4 105.8 131.9

aTV = traditional rice varieties, ITV = improved traditional rice varieties through pure-line selection and/or mutation breeding, MV = modern high-yielding varieties.Source: Sommut et al (2004).


of submergence tolerance or elongating ability, it can oc-casionally be damaged if flash floods occur at the seedling stage. The minimum growth period of FPR must be 30–45 days after emergence. Somehow, prolonged flooding at ma-turity may cause damage. Acid-sulfate soils are a problem in FPR areas mostly in the central-east areas of the country, especially at the early growth stage. The combined effect of drought and acid-sulfate soils makes rice yields in this area very low. Important diseases for FPR are blast (caused by Pyricularia grisea) occuring at the seedling stage as well as bacterial blight (caused by Xanthomonas oryzae) occurring at the flowering stage. Insect pests such as stem borers and brown planthopper are also constraints. Other considerations under the FPR ecosystem include the lack of adapted varieties since FPR varieties are usually location-specific and there is a lack of an efficient extension system to disseminate existing technologies, an insufficiency of good seed of new varieties and high price of good seeds, a lack of attractive rice policies to encourage farmers to grow FPR, and a lack of young FPR scientists, especially breeders, to work on FPR. Despite these constraints, the breeding programs in Thailand have many FPR lines under development, with different maturities and elongation ability, and with resist-ance to pests and diseases. Breeders working in collaboration with IRRI have released new varieties in Thailand, India,

Cambodia, Myanmar, and Indonesia, and other varieties have been exchanged between countries. Many new FPR varieties have performed better than the local varieties in Cambodia, Vietnam, Myanmar, Indonesia, and Africa, and emphasis on testing in the appropriate environments should speed up the development of better yielding varieties.

Collaborative project on FPR breeding

Varietal developmentVarietal improvement of FPR in Thailand dates back to 1916 when the first rice experiment station, Klong Rangsit Rice Experimental Farm, was established. The first step of rice improvement emphasized head selections for grain quality and yield in a large number of collected local varieties. Rice breeding continued until a formal linkage between Thailand and the International Rice Research Institute (IRRI) in 1960 that was further emphasized through the Thai-IRRI Deepwa-ter Rice Collaborative Project on varietal improvement and training started in 1986. During that period, up to 1990, three FPR varieties—RD 17, RD 19, and Huntra 60—suitable for 50–100-cm water depth and incorporating some desirable traits from IRRI lines and Thai traditional varieties were re-leased. Unfortunately, these varieties were not widely adopted by farmers due to the photoperiod insensitivity of RD 17 and high chalky grain of RD 19 and Huntra 60. However, RD

0

30

60

90

120

150

Rainfall (mm) Water depth (mm)

0

50

100

150

200

250

0

30

60

90

120

150


0

50

100

150

200

250

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

0

30

60

90

120

150


0

50

100

150

200

250

0

30

60

90

120

150


0

50

100

150

200

250

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov DecJan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

RainfallWater depth

Sowing

FloweringFlowering

Flowering

Sowing

Sowing

1994 1995

1996 1997

0

30

60

90

120

150Rainfall (mm) Water depth (mm)

0

50

100

150

200

250

0

30

60

90

120


0

50

100

150

200

250


0

30

60

90

120


0

50

100

150

200

250

0

30

60

90

120


0

50

100

150

200

250


FloweringFlowering

Flowering

Sowing

Sowing

Sowing

Sowing

1998 1999

2000 2001

RainfallWater depth

Fig. 1. Daily rainfall distribution and water depth at Prachinburi Rice Research Center, 1994-2001.

32 Sommut et al

19 is the most appropriate prototype for a high-yielding FPR cultivar and is still being used in our breeding program. The FPR collaborative project was strengthened further after the signing of a memorandum of understanding (MOU) between the Ministry of Agriculture and Cooperatives and IRRI on the occasion of the visit to IRRI of Her Royal High-ness Princess Maha Chakri Sirindhorn on 29 August 1991. It called for Thai and IRRI scientists to work together in improving FPR by developing new rice varieties and farming practices for farmers in Thailand and other countries of the region. As a result, a major part of the IRRI FPR breeding pro-gram was transferred to Thailand in 1992. IRRI and Thailand jointly developed and distributed plant breeding materials and associated technologies for improving production in FPR areas of South and Southeast Asia. The Rice Research Center

at Prachinburi (PRRC) was made responsible for improving FPR in the Central Plain of Thailand. PRRC is well staffed and equipped, with a satellite station at Huntra (HTA) that has other screening facilities and a different flooding pattern from that of PRRC. A summary of key elements and activities undertaken under the collaboration is given in Figure 3 and Table 4. IRRI conducted prebreeding research and provided assistance to collaborating NARES. Prebreeding research focused on submergence tolerance and elongation under slow- and fast-rising water. This included development/re-finement of screening methods, providing donors, studying the genetics and mechanisms of tolerance, and mapping and tagging of genes. Sturdy stems were deemed essential to FPR, particularly for medium water depths. IRRI aimed

Variation among years

050

100150200250300

23 S

ep

30 S

ep

7 O

ct

14 O

ct

21 O

ct

28 O

ct

4 N

ov

11 N

ov

18 N

ov

25 N

ov

2 D

ec

Prachantakam

Water depth (cm)

Flowering time

1996

20 J

ul

3 Au

g

17 A

ug

31 A

ug

9 S

ep

23 S

ep

7 O

ct

21 O

ct

4 N

ov

18 N

ov

2 D

ec

Kabinburi Flowering time

Variation among locations

050

100150200250300

Water depth (cm) 1998

28 A

ug

4 S

ep

11 S

ep

18 S

ep

25 S

ep

2 O

ct

9 O

ct

16 O

ct

23 O

ct

30 O

ct

7 N

ov

14 N

ov

21 N

ov

28 N

ov

Prachantakam

1998Flowering time

050

100150200250300

9 Ju

n

30 J

un

14 J

ul

28 J

ul

11 A

ug

25 A

ug

8 S

ep

22 S

ep

6 O

ct

20 O

ct

3 N

ov

17 N

ov

Prachantakam

1999Flowering time

050

100150200250300

7 Au

g

21 A

ug

4 S

ep

18 S

ep

2 O

ct

16 O

ct

30 O

ct

13 N

ov

27 N

ov

11 D

ec

Prachantakam

2000

Flowering time

050

100150200250300

28 A

ug

4 Se

p

11 S

ep

18 S

ep

25 S

ep

2 O

ct

9 O

ct

16 O

ct

23 O

ct

30 O

ct

7 N

ov

14 N

ov

21 N

ov

28 N

ov

Prachantakam

1998

Flowering time

050

100150200250300

15 J

ul

29 J

ul

12 A

ug

26 A

ug

9 Se

p

23 S

ep

7 O

ct

21 O

ct

4 N

ov

18 N

ov

Nakhonnayok

1998

Flowering time

050

100150200250300

23 A

ug

30 A

ug

6 Se

p

13 S

ep

20 S

ep

27 S

ep

4 O

ct

11 O

ct

18 O

ct

25 O

ct

1 N

ov

8 N

ov

15 N

ov

25 N

ov

Inburi

1998

Flowering time

050

100150200250300

Prachantakam15

Aug

29 A

ug

12 S

ep

26 S

ep

10 O

ct

24 O

ct

7 N

ov

21 N

ov

5 D

ec

1997

Flowering time

050

100150200250300

10

Jun

24

Jun

8 J

ul

22

Jul

5 A

ug

19

Aug

2 S

ep

16

Sep

30

Sep

14

Oct

28

Oct

11

Nov

25

Nov

Banna

050

100150200250300

1998

Flowering time

Fig. 2. Distribution of flooding patterns in farmers’ fields of flood-prone rice ecosystem of Thailand during 1996-2000.


at incorporating sturdy stems of some rainfed lowland and upland types into FPR. PRRC intensified its evaluation and selection of breed-ing materials. IRRI assisted in dry-season rapid generation advance, characterization of elite lines, multilocation testing of elite lines through the International Network for the Ge-netic Evaluation of Rice (INGER), and producing F1s that were difficult to handle at PRRC. Though flooding patterns were similar, soil and climatic (temperature and daylength) differences between FPR areas of South and Southeast Asian regions posed a selection problem at PRRC. To alleviate this, F2 populations evaluated in Thailand were selected for advancement at IRRI. Rigid selection is practiced for grain quality traits in Thailand but this procedure tends to discard very high yield-ing genotypes. Selection for grain quality was thus relaxed with two objectives: (1) to isolate lines that could demonstrate the yield potential of FPR and (2) to distribute materials to other countries where grain quality standards were not so high. The overlap between the rainfed lowland and FPR breeding programs at IRRI for the two ecosystems at water depths of around 50 cm was recognized and addressed.

Training, workshops, and visitsIn order to provide knowledge on FPR breeding and to gain useful experience for improving rice yield, training courses were held regularly from 1993 to 1997 at PRRC. The training courses had two main parts, one during crop establishment and the other during the maturity-harvesting period. For each part, both lectures from experts and practical work on-sta-tion and in farmers’ fields were conducted. Each part lasted about 3 weeks. Before conducting the FPR training course, PRRC staff attended a training course on Basic Training Skills Development provided by IRRI and held at PRRC in March 1993. A training course titled Principles of Research in Deepwater and Other FPR Ecosystems that was attended by scientists from 11 countries of South and Southeast Asia and Africa was also held in May-June 1993. Until 1997, training courses and annual meetings were held among breeders who came to PRRC and HTA to select FPR populations suitable for their own countries as well as to present results from their own work. Two rice breeders and a physiologist from PRRC were also able to pursue Ph.D. degrees at IRRI. Two Thai rice breeders also visited Vietnam, Cambodia, Myanmar, India, and Bangladesh to evaluate and select breeding lines that were sent from Thailand. Besides high yield, maturity was

Fig. 3. Key elements in the research program for Thailand’s FPR and linkages to the ir-rigated and rainfed lowland rice ecosystems.

FPR ecosystemResearch on flood and soil constraints to rice production

Zn and Pefficiency

Al toxicitytolerance

Stagnant floodtolerance

Fe toxicitytolerance

Prebreeding research at IRRI and other laboratories

Submergencetolerance

Salinitytolerance

Germplasm improvement research Thailand and India

Exchange breeding materials with NARES

In-country testing (NARES)

Irrigated rice

Socioeconomicpotential

Deepwater rice

Integration

Coastal riceBoro rice

Rainfed lowland rice

Soil and watermanagement

Nutrient andcarbohydrate

research

34 Sommut et al

the most essential character for selection. Ranges of flowering time from 1-30 November and 1-10 December were suitable for Indochina, and Assam, India, whereas the early-flowering group (October) was good for Uttar Pradesh, India.

Exchange of germplasmThough many breeding materials from the collaborative project have been entered in national testing programs, re-leases of FPR varieties have been few. Some were tested for more than 15 years before being released because (1) FPR plants require a long growing period of 6–8 months, of which 2–3 months must be before flooding; (2) different flooding regimes require different times of maturity in each location; (3) FPR varieties fall into two maturity groups according to latitude (Bangladesh and India 23–28oN; Myanmar, Cambo-dia, Thailand, and Vietnam 9–17oN) and the transfer of rice varieties between zones needs a long period of selection for maturity; (4) photoperiod sensitivity ensures flowering at the right time with respect to flooding, but only one generation can be harvested each year; and (5) field tests in the target environments have been very limited. Despite constraints, in 1996, the program exchanged germplasm as follows: (1) early generations of F2–F4 popula-tions (84) were sent to Ghagharahghat, India; (2) fixed and advanced lines were sent to Vietnam (18) and Cambodia (79) with the characteristics of promising lines summarized in Table 5; (3) 11 leading DWR Indian lines were evaluated and used in the FPR breeding program at PRRC but were found unsuitable for FPR areas of Thailand because of earliness; (4)

13 leading DWR and RLR lines/varieties from Vietnam were introduced and used in the Thai DWR breeding program; (5) lines possessing acid-sulfate tolerance from Vietnam were used in the acid-sulfate screening and breeding program of Thailand: Ca Dung Phen, Lua Thong, Nang Coi, Soc Nan, and Trang Hon Bihn; (6) 1,000 early-generation breeding lines were grown at PRRC and HTA for evaluation and selection by rice scientists from counterpart countries for their own target areas; (7) an acid-sulfate screening experiment was conducted in a farmer’s field at Prachinburi under soil pH 3.5 with 200 entries evaluated and the output used by col-

Table 4. FPR breeding: collaborative activities during 1992-97.

ActivityResponsibilitya

Thailand India IRRI

1. Submergence and elongation (fast & slow)a) Screening methods (lab/greenhouse)b) Screening (for donors) — lab/greenhouse — field +++c) Donor characterization ++d) Donor distributione) Genetic mechanismsf) Gene tagging

2. Hybridization +++ +++ ++3. Evaluation

a) Early bulks +++ +++b) RGA +++c) Adv = advanced pedigree +++ +++d) Observation nurseries +++ +++

4. Farm field tests +++ +++5. Elite line characterization (pests, diseases,

submergence, elongation, grain quality)+++

6. Multilocation national tests +++ +++7. INGER evaluation +++8. Seed increase +++ +++9. Release +++ +++

a+++ = high, ++ = complementary.

Table 5. List of promising lines from yield trials in Day Eth, Cambodia, and Vietnam, 1996.

Designation Height (cm) Flowering date

Yield (t ha–1)

Day Eth, CambodiaHTAFR84021-7-1-3 170 12-XII 3.0HTAFR87050-1-9 156 7-XII 3.3Mali Tawng 154 5-XII 3.0PCR87006-8-7 154 5-XII 2.6PCRFR84007-43-B 155 6-XII 2.6PCR89118-8B 154 5-XII 2.7VietnamDWCT82-B-107-7 170 4-XII 3.3HTAFR84035-B-18-7 167 19-XI 2.5HTAFR94096-B-4-8 171 10-XII 3.0HTAFR85035-B-20-5 167 10-XII 3.3


laborators and other countries encountering acid-sulfate soil problems; and (8) valuable germplasm exchanged through INGER in the International Deepwater Rice Observation Nursery (IDRON), International Blast Nursery (IRBN), and International Brown Planthopper Nursery (IRBPH) was sent to PRRC from 1992 to 1998. Out of 316 IDRON lines, one line (WAR87-7-R-10-2-3B-1-2B) was selected for testing in replicated yield trials. A few good lines resistant to BPH and blast such as IRAT2440, IRAT300, BR11, IAC47, and RT1031-69 (Acc. 15092) were used in our breeding program. PRRC and HTA regularly nominated a set of breeding lines to the three nurseries, particularly IDRON. Apart from germplasm exchange, FPR breeding lines from PRRC were sent to rainfed lowland areas of Thailand in order to strengthen and make a more efficient linkage of rice improvement between the FPR and rainfed lowland rice ecosystems. The materials sent included 9 glutinous and nonglutinous entries to the north in 2004, 200 glutinous and nonglutinous entries to the northeast between 1998 and 2004, and 246 nonglutinous entries to the south in 2003.

FPR breeding thrusts at PRRC

Submergence toleranceA simple mass screening method for submergence tolerance using a controlled water level in ponds was developed in Thailand (Boonwite et al 1977, Supapoj et al 1979). Mass screening in ponds is now the most common method for testing submergence tolerance. Seedlings 30 days old are transplanted in the pond and, 30 days after transplanting, the pond is filled to 150-cm water depth for 10 days or until the death of a susceptible check. Submergence tolerance scoring is done immediately after draining water and 14 days later. As far as our work is concerned, all FPR breeding lines, including local varieties, lack the trait of tolerance of complete submergence of 10 days, but the recovery abil-ity of some lines is good, including that for the donor line BKNFR76106-16-0-1-B.

Elongation abilityElongation ability is a stable and heritable trait. The number of internodes and length of culm elongation in FPR appear to be controlled by different genes (Morishima 1975). Supapoj et al (1977) reported that populations from crosses between floating and nonfloating varieties segregated with only a small portion of true floating types and more of intermediate types. Hamamura and Kupkanchanakul (1979) identified partial dominance and the multigenic nature of floating ability. Some researchers suggested multigenic control but were not certain of its degree of dominance. In some crosses, partial dominance was indicated, but in some cases dominance was found. They also suggested that hybrid populations should be flooded for a short period only and that the parents, F1, and F2 be grown in the same environment to avoid variation. Tripathi and Balakrishna Rao (1985) reported that a single dominant gene controlled early nodal differentiation. Each

year, a group of FPR breeding lines, include F2–F6 popula-tions, stable lines, elite lines, and local germplasm, were screened in ponds under target water depth of 90, 110, and 150 cm using 35-day-old seedlings at HTA. Many Thai breed-ing materials, especially local varieties, had good elongating stems above water at 150-cm depth.

Photoperiod sensitivityPhotoperiod sensitivity is an essential characteristic in all FPR areas in Thailand, mostly involving three groups of maturity. First, for FPR in central-east areas, flowering dates must be in early November. Second, in deeper low-lying areas of central-east Thailand, flowering should be at the end of November, synchronized with water recession. Third, for the late-maturity group suited for very deep low-lying central areas of Ayuthaya, Ang Thong, Singh Buri, and Suphan Buri, the appropriate flowering time should be early December. However, photoperiod-insensitive varieties are also desired for favorable areas needing short-duration high-yielding varieties (less than 100 days) for the recent conversion of traditional monocropping of FPR into double cropping of modern varieties in FPR areas.

Type of endosperm and grain qualityNonglutinous endosperm type with acceptable grain charac-teristics (long, slender, clear grain) and eating quality similar to those of Khao Dawk Mali 105 (soft and sticky after cook-ing), Khao Tah Haeng 17, or Leuang Pratew 123 (rather hard and fluffy after cooking) are essential. Currently, due to the linkage of the rainfed lowland and flood-prone ecosystems, glutinous varieties are also needed for northern and northeast Thailand. However, the greatest demand will still be for the nonglutinous type.

Other biotic and abiotic stressesDrought resistance at the vegetative stage is essential as well as adaptability to acid-sulfate soil conditions in central-east Thailand. To improve FPR yield, resistance to major diseases and insect pests should be incorporated into the varieties for release.

FPR breeding activities

Germplasm collection, evaluation, and useThailand is believed to be situated in the center of origin for rice; hence, high genetic diversity is visible. The col-laborative effort initiated in 1982 between Thailand and IRRI with support from the Japanese government collected 23,903 accessions of local varieties and other rice. During 1995-2000, a total of 557 accessions of local varieties were evaluated and characterized at PRRC under natural flood-ing. Table 6 summarizes some quantitative and qualitative characteristics. Collection of local varieties was repeated in 1992 and 2000.

36 Sommut et al

Table 6. Characteristics of 557 local rice varieties evaluated at Prachinburi Rice Research Center in wet season during 1995-2000.

Quantitative characteristics Qualitative characteristics

Characteristic Mean Range Characteristic Description

Height at maturity (cm)

174 ± 18.88 79–246 Culm habit Erect, semierect, open

Flowering date 2 Oct.-4 Jan. Spreading, prostrate

Days to 50% floweringa

167 ± 16.85 92–156 Has awn Absent

Number of tillers per hill

9 ± 2.39 4–21 Leaf blade color

Green, dark green

Number of panicles per hill

8 ± 1.84 3–19 Purple on tip

Number of spikelets per panicle

230 ± 56.75 27–353 Node color Green, purpleGreen with purple

linePanicle length (cm) 26.3 ± 2.23 19.0–33.3 Internode color GreenSpikelet fertility (%) 89.2 ± 4.72 76.4–95.8 Ligule color Green100-grain weight

(g)2.7 ± 0.36 1.3–4.3 Auricle color Green, purple

Grain size Green with purple line

Paddy rice Hull color Straw black, yellow, brown

Length (mm) 9.9 ± 0.67 7.4–11.5 Straw with brown line

Width (mm) 2.7 ± 0.22 2.2–3.8 Straw with brown spot

Thickness (mm) 2.1 ± 0.10 1.4–2.8 Straw with purple spot

Brown rice Brown rice color

White, red

Length (mm) 7.3 ± 0.55 5.4–8.8 Dark purple, light brown

Width (mm) 2.3 ± 0.15 1.8–2.9 Spikelet tip color

Pink, red, straw, purple

Thickness (mm) 1.8 ± 0.10 1.1–2.4 Stigma tip color White, light purpleAmylose content

(%)27.52 ± 5.05 5.56–31.63 Dark purple, yellow

aCounted from when seeded.

Varietal developmentGenetic variation in FPR populations was created by hy-bridization and induced mutation. Realizing the importance of photoperiod sensitivity, elongation ability, as well as sub-mergence tolerance, exotic and domestic donor parents were used. Most of the exotic breeding materials were sent through IRRI. Segregating populations were subjected to abiotic and biotic stress screening and then advanced to yield trials. Screening of segregating populations under natural flooding in farmers’ fields increased confidence in crop survival before entering materials in yield trials. In addition, other supportive information, for example, fertilizer response, reactions to diseases and insect pests, together with grain physical and chemical properties, was recorded. The participation of farm-ers in selection and conducting yield trials was successfully enlisted (Fig. 4) and has been used since 1998. From 1998 to 2004, yield trials were conducted at 6–8 sites every year in

flood-prone areas of Prachinburi, Nakon Nayok, Ayuthaya, Lopburi, Pitsanulok, and Pichit provinces. Maximum water depth was 70–190 cm. Some lines were tested in more than 30 trials and two lines, HTA88095-5B-4 and Pahn Tawng-67, outyielded the check by 18% (Table 7). An irradiated mutant of FR Plai Ngahm Prachinburi with grains more translucent than those of the original variety was obtained. The highlights from breeding FPR tolerant of acid-sul-fate conditions are shown in Table 8. DWCT82-2-2, a com-posite cross line, and pure-line-selected Khao Bahnnah-432 had more translucent grain and better yield than the check varieties. These two lines will be released soon to a target area of about 5,000 ha. Approximate grain yield and some agronomic characteristics together with reactions to major diseases and insects pests of some advanced elite lines from various sources are given in Table 9. The line WAR87-7-R-10-2-3B-1-2B was selected from the IDRON of INGER.


Table 7. Mean grain yield and some agronomic characteristics of flood-prone rice from farmers’ fields in yield trials during 1998-2004, Thailand.

Designation No. of trials Flowering date Height (cm) Grain yield (t ha–1)

Grain length (mm)

Chalkinessa Amyloseb (%)

HTA88026-5B-38 31 28 Nov 212 2.2 6.76 1.38 26.91HTA88026-5B-39 12 27 Nov 192 2.3 7.60 2.00 27.79HTA88037-5B-21 12 30 Nov 214 2.3 7.90 2.20 28.16HTA88095-5B-4 31 24 Nov 223 2.6 7.57 1.64 28.21HTA88110-5B-20 25 24 Nov 234 2.2 7.37 1.51 27.22Khao Luang-28 16 16 Nov 209 2.1 7.49 1.72 23.54Pahn Tawng-67 31 29 Nov 218 2.6 7.04 1.61 26.85PCR90155-3B-10-1 25 22 Nov 225 1.9 7.48 2.16 26.93PCR91038-8-3-3B 25 29 Nov 215 2.1 7.40 1.38 27.73PCR92058-R-1 25 23 Nov 218 2.4 7.60 2.12 27.46PCRC93018-48 8 23 Nov 273 2.4 7.35 0.86 26.78PCRFR84004-17-B 6 27 Nov 262 2.3 7.66 2.22 28.38PCRFR84007-48-B 6 27 Nov 255 2.3 8.03 1.07 27.87PCRFR84009-40-B 6 27 Nov 225 2.3 7.68 1.56 28.57PCRFR84013-43-B 39 26 Nov 242 2.3 7.95 2.38 29.04PCRFR84019-21-B 39 25 Nov 232 2.4 7.82 3.84 28.43PNG’93G2

60Co-17-7 8 4 Dec 296 2.1 7.62 1.49 29.56PNG’93G2

60Co-62-30-B 8 28 Nov 288 2.2 7.64 1.09 29.94Plai Ngahm Prachinburi

(check)39 28 Nov 265 2.2 7.62 2.13 27.11

aScale of 0–9: 0 = none, 1 = small (less than 10% of kernel area), 5 = medium (10–20% of kernel area), 9 = large (more than 20% of kernel area). bAmylose content (%): low = less than 20%, intermediate = 20–25%, high = more than 25%.

Table 8. Yield, flowering date, plant height, and other characteristics of floating rice in farmers’ fields in yield trials under acid-sulfate soil conditions conducted by Prachinburi Rice Research Center during 1998-2004 wet season.

Designation Flowering date Height (cm) Grain yield (t ha–1) Grain length (mm) Chalkinessa Amyloseb (%)

BKNFR82033-2-10-2-5 19 Nov 220 3.2 7.57 3.90 29.00DWCT82-2-2 19 Nov 214 3.1 7.56 2.52 28.29Khao Bahnna-432 12 Nov 223 3.1 7.70 2.60 29.41Khao Bahnna-184 19 Nov 226 2.7 7.46 4.36 28.77Khao Tah Petch-6 16 Nov 225 2.7 7.73 4.35 29.38BKNFR82033-2-10-2-

31621 Nov 225 2.6 7.43 3.86 28.81

BKNFR82033-2-10-2-3-6

18 Nov 241 2.5 7.54 3.98 28.92

BKNFR82033-2-10-2 21 Nov 222 2.5 7.47 2.51 29.00Khao Bahnna (check) 11 Nov 223 2.9 7.77 3.76 28.66Khao Tah Petch (check) 11 Nov 227 2.7 7.59 4.01 29.98aScale of 0–9: 0 = none, 1 = small (less than 10% of kernel area), 5 = medium (10–20% of kernel area), 9 = large (more than 20% of kernel area). bAmylose content (%): low = less than 20%, intermediate = 20–25%, high = more than 25%.

Varietal releaseVarietal improvement of FPR in Thailand began many years before the establishment of the Huntra Rice Experiment Sta-tion (HTA) in 1941. Simple breeding methods were used to collect and purify local varieties. Five recommended varieties were released: Mali Awng, Mali Tawng, Jampah 133, Kawd Pawm, and Khao Med Lek. The FPR varietal improvement program was strengthened in 1950 by using the pure-line selection method. Collecting rice germplasm was handled

by agricultural extension officers and researchers from HTA. The process of on-station and farmers’ field trials was used to evaluate yielding ability. Promising lines with high yield and good grain quality were proposed for release to farmers. In 1959, the FR varieties Nahng Chalong (glutinous rice), Chek Choey, Ta Pao Kaew 161, Leb Mue Nahng 111, and Pin Gaew 56 were released, followed by Khao Nahng Noey 11 and Khao Puang. BKN6986-66-2 and BKN6986-147-2 were released as RD 17 and RD 19, respectively, in 1979. RD 17

38 Sommut et al

Table 9. Approximate grain yield and some agronomic characteristics of flood-prone rice lines currently tested in replicated yield trials at Prachinburi Rice Research Center, Thailand, 2005 wet season.

DesignationFlowering

dateHeight (cm)

Yield(t ha–1)

Reactions toChalkinessb

BPHa GLH Bl/PCR Bl/KHS

GS.21054 21 Oct 178 3.5 S MR VS VS 1PCRC01146 21 Oct 184 4.5 S R VS VS 1–2PCRC01085 23 Oct 177 3.1 MS MR MR R 1GS.4937 24 Oct 181 3.9 MS S VS VS 1GS.4903 25 Oct 174 3.2 MS MS VS VS 1PCR93011-77-10-B 26 Oct 176 3.6 MR MR VS VS 0–1PCR93128-1-2B-1-2B 26 Oct 167 3.3 MR MR VS VS 0–1PCRC01037 28 Oct 166 3.4 MS MS VS MS 2PCR92082-B-3-3-B 29 Oct 172 3.1 MR MS MS VS 2PCR91052-4-B-1-B-1-3B 31 Oct 179 4.0 S MS MS MS 1–2PCR93093-55-1-2-2-2 1 Nov 182 3.3 S MR R S 1–2PCR93186-2-2-1-4B 5 Nov 172 3.6 MS MR VS VS 1–2PCRC01055 5 Nov 166 3.5 S R VS VS 1–2PCR93151-15-2B-2-B 7 Nov 182 3.3 MR R MS MS 1–2PCR92093-14-2-2B-290 10 Nov 187 3.4 S MS VS VS 1PCR93098-7-B-1 10 Nov 180 3.6 MR MS R R 1–2PCR94020-1-2-B 12 Nov 184 3.4 MR MS R MS 1–2SPR’76 Com4-100-1-3-21 12 Nov 181 3.2 MR R R R 1WAR87-7-R-10-2-3B-1-2B 12 Nov 197 3.5 MS MR R MR 1–2PCRC01109 13 Nov 176 3.1 S MS VS VS 0–1GS.20933 16 Nov 185 3.2 S S VS VS 1PCR92093-45-1-2B 16 Nov 179 3.2 S MS R R 1SPR’76 Com4-100-1-3-24 16 Nov 181 3.1 MS R MR MR 1–2HTA60’93G160Co-67-7-250 21 Nov 183 3.3 S S VS VS 0PCRC01001-13 23 Nov 201 3.5 MS S VS S 0–1PCRC01159 23 Nov 171 3.7 MS MR VS VS 1PCRC01001-39 24 Nov 370 3.1 S S S S 0–1GS.93018-48 25 Nov 185 3.3 MR MS VS VS 0–1GS.12386 26 Nov 183 3.7 MS S VS S 0–1GS.15927 28 Nov 178 3.2 S MS VS VS 0–1GS.9368 28 Nov 164 3.1 R S VS VS 0–1Krating Daeng 28 Nov 181 3.3 MR MS VS VS 1PNG’93G160Co-73-45 1 Dec 204 3.6 MS S MS VS 2PNG93G160Co-73-49 5 Dec 180 3.3 MR MS S S 2

aBPH = brown planthopper, Bl = blast disease, Bl/PCR = blast disease tested at Prachinburi Rice Research Center, Bl/KHS = blast disease tested at Royal Study Development Project, GLH = green leafhopper. R = resistant, MR = moderately resistant, S = susceptible, MS = moderately susceptible, VS = very susceptible. bScale of 0–9: 0 = none, 1 = small (less than 10% of kernel area), 5 = medium (10–20% of kernel area), 9 = large (more than 20% of kernel area).

and RD 19 yielded 3.0 and 3.9 t ha–1 and outyielded Pin Gaew 56 by 50% and 100%, respectively. In 1987, SPR7270-18 (Khao Nahng Noey 11/C4-63) was recommended as Huntra 60 for deepwater areas of 50–100-cm water depth, where it yielded as high as RD 19 but had better grain quality. The FPR varietal improvement program during 1992-2004 released five varieties through the Thai-IRRI Collaborative FPR Project (Table 10).

Seed production and distributionThe Rice Research Institute has the mandate to produce breeder and foundation seeds of newly released varieties as well as anticipated breeder seeds of a proposed variety in Thailand. Foundation seeds are sold to the Department of Agricultural Extension’s Center of Seed Multiplication. Only

100–150 tons of foundation seed of the five FPR varieties are produced annually due to budget limitations. Therefore, seed insufficiency occurs every year, which might cause seed mixtures in farmers’ fields. Farmers prefer to buy seeds from the Rice Research Center/Station rather than from the Center of Seed Multiplication. However, they often complain that seed prices are too high.

AchievementsDuring 1992 to 2005, more than 200 Thai breeding lines (with domestic and exotic parentages) were tested in India, Bangladesh, Vietnam, Cambodia, and Indonesia for adapt-ability. Three lines were officially released for and became popular in the deep-flooded areas of Cambodia—Don (HTAFR77022-45-3-2), Tewada, and Khao Tah Petch. Two


Table 10. Description of characteristics of released varieties.

Variety name Parenta Year of release

Flowering date

Yield (t ha–1)

Endosperm typeb

Target areas Distinguishable charactersitics

Plai Ngahm Prachinburi (PNG Prachinburi)

Mass selection from local variety

1994 28 Nov. 2.4 NG Water depth more than 100 cm

l Photoperiod sensitivel Excellent elongating typel Slender grain with rather

high chalkinessl Big panicle, good exsertionl Seed dormancy of 7 weeksl Moderately blast resistantl Good for rice products,

e.g., noodlesPrachinburi 1 (PCR 1) Composite

crosses of 29 F2 populations with IR lines

1998 25 Nov. 3.4 NG Water depth 50–120 cm

l Photoperiod sensitivel Intermediate grain sizel Sturdy culm, big panicle,

good exsertionl Moderately tolerant of

drought, submergence, acid-sulfate soil

l Good weed competition at early stage

Prachinburi 2 (PCR 2) (new plant type)

BKNFR80086*/HTAFR80038**

2002 18-23 Nov.

Shallow water: 5.3

NG Water depth 50–100-cm

l A new plant type, high-yield, flood-prone rice

Deepwater: 3.7

l Photoperiod sensitivel Slender grain with high

chalkiness

continued on next page

a* = derived from Rayada No. 14. ** = derived from IR8234-OT-20. bNG = nonglutinous endosperm.

l Moderately tolerant of submergence, acid-sulfate soil

l Resistant to blastl Good for rice products

Ayuthaya/(AY 1) (new plant type)

U-Ta Pao/ Khao Dawk Mali 105

2004 6-10 Nov.

Shallow water: 5.3

NG 50–100-cm water depth

l A new plant type, high-yield, flood-prone rice

Deepwater: 3.4

l Photoperiod sensitivel Earlier than PCR 1 and

PCR 2l Resistant to BPHl Moderately tolerant of

submergence, acid-sulfate soil

l Good for rice products

Bang Taen SPR60/IR60//IR64 2004 90–95 days (Nov.

to Mar.sown),

110–115 days (sown

May. to Aug.)

4.4 (wet season)5.2 (dry season)

NG l Flood-prone-areas where converted to double-crop of MV pre- and postflood

l With supplemen-tary irrigation

l Photoperiod insensitivel Sturdy, short stature,

100–110 cml Long, slender, and clear

grainl High yieldl High amylose contentl 15–20 days earlier than

current popular MV

40 Sommut et al

Variety name Parent Year of release

Flowering date

Yield (t ha–1) Endosperm type

Target areas Distinguishable characteristics

l Resistant to blast, BPH, and GLH in areas of Prachinburi Nakon Nayok, and Chachoengsazo provinces

l Tolerant of acid-sulfate soil

PCR89151-27-9-155 IR46/Hawm Nai Pon Expected in 2005

25-30 Oct

3.3 (rainfed lowland), 2.8 (deepwater

area)

NG 50–100-cm water depth

l Photoperiod sensitivel Good-quality, flood-prone

ricel Long, slender, clear grain

with aromal Low amylose contentl Sturdy culm

Table 10 continued.

varieties of new plant type DWR were also developed in Thailand, with one released as Prachinburi 2 in 2002 and another as Ayuthaya 1 in 2004. These two varieties performed very well in the flood-prone areas of the eastern plain of Thai-land, with yields of about 3.6 t ha–1 in deepwater areas and 5.0 t ha–1 in shallow-water areas. Another composite cross variety, Prachinburi 1, with excellent weed competitiveness in the early stage as well as good yield at 3.75 t ha–1, was also released in 1998 to replace local varieties in deepwater areas. Plai Ngahm Prachinburi, an FPR variety, was released in 1994. Moreover, a photoperiod-insensitive variety, Bang Taen, was released in 2004 to facilitate the change in cropping intensity in flood-prone areas. This variety has the shortest duration among modern varieties and is also suitable for fully irrigated areas.

Future prospectsThe national plan for rice production during 2004-08 envi-sions improving productivity and quality while maintaining planted area at 9.2 million hectares for the major crop and 1.44 million hectares for the second crop in order to support domestic consumption and keep the export market at about 10 million tons of milled rice annually. Therefore, average rice yields must be increased by 30%, from 2.2 t ha–1 in 2003 to 2.8 t ha–1. In addition, yield of the second crop must also be increased by 18%, from 4.2 t ha–1 to 5.0 t ha–1. To reach these targets, rice research in the FPR ecosystem should (1) strengthen research by a consortium approach and (2) clearly define its research priorities. Forming a consortium would benefit farmers in the 12 million hectares of FPR area in South and Southeast Asia. Despite constraints, this area has high potential for improv-ing rice productivity if scientists work closely together. The exchange of germplasm and technologies across ecosystems and within and across national borders should enhance tech-nology adoption and productivity of farmers. Although there is a declining trend in the growing of FPR and an increase

in the cultivation of other types of rice, research on the FPR ecosystem should be strengthened to retain Thailand’s posi-tion as the world’s best quality rice producer and leading rice exporter.Therefore, total productivity-enhancing initia-tives should focus on a systems approach rather on the FPR plant. For shallow-flooded areas (50 to 100 cm), activities should include (1) development of medium-high-yielding varieties with some essential agronomic traits—grain quality, intermediate height, new plant type DWR traits, submergence tolerance, and low fertilizer responsiveness; (2) improvement of submergence tolerance in glutinous rice for the northeast; and (3) continuation of pure-line selection using traditional varieties, especially for quality and specialty rice (for nutri-tion and processing). For deep-flooded areas (over 100 cm), research priori-ties can include (1) development of shorter duration (95–110 days) varieties with good grain quality and resistance to brown planthopper to facilitate the shift to double cropping; (2) development of appropriate technologies for good crop establishment of the major and second rice crops; (3) con-tinuation of pure-line selection of traditional varieties with the desirable traits of elongation ability, kneeing ability, as well as good grain quality, for both direct consumption and industrial use; and (4) proper water management.

References

Bekhasut P, Puckridge DW, Wiengweera A, Kupkanchanakul T. 1990. Sequential elongation of internodes of deepwater rice at different water depths. Field Crops Res. 24:195-209.

Boonwite C, Setabutara C, Jackson BR, Prechachart C, Anugul P. 1977. Screening for submergence tolerance using a deepwater pond. In: Proceedings of the 1976 Deepwater Rice Workshop, Bangkok, Thailand. Manila, (Philippines): International Rice Research Institute. p 101-108.


Catling HD, Puckridge DW, Boonyaviwatana S, Pattrasudhi R, Kupkan-chanakul T, Tanasate M, Supapornhaemin P. 1982. Deepwater rice crop-cuts for 1981. In: Report of Deepwater Rice Planning Meeting, March 1982. Department of Agriculture, Bangkok, Thailand. p 39-40.

Charoendham P, Palaklang W, Niyomvit L, Prommanart P, Charoen-silp N. 1995. Survey of deepwater rice cultivation in Thailand, 1992/1993. Prachinburi Rice Research Center. (In mimeo.)

Choudhary MA, Zaman SMH. 1970. Deepwater rice of East Pakistan. Paper presented at the International Rice Commission Working Party on Rice Production and Protection, 13th Session, Teheran, Iran. 19 p. (In mimeo.)

Hamamura K, Kupkanchanakul T. 1979. Inheritance of floating ability in rice. Jpn. J. Breed. 29(3):211-216.

Huke RE, Huke EH. 1997. Rice area by type of culture: South and Southeast Asia. Los Baños (Philippines): International Rice Research Institute.

Ministry of Agriculture and Cooperatives. 2004. Agricultural statistics of Thailand, crop year 2003/04. Bangkok, Thailand.

Morishima H. 1975. Floating ability as an adaptive character of the rice and its measuring method. In: Proceedings of the International Seminar on Deepwater Rice, August 1974. Bangladesh Rice Research Institute, Joydebpur, Dacca, Bangladesh. p 109-114.

Puckridge DW, Kongchum M, Thongbai P, Sattarasart A, Sinoupakarn S. 1994. Nitrogen uptake and yield of deepwater rice in the Central Plain of Thailand. Field Crops Res. 37:193-204.

Puckridge DW, Kupkanchanakul K, Peasaschar K, Sripongpankul S, Runtan S. 1990. Influence of time of flooding on development of basal internodes. In: Report of Deepwater Rice Planning Meeting 1990, Department of Agriculture, Bangkok, Thailand.

Setter TL, Waters I, Greenway H, Atwell BJ, Kupkanchanakul T. 1987. Carbohydrate status of terrestrial plants during flooding. In: Crawford RMM, editor. Plant life in aquatic and amphibious habitats. British Ecological Society Symposium No. 5. Oxford (UK): Blackwell Scientific Publications. p 411-433.

Sommut W, Bose LM, Singh VP, Hossain M. 2004. Agriculture and livelihoods in the flood-prone ecosystem in Thailand. Kasetsart J. (Soc. Sci.) 25:69-89.

Supapoj N, Setabutara C, Kupkanchanakul K, Shuwisiku E. 1977. Segregation for elongation ability in two crosses of floating rice with ordinary lowland rice. 1. Estimation of heritability and implication of selection efficiency. In: Proceedings of the 1976 Deepwater Rice Workshop, Bangkok, Thailand. Manila, (Philip-pines): International Rice Research Institute. p 29-36.

Supapoj N, Prechachat C, Kupkanchanakul K. 1979. Screening for flood tolerance in the field. In: Proceedings of the 1978 International Deepwater Rice Workshop, Calcutta, India. Manila (Philip-pines): International Rice Research Institute. p 135-138.

Tripathi RS, Balakrishna Rao M.J. 1985. Inheritance studies of charac-ters associated with floating habit and their linkage relationship in rice. Euphytica 34:875-881.

Vergara BS, Visperas R, Peralta J, Shuwisitkul E, Karin S, Sophonsakul-kaew S. 1977. Screening for kneeing ability. In: Proceedings of the 1976 Deepwater Rice Workshop, Bangkok, Thailand. Manila (Philippines): International Rice Research Institute. p 123-124

Vergara BS, Mazaredo A, De Datta SK, Abilay W. 1975. Plant age and internode elongation in floating rice varieties. In: Proceed-ings of the International Seminar on Deepwater Rice, August 1974. Bangladesh Rice Research Institute, Dacca, Bangladesh. p 178-183.

Vergara BS. 1985. Growth and development of the deepwater rice plant. IRRI Research Paper Series, Number 103. Manila, (Philippines): International Rice Research Institute. 30 p.

Notes

Authors’ address: Senior rice breeder, former senior rice breeder, former director, research assistant, and director, Prachinburi Rice Research Center, Thailand 25150. E-mail: [email protected].

Acknowledgments: Grateful acknowledgement is due to the govern-ment of Thailand and the International Rice Research Institute for useful collaboration on rice improvement. We owe much to the former program leader, Dr. D.W. Puckridge, former project leader, Dr. Drrk HelleRisLambers, and the late Dr. D. Senadhira for their kind effort and supportive collaboration. Our respec-tive thanks go to Dr. M. Hossain for his kind guidance on the socioeconomic aspects as well as Dr. V.P. Singh for his kind support. Special thanks also go to all scientists involved in the FPR collaborative project on germplasm exchange for South and Southeast Asia and to the staff of PRRC and HTA for their helpful assistance.

42 Sommut et al

Breeding rice for submergence-prone and aman areas of India 43

Shallow-flooded (submergence-prone) aman areas of South Asia

44 Mallik et al


The varietal development program for rainfed lowland rice, which occupies 17 million hectares in India, has lately received increased attention. Efforts were intensified only during the early 1990s through the launching of different projects, comple-mentary to the national and state programs, to enhance the productivity of this ecosystem. A shuttle-breeding program, a collaborative project between the Indian Council of Agricultural Research and the International Rice Research Institute, provided an opportunity for enhancing the flow of breeding materials of diverse origin among the eastern Indian states to strengthen the breeding program. The materials developed through this project also served as inputs to other breeding programs for this ecosystem. Many promising cultures having a diverse genetic base (15 to 20 parents from different countries), better plant type, and improved sink components have been developed. They have attained 20% higher yield than the best checks and, we hope, will increase the productivity of this ecosystem.

Breeding rice for submergence-prone and aman areas of IndiaS. Mallik, J. Ahmed, S.K. Bardhan Roy, J.N. Reddy, and G. Atlin

Rice is the most important food crop in India, occupying 44.5 million ha, with a total production of 85.5 million tons and average productivity of 1.9 t ha–1 (Tables 1 and 2). The area under rainfed lowland and deepwater conditions is 17 million ha (Table 2), of which 10 million ha face submergence at different growth stages. These areas are mainly concentrated in eastern India (Table 3). Rainfed lowland rice, accounting for nearly 40% of the rice area in India (Khush and Baenziger 1998), received less attention from the rice research community than it deserved compared with irrigated rice. The first coordinated trial for rainfed lowlands was constituted in 1976 with some pure-line selections from Uttar Pradesh (UP), Bihar, and West Bengal (WB). However, in the annual rice workshop of the All India Coordinated Rice Improvement Project (AICRIP), Hydera-bad, renamed as the Directorate of Rice Research (DRR) under the Indian Council of Agricultural Research (ICAR) in April 1979, it was decided to undertake a systematic multi-locational coordinated program for lowland conditions, after 14 years’ dominance of the semidwarf breeding program for the irrigated ecosystem. Initially, for the first seven years, the Rice Research Station (RRS), Chinsurah, was entrusted with the responsibility of coordinating and monitoring the rainfed lowland rice trials, but the program was taken over by the DRR in 1986 (Mallik 2000). The DRR program has aided actively in the devel-opment and release of 720 rice varieties in the country so far (DRR 2001a,b, 2002, 2003a,b, 2004a,b, 2005, Prasad et al 2001), of which 277 (38.5%) have been for rainfed ecosystems (Table 4). The number of released varieties for the rainfed lowland ecosystem is only 175 (24.3%), among which 129 (17.9%) are for the shallow ecosystem, 31 (4.3%) for the semideepwater ecosystem, and only 15 (2.1%) for

the deepwater ecosystem. In the classification used by DRR, shallow water represents a situation in which standing water generally rises up to 40 cm, for the semideepwater 41–75 cm, and for deep water more than 75 cm. The reasons for the slow progress in varietal develop-ment for the rainfed lowland ecosystem are numerous (Mallik 1995, Mallik et al 1995): l Harsh, heterogeneous, and unpredictable environ-

ment. l Wide variability, both between locations and within

locations over years, with large G × E interaction. l The crop is generally grown once a year, and most

varieties are long-duration/photoperiod-sensitive, thus limiting the advancement of generations to once in a year.

l Advancing the generations (F2 onward) did not use in situ testing in many of the locations.

Table 1. Area, production, and productivity of rice in India in different seasons during 1974-75 to 2000.

Year/season Area (million ha)

Productivity (t ha–1)

Production (million t)

Kharif (wet season)1974-75 35.96 1.0 35.931999-2000 40.67 1.9 76.71Rabi (dry season)1974-75 1.93 1.9 3.651999-2000 4.30 3.0 12.77Total1950-51 30.81 0.7 20.581999-2000 44.97 2.0 89.48

Source: Ratho (2004).

46 Mallik et al

l Inadequate generation of new materials with a wide genetic base.

l Inadequate donors for desirable traits, required for this ecosystem.

l Lack of proper screening methodologies. l Lack of a proper testing facility at many of the re-

search stations. l Lack of proper supporting environmental data such

as water depth, duration, and depth of submergence when presenting trial results by researchers—such as inadequate environmental characterization.

l Fewer researchers involved than with the irrigated ecosystem.

l Less involvement of NGOs, especially in varietal improvement, unlike the hybrid rice development program.

Out of 17 million ha of rainfed lowlands, 10 million ha are submergence-prone, where improvement has been mini-mal. There are two major subecosystems—semideep (41–75

cm of water) and deepwater (>75 cm of water)—and the average productivity of this ecosystem is relatively low (<1 t ha–1) compared with that of other rice-growing ecosystems (Singh 2000). Out of 277 varieties for the rainfed ecosystem released in India, only 46 (6.4% of the total varieties and 16.2% of the varieties released for the rainfed ecosystem) have been released for submergence-prone areas. Four of these, Nalini, Amulya, Purnendu, and Jitendra, were released by the Central Variety Release Committee (CVRC) of ICAR, while the remaining 42 were released by the State Variety Release Committees (SVRC) of different states. During 2000-04, a total of 88 varieties were released in India, of which only two are for the submergence-prone ecosystem. Of 46 varieties released so far, 19 have been developed from single crosses, 5 from three-way crosses, 2 through mutation breeding, and the remaining 20 through pure-line selection from landraces. The varieties thus far developed, except for Bhudeb (Mallik et al 2003), for this ecology have a narrow genetic base as a limited number of parents (2–3, mostly of Indian origin) were used in their breeding ancestry. Conse-quently, these varieties have a limited area of adaptability and are not providing higher yields in this harsh and variable growing environment. This paper summarizes the varietal improvement efforts for this ecosystem.

Programs for enhancing the productivity of rainfed lowland rice in eastern India

Besides the state and national programs, the following projects are either in operation or completed with different objectives for increasing the productivity of rainfed lowland rice.

The Rainfed Lowland ConsortiumThis collaborative project between the International Rice Research Institute (IRRI) and the National Agricultural Re-search Programme (ICAR) started in 1991. The objectives were to expand and strengthen the research base for rainfed rice ecosystems. The emphasis was on strategic and applied research to generate new technologies that will have both local and regional application.

Table 2. Rice ecologies, area, production, and productivity in India (2000-01).

Rice ecologiesArea Production

Yield (t ha–1)Million

ha% Million

tons%

Irrigated 20.5 46 60.0 70 2.9Wet season 16.5 36 46.0 50 2.8Dry season 4.0 9 14.0 20 3.5Upland 6.0 14 5.5 6 0.9Favorable 2.0 5 3.0 4 1.5Drought-prone 4.0 9 2.5 2 0.6Rainfed lowlands (0–50 cm) 13.0 29 16.0 19 1.2Drought-prone 4.0 9 6.0 7 1.5Favorable 3.0 7 6.0 7 2.0Medium deep, waterlogged 3.0 7 2.5 3 0.8Submergence-/flood-prone 3.0 7 1.5 2 0.5Deepwater (>50 cm) 4.0 9 3.0 4 0.8Deepwater 3.0 6 2.5 3 0.8Floating rice 1.0 2 0.5 1 0.5Coastal wetlands 1.0 2 1.0 1 1.0Total 44.5 100 85.5 100 1.9

Source: Singh (2002).

Table 3. Rainfed lowland rice area (million ha) in different subecosystems in east-ern India.

Subecosystem Assam Bihar Orissa Madhya Pradesh

Uttar Pradesh

West Bengal

Total

Lowland (0–30 cm) 0.9 1.7 1.7 2.7 1.9 1.7 10.6Lowland (30–50 cm) 0.5 0.5 0.5 – 0.3 0.5 2.3Deepwater (>50 cm) 0.4 0.4 0.4 – 0.2 0.4 1.8Floating (>100 cm) 0.1 0.7 0.1 – 0.5 0.7 2.1Total (lowland and

deepwater)1.9 3.3 2.7 2.7 2.9 3.3 16.8

Source: Singh (2000).


Initially, Masodha under NDUAT, Faizabad, UP, and Polba-Chinsurah under the Department of Agriculture, government of WB, were the two key sites in India among eight locations in five countries (IRRI 1992). The Central Rice Research Institute (CRRI), Cuttack, was later included as another site for the consortium. The main research focus for Masodha, Polba-Chinsurah, and CRRI was flash flood, stagnant flooding, and mechanization, respectively. Informa-tion regarding environmental characterization, mechanisms of submergence tolerance, and germplasm improvement was generated for the ecosystem (below 50 cm water depth) through this project (Mallik et al 1995). Activities and ac-complishments included l Mechanisms of submergence tolerance character-

ized. l Characterization of floodwater in rice fields. l Testing of early-generation breeding lines in farmers’

fields and participation of farmers in varietal devel-opment and selection through farmer participatory breeding.

l Identification of three photoperiod-sensitive groups for eastern India (second week of October for Chat-tishgarh, third week of October for Bihar, eastern UP, and Assam, and first week of November for Orissa and WB).

l Donors such as Dhulia, Gangasuili, Kusuma, Ra-vana, Luni-form, etc., from local varieties of Orissa and other improved lines such as IR38784-15-19, IR67709-AC57-4, TCA 48 (Vaidehi), and CN 718-21-10-8 (Sudhir), for submergence tolerance, comparable to FR13A.

l Suitable genotypes such as CR 683-123 (Durga), CR 780-1937, IR49745-CPA-42-B-1-5-1-CR1-1, IR54112-B2-1-6-2-2-2-CR2-1, OR 1334-8, OR 1537-6, PSR 1119-13-3 (Kishori), NDR 96002, NDR 960014, NDR SB 9730020, and NDR SB 9730018, etc., for delayed transplanting.

l A plow with a seeding attachment has been found useful by farmers because it increases the plant population and reduces the cost of weeding.

l Opportunities for scientists to participate in meet-ings, workshops, conferences, and symposia for better understanding of the rainfed lowland ecosys-tem.

l Human resource development through training on experimental design and data analysis, G × E analy-sis, etc.

l Infrastructure development.

Eastern India Rainfed Shuttle Breeding ProgramA significant change in the breeding approach in the early 1990s was the use of a shuttle-breeding program. To improve the genetic yield potential of rainfed lowland rice varieties of eastern India, an ICAR-IRRI Collaborative Shuttle Breeding Programme was begun in 1992 mainly for shallow-water (up to 50 cm water depth) conditions. Different cooperating centers are Titabar, North Lakhimpur, and Gerua in Assam; Pusa and Patna in Bihar; Motto and Bhabanipatna in Orissa; Raipur in Chhattisgarh; Masodha in UP; Chinsurah in WB; and CRRI, Cuttack, being the coordinating center. The main objectives of the project are as follows (CRRI 2001d): l To make available diversified donors/improved

breeding lines suitable for rainfed lowlands for each cooperating center.

l To provide segregating populations with a broad genetic background to all centers for effective selec-tion as per location-specific requirements.

l To evaluate elite breeding lines, developed by the cooperating centers and IRRI, especially for sub-mergence tolerance, photoperiod sensitivity, yield potential, and adaptability in eastern India.

l To organize a breeders’ workshop for eastern India to evaluate and select breeding materials at key sites.

Table 4. Varieties released in India for different ecosystems up to 2004.

Ecosystem No. of varieties Percentage National checks

RainfedUpland 95 13.2 Heera, Vandana, Annada, PS Dhan1Shallow 129 17.9 Savithri, Pooja, SalivahanaSemideep 31 4.3 Sabita, PurnenduDeep 15 2.1 Jalmagna, DineshHill 7 1.0 VLDhan221, CH988Total 277 38.5IrrigatedEarly 139 19.3 Heera, Vandana, Annada, PS Dhan1Medium-early 43 6.0 Sasyasree, IR64, RH204Medium 178 24.7 Jaya, KRH2Saline-alkaline 21 2.9 CSR13, CSR27, JayaScented 30 4.2 P. Basmati1, T. BasmatiHill 32 4.4 VL Dhan 61, RP2421Total 443 61.5Grand total 720 100

48 Mallik et al

l To conduct on-farm evaluation of promising culti-vars to study their adaptability and acceptability to farmers.

RRS, Chinsurah, is one of the founding cooperators of this project. Initially, there were only five centers. The number of locations increased as the program became stronger. Gerua in Assam was included as one of the cooperating centers during kharif 2001. Figure 1 depicts the linkage between the station and shuttle-breeding program at RRS, Chinsurah. Parents are selected through either artificial or field screening. Natural calamities such as drought in 1979 and unprecedented floods in 1978 and 2000 in WB also helped to identify donors. The F1s are grown in shallow water. From F2 onward, dry seeds are sown in the field during late April to early May, prior to premonsoon rain. Drought commonly occurrs for about a month or so as the monsoon generally sets in during mid-June. Water starts accumulating in the field from mid-July and increases gradually, depending upon rain-fall and surface runoff from neighboring fields and remains

standing until harvest. As a result, breeding materials during generation advance have been exposed to several abiotic stresses such as drought at early vegetative stage, flooding of varying depth and duration at different growth stages, and biotic stresses such as aquatic weeds, insects, and diseases. Preliminary yield trials are conducted using F6 under trans-planted conditions before replicated yield trials in F7. A segregating population, mainly F2s received from IRRI, is also assessed and advanced following the procedure mentioned above. RRS, Chinsurah, shares its own F1s and F2s with IRRI and other collaborators to facilitate the exchange of breeding materials. It is important to mention that locally adapted traditional and long-duration cultivars were used in IRRI’s breeding program. For example, Sabita, a predominant variety for the rainfed lowland ecosystem in WB, has been used in more than 50 crosses by IRRI (Mallik et al 1999). Exchange of breeding materials (F3/F4) among breeders also takes place at breeders’ workshops at different sites, which

×

Parents are selected through field and artificial screening(Single, double, or three-way cross)

P1 P2

Shuttle-breeding program State and national trials

Grown in shallow water (5–20 cm) to prevent

loss of genotypes

Direct-seeded (40–70 cmdeep). Panicle selection

on the basis of plant type, duration, better grain type, etc.

F3–F5, progeny row, direct-seeded (40–70 cm), selection for abiotic and biotic stresses

Bulk-transplanted and yieldtrial, selection against

diseases and pests along withselection for other desirable

characters

National trial (DRR),AP Cess Fund Scheme

Multiplication and state adaptive trials

Field/on-farm

Release

Promising entry

Nomination(2–3)

Yield trials

F6

F2F2

F3/F4

F1

On-farm

Shuttle-breeding testing network

Sent to collaborating centers and also receive

from them

Sent to IRRI and otherbreeders-collaborators

of shuttle-breedingprogram in

eastern India

Received from IRRIand also sent to IRRI and collaborators

Fig. 1. Linkage between station and shuttle-breeding program at Rice Research Station, Chinsurah.


facilitates the selection of breeding materials that are suited to local conditions (Fig. 2 and Table 5). Promising advanced breeding lines, identified after replicated yield trials, are nominated to national trials (DRR), the A.P. Cess Fund Scheme trial (described below), and shut-tle-breeding trial to consider suitability and the merits of the material. Some 194 materials were nominated to the national program for the rainfed lowland ecosystem during 1995 to 2004 by three centers (Table 6). Through a shuttle-breeding program, several rice varieties have either been released or are in the prerelease stage in different states (Table 7). Kishori and Satyam were released in Bihar and Bhudeb in WB in 2002. OR 1206-25-1 in OUAT, Orissa; NDR 96005(IR66363-10-NDR-1-1-1-1), NDR 8002 (IR67493-M-2), and NDR 96006 (IR67440-15-NDR-1-1-1-1) in UP; Prafulla (TTB 238-3-38-3) in Assam; R 650-1817 and IR42342 in Chhattisgarh; and IR54112-B-1-1-6-2-2-CR-1 and IR49745-CPA-42-B-1-5-1, developed at CRRI in Orissa, are in the prerelease stage (Mohanty et al 2000).

A.P. Cess Fund Research SchemeICAR sanctioned a scheme titled “Enhancement of genetic yield potential of rainfed lowland rice with emphasis on semi-deepwater ecology” starting in 1997. CRRI is the coordinat-ing center with six other centers, one each in Assam (North Lakhimpur, AAU), Bihar (Pusa, RAU), Orissa (Bhubaneswar, OUAT), eastern UP (Masodha, NDUAT), Chhattisgarh (Am-bikapur), and WB (RRS, Chinsurah). The main objectives of

the project are as follows (CRRI 2001a,b,c,d): l Characterization of the target environment. l Collection and evaluation of germplasm for use in

breeding programs. l Genetical, morphological, and physiological studies

on submergence tolerance and adaptation to low light conditions.

l Multilocational evaluation of promising lines in the respective location under the target environment.

l Reorientation of breeding objectives to develop plant types adapted to local conditions.

Some 153 breeding lines of different generations (F3 to F9), contributed by the cooperating centers, were evaluated in six locations in five states—Bhubaneswar, OUAT and CRRI, Cuttack; in Orissa; Pusa in Bihar; North Lakhimpur in Assam; Masodha in UP; and Chinsurah in WB—through the above project, and 2,995 materials were selected during the wet season (kharif) of 2000. Out of 153 breeding mate-rials, 142 nominations were from RRS, Chinsurah (CRRI 2001b). The number of selections made by the individual location was 1,180 at OUAT, 273 at CRRI, 452 at Pusa, 692 at North Lakhimpur, 268 at Masodha, and 130 at Chinsurah. The maximum water depth was 160 cm at Chinsurah, 85 cm at North Lakhimpur, and, for the rest of the locations, it was below 50 cm. The maximum number of selections (45) was made from two crosses: IR67431-CN 7-1 (Biraj/IR53479-B-45-3-2-3) and IR72646-CN 6-1-16 (Banla Phdao/CN 646-6-6//IR40931-33-2-3-2), and these were in the F7 and F5 stage, respectively, during kharif 2000. Biraj and CN 646-6-6,

Motto,Orissa

CRRI,Cuttack,Orissa

RRS,Chinsurah,

West Bengal

North Lakhimpur,

Assam

F2 from IRRI, Philippines

OUAT,Orissa

AAU,Titabar,Assam

RAU, Pusa, Bihar

Raipur,Chhattisgarh

RAU, Patna,Bihar

NDUAT,Masodha,

Uttar Pradesh

F3/F4

F3/F4

F3/F4F3/F4

Fig. 2. Flow of breeding materials to different sites of eastern India through shuttle-breeding program.

50 Mallik et al

developed from RRS, Chinsurah, were one of the parents in the two breeding lines.

On-farm evaluation of deepwater (up to 100 cm water depth) rice varieties and production technologies in the rainfed ecosystem of eastern IndiaStarting in kharif 2001, under the National Agricultural Technology Project (NATP), ICAR has sanctioned a project with five centers in five eastern Indian states: CRRI, Cuttack (Orissa); RRS, Chinsurah (WB); Titabar, AAU (Assam),

Tabl

e 5.

Bre

edin

g m

ater

ials

eva

luat

ed a

nd s

elec

tion

s m

ade

unde

r sh

uttl

e-br

eedi

ng p

rogr

am f

rom

fiv

e ce

nter

s—C

hins

urah

, C

utta

ck,

Rai

pur,

Patn

a, a

nd T

itab

ar—

duri

ng 1

997-

2004

.

Year

F 2F 3

F 4F 5

F 6F 7

Adva

nced

lin

esTo

tal

Evtd

aS

ecEv

tdS

ecEv

tdS

ecEv

tdS

ecEv

tdS

ecEv

tdS

ecEv

tdS

ecEv

tdS

ec

1997

1,67

51,

927

1,95

653

41,

707

593

1,88

876

768

41,

115

102

––

7,92

04,

938

1998

748

1,18

43,

106

654

817

581

890

682

902

539

893

641

––

7,35

64,

281

1999

730

522

889

446

962

419

572

221

444

187

462

3484

2–

4,90

11,

829

2000

229

812

1,18

669

573

275

157

434

127

312

816

211

153

116

83,

687

3,00

620

0126

053

279

838

799

443

585

542

150

018

618

157

669

264

4,25

72,

282

2002

––

800

419

427

279

592

320

281

158

217

104

579

182

2,89

61,

462

2003

106

1,22

280

21,

033

872

595

299

145

1,25

954

436

1559

720

93,

971

3,76

320

0442

91,

807

1,18

599

686

756

845

432

226

711

823

518

521

916

03,

656

4,15

6To

tal

4,17

78,

006

10,7

225,

164

7,37

84,

221

6,12

43,

219

4,61

02,

975

2,19

61,

149

3,43

798

338

,644

25,7

17

a Evt

d =

eva

luat

ed; S

ec =

sel

ecte

d.

Table 6. Number of nominations to different lowland trials in the DRR program.

Trial Masodha CRRI Chinsurah Total

AVT-ShW/AVT-L 6 – – 6IVT-ShW/IVT-L 8 – 11 19AVT-SDW – – 2 2IVT-SDW 9 11 30 50NSDWSN 2 24 59 85AVT-DW – – 12 12IVT-DW – – 20 20Total 25 35 134 194Period 1998-2002 1997-2004 1995-2004

aAVT = advanced variety trial; IVT = initial variety trial; NSDWSN = national semideepwater screening nursery; ShW = shallow; SDW = semideepwater; DW = deepwater; L = late.

Table. 7. Promising materials identified from shut-tle-breeding program in different states.

State Variety/cultivar

Assam Nandang (PSR 1119), Gakapani, and Prafulla (TTB 238-3-38-3)

Bihar Kishori, Satyam, Shakuntala, Rajen-dra Mahsuri-1, and Rajendra Sweta

Orissa Durga, Jagabandhu (OR 1206-25-1), CRLC 899 (IR54112- B2-1-6-2-2-2-CR 2-1), CR 2003-2 (IR67638-CR-15-1-6-1-4), and OR 1898-8-21

Uttar Pradesh NDR 8002 (IR67493-M-2), NDR 96005 (IR66363-10-NDR-1-1-1-1), and NDR 96006 (IR67440-15-NDR-1-1-1-1)

West Bengal Bhudev (CN 1035-61), CN 1230-12-2, IR 72035-CN 32-8, and CN 1231-11-7

Pusa; RAU (Bihar); and Ghagraghat, NDUAT (UP). The main objectives are as follows (CRRI 2001c): l To test promising rice varieties in multilocation trials

with farmers’ participation. l To analyze farmers’ perceptions on adaptability and

acceptability of rice varieties. l To expedite interstate flow of promising rice varie-

ties. l To verify and refine location-specific technology

packages for improvement of rice yield. l To exploit production potential employing improved

technology for realizing maximum yield. Recently released rice varieties for the rainfed lowland ecosystem, from the five states, Sarala and Durga (Orissa); Barhaborodhi and Jallahari (UP); Rajshree and Vaidehi (Bi-har); Ranjit, Bahadur, Padmapani, and Panindra (Assam); and Bhudeb, Mahananda, Ambika, and Hanseshwari (WB), are being tested in farmers’ fields on a large scale through NATP starting in kharif 2001. Five hundred farmers in each state (100 in one district) have been provided with 5 kg of seed


of new varieties to expedite the interstate flow of promising rice varieties. During 2001 to 2003, several varieties were identified as promising in different states (Table 8) through evaluation and they were higher yielding than farmers’ varieties by 10–20%. Varieties Rajshree, Ranjit, Durga, Sarala, Ambika, and Bhudeb performed well in more than one state and were accepted by farmers. One interesting finding was that Mahananda, a variety developed from WB, was accepted by the farmers of only Bihar and was found to be consistently good in all three years. Experiments were also conducted in farmers’ fields to improve productivity through improved agronomic practices. Improved management, which included management of the seedbed (seeding time, seeding rate, application of fertilizer, and seedling age) and maintaining an optimum plant popula-tion in the main field, increased production by 7.4–19.4% in different states (Table 9). Application of fertilizer (N, P, and K) also increased production by 5–20% (Fig. 3) in all the states (CRRI 2002, 2003, 2004).

The national program

The initial effort for varietal improvement for the rainfed lowland ecosystem was mainly confined to the purification of landraces through pure-line selection. Most of the varieties developed up to the mid-1980s were pure-line selections. Due to increased emphasis for this ecosystem by breeders during the mid-1980s, 60 varieties, 42 for shallow, 10 for semideep water, and 8 for deep water, were released during 1991-2000 (Table 10) and 10 varieties were released during 2001-04 (Table 11). The operation of the shuttle-breeding program in the early 1990s was a significant change in breeding approach. Materials with a broad genetic base were available. Exchange of breeding materials among breeders in eastern India pro-vided a better opportunity for selection of genotypes suited to local conditions. Normally, 10–15 years are needed to release a variety for this ecosystem starting from the year of crossing. The shuttle-breeding program helped to reduce the time for varietal development by 3–4 years as breeders received F2 or even F3/F4 populations (Fig. 3). It is expected

that more productive varieties will be released for this harsh and unpredictable ecosystem within the next five years, as they are now in an advanced stage (F7 to F9) at different centers of eastern India. The materials that were nominated by different cent-ers for a semideepwater and deepwater ecosystem and were identified as promising through the national testing program during the last five years are presented in Tables 12 and 13, respectively. The average yield of the promising entries in the semideepwater ecology was more than 3 t ha–1 and they were superior to national, regional, and local checks in most cases (Table 12). The traditional varieties grown in this ecosystem have mostly short bold grains with red kernels, whereas the new promising materials have long to medium grains. In the deepwater ecosystem, the average yield of the promising entries was more than 2 t ha–1 though a few entries, such as CN1234-7-34, CN1230-12-2, and CN1165-6-3, yielded more than 3 t ha–1 (Table 13). They were also superior to national, regional, and local checks.

The state program

Considering the magnitude of the problem and potentialities, efforts for germplasm improvement for the deepwater eco-system started in the early 1970s at the Rice Research Station

Table. 8. Promising varieties identified through varietal evaluation trials during 2001-03.

State 2001 2002 2003

Orissa Durga, Sarala, Rajshree, Bhudeb, and Ranjit

Durga, Sarala, Bhudeb, and Rajshree

Durga, Ambika, and Sabita

Assam Ranjit, Bahadur, Bhudeb, and Barh Avarodhi

Ranjit, Padmapani, Panin-dra, and Bahadur

KDML105, Panindra, and Padmapani

Bihar Rajshree, Mahananda, Durga, and Sarala

Rajshree, Mahananda, Durga, and Vaidehi

Mahananda, Ambika, and Vaidehi

West Bengal Rajshree, Durga, Ranjit, and Bhudeb

Ambika, Ranjit, Bhudeb, and Rajshree

Ambika, Hanseshwari, and Bhudeb

Uttar Pradesh Jal Lahari, Barh Ava-rodhi, Rajshree, and Bhudeb

Barh Avarodhi, Ranjit, Jal Lahari, and Rajshree

Barh Avarodhi, Jalpriya, and Jal Lahari

Table. 9. Performance of agro-management trial in farmers’ fields in different states during kharif (wet) season, 2003.a

State LocationYield (t ha–1) Increase (%) over

T1 T2 T3 T4 T1 T2 T3

Orissa 17 2.05 2.21 2.48 2.66 29.75 20.55 7.38Assam 24 2.85 3.50 3.37 4.01 40.70 14.57 18.99Bihar – 1.49 1.82 1.98 2.25 51.00 23.63 13.64WB 9 2.44 2.67 2.64 2.88 18.03 7.86 9.09UP 40 2.26 2.59 2.77 3.02 22.70 16.37 19.38Mean 2.22 2.56 2.65 2.96 32.44 16.60 13.70aT1 = local variety with traditional practice, T2 = local variety with improved practice, T3 = improved variety with traditional practice, and T4 = improved variety with improved practice.

52 Mallik et al

(RRS), Chinsurah, and were subsequently intensified during the early 1990s through the launching of different projects. Crosses were made to generate new breeding materials with a broad genetic base, and also under a shuttle-breeding pro-gram, the F2s/F3s received from IRRI were exposed to real growing conditions for generation advancement, assessment, and selection, adopting the method suggested by Mallik et al (2002). The segregating generations from F2s were exposed to several abiotic stresses, such as drought, at the vegetative stage, flooding of varying depth and duration at different growth stages, and biotic stresses such as aquatic weeds, insect pests, and diseases in situ. One of the parents used in generating such F2s was generally a locally adapted variety of

3.5

3.0

2.5

2.0

1.5

1.0

0.5

0Orissa Assam Bihar West

BengalUttar

Pradesh

Yield (t ha–1)

Orissa N4O PO KONO PO KO N40 P2O KON4O PO KO N40 P2O K2ON40 P2O KO West BengalN4O P20 K2O NO PO KO Assam N4O P2O K2O NO PO KO Uttar PradeshN4O PO KO NO PO KON40 P20 KO N4O PO KON40 P2O K2O N4O P2O KO Bihar N4O P2O K2O NO PO KO

Fig. 3. Effect of fertilizer application in deepwater ecosystem, 2003.

Table. 10. Number of varieties released during 1991-2000 for the lowland eco-system.

State ShWa SDW DW Total

Assam 11 – 1 12Bihar 1 2 – 3Orissa 10 – – 10Uttar Pradesh – 2 2 4West Bengal 3 6 3 12Other 17 – 2 19Total 42 10 8 60

aShW = shallow, SDW = semideepwater, DW = deepwater.

Table 11. Varieties released for lowland ecosystem during 2001-04.a

Name/designation IET no. Yield range (t ha–1)

Ecosystem Resistance Grain type

Hemavati 13943 2.5–3.0 DW R to BL MSADT-44 14099 5.5–6.0 ShW R to BL, GLH, ShB SBSKL-8 – 5.0–5.5 ShW R to GM, ShB, MR to BL LSC 11-A-41 15358 5.5–6.5 ShW R to BL, GM, LF;

MR to BLB, ShR, ShB

SB

Bapatla Sannalu 16305 5.0–5.6 ShW R to BLB MSTholakari 16672 5.5–6.0 ShW MR to BLB, BPH MSGodavari 16673 5.5–6.0 ShW MR to BLB, BPH MSGiri 14400 4.5–5.0 ShW R to GM LSBhudeb 14496 4.5–5.5 SDW R to BPH, GM,

BLB; MR to ShB, ShR,

MS

Rajendra Mahsuri – 5.5–6.0 ShW MR to BPH, LF, BLB, ShB, BS, ShR

MS

aShW = shallow, SDW = semideepwater, DW = deepwater, R = resistant, MR = moderately resis-tant, BL = blast, GLH = green leafhopper, ShB = sheath blight, BLB = bacterial leaf blight, BPH = brown planthopper, GM = gall midge, ShR = sheath rot, MS = medium slender, SB = short bold, LS = long slender.


Table 12. Promising entries in semideepwater (SDW) ecosystem in national trials dur-ing 1999-2004.

Year Designation Triala IET no.Yield

(t ha–1)Increase (%) overb Grain

typec

NC RC LC

1999 PLA750 AVT1 (SDW) 16209 3.4 11.9 16.5 12.0 MSPLA10934 15514 3.1 4.3 8.6 4.6 LS

2000 OR1535-3 16473 2.7 53.6 32.0 44.5 MBOR1551-6-2 16472 2.4 39.1 19.6 30.9 MB

2001 IR53487-141-3-3-2-1

AVT1 (SDW) 16958 3.3 7.0 5.7 – LS

NDR40053-3-1 16955 3.2 2.5 1.3 – LBCRLC899 AVT2 (SDW) 16481 3.3 3.4 11.5 – LSPLA787 16488 3.1 – 6.1 – LS

2002 OR1234-12-1 AVT1 (SDW) 17318 4.3 32.5 34.5 – SBNDRSB9830121 16913 3.2 – – – LB

2003 OR1234-12-1 17318 3.5 82.6 62.8 20.0 SB2004 NDR9930029 AVT1 (SDW) 17786 2.6 12.1 34.4 24.0 LS

OR1898-8-21 17807 2.5 7.5 28.9 18.9 MSCR978-8-2 17402 2.4 5.5 26.6 16.8 MSMean 3.1 29.1 22.2 21.4

1999 NDR4209 IVT (SDW) 16477 3.2 48.0 57.0 28.0 LSNDR40075 16237 3.2 48.0 56.0 27.0 LB

2000 NDR40053-3-1 16955 2.6 30.3 24.1 22.9 LBCN1057-114 16952 2.1 8.9 3.8 2.8 LS

2001 OR1234-12-1 17318 3.1 31.9 29.2 23.5 MBNDRSB9830121 16913 2.9 25.5 22.8 17.4 LB

2002 CN1231-11-7 17792 5.0 19.6 30.4 14.2 LSOR1550-10-2 17804 4.9 18.5 29.1 13.1 MS

2004 CN1414-1 IVT (SDW) 18130 2.5 17.2 40.8 17.9 LSIR53945-CN35-8-3 18193 2.5 15.2 38.4 15.9 MBCN1230-20-1-16-1 2.3 5.4 26.7 6.1 MSMean 3.1 24.4 32.6 17.2

aAVT = advanced variety trial, IVT = initial variety trial, bNC = national check (Sabita), RC = regional check (Purnendu), and LC = local check (recently released best adapted variety of a particular location). cMS = medium slender, LS = long slender, LB = long bold, MB = medium bold.

Table. 13. Promising entries in deepwater (DW) ecosystem in national trials during 1999-2003.

Year Designation Triala IET no.Yield

(t ha–1)Increase (%) overb Grain

typea

NC RC LC

1999 CN53 AVT (DW) 16115 2.6 15.4 3.5 – LBDWR4107 13493 2.5 11.0 – – LS

2000 CN1414 16847 1.2 36.0 – – LSNDGR448 16850 1.0 6.9 – – SB

2002 CN1234-7-34 17683 3.6 42.1 61.2 28.1 LSCN1230-12-2 17684 3.1 22.3 38.8 10.3 LSCN1165-6-3 17674 3.1 22.0 38.4 10.1 LSIR72035-CN32-8 17686 3.0 18.1 34.0 6.6 LS

2003 IR72035-CN32-8 17686 2.2 58.4 14.8 12.5 LSCN1230-12-2 17684 2.1 54.0 11.7 9.4 LSMean 2.4 28.6 28.9 12.8

aAVT = advanced variety trial, bNC = national check (Jalmagna), RC = regional check (Dinesh), LC = local check (recently released best adapted variety of a particular location). cLB = long bold, SB = short bold.

54 Mallik et al

eastern India such as Sabita, TCA 48, Rajshree, and Mashuri. The breeding materials, including F1s, thus developed were shared with other research centers in eastern India, including IRRI. IRRI used these materials as one of the parents in a breeding program as in IR76210 (CN1205/TCA48), IR76219 (CN1217/IR67632-4-1-1), IR76220 (CN1217/IR68087-55-1), and IR76221 (CN1217/TCA48). The F1 seeds of CN1205 (CRK2-6/CN1035-36) and CN1217 (CRK2-6/CN1035-15) were provided by RRS, Chinsurah. Several promising lines having better plant type and improved sink components have been developed. They were used as parents or nominated to the national program. To estimate yield-attributing traits and sink components, seven promising lines—CN1231-10-7-6-1, CN1231-16-3-1-1 (IR73232: IR57519-PMI-4-1-1-3-1/CN846-6-6//IR58910-202-1-3-2-2), CN1233-31-5-1-1, CN1233-12-1-1 (IR73236: IR58895-PMI-5-1-3-3/Sabita//Rajshree), IR67624-CN6-1-1-1 (IR40931-33-1-3-2/IR51089-65-1-1-3//IR43506-UBN-520-1-3-1-1), IR70242-CN36-10-3-2-1 (Kong Phlouk/IR52555-UBN-3-2-1//Sabita), and IR70418-CN11-26-2-1 (IR66295-71-2/IR67016-45-6-3)—along with three checks, Sabita, Purnendu, and Jitendra, were sown during May 2001 and 2002 at RRS, Chinsurah. Sabita and Purnendu were the national checks for the semideepwater ecosystem, while Jitendra was the only deepwater (>75 cm of water) variety released by CVRC. The two lines, CN 1231 and CN 1233, were selected from IRRI F2s—IR73232 and IR73236, respectively. The mean number of filled grains per panicle was more than 300 in CN1231-16-3-1-1 and IR70418-CN11-26-2-1 compared with 137 in Sabita and 217 in Purnendu (Table 14). All the test cultures had more primary (PB) and secondary (SB) branches in the panicles to support more grains than the checks. Though the fertility percentage was higher in CN1233-31-5-1-1 (92.4%), the number of filled grains per panicle was high-est in CN1231-16-3-1-1 (370), which is more important for higher yield. The mean fertility percentage on PB was higher (85.8) than that of SB (74.2). The test weight (1,000-grain weight) of the cultivars varied from 19.8 g for CN1231-16-3-1-1 to 32.9 g for Sabita. Out of the seven test cultivars, four have long slender grains with an L/B ratio of kernel more than 3 and the other three have medium slender grains. There is an increasing demand for varieties with improved grain quality, preferably hav-ing medium slender to long slender grains, which can fetch premium prices in the market. Therefore, the identified elite lines combining high yield with better grain type are likely to be accepted quickly by farmers and consumers. The new cultivars identified through an evaluation test are semitall (140–145 cm), with moderate tillering ability (3–4 effective tillers per hill), possess stiff straw, and have high yield potential (3.75–4.68 t ha–1), while current deep-water varieties such as Sabita, Purnendu, and Jitendra are tall (170–180 cm), susceptible to lodging, and have low yield potential (2.21–3.38 t ha–1) with inferior sink capacity. The cultivars are being multiplied for nomination to the national

program for testing across diverse growing conditions. They have a diverse genetic base involving 15 to 20 different parents from various countries, such as India, Thai-land, Sri Lanka, Cambodia, Vietnam, and Myanmar. They also possess resistance genes inherited from Oryza nivara and many useful genes for biotic and abiotic stresses and superior grain quality from different landraces.

Conclusions

On-farm testing serves as a venue for visual assessment by farmers and for selecting location-specific varieties as well as a source of widespread dissemination of seeds, which is one of the major constraints to enhancing the productivity of this ecosystem. Results have shown that new varieties coupled with improved management and fertilizer application can improve productivity by 10–20% or even more. IR36 is the most adapted, stable, and successful variety developed at IRRI for the irrigated ecosystem and it is grown widely in many rice-growing countries. Therefore, it is likely that the promising cultivars identified during this investiga-tion with inherent high yield potential due to increased sink capacity, and a diverse and broad genetic base, are expected to express greater adaptability and stability of performance over the heterogeneous, harsh, and unpredictable environ-ments of the submergence-prone ecology. A future breeding program should emphasize the fol-lowing objectives: l Identification of suitable donors l A broad and diverse genetic base l Submergence tolerance l Late planting l Photoperiod sensitivity l Thermo-insensitivity l Strong seed dormancy l Semitall and stiff straw l Higher panicle weight, with higher grain number

(200) l Drought tolerance at early and flowering stages l Tolerance of major pests and diseases l Higher grain yield

References

CRRI (Central Rice Research Institute). 2001a. Annual progress report of the A.P. Cess Fund research scheme for the year 2000-01. CRRI, Cuttack, Orissa, India. p 1-54.

CRRI (Central Rice Research Institute). 2001b. Annual progress re-port of the A.P. Cess Fund research scheme for 2000 for CRRI, Cuttack, Pusa, Masodha, North Lakhimpur, Chinsurah, and Bhubaneswar, India.

CRRI (Central Rice Research Institute). 2001c. Research proposal: on-farm evaluations of deep water rice varieties and produc-tion technologies in rainfed ecosystem of eastern India. CRRI, Cuttack, Orissa, India. p 1-41.


Tabl

e 14

. Yie

ld, y

ield

com

pone

nts,

and

pan

icle

cha

ract

ers

of n

ewly

dev

elop

ed e

ntri

es f

or t

he d

eepw

ater

eco

syst

em.

Cha

ract

ersa

CN

1231

-10

-7-6

-1C

N12

31-

16-3

-1-1

CN

1233

-31

-5-1

-1C

N12

33-

12-1

-1IR

6762

4-C

N6-

1-1-

1IR

7024

2-C

N36

-10-

3-2-

1

IR70

418-

CN

11-2

6-2-

1

Sab

ita

(che

ck)

Jite

ndra

(c

heck

)Pu

rnen

du

(che

ck)

Mea

nC

D

(0.0

5)

Yiel

d (t

ha–

1 )4.

363.

754.

434.

374.

054.

684.

413.

382.

213.

253.

890.

74Ef

fect

ive

tille

rs p

er

plan

t (n

o.)

4.0

3.0

4.0

4.0

4.0

3.0

3.0

6.0

5.0

6.0

4.2

1.11

Fille

d gr

ains

per

pa

nicl

e (n

o.)

261

370

219

297

210

292

313

137

109

217

242.

568

.03

Fert

ility

(%

)78

.85

84.2

892

.40

66.2

956

.30

74.8

779

.64

85.0

990

.83

61.8

277

.04

8.90

PB p

er p

anic

le (

no.)

1515

1416

1515

1612

1114

141.

99Fi

lled

grai

ns o

n PB

(n

o.)

6368

7958

6272

7958

5459

65.2

08.

16

Fert

ility

(%

) of

PB

82.8

990

.67

96.3

466

.67

77.5

085

.71

90.8

092

.06

93.1

081

.94

85.7

66.

76S

B p

er p

anic

le (

no.)

7083

4389

7173

7531

2270

62.7

19.5

Fille

d gr

ains

on

SB

(n

o.)

198

302

140

239

148

220

234

7955

158

177.

363

.02

Fert

ility

(%

) of

SB

77.6

482

.79

90.3

266

.20

50.5

171

.90

76.4

780

.61

88.7

156

.63

74.2

09.

49PB

:SB

1:4.

671:

5.53

1:3.

071:

5.56

1:4.

731:

4.87

1:4.

691:

2.58

1:2

1:5

––

Fille

d gr

ains

on

PB:S

B1:

3.14

1:4.

441:

1.77

1:4.

121:

2.39

1:3.

051:

2.96

1:1.

361:

1.02

1:2.

68–

–Te

st w

eigh

t (g

)25

.80

19.8

031

.40

29.4

029

.10

30.9

020

.50

32.9

032

.50

20.0

027

.23

4.32

Kern

el le

ngth

(m

m)

7.04

5.83

7.46

7.05

7.05

7.32

6.11

7.98

7.60

5.62

6.91

0.67

Kern

el b

read

th (

mm

)2.

082.

122.

242.

402.

342.

362.

112.

242.

332.

092.

230.

21L/

B r

atio

3.38

2.75

3.33

2.94

3.01

3.10

2.90

3.56

3.26

2.84

3.11

0.38

a PB

= p

rimar

y br

anch

, SB

= s

econ

dary

bra

nch,

L =

leng

th, a

nd B

= b

read

th.

CRRI (Central Rice Research Institute). 2001d. Annual progress report for eastern India rainfed lowland shuttle breeding programme for 2000-01. CRRI, Cuttack, Orissa, India. p 1-55.

CRRI (Central Rice Research Institute). 2002. Annual progress report for the project “On-farm evaluations of deep-water rice variet-ies and production technologies in rainfed ecosystem of eastern India, 2001. CRRI, Cuttack, Orissa, India.

CRRI (Central Rice Research Institute). 2003. Annual progress report for the project “On-farm evaluations of deep-water rice variet-ies and production technologies in rainfed ecosystem of eastern India,” 2002. CRRI, Cuttack, Orissa, India.

CRRI (Central Rice Research Institute). 2004. Annual progress report for the project “On-farm evaluations of deep-water rice variet-ies and production technologies in rainfed ecosystem of eastern India,” 2003. CRRI, Cuttack, Orissa, India.

56 Mallik et al

DRR (Directorate of Rice Research). 2001a. High yielding rice va-rieties of India—2000. Bulletin 2001-1. DRR, Rajendranagar, Hyderabad, India. p 1-102.

DRR (Directorate of Rice Research). 2001b. Draft proceedings—36th annual rice group meetings. All India Coordinated Rice Im-provement Programme, Indian Council of Agricultural Research (ICAR), 9-12 April 2001.

DRR (Directorate of Rice Research). 2002. Draft proceedings—37th annual rice group meetings. All India Coordinated Rice Improve-ment Programme, Indian Council of Agricultural Research, 8-10 April 2002.

DRR (Directorate of Rice Research). 2003a. Draft proceedings—38th annual rice group meetings. All India Coordinated Rice Improve-ment Programme, Indian Council of Agricultural Research, 7-10 April 2003.

DRR (Directorate of Rice Research). 2003b. Progress report, 2002. Vol. 1. Varietal improvement. All India Coordinated Rice Improve-ment Programme (ICAR).

DRR (Directorate of Rice Research). 2004a. Draft proceedings—39th annual rice group meetings. All India Coordinated Rice Improve-ment Programme, Indian Council of Agricultural Research, 11-14 April 2004.

DRR (Directorate of Rice Research). 2004b. Progress report, 2003. Vol. 1. Varietal improvement. All India Coordinated Rice Improve-ment Programme (ICAR).

DRR (Directorate of Rice Research). 2005. Draft proceedings—40th annual rice group meetings. All India Coordinated Rice Improve-ment Programme, Indian Council of Agricultural Research, 8-11 April 2005.

IRRI (International Rice Research Institute). 1992. Steering committee meeting–rainfed lowland rice research consortium–reports from sites for 1991. International Rice Research Institute, Manila, Philippines.

Khush GS, Baenziger PS. 1998. Crop improvement: emerging trends in rice and wheat. In: Chopra VL, Singh RB, Varma A, editors. Crop productivity and sustainability–shaping the future. New Delhi (India): Oxford and IBH Publishers. p 113-125.

Mallik S, Kundu C, Banerji C, Chatterjee SD, Ingram KT, Setter TL. 1995. Seedling vigor: screening, physiology, and relationship to submergence tolerance. In: Ingram KT, editor. Rainfed lowland rice: agricultural research for high-risk environments. Manila (Philippines): International Rice Research Institute. p 111-118.

Mallik S.1995. Recent efforts in genotype improvement for rainfed lowland. In: Deb DL, editor. Sustaining crop and animal produc-tivity–the challenge of the decade. New Delhi (India): Associated Publishing Co. p 37-46.

Mallik S. 2000. Rainfed lowland rice research in India: perspectives and future projections. In: Sharma RD, Gahlot P, Gahlot M, editors. Advances in agricultural research in India. Dehradun (India): International Book Distributors. Vol. XIII. p 1-32.

Mallik S, Banerji C, Kundu C, Nayak DK, Chatterjee SD, Nanda PK, Ingram KT, Setter TL. 1995. Rice germplasm evaluation and improvement for stagnant flooding. In: Ingram KT, editor. Rainfed lowland rice: agricultural research for high-risk envi-ronments. Manila (Philippines): International Rice Research Institute. p 97-109.

Mallik S, Kundu C, Mandal BK, Chatterjee SD, Sen SN, Maiti PK, Bose S. 2003. Bhudeb, a new variety for the rainfed lowland ecosystem in eastern India. Int. Rice Res. Notes 28(1):35-36.

Mallik S, Mandal BK, Sen SN, Sarkarung S. 1999. Sabita, a donor for rainfed lowland ecosystem. Oryza 36:149-150.

Mallik S, Mandal BK, Sen SN, Sarkarung S. 2002. Shuttle breeding: an effective tool for rice varietal improvement in rainfed lowland ecosystem in eastern India. Curr. Sci. 83(9):1097-1102.

Mohanty HK, Mallik S, Grover Anil. 2000. Curr. Sci. 78:132-137.Prasad GSV, Prasadarao U, Sobha Rani N, Rao LVS, Pasalu IC, Mu-

ralidharan K. 2001. Indian rice varieties released in countries around the world. Curr. Sci. 80(12):1508-1511.

Ratho SN. 2004. In: Sharma SD, Prasada Rao U, editors. Ge-netic improvement of rice varieties of India. New Delhi (India): Today & Tomorrow Printers and Publ., Part I. p 39-99.

Singh BN. 2002. High yielding rice varieties in India. Rice India (March issue):5-6.

Singh VP. 2000. Sustaining rice farming in eastern India. In: Singh VP, Singh RK, editors. Rainfed rice: a sourcebook of best practices and strategies in eastern India. International Rice Research Institute, Philippines. p 41-45.

Notes

Authors’ addresses: S. Mallik, J. Ahmed, and S.K. Bardhan Roy, Rice Research Station, Chinsurah, 712 102; J.N. Reddy, Central Rice Research Institute, Cuttack, 753 006, India; G. Atlin, International Rice Research Institute, DAPO Box 7777, Metro Manila, Philippines; e-mail: [email protected]; [email protected].

Acknowledgment: The authors gratefully acknowledge the facilities provided by the principal secretary and director of agriculture & ex-officio secretary, Department of Agriculture, government of West Bengal. They also thank ICAR and IRRI for finan-cial support and the collaborators for their cooperation. The help rendered by Dr. S. Islam, assistant entomologist, RRS, Chinsurah, during preparation of the manuscript is also duly acknowledged.

Breeding rice for submergence-prone and aman areas of Bangladesh 57

Composite cross, backcross, and shuttle-breeding programs have been carried out to improve rice varieties for the submer-gence-prone and aman areas of Bangladesh. BRRI dhan33 was released in 1997 and BRRI dhan39 in 1999 as replacements for the early photoperiod-sensitive landraces. BR5226-3-2, a strongly photoperiod-sensitive line, has been isolated for delayed planting up to 15 September coinciding with the recession of floodwater. Also, BR6110-10-1-2 was found to be adaptable to tidal wetland areas up to 60-cm tidal depth. These two lines are being placed in the variety release pipeline. Under a shuttle-breeding program, IR70175-54-1-1-2-3-HR2 showed an advantage of earliness by 10 days and lodging tolerance over the standard varieties BR11 and BRRI dhan32. Also, slow-elongating IR lines have been evaluated in shallow flooded areas, but they showed poor adaptability. The evaluation of IR lines for flash-flood submergence isolated tolerant genotypes but these were prone to lodging. Moreover, 53 landraces were collected from the submergence-prone areas of the country for characterization and identification of new gene sources for submergence. Several breeding lines have been developed through composite crosses having fine grain, improved plant type, and insensitivity to photoperiod. Short-duration near-isogenic lines have been developed using the recurrent parents BR11 and Swarna under a backcross-breeding program. Another backcross program has been progressing in applying the Sub1 gene to develop varieties for flash-flood submergence.

Breeding rice for submergence-prone and aman areas of BangladeshM.A. Salam

Aman refers to two types of rice in Bangladesh: (1) trans-planted aman (T. aman) and (2) broadcast aman (B. aman). T. aman is a synonym for rainfed lowland rice (RLR) and B. aman for deepwater rice (DWR) or floating rice. Submer-gence is a common phenomenon in both production systems from flash flood to complete flooding. Also, RLR suffers from drought stress. This paper therefore intends to review progress in breeding for the drought-prone and flood-prone aman rice production environments of Bangladesh. The development of varieties for unfavorable rice-growing environments such as drought and flood has been an important part of the national program of the Bangladesh Rice Research Institute (BRRI). BRRI joined the Rainfed Lowland Rice Research Consortium (RLRRC) and the International Fund for Agricultural Development (IFAD) program targeting drought-prone and flood-prone environ-ments, respectively. Bangladesh is now a member of the Consortium for Unfavorable Rice Environments (CURE) Working Group 2 (WG 2, flood-prone) and Bundesminister fur Wirtschastliche Zusammenarbeit/Federal Ministry for Economic Cooperation (BMZ) for submergence tolerance. These collaborations have strengthened BRRI’s role in the national programs of Bangladesh.

RLRRC collaboration

Five distinct breeding strategies have been used to improve rice varieties for the drought-prone and flood-prone RLR (T. aman) growing systems of northern Bangladesh.

Strongly photoperiod-sensitive riceThe advantage of this type of rice is its flexibility of seeding from June to August and transplanting with aged seedlings (> 60 days) to cope with early-season drought. The use of modern photoperiod-sensitive genotypes did not show a yield advantage in these conditions due to drought at the reproductive phase. Moreover, strong photoperiod sensitiv-ity is a requirement for flood-prone RLR areas where aged seedlings are transplanted coinciding with the recession of floodwater from late August. Transplanting is recommended up to mid-September in this environment such that flowering is completed in or before mid-November to avoid the pre-vailing low temperature (<18 °C) and to avoid late-season moisture stress. BR5226-3-2 has been identified as a superior genotype over the standard variety BR22 and landraces (Table 1). It is now in the pipeline for release as a variety.

Early photoperiod-sensitive riceThese varieties are usually transplanted in early August and the crop flowers in early October before the onset of drought. Short-duration photoperiod-sensitive genotypes were evalu-

58 Salam

ated and three improved lines were successfully isolated (Table 2). Among these lines, two were released (as BRRI dhan33 and BRRI dhan39) for the drought-prone RLR areas (Table 2, Fig. 1).

Shuttle breedingA resource scientist of the RLRRC for Bangladesh shared 635 breeding lines from Thailand. Screening and evaluation of these lines were done for several years in the drought-prone environment. The major disadvantages of these lines were very broad flag leaf, long growth duration, and susceptibility to leaf diseases. Only a few genotypes showed adaptabil-

ity in the growing conditions of Bangladesh. Nonetheless, IR71075-54-1-1-2-3-HR2 was selected for evaluation in on-farm trials, the last step of the variety release pipeline (Table 3).

Population improvementBoth aromatic and nonaromatic fine-grain rice varieties are popular in the RLR ecosystem of northern Bangladesh. To develop high-yielding fine-grain quality rice, a population improvement tool or composite cross was designed involving BR, IR, Basmati 370, KDML 105, and popular aromatic and nonaromatic fine-grain rice varieties of northern Bangladesh. Seven breeding populations consisting of 5–7 parental crosses were developed. The major problem was the heterogeneity of the breeding populations even in the F10 generation. Above all, two breeding populations, BR6817 and BR6818, showed promising results and several fixed lines were isolated with grain type similar to that of the popular aromatic variety Kataribhog, along with improved plant type and insensitivity to photoperiod (Table 4, Fig. 2).

Backcross breedingBR11 has been the most popular variety of the RLR eco-system throughout Bangladesh. On the other hand, Swarna is popular in the highlands of Rajshahi region. The need is to develop varieties similar to BR11 and Swarna with short growth duration. Because of long growth duration, the crop suffers from drought at the heading stage and seeding of winter crops is delayed. To develop short-duration BR11 and Swarna, backcross breeding was employed involving four donors and four recurrent parents. This program has advanced to the BC5 and near-isogenic lines (NILs) similar to recurrent parents BR11, Swarna, and Dadkhani have been successfully developed with the same grain types along with 7–10 days earlier growth duration (Table 5).

Table 1. Breeding lines in the pipeline for release as varieties for de-layed planting up to mid-September in flood-prone T. aman areas.

Genotype Height (cm)

Date of flowering

Yield (t ha–1)

Remarks

BR5226-3-2 105 14 Nov. 4.7 Medium-bold grainBR22 (standard check) 105 13 Nov. 4.3 Good grainTalosh (local check) 115 16 Nov. 3.5 Prone to lodging

Table 2. Short-duration varieties released for the drought-prone RLR (T. aman) ecosystem.

Line Origin Growth duration (days)

Yield (t ha–1)

Remarks

BG850-2 Sri Lanka 118 4.5 Released in 1997 as BRRI dhan33BR5969-3-2 Bangladesh 122 4.5 Released in 1999 as BRRI dhan39IR33380-7-2-1-3 IRRI 122 4.5 Not released by National Seed

Board

Fig. 1. National Seed Board of Bangladesh released BR5969-3-2 as BRRI dhan39.


IFAD collaboration

Breeding programs concentrated on two types of flood-prone environments: (1) tidal submergence and (2) shallow flooded DWR.

Tidal submergenceThis program was launched for the nonsaline tidal wetlands of southern Bangladesh. Landraces have been grown at flooding depths from 30 to 100 cm. Evaluation of breeding lines at 30–60-cm tidal depths showed the adaptability of several genotypes. Furthermore, BR6110-10-2-1 has been identified as superior to a popularly grown landrace (Table 6). Its adaptability was verified up to 60-cm tidal depth and

Table 3. Advanced lines isolated from the shuttle-breeding program of RLRRC for including in the variety release pipeline, T. aman, 2003-04.

Breeding linePlant

height (cm)Growth duration (days)

Yield (t ha–1) Remarks

IR70175-54-1-1-2-3-HR2 95 130 4.7 Lodging tolerant, good grains (long slender)

BR11 (standard check) 105 145 4.5 Popular variety of T. aman, medium-bold grain

BR32 (standard check) 113 130 4.5 Lodging susceptible, medium-bold grain

Table 4. Breeding lines developed through composite crosses, T. aman, 2003-04.

Breeding population Cross combination No. of promising lines

BR6817 Basmati370/BR1867-20-1//Kataribhog/IR46288-22-6-5-8///Kataribhog/IR46288-22-2-6-5-8// Horinandi/IR33380-7-2-1-3

5

BR6818 BR4970-42-1-2/Basmati 370//IR46288-22-6-5-8/Basmati 370///Kataribhog/IR46288-22-2-6-5-8//Horinandi/IR33380-7-2-1-3

10

Fig. 2. Advanced lines developed with fine-quality grain through composite cross: A—BR6818-25-3-2-3, B—BR6817-5-4-5-2.

it showed yield potential of 4.30 to 4.46 t ha–1. This genotype is now in the pipeline to be released as a variety.

Shallow flooded DWRLong periods of submergence happen to occur due to a usual increase in flood level day by day in the DWR ecosystem. Flood depths and patterns are highly variable within a loca-tion in succeeding years. This causes vulnerability of the crop and hampers the proper evaluation of breeding materi-als. However, the infrastructure, in particular embankments, roads, and road-cum-embankments, has been stabilized at water depth within 1.0 m in some DWR-growing areas. This shallow flooded area of the country was targeted to evaluate 50 improved slow-elongating lines from IRRI. The top five

60 Salam

lines were evaluated in an on-farm trial at 10 locations of the shallow flooded areas. The flooding patterns at trial sites showed a wide variation in depth and longevity of the flood-ing period (Fig. 3). These lines showed some yield advantage over landraces and had maturity duration similar to that of the local varieties (Table 7). The adaptability of these lines is still in question because of their inability to cope with sub-mergence at high rates of increasing flood depth due to slow elongation capacity. Therefore, introgression of submergence tolerance into these lines has been suggested.

BMZ and CURE collaboration

This program centered on breeding rice for flash-flood sub-mergence and medium stagnant water environments.

Shuttle breedingA total of 700 F7 lines and 24 advanced lines from IRRI were evaluated in an on-station trial in T. aman 2003. Thirty-day-old seedlings were transplanted and the crop was submerged for 10 days after 3 days of transplanting. The lines that re-covered from submergence were tall in stature and prone to

Table 5. Generation of NILs after selfing of BC5F1, T. aman, 2004.

Original cross Recurrent parent

Characters

Swarna/BR6389-12-2 Swarna Swarna type but short durationBR11/Minikit BR11 BR11 type but short durationRajshahi1042/BR6398-12-2 BR6398-12-2 Short duration with long slender grainDadkhani/BR6398-12-2 BR6398-12-2 Improved plant type but dadkhani grain

Table 6. Yield and growth duration of BR6110-10-1-2 in variety trials in nonsaline tidal wetlands, T. aman, 2002-03.

Genotype Yield (t ha–1)

Pirozpura Barisal Bakerganj Jhalakathi Mean

BR6110-10-1-2 5.7 5.5 5.8 4.0 5.3(148) (147) (144) (148) (147)

BR11 (check) 3.6 4.6 4.0 3.4 3.9(146) (145) (144) (145) (145)

Moulata (local check) 2.24 2.42 2.19 2.40 2.31(160) (160) (160) (160) (160)

aNumbers within parentheses are days to maturity.

Fig. 3. Flooding patterns in experimental fields, IFAD-DWR trial sites, 2001.

0

10

20

30

40

50

60

70

Date of flooding

Flooding depth (cm)

Maligram GangabordiShingria SuterkandiMazikanda BarokotaVennabari Nagbari

9-XI

-01

12-V

-01

22-V

-01

5-VI

-01

11-V

I-01

17-V

I-01

25-V

I-01

5-VI

I-01

10-V

II-01

15-V

II-01

26-V

III-0

11-

VIII-

016-

VIII-

0111

-VIII

-01

15-V

III-0

122

-VIII

-01

9-VI

II-01

7-IX

-01

17-IX

-01

22-IX

-01

25-IX

-01

5-X-

0114

-X-0

129

-X-0

19-

X-01

5-XI

-01

19-X

I-01

24-X

I-01

9-XI

-01


Table 7. Yield potential of slow-elongating lines at IFAD sites, DWR (B. aman), 2001-02.a

Genotypes Yield (t ha–1)

L1 L2 L3 L4 L5 L6 L7 L8 L9 L10 Mean

IR64588-47-3-2-2B- 12-1-2-3

2.3 2.2 2.1 2.4 2.5 2.4 2.8 4.5 1.7 1.1 2.5

IR64588-47-3-2-2B- 9-2-2-3

2.2 3.0 1.9 1.8 2.7 2.0 2.7 2.5 1.6 1.5 2.3

IR60436-B-65-2 1.5 2.2 1.2 1.2 2.0 2.2 2.8 1.5 1.2 1.5 1.7PCR89114-B-R-2-

2-2-12.0 2.5 1.9 1.0 2.3 2.4 2.7 1.9 1.3 1.5 2.0

IR62653-8-3-3 1.6 1.8 1.4 0.9 2.1 1.9 2.5 1.8 1.4 0.9 1.6Farmers’ variety 1.8a 2.3b 1.8c 1.2d 2.5e 2.7f 1.8g 1.5h 1.2i 1.0j 1.8

aL1 (Maligram), L2 (Shingria), and L3 (Mazikanda) under Faridpur District; L4 (Vennabari), L5 (Gangabordi), and L6 (Suterkandi) under Madaripur District; L7 (Duepara-1) and L8 (Duepara-2) under Tangail District; L9 (Nagbari) and L10 (Barokota) under Comilla District.

a = Laxmidigha, b = Laxmidigha, c = Laldigha, d = Karticksail, e = Gabura, f = Karticksail, g = Chamara,h = Chamara, i = Chamara, and j = Hijoli khama.

Table 8. List of selected genotypes tolerant of flash-flood submergence, T. aman, 2004.

Designation

Seedling height (cm)

Phenotypic score at flowering

Before submergence (30-day-old seedlings)

Aftersubmergence

(15-days’ recovery)

Elongation Survival (%)

IR66036-3B-12-2-B 32 55 23 94 2IR66036-3B-13-2-B 28 48 20 94 2IR7518-B-11-2-B 31 47 16 97 3IR7518-B-11-3-B 31 54 23 96 3IR7518-B1-3-B 34 54 20 97 3IR75407-R-R-R-R-5 34 47 13 94 2IR75407-R-R-R-R-7 34 47 13 93 2IR75407-R-R-R-R-8 28 54 26 88 3IR75407-R-R-R-R-10 33 47 14 94 3IR75407-R-R-R-R-11 33 50 17 94 3FR13A (check) 41 58 17 99 2

lodging at maturity. However, 27 lines were selected with relatively short plant height. These lines were evaluated under controlled submergence in an on-station trial (Table 8). The best lines will be evaluated for yield potential.

Collection of landracesA mission was launched in cooperation with the Depart-ment of Agricultural Extension (DAE) to collect the rice

germplasm presently grown in the submergence-prone and medium stagnant water environments of the country. In total, 53 landraces were collected from 10 districts.

Application of marker-assisted selectionTwo advanced lines were selected from IRRI lines to gener-ate BC2F2 populations for the application of marker-assisted selection (MAS) for the Sub1 gene. IR67518-B-1-2-B and

Table 9. List of BC1F1s to generate BC2F2s for marker-assisted selection, breeding for submergence tolerance, T. aman, 2005.

Original cross Recurrent parent No. of BC1F1 seeds

Target no. of seeds for BC2F1

IR67518-B-1-2-B/BR11 BR11 52 300IR67520-B-14-3-4/BR11 BR11 149 300

62 Salam

IR67520-B-14-3-4 were used, involving Godaheenati and Kurkaruppan-1 as sources of Sub1, respectively. These lines were crossed with BR11, the most popular T. aman variety in Bangladesh. BC1F1s were developed using BR11 as a recurrent parent (Table 9). The development of flash-flood submergence-tolerant rice varieties has been a research priority at BRRI. Efforts of conventional and marker-assisted breeding have been emphasized. The program for introgression of Sub1 into mega-varieties is strengthened to increase the productivity of submergence-prone areas in Bangladesh.

Notes

Author’s address: Chief scientific officer, Plant Breeding Division, BRRI, Bangladesh.

Rice breeding for the tidal wetlands of Indonesia 63

Tidal wetlands/problem-soilecosystems

64 Sulaiman et al


Rice breeding for the tidal wetlands of IndonesiaS. Sulaiman, I. Khairullah, and T. Alihamsyah

Tidal wetlands or tidal swamplands are becoming important land resources for agricultural production and employment opportunity in Indonesia because of the recent increase in human population and the increasing conversion of fertile lands to other nonagricultural purposes. Of the approximately 9 million hectares of tidal swamplands considered suitable for agricultural use, about 4,200,000 hectares have already been reclaimed (Widjaya Adhi et al 1992, Direktorat Bina Rehabilitasi dan Pengembangan Lahan 1995). Some tidal swampy areas have been cultivated for many years with rice and a range of upland crops as well as coconut, coffee, and citrus. Single cropping of rice using traditional photoperiod-sensitive varieties (6–8 months) is commonly practiced by farmers in these areas; however, rice grain yield is very low, usually varying from 1 to 2.5 tons per hectare. Moreover, the adoption of improved high-yielding, early-maturing, and pho-toinsensitive varieties is still very low in this ecosystem. Farmers in tidal swamplands face various problems in developing an efficient rice production system. These problems include poor water control facilities, environmental problems (low soil pH, iron and aluminum toxicity, peat soil, salinity, deepwater flooding, fluctuating water regimes), and socioeconomic problems such as a lack of labor and capital. To support agricultural development in tidal swampland ar-eas, the government of Indonesia has made various efforts, including developing drainage facilities and other physical infrastructure, and funding some research activities (Aliham-syah 2004). Several research institutions worked to develop effective management options for various aspects of this ecosystem such as proper soil, water, and crop management strategies as well as rice varietal improvement. Some research outcomes have also been tested and are being adopted by farmers in these areas. Rice production in tidal swampland areas can be in-creased by practicing proper land reclamation techniques and growing improved tolerant high-yielding varieties. However, the application of land reclamation techniques in tidal swampland is costly and needs various inputs such as soil amelioration material and fertilizer, because of the low soil quality and lack of water control. Therefore, the adop-tion of land reclamation techniques by farmers is progress-ing very slowly. This condition makes the use of improved stress-tolerant and high-yielding varieties a very effective and important way to increase and sustain rice production in tidal

swamplands, because it is cheaper and easier to be adopted by farmers once suitable varieties are developed. This paper presents an overview of the rice-growing environment and cultivation systems, the rice breeding strategy and programs, progress made so far through rice breeding, and the current challenges and future directions for rice breeding for the tidal swamplands of Indonesia.

Rice-growing environments and cropping systems

Rice-growing environmentsRice cultivation in tidal swampland areas is mainly along coastal areas where the sea tides fluctuate through rivers and canals in the fields during part or all of the growing season. The high tides normally cover the highest part of the land. These areas are characterized by shallow inundation in the wet season, caused mainly by stagnant rainwater. Tidal swampland of Indonesia occurs mostly along the coasts or the large rivers of Sumatra, Kalimantan, Sulawesi, and Irian Jaya. The water regime in these areas is dominated by daily tidal fluctuation of the rivers and water depth is influenced by the tide and rainfall. Based on the influence of the tides, tidal swamplands can be divided into four types (Noorsyamsi et al 1984, Widjaya Adhi 1986): (1) the area with tidal type A is flooded by tidal water during spring and neap tides, (2) the area with tidal type B is flooded by tidal water only dur-ing spring tide, (3) the area with tidal type C is not flooded by tidal water but its groundwater table is less than 50 cm deep, and (4) the area with tidal type D is also not affected by tidal water but its groundwater table is more than 50 cm deep. These tidal types are illustrated in Figure 1. Brown and Sulaiman (1984) divided tidal swampland into four major groups based on prevailing abiotic stresses: acid-sulfate land, non-acid-sulfate land, peat land, and saline land. The majority of acid-sulfate lands in Indonesia fall into the soil family Sulfic Tropo Fluaquent and they can be Sul-faquents for the potential acid-sulfate soils and Sulfaquepts for the actual acid-sulfate soils. These soils are also character-ized by pH of less than 4.5, high iron and aluminum contents, and low base saturation, which could cause problems for rice. Non-acid-sulfate lands occur in several soil groups, including extensively cultivated Troposulfaquents, Tropo Fluaquents, and ferrolyzed soils, with soil pH of 4.5 or higher. These soils normally suffer from phosphorus deficiency and low nutri-

66 Sulaiman et al

ent status, which could hinder rice production. Peat soils are characterized by high organic matter or a peat layer of more than 20 cm thickness. In soil taxonomy, this soil is mainly classified as Tropaquepts. It could be peat land (peat layer around 50 cm thick) and peaty land (peat layer <50 cm thick). The major problems of these soils are low base saturation, organic acid toxicities, and poor root anchorage. Organic layers are often underlain by potential acid-sulfate soils, which may also contribute to iron and aluminum toxicities. Saline lands occur mostly in coastal areas that are subject to sea-water intrusion. Similar to the above classification and for practical pur-poses, Widjaya Adhi (1986) grouped tidal swamplands into four land types: deep acid-sulfate lands, shallow acid-sulfate lands, peat lands, and saline lands. Deep acid-sulfate lands are ones having a pyrite (FeS2) layer with a concentration of less than 2% at a depth of more than 50 cm. Shallow acid-sulfate lands are ones having a pyrite layer with a concentration of greater than 2% at a depth of less than 50 cm. Peat lands are lands having a peat layer in their surface horizon with more than 20 cm thickness, whereas saline lands are lands subject to salt-water intrusion for more than 3 consecutive months in a year. Based on land conditions, the main problems for de-veloping rice production systems in tidal swampland areas include high soil acidity and toxic element concentrations, especially iron and aluminum, fluctuating water regimes and a lack of water control facilities, water flooding or submergence during the rainy season, and drought during the dry season. The main problems for practicing double rice-cropping are rat and bird damage for the first crop, poor water control facilities, and limited farmer capital and labor. Research is still needed for developing better varieties and management practices for rice double-cropping systems in tidal swamp-lands.

Disease incidence is normally not serious under the existing single rice-cropping pattern using traditional local rice varieties. However, rats are considered to be one of the most serious pests for rice in many tidal swampy areas. The most important diseases observed in farmers’ fields are brown spot, Cercospora leaf spot, bacterial leaf blight, sheath blight, and blast. Tungro virus occasionally causes moderate damage at a few sites in South Kalimantan. Grassy stunt and ragged stunt viruses are less common than tungro under single rice-cropping, but they should be recognized as potentially serious under a more intensive two-rice-crop production system. Stem borer, rice bug, green leafhopper, and brown planthopper are considered potential problems in these areas. The socioeconomic environment in the tidal swamps also affects the breeding program. Farmers’ incomes are generally low and farmers use little or no external inputs in their rice production systems. Available farm labor and capital are also limited and seasonal. To raise farmers’ income by introducing improved rice varieties will require that the increased yields of the improved varieties be more than the added costs of fertilizer, pesticide, and labor.

Cropping systemsSingle cropping of rice using long-duration varieties is commonly practiced. By using photoperiod-sensitive rice varieties, farmers can expand the period of land preparation or transplanting activity, which helps them to use their own labor resources and reduce the need for hired labor. In addi-tion, local rice varieties do not require high external inputs and have good grain cooking and eating quality with a high price and low risk of failure. When the water level is not too high, early-maturing improved rice varieties can be grown as the first crop, followed by a second planting using local varieties. Only a few farmers practice double- or triple-cropping systems because of poor water control facilities in

Fig. 1. Illustration of tidal types in swampy areas.

Minimum

Marine

Peat

50 cm

Tidalmaximum

IIC B A I B C D

50 cm

Flufiatil

---


most tidal swampland areas and limited labor and capital. Inorganic fertilizers or insecticides are rarely applied. On the other hand, incorporating grassy weeds in the soil during land preparation provides some organic nutrients. Weeds are not a major problem since farmers grow traditional varieties, which are tall and have droopy leaves, besides the effect of flooding, which helps in weed control. According to Alihamsyah (2004), farmers practice four rice cultivation systems in tidal swampland areas: tradi-tional transplanting, improved transplanting, a combination between traditional and improved transplanting, and direct seeding. The traditional transplanting system is practiced by the majority of farmers in tidal swampland areas, espe-cially local farmers such as in Banjarese and Bugise. This transplanting system uses local photoperiod-sensitive rice varieties, which are mostly adapted to local conditions. A high water level during the early vegetative stage requires seedlings that can be transplanted up to 3 times at intervals of 40 to 50 days to produce seedlings that are tall and strong enough to withstand the high water after the third transplant-ing in the main field. Rice seeds at the rate of 5 kg of seed per 150 m2 of land are usually sown on dry seedbeds (locally called teradakan) at the beginning of the wet season (October-November). After 30–40 days, the seedlings are transplanted to the lower parts of the rice field (locally called ampakan). Forty days later, the seedlings are transplanted again on larger areas in the field (locally called lacakan), which still cover only about one-third of the total field to be planted. The time for final transplanting depends on the water level in the main rice field but this is mostly done around 40 days after the second transplanting. Most farmers do not apply any kind of chemical fertilizer and this traditional rice cultivation system seems to be suitable for these tidal swampland situations. The system also has some advantages, such as preventing the oxidation of pyrite, alleviating iron toxicity problems, escaping from seedling submergence, reducing fertilizer application, and spreading the labor requirements for farming activities. Rice yields for this cultivation system are reported to be about 1.5–3 t ha–1. The improved rice transplanting system is mostly prac-ticed by some trans-immigrant farmers in several swampland areas after reclaiming their agricultural lands and applying proper water management techniques, besides using im-proved tolerant rice varieties and cultural practices. Rice seeds at 25–30 kg ha–1 are usually sown on dry seedbeds at the beginning of the rainy season (around October). After 20–25 days, seedlings are transplanted in rice fields previously well prepared using hand hoes or tractors. The improved tolerant rice varieties usually mature within 120 to 140 days, with a grain yield of 4–6 t ha–1. By using this rice transplanting system, the farmers can grow two rice crops each year. The first planting period is October-February or March and the second planting period is April-August. The combination of traditional and improved trans-planting systems (locally called sawitdupa) uses an improved

short-maturing tolerant rice variety and a local photoperiod-sensitive rice variety. Seeds of both varieties are usually sown on dry seedbeds at the beginning of the rainy season (around October). Twenty to 25 days after sowing the improved vari-ety and 40 days after sowing the local variety, the seedlings are transplanted in fields that were previously well prepared using a hoe or a tractor. The area covered by the improved rice variety is about 80% of the rice field and that of the local variety about 20%. After the improved variety is harvested, usually in February or March, seedlings of the local variety are re-transplanted over the whole field. The local rice variety is usually harvested in July or August. Therefore, the crop-ping intensity in this rice cultivation system is 180%. This system allows crop intensification without greatly altering the traditional system, so that it could be easily adopted by farmers. The direct-seeded rice system is usually practiced by some farmers using improved rice varieties in areas having a good water management system or in areas with C and D tidal type (Fig. 1) as well as in shallow freshwater swamp-lands. In this rice cultivation system, rice seeds are sown in well-prepared wet or dry soils. In wetland conditions, the rice seeds are broadcast by hand in the field, whereas in dryland sowing, the seeds are sown in a hole made by a wooden stick. Rice seeds at 25–30 kg ha–1 are usually sown at the beginning of the rainy season (around October) and the crop is harvested in February or March. Fields are kept flooded for up to 4 weeks before harvest, after which the crop is harvested. Farmers usually use an improved transplanting system so that they can grow two rice crops in a year.

Rice breeding for tidal wetlands

Rice breeding strategy and prioritiesThe rice breeding program for tidal swamplands should be intensified in order to support rice production in these highly unexploited lands. Based on the specific local soil problems and farmers’ preferences and constraints, a rice varietal im-provement program should focus on (1) developing improved rice varieties to replace the current long-duration local rice varieties and (2) developing improved rice varieties with better performance than the existing high-yielding early-ma-turity rice varieties, with particular focus on their adaptability to local conditions and acceptability to farmers. The varietal improvement program for tidal swamplands should focus on developing high-yielding varieties that are more adapted to fluctuating water regimes and transient submergence, and the various soil problems, especially high soil acidity and iron and aluminum toxicity. An effective breeding program for these diverse ecosys-tems needs to be target-specific to develop varieties that can tolerate the existing abiotic stresses and meet local farmers’ needs. Of particular importance are crop duration to fit in the existing cropping patterns and local farmers’ preferences. Target ecosystems are grouped by soil type rather than by water regime because water problems are common across all

68 Sulaiman et al

sites. There are four target ecosystems: saline soils, acid-sul-fate soils, non-acid soils, and organic soils. The first priority should be given to acid-sulfate soils and organic soils because of their relatively large area and their occupation mostly by transmigration farmers. In term of duration, priority should be given to early (105–115 days) and intermediate (115–135 days) maturing varieties. These varieties could then be used in double or triple rice-cropping systems, whereas the improved high-yielding photoperiod-sensitive varieties can replace the existing local rice varieties for the single rice-cropping system.

Rice breeding program and implementationDeveloping a suitable variety is an important component for enhancing the productivity of the swampland rice farming system. The varietal improvement program for tidal swamp-land is being handled by breeders at the Indonesian Center for Rice Research (ICRR) in Sukamandi, West Java, and is mostly focused on developing high-yielding and early- or intermediate-maturity rice varieties with resistance to pests as well as iron toxicity and soil acidity. Breeding of early-maturity rice varieties with higher yield and better adaptation to tidal swampland conditions has been handled by research-ers at the Research Institute for Food Crops in Swamp Areas since 1993. The breeding program aimed to develop acceptable rice varieties that meet the need of local farmers and consumers, especially in Central and South Kalimantan. Therefore, the adoption of those rice varieties by farmers in tidal swamp areas is expected to be faster. Breeding varieties for tidal swamplands is being conducted under the Indonesian Agency for Agricultural Research and Development and this involves two institu-tions, the Research Institute for Food Crops in Swamp Areas (Banjarbaru, South Kalimantan) and the Indonesian Center for Rice Research (Sukamandi, West Java). Some research activities are being handled in collaboration with the Inter-national Rice Research Institute, especially in the exchange of germplasm for testing in target areas. The following traits should be considered in developing rice varieties for tidal swampland rice-farming systems: l Moderately tall varieties (90–130 cm) that are leafy

to reduce the labor required for weeding. l Perform well under low fertilizer input. l Easily threshed, without awns and with good panicle

exsertion. l Have some amount of submergence tolerance be-

cause the crop is occasionally submerged, especially at the seedling stage.

l Resistance to common rice diseases and insects. l Tolerance of drought for varieties planted as a second

crop in the dry season. l Tolerance of adverse soil conditions and locally

adapted. This may be sought in crosses with tradi-tional local rice varieties.

l Have small slender grain (local variety type) or medium slender grain (IR66 type).

l Good eating and cooking quality with high amylose content similar to that of local varieties to suit the taste of local people, and intermediate amylose content for transmigration farmers.

Achievements in rice breeding

Released varietiesGood progress has been achieved in developing varieties adapted to this ecosystem. Between 1981 and 2003, the government of Indonesia, through the Department of Agri-culture, released 18 improved rice varieties adapted to tidal swampland conditions (Table 1). However, most of these varieties are only tolerant of certain problems encountered in these areas, such as high soil acidity and iron toxicity under good management practices. Most of the improved rice varieties also have low or medium texture and cooking quality, with large grain size, and are suited only for certain farmers or consumers. Some of those rice varieties, such as Kapuas, Lematang, Lalan, Batanghari, and Banyuasin, have been grown by farmers but in limited areas. The slow adoption of those improved rice varieties was attributed to insufficient seed availability, higher input and labor require-ments, and inadequate quality to meet farmers’ preferences. For example, Banjarese farmers, especially in Central and South Kalimantan, prefer a tall plant with long and slender grain and medium rice texture or cooking quality. Margasari and Martapura are two rice varieties devel-oped from crosses between the local rice variety Siam Unus with Cisokan and Siam Unus with Dodokan, respectively, in order to meet the needs of local farmers and consumers, especially in Central and South Kalimantan. The major characteristics of those rice varieties (Table 2) are similar to those of local rice varieties, such as slender and long grain with high amylose content (27–28%) or medium texture, moderately tall plants (90–130 cm), moderately leafy plants, and tolerance of iron toxicity. These characters seem to match the requirements of the local farmers and consumers. These varieties are expected to be easily adopted by farmers, with a yield potential of 3–4 t ha–1, and shorter duration of 120–125 days compared with the yield of local varieties of 1–3 t ha–1 and a much longer duration of 6 months. In 2002, the local extension personnel together with farmers successfully tested a double-cropping system using Margasari and Martapura as the first crop on about 400 ha of rice fields in Barito Kuala District, South Kalimantan. The farmers obtained a rice yield of about 4 t ha–1. Tidal swampland areas planted with those varieties are progressively increasing, especially in the south and central parts of Kalimantan. It is now estimated that more than 1,000 ha of swampland area in Central and South Ka-limantan are being planted with Margasari and Martapura.

Promising rice lines for adverse soil conditionsAs mentioned in an earlier section, the adverse soil chemi-cal characteristics that constrain rice cultivation in tidal swamplands include a high concentration of soluble iron


and aluminum, salinity, and soil acidity. New rice varieties should therefore have tolerance of those adverse soil char-acteristics. From a series of breeding activities conducted at the Bogor Research Institute for Food Crops (BORIF), some promising rice lines adapted to tidal swampland conditions were identified, particularly with tolerance of soil acidity and iron and aluminum toxicity. For the deep acid-sulfate

swamplands of South Sumatera, three promising rice lines were developed with yield significantly higher than that of Lematang (Tumarlan et al 2000): IR48948-B-2-Mr-1, IR51471-2B-2-2B-Mr-1, and B-655f-KA-141-6, with mean grain yield of 4.5, 4.0, and 3.8 t ha–1, respectively, whereas Lematang yielded only 3.0 t ha–1. Rice breeding activities at RIFSA or RISA (Banjarbaru) mostly focus on developing photosensitive rice varieties for acid-sulfate and peat tidal swamplands. From a series of breeding activities in Kalimantan, breeders at RIFSA or RISA (Banjarbaru) identified some promising rice breeding lines adapted to tidal swampland conditions, particularly with tolerance of soil acidity and iron toxicity. Eight promising rice lines were identified: GH47, GH137, GH149, GH173, GH460, GH493, GH505, and GH591. However, these eight lines are susceptible to brown planthopper biotype 2, moderately resistant to blast disease (except GH460), and moderately susceptible to moderately resistant to sheath blight disease (Table 3). They have intermediate plant height, with long slender grains, high amylose content, medium rice texture, and tolerance of Fe toxicity, so they are considered suitable for the local conditions of South and Central Kali-mantan. For the shallow acid-sulfate soils of Belandean and Oanda Jaya, most of the 15 lines tested were found to perform well, with tolerance of iron toxicity. However, of these lines, only four produced higher grain yield than Margasari and IR64 (Table 4). Those are GH173, IR53709-36-10-2, TOx3118-B-E2-3-2, and B10179B-MR-1-4-1. In

Table 1. Improved rice varieties adapted to swampland released during 1981–2001.a

NameYear

releasedMaturity (days)

Yield (t ha–1)

Rice texture

Fe toxicity

Resistant to

BPH SB BS BL BB PB

Barito 1981 140–145 3 Loose – R-1 MR – – MR –Mahakam 1 1983 135–140 3–4 Loose – S-123 MR – – MR –Kapuas 1 1984 127 4–5 Medium R R-1 MR – – MR –Musi 1988 135–140 4–5 Loose – R-2 R – R R –SeiLilin 1991 115–125 4–6 Loose – MR-2 – – MS – –Lematang 1991 125–130 4–6 Loose – R-1 – – MR – –Lalan 1997 125–130 4–6 Loose – R-123 – – R – –Banyuasin 1997 115–120 4–6 Sticky MR R-3 – R R – –Batanghari 1999 125 4–6 Loose R R-12 R – R R –Dendang 1999 125 3–5 Sticky R R-12 – MR MR – –Indragiri 2000 117 4.5–

5.5Medium R R-2 R – R R –

Punggur 2000 117 4.5–5 Medium R R-23 – – R – –Margasari 1 2000 120–125 3–4 Medium R MR-2 – – R – RMartapura 1 2000 120–125 3–4 Medium R MS-2 – – R – MRAir Tenggulang 2001 125 5 Loose R R-123 R – R R –Siak Raya 2001 125 5 Loose R R-2 – R R – –Lambur 2001 120 4 Sticky R R-2 – – R – –Mendawak 2001 115 4 Sticky R MR–3 – – MR – –

aResults by Research Institute for Food Crops in Swamp Areas, Banjarbaru. bR = resistant, MR = moderately resistant, MS = moderately susceptible, S = susceptible, BPH = brown planthopper, SB = sheath blight, BL = blast, BS = brown spot, BB = bacterial blight, PB = panicle blast.Source: Lesmana et al (2002).

Table 2. Some characteristics of Margasari and Martapura.

Characteristic Margasari Martapura

Age at maturity 120–125 days 120–125 days Plant type Moderately leafy Moderately leafy Plant height 120–130 cm 120–130 cm Tillering ability Medium Medium Brown rice shape Slender-small Slender-small Panicle threshability Easy Easy Panicle exsertion Well exserted Well exserted Lodging resistance Moderately resistant Moderately resistant 1,000-grain weight 21 g 22 g Amylose content 27% 28% Cooking-rice structure Medium Medium Eating quality Good Good Grain yield 3–4 t ha–1 3–4 t ha–1 Resistance to Bph2 Moderately resistant Moderately resistant Sheath blight Moderately susceptible Moderately resistant Panicle blast Resistant Moderately resistant Fe toxicity Tolerant Tolerant

70 Sulaiman et al

deep acid-sulfate soils of Handil Manarap, no iron toxicity was observed, and comparable rice yields were given by GH47, GH149, GH460, BW307-6, and B10277B-MR-1-4-3. Results of field trials in three tidal swampland sites showed that only four rice lines have high tolerance of acid-sulfate soils and grain yield of more than 3 t ha–1 (Table 5). Those lines are GH47, GH137, TOx3118B-E-2-3-2, and IR58511-4B-4. Rice lines GH47 and GH137 are not only tolerant of high soil iron concentration and acidity, but they also have

long slender grains and medium rice texture, which match local farmers’ preferences. Data on genotype by environment interaction from trials conducted in some tidal swampland sites of South Sumatera and Central and South Kalimantan showed a significant inter-action between rice genotype and environment. This indicated that some of the promising lines were well adapted only in certain locations or soil conditions. Soil pH at Kalimantan locations is lower (pH 3.3–3.8) than at Sumatera locations (4.6–4.8), and this affects crop performance. Based on the average yield of the six locations, five promising rice lines produced grain yield at least 10% higher than the yield of Margasari (Table 6). Those lines are GH47, TOx3118B-E-2-3-2, IR61242-3B-B-2, IR58511-4B-4, and B9852E-35-KA-66. IR61242-3B-B-2 was the only line that yielded more than 4 t ha–1. However, GH47 was considered to be the most stable high-yielding rice line. Based on farmers’ preferences, GH47 was selected by farmers at three locations in Kalimantan and one location in Sumatera, while GH137 was selected by farmers at three locations in Kalimantan and two locations in Sumatera. IR61242-3B-B-2 was selected at three locations (Sumatera). TOx3118B-E2-3-2 and IR58511-4B-4 were selected by farmers at only one location in Sumatera. GH47 and GH137 have intermediate plant height and small-slender grain type, while TOx3118-B-E2-3-2, IR61242-3B-B-2, and IR58511-4B-4 are semidwarf with long-slender grain type. GH137 was selected at five locations because of its high yield but lower stability. This indicates that farmers’ preferences for different lines vary across locations.

Promising rice lines with submergence toleranceFlooding or submergence occurs in tidal swamplands with tidal type A and B along the rivers. Until 2003, breeding to enhance submergence tolerance for tidal swamplands of Indonesia was very limited, and not much information was available. A submergence trial was conducted in an artificial pond at RISA, Banjarbaru, during the 2003-04 rainy season. Twenty-one-day-old seedlings were submerged for three different durations of 1, 2, and 3 weeks with 125-cm water

Table 3. Characteristics of promising rice lines selected for tidal swamplands. Data are from RIFSA.

Code of lines

Plant height (cm)

Maturity (days)

Brown rice Amylose content (%) Fe toxicity

Resistancea to pests

Length L/W ratio Bph2 SB PB Tungro

GH47 126 120 M Slender 28 T – – – MSGH137 112 118 M Slender 28 T S MR R SGH149 116 123 M Slender 27 T S MS MR SGH173 124 122 M Slender 27 T S MR MR SGH460 114 123 M Slender 27 T S MR S SGH493 110 124 M Slender 29 T S MS MR MSGH505 113 123 M Slender 29 T S MS R SGH591 116 121 M Slender 28 T S MR S S

aR = resistant, MR = moderately resistant, MS = moderately susceptible, S = susceptible, T = tolerant, M = medium.Source: RISA (2002).

Table 4. Yield and Fe toxicity scores of rice lines in shallow acid-sul-fate soils of Belandean and Oanda Jaya, South Kalimantan, 2002.a

LineBelandean Oanda Jaya

Yield (t ha–1)

Fe toxicity score

Yield (t ha–1)

Fe toxicity score

GH47 (KaI9407d-Bj-18-2) 3.0 1 3.0 cde 1GH137 (KaI9408d-Bj-28-4) 3.1 2 3.8 def 1GH149 (KaI9408d-Bj-38-1) 3.0 2 3.1 abc 2GH173 (KaI9408d-Bj-70-4) 3.8* 1 4.0 efg 1GH460 (KaI9414d-Bj-110-

1) 3.3 2 3.1 abc 3

GH493 (KaI9420d-Bj-6-2) 2.9 2 3.6 cde 3GH505 (KaI9420d-Bj-14-1) 3.4 2 3.5 bc 3GH591 (KaI9420d-Bj-11

0-2) 2.8 3 3.4 bc 2

IR53709-36-10-2 3.5 3 4.4 fg 3BW307-6 3.3 3 3.3 bcd 3B10278B-MR-3-3-1 2.7 3 3.4 bc 3B10179B-MR-1-4-2 2.6 5 2.9 ab 3B 1 0277B-MR-1-4-3 2.6 5 2.6 a 3TOX3118B-E-2-3-2 3.1 3 5.0 h 2B10179B-MR-1-4-1 3.2 3 4.4 fg 2Margasari (check) 3.0 1 3.6 cde 3IR64 (check) 2.3* 5 2.9 ab 4Kapuas (check) 2.8 1 4.4 9 3

ans = nonsignificant, * = significant at the P = 0.05 level. Means followed by the same letters are not significantly different according to Duncan’s multiple range test.Source: RISA (2003).


depth. A total of 70 rice genotypes were tested in this trial, of which 40 were from IRRI, 24 from RISA, and 5 from improved varieties. The results showed that after 1 week of submergence, 24 genotypes showed 100% survival, 20 genotypes showed 92% survival, and 14 genotypes showed 85% survival. However, after 2 weeks of submergence, only 3 rice genotypes showed 100% survival, 10 genotypes showed 92% survival, and another 10 showed 85% survival. The three rice genotypes are IR66036-3B-13-2-B, IR70215-2-CPA-2-1-B-1-2, and IR73047-6-1-1-1-B-2-B. All rice seedlings died after 3 weeks of submergence. Two similar trials were also conducted in an artificial pond at RISA, Banjarbaru, during the 2004 dry season using

clear water and during the 2004-05 rainy season using turbid water. Twenty-one-day-old seedlings of 16 rice lines from IRRI, including IR66 and Tapus as check varieties, were transplanted in rows with 28 plants per entry. One week after transplanting, all plants were completely submerged with a water depth of about 125 cm for durations of 1, 2, and 3 weeks in clear water and 6, 12, and 18 days in turbid water. Survival was counted 1 week after submergence. After 1 week of submergence in clear water, 11 rice genotypes showed 100% survival, but, after 2 weeks of submergence, only 3 rice genotypes had 100% survival (Table 7). These three rice genotypes are IR73047-6-1-1-1-B-2-B, IR70215-2-CPA-2-1-B-1-2, and Tapus. After 6 days of submergence in turbid

Table 5. The performance of rice genotypes in shallow acid-sulfate land of Bari-to Kuala, South Kalimantan, 2003.

Rice genotypesBelandean Terantang Oanda Jaya

Yield (t ha–1)

Fe toxicity score

Yield (t ha–1)

Fe toxicity score

Yield (t ha–1)

Fe toxicity score

GH47 4.38 3 3.64 2 3.67 2GH137 3.70 2 2.88 2 3.58 2GH173 2.45 4 3.22 3 4.21 3GH460 2.95 4 4.08 4 3.60 4TOX3118B-E-2-3-2 3.55 4 3.12 2 3.69 4IR58511-4B-4 4.11 4 3.15 3 4.50 4Mendawak 3.72 4 4.70 2 4.47 4Margasari 2.95 3 3.07 2 3.73 2CV (%) 4.7 16.0 9.9Soil pH 3.52 3.57 3.93Fe content (ppm) 297 206 153 Source: RISA (2004).

Table 6. Yield of rice breeding lines in acid-sulfate soils of Belandean and Unit Tatas, in peat soils of Pinang Ha-bang, Kalimantan, in acid land of Jambi and Karang Agung I, and in peat soils of Karang Agung II, Sumatera.

GenotypeYield (t ha–1)

Belandean U. Tatas P. Habang Jambi K. Agung K. Agung II Mean W2a

GH47 (KaI9407d-Bj-18-2) 3.16 abc 3.05 bcd 3.14 4.09 b 4.37 ab 3.06 cd 3.48 0.0755GH137(KaI9408d-Bj-28-4) 3.13 abc 3.36 cd 3.17 2.59 a 4.22 ab 3.18 cd 3.28 2.3777GH460(KaI9414d-Bj-110-1) 3.30 be 3.12 bcd 3.02 2.72 a 3.90 a 1.78 a 2.98 1.8158GH505(KaI9420d-Bj-14-1) 3.27abc 3.13 bcd 2.81 2.92 a 4.11 ab 2.63 be 3.15 1.0341BW307-6 3.17 abc 3.32 cd 2.88 4.62 bcd 4.33 ab 2.36 ab 3.41 0.5772Tox3118b-E-2-3-2 3.05 ab 3.17 bcd 3.16 5.34 d 4.14 ab 2.56 be 3.50 1.6008B10179b-Mr-1-4-1 2.81 a 2.99 be 3.10 4.15 be 4.21 ab 2.17 ab 3.24 0.3474IR61242-3B-B-2 3.62 3.57 d 3.44 5.15 d 5.20 d 3.59 d 4.09 0.4131IR58511-4B-4 3.55 c 2.32 a 3.21 5.24 d 4.56 bc 3.43 d 3.72 1.9711B9852E-35-KA-66 3.53 be 2.76 ab 3.23 4.84 cd 4.84 cd 3.61 d 3.79 0.8795Margasari (control) 3.18 abc 3.11 bcd 2.85 3.79 b 2.09 ab 3.19 3.19 0.3270Mendawak (control) 3.21 abc 3.02 bc 3.37 4.16 be 3.15 cd 3.57 3.57 0.1322CV (%) 7.6 9.2 11.7 11.8 13.3Soil pH (H20) 3.33 3.50 3.8 4.69 4.48Fe content (ppm) 527.0 1098.98 153.79 42.50 1157.33

aW2 = stability index of “Wricke’s ecovalence.” Means followed by the same letters are not significantly different according to Duncan’s multiple range test.Source: RISA (2003).

72 Sulaiman et al

water, only 2 rice genotypes (IR70213-10CPA-2-3-2-1 and IR70215-2-CPA-2-1-B-1-2) had 100% survival, and, after 12 days of submergence, only IR70215-2CPA-2-1-B-1-2 showed 100% survival. Genotypes IR70213-10-CPA-4-2-1-1-3, IR66036-3B-13-2-B, and IR70213-9-CPA-12-UBN-2-1-3-1 were considered moderately tolerant.

Collection of rice cultivarsTo support breeding research activities and conserve rice cultivars and genetic resources in tidal swamplands, re-searchers, especially at RIFSA or RISA, Banjarbaru, South

Table 7. Percentage survival of rice genotypes after submergence in clear water (DS 2004) and in turbid water for different periods (WS 2004-05).a

GenotypeClear-water survival (%)

Muddy-water survival (%)

1 week 2 weeks 6 days 12 days 18 days

IR69502-6-SRN-3-UBN-1-B-1-2 85 (5) 46 (9) 46.4 (9) 0 0 IR69502-6-SRN-3-UBN-1-B-1-3 100 (1) 85 (5) 35.7 (9) 3.6 0 IR70181-5-PMI-1-2-B-1 100 (1) 62 (7) 85.7 (5) 35.7 3.6 IR70213-9-CPA-12-UBN-2-1-3-1 100 (1) 77 (5) 92.9 (5) 67.9 17.9 IR70213-10-CPA-2-3-2-1 100 (1) 69 (7) 100 (1) 85.9 85.7 IR68835-44-8-B-B-4-1 100 (1) 62 (7) 92.9 (5) 53.6 7.1 IR70181-32-PMI-1-1-5-1 100 (1) 69 (7) 96.4 (1) 82.1 78.6 IR66036-3B-13-2-B 100 (1) 85 (5) 96.4 (1) 75.0 39.3 IR70213-10-CPA-4-2-1-1-3 100 (1) 92 (5) 96.4 (1) 92.9 82.1 IR70215-2-CPA-2-1-B-1-2 100 (1) 100 (1) 100 (1) 100 39.3 WAR115-1-2-1-3-6-B-B-2 85 (5) 8 (9) 92.9 (5) 96.4 0 WAR115-1-2-4-2-4-B-B-4 62 (7) 15 (9) 64.3 (5) 64.3 0 IR73047-6-1-1-1-B-2-B 100 (1) 100 (1) 96.4 (1) 46.4 3.6 GH137 62 (7) 46 (9) – – –TOX3118B-E-2-3-2 69 (7) 46 (9) – – –IR66 92 (5) 8 (9) – – –Tapus (check) 100 (1) 100 (1) 82.1 3.6 0

aNumbers in parentheses = % comparative survival score (CS) = (% survival of entry/% survival of con-trol) × 100%. CS 100 = 1; CS 75–94 = 5; CS 50–74 = 7; CS 0–49 = 9.Source: RISA (2005).

Table 8. Insect and disease resistance of some local rice varieties collected from tidal swamplands in South Kalimantan.a

Variety Location Brownspot

Sheath blight

Leaf blast

Neck blast

Stem borer

Bph1

Runtai Hulu Sungai MR MS MS HS S MS Bayar Pahit Sei Tabuk MR MS MS S S MS Siam Unus Sei Tabuk MR MS MR MR S MS Siam Arjan Kalumpang R S R MS S MS Siam Sanah Sungai Raya MR S MR S S MS Ketan Siam Sei Tabuk S MS S R S MS Runut Simpur S MS S MR S MS Siam Bam-

ban Simpur MS MS R MS S S

Siam Cinta Simpur S MS R MS S MS

aR = resistant, MR = moderately resistant, HS = highly susceptible, MS = moderately sus-ceptible, S = susceptible.Sources: Balittan Banjarbaru (1995) and Mukhlis and Imberan (1998).

Kalimantan, collected, characterized, and conserved a large number of local rice cultivars. By 2001, about 239 lines were collected, out of which 107 cultivars were selected, consisting of 26 cultivars for deep acid-sulfate lands, 63 for shallow acid-sulfate lands, 6 for peat lands, and 12 for freshwater swamplands (RIFSA 2001). Screening of 58 lo-cal rice varieties was done for brown planthopper resistance (Bph1), stem borer, brown spot, sheath blight, leaf blast, and neck blast. Most of them were susceptible to these insects and diseases (Table 8). Prayudi et al (2002) reported ten local rice varieties in tidal swamplands of South Kalimantan that


are susceptible to tungro disease: Siam Unus, Pandak, Siam Raden, Lemo, Bayar Pahit, Siam Perak, Siam Teladan, Siam Adil, Kandangdukuh, and Siam Putih. In 2001 and 2002, researchers at RISA collected an additional 182 local rice cultivars from the swamplands in South and Central Kalimantan as well as from Lampung and South Sumatera. Most of the local rice varieties are photoperiod-sensitive, with yields varying from 1 to 4 t ha–1, and they require 4 to 9 months to reach maturity. The local rice varieties have well-exserted panicles, are awnless, and have intermediate plant height, 7–17 panicles per hill, 17–23 grams per 1,000 grains, slender grain type, and medium to short brown rice length. Seventy-one rice cultivars had been characterized for their agronomic and growth characters (RISA 2003). Some of them, such as Bayar Palas, Pandak Putih, Siam Unus, and Lemo Putih, were found to be relatively resistant to lodging, whereas Siam Puntal, Siam Adus, Siam Suruk, and Siam Lantik were resistant to tungro. Six rice varieties were found to have iron content of more than 66 ppm, and only Siam Pandak and Siam Wol had higher iron content of 83 and 70 ppm, respectively. Fifteen rice cultivars had zinc content of more than 72 ppm; one of them, Siam Panangah, had zinc content of 108 ppm.

Challenges and future prospects for rice breeding

Challenges for rice breeding for tidal wetlandsDeveloping adapted high-yielding varieties is an important component for enhancing and sustaining productivity in swampland rice-farming systems. Eighteen improved rice varieties for tidal swamplands had been released and some promising rice lines had been identified, but their adop-tion by farmers was still very slow. This is probably due to several reasons, including high external input requirements, low farmers’ or consumers’ preference, limited adaptation to only one or a few existing problems, and greater demand for labor, land, and water management. On the other hand, local rice varieties commonly grown in tidal swampland areas have wider adaptation to these conditions and still perform well with low external inputs, but they have low yield and long duration. These situations could be challenging for rice breeding programs to support the increasing demand for rice production in tidal swamplands. Farmers in tidal swamplands benefited very little from modern agricultural systems, including improved rice varie-ties. Most of these farmers still grow traditional local rice varieties using traditional production systems. Strong and sustained efforts are needed to improve the rice production systems and increase rice production in tidal swampland areas. These efforts include considerable investments in research to develop adapted high-yielding rice varieties with acceptable grain quality, together with proper management practices. In most instances, highly site-specific management practices are required because of the specificity of these stresses, and suitable varieties for such specific situations

become the key entry point for developing new manage-ment practices. The eating and cooking qualities of the new rice varieties should also be given a high priority. Since research facilities and funds are limited, well-coordinated and collaborative research activities, involving both regional and international networks, are needed to optimize the use of available resources. Collaboration and coordination be-tween breeders and other scientists, extension workers, and policymakers are needed to develop a cirtical mass for such concerted efforts. Some of the future challenges for varietal improvement for tidal swamplands include 1. Developing better varieties for the major cropping

systems and environments. 2. Continuously evaluating and improving the breeding

objectives, priorities, and strategies. 3. Forming stronger linkages with other institutions

and agencies in Indonesia and abroad to improve the testing and extension of improved tidal swampland rice varieties.

4. Improved facilities and well-trained breeders to handle the expanding research activities.

Future prospects for rice breeding for tidal wetlands

Rice breeding programs for tidal swamplands should be intensified in order to support the increasing demand for rice production in these areas, and to resolve the existing challenges. Future rice breeding programs for these areas must be directed not only to solve the existing problems of rice production systems but also to fulfill farmers’ require-ments and preferences. Therefore, the aim of rice varietal improvement programs for tidal swamplands should be to (1) develop high-yielding photoperiod-sensitive rice varieties to replace the existing long-season local rice varieties and (2) improve existing early-maturity high-yielding rice varieties, particularly for their adaptability or tolerance of specific local problems, yield potential, and grain quality. With limited research facilities and resources, more coordinated and collaborative efforts are needed, at both the national and international level. Moreover, careful analysis and prioritization are needed to develop more focused and efficient research plans to tackle these complex issues. Mod-ern rice breeding techniques such as biotechnology, mutation breeding, and molecular breeding should be explored together with conventional methods in order to streamline breeding programs and speed up the development of suitable new rice varieties.

References

Alihamsyah L. 2004. Agricultural research and development on swampland or flood-prone ecosystem in Indonesia. Paper pre-sented at the International Workshop on Flood-Prone Rice in Asia. Cantho, Vietnam, 8-9 February 2004.

74 Sulaiman et al

Balittan Banjarbaru. 1995. Laporan Hasil Penelitian. Balai Penelitian Tanaman Pangan, Banjarbaru, Indonesia 1994/95. 554 p. (In Indonesian.)

Brown KD, Sulaiman S. 1984. Present state and future challenges of the BARIF Indonesia swamp rice breeding. I. Tidal swamps. Indonesia Agric. Res. Dev. J. 6(1 & 2):1-9.

Direktorat Bina Rehabilitasi dan Pengembangan Lahan. 1995. Luas penggunaan lahan rawa pasang surut, lebak, polder dan rawa lainnya di tujuh propinsi. Dirjen Tanaman Pangan dan Hortikul-tura. Departemen Pertanian, Jakarta. (In Indonesian.)

Lesmana OS, Toha HM, Las I. 2002. Deskripsi varietas unggul baru padi. Research Institute for Rice (RIR), Sukamandi, Indone-sia.

Mukhlis M, Imberan. 1998. Pengujian ketahanan varietas padi lokal terhadap penyakit bias. Kalimantan Agrikultural, IImiah Fak. Pertanian Lambung Mangkurat 5(1):12-15.

Noorsyamsi H, Anwarhan H, Sulaiman S, Beachell HM. 1984. Rice cultivation of the tidal swamps of Kalimantan. In: Workshop on Research Priorities in Tidal Swamp Rice. Los Baños (Philippi-nes): International Rice Research Institute. p 17-28.

Prayudi B, Widiarto IN, Sulaiman S. 2002. Ketahanan genotipe padi rawa pasang surut terhadap penyakit tungro. Oalam Prosiding Seminar Nasional Pertanian Lahan Kering dan Lahan Rawa. Balai Penelitian Pertanian Lahan Rawa. p 41-48. (In Indone-sian.)

RIFSA (Research Institute for Food Crops in Swamp Area). 2001. Em-pat Puluh Tahun Balittra: Perkembangan dan Program Penelitian Ke Depan. Banjarbaru, South Kalimantan, Indonesia.

RISA (Research Institute for Swampland Agriculture). 2002. Annual report for 2001. Banjarbaru, South Kalimantan, Indonesia.




Tumarlan IW, Suastika, Nasution B. 2000. Uji daya hasil padi di lahan pasang surut Karang Agung Ulu, Sumatera Selatan. Proceedings of National Seminar on Agricultural Research and Development in Swampland. Cipayung, 25-27 July 2000, Indonesia.

Widjaya Adhi IPG. 1986. Pengelolaan lahan pasang surut dan lebak. Jurnal Litbang Pertanian V(1), Januari 1986. Badan Litbang Pertanian.

Widjaya Adhi IPG, Nugroho K, Ardi DS, Karamadan AS. 1992. Sumber daya Lahan Pasang Surut, Rawa, dan Pantai : Potensi, Keterbatasan dan Pemanfaatan. Oalam proceedings Pertemuan Nasional Pengembangan Lahan Pertanian Pasang Surut dan Rawa. Cisarua, 3-4 Maret 1992.

Notes

Authors’ address: S. Sulaiman, I. Khairullah, and T. Alihamsyah, Plant breeder, assistant plant breeder, and director, respectively, of Research Institute for Upland Agriculture, Banjarbaru, South Kalimantan, Indonesia.

Rice breeding for acid-sulfate soils in Vietnam 75

Rice breeding for acid-sulfate soils in VietnamBui Chi Buu and Nguyen Thi Lang

To feed the current population (81 million) of Vietnam, our farmers cultivate nearly 4.3 million ha of rice (total agri-cultural area of 7.4–7.6 million ha). Rice land is declining because of urbanization, industrialization, and other reasons to about 4.0 million ha. In addition, water shortage and salt intrusion will become serious problems and large challenges for human food security, even though Vietnam is currently a net rice exporter. Rice is still the most important crop as it supplies about 67% of the calorie intake in the country. Rice production was about 15.7 million tons in 1985 but doubled to about 35.9 million tons in 2004. Current average yield is about 4.87 t ha–1 (MARD 2005). Acid-sulfate soils account for 40.8% and 5.6% and alluvial soils for 30.1% and 48.5%, respectively, in the Mekong Delta and the Red River Delta. Clearly, acid-sulfate soils have become the main constraint to rice production in the Mekong Delta (Table 1).

Three main problems in acid-sulfate soils have been targeted: (1) aluminum toxicity, (2) iron toxicity, and (3) phosphorus deficiency. The Cuu Long Rice Research Institute (CLRRI) has focused its research on the first two problems. For iron-toxicity tolerance, we initiated a crossing program to develop breeding populations for QTL mapping and selec-tion. The problems and constraints encountered in these soils vary across ecosystems, which require different breeding and management strategies to solve particular problems. Cur-rently, suitable water management and agronomic practices have been recommended and rice varietal improvement is also considered a key approach for enhancing yield on these soils.

Tolerance of aluminum toxicity

The exploitation of genes from wild rice relatives has been successful in introducing many useful genes into the culti-vated gene pool. A good example is the development of the high-yielding variety AS996 (IR64/Oryza rufipogon), which is tolerant of aluminum toxicity, is of short duration, and is adapted to acid-sulfate soils. Major QTLs on chromosome 3 were identified that are found to be associated with aluminum toxicity (Fig. 1). This is achieved through the development and analysis of a recombinant inbred line (RIL) population developed from the cross of IR64/O. rufipogon. QTLs as-sociated with control root length (CRL), stress root length (SRL), and relative root length (RRL) were mapped (Nguyen et al 2002).

Fig. 1. Comparison of the QTL identified on chromosome 3 in rice with that of wheat, barley, and rye (Nguyen et al 2002).

0.6

12.7

111.9

Chr 3RZ448

RG745

RG391CDO1395

Rice

Chr 3

13.1

15.7

104.8

RZ448

RG191

ACC-CTG2

CDO1395

Rice

Chr 4H

140.4

WG114

BCD1172.12.1 CDO1395

Alp

Barley

Chr 4DL

1.110.2 BCD1230

CDO1395

AltB

Wheat

Chr 4R

11.9 0.0 Alp

BCD1230CDO836

Rye

Table 1. Problem soils in rice cultivation areas in the Mekong Delta (MD) and Red River Delta (RRD).

Soil typeArea (million

ha)Percentage of total area

MD RRD MD RRD

Acid-sulfate 1.60 0.11 40.8 6.0Salinity 0.74 0.09 18.9 5.2Peat soils 0.02 0.002 20.0 0.1Gray soils 0.13 0.15 3.4 8.0

76 Bui Chi Buu and Nguyen Thi Lang

Three populations of O. rufipogon were collected from Tram Chim (see Table 2) by Duncan Vaughan and Bui Chi Buu in 1989. The soil of Tram Chim (bird sanctuary) is strongly acid-sulfate, with pH from 2.8 to 3.2. Later, three crosses were developed (D.S. Brar) as follows:

AS1007 IR64/O. rufipogon (acc. 106412) AS833 IR64/O. rufipogon (acc. 106424) AS996 IR64/O. rufipogon (acc. 106412)

All the O. rufipogon accessions were collected from Tram Chim. The F2 progenies were sent to Vietnam for test-ing in the target areas. AS996 was subsequently selected and

later released as a national variety in 2002 for cultivation in less favorable acid-sulfate soils (Tables 3 and 4).

Tolerance of phosphorus

Genotypes with promise for tolerance of phosphorus deficiency came from the following crosses: Kasalath 47/OM4495 (BC2F2), AS996/OM2395 (RILs F6), M23/AS996, M362/AS996, M379/AS996, M382/AS996, and Kasalath 47/AS996. OM4498 is a backcross line developed from IR64/OMCS2000//IR64 using marker-assisted selection to incorporate two QTLs for salinity tolerance, one each on chromosomes 1 and 8 (with markers RM315 on chromosome 1 and RM223 on chromosome 8). This breeding line has been extensively tested and will be released soon due to its large-scale adoption in large areas in less favorable regions. Four traits related to P-deficiency tolerance were considered (Table 5). AS996 and IR26 had a higher relative index of shoot length, root length, shoot dry weight, and root dry weight than OM2395 and IR36. The genetic control of some characters related to P-deficiency tolerance was studied through diallel analysis. Suitable materials chosen were OM723-11, OM850, IR64, IR50404, OM997, and

Table 2. Phenotyping for Al toxicity tolerance us-ing Yoshida solution and two concentrations of aluminum, 0 and 30 ppm.

Material Relative root length

Oryza rufipogon (acc. 106412) 1.158AS996 1.133OM1490 1.058OM1314 0.899Ca Dung Do (tolerant check) 0.843IR29 (sensitive check) 0.514LSD0.05 0.230

Table 3. Area covered by AS996 during 2000 to 2004.

Year Area (ha)

2000 38,0642001 93,6662002 125,3202003 21,2062004 23,743

Table 4. Rice varieties adapted to acid-sulfate soils in the Me-kong Delta in 2003 and 2004.

Designation OriginArea covered (ha)

2003 2004

OM1490 OM606/IR44592-62-1-1-3 213,157 344,263VND95-20 IR64 mutant 153,994 185,201AS996 IR64/O. rufipogon 21,206 23,743OM3536a TD8/OM1738 37,449 88,693IR64 (check) IR18398-36-3-3 101,211 98,747

aOM3536 is intended to replace AS996 due to its better grain quality in less favorable areas.

Table 5. Phenotyping for P-deficiency tolerance in Yoshida solution using two tests of 5 and 10 ppm of P2O5.

LineShoot length (cm) Root length (cm) Shoot dry weight

(mg)Root dry weight

(mg)

5 ppm 10 ppm 5 ppm 10 ppm 5 ppm 10 ppm 5 ppm 10 ppm

AS996 29.9 32.7 9.14 9.70 19.4 20.6 2.5 2.8 0.91a 0.94 0.94 0.89

OM2395 28.7 31.9 8.40 11.10 20.7 26.8 2.3 2.90.90 0.76 0.77 0.79

IR26 27.4 29.2 9.76 10.72 18.2 22.6 2.4 2.60.94 0.91 0.81 0.92

IR36 27.8 32.9 8.04 10.06 15.3 29.2 2.1 2.50.84 0.80 0.52 0.84

aRelative index = value at 5 ppm/value at 10 ppm × 100. Data are means of 5 replications.

Rice breeding for acid-sulfate soils in Vietnam 77

IR59606. Tillering ability was considered a good indicator of performance under P deficiency and was used as a selection criterion. Maximum tiller numbers were scored at 45 days after transplanting in the hybrids and the parents, constituting a 6 × 6 diallel set. From this analysis, it was observed that shoot dry weight was the most sensitive growth parameter affected by P deficiency, followed by root dry weight and then number of tillers. The proportion of dominant and recessive genes in the parent (KD/KR = 1.6) was more than one unit. This means that dominant gene actions are more important under P-deficiency stress. The effect of positive alleles was clear (H2/4H1 = 0.37), showing that the higher the root dry weight, the better tolerance of P deficiency. The variance ratio 2 σ2gca/(2 σ2gca + σ2sca) was computed from the expected components of the mean square assuming a fixed model to assess the relative importance of additive and nonadditive gene effects in predicting progeny performance. Data in Table 6 demonstrate the effect of nonadditive gene action in the inheritance of the characters, except for growth duration.

Work plans

The following objectives were set: l To perform QTL analysis for P-deficiency toler-

ance. l To select promising lines from the current crosses at

CLRRI’s experimental field and target areas (Long An, Tien Giang, Dong Thap, Kien Giang) to go into observation yield nurseries, preliminary yield trials (PYTs), and advanced yield trials (AYTs).

l To create new crosses based on phenotyping and a DNA survey among gene pool materials.

l To do G × E interaction analysis (9–15 sites). l To analyze iron deficiency through both phenotyping

and genotyping. Rice production in Vietnam was successfully increased because of the effective collaboration with IRRI and others in various research areas. Rice breeding for acid-sulfate soils will be considered as a key activity in the coming years to help narrow the yield gap in these less favorable areas. Priorities considered will be the development and use of marker-assisted selection combined with the advantages of conventional breeding methods. Vietnam needs to increase and strengthen its capacity in biotechnology for further rice improvement and will need further assistance from IRRI to

develop prebreeding materials. The integration of biotech-nology tools with conventional breeding methods offers new opportunities to increase rice productivity and sustainability and develop better varieties with higher tolerance of acid-sulfate toxicity. The potential of the existing genetic diversity for enhancing the productivity of acid-sulfate soils has not yet been adequately explored. We need further collaboration to make better use of the latest biotechnological developments to be employed in conjunction with conventional rice breeding programs for germplasm improvement.

References

MARD (Ministry of Agriculture and Rural Development). 2005. Annual report of 2004. Hanoi (Vietnam): MARD.

Nguyen DB, Brar DS, Buu BC, Tao NV, Luong PN, Nguyen HT. 2002. Identification and mapping of the QTL for aluminum tolerance introgressed from wild sources, Oryza rufipogon Griff., into in-dica rice (Oryza sativa L.). Theor. Appl. Genet. 106:583-593.

Notes

Authors’ address: Cuu Long Rice Research Institute, Vietnam.

Table 6. Inheritance of some growth and yield traits under P-deficiency stress.

Trait (H1/D)1/2 2 σ2gca/(2 σ2gca + σ2sca)

H2ns (%) (narrow-sense

heritability)

Tilling capacity 1.94 0.16 19.70Growth duration 0.98 0.56 33.90Filled grains per panicle 5.80 0.01 3.10Root dry weight 0.81 0.03 20.90

78 Singh et al

Breeding rice for salt-affected areas of IndiaR.K. Singh, B. Mishra, A.M. Ismail, and G.B. Gregorio

Breeding of rice varieties for salt-affected areas remained elusive for a long time because of the poor understanding of the problem and the complexities associated with it in real farmers’ fields, where multiple stresses are common. Recently, considerable progress has been made by different organizations in developing suitable salt-tolerant rice vari-eties for target-specific areas after considerable efforts and investments were made on characterization of these problem soils. To some extent, almost all soils contain some salts but they are not designated as salt affected until the concentra-tion of specific harmful salt becomes too high and hinders plant growth and development. Sodium-based salts usually predominate in salt-affected areas. Before addressing the needs and strategies for breeding rice varieties suitable for salt-affected ecologies of India, we will provide an overview of the type and extent of salt-affected areas.

What are salt-affected soils?

Salt-affected soils are classified broadly into two categories: saline and sodic (or alkali) soils. However, another category of salt-affected soils is also sporadically found and is referred to as saline-sodic soils. Sodium is the dominant cation in all salt-affected soils, with soluble chloride and sulfates as the corresponding anions in saline soils, resulting in an electrical conductivity (EC) of the soil solution of greater than 4 dS m–1. In sodic soils, however, sodium also predominates over other cations but rather more in the exchange complex than in soil solution, together with high concentrations of carbon-ate/bicarbonate anions. Such soils have high exchangeable sodium percentage (ESP > 15) and pH (>8.5 and sometimes up to 10.7) and poor soil structure. Saline-sodic soils, also called saline-alkali soils, have both high ESP (>15) and EC (>4 dS m–1).

Extent of salt-affected soils in India

Abrol and Bhumbla (1971) estimated the country’s salt-af-fected soils as 7 million ha in 1971. This was followed by various estimates by different scientists and agencies, and, after critical evaluation of these varying estimates, the Central Soil Salinity Research Institute (CSSRI), Karnal, endorsed the figure of 8.6 million ha from Singh (1992) as the latest reliable figure. Broadly based on chemical characteristics

and place of occurrence, the salt-affected soils of India are grouped into three major problem areas: sodic soils, inland saline soils, and coastal saline soils, with estimated areas of 3.4, 3.1, and 2.1 million ha, respectively (CSSRI 1997). With the introduction of large-scale irrigation canal networks throughout the country, a majority of the salt-related problems are emerging under the canal command areas (Table 1), most of which still remain potential sources of secondary saliniza-tion. Although both salinity and sodicity can be found in a canal command area as well as outside of it, most of these command areas are saline (inland). A majority of sodic soils are located in Uttar Pradesh (1.1 million ha), Punjab, (0.35 million ha), Haryana (0.18 million ha), and some parts of Bihar, Rajasthan, and Karnataka, whereas inland saline ar-eas are predominant in Rajasthan, Gujarat, Andhra Pradesh, Karnataka, Punjab, and Haryana. Coastal salinity is found along the 5,734-km stretch of coastline of the Indian mainland, which can be divided into four coastal regions: Gujarat coast plains, West coast plains, East coast plain, and Midnapur coast plain. The ecology of the coastal region is very unusual because of several com-pounded problems such as salt-water intrusion, salt sprays, bank erosion, impeded drainage, and other climate-induced factors, which eventually have a drastic effect on agricultural production. Soils of the coastal regions vary considerably, but generally are medium to heavy textured with NaCl as a dominant soluble salt, except in Kerala and some areas of West Bengal, where soils also contain large quantities of Na2SO4. Soil salinity status varies by season and is at its peak in May. It decreases with the onset of the monsoon and is generally the lowest during August-September. Depth of groundwater and its salinity also vary by season.

Projected losses in salt-affected soils

India incurs about a 12% loss in its production system because of degraded lands, which includes salt-affected soils (Singh et al 2003). This loss is estimated at about US$291 million per annum in salt-affected areas alone. The states with larger salt-affected areas share most of the losses such as Gujarat, Rajasthan, Uttar Pradesh, Andhra Pradesh, and West Bengal, when compared with other states with fewer affected areas. Overall losses nationally are projected as

Breeding rice for salt-affected areas of India 79

Country’s productivity at 2002 prices: Rs. 12,288 per ha ($273 per ha)Losses due to degraded lands: Rs. 1,521 per ha ($34 per ha)Value of productivity losses due to salt-affected soils: 8.6 × 106 × 1,521 Rs. 13.1 billion per annum $291 million per annum

The rainfed rice ecosystem in salt-affected areas of the country

Limited freshwater availability is a serious constraint along the coastal belt of India. Excess waterlogging and stagnant partial flooding due to high rainfall within a short span in the monsoon season, coupled with poor drainage, particularly along the eastern coast, create serious problems for rice growth and yield. Water shortage, on the other hand, caused by scanty and uncertain rainfall limits cropping intensity and productivity during the postmonsoon season. Consequently, coastal area is mostly monocropped with rice, except in some regions of the Gujarat plains. About 82–85% of the area is covered by rice during the wet (kharif) season under rainfed conditions, whereas summer/boro (rabi/dry season) paddy cultivation is restricted to areas where sufficient fresh water is available for irrigation. Cropping in inland sodic or saline areas mostly depends on the availability of irrigation facilities, and most of the sodic and saline-sodic areas where these facilities are not available remain barren. People try to bring them back into cultiva-tion using various reclamation strategies in which relatively good-quality irrigation water is a prerequisite. Similarly, in inland saline areas where either rice- or non-rice-based crop-ping systems are being practiced, farming depends mostly

on the quantity and quality of irrigation water, if available. Therefore, the only ecology that is totally rainfed in salt-af-fected soils is predominantly in coastal saline areas during the wet season. However, dry-season cultivation in these areas invariably depends upon ensured irrigation for a major portion of the season.

Why is rice the major crop in salt-affected areas?

Almost the entire coastal belt of the Indian mainland is prone to inundation by seawater during high tides, which results in salinity intrusion and occasional submergence. Thus, only salt- and submergence-tolerant crop varieties, such as rice, are economically viable in these areas. Rice is the only economic crop that can grow well in waterlogged environ-ments and it also tolerates salinity to some extent. In sodic soils, low infiltration rate due to poor hydraulic conductivity and poor soil physical and chemical conditions (high ESP) causes stagnation of water on the soil surface, a condition that cannot be tolerated by any other crop except rice. Thus, rice is recommended as the first crop to start with during reclamation of sodic soils. For these reasons, rice remains the most preferred crop in salt-affected areas.

Challenges and constraints facing rice production in salt-affected areas

Despite the large area under rice in the coastal belt, the average yield is far lower (1.2–1.4 t ha–1) than the national average (approx. 2.0 t ha–1). Possible causes for such low productivity are 1. Occasional flooding, cropping in medium lands/

lowlands, deep waterlogged areas, and floodplains under the rainfed rice ecosystem.

Table 1. Extent and distribution of salt-affected soils in India (000 ha).

StateSalt-affected areas

Canal command area

Outside canal command area

Coastal Total

Andhra Pradesh 139.4 390.6 283.3 813.3Bihar 224.0 176.0 0.0 400.0Gujarat 540.0 372.1 302.3 1,214.0Haryana 455.0 na 0.0 455.0Karnataka 51.4 266.6 86.0 404.0Kerala na na 26.0 26.0Madhya Pradesh 220.0 22.0 0.0 242.0Maharashtra & Goa 446.0 na 88.0 534.0Orissa na na 400.0 400.0Punjab 392.6 126.9 0.0 519.5Rajasthan 138.2 983.8 0.0 1,122.0Tamil Nadu 256.5 na 83.5 340.0Uttar Pradesh 606.0 689.0 0.0 1,295.0West Bengal 0.0 na 800.0 800.0Total 3,469.1 3,027.0 2,069.1 8,562.2

Source: Singh (1992), na - not available.

80 Singh et al

2. Frequent water-deficit stress coupled with salinity. 3. Continued cultivation of traditional low-yielding

local rice varieties and landraces. Under these circumstances, losses of crop yield are 10–80% and sometimes even 100% when rainfall is erratic. During the dry (rabi) season, farmers are compelled to keep land fallow because of high salinity, lack of good-quality irrigation water, and, above all, lack of suitable crops and varieties with high salinity tolerance. Current well-adapted rice varieties grown in coastal areas are mostly traditional photoperiod-sensitive, fertilizer-non-responsive, tall (prone to lodging) varieties with low yield potential (1.5–2.5 t ha–1) and poor grain type. The major hurdle for the reclamation of sodic and in-land saline soils is the high initial cost. Most such areas are either owned by the local government or by poor farmers who do not have enough resources and risk-bearing capacity to reclaim or amend these soils; hence, such lands remain bar-ren, particularly during the dry season. However, once these soils are reclaimed, they remain very productive, reflecting a vast scope for bringing back such lands into the produc-tion chain, provided resources and technical help become available. Improved rice varieties with a reasonable amount of salinity tolerance and adaptation to various ecologies are becoming available, but their spread to farmers’ fields is not yet adequate because of the lack of sufficient seeds and knowledge. A lack of adequate communication facilities and poor marketing infrastructure in general further add to the slow spread of these varieties. With the recent expanding urbanization and population growth, a minimum of 2.5% annual growth in food grain production will be needed to meet the increasing demand. India’s rice production target for 2020 is 140 million tons (Paroda 1998). Rice occupies 42 million ha under diverse ecologies and has large untapped yield potential even with current available modern production technologies. However, the growth rate in rice production, which reached 4.4% in the 1980s, declined to only 2.3% in the 1990s. Therefore, if the country has to maintain self-reliance in food production, it would have to strive to achieve and sustain a minimum an-nual growth of 2.5% in rice production. However, a growth rate of at least 3.1% in production needs to be regained in order to have sufficient rice for both home consumption and exports (Swaminathan 1999).

Overcoming the challenges

Crop productivity from salt-affected areas in general and from coastal areas in particular is lagging behind most other inland areas. Serious thought should be given to increasing the share of coastal areas for the country’s overall economic growth as the contribution of coastal areas to rice production is the lowest. Out of a total of about 92 million tons of rice being produced in India, only about one-fourth is from the long coastal belt of the country. The production growth rate in coastal states is much lower than that of irrigated inland

areas. Concerted efforts are needed to bridge the produc-tion gap in these areas given that they are highly populated with impoverished communities. To achieve the production targeted for 2020, the following important areas of research need further attention: l Steadily raising the genetic yield ceiling in rice

varieties. l Reducing the gap between potential and actual

yield. There is an immediate need to devote greater efforts and resources to develop target-specific, locally adapted high-yielding varieties for the diverse rainfed ecologies (Paroda 1998). The gap between potential and actual yield in these coastal areas is immense, and is widening because of the development of many improved salt-tolerant, fertilizer-responsive, and intermediate stature high-yielding varieties. However, the potential of these varieties is not yet being realized in farmers’ fields due to a lack of knowledge and seeds of these varieties in target areas. Therefore, for major yield breakthroughs in the rainfed coastal ecologies, both rice varietal improvement and ensured availability of quality seed along with technological awareness are needed.

Linkage established between IRRI and network sites of India

No single institute is capable of undertaking all the complex tasks associated with problem soils. Therefore, to develop a collaborative research program for the improvement of rice yields in problem soils, the International Rice Research Institute (IRRI) convened a workshop in November 1986, attended by participants from South and Southeast Asian rice-growing countries. After reviewing the status of research and future needs, it was decided that concerted efforts should be devoted to expedite progress through a network program. A cooperative was established to share virtual resources and responsibilities in research on problem soils. IRRI agreed to undertake the prebreeding research and to strengthen the national program for each type of problem soil, so it could assume regional responsibilities. India was also selected as a partner for cooperative network research on salinity. Before the formal establishment of IRRI’s Consortium for Unfavorable Rice Environments (CURE), the ICAR-IRRI collaborative research program on “Germplasm Improvement for Saline Soils” began in 1987. This project aimed to develop salt-tolerant varieties suited to target environments in the country through shuttle breeding. The Central Soil Salinity Research Institute, Karnal, was identified as the lead center for collaboration. Based on the variability and complexity of adverse soils, which are invariably compounded by climatic hazards, and nutritional toxicity or deficiency, and mostly influenced by the interactions among these factors, target network sites were identified and characterized in the country. The major theme of the network approach is the sharing of resources (in terms of their in situ stress conditions) to iden-tify the best adapted material for target areas. The selected materials were advanced and used in breeding programs at


all the network centers for the development of target-specific as well as widely adapted genotypes. The objectives of the network were set as follows: l To develop high-yielding salt-tolerant rice varieties

for enhancing productivity for sustainable agricul-ture on salt-affected soils.

l To collect and evaluate the germplasm available from traditional areas.

l To generate breeding materials as per the need of target sites.

l To enhance the amount of salt tolerance in rice genotypes.

l To strengthen the research capacity of the scientists involved in the research network.

Major accomplishments

The progress made under the ICAR-IRRI collaborative re-search program, considering the low budget of the project, is quite remarkable in terms of the progress made in developing improved genotypes. Out of nine varieties (CSR21 to CSR29) developed by CSSRI using material exchanged through this network, six (CSR21, CSR23, CSR26, CSR27, CSR28, and CSR29) were recommended by the All India Rice Workshop as promising genotypes for various salt-affected ecologies. Out of these varieties, CSR27 and CSR23 have already been released by the Central Varietal Release Committee (Table 2). Similarly, other collaborating centers also screened large sets of germplasm at specific target sites over the years and used the advanced material for direct release or indirectly in their breeding. Recently, CSR22 was released as a high-yielding rice variety for salt-affected areas of Karnataka.

Table 2. Varieties developed at CSSRI-Karnal through ICAR-IRRI collaborative research.

VarietyStress adaptation Duration

(days)Grain typea Remarks

pH1:2 ECe (dSm–1)

CSR21 9.8–10.0 < 9.0 125–130 LB Recommended by the 30th AIRW for sodic soils

CSR22 9.8– 9.9 < 10.0 135–140 MS Released by state variety release commit-tee for salt-affected soils of Karnataka

CSR23 9.8–10.0 < 10.0 135–140 MS Released by CVRC for alkaline and coastal saline soils

CSR26 9.8–10.0 < 9.0 125–130 LS Recommended by the 33rd AIRW for alkaline and coastal saline soils

CSR27 9.6– 9.9 < 10.0 125–130 LS Released by CVRC for alkaline and coastal saline soils

CSR28 9.8–10.0 < 10.0 130–135 SB Recommended by the 31st AIRW for sodic and inland saline soils

CSR29 9.8–10.0 < 10.0 115–120 LS Recommended by the 31st AIRW for sodic and inland saline soils

aLB = long bold, MS = medium slender, LS = long slender, SB = short bold.

Site and plant type characterization

One of the most important reasons for the early and consider-able success of the project was the initial site characteriza-tion undertaken at the beginning of the project. The soil and water quality and plant type requirements were worked out to supply the best genotypes to the specific network site for further testing of their suitability and adaptability. Hot spots were identified all over the country (Table 3, Fig. 1), and then a shuttle-breeding approach was adopted to evaluate the advanced breeding lines and populations. Materi-als were selected, advanced, and used for breeding programs at all the network centers to select genotypes suitable for their particular target sites and plant-type preferences (Table 4). Germplasm exchange was the main component of the collaborative research program of this network. IRRI supplied 978 advanced breeding lines distributed to all collaborators of the country after multiplication. The year-wise distribution appears in Table 5. This advanced rice germplasm received from IRRI under the project was in addition to the regular IRSSTON trial received through INGER (the International Network for the Genetic Evaluation of Rice). The breeding materials were multiplied, evaluated, selected, advanced, and distributed to the network centers. Promising lines were nominated to the national nursery and trials by the different target centers as per their plant-type requirements. The number of selected entries nominated by CSSRI, Karnal, for the national nursery (NSASN) and trial (SATVT) is presented in Table 6.

The development of rice varieties for salt-affected areas in India—a look back

The development of suitable salt-tolerant rice varieties for salt-affected soils in India dates back to the pre-Independ-ence era. A majority of such rice varieties were from the

82 Singh et al

Table 3. Test sites of the ICAR-IRRI national network for the development of salt-tolerant improved rice genotypes.

Site (state) Coordinating research institute Ecosystem, type of salt stress, and other stressesa

Kaithal (Haryana) Central Soil Salinity Research Institute

Irrigated, sodic (high RSC)

Karnal (Haryana) Central Soil Salinity Research Institute

Irrigated, nonstress for yield com-parison

Lucknow (U.P.) Central Soil Salinity Research Institute

Irrigated, high sodicity

Canning Town (W. Bengal) Central Soil Salinity Research Institute

Rainfed, coastal saline, waterlog-ging

Anand Central Soil Salinity Research Institute

Irrigated, inland saline

Kanpur (U.P.) C.S. Azad University for Agri-culture & Technology

Irrigated, sodic

Faizabad (U.P.) Narendra Deva University for Agriculture & Technology

Irrigated, sodic

Vyttila (Kerala) Kerala Agricultural University Rainfed, coastal deepwater, acid saline soils

Panvel (Maharashtra) Konkan Krishi Vidyapeeth Rainfed, coastal saline drought-prone

Machilipatnam (A.P.) Andhra Pradesh Agricultural University

Rainfed, coastal saline

Tiruchirapally (Tamil Nadu) Tamil Nadu Agricultural University

Irrigated, sodic

Motto (Orissa) Orissa University of Agriculture and Technology

Rainfed, coastal saline

Nawagam (Gujarat) Gujarat Agricultural University Irrigated, inland salinityKaraikal (Pondichery) Pandit Jawaharlal Nehru

College of Agriculture and Research Institute

Sodic (high RSC)

aRSC = relative sodium content.

coastal belt due to the perpetual nature of salinity problems, where farmers do not have many options except to grow rice. Local rice germplasm grown in saline and sodic soils was very poor in yield and quality. Regional rice stations were established in different rice-growing states with the help of ICAR to tackle these regional problems. Scientific efforts in different ecosystems resulted in the development of salt-tolerant rice varieties, though most of them were selections from the locally adapted landraces prevailing for a long time in the target environments. Details on these local genotypes from different parts of the country and the early selections are given in Table 7. Some of these selections were released by the states, whereas most of them remained as selections in very specific and small isolated areas. Lately, efforts from some regional rice research sta-tions resulted in the development of a few salt-tolerant rice varieties. The varietal development program for salt-affected areas progressed very slowly for numerous reasons: (1) the complex nature of this ecosystem with high temporal and spatial variability, (2) the complex nature of inheritance of the tolerance trait, (3) a lack of sufficient resources and research capacity, and (4) a lack of precise and repeatable screening techniques. Lately, more attention has been devoted, which

yielded some results when rice researchers attempted a few crosses before the 1970s, involving salt-tolerant landraces. But, unfortunately, the rice varieties developed were not accepted largely due to insufficient salt tolerance and poor and unacceptable grain quality. The genetic drag of the local traditional types was large, resulting in tall genotypes with poor grain quality that were nonresponsive to fertilizers (Table 8).

Post-1970 breeding progress—the development of new rice varieties

In the post-1970 era, the Indian Council of Agricultural Research (ICAR) entrusted this responsibility to national institutes such as CSSRI, Karnal, with a mandate to work on salt-affected areas at the national level. Similarly, some of the state-owned research stations also devoted some resources to develop target-specific salt-tolerant rice varieties for various environments. The concerted efforts to develop target-specific as well as widely adapted salt-tolerant rice varieties resulted in the release of numerous varieties. Initially, most of the re-leased varieties were from different states for target areas, that is, specific soil-stress conditions such as CSR 5 and Usar 1


6. Canning Town

7. Motto

1. IRRI (Philippines)3. Kanpur 4. Lucknow 5. Faizabad

2. CSSRI, Karnal

INDIA15. Nawagam

14. Anand

13. Panvel

8. Machilipatnam

10. Port Blair

9. Karaikal11. Tiruchirapalli12. Vyttila

Fig. 1. ICAR-IRRI collaborative research network of India.

IndexCollaborating centers 1. IRRI–Philippines 2. CSSRI–Karnal

Network sites 3. CSAU Kanpur (U.P.) 4. CSSRI-RRS Lucknow (U.P.) 5. NDAUT Faizabad (U.P.) 6. CSSRI–RRS Canning Town (W.B.) 7. NARP, Motto (Orissa) 8. ARS, Machilipatnam (A.P.) 9. Pajancoa, Karaikal10. CARI, Port Blair11. SSRC, Tiruchirapalli12. RRS, Vyttila (Kerala)13. KRS, Panvel (Maharashtra)14. CSSRI–RRS, Anand (Gujarat)15. GAU, Nawagam

for the sodic soils of Haryana and Uttar Pradesh, respectively; Vyttila 1 and Panvel 1 for the coastal saline soils of Kerala and Maharashtra, respectively; and Co 43 for the saline so-dic soils of Tamil Nadu. Many target-specific improved rice varieties were also developed by different research stations and released by their respective states.

National releases of rice varieties

Considering its national mandate and responsibility, CSSRI took the lead by developing the first salt-tolerant rice variety for sodic and inland saline soils at the national level that was released in 1989. So far, only 12 salt-tolerant rice varieties were released nationally through CVRC, of which six were developed by CSSRI at Karnal, four by the CSSRI Regional Research Station at Canning Town, and one each by the Central Rice Research Institute, Cuttack, and Directorate of Rice Research, Hyderabad. Some characteristics of these rice varieties are given in Table 9.

State releases of rice varieties

During the post-1970 era, various states started giving im-petus to the development of salt-tolerant crop varieties for their specific problem areas. These intensified efforts gave

dividends in terms of the release of many target-specific varie-ties by the state governments for specific niches. State-wise details of these varieties are given in Table 9.

Infrastructure and screening facilities

The ICAR-IRRI collaborative research project was mainly committed to developing and distributing the germplasm to various collaborating centers due to limited resources. Subsequently, various centers were able to develop their own facilities by capitalizing on the knowledge and experiences of other partners. Most of the collaborators used in situ field tri-als to screen germplasm due to the paucity of funds. However, CSSRI, Karnal, and a few other centers were able to upgrade and expand their existing screening facilities with the help of some other projects. Screening facilities and procedures for salt-tolerance studies are enumerated as follows.

In situ field evaluationScreening of genotypes under an actual field-stress environ-ment in 2 to 5 long rows of 8 to 20 meters to minimize the effects of spatial variability was followed at most of the target sites, including CSSRI, depending on the availability of suitable fields. Replicated trials are always preferred for advanced stabilized material. Field screening is usually the

84 Singh et al

Table 4. Survey of specific plant-type requirements of the target sites.

Target site Plant-type characterization

1. Vyttila (Kerala) l Plant height not less than 125 cml Total duration not more than 125

daysl Tolerance of flooding (early seedling

vigor)l Tolerance of salinityl Tolerance of acidity

2. Machilipatnam (A.P.) and Motto (Orissa)

l Semidwarf to medium talll Total duration more than 140 daysl Tolerance of salinityl Fertilizer responsive and high yield-

ing (traditional varieties have disap-peared from the problem areas)

3. Panvel (Maharashtra) l Dwarf and semidwarf l Fine grainl Salinity tolerant l Early maturing (less than 125 days)

4. Canning Town, Chinsurah (W. Bengal), and Port Blair

l Medium tall l Tolerant of salinity l Tolerant of waterlogging (early

seedling vigor)l Weakly photosensitivel Medium growth duration For upland and well-drained condi-

tions: l Early maturingl Height 60 to 90 cm l Fertilizer responsive and photo-in-

sensitive genotypes5. Haryana, Punjab, and Uttar

Pradeshl Dwarf/semidwarf l High yieldingl Early maturing (125 days)l Fertilizer responsivel Slender to medium coarse grain l High tolerance of sodicity and

salinity6. Anand and Nawagam

(Gujarat)l Semidwarf l High yieldingl Medium growth durationl Fertilizer responsivel Slender to medium coarse grain l High tolerance of salinity

7. Karaikal (Pondicherry) l Semidwarf l High yieldingl Early maturing (125 days)l Fertilizer responsivel Medium coarse grain l High tolerance of sodicity and

salinity

Table 5. Breeding lines evaluated through the network during 1987 to 2000.

Year Germplasm Number

1987 Advanced breeding lines 501987 Anther culture derivatives 401988 Advanced bulk populations 1311988 Anther culture derivatives 71989 Advanced bulk populations 1061990 Advanced bulk populations 381991 Advanced bulk populations 401992 Advanced bulk populations 511993 Advanced bulk populations 291994 Advanced bulk populations 601995 Advanced bulk populations 161996 Advanced bulk populations 281997 Advanced bulk populations 851997 Recombinant Inbred lines (IR29 × Pokkali) 891998 Advanced bulk populations 542000 Advanced bulk populations 702000 Advanced breeding lines 84

Total germplasm received under the project 978

Table 6. Genotypes and vari-eties nominated for the na-tional and nursery trials from 1991 to 2005.

Year No. of entries

National Saline Alkaline (Tolerant) Screening Nursery (NSASN)1991 51992 101993 141994 151995 101996 101997 –1998 91999 42000 42001 12002 32003 32004 –2005 1Total 89

Saline Alkaline Tolerant Vari-etal Trial (SATVT)1993 41994 81995 81996 41997 71998 41999 52000 32001 –2002 –2003 2Total 45


Table 7. Traditional landraces and subsequent selections adapted to salt-affected soils in India.

Landrace Selection State Area of adaptation(local variety)

Jhona Jhona 349 Punjab Sodic soils of Punjab and HaryanaChaul local T 21 Uttar Pradesh Sodic soils of Uttar PradeshKashi Lakra or T 22A Uttar Pradesh Sodic soils of Uttar PradeshBejhari Uttar Pradesh Sodic soils of Uttar PradeshKalambank SR 26B Orissa Coastal saline soils of east coastLocal collection SR 8 Orissa Coastal saline soils of OrissaGeti Orissa Coastal saline soils of OrissaRaspanjore Orissa Coastal saline soils of OrissaLunpatali Orissa Coastal saline soils of OrissaChoudusannalu Andhra Pradesh Saline soils of Andhra PradeshBudda Mologolakulu MCM 2 Andhra Pradesh Coastal saline soils of Andhra

PradeshTellathokavadlu Andhra Pradesh Coastal saline and sodic soils of

Tamil NaduKalar samba Tamil Nadu Coastal saline and sodic soils of

Tamil Nadu Chuvada samba Tamil Nadu Coastal saline and sodic soils of

Tamil NaduKullakar Tamil Nadu Coastal saline and sodic soils of

Tamil NaduVellaikattai Tamil Nadu Coastal saline and sodic soils of

Tamil Nadu Kallimadyan Tamil Nadu Coastal saline and sodic soils of

Tamil NaduLocal Rata Kala Rata 1-24 Maharashtra Coastal saline soils of MaharasthraLocal Rata Bhura Rata 4-10 Maharashtra Coastal saline soils of MaharasthraMorchuku Maharashtra Coastal saline soils of MaharasthraDodka Maharashtra Coastal saline soils of MaharasthraKhare Bhat Maharashtra Coastal saline soils of MaharasthraMahadi Rata Maharashtra Coastal saline soils of MaharasthraPatnai Patnai 23 West Bengal Coastal saline soils of West BengalPatnai Patnai 298 West Bengal Coastal saline soils of West BengalDamodar CSR1 West Bengal Coastal saline soils of West BengalDasal CSR2 West Bengal Coastal saline soils of West BengalGetu CSR3 West Bengal Coastal saline soils of West BengalBenisail Matla West Bengal Coastal saline soils of West BengalNona Bokra Hamilton West Bengal Coastal saline soils of West BengalVelki (Bhaluki) West Bengal Coastal saline soils of West Bengal

and OrissaRupsal West Bengal Coastal saline soils of West BengalKumargore West Bengal Coastal saline soils of West BengalNonasail CSR 6 West Bengal Coastal saline soils of Orissa and

West BengalArya Arya 33 Karnataka Saline and sodic soils of KarnatakaKarekagga Ankola Karnataka Saline and sodic soils of KarnatakaBilekagga Karnataka Saline and sodic soils of KarnatakaBali Kerala Coastal saline soils of KeralaOarkzhama Kerala Coastal saline soils of KeralaPokkali Kerala Coastal saline soils of KeralaOarpandy Kerala Coastal saline soils of KeralaOrmundakan Kerala Coastal saline soils of KeralaOdacheera Kerala Coastal saline soils of KeralaChettivirippu Mo 1 Kerala Coastal saline soils of KeralaKalladachampavu Mo 2 Kerala Coastal saline soils of KeralaKunjathikkara Mo 3 Kerala Coastal saline soils of KeralaChottupokkali Vytilla 1 Kerala Coastal saline soils of KeralaKorgut Goa Coastal saline soils of GoaAzgo Goa Coastal saline soils of Goa

86 Singh et al

Table 8. Salt-tolerant rice varieties evolved through recombina-tion—past efforts (before 1970).

Variety Parentage Institution

PVR 1 MTU 1/SR 26B Rice Research Station, Peravoorani, Tamil Nadu

MCM 1 CO 18/Kuthir Rice Research Station, Machlipatnam, Andhra Pradesh

MK47-22 Malkudai/KR 1-24 Kharland Research Station, Panvel, Maharashtra

SR 3-9 KR 1-4/Zinnya 149 Kharland Research Station, Panvel, Maharashtra

MR 18 SR 26B/Wannar-1 Rice Research Station, Mandya, Karnataka

best way to select adapted genotypes and do final evaluation. Spatial variability is determined by extensive soil testing of the trial fields.

Screening in microplotsSoil heterogeneity and spatial variability are the major factors that sometimes limit the reliability of field-testing data. This attracted the development of mini-field environments with varying controlled salinity and sodicity in plots built out of brick-mortar-concrete materials, measuring 6 m by 3 m, with

Table 9. Rice varieties released for salt tolerance in India.

Name of variety Parentage

Stress tolerance limit Stress adaptation

(type of soil)Year of release

Days to 50% flowering Grain typea

pH2 ECe (dS m–1)

Notified and released by centerCSR 10 M40-431-24-114 / Jaya 10.2 11 Alkaline and

inland saline1989 100 SB

CST 7-1 Damodar/IR24 – 8 Coastal saline 1991 115 LBLunishree Mutant of Nonasal – 8 Coastal saline 1992 115 LSCSR 13 CSR 1/Bas. 370//CSR 5 9.9 8 Alkaline and

inland saline1999 115 LS

CSR 27 (derived from

IR51471)Nona Bokra/IR5657-33-2 9.9 8 Alkaline and

coastal saline1999 95 LS

CSR 30 (Yamini) BR 4-10 / Pak. Basmati 9.5 – Normal and Alkaline

2001 120 LS (Basmati)

CSR 23 (derived from IR52713)

IR64//IR4630-22-2-5-1-3/IR9764-45-2-2

10.0 8 Alkaline and coastal saline

2004 105 MS

Sumati Pankaj/NC 678 – 8 Coastal saline 2004 105 SBBhutnath SR26B/Pankaj – 8 Coastal saline 2004 Photo-senst. SBJarava (B90-15) B32-Sel-4/Spontanea 4 – 8 Coastal saline 2005 Photo senst. SB (IET 15420) //B296CSR36 (Naina) CSR 13/Panvel 2//IR36 9.9 – Alkaline 2005 110 LS (IET 17340)CSRC(S)7-1-4 Pankaj/SR26B – 9 Coastal saline 2008 115 LB (IET 14199/ 18250)

Continued on next page

a depth of about 0.8 to 1 m. Artificially prepared soil of dif-ferent EC (maintained by irrigating with saline water) or pH is maintained in these microplots in a manner that very much reflects field conditions except for the better control of soil heterogeneity. The plot size is usually very small (a single row) but, because of good control over micro environment, the screening is highly reliable and repetitive. The microplots are being used to screen mostly early segregating and stable populations besides genetic studies.

Screening in potsPot screening is mostly conducted for precise, controlled studies of individual plant response under a constant stress. The edaphic environment in pots is more or less uniform throughout the plant growth period with respect to the extent of stress.

Screening in solution cultureSalinity. Solution culture techniques are being used in two ways, one for screening during the seedling stage and the second for screening up to maturity. In the first category, 7-day-old seedlings grown on nonstress modified Yoshida culture solution (Yoshida et al 1976) are transferred to the desired amount of stress either in a bread box with a perfo-rated lid or perforated styrofoam with a bottom mesh. This modified culture solution has KH2PO4 and K2HPO4 in place


Developed/released by stateHaryana/Punjab CSR 5 (Vikas) TKM6/IR8

9.7 7 Alkaline 1979 95 MS

KeralaVyttila 1 Selection from Chotupok-

kali4.5–9.7 9 Coastal saline – 90 MB

Vyttila 2 Selection from Cheruvir-uppu

4.5–9.7 8 Coastal saline 1980 95 LB

Vyttila 3 Vyttila 1/TN 1 4.5–9.9 9 Coastal saline 1987 85 MBVyttila 4 Chethivirippan/IR4630-

22-2-174.5–9.7 10 Coastal saline 1993 95 LB

Vyttila 5 Mutant of Mahsuri 4.5–9.5 8 Coastal saline 1996 95 MBVyttila 6 Cheruviruppu/IR 5//Jaya 4.5–9.7 9 Coastal saline 2004 85 MB

MaharasthraPanvel 1 IR8/BR4-10 9.7 8 Coastal saline 1984 95 SBPanvel 2 BR4-10/IR8 9.8 9 Coastal saline 1988 90 LSPanvel 3 Damodar/Pankaj 9.8 9 Coastal saline 2000 100 LB

Andhra PradeshMCM 1 CO 18/Kuthir – 8 Coastal saline 100 SBWest BengalMohan (CSR 4) Mutant of IR8 – 7 Coastal saline 1981 115 SB

Name of variety Parentage

Stress tolerance limit Stress adaptation

(type of soil)Year of release

Days to 50% flowering Grain typea

pH2 ECe (dS m–1)

Table 9 continued.

KarnatakaCSR 22 IR64//IR 4630-22-2-5-1-

3/IR 9764-45-2-2

9.9

8 Inland saline 2006 105 MS

OrissaSR 26B Selection from Kalambank 9.7 8 Coastal saline 1988 115 LBTamilNaduPVR-1 MTU-1/SR 26B 9.4 4 Coastal saline 1964 115 SBCO 43 IR20/Dasal 9.7 8 Alkaline and

coastal saline1982 110 MS

ASD 16 ADT 31/CO 39 9.6 7 Coastal saline 1986 95 SBTRY 1 IR578-172-2-2/BR 1-2-

B-1910.0 10 Alkaline and

coastal saline1995 110 SB

TRY 2 IET 6238/IR36 9.9 8 Alkaline and coastal saline

2001 85 LS

GujaratSLR51214 Vijay/PTB-21 – 8 Inland and coastal

saline1982 100 LS

Dandi (IET 14906) Panvel 2/IET 8320 – 9 Inland and coastal saline

2003 95 LS

Uttar PradeshUsar 1 Jaya/Getu 9.9 8 Alkaline 1985 100 SBNarendra Usar 2 Selection from the IR

2058 derivative9.8 8 Alkaline 1998 95 LB

Narendra Usar 3 Selection from IR 46330 derivative

9.9 8 Alkaline 2000 100 LS

aSB = short bold, LB = long bold, LS = long slender, MS = medium slender, MB = medium bold.

88 Singh et al

of NaH2PO4 as sodium salt within culture solution and may increase the Na ion concentration (T.J. Flowers, personal communication). However, for screening for tolerance during the reproductive stage, 2–3-week-old seedlings are trans-planted in pots filled with ½ kg of soil and irrigated through an automated circulating irrigation system. After establishment of the seedlings for 3–5 days, the salinity is raised to 50 mM in culture solution stored in an underneath tank using NaCl. This is further raised to 70 mM salinity after 10–15 days. The automated circulatory irrigation system maintains the desired exact salinity in the rhizosphere. Plants are irrigated from the lower tank through a timer-controlled pump. Sodicity. A TRIS buffer-based system has been devised to screen rice genotypes, in which a certain pH is maintained in solution and micronutrients are supplied through foliar sprays. Genotypic performance is evaluated from the change in seedling weight and phenotypic performance. A TRIS (Tris hydroxymethyl aminomethane) concentration of 4 mM (pH 8.6) provides a means for evaluating genotypic differ-ences in response to sodicity. Details are given in Singh et al (2002). Screening in trays. For large-scale screening of varie-ties at germination and seedling stage, salt-affected soil filled in shallow-depth wooden or metal germination trays with a polythene sheet lining on the inner face is used. This is quite convenient for control of salinity, sodicity, and moisture.

Breeding methods used for developing salt-tolerant rice varieties

Conventional breeding Mostly conventional breeding techniques such as introduc-tion, selection, hybridization, mutation, and a shuttle-breed-ing approach were used for developing salt-tolerant rice varieties. Varieties such as Damodar (CSR 1), Dasal (CSR 2), and Getu (CSR 3) are pure-line selections of the local salt-tolerant traditional cultivars that prevailed in the Sunderban areas in West Bengal. These varieties were adapted to salinity stress but, when they were first introduced in sodic areas at Karnal, they were found to be tolerant of sodicity as well. Almost all other released varieties of rice were developed using one or more of the following techniques: l Pedigree method l Modified bulk pedigree method: Individual selected

F2 plants are bulked up to the F4 generation, followed by panicle selection and then advancing the popula-tion as in the pedigree method.

l Shuttle breeding The segregating materials are grown in long rows under salt stress, with space planting, particularly in F2. Selection pressure is gradually increased with generation advance-ment simultaneously under moderate stress and high stress of sodicity or salinity. At the F5 or F6 generation, a part of the material is screened at high pH of up to 9.9–10.0 and salinity of about 10–12 dS m–1.

Mutation breeding was also used before, as in the development of the first salt-tolerant variety, CSR 10. The female parent of CSR 10, M40-431-24-114, was derived from the γ-irradiated F1 seed of the cross CSR 1/IR8. Similarly, Lunishree and CSR 4 were also developed through mutation breeding. Shuttle breeding was adopted in the development of salt-tolerant varieties such as CSR 23 and CSR 27 under the ICAR-IRRI collaborative project for salt tolerance (Mishra 1994).

Nonconventional breeding Anther culture derivatives generated at IRRI and evaluated under the ICAR-IRRI collaborative network in India by CSSRI led to the identification of promising breeding lines IR51500-AC-17, IR51485-AC-1, and AC6534-4 for salin-ity; AC6533-3 for sodicity; and AC6534-1 for dual tolerance (Singh et al 1992, Singh and Mishra 1995). IR51500-AC-17 and AC6534-1 were later named as CSR 21 and CSR 28, and were recommended by the All India Coordinated Rice Improvement Program for release. Rice breeding line IR51500-AC-17, selected in India and identified as CSR 21, was released as PSBRc50 or Bicol in the Philippines (Senadhira et al 2002).

Impact of salt-tolerant rice varieties: geographical distribution, area covered, and obstacles in uptake and seed production

Salt-tolerant rice varieties are widely grown in the salt-af-fected soils of different states and are in great demand by farmers. It is difficult to realize the direct impact of these varieties but their adoption, popularity, and impact could be measured through their breeder and certified seed production and distribution chains. To quantify the impact of salt-tolerant rice varieties, seed distributed by CSSRI, Karnal, to various seed-multiplying agencies was taken as a case study. Until 2004, CSSRI supplied 124.42 quintals of breeder seed and 2,387 quintals of certified seed of all the released salt-tolerant rice varieties to various seed multiplication agencies and land development corporations of the different states. The seed multiplication ratio of breeder seed to certified seed is taken as 1:150 in rice (when the breeder to founda-tion and foundation to certified seed chain is not included in this case study; otherwise, it would be many thousand-fold). If we extrapolate the multiplication ratio of rice, then 124 quintals of breeder seed would have produced a minimum of 18,600 quintals of certified seed in one production season. In addition to this, CSSRI supplied 2,387 quintals of certified seed, so the total certified seed quantity would be 20,987 quintals within one season. The seed produced for sowing is always much more expensive than that used for consump-tion. Even with conservative estimates, if we consider only 20,000 quintals of total certified seed (instead of 20,887 q) in the production chain, which might have covered more than 50,000 ha of area, this would have produced more than 1.5 million quintals of paddy (taking average production of


only 30 q ha–1). This additional food grain from salt-affected soils must be worth several million rupees, besides providing employment opportunities. Indian farmers usually do not replace their seed every season and keep it for a few years before seed replacement. The above estimate is only a one-season scenario; therefore, we could expect an even greater impact after a few seasons. Besides CSSRI, Karnal, many seed multiplication agencies are continuously engaged in seed supply to the land development corporation of different states, other agencies, and directly to farmers. Because of this scenario, the actual impact of salt-tolerant rice varieties will be much greater than what was predicted above. Another notable impact is bringing back unproductive barren soils into rice farming or enhancing the yield on soils with very low productivity. Two scenarios are possible. In the first, a new high-yielding variety is replacing another old va-riety and increasing yield from 5 to 6 t ha–1 with an advantage of 1 t ha–1 only. In the second scenario, unproductive barren land is cultivated using a salt-tolerant rice variety with no or a small amount of gypsum, which gives yield of about 3.0 t ha–1 in the first year and probably more in subsequent years. Comparing both scenarios, though productivity is more in the first scenario, the yield advantage is about 3 times more in the second scenario in addition to more land being put under cultivation. Clearly, the development and deployment of salt-tolerant rice varieties could allow both horizontal and vertical expansion in food production.

Seed production and availability problems

Despite the urgent need of seed of salt-tolerant rice varie-ties, seed-producing agencies are not effective in producing sufficient seed to meet local demand except in a few areas. The major bottlenecks are l Risk aversion by farmers. Because of uncertainty of

production in these unfavorable areas, farmers do not want to invest in costly inputs, including seed.

l Low demand for seed as landowners of salt-affected areas are mostly resource poor.

l Problem areas need mostly site-specific varieties. Seed-multiplying agencies do not want to invest in producing a relatively small amount of seed of many target-specific varieties.

l Salt-tolerant varieties adapted to multiple stresses such as salinity and submergence or sodicity and tolerance of P and Zn deficiency are not yet avail-able.

These problems could be solved to some extent if the state government-owned land development corporations could ensure the procurement of seed of salt-tolerant rice varieties, and help resource-poor farmers with substantial backup of technological know-how and access to other re-sources needed for reclamation or as an input.

Example of the UP Land Development Corporation (UPLDC) in disseminating salt-tolerant varieties

Land development programs in Uttar Pradesh were started in the 1990s. A land development corporation oversees the overall availability of farm inputs, including seeds of salt-tolerant varieties, and with back-up support of technological know-how. In 1993, the UPLDC started an ambitious project of reclaiming 47,000 ha of sodic land in phase I with help from the World Bank and European Economic Community (EEC). Through concerted efforts, UPLDC successfully doubled the targets by reclaiming 97,056 ha of sodic land within 1993-2000 (UPBSN Status Report, April 1999, Am-bekar 2004). UPLDC targeted 150,000 ha of sodic land for reclamation in its phase II program (2000-05). Up to March 2002, it had already reclaimed 126,861 ha, which is also ahead of its scheduled targets. In both phases, the major cause of this great success was the concerted efforts and timely availability of resources. As per the UPLDC Status Report (April 1995), rice variety CSR 10 produced its maximum yield under C class soils in comparison with other rice varieties. C class soils are those where no crop has been grown before during both the wet and dry seasons. Until now, UPBSN, Lucknow, is the bulk seed purchaser of salt-tolerant rice varieties from CSSRI and other seed-producing agencies for its sodic soil reclamation program. Use of these salt-tolerant varieties is an integral component of its reclamation technology in farmers’ fields.

Breeding objectives and challenges for the coming 5–10 years—opportunities for new and enhanced strategies

If we compare the progress in breeding for salt tolerance with that for irrigated ecosystems, it is apparent that the first pro-gram is still far behind. The immediate concern is to raise the average yield of salt-affected areas from the current 1.0–1.2 t ha–1 to about 2.0 t ha–1. As such, rice production will increase by more than 8.5 million tons. Salinity, sodicity, or any other salt-related stress seldom happens in isolation and more ef-forts are needed to ensure site-specific evaluation of breeding lines at target sites. Sharing of resources through germplasm exchange networks is the first step toward a proper diagnosis of problems specific to each site and also for the develop-ment of adapted varieties. As the latest technologies such as marker-assisted selection become available to breeders, more target-specific varieties with enhanced tolerance of salt stress as well as other prevailing stresses will be developed. With the tagging of agronomically important QTLs such as Saltol, Sub1, and Pup1, the complexities of breeding for these environments could be eased and future programs could focus on developing marker systems and strategies to introgress these multiple major QTLs and genes into agronomically superior backgrounds that meet the quality requirements of farmers in any target area.

90 Singh et al

References

Abrol IP, Bhumbla DR. 1971. Saline and alkali soils in India – their occurrence and management. FAO World Soil Resources Rep. p 41:42-51.

Ambekar VW. 2004. Sodic land reclamation in Uttar Pradesh: some policy issues. In: Report of International Conference on Sustain-able Management of Sodic Soils held at Lucknow, India, 9-14 February 2004. p 14-18.

CSSRI. 1997. Vision 2020: CSSRI Perspective Plan. CSSRI, Karnal. 95 p.

Directorate of Rice Research. 2000. High-yielding rice varieties of India—2000. DRR, Hyderabad. 102 p.

Mishra B. 1994. Breeding for salt tolerance in crops. In: Rao et al, edi-tors. Salinity management for sustainable agriculture—25 years of research at CSSRI. Central Soil Salinity Research Institute, Karnal, India. p 226-259.

Paroda RS. 1998. Sustaining the Green Revolution: new paradigms. Dr. B.P. Pal Memorial Lecture. In: Proceedings of the 2nd In-ternational Crop Science Congress. New Delhi (India): Oxford & IBH. p 79-110.

Senadhira D, Zapata FJ, Gregorio GB, Alezar MS, de la Cruz HC, Padolina TF, Galvez AM. 2002. Development of the first salt-tolerant rice cultivar through indica/indica anther culture. Field Crops Res. 76:103-110.

Singh NT. 1992. Land degradation and remedial measures with reference to salinity, alkalinity, waterlogging and acidity. In: Deb DL, editor. Natural resources management for sustain-able agriculture and environment. New Delhi (India): Angkor Publications. 442 p.

Singh RK, Mishra B. 1995. Screening F1 anther culture derivatives of rice for salt tolerance. In: Sharma B et al, editors. Genetic research and education: current trends and the next fifty years. New Delhi (India): Indian Society of Genetics and Plant Breed-ing. p 509-513.

Singh RK, Mishra B, Senadhira D. 1992. Promising salt tolerant F1 anther culture derivatives (ACDs). Int Rice Res Newsl. 17(1):17.

Singh RK, Mishra B, Chauhan MS, Yeo AR, Flowers SA, Flowers TJ. 2002. Solution culture for screening rice varieties for sodicity tolerance. J. Agric. Sci. Cambridge (UK) 139:327-333.

Singh RP, Vasisht AK, Mathur VC. 2003. Quantitative assessment of economic losses of degraded land in India. Division of Agri-cultural Economics, IARI, and Advance Publishing Concept, New Delhi. 51 p.

Swaminathan MS.1999. Predict: century of hope. Towards an era of harmony with nature and freedom from hunger. East West Books (Madras) Pvt. Ltd. p 155-199.

UPBSN. 1999. Status report: UP sodic lands reclamation project. UP Bhumi Sudhar Nigam (erstwhile UPLDC), Lucknow. 53 p.

UPLDC. 1995. Status Report: UP sodic lands reclamation project. Part I. UP Land Development Corporation, Lucknow. 78 p.

Yoshida S, Forna, DA. Cock JH, Gomez KA. 1976. Laboratory manual for physiological studies of rice. Manila (Philippines): Interna-tional Rice Research Institute. 34 p.

Notes

Authors’ addresses: R.K. Singh, A.M. Ismail, and G.B. Gregorio, International Rice Research Institute, Los Baños, Philippines; B. Mishra, project director, Directorate of Rice Research, Hy-derabad, India.

Breeding rice for the sloping uplands of Yunan 91

Upland ecosystems

92 Tao et al


Breeding rice for the sloping uplands of YunnanD. Tao, F. Hu, G.N. Atlin, S. Pandey, P. Xu, J. Zhou, J. Li, and X. Deng

Farmers have been growing upland rice for more than 4,000 years in Yunnan Province of China using traditional varieties under shifting cultivation systems that are based on slash-and-burn practices. Even now, upland rice production is still pri-marily a subsistence-oriented activity in southern Yunnan, with farmers using a large share of their resources to meet food needs. Our efforts over the years have been to develop japonica cultivars with strong seedling vigor, drought tolerance, blast resistance, adaptation to the intermediate elevation of southern Yunnan, and weed competitiveness. In the meantime, germ-plasm introduction from outside Yunnan has been kept through active domestic and international cooperation since 1989. These endeavors have led to the release of varieties IRAT104, B6144F-MR-6, Yunlu 29, and Yunlu 52 in Yunnan. The goal of hybridization breeding in Yunnan is to combine the good adaptation of Yunnan traditional upland rice cultivars with good plant type, blast resistance, and the drought tolerance of foreign upland rice accessions. The keys in our inbred and hybrid breeding programs are the identification of donors for intermediate stature and blast resistance, of recombinants of aerobic adaptation and high yield, and of promising crosses. A traditional upland cultivar, Mengwanggu, usually yields about 2 t ha–1 with optimized fertilizer management on terraces and favorable lower slopes, whereas improved cultivars Yunlu 29, Yunlu 52, and B6144F-MR-6 yield 4 t ha–1. The yield advantage of the improved cultivars results mainly from the significant increase in tiller and panicle size. In 2002, breeding research shifted from breeding rice for sloping uplands to developing aerobic rice with high yield, blast resistance, and fine grain quality under favorable environments and good management. Indica type, because of its input response, became our major target type. One strategy for improving the yield potential of aerobic rice cultivars is to transfer their aerobic adaptation to lowland high-yielding varieties (HYVs), which partition up to 50% of total dry matter to grain, but do not perform well under upland conditions.

Inthisregion,uplandriceplantingareahasbeenstableat100,000hectares,withanaverageyieldofabout2.5tha–1.Asuplandricecultivationisrelatedtothefoodsecurityofthepeoplebasedinhillyregions,itisnotpossibletoreplaceuplandricebyothercropsinashorttime(Zhou1987).How-ever, sloping fields can be changed into terraces to control waterandsoilerosion,fertilizesoil,useagriculturalimple-ments, use improved input-responsive cultivars, and finally to increaseyieldperha.Amongthenumerousfactorsleadingtoincreaseduplandriceyield,breedingforimprovedcultivarsandtheircultivationhaveanimportantroletoplay.However,theYunnanSeedManagementStation(1992)reportedthataround 129 accessions are the most common upland ricecultivarsgrownbyfarmersinuplandareas,outofwhich119arejaponicasand10areindicas.Noneofthesecultivarshasaplantingareaofmorethan6,700ha.Since1980,breedingandtheintroductionofuplandricecultivarshavebeenoneoftheleadingresearchprojectsattheYunnanAcademyofAgriculturalSciences(YAAS).Ourmanuscriptsummarizestheachievementsandexperiencesof25yearsofresearchonuplandriceimprovementatYAAS.

UplandareasinYunnanarecharacterizedbyruggedterrain,pooraccess tomarkets,environmentaldegradation,andahighincidenceofpoverty.Thesearealsoborderareasinhab-itedbyethnicminorities.Farmershavebeengrowinguplandriceformorethan4,000yearsusingtraditionalvarietiesundershiftingcultivationsystemsthatarebasedonslash-and-burnpractices(Zhou1982).Despiteincreasingcommercializationofagriculture,uplandriceproductionisstillprimarilyasub-sistence-orientedactivityinsouthernYunnan,withfarmersusing a large shareof their resources (primarily land andlabor) to meet food needs. Self-sufficiency in food production isadominantobjectiveofuplandfarmers.Income-generatingcommercialactivitiesandeco-reservationactivitiesthusneedtobebasedontheimprovementofhouseholdfoodsecurity,whichoftenmeansthathouseholdsareabletoproduceabulkoftheirownfoodrequirements. In five prefectures (Wenshan, Honghe, Puer, Xishuang-bana,andLincang)ofsouthernYunnan,uplandriceisoneofthethreestaplefoodcrops(irrigatedrice,maize,anduplandrice).Thegovernmentispayingmoreandmoreattentiontopovertyalleviationforpeoplelocatedintheseregionsandthedevelopmentofuplandriceisoneoftheintegratedmeasuresfor becoming self-sufficient in grain production.

94 Tao et al

Target environments and breeding objectives

UplandricecultivationinYunnanextendsfrom21to28οNandfrom73.2to2,200mabovesealevel.Among129rice-growingcountiesinYunnan,uplandricecultivationisprac-ticedin50to60counties,accountingfor13%ofthetotalrice area (Xu 1993). Of 179,866 ha of upland rice in 1982 in Yunnan, 87% was distributed in five prefectures (Puer, Xishuangbana, Wenshan, Lincang, and Honghe) in south andsouthwestYunnanbetween21and24οN,and80%wasgrowninareas800to1,600mabovesealevel.Theseareasaremountainousregionsinhabitedbyminorethnicgroups,andtheyarethetargetpopulationofenvironmentsofouruplandbreedingprograms(Tao1995,Taoetal1993,1996). ThepredominategermplasmforuplandsinYunnanisjaponica.However,indica hybridsfortheirrigatedecosystemhavebeensuccessfullyintroducedtouplandsbelow1,300m(Taoetal1993,1998,1999).TheconstraintstouplandriceproductioninYunnanarepoorsoilfertility,weeds,drought,andblastdisease.Therefore,selectionforimprovedseedlingvigor,droughttolerance,blastresistance,andweedcompeti-tivenessisconsideredinouruplandricebreedingprograms.Ourpreliminaryresearchrevealedthatthebiomassoftra-ditionaluplandricecultivarswasaboutthesameasthatofirrigatedrice,buttheirharvestindexandspikeletsm–2weresignificantly lower than those of irrigated rice. Thus, the ap-proachweapplytoincreasingyieldpotentialistodecreasetheplantheightoftraditionaluplandricecultivarstoavoidlodging,while increasing tilleringabilityandpaniclesize(Taoetal1996).

Approaches

Besidessomeextensionmeasuressuchaschangingslopingfields into terraces, reforming cropping systems, applying chemical fertilizer and herbicides, using irrigated indicahybridsforuplandplanting,andusingbettertraditionalvarie-ties,higherpriorityisalwaysgiventocultivarimprovement.We have had a policy of simultaneous introduction from abroadandbreedingbyourselvessince1989(Huetal1996,1997,Taoetal1993,1997,1998).

Germplasm introduction from abroad

YAAShasbeenanactivememberofINGER(InternationalNetwork for Genetic Evaluation of Rice) and the URRC(UplandRiceResearchConsortium),andthishasprovideda good opportunity for exchange of information, breed-ing materials, and donors; scientific visits; and training of scientists fromYAAS at the International Rice ResearchInstitute (IRRI), Philippines, and at other institutes. This hasbeenanimportantcomponentofourbreedingstrategy,andhasstronglyimproveduplandricebreedinginYunnan.Until2004,3,158accessionsofuplandricewereintroducedfromIRRI(mainlyviaINGERorURRC),theInternationalInstitute of Tropical Agriculture (IITA), the West Africa Rice

Development Association (WARDA), Centro Internacional deAgriculturaTropical (CIAT), Institute for Research inTropicalAgriculture (IRAT, now CIRAD), and Brazil.UplandricebreedersfromIRRI,CIAT,CIRAD,IITA,IRD(Institut de recherche pour le développement), WARDA, and SoutheastAsiavisitedYunnan,anduplandricebreedersfromYAAS visited IRRI, CIAT, CIRAD, IRD, WARDA, Brazil, CostaRica,Côted’Ivoire,Madagascar,Thailand,Indone-sia,andVietnam.SevenyoungscientistsfromYAASweretrainedatinternationalagriculturalresearchcenters.Active,strongcooperationandexchangebetweenYAASandforeignresearchorganizationshavebeensetup. As a result, IRAT104, bred by IRAT/CIRAD andintroduced from IRRI in1991,was releasedby theAgri-cultural Department of Yunnan Province formally in 1995, and it became the first improved upland rice variety released formallybyauthoritiesinYunnan(Huetal1997,Taoetal1997).B6144F-MR-6wasreleasedinYunnanin2000,andhasnowbecomeoneofthemostpopularcultivarsforuplandsinareasbelow1,200m.

Hybridization breeding

ThestrategyofhybridizationbreedinginYunnanistocom-bine theadvantageofgoodadaptationandcold toleranceofYunnantraditionaluplandricecultivarsandgoodplanttype,blastresistance,anddroughttoleranceofuplandriceaccessions(Tao1995,Taoetal1996,1998,1999)introducedfromoutside. Since1989,2,032combinationsofF1havebeenmade,and207advancedbreedinglinesweredeveloped(Taoetal1998).

Donors for intermediate statureFrom207advancedbreedinglinesbred,uplandriceacces-sionsIRAT104,IRAT216,CT8402-27-M-4-2-3M,CT6947-7-1-1-1-7,IR47719-2-2-1-2,CT7979-4-M-7-5-3M,CT9278-13-8-9-3-M,andYunnanirrigatedjaponicaYunjing136werepromising accessions of intermediate stature, especiallyIRAT104andIRAT216,whichwerereallyeffectiveinreduc-ingtheheightofYunnantraditionaluplandrice(Table1).

Donors for blast resistanceScreeningfor resistanceagainstblasthasbeenoneof themajorobjectivesofourbreedingprogram.Severalyearsofscreening at our breeding site in Puer led to the identification ofnineintroducedaccessionsand16Yunnanlocaluplandricecultivarswithresistancetoblast(Table2).Ancestoranalysisof 207 advanced breeding lines indicated that IRAT104,IRAT216,Bayuenou,Boyiegu,Dahongguxuan-3,Huanpigu,Yuanjingdao, and Zaxima were important donors of blastresistance(Table3),andimprovementinblastresistanceisanimportantareaofourbreedingprogram(Table4).


Breeding site selection (hot spot)Abreeding site should closely resemble the targetedareaof cultivation, and constraints such asdrought, blast, andproblemsoilshouldappeareveryyear.Anotherimportantissue is the convenience of operations. Puer is an ideal site forallofthis.

Cross slection of F1 generationGenetic research indicated that general combining ability(gca)wasimportantfortheinheritanceofheadingdate,plantheight,paniclenumberperplant,andspikeletnumberperpanicle(Yangetal1997).Meanwhile,ourbreedingexperi-encealsoindicatedthatcrossselectionfromtheF1generationwassimple,convenient,andeffective(Fig.1).

Generation accelerationIn Yunnan, upland rice could finish one generation within a year.To accelerate breeding progress, in October afterharvest from Puer, seeds can be sown in Sanya, Hainan, and harvested again in March of the next year. In Puer, normal sowingtimeisinApril.Inthisway,twogenerationscanbecompletedwithinayear(Yangetal1995).

Adaptation testSinceG×Eisobviousandimportantforuplandrice,webegan multisite testing in Yunnan (5–10 sites) when a fixed linewasbredtomakeselectionforadaptation.

Performance of improved cultivars

Theresultsofabout25yearsofpracticeindicatedthatwehave partially reached our breeding objectives.Yunlu 29andYunlu52werereleasedin1999and2004,respectively.Bothbecameimportantimprovedcultivarsusedbyfarmersinuplandsabove1,200m. On terraces and favorable lower slopes, and withoptimizedfertilitymanagement,Mengwanggu(traditionaluplandcultivar)usuallyhasabout2 tha–1yield,whereasYunlu29,Yunlu52,andB6144F-MR-6usuallyyield4tha–1(Table5).Improveduplandricecultivars(aerobicrice)havebeenshown tooutyield traditionalvarietiesbymore than1tha–1.Thisdifferenceismainlybecauseoftheincreaseinpaniclesm–2andspikeletspanicle–1(Atlinetal2006,Taoetal1999).

New challenges

AccordingtoYunnangovernmentpolicy,in2005,uplandriceareawastobe667,000hainsteadofthe100,000hain2001.However, all rice was to be planted on terraces or flat fields withnewcultivars,suitableinputs,andappropriatemanage-ment.Averageyieldwasprojectedtobe3tha–1. Traditionally, upland rice breeding was for slopinguplands, emphasizing yield stability, blast resistance, andtoleranceforlowsoilfertility.Foodsecurityoffarmersde-pendsmainlyonplantingareaandtheapplicationofchemical

Table 1. Frequency of intermediate status ances-tors in 207 advanced breeding lines.

Ecotype Cultivar Frequency (%)

Upland rice IRAT104 29.5IRAT216 19.8CT8402-27-M-4-2-3M 8.7CT6947-7-1-1-1-7 6.3IR47719-2-2-1-2 6.3CT7979-4-M-7-5-3M 5.8CT9278-13-8-9-3-M 5.3

Irrigated rice Yunjing 136 16.4IR28 1.9

Table 2. Upland rice resources with blast resistance (scored accord-ing to Standard Evaluation System for Rice).a

Cultivar 1991 1992 1993 1994 1995

BL PB BL PB BL PB BL PB BL PB

OS6 3 3 3 5 – – 1 1 – –IRAT8 – 5 – 5 – – – – – –IRAT104 5 1 1 5 5 – 1 3 3 3IRAT216 3 3 1 5 1 – 1 1 3 3CT9993-5-10-

1-M3 3 1 5 – – 3 1 3 –

CT10035-32-4-4-M

5 3 3 5 – – 1 1 3 –

CT10035-33-6-2-M

3 1 1 5 – – 3 1 – –

CT9278-11-1-1-2-M

1 3 3 5 – – 3 7 3 –

CT9278-11-5-8-2-M

1 1 3 5 – – – 5 3 –

Haobingxuan – 1 – 5 – – – 5 – –Epian – 1 – 5 – – – – – –Bayuenou – 1 – 5 – – – 0 3 3Xiaobaigu – 1 – 5 – – – 1 3 3Zibo – 5 – 5 – – – 1 1 1Ela – 1 – 5 – – – 1 3 1Huanpigu – 1 – 5 – – – 1 3 1Zaxima – 3 – 5 – – – 1 3 1Biangyidabaigu – 3 – 5 – – – – 3 3Haohai – 1 – 5 – – – 0 1 1Dahongguxuan-3 – 3 – 5 – – – 0 – –Hongxuan No.4 – 1 – 5 – – – 0 – –Boyiegu – 3 – 5 – – – – – –Sanbang 70 Lo – 1 – 5 – – – 1 – –Yuanjingdao – 3 – 5 – – – 0 3 1Sanlicun – 3 – 5 – – – – – –

aBL = blast, PB = panicle blast.

96 Tao et al

fertilizersandherbicides.Now,farmershavetoimprovetheirfoodsecurityfromlessland,butwithhighyieldpotential,input-responsivecultivars,andgoodresourcemanagementbecauseofhighpopulationpressureandruraldevelopment.Thus, starting in 2002, all trials had to be conducted onterraces, and the breeding objective shifted to high yield(nitrogen-responsive, lodging-resistant), fine grain quality, andblastresistance.Ofcourse,theindicatype,becauseofitshighinputresponsiveness,becameourmaintargettype. Within a package for upland rice development in Yun-nan,improvedcultivarsplayabasicandirreplaceablerole.

Improvedcultivarscoverabout30%oftheuplandriceareainYunnan. In recent years, we have paid special attention toreturning slope fields to forestry and grasses, constructing terraces,andalleviatingpovertyviatheextensionofscienceandtechnology. Slash-and-burn systems for upland rice have beenreplacedbyintensivemanagementwithnewcultivarsandnew technologies on terraces or flat fields. The food security ofuplandricefarmershasgreatlyimproved.Thus,farmerscouldhavesurplusresourcesforplantingcashcropsanddo-inglivestockfarming.Theeconomicsituationinruralareashasimprovedgreatlybecauseoftheimprovementinfoodsecurity (Kam 2003). But, eco-environment deteriorationhasnotbeenthoroughlyreversed.Theadoptionofapackageforuplandricedevelopmentisstilllow,andthenatureofdiversification inYunnanislittleunderstood. Onterracesandfavorablelowerslopes,andwithop-timizedfertilitymanagement,improved(aerobic)cultivarshavebeenshowntooutyieldtraditionalvarietiesbymorethan1tha–1.But,currently,yieldsofimprovedricecultivarshavereachedaplateauof4to5tha–1 in farmers’ fields (Atlin etal2006).Toconsistentlyobtainhighermaximumyields,however, the current plant type of aerobic rice cultivarswillhavetobefurtherimproved.CurrentlyusedvarietiesB6144-MR-6-0-0andYunlu52arerelativelytall,leafy,andlow-tillering, rarely achievingaharvest indexof0.4or apaniclenumberabove250perm2.Onestrategyforimprov-ingtheyieldpotentialofaerobicricecultivarsistotransfertheiraerobicadaptationtolowlandhigh-yieldingvarieties,whichpartitionupto50%oftotaldrymattertograin,butwhichdonotperformwellunderuplandconditions.

References

Atlin GN, Lafitte HR, Tao D, Laza M, Amante M, Courtois B. 2006. Developingricecultivars forhigh-fertilityuplandsystems intheAsiantropics.FieldCropsRes.97:43-52.

Hu F, Tao D, Yang G, Yang J. 1996. Preliminary studies on the intro-ductionofuplandricefromabroad.SouthwestChinaJ.Agric.Sci.9(2):8-11.

Table 3. Frequency of blast resis-tance ancestors in 207 advanced breeding lines.

Donor Frequency (%)

IRAT104 29.5IRAT216 19.8Bayuenou 13.2Boyiegu 9.2Dahongguxuan-3 6.3Huanpigu 3.9Yuanjingdao 3.9Zaxima 2.4

Fig. 1. Effect of F1 selection. (In 1990 in Simao, 37 crosses with PAcp scale 1–3 were called a better cross, and 168 combinations with PAcp scale 5 were called ok crosses and are shown here. The other 560 crosses with PAcp 7–9 were discarded from 765 F1 cross combinations.)

1990 Summer 1990 Winter

1991 Summer

1991 Winter

1992 Summer

1992 Winter

1993 Summer

1994Summer

1995 Summer

37 F2 (Better)

32 F3 16 F4 15 F5 12 F6 10 F7 8 F8 7 F9

765 F1

168 F2 (OK)

6 F3 6 F4 4 F5 4 F6 4 F7 3 F8

Table 4. Progress of blast resistance improvement (scored according to Standard Evaluation System for Rice).

Year Generation Planted Selected

1992 F2 7.59 7.27F3 7.04 6.29F4 6.73 6.68

1994 F2 5.48 3.76F3 4.27 3.17F4 5.00 5.00F5 3.66 3.00F6 3.88 3.66F7 3.67 2.30


HuF,TaoD,YangG,YangJ.1997.GenealogicalanalysisofIRATupland rice cultivars. In: Poisson C, Rakotoarisoa J, editors. Rizicultured’altitude,ActesduSeminairerizicultured’altitude.Montpellier(France):CIRAD-CA.p181-184.

Kam SP. 2003. Integrated natural resource management for rice produc-tion.Int.RiceRes.Notes28(2):12-18.

TaoD,HuF,NengN,HeY.1993.Introductionofuplandricefromabroad. J. Shanxi Normal Univ. (Nat. Sci. Ed.) 21(Sup.):32-38.

TaoD.1995.Uplandricebreeding.In:JiangZ,editor.RiceinYun-nan. Kunming (China): Yunnan Science and Technology Press. p221-230.

TaoD,HuF,YangG,YangJ.1996.UplandriceresearchinYunnan,China. In: Piggin C, Courtois B, Schmit V, editors. Upland rice research in partnership. IRRI Discussion Paper Series No. 16. Manila (Philippines): International Rice Research Institute. p 96-102.

TaoD,HuF,YangG,YangJ.1997.IntroductionandutilizationofIRAT/CIRAD upland rice cultivars. In: Poisson C, Rakotoarisoa J,editors.Rizicultured’altitude,ActesduSeminairerizicultured’altitude.Montpellier(France):CIRAD-CA.p35-38.

TaoD,HuF,YangJ,ChengH.1998.Uplandricebreedingstrategyin Yunnan, China. Presented at the VI National Rice Research Conference and the Third Upland Rice Breeders Workshop, Goiânia,Goiás,Brazil,9-13March1998.

TaoD,HuF,YangY.1999.Newuplandricecultivarsandtheirculti-vation.ReportedinChina/MyanmarAlternativeDevelopmentCooperationMeeting,5-9April1999,Simao,Yunnan,China.

Xu P. 1993. Eco-characterization of upland rice in Yunnan. J. Shanxi NormalUniv.(Nat.Sci.Ed.)21(Sup.):8-12.

YangG,TaoD,HuF.1995.AresearchonthechangeoftheheadingperiodandplantheightofvariouspaddyriceanduplandriceinSanyaandSimao.J.SouthwestAgric.Univ.17(4):338-343.

YangG,TaoD,HuF,YangJ.1997.Studiesoncombiningabilityofthemaineconomiccharactersinuplandrice.ChineseJ.RiceSci.11(2):77-82.

YunnanSeedManagementStation.1992.RecordsofricecultivarsinYunnan..YunnanSeedManagementStation,Kunming,China.p179-220.

ZhouJ.1982.HistoryandpresenceofuplandriceinYunnan.YunnanAgric.Sci.Technol.5:22-26.

ZhouJ.1987.Regrowingupofuplandriceplanting.YunnanAgric.Sci.Technol.4:22-24.

Notes

Authors’ addresses: D.Tao,F.Hu,,P. Xu, J. Zhou, J. Li, and X. Deng, FoodCropsResearchInstitute,YunnanAcademyofAgriculturalSciences,Kunming650205,China;G.N.Atlin, S. Pandey, Inter-national Rice Research Institute, Los Baños, Philippines.

Table 5. Performance of improved upland rice cultivars in Yunnan (summarized from reports of Yunnan Province Regional Trial for Upland Rice, 1993-2002).

CultivarDays to heading

Plant ht. (cm)

Panicles (no. m–2)

Spikelets (no. panicle–1)

Fertility (%)

1,000-grain weight (g)

Grain yield (t ha–1)

Short-season Long-season

Mengwanggu (check) 103 121 168 119 77 30 2.57 ± 0.10 2.92 ± 0.12Yunlu 29 103 111 245 123 80 27 3.34 ± 0.20 3.29 ± 0.24Yunlu 52 109 114 191 172 78 27 2.86 ± 0.23 4.00 ± 0.28B6144F-MR-6 105 89 266 120 71 23 1.70 ± 0.17 4.13 ± 0.20

98 Suwarno et al

Progress of upland rice breeding in Indonesia since 1991Suwarno, E. Lubis, and B. Kustiano

The major constraints to high yield of upland rice are blast disease and aluminum toxicity in humid regions, drought and short duration of the wet season in arid regions, and shading in interplanting systems. Breeding programs for upland rice to develop improved varieties with high yield and resistance to or tolerance of the respective major constraints have been conducted continuously in Indonesia. Traditional local varieties and introduced varieties from IRRI and other countries were used as genetic sources for the desirable characteristics. Several improved varieties have been released with major improvements in yield, maturity, blast resistance, and tolerance of Al toxicity, drought, and shading. Some promising lines containing two blast resistance genes and tolerance of Al toxicity have been developed and breeding lines with different blast resistance have also been selected. Different sources for blast resistance, including monogenic lines and traditional varieties, were used to diversify the genetic base for blast resistance of the improved varieties. Cultivation of the improved varieties with proper crop management to achieve high yield was demonstrated. Further breeding programs will emphasize blast resistance, Al-toxicity tolerance, drought tolerance, very early maturity, and grain quality. Diversification of blast resistance has been put into action in the breeding program.

the production of upland rice could therefore be achieved through an intensification program using existing cultivation technologies to increase yield and an intensification program for increasing upland rice cultivation area.

A breeding program

The objective of a breeding program for upland rice in Indonesia is to develop improved varieties with high yield potential, good grain quality, early maturity, and resistance to or tolerance of biotic and abiotic stresses that are present in the targeted ecosystems. The breeding program for the arid region, for example, has emphasized drought tolerance and early maturity; for the humid region, blast resistance and tolerance of problems associated with acid soil such as Al toxicity and P deficiency have been the focus; and for inter-planting, emphasis has been on tolerance of shading. All of the programs are conducted continuously at the Indonesian Institute for Rice Research (IIRR); however, the prioritization of objectives has changed over the years. Resistance to blast disease has always had a high priority in the upland rice breeding program. Very early maturity and tolerance of drought had high priority during 1991-94, toler-ance of shading during 1991-96, and tolerance of Al toxicity from 1991 to 2000. Rice research on suboptimal ecosystems, including breeding for upland rice, was intensified recently. The previous breeding objectives were continued but breed-ing for blast resistance, Al toxicity, drought tolerance, and good to premium grain quality were accorded high prior-ity.

Indonesia has an upland area of about 5 million hectares, which can potentially be used for producing food crops. The upland rice cultivation area of the country, however, has not developed rapidly and has remained almost constant at 1.2 million hectares. Farmers in these areas cultivate traditional varieties under low-input conditions and obtain low yields, averaging 2.3 t ha–1. With this condition, upland rice con-tributes only 5% to national rice production. Despite this low contribution, however, upland rice is very important in providing food and as a source of income for people living in dryland areas. The cultivation area of upland rice is distributed over a wide range of agroclimatological zones—from humid areas with eight or more wet months to arid areas with four or fewer wet months. In the humid area of West Java, Sumatra, and Kalimantan, blast and brown spot diseases, soil acidity, Al toxicity, and P deficiency are the major constraints to high yield. In contrast, in the arid areas of Central and East Java, NTB (Western Nusa Tenggara Islands), and NTT (Eastern Nusa Tenggara Islands) provinces, drought is predominant. Breeding programs for upland rice to develop improved varieties with high yield and resistance to or tolerance of the corresponding major constraints have been conducted con-tinuously over a long period of time in Indonesia. Improved varieties have been released and several promising breeding lines are being further evaluated to select better varieties. Cultivation of improved varieties with proper manage-ment has been demonstrated in some locations to achieve high yields of more than 5 t ha–1. This indicated a high yield gap between research and farm levels for upland rice. Increasing

Progress of upland rice breeding in Indonesia since 1991 99

Most of the breeding programs were conducted follow-ing a modified bulk method in which populations of early F2–F5 generations were grown with closed spacing. Panicle selection was conducted on F5 populations followed by pedi-gree selection and observation yield trials, preliminary yield trials, and advanced yield trials. To speed up the breeding process, the bulk populations were grown under irrigated lowland conditions during the dry season. Various genetic sources have been used in breeding for important characteristics of upland rice (Table 1). The approach of breeding for blast resistance was modified from developing varieties with resistance to many blast races or pyramided resistance genes to varieties with different resist-ance genes, that is, diversification of blast resistance. Differ-ent sources for blast resistance, including those with identified genes and traditional varieties, were used for crossing (Table 2).

Released varieties

Thirteen improved varieties of upland rice have been of-ficially released in Indonesia since 1991 (Table 3). All of these varieties had resistance to blast disease and other de-sirable characteristics related to the targeted environments. The approach taken in breeding for blast resistance was to pyramid resistance genes. Danau Tempe, which was resistant to six races of blast pathogen, was released in 1991. Varie-ties with a wider spectrum of blast resistance were released later, namely, Way Rarem, Jatiluhur, and Batutegi, which were resistant to 10 races of the pathogen. The cooked-rice texture of the varieties differs in accordance with the diverse preference of consumers. Silugonggo was selected from a breeding line in-troduced from IRRI: IR39357-71-1-2-2. This variety had resistance to blast and bacterial leaf blight and had a very early maturity of 90 days. It could be adapted to upland areas mainly in Java, where Al toxicity is not prominent. Some programs to intensify upland rice cultivation by introducing improved varieties and cultivation technologies have been conducted. The introduction of newly released up-land rice with drought tolerance and early maturity was done in arid regions of East Java in 1994. Programs to introduce the improved varieties Way Rarem and Jatiluhur, which were tolerant of blast, Al toxicity, and shading, were implemented in the newly developed area for estate crops and Imperata sp. involving idle lands with a total area of 100,000 ha and 215,000 ha during the planting seasons of 1994 and 1995, respectively. The program succeeded and high yields of 3–4 t ha–1 fresh rice were reported (Directorate of Estate Crop Production 1995). Most of the programs, however, were not sustained and farmers in upland areas still cultivate traditional varieties with low inputs, leading to low yields. These low yields might also be caused by a breakdown in disease resist-ance (Amir and Nasution 1995) or a decrease in soil fertility due to a reduction in soil organic matter content (Adiningsih and Mulyadi 1993, Hidayat et al 2000).

Table 1. Genetic sources for some desirable characteristics used in the breeding program for upland rice in Indonesia.

Desirable characteristic Varieties

Drought tolerance Salumpikit, Cabacu, Gajah Mungkur, Kalimutu, ICOXI-B-66, Lagos

Blast resistance Carreon, Tetep, Tadukan, Klemas, Jambu, Mat Embun, Cabacu, Cuil, Danautempe, C101LAC, C105 A51, Sayap, Bonti, Jadah, Perak, Raden Intan, Sejang Ungu, Grogol, Dupa

Tolerance of low pH and low pH and Al toxicity

Seratus Malam, Hawara Bunar, Ketombol, Lawean, Meulaboh, Dupa, Simariti, IRAT352, IRAT379, Gro-gol, Simedan

Grain quality Cabacu, Gajah Mungkur, Membramo, CT6510-24, Ketan Tuban, Seratus Malam

Aromatic rice Simariti, Dupa, Mesir, Bengawan Solo

Table 2. Rice varieties used as genetic donors for blast resistance in the breeding program for upland rice in Indonesia.

Name Type of variety

Name Type of variety

Bonti Traditional Sirendah Traditional Bulan Sabit Traditional Sirendah Pulen Traditional Cabacu Introduced Tetep Traditional Dupa Traditional Asahan Improved Grendel Traditional Limboto Improved Grogol Traditional Wayrarem Improved Jambu Traditional Batu tugi Improved Ketombol Traditional IRBL 8 Monogenic (Pik-h) Klemas Traditional IRBL 10 Monogenic (Piz-5) Lampung Arak Traditional IRBL 19 Monogenic (Pi3) Lampung Kuning Traditional IRBL 23 Monogenic (Pi12(t)) Lampung Putih Traditional IUF5 1 Monogenic (Pii)Malio Traditional IUF5 70 Monogenic (Pi7(t))Simacan Traditional

100 Suwarno et al

Technologies for rice cultivation as a component of a farming system aimed at sustaining or improving soil productivity are available (Toha and Fagi 1995). Upland rice cultivation based on integrated crop management with improved varieties has also been tested in farmers’ fields, obtaining high yields of 3.88–4.74 t ha–1 in the wet season (WS) of 2002-03, which increased to 5.52–6.20 t ha–1 in the WS of 2003-04 (Table 4) (Toha, H.M., personal communica-tion). Blast resistance of varieties broke down after they were cultivated in a wide area for several consecutive years (Amir and Nasution 1995). The development of improved varieties with different blast resistance genes could be an appropriate approach for a breeding program on upland rice. Thus, such a program was started and some selected breeding lines have been obtained.

Table 3. Improved varieties of upland rice officially released in Indonesia since 1991.

Variety Year of release

Maturity (days)

Cooked-rice texture

Yield (t ha–1)

Resistance and tolerancea

Danautempe 1991 115 Hard 2.5–3.5 Bl, AlGajah Mungkur 1994 95 Medium 2.5–3.5 Bl, DrWay Rarem 1994 105 Hard 3.0–4.0 Bl, Al, FeKalimutu 1994 95 Medium 2.5–3.5 Bl, DrJatiluhur 1994 115 Hard 2.5–4.5 Bl, ShCirata 1996 120 Medium 3.0–5.0 Bl, BPH, ShLimboto 1999 105 Hard 3.0–5.0 Bl, Al, DrTowuti 1999 120 Soft 3.0–5.0 Bl, BPH, BLBBatutugi 2001 116 Medium 3.0–5.0 Bl, Al, DrDanau Gaung 2001 113 Medium 3.4–5.0 Bl, BS, Al, FeSilugonggo 2001 90 Medium 4.5–5.5 Bl, BLBBatang Gadis 2002 122 Medium 3.6–5.6 BlSitu Bagendit 2002 105 Soft 3.0–5.0 Bl, BLB

aBl = blast, BS = brown spot, BPH = brown planthopper, Dr = drought, Al = aluminum toxicity, Fe = iron toxicity, Sh = shading.

Table 4. Average yield of improved upland rice va-rieties cultivated with integrated crop management technology in Seputih Rahman, Lampung, WS 2003-04.

Name of farmer

Yield (t ha–1)

Batutegi Limboto Situ Patenggang

Average

Wira 6.24 5.81 5.71 5.92Era 6.37 5.25 5.67 5.76Eri 6.20 5.72 5.29 5.74Pt. Sukasi 5.98 5.54 6.02 5.85Ayu 6.29 5.93 4.47 5.56Oka 6.31 5.95 5.20 5.82Budiarte 6.25 6.52 6.33 6.37Agus P 5.96 7.02 5.52 6.17 Average 6.20 5.97 5.52 5.90

Promising lines and other breeding materials

Breeding for upland rice is conducted continuously and pro-duces promising lines and other materials at different stages in the breeding pipeline. All of the promising lines tested in yield trials at Tamanbogo and Sukadana, Lampung (Table 5), were resistant to blast, with some of them containing the two resist-ance genes, Pi-1 and Pi-2, and others also having tolerance of Al toxicity. Eight promising lines had high yield above 4.0 t ha–1, which was comparable with that of the improved check variety Limboto and higher than the yield of traditional va-riety Sirendah. Participatory varietal selection that began in the wet season of 2001-02, in which farmers were allowed to select their preferred promising lines (Suwarno et al 2002), was also applied to the promising lines. Most of the released varieties have cooked-rice texture of medium to hard. The preferences of rice consumers in Indonesia differ across regions. In Java and Bali, consumers prefer soft rice, while in Sumatera and Kalimantan they prefer hard rice. Many farmers living in Sumatera and Kalimantan came from Java and Bali and thus prefer soft rice. Breeding lines with good phenotypic performance and amylose content associated with soft to medium rice texture were selected (Table 6). These lines were resistant to both leaf and neck blast and tolerant of Al toxicity. The lines had medium grain shape, medium to long grain length, and small to medium chalkiness. The most popular variety in Indonesia, IR64, had long grain with medium shape and medium chalkiness. The premium quality rice Cianjur, on the other hand, had intermediate length, shape, and chalkiness. To develop improved varieties with different blast resistance, the existing breeding lines were screened by inoculating them with combinations of two blast races. A total of 479 breeding lines derived from different crosses


were obtained (Table 7). More systematic crosses involving monogenic lines to develop improved varieties, each contain-ing two resistance genes, were undertaken.

Exchange of breeding materials

The exchange of breeding materials was conducted mainly through activities or networks coordinated by IRRI, includ-ing the International Network for Genetic Evaluation of Rice (INGER), the Asian Rice Biotechnology Network (ARBN), the Upland Rice Research Consortium (URRC), and the Con-sortium for Unfavorable Rice Environments (CURE). Many varieties and breeding lines exchanged were evaluated and used in Indonesia. For example, improved variety Silugonggo was selected from a breeding material obtained from IRRI. Many other varieties and lines received from IRRI were also used as genetic donors in the breeding program for upland rice in Indonesia. Some of these were Cabacu, ICOXI-B-66, and Lagos for drought tolerance; IR60080-23, IRAT352, and IRAT379 for Al-toxicity tolerance; and Cabacu and CT6510-24 for grain quality.

Table 5. Promising lines and check varieties of upland rice tested in yield trials in Tamanbogo and Sukadana, Lampung, during WS 2004-05.

Line/variety Maturity (days)

Yield (t ha–1)

Resistance/tolerance trait

Bio528B-TB-12-1-1 102 4.76 R to blast; Pi-1 and Pi-2 genes

Bio511B-61-2-3-1 110 4.07 R to blast; Al tox.; drought Bio511B-5-12-5-1 111 3.63 R to blast; Al tox.B9071F-B-7 103 4.64 R to blast; Al tox.Bio511B-61-2-4-1 109 2.98 R to blast; Al tox.Bio512-MR-1-4-PN-26 109 3.33 R to blast; Al tox.Bio530B-39-3-6 108 3.08 R to blast; Pi-1 and Pi-2

genesBio530B-5-6-5-4 109 4.51 R to blast; Pi-1 and Pi-2

genesBio530A-5-14-2-2-8 107 4.01 R to blast; Pi-1 and Pi-2

genesTB360B-TB-26-1 103 3.86 R to blast; Al tox.TB393B-TB-17-1 108 4.73 R to blast; Al tox.TB396B-TB-14 111 4.40 R to blast; Al tox.TB437B-TB-1 109 2.76 R to blast; Al tox.TB356B-TB-18-3 103 4.30 R to blast; Al tox.Limboto 105 4.42Sirendah 116 3.33

Table 6. Selected breeding lines resistant to leaf and neck blast and tolerant of Al toxicity en-tered in preliminary yield trials during WS 2005.

LinePlant

height (cm)Maturity (days)

Resistancea Grain characterb

LB NB Al Amylose (%)

Length Shape Chalkiness

TB364C-TB-12-2 121.4 100 R R R 21.4 M M MTB393C-TB-2-2 109.0 102 R R R 21.4 M M MIR55423-01 103 104 R R R 16.8 M M MTB356B-TB-20-3-1 106 98 R R R 21.2 M M MTB356B-TB-47-1-1 127 102 R R R 21.2 M M MTB356B-TB-47-3-2 98 96 R R R 19.2 M M MTB361B-TB-17-2-2 14 100 R R R 21.6 M M STB361B-TB-30-6-2 109 100 R R R 24.1 M M MTB366B-TB-2-1-3 124 104 R R R 21.3 M M MTB425B-TB-6-4-3 119 101 R R R 23.0 M M MBP1966B-12-5-TB-1 110 102 R R R 21.5 M M SBP1978B-24-5-TB-1 114 103 R R R 21.5 L M STB59-TB-16-1-1 111 103 R R R 24.9 L M SBP241D-TB-18-6-1 110 97 R R R 20.1 M M MBP1351D-1-2-PK-3 102 102 R R R 20.1 M M MTB356B-TB-20-4 101 100 R R R 18.2 M M MTB356B-TB-47-3 94 104 R R R 19.2 M M MTB356B-TB-52-2 103 104 R R R 20.7 M M MTB360B-TB-26-1 93 102 R R R 20.7 M M MTB396B-TB-14 94 110 R R R 19.1 M M MBP1976B-23-7-TB-1 100 102 R R R 20.8 M M MTB437B-TB-5 80 106 R R R 20.1 M M MTB457B-TB-2 133 102 R R R 17.8 M M SIR64 L M MCianjur rice M M M

aLB = leaf blight, NB = neck blast, Al = aluminum. bM = medium, L = long, S = small.

102 Suwarno et al

Monogenic lines for blast resistance obtained from IRRI were also evaluated for their reaction to different blast races (Table 8). These monogenic lines were useful not only as genetic donors in the upland rice breeding program but also in the characterization of blast pathogen populations in Indonesia.

A future breeding program

The constraints of blast disease and acid soil–related prob-lems are still prominent in the existing and potential areas for upland rice cultivation in humid regions. Thus, the breeding program for upland rice in Indonesia has put great emphasis on blast resistance and Al-toxicity tolerance. The blast pathogen has many races that are capable of breaking down the resistance of improved rice varieties after cultivating them for several consecutive planting seasons. Diversification of blast resistance through the development of improved varieties containing two resistance genes originat-ing from selected monogenic lines and varieties with blast resistance originating from different traditional varieties will be continued. In the arid region, blast disease is not severe but drought and the short duration of the rainy season are major problems. To provide better and improved varieties for this region, the breeding program for drought tolerance and very early maturity will be intensified. One of the reasons that high-yielding and improved varieties could not be adopted by farmers in upland areas is their grain quality, which does not satisfy farmers’ require-ments. Traditional varieties cultivated by farmers have good grain quality, are preferred by consumers, and are frequently

more expensive than improved varieties. The incorporation of good grain quality should be included in the development of improved upland rice varieties for both humid and arid regions. Moreover, farmers’ preferences could be different not only in terms of grain quality but also in terms of the performance of an improved variety itself. To meet farm-ers’ preferences, participatory varietal selection that already started will be continued.

References

Adiningsih SJ, Mulyadi. 1993. Alternatif teknik rehabilitasi dan pe-manfaatan lahan alabg-alang. Prosiding Pemanfaatan Lahan Alang-Alang untuk Usahatani Berkelanjutan. Pusat Penelitian Tanah dan Agroklimat, Bogor. Hal. 29-49.

Amir M, Nasution A. 1995. Status dan pengendalian blas di Indonesia. Hal. 583-592. In: Dalam Syam M et al, editors. Kinerja Penelitian Tanaman Pangan. Buku Puslitbangtan. Badan Penelitian dan Pengembangan Pertanian.

Amir M, Nasution A. 1997. Pemanfaatan galur-galur monogenik padi (Oryza sativa L.) untuk pengendalian penyalit blas (Pyricularia grisea). In: Proceedings of the Indonesian Phytopathological So-ciety Congress and National Seminar, Palembang, 27-29 October 1997. Sriwijaya University, Palembang. p 171-177

Hidayat A, Hikmatullah, Santoso D. 2000. Potensi dan pengelolaan la-han kering dataran rendah. In: Adimihardja, editor. Sumberdaya Lahan Indonesia dan Pengelolaanya. Pusat Penelitian Tanah dan Agroklimat, Badan Litbang Pertanian. Hal. 197-225.

Tabel 7. Number of F6 selected breeding lines resistant to the respective combination of P. grisea races and their major parental varieties, WS 2003-04.

Race combination Number of selected lines

Major parental varietiesa

123 and 173 182 TB154E-TB-2/IRAT144//IRAT379 (18); IR60080-23/BP303 (17); Memberamo/TB154E-TB-2 (17); G.M./Cabacu//B. Sabit/Memberamo (16)

001 and 031 36 Asahan/Dupa//GM/Cabacu (6); G.M./Cabacu//B. Sabit/ Mem-beramo (5); Bonti/G.M. (4)

021 and 033 120 G.M./Cabacu//B. Sabit/Memberamo (18) Memberamo/IR60080-23 (16); TB154E-TB-2/ IRAT144//IRAT379 (12)

023 and 100 141 Bonti/Malio//G.Mungkur/IRAT144 (16 galur) Memberamo/TB154E-TB-2 (15) B8503E/IAC25 (14); Asahan/Dupa//GM/Cabacu (14)

aNumber in parentheses indicates the number of lines.


Table 8. Reaction of monogenic lines to 20 races of the blast pathogen.

LineP. grisea racea

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

1. C101 A51 (p-1) R R R R R R R R R R R R R R R R R R R R 2. C102 A51 (p-2) R S R R R R R R R R R R R R S R R R S S 3. C103 A51 (p-3) S S R R R R S R R R R R R R R R R R R R 4. C104 A51 (p-4) S S R R S R R R R R R R R R R R R R R R 5. C105 A51 (p-5) R R R R S R R R R R R R R R S R R R S R 6. C 101 LAC (p-6) R R S R R R R R R R R R R R R R R R R R 7. C104 LAC (p-6) R R S R R R R R R R R R R R R S R R R S 8. C101 PKR (p-8) R R R R R R R R S R R R R R R S S S S S 9. C102 PKR (p-9) R R R R S S R R R R R R R R R R S S S S10. C104 PKR (p-10) R S S R S S R S R R R R S S R R R S S S11. C. 101 RRP-1 (p-11) R S S S R R R R S R S R S S S S S S S S12. C101 RRP-2 (p-12) R S S R R S R R R R R R S R R S S S R S13. C101 RRP-3 (p-13) S S S R S S R S S R S R S S S S S S S S14. C101 RRP-4 (p-14) S S S R S R R R S R R R R S S S S S S S15. C101 RRP-6 (p-15) S S S R S R R S R R R R R R R S R R S S16. C101 RRP-6 (p-16) R S S R R R R R R R R R S R S S S S R S17. C103 RRP (p-17) R R R R R R R R R R R R R R S R R R R R18. C105 RRP -1 (p-18) S R S R S R R R S R R R R R R S S S S S19. C105 RRP-1 (-19L9) R R S R S R R R S R R R R R R S S S S S20. C105 RRP-2 (p-19L23) S S S S S S R S R R R R R S R S R R S S21. C105 RRP-4 (p-20L6) R R S R R R R S R R R R R R R S S S S R22. C101 RRP-4 (p-20L6) S S S S S R S S R R S R R R R S S S S S23. Asahan R R R R R R R R R R R R R R R R R R R S24. Kencana Bali S S S S S S S S S S S S S S S S S S S S

a1. ID-14 (15) Lampung, 2. IG-2 (24) Sukabumi, 3. IG-1 (260) Bandung, 4. IC-15 (39) U. Pandang, 5. ID-13 (60) Cianjur, 6. 001 Sukabumi, 7. 003 Sukabumi, 8. 013 Sitiung, 9. 021 Sukabumi, 10. 023 Sukabumi, 11. 031 Sitiung, 12. 033 Sitiung, 13. 100 Kalimantan, 14. 103 Kali-mantan, 15. 121 Kalimantan, 16. 123 Sukabumi, 17. 133 Kalimantan, 18. 161 Sukabumi, 19. 173 Kalimantan, 20. Sitiung.Source: Amir and Nasution (1997).

Suwarno, Kustianto B, Arjasa WS, Atlin GN. 2002. Participatory selection on upland rice in Sumatera. In: Witcombe JR, Parr LB, Atlin GN, editors. Breeding rainfed rice for drought-prone environments: integrating conventional and participatory plant breeding in South and Southeast Asia. Los Baños (Philippines): International Rice Research Institute. p 61-63.

Toha HM, Fagi AM. 1995. Budidaya tanaman pangan dan sistem usahatani konservasi di DAS Jratunseluna bagian hulu. Kinerja Penelitian Tanaman Pangan. Badan Litbang Pertanian. Buku 3:810-824.

Notes

Authors’ address: Indonesian Center for Rice Research (ICRR), Suka-mandi, West Java, Indonesia.

104 Sinha et al

Breeding rice for the Indian plateau uplandsP.K. Sinha, M. Variar, and N.P. Mandal

trient management, and blast. Subsequently, in the second and third phases of the consortium, the scope was widened to incorporate research lessons from other consortium sites to develop drought-tolerant breeding lines as well as crop resource management techniques.

Drought occurrence

Spatial heterogeneity and seasonal variation in rainfall pat-terns profoundly affect the growth of upland rice, especially in the plateau regions of Jharkhand, Orissa, and Chhattisgarh. One of the activities in the first phase of the collaborative program was, therefore, site characterization, especially quantification of the magnitude of drought at various phe-nological phases of upland rice at Hazaribag. The monsoon months of June-October (22–43 metro-wk) were divided into initial (7 wk), intermediate (8 wk), and terminal (7 wk) phases and the frequency of occurrence of drought (number of weeks in which no rainfall was recorded) in each phase was analyzed covering 1972 to 1991. The probability of drought spells of 2−4-wk duration in the initial phase and of 3−6-wk duration in the terminal phase was high. Drought spells of 7-wk duration occurred rarely (5% probability) in the initial or terminal phase. Drought spells were infrequent in the intermediate phase, with low variability, whereas drought spells coincid-ing with germination and establishment of the rice crop and at the reproductive stage were frequent and highly variable. At 50% probability level, in this agro-climatic region, 14 d of maximum drought spell were generally expected during the latter part of the crop-growing period. Though the number of rainless days or the percentage of below mean rainfall described only part of the complex phenomenon of water balance, it was nevertheless an important criterion in the development of a crop management strategy.

Drought-tolerant genotypes

Assessment of genetic variability for drought tolerance in upland rice and the identification of traits associated with drought tolerance were taken up initially with selected aus, indica, and japonica genotypes with known drought toler-ance and later with mapping populations for marker-aided selection of quantitative traits associated with drought tol-

Rice area in India has fluctuated between 40.2 million ha (2002-03) and 45.1 million ha (an all-time high in 1999-2000) since 1994 (DES 2004), while production varied between 72.65 million t in 2002-03 and 93.34 million t in 2001-02. Variation in rice production was closely related to total annual rainfall. Nearly half of the total rice area is rainfed, located mostly in the eastern Indian states of Assam, Chhattisgarh, Jharkhand, Madhya Pradesh, and Orissa; altogether, they have less than 25% of the area under irrigation. Rice produc-tion dropped by more than 20 million t in the drought year of 2002-03, when deficiency in rainfall was about 19% of the normal. Of this, about 14 million t were lost in the pre-dominantly rainfed eastern Indian states. Uplands constitute about 15% of the rice grown under rainfed conditions and the rest is lowland (30%) and flood-prone (7%). Rice yields in the rainfed ecosystems are low and fluctuate greatly from year to year. Yield in regions with irrigation is 3.2 t ha–1; the range is from 0.6 to 1.5 t ha–1 in uplands, 0.9 to 2.4 t ha–1 in rainfed lowlands, and 0.9 to 2.0 t ha–1 in deepwater areas. Average rice yields in eastern India vary from 0.8 to 1.96 t ha–1 (Shobha Rani et al 2002). Upland rice is grown on flat lands in coastal Orissa, Assam, and eastern Uttar Pradesh; on gently rolling lands (up to 8% slope) in Chhattisgarh and Madhya Pradesh; and on sloping lands (>30% slope) in Jharkhand, western Orissa, Meghalaya, and Uttaranchal hills. Around 70% of the upland rice area is drought-prone, with the other 30% in coastal Orissa, Himachal Pradesh, Uttaranchal, and Bihar having access to irrigation. In northeastern India, around 550,000 ha are under shifting cultivation (Singh 2002). Intermittent drought spells, weed problems, declining soil fertility and productivity, and inability of farmers to invest in inputs are the major constraints to productivity in the upland ecosys-tem. These technical constraints are important not only in India but also in other upland rice-growing regions of South and Southeast Asia. The commonality of constraints across regions, in rainfed uplands as well as in lowlands, was the raison d’être for the establishment of the upland and lowland rice research consortia (URRC and LRRC) by the Philippines-based International Rice Research Institute. The Central Rainfed Upland Rice Research Station in Hazaribag is the lead center for consortium activities on upland rice in India, with drought as the major mandate for research. Other consortium partners focused on weed problems, nu-

Breeding rice for the Indian plateau uplands 105

erance. A repeatable drought condition in a defined growth stage was difficult to achieve in outdoor screening nurseries during the crop season. Delayed sowing to coincide with terminal drought was compounded by other stresses—for example, leaf and neck blast interfered with the evaluation of leaf drying under drought stress. The success of selecting tolerant genotypes, therefore, depended on the incidence and intensity of drought during the season. Among the selection parameters, delayed rolling/absence of rolling and wilting and drying under prolonged drought were the most discernible features easily scored in the field. Several upland indica geno-types (CR143-2-2, RR18-3) did not wilt and had low grain sterility under drought, comparable with known drought-tolerant tropical japonicas (IAC25, IAC1131, Kinandang Patong, Salumpikit). Several breeding lines from IRRI also had good tolerance for drought and had the same yields as local checks, but they were longer in duration (Table 1).

Drought-tolerance traits

Root parameters obviously were the most difficult to measure in the field. Selected upland rice genotypes were grown in polyethylene tubes (14 cm diameter, 1 m length) filled with compacted field soil and placed in similarly sized holes. Two sets were maintained: drought was induced in one set by withholding watering 35 d after sowing; the other set was watered regularly. Root length was measured after the stress period by washing off the soil from the plants. Root and shoot volume were also recorded in the stressed and nonstressed plants. In general, drought stimulated root growth in tolerant genotypes; Vandana recorded a 15% increase in root length under stress, whereas susceptible check Co 13 and traditional upland drought-tolerant cultivar Brown Gora had less root length under stress. Root volume was highest in IRAT112 (not stressed) and Salumpikit (stressed), mostly distributed in the upper 30-cm segment. Root thickness in 60-d-old plants, measured with a magnifier, also differed significantly with some drought-tolerant genotypes (OS6, VHC1253, IRAT112,

Sathi 34-36) exceeding 1.5 mm thickness, 5 cm below the first node. The ability of tissues to withstand dehydration (dehy-dration tolerance) was measured in several genotypes by ex-posing detached leaves to desiccation. Leaf water content in excised leaves, measured as the difference in weight between turgid and rolled leaves, was also correlated with drought scores of the genotypes. Japonica and drought-tolerant indica genotypes maintained high leaf water content up to 145 min of excision. The rate of water loss decreased after comple-tion of leaf rolling. A low rate of percentage water loss was recorded in moderately drought-tolerant genotype Azucena and drought-tolerant genotypes Annada and IRAT112. An-nada (indica genotype) transpired small amounts of water, whereas japonica varieties Azucena, IRAT112, and IRAT216 transpired a greater quantity of water during the period of drought stress. Courtois and Lafitte (1999) compared earlier reports of root system development and osmotic adjustment in a broad sample of varieties and concluded that japonica types generally have a deep, thick root system and good penetra-tion but poor dehydration tolerance and osmotic adjustment. Indica types have shallow, thin root systems but they show better dehydration tolerance and osmotic adjustment. Even though the evaluation at Hazaribag was limited to a few genotypes, the results confirmed the findings of Courtois and Lafitte earlier. Since dehydration tolerance is readily measured in the laboratory, mass screening of genotypes is feasible.

Weed competitiveness

Weeds are next only to drought as a limiting factor in upland rice production. Consortium activities therefore emphasized the identification of weed-competitive genotypes and traits associated with competitiveness. Aus genotypes were more weed-competitive (Brown Gora, Black Gora, Birsa dhan 102, Saita, Kalakeri, Sathi 34-36, VHC1253, RR174-1) and tropical japonicas were the least competitive when they were screened under natural populations of weeds in unweeded

Table 1. Drought-resistant breeding lines of intermediate stature.

Designation Plant height (cm)

Duration (d)

Grain typea

Drought reactionb

IR58662-05 97 120 MS RB2997c-TB-60-3-(ACI-2)-C 81 116 LS MRIR57918(ACI-6)-C 85 112 LS MRIR57920(AC25-2)-C 87 114 LS MRCT6510-24-1-2-19 73 114 LB MRIR60080-32 87 112 LB RIR55435-05 70 112 LB MRIR55419-04 86 108 LB MRKalinga III (high-yielding check) 108 92 LS SBrown Gora (traditional check) 118 95 LB MR

aLB = long, bold, LS = long, slender, MS = medium, slender. bR = resistant, MR = moderately resistant, S = susceptible.

106 Sinha et al

upland plots. A few indica genotypes also showed weed com-petitiveness (Salumpikit, Vandana). Experiments conducted in 1991-92 also revealed that, besides Brown Gora, improved genotypes Vandana and RR51-1 were also weed-competitive in terms of early emergence of plumule and initial higher leaf area index, tiller number, number of roots per plant, and root depth. However, no genotype had a weed-suppressing effect in terms of a reduction in weed biomass. The traits contribut-ing to weed competitiveness could not be identified under unweeded conditions, but, under once-weeded conditions, weed-competitive genotypes had early seedling vigor, greater plant height, and higher biomass. Differences in weed biomass at different crop growth stages in competitive (Brown Gora) and noncompetitive (Heera) cultivars showed that the former had better weed-suppressing ability at all stages. The slope of the weed sup-pression curve, in the case of the weed-competitive cultivar at 35, 50, and 65 DAS, was significantly higher when seed density increased from 100 to 300 seeds m–2. There was a corresponding increase in rice biomass at these seed rates, with a steep slope. Genotypes with similar weed suppression curves at normal seed rates would be expected to have good rice competitiveness. In the case of the noncompetitive cul-tivar, however, the total weed biomass was higher than that of rice biomass at all seed densities of rice. An analysis of traits contributing to rice biomass indicated that, in the case of the weed-competitive cultivar, a maximum reduction in plant height occurred at normal seed rates, whereas, at higher seed rates (400–500 seeds m–2), plant height increased, espe-cially at the later stages of crop growth. Plant height generally increased in the case of the noncompetitive cultivar at higher seed densities. The reduction in tiller number was highest at lower plant densities and minimum at 500 seeds m–2. Changes in leaf number and area also showed a similar trend.

Techniques for mass screening for rooting depth under drought

Reliable field-screening techniques for mass screening of populations for drought are yet to be developed. A herbicide injection technique to measure rooting depth was evaluated. Here, herbicides were injected at specific depths (50, 75, and 100 cm) and symptoms of injury on the canopy were scored, an indication that roots have reached those depths. Based on symptom appearance at 50 and 75 cm, Moroberekan, Vandana, CR143-2-2, Sathi 34-36, Saita, Black Gora, and Kalakeri were identified as deep-rooted genotypes. Minor modifications in the methodology were attempted to improve repeatability, but the technique was not standardized and did not appear feasible for screening large populations. Depletion of water supply in the reproductive phase due to drought spells can reduce reproductive growth and cause severe yield reduction. It is reported that, under such situations, transfer of pre-anthesis assimilates to the grain increases two- to threefold and that variation for this character might exist in a broadly based gene pool (Turner 1982). An

indirect screening technique to evaluate genotypes for their ability to translocate carbohydrates from the stem to the grains was examined. Chemical and mechanical defoliation at 5 d after anthesis in 15 selected upland rice genotypes revealed highly significant differences among the genotypes in terms of reduced grain number, test weight, and yield. In general, aus genotypes were more efficient in translocating stem-reserve carbohydrates to grain in upland genotypes than indica or japonica ones. The reduction in grain yield (Fig. 1A) and thousand-grain weight (Fig. 1B), as a result of defoliation treatment, was lowest in Kalakeri and Sathi 34-36, respectively. The magnitude of the percentage reduc-tion in yield due to defoliation was greater for yield than for thousand-grain weight. This may indicate that defoliation treatment causes complete sterility of the florets of late-formed tillers or induced floret abortion in the higher-order florets. The highly significant genotype × treatment interac-tion indicates that this technique can be used to discriminate genotypes based on their response to post-anthesis drought. Both techniques appeared effective in revealing genetic variation in post-anthesis stress tolerance. Refinement and precision in the methodology were essential in the case of chemical defoliation. Mechanical defoliation was, on the other hand, labor-intensive and less likely to be effective for mass-screening populations.

Marker-aided selection for drought and blast

Work with two crosses (IR64/Azucena and Co 39/Morob-erekan) developed at IRRI began at Hazaribag (Table 2) in 1995. This site had natural drought occurrence and its soil type differed from that in the Philippines. Drought scores, leaf rolling, and root thickness were recorded in these populations. Eight markers were identified for root thickness, measured at 30, 40, and 60 DAS. Regression on flanking markers resolved six segments, two on chromosome 4 and one each on chromosomes 1, 3, 6, and 12. Four of these segments overlapped with those identified for drought scores and leaf rolling. Alleles coming from the best parent (Moroberekan) exerted a positive effect in all cases. Champoux et al (1995) reported 26 markers associated with root morphology in Co 39/Moroberekan in greenhouse experiments at IRRI. At 30, 40, and/or 60 DAS, most of the markers associated with root thickness were commonly identified in our investigation. QTLs associated with root thickness were located on the same chromosomal region as leaf rolling and drought avoidance, suggesting a clustering of different QTLs. Major genes and QTLs associated with blast resistance, effective under rainfed upland conditions at Hazaribag, were first detected in Co 39/Moroberekan recombinant inbred lines (Variar et al 2002). The stability of these QTLs was studied in multilocation testing in eastern India. Further, a Vandana/Moroberekan advanced backcross population was developed to improve the blast tolerance of Vandana and identify QTLs associated with blast resistance. This population was phenotyped in several locations in India and also at IRRI,


Philippines, for drought and blast. Field-testing of BC3F4 lines identified several lines with partial resistance to blast and good agronomic acceptability. Some of the lines showing partial resistance and carrying different introgression region genes were intercrossed to accumulate the introgression region in a common background. Multilocation testing of the derived lines in blast-endemic locations in India and assessment of drought and grain quality characteristics at IRRI in 2004 led to the selection of several agronomically acceptable lines with blast resistance.

Germplasm exchange

The upland ecosystem is structured into three zones climati-cally—subhumid mainland Southeast Asia, dry plateaus of eastern India and Bangladesh, and equatorial humid areas of Indonesia, South Vietnam, and southern Philippines. Upland rice varieties grown in the semihumid subecosystem are generally traditional tropical japonicas (panicle-weight type, glutinous rice), whereas, in the eastern Indian plateaus, tra-ditional varieties are aus and the few high-yielding types are indica (short-duration, panicle-number type). In the equato-rial humid subecosystem, farmers grow indicas and also some improved japonicas, but tropical japonicas are not adapted

Table 2. Mapping populations studied at CRURRS, Hazaribag, for drought- and blast-toler-ance QTLs.

Designation Population Characters

IR64/Azucena Doubled-haploid lines Components of drought toleranceCo 39/Moroberekan Recombinant inbred lines Components of drought tolerance and

blast resistanceAzucena/Bala Recombinant inbred lines Blast resistanceVandana/Moroberekan Advanced backcross (BC2F4

and BC3F4)Blast resistance, identification of superior

transgressed lines with breeding value. Components of drought tolerance

Table 3. Donors that confer resistance to abiotic and biotic stresses (selected from breeding network, 1999-2000).

Variety OriginDuration

(d)Plant height (cm)

Drought. score (0−9)

Brown spot score (0−9)

Blast score (0−9)

CT10006-7-2-M-5-1P-3-M Colombia 82 68.8 6 4 0CT11891-3-3-3-M Colombia 72 65.4 4 2 0CT13370-2-1-M Colombia 82 72.8 6 2 0CT13366-8-2-M Colombia 74 64.8 4 6 0CT13370-8-M Colombia 87 70.0 4 3 0SMGC89001-6 Thailand 88 89.6 5 5 3TRI 8409178 Thailand 89 95.7 6 4 3IRAT144 Côte d’Ivoire 76 74.2 3 5 0PCT-4\AV \0\0>IR11-1-1 Philippines 74 79.3 3 7 0PCT-4\AV \0\0>IR4-1-1 Philippines 73 74.9 2.5 6 2RR166-645 India 81 79.3 6 3 0RR20-5 India 75 50.8 3.5 2 4RR36-141 India 68 83.2 2.5 3 4Vandana India 73 77.5 3.5 5 6

to this ecosystem. Breeding priorities for subecosystems are therefore different and germplasm exchange is favored only for the identification of specific traits and their introgres-sion into locally adapted cultivars at each site. Tolerance for drought and blast are traits that are needed in every subsystem and evaluation of a common set of genotypes at different consortium sites led to the identification of several donors with the desired characteristics. Table 3 lists the genotypes identified for improved drought and blast resistance. Since the preferred breeding strategy was to introgress these traits into locally adapted traditional or yield-improved cultivars, new crosses were attempted to exploit the drought and blast tolerance traits of these donors. Breeding in the target zones with farmer participatory selection was emphasized for these populations to improve the adoption rate of high-yielding genotypes among farmers in eastern India.

Varieties developed and released

The development of varieties using drought-tolerant aus and high-yielding indica genotypes was simultaneouly un-dertaken at CRURRS, Hazaribag. Vandana, which is highly drought-tolerant and adaptable to upland conditions, was developed from one such cross (C22/Kalakeri). C22 is a

108 Sinha et al

semitall, medium-duration variety from the Philippines and Kalakeri is a tall, drought-tolerant, traditional cultivar from Orissa. Vandana was released by the Bihar State Variety Release Committee in 1992 for the plateau region (Sinha et al 1994) and later for uplands of Orissa in 2002. Vandana is tall, matures in 95 d, and has moderately acceptable grain quality. It has a very good (deep) root system and is highly tolerant of drought. Vandana is moderately resistant to leaf blast and brown spot and has moderate tolerance for major insect pests. The average grain yield of Vandana is 2.5−3.0 t ha–1, but it can yield up to 5.0 t ha–1 when conditions are favorable. The breeding program at Hazaribag resulted in the development of another variety, Anjali, in 2002. Anjali was identified through national coordinated trials (DRR) for the uplands of Jharkhand, Bihar, Orissa, Assam, and Tripura and was released by CVRC in 2002. It is moderately drought-tolerant with thick, broad, erect leaves and high vegetative vigor. Anjali has short bold grains; its eating quality is highly acceptable to farmers (Singh et al 2000). This variety is semi-

tall and matures in 95 d when direct-seeded. Anjali is resist-ant to brown spot and gall midge biotype 5 and moderately resistant to leaf blast and sheath rot. RR345-2 was another highly promising, high-yielding, drought-tolerant breeding line developed at this center, identified for release in the All India Coordinated Varietal Trials. Several other breeding lines developed at Hazaribag—RR18-3, RR165-1160, RR348-1, RR348-6, and RR354-1—were found to have good drought tolerance. A sound breeding program is under way to develop drought-tolerant, productive genotypes at BAU, Ranchi. Under this program, besides drought tolerance, early vigor, plant height, and earliness were emphasized during selection. These efforts have led to the development of several promis-ing varieties. A list of drought-tolerant varieties developed is given in Table 4.

Table 4. Drought-tolerant rice varieties bred for Jharkhand.

Variety Parentage Duration (d)

Salient features

Vandana C22/Kalakeri 90 Tall, long bold grains, deep root system, moderately resistant to leaf blast and brown spot, suit-able for tanr land and don 3

Anjali Sneha/RR149-1149 95 Semitall, short bold grains, broad erect leaves, resistant to brown spot and gall midge biotype 5, moderately resistant to leaf blast and sheath rot, suitable for tanr land and don 3

Birsadhan 101 Fine Gora/IET 2832 85 Semidwarf, long bold grains, resistant to blast, suitable for intercropping with pigeon pea and soybean

Birsa Gora 102 Selection from Brown Gora 100 Tall, long bold grains, red kernel, suitable for tanr land

Birsadhan 105 Fine Gora/IET 2832 90 Semidwarf, short bold grains, resistant to blast, brown spot, bacterial leaf blight, and gall midge

Birsadhan 106 Bala/Black Gora//OS36/CH1039 95 Semidwarf, short bold grains, resistant to blast, brown spot, bacterial leaf blight, sheath blight, and gall midge

Birsadhan 107 Gora mutant/ IAC25 95 Semidwarf, compact and long panicles, short bold grains, resistant to blast and bacterial leaf blight, moderately resistant to brown spot

Sadabahar BRRI SAIL/IR10181-58-3-1 105 Semitall (90–105 cm) variety, stiff straw, very good early and late vigor, long bold grains, suitable for don 3 and 2

Hazaridhan IR42/IR5853-118-5 120 Semidwarf, erect leaves, stiff straw, long slender grains, white kernel, good eating quality, blast resistant, suitable for don 2


Participatory research

Though several new varieties have been released for the rainfed uplands, their adoption has not been to the level ex-pected. The reasons for the low adoption are a lack of access to seeds and, more important, the unsuitability of the new varieties to specific local conditions and farmers’ needs. In addition, a range of socioeconomic constraints may have also influenced farmers’ ability to adopt the new varieties. While the first phase of the consortium focused on strategic research to tackle problems of common impor-tance across regions, the second phase emphasized farmer participatory approaches. These involved selecting and testing new genotypes (participatory variety selection, PVS) and populations (participatory plant breeding) derived from parental lines identified in the first and second phases. Yielding ability of drought-tolerant genotypes under stress in drought-prone environments was also evaluated in the second phase to validate the strength of the tolerance traits hitherto identified. Three villages in Hazaribag District were selected for this project on the basis of the representativeness of the environments targeted in the breeding work, diversity and range of ecological conditions, involvement of women in farm activities, availability of previous survey data, and easy access to the site. Ten farmers from each village were selected for the PVS trials based on their primary occupation (rice farming, especially upland rice), landholding, and some degree of literacy.

Participatory varietal selectionPVS trials were conducted for three consecutive years from 1997 to 1999, with 15−16 upland rice genotypes (including local check Brown Gora). The participating farmers of a given village and the breeders ranked the genotypes in the village and also at the research station at two phenological stages (vegetative and reproductive), from “most liked” to “least liked,” on the basis of their selection criteria. In addi-tion, breeders also recorded duration, height, yield and yield components, and reaction to diseases and pests in all trials. To compare the ranks given by farmers and breeders, the Kendall coefficient of concordance (W) was used (Siegel 1956). Farmers’ and breeders’ rankings were compared fol-lowing the Spearman rank coefficient of correlation (Courtois et al 2001).

Participatory plant breeding (PPB)Participatory plant breeding started 1 year later after farm-ers were given some basic training on the objectives and methodology of single-plant/line selection in the segregating population. One hundred segregating lines from 12 crosses in the F5 generation were grown at one on-farm site (Khorahar) and at the research station. Lines were scored by farmers and breeders at both sites at different growth stages. At maturity, single-plant selections were made separately by farmers and breeders at both sites. The plants selected by farmers and breeders on-station and at Khorahar were grown in separate

blocks in the same field at the respective sites in the 1999 wet season (WS). The farmers and breeders continued their selection of materials and places and, during the 2000 WS, four sets of final (bulk) selections were made—breeders’ selection on-station, farmers’ selection on-station, breeders’ selection on-farm, and farmers’ selection on-farm. These four sets were pooled together along with three checks (Brown Gora, Kalinga III, and Vandana) and evaluated both on-farm and on-station during the 2001 WS using an alpha-lattice de-sign with two replications. The PPB trial data were analyzed using a mixed model, with selector and selection environ-ments taken as fixed variables, and lines within selection environment × selector combinations considered random. The analysis was conducted with the REML algorithm of SAS PROC MIXED. The coefficient of concordance among farmers was highly significant in all the trials conducted in all 3 years. This indicated that farmers’ rankings were not randomly attributed and there was a good agreement among them. The concordance among breeders’ rankings was also high but often not significant because of the small number of breeders. There was a good agreement between farmers and breeders in genotype choice. The farmers’ or breeders’ rankings at maturity were not always highly correlated with yield, especially the farmers’ ranking in low-yielding trials. This may be because farmers considered criteria other than grain yield when selecting varieties for their farms. There was no correlation between farmers’ and breed-ers’ rankings and duration or plant height, with a few excep-tions. The variance component analysis for the on-farm trials showed large and significant differences among cultivars in grain yield under farmer management. The grain yield data averaged over the four on-farm sites and 3 years showed that one elite line (RR348-5) significantly outyielded Vandana and recorded about 100% more yield than Brown Gora (Table 5). No variety × location or variety × year interaction was detected. The three-way interaction, variety × location × year, was large (0.246) but cannot be separated from the within-trial error in this analysis as the trials were unreplicated. This result does not support the hypothesis that varieties exhibit specific adaptation to particular subenvironments within the target region of the research station. The repeatability or broad-sense heritability of different on-farm and on-station trials was estimated to judge the precision of the trial. It was found that the repeatability of the on-station trial was poorer than that of the on-farm trial. This may be because of the low yield of the trials on-station as upland rice is being grown continuously. This resulted in poor soil fertility. The regression of on-farm performance on on-station performance was not significant. This meant that on-station testing is not able to predict on-farm performance of the geno-type and there is a need for integration of on-farm testing at an early stage of the breeding program. At the end of 3 years of PVS trials, farmers selected five genotypes—RR354-1, RR347-166, RR151-3, RR166-645, and RR 51-1—us-

110 Sinha et al

Table 5. Mean grain yield (t ha–1) of upland cultivars and breeding lines evaluated under farmer manage-ment in four villages near Hazaribag, Jharkhand, over three wet seasons.

Cultivar Trials (no.)

Mean over years and locations

Brown Gora 12 1.12Vandana 11 1.74RR347-166 12 1.92RR348-5 12 2.25RR151-4 9 1.11RR203-16 9 1.47RR354-1 12 2.12RR51-1 12 1.78RR151-3 12 1.68RR166-645 12 1.49RR50-5 9 1.63RR139-1 12 1.25LSD0.05 for means

over nine trials0.47

ing their own selection criteria. The highest-yielding line, RR348-5, was not selected by the farmers. Farmers did not consider yield as the only important trait. Long slender grains (RR166-645), high tillering (RR51-1), tall stature, and good cooking quality (RR354-1, RR347-166) appealed to most farmers. These varieties, given to farmers in the following year (2000), performed better than the local check (Brown Gora) during that drought year, except for RR51-1, which failed as it is susceptible to drought. Drought tolerance was not taken into account by farmers as previous seasons were favorable. Mother-baby trials began during the 2001 WS. The analysis of mother-trial data at three on-farm sites showed large and significant differences among cultivars for grain yield under farmer management (Table 5). Averaged over the three sites, RR347-1, RR354-1, and RR 347-166 significantly outyielded Vandana (Table 6). These lines also performed well during 1999-2000. They yielded about 40% more than Brown Gora and about 30% more than Kalinga III in 2001 and have consistently outperformed these lines in previous years. Variance component analysis was conducted to estimate the precision of on-farm trials through estimation of repeatability or broad-sense heritability (H). The H of grain yield measured in on-farm trials was relatively high, as in previous years. The results showed that evaluation of genotypes at three sites, with three replicates per site, would give adequate precision in detecting cultivar differences (Table 7). In baby trials, three varieties, along with a local variety, were given to each farmer to grow. Farmers’ perceptions of the different varieties were recorded. At Chichi, two farmers preferred RR354-1 and one farmer each preferred RR347-1, RR363-737, Vandana, and Kalinga III, respectively. In most of the trials, severe weed infestation was the reason given by

Table 6. Mean grain yield (t ha–1) of upland cultivars and breed-ing lines evaluated under farmer management in three villages and on-station in Hazaribag, Jharkhand, 2001 WS.

Cultivar Chichi Khorahar Peto Mean over villages

CRURRS (on-station)

RR354-1 1.32 1.47 1.70 1.49 2.67RR347-1 1.46 1.53 1.30 1.43 2.92RR347-166 1.61 1.55 1.10 1.42 3.25RR348-5 0.74 1.65 1.50 1.30 2.78RR345-2 1.47 1.38 1.00 1.29 2.30Kalinga III 1.13 1.30 0.80 1.08 1.83RR361-1 1.03 1.28 0.90 1.07 2.35CR876-6 1.29 1.20 0.70 1.06 1.95RR51-1 1.44 0.87 0.80 1.04 2.67Brown Gora 1.26 1.22 0.60 1.03 1.85RR151-3 1.09 1.10 0.50 0.90 2.10Vandana 1.01 0.95 0.70 0.89 2.43RR166-645 1.24 0.87 0.40 0.84 2.25RR363-737 0.90 0.92 0.60 0.81 2.38RR139-1 0.55 1.32 0.35 0.74 1.87RR361-783 0.51 0.90 0.33 0.58 1.15LSD0.05 0.70 0.70 0.70 0.42 0.76

Table 7. Predicted broad-sense heri-tability (H) of grain yield for cultivars evaluated under farmer manage-ment in three villages near Hazari-bag, Jharkhand, 2001 WS.

Sites (no.) Replicates (no.) H

1 1 0.231 3 0.443 1 0.503 3 0.70

farmers in explaining the trial failure. In a few cases, drought damaged the crop at key phenological stages. Farmers concentrated on fewer crosses than breeders. They rejected crosses that did not produce plants that met their criteria. In general, farmers preferred breeding lines with tall stature, high tillering ability, and long panicles. Plants generated from certain crosses (VHC1253/Sathi 34-36, N22/RR20-5, Annada/RR151-3, and RR139-1/IR57893-08) were preferred by both farmers and breeders. The results indicate that, on average, farmer selections significantly outperformed breeder selections, and that selection on-station was superior to selection on-farm (Table 8). Farmer selections yielded, on average, almost twice as much as breeder selections (Table 9). Of the five highest-yielding lines, four were selected by farmers on the research station (Table 10). These were also superior to check varieties. Farmer agreement on ranking varieties was highly significant, although differences in opinion occurred. The relatively high agreement contradicted our initial assumptions that farmers’ preferences will vary because of their diverse


Table 9. Effecta of selection by farm-ers versus breeders and selection on-farm versus on-station on grain yield (g plot–1) of selected lines evaluated under farmer management at Khorahar, Jharkhand, 2001.

Selection environment

Selector

Farmer Breeder Mean

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

aF tests of main effects of selector and selection environment were significant (Pr >F 0.0002 and 0.0001, respectively).

Table 8. F tests for effects of selection by farmers ver-sus breeders and selection on-farm versus on-station on grain yield of selected lines evaluated under farmer man-agement at Khorahar, Jharkhand, 2001.

Source F value Pr >F

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

Table 10. Means and selection history of the checks and five highest-yielding lines.

Line Selection history Grain yield (g plot–1)

RR356-77 Farmer selected on-station 831RR356-72 Farmer selected on-station 825RR356-74 Farmer selected on-station 769RR356-71 Farmer selected on-station 569RR356-51 Breeder selected on-station 519Kalinga III 413Vandana 413Brown Gora 256LSD0.05 412

socioeconomic backgrounds. This may also be due to the limited number of farmers involved in the project or to the low diversity in wealth, caste, and ethnicity of the sample farmers. The high agreement among breeders indicated similarity in selection. The agreement between farmers and breeders was good in most cases. Participation will bring little improvement when there is close agreement between farmers and breeders. The degree of agreement is highly influenced by the materials used for selection. For example, in the uplands of eastern India, the variety has to be tall and has to mature within 100 d. Less variability within the tested genotypes perhaps influenced the farmers’ and breeders’ preferences for variety or trait. The rankings of farmers and breeders were correlated with yield in only a few cases, as farm-ers considered other factors also in their decision. Though duration and stature had little association with yield, these are important characters in varietal choice. Again, because of the low variability in the tested material, these were not identified as important. Therefore, the ranking of varieties must be combined with the survey data on farmers’ selection criteria to get the actual information (Courtois et al 2001). The results of the 3-year on-farm testing indicated that varieties tended to be ranked similarly across farms. There-fore, genotype × environment interaction is not a significant factor in influencing varietal performance. This indicates that an on-station breeding program can serve specific purposes

for the targeted region. But, the precision of on-farm trials is no less than that of on-station trials and 3 years of on-sta-tion trials failed to predict on-farm performance. This clearly indicates that PVS should be integrated at an early stage of testing to better predict varietal performance. The farmers’ intention of voluntarily testing new varie-ties in their own fields indicates that access to new varieties or information is a major constraint as the local seed system is not very effective. Out of five varieties farmers tested during the 2000 WS, two did not perform well and were rejected by farmers. At the early stage, farmers may adopt more varieties, but some will be dropped as testing is done for a number of years. Once a variety is adopted by farmers, it can be disseminated to the neighboring villages (Joshi et al 2001). In PPB, in which both farmers and breeders make selec-tion from segregating populations, farmers selected progenies from fewer crosses. Farmers immediately rejected crosses that did not produce progenies according to their set criteria. Because breeders know the background of these crosses, they expected desirable segregants in later generations. The significant superiority of farmers’ selection over breeders’ selection is quite unexpected. It has established that farmers are also able to identify high-yielding entries (Ceccarelli et al. 2001). The better performance of on-station selection over on-farm selection shows that selections at the research station can serve the target region. This may be due to the consistently low yield at the research station. The PVS program has given breeders a systematic way to approach farmers. The interaction with farmers and social scientists involved in the project helped all stakeholders gain a better understanding of the complexity of the problem. As the local seed system is not functioning, access to new varie-ties or information about the new technology seemed to be major constraints. Proper allocation of resources for varietal testing is needed as on-station evaluation alone cannot pre-dict cultivar performance in the target region. PPB work has produced very interesting results. Though it is still early to make conclusions, we can surmise that farmers can handle advanced-stage segregating populations. Farmers were able

112 Sinha et al

to select superior progenies from segregating generations and it is thus possible to expose breeding materials at the advanced segregating stage to farmers.

The Upland Rice Shuttle Breeding Network

To strengthen the exchange of genetic materials among dif-ferent research institutions, the Upland Rice Shuttle Breeding Network was established under the auspices of the ICAR-IRRI collaborative program. This network greatly facilitated the generation of information on varietal performance in a range of soil and climatic conditions, with better precision. A major goal was to address location specificity in the deployment of genotypes by developing and evaluating breeding lines under naturally occurring stress in the target environment. The breeding lines identified at Hazaribag and the adapted germplasm from individual centers were pooled and evalu-ated under the network as preliminary and advanced multi-environment trials (2002). In subsequent years (2003, 2004), the centers also received a set of fixed lines originating from IRRI, which were selected along with the center’s own breeding lines. Seed increase was done at Hazaribag in the dry season and seeds were redistributed for further selection. Promising entries from the advanced multilocation trials were contributed to the PVS trials and to AICRIP testing.

References

Ceccarelli et al. 2001. Farmer participation in barley breeding in Syria, Morocco and Tunisia. Euphytica 122:521-536.

Champoux MC, Wang G, Sarkarung S, Mackill DJ, O’Toole JC, Huang N, McCouch S. 1995. Locating genes associated with root morphology and drought avoidance. Theor. Appl. Genet. 90:969-981.

Courtois B, Lafitte R. 1999. Improving rice for drought-prone environments. In: Ito O, O’Toole J, Hardy B, editors. Ge-netic improvement of rice for water-limited environments. Los Baños (Philippines): International Rice Research Institute. p 35-56.

Courtois et al. 2001. Comparing farmers’ and breeders’ rankings in varietal selection for low-input environment: a case study of rainfed rice in eastern India. Euphytica 122:537-550.

Government of India. 1984. Area, production and productivity of rice in India. Directorate of economics and statistics. Ministry of Agriculture and Cooperation, New Delhi, India.

Joshi et al. 2001. How narrowly adapted are the products of decentral-ized breeding? The spread of rice varieties from a participatory plant breeding programme in Nepal. Euphytica 122:589-597.

Siegel S. 1956. Non-parametric statistics for the behavirol sciences. New York (USA): MacGraw Hill. 312 p.

Singh BN. 2002. High yielding rice varieties in India, Rice India (March), p 5-6.

Singh RK, Prasad K, Mandal NP, Singh RK, Courtois B, Singh VP. 2001. Sensory evaluation of upland rice varieties with farmers: an experience in eastern India. In: An exchange of experiences from South and Southeast Asia. Proceedings of the International Symposium on Participatory Plant Breeding and Participatory Plant Genetic Resources Enhancement, Pokhara, Nepal, 1-5 May 2000. Cali (Colombia): Centro Internacional de Agricultura Tropical. p 319-327.

Sinha PK, Variar M, Singh CV, Prasad K, Singh RK.1994. A new upland rice variety ‘Vandana’ for Bihar plateau. Indian Farm-ing 44(3):1-3.

Variar M, Sinha PK, Mandal NP, Maiti D, Shukla VD, Sridhar R, Dash AB, Bhat JC, Sengar RBS, Veracruz CM, Atlin G, Courtois B. 2002. Linking QTL detection with varietal development. Paper presented at National Symposium on Upland Rice Production Syetems, 26-28 September 2002, CRURRS, Hazaribag, India.

Notes

Authors’ address: Central Rainfed Upland Rice Research Station, Hazaribag, India.