Rice research for food security and poverty alleviation

Proceedings of the International Rice Research Conference

Book is available

Session 1: 
Increasing Yield Potential in Irrigated Rice: Breaking the Barrier
 

Increasing yield potential in irrigated rice: breaking the yield barrier 

Takeshi Horie

A stagnation in increases in rice yield under intensive crop management has been
recognized in recent decades, especially in tropical dry-season rice. To break the
current barrier to the yield potential of rice, it is necessary to identify yield-limiting
processes and associated traits, and to quantify their genetic variability. For this, a
brief review was made on previous studies on yield formation processes of rice and
their genetic variability. The results of our recent experiment in Kyoto, Japan,
indicated that the most critical growth attribute for rice yield under intensive
management is crop growth rate (CGR) during the latter half of the reproductive
period (15 to 0 d before heading). The CGR at this stage critically affects final
spikelet number by regulating spikelet degeneration, potential single-grain weight by
determining husk size, and grain filling by forming active sinks and determining
endosperm cell number at initial grain filling. A considerably large genotypic
difference existed in CGR during this period and, hence, in yield; however, higher
CGR and yield were not necessarily associated with higher leaf N content. A remote-
sensing technology for evaluating canopy diffusive conductance (gc) was developed
on the basis of simultaneous measurements of the temperature of sunlit and
suddenly shaded canopies. The gc obtained by this method for various genotypes at
midday in the latter half of the reproductive period showed a significant linear
relation to their CGR and yield, which suggests that gc can be an effective criterion
for selecting high-yielding genotypes under field conditions. The existence of a
considerably large genotypic difference in gc, together with its rapid evaluation by
remote sensing, may encourage efforts to break the barrier to increases in yield
potential in rice by breeding inbred and hybrid cultivars having a wide genetic
background.

Closing the rice yield gap for food security 

Dat Van Tran 

The Green Revolution has enabled rice production to meet the demands of the
growing population. Since 1990, however, rice production has increased at a lower
rate than the population. This deceleration in the growth of rice production is a cause
for concern in terms of world food security. It has been the topic of numerous
reviews and several rice scientists have alerted those concerned of the risk of a
pending food crisis. Yield gaps can still be observed in several countries, while
evidence of a productivity decline in intensive rice production has been increasingly
noticed both on research stations and in farmers' fields. An increase in rice yield can
lead to improved farmers' income and food security. Because of the complexity of
this problem, different points of view exist regarding the possibility of narrowing
yield gaps as a tool for increasing rice production. Several experts believe that yield
gaps in favorable rice ecosystems are not significant for improving rice yield
potential. Others believe that large and exploitable rice yield gaps still exist. In 1998,
the average rice yields in 81 countries were less than the world average yield of 3.8 t
ha-1, indicating the existence of yield gaps. Also, progressive farmers usually obtain
higher yields and more profits than ordinary farmers, indicating the presence of
knowledge gaps. The yield gap in rainfed rice, which is usually larger than in
irrigated rice, suggests the potential for increasing rice production. Closing the rice
yield gap aims not only to increase rice yield and production but also to improve the
efficiency of land and labor use, reduce production costs, and increase food security.
Various factors cause exploitable yield gaps in rice, such as physical, biological,
socioeconomic, and institutional constraints, which can be effectively improved
through participatory research and government attention. In practice, yield gaps are
classified into agronomic gaps, socioeconomic gaps, institutional gaps, and mixed
gaps according to constraints in improvement programs. The narrowing of the yield
gap is not static but dynamic with technological development in rice production, as
the gap tends to enlarge with the improvement of the yield potential of rice varieties.
This requires integrated and holistic approaches, including appropriate concepts,
policy interventions, understanding of farmers' actual constraints to high yield,
deploying new proven technologies for production and postproduction in an
integrated approach, and adequate institutional support to farmers.

Development of the new rice plant type and advances in research on

breeding for super high yield 
Chen Wenfu, Xu Zhenjin, Zhang Longbu, and Yang Shouren

The history, present status, and research trends of breeding for super high rice yield
are reviewed and analyzed comprehensively in this chapter. We believe that, even
after two major breakthroughs, one through dwarf breeding and the other through
the use of heterosis, rice yield has remained stagnant for a long time. The third
breakthrough in yield potential is expected to be realized by combining the ideal
plant type with growth vigor. Research findings have shown that increasing the
biological yield is the foundation, whereas optimizing yield components is the
precondition for breaking the yield barrier. Results have also indicated that an
effective way to develop super rice lies first in developing the new plant type and
strong vigor by crossing indica with japonica subspecies, or crossing geographically
distant varieties, and then consolidating the two advantages by optimizing the
combination of desirable traits via multiple crossing or backcrossing. Based on this
theory and method, researchers have developed elite germplasm of the new plant
type such as Shennong 89366 and super high-yielding japonica varieties (Shennong
265 and Shennong 606) with large, erect panicles. In terms of photosynthetic
efficiency and biomass production, an erect panicle type could be another
phenomenal change in adjusting the japonica rice plant type to achieve maximum
yield, following breeding for dwarfness and the ideal plant type.

Increasing yield in irrigated boro rice through indica/japonica improved

lines in West Bengal, India 
S.K. Bardhan Roy 
Boro (dry season) rice has made a significant contribution to the growth in rice
production of West Bengal. Its main contribution has come from growth in area,
whereas the growth rate of productivity has remained only 2.9%. However, the
growing environment of boro rice provides an opportunity to reach a yield of 13 t ha-
1 with a suitable plant type. Several indica/japonica (I/J) lines were screened and
evaluated for high grain yield in the boro season. The highest grain yield of 9.2 t ha-
1 was obtained from selected I/J line IR58565-2B-12-2-2, which was equal to that of
indica hybrid CNHR3 and significantly higher than that of modern variety IR36.
Panicle length and spikelet number were higher in selected I/J lines. On-farm tests
revealed a 21% higher grain yield with superior grain type. Selected I/J lines are
suitable for increasing the productivity of boro rice in West Bengal.

Breeding rice varieties for higher productivity in the Mekong Delta of

Vietnam 
Bui Ba Bong

The Mekong Delta (Cuu Long), with a rice area of 4.0 million ha, produced half of the
total rice production in Vietnam and supplied 4.5 million tons of milled rice for export
in 1999. Sixty percent of the rice area is irrigated with intensive cultivation of two to
three crops per year. The average yields of improved varieties are now 6 t ha-1 in
the dry season and 4 t ha-1 in the wet season. During the past ten years, yield
increased 1 t ha-1 or 2.4% per year. To sustain this increased rate, we have focused
on breeding rice varieties for high yield potential, resistance to brown planthopper
and blast disease, and good grain quality. The target is to increase the average yield
from 6 to 7 t ha-1 in the dry season and from 4 to 5 t ha-1 in the wet season.

What governs ceiling yield and its rate of attainment? 

J.E. Sheehy, P.L. Mitchell, J. Dionora, and A. Ferrer 

The goal in an ideal agricultural system would be to achieve the greatest possible
yield in the shortest possible time using minimum inputs. There are perhaps an
unusually large number of uncertainties in the area of research concerned with
breeding for higher yield and they give rise to a number of important questions.
What is the maximum yield achievable in a given growing period? What are the
uncertainties surrounding estimates of yield limits? Can we construct rice ideotypes
that are useful in the pursuit of ceiling yields? Are yield and quality inversely related?
Does the concept of hybrid vigor become invalid as yield limits are approached?
What is the likely role of genetic engineering in the pursuit of higher yields? The
above questions are addressed in this paper.

Scaling up from quantum yield to grain yield to identify critical features of

rice with C4 photosynthesis
P.L. Mitchell, J.E. Sheehy, and F.I. Woodward 

Given the need to increase rice yields substantially, a rice plant with C4
photosynthesis produced by genetic engineering could be of value. We analyzed the
losses of potential fixed carbon when scaling up from theoretical quantum yield to
grain yield to identify the differences between rice and maize. Maize has a higher
measured quantum yield since photorespiration is suppressed and canopy
photosynthesis is increased by higher maximum rates of leaf photosynthesis arising
from the higher concentration of carbon dioxide around Rubisco. A successful C4 rice
plant must possess both these features, which are associated with the combination
of the C4 biochemical pathway and specialized (Kranz) anatomy. In addition,
improvements in primary partitioning (from initial products of photosynthesis to
sucrose, starch, and amino acids) are probably necessary to fully exploit higher rates
of photosynthesis.

Tillering duration and yielding ability of rice varieties in the winter rice
season of Bangladesh
A.R. Gomosta, H.A. Quayyum, and A.A. Mahbub

In Bangladesh, the winter rice crop (boro) is an irrigated crop. Among the three rice
seasons of Bangladesh, it is the longest rice season, producing the highest grain
yield. The winter season crop begins in November and ends in May. Farmers sow
their seeds, however, between November and February. The tillering duration of the
crop varies because of different sowing dates or transplanting dates in the winter
season. The different duration of tillering has a positive association with the duration
of low temperature (below 20 ºC), at which the vegetative phase of the crop is
exposed. A longer crop duration allowed the tillers to become more mature,
producing a higher number of panicles in the winter season. In the winter-season
crop, the use of different-aged seedlings and time of seeding was more flexible for
long-duration varieties than for short-duration varieties. A short-duration variety
such as BR28 could produce 6 t of grain yield when 30-d-old seedlings from a
November-seeded bed were used. Seedlings from an October-seeded bed produced
only 1-2 t of grain yield. However, 30-90-d-old seedlings of BR29, a long-duration
variety, produced 5-7 t of grain yield when seedlings from an October-seeded bed
were used.

Varietal differences in specific leaf area: a common physiological

determinant of tillering ability and early growth vigor?
M. Dingkuhn, F. Tivet, P. Siband, F. Asch, A. Audebert, and A. Sow

Specific leaf area (SLA), defined as the ratio between leaf blade area and its dry
weight, determines the physiological cost of producing leaf area. Theoretically, at a
given pattern of assimilate partitioning among plant organs, varieties having high
SLA should produce a leaf area more rapidly than those having low SLA, or thick
leaves. This is because the reduction in leaf photosynthesis caused by high SLA is
more than offset by the gain in light interception during exponential growth. The
same reasoning would call for higher relative growth rates (RGR) during exponential
growth for varieties having high SLA. According to studies in 1986-90 at the
International Rice Research Institute, RGR is correlated with tillering. One can
therefore argue that SLA is physiologically linked with tillering ability. Studies at the
West Africa Rice Development Association in 1995-98 indicated a positive correlation
across genotypes between SLA and tillering ability, leaf area index (LAI), and weed
competitiveness. The study used diverse genotypes, including traditional and
improved Oryza sativa indica (lowland-adapted) and japonica (upland) types, as well
as O. glaberrima landraces and interspecific crosses. A subset of these materials
showed that varietal differences in SLA were already expressed before the plant
became autotrophic and were carried through to the flowering stage. Recent studies
at the Centre de coopération internationale en recherche agronomique pour le
développement disaggregated the components of leaf area development, such as
tiller and leaf appearance rate and genealogy, as well as leaf extension rates and
final dimensions for an O. sativa indica (high tillering, small leaves), an O. sativa
japonica (low tillering, large leaves), and an O. glaberrima cultivar (high tillering,
intermediate leaves). For all cultivars, a common linear relationship was observed
between relative tillering rate (RTR) and RGR. This relationship was quite robust
across development stages, except the reproductive stage, and across different
levels of temperature and drought. Differential tillering rates and leaf area
development among varieties and environments were therefore related to dry matter
acquisition, based on a common relationship for morphologically and phenologically
different materials. We conclude that there is a strong case for SLA being responsible
to a large extent for varietal differences in tillering ability and LAI in rice. Varietal
differences in tillering ability may in fact be due to inherent growth ability. High SLA
(thin leaves) is a major factor enabling early ground cover and light interception;
therefore, it determines potential growth in many situations. Once ground cover is
complete, however, high SLA becomes disadvantageous because leaf photosynthetic
rates then limit growth. Consequently, the most efficient plant type in terms of
growth would have high SLA during early growth stages and low SLA for leaves
produced later on. Such a plant type would have high and early tillering. This
reasoning, if accurate, has far-reaching consequences for plant-type concepts and
selection criteria because selection for low-tillering types would most likely favor low-
SLA types, associated with poor vegetative growth vigor and weed competitiveness.
The new plant type is discussed in the light of these results.

Structure of rice caryopsis in relation to strategies for enhancing yield

G.A.I. Ebenezer, S. Krishnan, and P. Dayanandan

Rice varieties with higher yield potential can be developed by increasing the number
of grains per area, by increasing the grain weight, or by a combination of both. Our
studies on the structure, histochemistry, grain filling, and response to plant growth
regulators (PGRs) have clarified certain unique features of the spikelet of rice grains
and have identified strategies required to enhance the weight of rice grains. The
most important barrier to grain weight is the space limitation imposed by the fertile
glumes, the palea and lemma within which the caryopsis should develop.
Experiments with intact plants and excised spikelets reveal that most PGRs, singly or
in combination, do not influence the growth of glumes. Brassinoloide (BR) promotes
growth of the palea and lemma when applied during the panicle initiation stage. BR
at 10-7 M and benzylaminopurine (BAP) at 10-5 M applied together as a soaking
spray increase grain size and dry weight by 39%. This is achieved through promotion
of cell size and cell number, of both the palea and lemma, and the caryopsis within.
The aleurone cells of the treated grains increase to 160,000 from 75,400 in control
grains. Strategies for yield improvement could aim at increasing the dry grain
weight, either by (1) breeding rice varieties with a larger palea and lemma or by (2)
producing varieties that respond to PGR application by rapidly increasing the size of
the palea and lemma.

Raising rice yield potential by manipulating photosynthesis: improvement of

radiation conversion efficiency
Erik H. Murchie, Stella Hubbart, Shaobing Peng, and Peter Horton
Projected increases in rice grain yield are likely to require an increased rate of
biomass production, which will demand a greater canopy photosynthetic assimilation
rate (CAP). There is evidence that the radiation-conversion efficiency (RCF: the
aboveground biomass produced per unit of radiation energy intercepted) for rice is
lower than that of other C3 crops. One of the components of CAP is the leaf-level
photosynthetic rate (P). We have investigated variation in light-saturated
photosynthetic rates (Pmax) of rice plants in the field. We compared IR72 and the
new plant type (NPT) and examined responses to (1) leaf age during grain filling and
(2) position within the rice canopy. Photosynthesis during grain filling accounts for
between 60% and 100% of the carbon in mature rice grain: following flowering,
Pmax in the flag leaf either declined gradually (IR72) or was generally sustained for
the greater part of the grain-filling period (NPT). In both cases, it was difficult to link
changes in photosynthesis to the decline in Rubisco content. In fact, a rise in Rubisco
content was noted after flowering, which did not correspond to a change in Pmax.
Pmax declined according to leaf position within the canopy; this was linked to
available light at each position and also occurred alongside a reduction in Rubisco
content. However, Pmax of the NPT was considerably lower than that of IR72. An
improved RCF could occur by raising leaf-level Pmax and by improving the
photosynthetic responses of leaves lower in the canopy to high light. A coordination
of biochemical and whole-plant studies will be needed.

Agrobacterium-mediated transformation of elite Chinese rice cultivars with

antisense waxy gene to reduce amylose content in the endosperm and
improve rice quality
Qiao-quan Liu, Zong-yang Wang, Xiu-hua Chen, Xiu-lin Cai, Men-ming Hong, and
Ming-hong Gu

Rice (Oryza sativa) grain starch consists of amylose and amylopectin. The amylose
content in the endosperm is a key determinant of rice eating and cooking quality.
High amylose levels are usually associated with dry, fluffy, and separate cooked rice.
The waxy (Wx) gene is known to encode the granule-bound starch synthase
responsible for the synthesis of amylose in the rice endosperm. To improve rice
cooking quality, several elite rice cultivars were selected for transformation with the
antisense Wx gene using the Agrobacterium-mediated method. A chimeric antisense
gene containing the 3.1-kb portion of the 5¢ upstream region from the translational
starting site of the rice Wx gene, a 0.75-kb antisense waxy DNA fragment, the gus
reporter gene, and the nos terminator were constructed and inserted in the T-DNA of
the binary vector pCAMBIA1300. The resultant Ti-plasmid, p13W4, was used to
transform calli derived from immature embryos 12-15 d after pollination via
Agrobacterium tumefaciens. More than 1,000 transgenic plants have been
regenerated from several elite cultivars including japonica and indica. The analyses
of polymerase chain reaction (PCR) and Southern blot hybridization confirmed that
the T-DNA containing the antisense construct has been integrated into the genome
of the transgenic rice plants. Data from GUS activity showed that the GUS fusion
gene could express efficiently in the endosperm, but was less detected in the leaves,
stems, and roots of the transgenic rice plants. In some of the seeds from these
transformants, a range of reduction in the apparent amylose content was found (up
to a maximum of nearly 100% reduction). A reduction in the kernel weight of some
transgenic plants was also detected (up to a maximum of 7-8% reduction in several
transgenic plants, whereas the amylose content dropped below 1%). Genetic
analysis suggested that the foreign gene and amylose content in these transgenic
plants could be transmitted stably. Now, a lot of elite transgenic lines (T5-T7
generation) with lower amylose have been selected for a continuous test.

Session 2: 
Exploitation and Use of Heterosis in Rice

Breeding of super hybrid rice

Yuan Longping

Plant breeders realize that there are only two effective ways to increase the yield
potential of crops through breeding: improving morphological traits and using
heterosis. However, the potential is limited when improving morphological traits
alone and heterosis breeding will produce undesirable results if it is not combined
with improved plant morphology. Any other breeding approaches and methods, such
as molecular breeding, must be incorporated along with good morphological
characters and strong heterosis, otherwise there will be no real contribution to
increasing yield. On the other hand, further developing plant breeding for high
targets must rely on the progress of biotechnology.

Opportunities and challenges of developing and using hybrid rice technology

in the tropics
S.S. Virmani

By 2025, we anticipate a 40% increase in global demand for rice. Therefore, we need
to produce more rice per unit area per unit time. The success of hybrid rice in China
has encouraged IRRI and several tropical countries to explore the prospects of this
technology to increase their rice yields. Although Chinese rice hybrids and parental
lines were not usable as such in the tropics, the presence of significant heterosis in
tropical rice was amply demonstrated. IRRI scientists developed several cytoplasmic
male sterile, maintainer, and restorer lines and heterotic combinations and shared
these with NARS. Some of these have been released for commercial cultivation as
such and others with some local modifications. Farmers have also reported a yield
advantage of at least 1 t ha-1 from some of the released hybrids. Concurrently,
hybrid rice seed production technology was also developed for the tropics, which
allowed seed yields of 0.5-3 t ha-1. Hybrid rice is now being grown commercially in
India, the Philippines, and Vietnam, and several other countries are exploring the
prospects of its commercialization. These developments have encouraged private
seed industries (both national and multinational) to invest in hybrid rice technology
generation and/or seed production and marketing. IRRI and national agricultural
research systems (NARS) have been collaborating to harness human resources to
develop and use the technology. Major challenges to the large-scale adoption of
hybrid rice technology in the tropics are very high expectations of farmers,
inconsistent performance of the first set of released hybrids, inadequate
understanding of agronomic management of hybrids, inadequate availability of pure
seeds of parental lines and hybrids, poor grain quality of hybrids in comparison with
premier-quality rice, inadequate level of disease/insect resistance in the released
hybrids, inconsistent seed yields, high cost of hybrid seeds, and the habit of rice
farmers to use their own seed. Inadequate linkage and understanding between
research and seed production agencies and inadequate coordination among research,
seed production, and technology transfer agencies also constrain the large-scale
adoption of the technology. Future opportunities exist for enhancing the level of
heterosis through indica/tropical japonica hybrids and the selective use of heterotic
groups and gene blocks; developing better agronomic, nutrient, and pest
management practices; improving grain quality and disease/insect resistance of
hybrids; developing hybrids for stress-prone environments; and direct seeding.
Prospects of further increasing seed yields and reducing input costs in hybrid seed
production plots are also real. Work is in progress in all these areas. Besides
developing better rice hybrids, the transfer of this technology can also be facilitated
by establishing effective mechanisms to link research and seed production and
technology transfer systems. To strengthen these systems, IRRI collaborates with
seven NARS, the Food and Agriculture Organization (FAO), and Asia Pacific Seed
Association (APSA) in an ADB-funded project "Development and Use of Hybrid Rice in
Asia." Hybrid rice technology in China helped to increase its global rice production by
about 20 million t per annum. This spares an estimated 6 million ha of land every
year from rice production. Any additional area under this technology would further
contribute to global food security and environmental protection.

Hybrid rice in Bangladesh: opportunities and challenges

A.W. Julfiquar, S.S. Virmani, M. Mahiul Haque, M.A. Mazid, and M. Mostafa Kamal

Hybrid rice is one option for increasing the yield ceiling in rice by 15-20% over the
best modern varieties. The development of hybrid rice technology in Bangladesh
began in 1993. Initial work on hybrid rice involved testing F1 hybrids and evaluating
cytoplasmic male sterile (CMS) and restorer lines received from IRRI. The CMS lines
IR58025A and IR62829A were adaptable and stable with an outcrossing rate as high
as 33.8% and 41.0%, respectively. Some well-adapted varieties/lines were identified
as maintainers or good restorers for the wild abortive cytosterility system. Some elite
maintainer lines identified are BR29, BR5876-6-2-1, BR4839-17-2-2-HR42, BR5690-
62-23, BR5882-12-2-1, and BR5892-32-5-3. Some CMS lines and their maintainers
from China were also evaluated for their adaptability and performance. However,
these lines were not adapted to Bangladesh conditions and were highly susceptible to
disease and insects. Several IRRI-developed hybrids, tested in multilocation yield
trials during the 1995-96 boro season at Gazipur, Comilla, Bhanga, and Habiganj,
outyielded the standard check variety of the same duration by more than 1 t ha-1.
Grain quality characteristics of the tested hybrids were comparable with those of our
recommended check varieties. Two hybrids, IR69690H and IR68877H, are being
tested in the National Hybrid Rice Trial during the boro season 1999-2000 for
possible recommendation by the National Seed Board. IR69690H has growth
duration similar to that of BRRI Dhan 29 and IR68877H has growth duration
comparable with that of BRRI Dhan 28. These two hybrids have shown about a 1 t
ha-1 yield advantage over the check variety of similar duration in multilocation trials
and on-station trials in different regions of Bangladesh. Trials conducted on seed
production during the past two years have shown that six rows of a CMS line and two
rows of a maintainer line appeared to be the optimum row ratio for obtaining the
highest seed yield for CMS multiplication. For hybrid seed production, eight rows of a
CMS line and two rows of a restorer line seem to be the optimum row ratio. Hybrid
seed yield increased significantly over seasons from 24 kg ha-1 to 1.08 t ha-1. F1
seed yield was higher during the boro season than in the T. aman season.

Recent developments in hybrid rice research in India

K. Krishnaiah, B.C. Viraktamath, M. Ilyas Ahmed, C.H.M. Vijayakumar, and M.S.
Ramesha 

Rice occupies a pivotal place in India's food security and livelihood system. The
country has to produce about 135-140 million tons of rice by 2020 to meet its ever-
increasing food requirements. Recognizing the potential of hybrid rice to enhance
production and productivity, a goal-oriented project on hybrid rice was launched by
the Indian Council of Agricultural Research (ICAR) in 1989. Since then, significant
progress has been made in developing hybrid rice. More than 1,000 experimental
hybrids were evaluated and 13 hybrids (12 from the public sector and one from the
private sector) have been released for commercial cultivation. Many cytoplasmic
male sterile (CMS) lines and restorers were evaluated and promising lines have been
identified. These lines are now being used to develop new hybrids. Experimental
hybrids were screened for quality and resistance to major pests and diseases and
promising hybrids were identified. Major efforts have been made for efficient and
economic seed production. A series of trials were conducted to determine optimum
row ratio, planting pattern, GA3 requirement, supplementary pollination techniques,
etc. A seed production package has been developed based on these results. Seed
yields that were <100 kg ha-1 during the initial stages have risen to 1.0-1.5 t ha-1
because of the advent of this package. Specific breeding programs were begun to
improve parental lines. New approaches such as the development of iso-cytoplasmic
restorers, cross breeding of restorers and maintainers, and population improvement
by using male sterility-facilitated recurrent selection have helped to isolate 1,400-
1,500 derivatives that will be useful in hybrid rice breeding. To enhance the levels of
heterosis, two-line breeding and indica/tropical japonica hybridization methods are
being deployed. More than 50 thermosensitive genetic male sterile lines were
screened and five lines have been identified as promising based on their distinct
transformation behavior and other desirable agronomic features. Fifty experimental
two-line hybrids were evaluated and promising combinations were selected for
further testing. To develop indica/tropical japonica hybrids, promising tropical
japonica lines were screened and those possessing wide compatibility (WC) genes
have been identified. Emphasis is being given to improving grain quality and
incorporating resistance to major pests and diseases.

Exploiting rice heterosis in the Philippines

Edilberto D. Redoña, Frisco M. Malabanan, Leo J. Javier, Salvador I. Yabes, Samuel
A. Ordoñez, and Santiago R. Obien

The use of hybrid rice technology as a major approach for further increasing rice
production has been recognized in the Philippines with the inclusion of hybrid rice
technology as an integral component of the Department of Agriculture's rice
production program in January 1998. Using three released hybrids, PSBRc26H,
PSBRc72H, and PSBRc76H, large-scale seed production and technology
demonstration activities began in the 1998 dry season. National capacity for hybrid
rice development and use is continuously being strengthened through training
activities implemented nationwide, with more than 800 agricultural technicians, seed
inspectors, local government officers, and potential seed growers participating in
more than 20 training courses on hybrid rice seed production to date. Hybrid rice
research and development, on the other hand, has been strengthened with the
establishment of a multidisciplinary PhilRice Hybrid Rice Program, the conversion of
the PhilRice branch station in Isabela into a hybrid rice center, and the forging of
research collaboration with IRRI, five Chinese hybrid rice research institutions, and
hybrid rice programs of other countries. The participation of the private sector is
encouraged to expedite the development and farmer adoption of the technology in
the country.

Research and development of hybrid rice technology in Sri Lanka

S.W. Abeysekera, D.S. de Z Abeysiriwardena, and M.P. Dhanapala

Research and development for hybrid rice in Sri Lanka passed through three main
stages: the evaluation of F1 hybrids for heterosis, the transfer of cytoplasmic male
sterility (CMS) from adapted CMS lines to promising Sri Lankan rice varieties, and
development of a locally adapted package of management practices for hybrid rice
seed production. During the past four seasons, 26 IRRI-bred CMS lines were
evaluated and their pollen sterility ranged from 87% to 100% and only the lines with
100% sterility were used. Of these, 25 were uniform and adaptable to local
conditions. Outcrossing was very high in IR62829A, IR69625A, IR68899A, and
IR58025A (26-40%). Out of the 156 heterotic combinations made using adaptable
CMS lines, 11 combinations were promising. Out of 1,271 test crosses evaluated, 56
were highly sterile. These potential CMS lines are now in the BC2 and BC3
generations. Nine F1 hybrids developed at the Rice Research and Development
Institute (RRDI) have been evaluated in yield trials at target locations during 1999.
Some hybrids were promising with a yield advantage of more than 1 t ha-1 over
inbred checks. Ten IRRI-bred rice hybrids were tested in an international coordinated
hybrid rice yield trial during the 1999 major (wet) season at two target locations.
Three F1 hybrids (IR71092H, IR67693H, and IR64616H) outyielded the local and
international checks by 26-33%. Leaf CO2 assimilation of IR71627H was higher than
that of Bg357 (the highest-yielding inbred variety in Sri Lanka). The yield of
IR71627H was always greater than that of Bg357 by 20-30% at all levels of fertilizer
tested. Seed of some of the promising F1 hybrids was produced by the isolation-free
method and strict-isolation method. With these methods, 0.75-1 t ha-1 seed yield
was obtained for each combination.

The commercialization and performance of hybrid rice in the United States

Robin D. Andrews

The development of hybrid rice for the mechanized farming system used in the
United States began in 1980, in the private sector, with a contractual arrangement
between the China National Seed Corporation and a U.S. company. Subsequent
attempts to grow hybrid rice in the U.S. on a farm scale were unsuccessful. In the
early 1990s, the partly developed knowhow and germplasm were acquired by
RiceTec AG (Vaduz, Liechtenstein), which invested heavily in solving the economic
and quality problems involved in meeting U.S. farmer, processor, and consumer
requirements. RiceTec's germplasm development program and technical work were
supported by a collaborative agreement with the Hunan Hybrid Rice Research Center
in Changsha, Hunan Province, China. The development effort has been successful. In
December 1999, RiceTec, Inc. (Alvin, Texas, USA) began selling the first hybrid seed
in the U.S. Hybrid seed for approximately 6,000 ha is available for sale for planting
in 2000 in the states of Arkansas and Missouri, located at approximately 36ºN. The
price of hybrid seed on a per hectare basis is about three times that of varietal seed.
In this chapter, we discuss the performance of this first commercial U.S. hybrid-
denoted as XL6-relative to U.S. varieties in terms of yield, quality, and grower
economics. XL6 is a conventional U.S. long grain in terms of grain dimensions and
starch properties with a grain quality well suited to parboiling. In 1999, XL6 exhibited
more than a 30% yield advantage over local varieties in large-scale field trials in
Arkansas. The production of hybrid grain requires a change in cultivation practices on
the part of the U.S. rice farmer, primarily related to substantially lower seeding rates
and lower fertilization requirements. We describe the manner in which the added
value of the hybrid is shared between RiceTec and farmers.

Improving seed production to speed up the global commercialization of

hybrid rice
Changxiang Mao

Hybrid rice in China and other countries has yielded 20-30% higher than the best
inbred varieties. This yield advantage can contribute considerably to national and
international efforts to achieve food security and alleviate poverty. In recent years,
the hybrid rice-growing area in China has become more than 50% of the total rice
area. Its commercial cultivation in developing countries outside China has expanded
rapidly also, especially in Vietnam and India. Many stable cytoplasmic male sterile
(CMS) lines and restorer lines have been developed by scientists from China, the
International Rice Research Institute, India, Vietnam, and other countries. By using
CMS lines with good general combining ability and specific combining ability
(including environmental genetic male sterile lines) and R lines, more and more rice
hybrids have been released and commercialized not only by breeding institutes but
also by the seed sector. Breeding heterotic rice hybrids has thus become easier. But
the major barrier to the further expansion of hybrid rice in many developing
countries is its high price and poor quality of seeds. Chinese experience showed that,
with efforts to improve seed production techniques, the nationwide average yield of
hybrid seed in China increased from 0.27 t ha-1 in 1976 to 2.72 t ha-1 in 1997, with
a record of 7.39 t ha-1. The cost of producing F1 seed has decreased from US$6 to
$0.80 kg-1 during the same period and seed quality has also improved markedly.
The refinement of seed production techniques, common efforts from the public and
private sector and national and international seed sectors, and policy support from
the government are solutions for further improving hybrid rice seed production,
increasing its yield, and decreasing the cost of F1 hybrid rice seed in the 21st
century.

Adoption of hybrid rice technology in India: an economic assessment of

early farm-level experiences
A. Janaiah and Mahabub Hossain

Policymakers and research managers in India considered hybrid rice, an innovative

technology in the late 1980s, as an option for raising the yield frontier under
irrigated environments. However, concerns are being raised about its acceptability on
the part of hybrid rice farmers in view of its slow adoption in the target
environments. This chapter aims to discuss this issue based on empirical evidence
obtained from sample farmers, seed growers, traders, and millers of hybrid rice in
irrigated rice systems in India during the 1997-98 crop year. On average, the yield
gain of hybrid rice over that of popular inbred varieties was 16%. Under current
pricing, hybrid seed production for rice is highly profitable. But hybrid rice grain
brought an 11% lower price in the market than inbred varieties because of sluggish
consumer demand, largely on account of inferior grain quality and lower head rice
recovery. Therefore, the net return to hybrid rice production was 5% lower than that
of inbred rice. Higher seed cost and lower output price brought about the lower
profitability of hybrid rice production. To further develop hybrid rice, breeding for
value addition through quality improvement, risk reduction, and increased heterosis
was identified as a research priority.

Session 3: 
Breeding for Abiotic Stress Tolerance
 

Plant breeding for stress environments: Are we making progress?

A. Blum

Although conventional breeding has been attaining slow but steady progress in plant
breeding for stress environments, an impressive effort is now being made to apply
molecular biology as a means for accelerating progress. In this respect, molecular
biology is applied via two main channels: (1) marker-assisted selection of
physiological and developmental stress-resistance traits to augment conventional
yield-based selection programs, and (2) genetic engineering consisting mainly of
"mining" and "inserting" stress-responsive/adaptive genes that are expected to
transform plants into being more stress resistant in the field. The first channel
appears to have produced an increasing number of successful cases in recent years.
It is presently limited mainly by the often poor quality of phenotyping of the mapping
populations for stress-resistance quantitative trait loci. It might be limited in the
future by the presently insufficient research in crop stress physiology. The genetic
engineering approach is still wrestling with problems of concept and procedure.
Advanced technologies are turning the identification and cataloguing of stress-
responsive genes into a routine. The dilemma is mainly in identifying and quantifying
the function and value of the candidate gene-in the whole plant and under the
relevant stress scenario in the field. The difficulties stem from known problems of
gene expression in the transgenic plant, which are exacerbated by unnecessary
faults in testing plants for stress resistance. An examination of the main components
of the problem leads to the conclusion that genetic engineering for improved crop
stress resistance carries a proven potential, but the realization of this potential in the
field requires close collaboration between the molecular and physiological/agronomic
disciplines. Progress being made in the application of marker-assisted selection to
plant breeding for stress environments is a successful case of such collaboration.

Development of rice yield components in irrigated upland and lowland

environments
H.R. Lafitte and B. Courtois

As water for agriculture becomes scarcer, some lowland areas will be forced to shift
to upland rice cultivation. Rice generally yields less in aerobic systems. To obtain
adequate yields, we need to understand the causes of reduced productivity in upland
conditions and their genetic control. This chapter combines results from variety
comparisons, a doubled-haploid line (DHL) mapping population, and near-isogenic
lines (NILs) to address these questions. The development of leaf area index was
greater in aerobic conditions for an upland variety, Azucena, and also a lowland
variety, IR72. For both varieties, the percentage of sterile grain was greater and the
1,000-grain weight less in aerobic conditions, leading to a lower harvest index in the
upland field. The reduction in harvest index was more severe in the lowland cultivar.
We also compared yield components measured in sprinkler-irrigated upland, furrow-
irrigated upland, and lowland conditions for 85 DHLs derived from a cross between
another semidwarf indica lowland cultivar, IR64, and Azucena. Yields and most yield
components were poorly correlated across upland and lowland environments.
Common quantitative trait loci (QTLs) were identified across water levels for some
traits, but QTLs for yield and yield components were generally specific to one
environment. Additional studies examined the performance of NILs of IR64 that had
chromosomal segments from Azucena introgressed through marker-aided selection.
The introgressed segments were chosen because they contained QTLs for root traits
(e.g., root thickness, length), but some yield component QTLs had also been mapped
near those segments. When the NILs were cultivated under lowland conditions, they
produced fewer tillers per plant than did IR64, but some produced more tillers m-2 in
the uplands. Preliminary yield data indicate that some NILs can outyield IR64 in
lowland and/or upland conditions. These results provide insights into the genetic
control of the development of rice yield components in aerobic and anaerobic
environments.

Combining ability analysis for physico-biochemical traits influencing

drought tolerance in rice 
H.D. Mohan Kumar, P.M. Salimath, S.A. Patil, M.B. Chetti, V.V. Shenoy, and H.M.
Vamadevaiah

Combining ability analysis was carried out in rice through line × tester analysis for
physico-biochemical traits influencing drought tolerance such as relative water
content (RWC), proline accumulation (PRO), and epicuticular wax (EW) content at
the 95th day (S1 stage) and 105th day (S2 stage) under a moisture stress-free
environment (E1) and moisture stress at the reproductive stage (E2). The
inheritance of these characters in both environments (E1 and E2) and at both stages
(S1 and S2) was primarily under the control of additive gene action. The ratio of
variances due to general combining ability and specific combining ability (s 2 GCA/s
2 SCA) was more than 1, indicating the preponderance of additive genes in
controlling these characters. The genotypes Rasi, MTU-1001, and Annada for RWC,
Rasi, Amruth, and Sona Mahsuri for proline accumulation, and Bilekalavi,
Kempukalavi, and Amruth for epicuticular wax content emerged as the best general
combiners. Under the stress condition (E2), four hybrids each for RWC and PRO and
22 hybrids for EW content expressed significant positive SCA effects. The superior
cross combinations were obtained from high × high and high × low general
combiners. In view of the preponderance of additive genetic components in
controlling these characters, it is expected that their genetic advance through single-
plant selection would be quite rewarding.

Drought risk, farmers' coping mechanisms, and poverty: a study of the

rainfed rice system in eastern India
S. Pandey, D. Behura, R. Villano, and D. Naik
Drought is one of the major production constraints to rice production in eastern
India, which has almost half of the country's total rice area. Severe droughts in this
poverty-stricken region adversely affect the livelihoods of millions of people. In
addition to losses in rice production during the rainy season, food grain production
during the postrainy season can also be severely affected. The second-round effects
of drought are losses in employment in farm and rural nonfarm industries, losses in
livestock production because of poor health and mortality, loss of land, migration,
hunger, and social deprivation. The impact is highest among the most vulnerable
groups of people who, even during normal times, are among the poorest. Even
though drought may not be of intensity severe enough to result in famine, many
poor people suffer economically and socially because of their inadequate access to
institutional mechanisms that help smooth consumption over time. This chapter
analyzes the economic cost of drought in eastern India. The value of rice and nonrice
output lost in drought years was calculated based on an analysis of state-level data
for the period 1970-96. The total value of output lost averaged over the 27-year
period was estimated to be as high as $400 million per year. This is certainly a
substantial cost, accounting for about 8% of the total value of agricultural
production. Using panel data of farm households, we analyzed various strategies that
farmers have adopted to reduce the loss from drought. Changes in cropping
patterns, reduction in rice area, and changes in rice varieties and input usage were
the major strategies employed when drought was of milder intensity and occurred
during the early part of the growing season. During severe droughts when crops
failed completely, a range of consumption-smoothing strategies such as local
nonfarm income, sale of assets, borrowing, and migration were used. A substantial
proportion of households suffered a major reduction in consumption despite these
strategies. Vulnerable groups such as women and children suffered the most. The
effect of drought was not limited to the year when drought occurred but it lingered
on for several years as farmers attempted to rebuild the stock of productive farm
assets such as livestock depleted during drought. These economic effects are
quantified based on farm data from eastern Uttar Pradesh and Orissa. During severe
droughts, informal arrangements based on mutual help were inadequate to prevent a
drop in consumption of the poor. As inability of the poor to gain access to formal
sources of credit and other types of formal mechanisms to smooth consumption is a
major reason for the persistence of poverty in these drought-prone environments, a
range of technological and policy interventions that are needed to improve risk-
coping mechanisms are suggested.

Recent advances and future strategies for breeding salt-tolerant rice

varieties
B. Mishra, R.K. Singh, and D. Senadhira

Major breakthroughs have been achieved in breeding of high-yielding salt-tolerant

rice varieties for various inland saline, coastal saline, and alkaline soils of fragile
ecosystems. Out of 32 salt-tolerant rice varieties developed by the Central Soil
Salinity Research Institute (CSSRI), the first dwarf high-yielding salt-tolerant early
maturing rice varieties, CSR10 and CSR11, are popular as biological amendments for
resource-poor farmers and are being used as low-cost technology for salt-affected
soils. Fine-grain salt-tolerant rice varieties CSR13 and CSR27 were also developed.
CSR13 (CSR1/Bas 370//CSR5) is adapted to alkaline (pH 9.2-10.0) and inland saline
(up to EC 9 dSm-1)-stressed soils and resistant to major pests and diseases. CSR27
(Nona Bokra/IR5657-33-2) possessing dual tolerance for coastal salinity (up to EC
10 dS m-1) and sodicity (pH 9.2-9.9) and resistance to major pests and diseases
was released across the country in 1999. Variety CSR27 possesses high tissue
tolerance and high tolerance for K+ and phosphorus deficiency. We have developed
basmati qualities along with salt tolerance in CSR30. It takes 113-116 d to flower
and is recommended for partially reclaimed sodic soils of Uttar Pradesh and Haryana.
It has yielded more than Taraori Basmati in normal soils with the same quality
parameters and has better resistance to diseases and pests. The F1 anther culture-
derived varieties CSR21 and CSR28, along with rice varieties CSR11, CSR18, CSR23,
CSR24, CSR26, CSR29, and CSR30 having wider adaptability, are in the pipeline for
release. CSSRI-IRRI partnership efforts through shuttle breeding have been
successful and more than a dozen rice varieties have been developed for various
problem soils. A wide spectrum of rice germplasm has been evaluated and
categorized based on tissue tolerance, Na+ exclusion, K+ and P uptake, and
reproductive-stage tolerance. Results showed that no single mechanism was
responsible for absolute salt tolerance. We have been able to combine different
physiological mechanisms into one genetic background. The advanced progenies
have more than two mechanisms of increased tolerance for P, K, and Zn deficiency.
There was no correlation of vegetative-stage salinity score with reproductive-stage
salinity score and grain yield. K+ content exhibited a strong positive correlation with
grain yield and the ratio of Na-K revealed significant negative correlation with grain
yield. Genetic components of salinity tolerance, K+, and Na-K ratio were
investigated. The significant specific combining ability effect for salinity tolerance can
be exploited by hybrid rice breeding.

Identifying rice lines for improved salt tolerance from a mapping population
Raziuddin Ansari, Aisha Shereen, T.J. Flowers, and A.R. Yeo

Rice lines subjected to 50 mM NaCl (corresponding to EC of 6 dS m-1) generally

showed an adverse effect of salts on plant height, total and productive tillers, days to
flowering, leaf area, straw and grain weight, panicle length, and sterility. These
characters exhibited large variations between lines. In the F9 generation, phenotype
stability is evident within individual lines. Calculated as percent reduction in grain
weight compared with the respective controls, there was no adverse effect of salinity
on 20 lines, 30 lines suffered less than 30% reduction, 30-50% reduction was
observed in 50 lines, and the rest were more sensitive. Keeping high-yielding but
salt-sensitive local cultivar Khushboo 95 as a check, ten promising lines have been
selected that show satisfactory yield under the control and suffer less than 50%
reduction at 50 mM NaCl salinity.

Using modified bulk and single panicle selection for improving yield and
submergence tolerance in rice
Sanjay Singh, O.N. Singh, R.K. Singh, and S. Sarkarung

Efficiencies of modified bulk (MB) and single panicle selection (SPS) were compared
following selection in the F3 to F6 generations of five rice crosses. In both cases, 21-
d-old seedlings of the F3 generation were submerged for 7 d and selection was made
from the surviving plants. The later generations were grown in natural flood-prone
conditions. In all, 100 single panicles were selected in F3. For SPS, the seeds of each
of 100 panicles were collected separately and sown using the panicle-progeny-row
method. Again in F4, one panicle each from the best 100 plants was selected and
seeds from each panicle were collected separately for the next cycle of selection. The
process continued until the F6 generation. For MB, five seeds from each of the 100
individual panicles were bulked to be space-planted in F4 and bulks continued until
the F6 generation. From both SPS and MB populations, 20 progenies were selected in
F6 and planted in F7 in a compact family block design in two replications with four
checks, and observations were recorded on yield and other traits. Analysis of
variance showed significant variances due to crosses and selected progenies within
crosses for the SPS population for all the traits. MB progenies, however, showed no
significant differences due to crosses and crosses versus selection. The SPS
progenies had significantly higher yields than the MB progenies. Similarly, the
number of lines superior to the best check was much larger for the SPS progenies
than for the MB group. The SPS progenies also showed higher genotypic variance,
heritability, and expected genetic advance.

Factors influencing rice seedling vigor under different temperature

treatments
R.F. Reinke

Rice cultivars with increased seedling vigor have been shown to accumulate biomass
more rapidly during the vegetative stage under New South Wales (NSW) conditions.
Rapid biomass accumulation is a necessary element in improving the productivity of
short-season cultivars. The rice industry in NSW, Australia, requires cultivars with
reduced growth duration. Such cultivars are essential for seasons in which the total
amount of irrigation water allocated to farmers is reduced or when sowing is delayed.
Given the association between seedling vigor and rapid biomass accumulation, this
study investigated three variables influencing seedling vigor in 17 rice cultivars of
diverse origins under three diurnal temperature regimes (8/22 ºC, 11/25 ºC, and
14/28 ºC), with the aim of assessing their relative importance and identifying traits
that may be used to improve vigor. The variables were seed size, days from sowing
to emergence, and leaf elongation rate (LER) immediately following emergence.
Seedling vigor was measured as seedling biomass when the third leaf had fully
expanded. The correlation coefficients between seedling biomass and each of the
explanatory traits varied across the temperature treatments. Under the lowest
temperature, LER had the highest correlation with biomass; at the intermediate
temperature, the correlations between biomass and LER and between biomass and
seed size were approximately equal; whereas, at the high temperature, emergence
and seed size had the highest correlations with biomass. After accounting for
correlations between explanatory traits, LER was still most closely associated with
seedling biomass under both the low and intermediate temperature treatments,
whereas, under the high-temperature treatment, the correlations of seed size and
LER with seedling biomass were similar. Time to emergence had the lowest
correlation with seedling biomass under all temperature treatments. These results
indicate that seed size and LER both played significant roles in determining seedling
vigor at the three-leaf stage. However, seed size is an important quality
consideration for Australian rice, with increased seed size likely to prove
unacceptable in many export markets. These results indicate that, under low
temperatures, LER was more closely associated with seedling vigor than seed size,
raising the prospect of indirect selection for seedling vigor on the basis of LER. The
success of this approach rests on the genetic correlation of LER and seedling vigor,
the narrow-sense heritability of LER, and the extent to which LER is associated with
other traits such as plant height.
Protection of ribulose-1,5-bisphosphate carboxylase-oxygenase from
oxidative stress-induced degradation in rice
Peng Xinxiang, Liu Yonghai, and Li Mingqi

Oxidative stress during the ripening stage is considered to be a major cause of

Rubisco (ribulose-1,5-bisphosphate carboxylase-oxygenase) breakdown. This stress
condition can be artificially created by treating plants with a well-known active
oxygen inducer, methyl viologen (MV). In this experiment, MV was supplied through
the roots of hydroponically grown rice seedlings (Oryza sativa L. cv. Shanyou 63).
Under controlled conditions (29 ± 1 ° C and 135 m mol m-2 s-1 light intensity),
Rubisco was lost at a rate of 1.2-1.4 mg g-1 fresh weight (FW) h-1 when 0.1 mmol
L-1 of MV was applied. When some active oxygen scavengers were added to the
oxidative treatment, the induced Rubisco breakdown was significantly inhibited. The
degradation rate was decreased from 1.2 to 0.3 mg g-1 FW h-1 by ascorbate, from
1.2 to 0.6 by mannitol, and from 1.2 to 0.9 by benzoate. In addition, oxalate was
effective in inducing resistance to oxidative stress. The degradation rate decreased
from 1.3 to 0.8 mg g-1 FW h-1 when the plants were pretreated with 5 mmol L-1 for
2 d. The result further revealed that oxalate was able to induce the activities of some
antioxidant enzymes. Pretreatment with oxalate doubled the activities of ascorbase
peroxidase and superoxide dismutase, and increased catalase activity 25% compared
with the control. This induction may be the mechanism for oxalate to induce
resistance to oxidative stress. In conclusion, Rubisco breakdown, and photosynthesis
as well, may be improved by exogenously applying some active oxygen scavengers
or defense-response elicitors to the senescing leaves, or by selecting genotypes with
a high level of endogenous antioxidants or antioxidant enzymes during the ripening
stage.

Using molecular markers to increase tolerance for phosphorus deficiency in

rice (Oryza sativa L.)
Matthias Wissuwa and Noriharu Ae

Rice yields are frequently limited by soil deficiency for phosphorus (P). Increasing the
P deficiency tolerance of rice cultivars may represent a more cost-effective solution
to this problem than reliance on fertilizer application. Experimental evidence
suggests that considerable genotypic variation for P uptake exists and that traditional
varieties are likely to carry genes for high P uptake that have not been exploited in
rice improvement. This chapter explores the possibility for transferring genes for
high P uptake from the traditional variety Kasalath to a modern rice cultivar. The
strategy followed here involves identifying molecular markers linked to genes for P
uptake. The advantage of such markers in selecting genotypes with superior P
uptake as well as in providing ideal plant material for physiological studies will be
discussed. A population of 98 backcross inbred lines derived from a cross of the
modern japonica cultivar Nipponbare with the indica landrace Kasalath was grown on
a highly P-deficient volcanic ash soil (Andosol). Lines were genotyped at 245
restriction fragment length polymorphism (RFLP) marker loci and, by relating
genotypic information to P uptake, four quantitative trait loci (QTL) for P uptake
could be identified. One QTL linked to marker C443 on chromosome 12 had a major
effect and lines carrying Kasalath alleles at this QTL had twice the P uptake as lines
with Nipponbare alleles. Based on results of the QTL study, near-isogenic lines (NILs)
for QTL C443 (chromosome 12) and QTL C498 (chromosome 6) were developed by
three additional backcrosses to Nipponbare and subsequent selection for the desired
RFLP markers. NILs genetically resembled Nipponbare to more than 90% but carried
chromosomal segments from Kasalath at the respective QTLs. The effect of both
QTLs could be confirmed using these NILs. P uptake of lines was not different with
optimum P supply, but under P deficiency both NILs had significantly higher P uptake
than Nipponbare. Genes at these QTLs therefore appear to encode for mechanisms
specific to P deficiency. Initial physiological studies imply that Kasalath alleles at
C443 and C498 enable lines to maintain relatively high rates of root growth under P
deficiency, whereas root development is severely inhibited in Nipponbare. Genes at
both QTLs can, however, not be considered general root size genes because the
effect on root size can only be observed under P deficiency. The potential mechanism
responsible for this effect will be discussed. These results show that genotypic
variation for a quantitative trait such as P uptake can be attributed to several loci of
known chromosomal location. Furthermore, it was demonstrated that it is possible to
triple the P uptake of a modern rice cultivar by transferring these loci from a donor of
low agronomic value. After fine mapping both QTLs, efforts will concentrate on
cloning the underlying gene or genes. Once this has been accomplished, the
transformation of modern cultivars promises to be far quicker than with traditional
plant breeding methods.

Session 4: 
Durable Host-Plant Resistance
 

Durable host-plant resistance, a desirable trait for integrated disease

management
Anupam Varma and Neena Mitter

The intensification of agriculture and introduction of new cropping systems have

resulted in a change in disease scenario in various parts of the world. The emergence
of new and damaging diseases has posed a major challenge to scientists for their
eco-friendly management to achieve full potential of crop yield so that increasing
demands for food and feed can be met in the 21st century. In the past, plant
diseases have been managed effectively by host-plant resistance. Some of these
resistances have been very stable, whereas pathogens overcome others in a short
time. It is therefore desirable to have durable host-plant resistance to minimize
losses caused by plant diseases. A resistance gene may have a durable effect for two
reasons. First, a resistant variety may not be exposed to large amounts of inoculum,
affording little opportunity for pathogen adaptation to overcome the resistance.
Second, a genetic factor may prevent pathogen adaptation to a resistant variety
despite a large population size. In different pathosystems, durable resistance could
be controlled by single genes or multiple genes. Monogenic resistance could be based
on a host component interfering with the pathogen or interaction of a host
component with another host component or absence of a host component. Polygenic
durable resistance is based on several host components that interfere with the
pathogen. Genetically durable resistance in most cases is of a quantitative nature as
reported for wheat3/4 P. graminis f. sp. tritici; maize3/4 P. sorghi; barley3/4
Erysiphe graminis f. sp. hordei; potato3/4 Phytophthora infestans; rice3/4
Magnaporthe grisea and Xanthomonas campestris pv. oryzae; and potato3/4 virus X
and virus Y. Durability of resistance is a highly variable phenomenon. Recent
developments in biotechnology have opened up new possibilities of incorporating
durable resistance to pathogens in plants. For example, the introduction of both
resistance and avirulence genes in plants could help in developing effective durable
resistance in host plants. Transgenics for resistance to bacterial diseases are being
produced. In viruses, genetically engineered resistance using coat protein,
movement, and replicase genes is being exploited to produce resistant plants.
Biotechnological approaches are also useful in marker-aided selection and RFLP
mapping, identification of R genes, and gene pyramiding. Durability of the host
against a pathogen cannot be predicted and can be seen only in retrospect. The
strategy should be to manage resistance in such a way that the pathogen cannot
overcome it. Recurrent selection of the host population against susceptibility and
presumed major genes should be carried out. Pathologists and breeders should aim
to achieve durable resistance in which the host plant is resistant in time and space.

Insect-resistant transgenic aromatic rice expresses a barley chitinase gene

and is resistant against sheath blight
B. Ghareyazie, C. Menguito, L.G. Rubia, J. De Palma, A. Ona, T. Mew, S.
Muthukrishnan, R. Velazhahan, G.S. Khush, and J. Bennett

Enhanced levels of sheath blight resistance are reported in the advanced progenies
of transgenic insect-resistant line Tarom Molaii containing a synthetic cryIA(b) and
barley chitinase class II (CHI) gene. Microprojectile bombardment was used for
transformation. Polymerase chain reaction (PCR) and Southern analysis showed the
integration of the chitinase gene into the rice genome. The chitinase gene was
inherited by the progeny of line #827. Immunoblot analyses of protein extracted
from leaves of transgenic (PCR-positive) and control plants (PCR-negative) showed
enhanced levels of CHI protein expression in this transgenic line. The eighth
generation of line #827 was tested for resistance to sheath blight by artificial
infection of plants with the causal fungus (Rhizoctonia solani). Resistance to the
sheath blight pathogen was higher in the progeny of line #827 than in the
nontransgenic control plants (Tarom Molaii).

Transgenic basmati rice carrying genes for stem borer and bacterial leaf
blight resistance
S.S. Gosal, Raman Gill, A.S. Sindhu, Deepinder Kaur, Navraj Kaur, H.S. Dhaliwal,
and Paul Christou

Genetic engineering of basmati rice using synthetic cry1Ac and Xa21 genes has been
achieved. Scutellar tissues of Pusa basmati 1, excised after 6-7 d from mature seeds,
cultured on proliferation medium (MS + 2,4-D [2.5 mg L-1] + proline [560 mg L-1]),
were bombarded using gold particles coated with a mixture of cry1Ac + Xa21 DNA
on osmoticum medium (MS + 2,4-D [2.5 mg L-1] + mannitol [0.4M]). Bombarded
tissues, after 2 d growth on proliferation medium, were subjected to hygromycin
selection (30 mg L-1) for 2 cycles, 2 wk each. Plants regenerated on hygromycin (30
mg L-1) containing medium were screened using the histochemical GUS assay,
polymerase chain reaction (PCR), and reverse transcript-PCR. Southern and western
analysis of T0 and T1 plants confirmed the stable integration and expression of the
cry1Ac. Transgenics included plants containing one or both of the two genes used for
transformation. Segregation analysis during the T1 and T2 generations confirmed the
Mendelian inheritance for marker, reporter, and cry1Ac genes. Insect bioassays
during the T1 and T2 generations have shown enhanced resistance to yellow stem
borer.

Analysis of diversity in pedigrees of blast-resistant rice varieties using PCR-

based resistance-gene analog markers
Shaohong Zhang, Xiaoyuan Zhu, Bin Liu, Qiyun Yang, Lin Luo, Shangzhong Wu, and
H. Leung

Sanhuangzhan 2 (SHZ-2), an indica rice variety, has been widely planted in southern
China for more than 10 years and has showed durable resistance to blast. The
planted area for SHZ-2 and its derived resistant varieties has reached 660,000 ha.
But there is a difference in disease resistance to blast among these derived varieties,
and resistance in some of them has broken down in blast-conducive areas. In this
research, seven pairs of primers designed from the conserved regions of cloned
disease-resistance genes were used to analyze the differences among SHZ-2 and its
derived varieties. These varieties could be clustered into different groups according
to their DNA banding patterns. The clustering matched well with their disease
reactions to blast pathogen isolates. Four varieties (SHZ-2, Waixuan 35, Sanluzhan
7, and Qingluzhan 11) with broad-spectrum resistance to blast were clustered into
one group. The susceptible varieties are far from this group. Results indicated that
resistance-gene analog genome scanning could reflect the difference in disease
resistance to blast among these varieties. The short-lived resistance to blast for the
derived varieties may be attributed to inadequate methods for identifying resistance
genes in breeding. As a second component of this research, 250 recombinant inbred
lines were developed using SHZ-2 (resistant) and Lijiangxintunheigu (susceptible
japonica cultivar) to map the genes conferring durable resistance to blast in SHZ-2
using candidate defense genes. Backcrossing and marker-assisted selection were
used to introgress the durable resistance from SHZ-2 into two popular indica
cultivars in southern China: Texianzhan 13 (high-yielding, with good quality, but
susceptible to blast) and Qisizhan (medium-yielding, with superior quality, but
susceptible to blast). Two backcross populations (BC3F2) each with more than 100
lines have been obtained and molecular analysis and field tests for these lines are
under way.

Biotechnology for gall midge resistance: from molecular tagging to gene

pyramiding
Sanjay K. Katiyar and John Bennett

Asian rice gall midge (Orseolia oryzae) is a major pest across much of South and
Southeast Asia. This pest is genetically diverse and several gall midge biotypes are
known to exist in each country. During the past three decades, host-plant resistance
has proved to be the most effective mechanism for controlling the Asian rice gall
midge. The appearance of different biotypes of this insect at different times and in
different localities prompted the search for new resistance genes for use in breeding
programs. Although several genes have been identified against Indian gall midge
biotype I, there are still some biotypes for which either effective resistance genes
have not been found or the resistance relies mainly on a single major gene. Some of
the genes are rendered ineffective by the evolution of new gall midge biotypes.
There is an urgent need to speed up the process of identifying, mapping, and using
new genes and to maximize the durability of major gene resistance by gene
pyramiding. DNA markers linked to resistance genes are useful for facilitating the
introgression of one or more of these genes in breeding materials. Our efforts in the
area of gene tagging, mapping and pyramiding resulted in the tagging of five gall
midge resistance genes (Gm-2, gm-3, Gm-4, Gm-5, and Gm-6t) with several
randomly amplified polymorphic DNA (RAPD) and sequence-tagged site (STS)
markers and in the development of marker-aided selection kits. These kits are being
used in China to transfer the Gm-6t gene to the parents of hybrid rice. DNA markers
linked to these gall midge resistance genes facilitate the pyramiding of gall midge
resistance genes to achieve durable and broader resistance against this pest.

Induction and use of japonica rice mutant R917 with broad-spectrum

resistance to blast
Mingxian Zhang, Jianlong Xu, Rongting Luo, De Shi, and Zhikang Li

R917, a japonica rice mutant with broad-spectrum resistance to blast, was selected
by treatment of the F1 seeds from the cross between Chengte 232 and Xiushui 37
with 10 krad of 60Co g -ray. R917 was evaluated for blast resistance in multiple
years and at multiple locations. It was resistant to 136 of the 138 isolates collected
from different regions in China, with a resistance frequency of 98%, much higher
than the 86% and 10% of Chengte 232 and Xiushui 37. To analyze its inheritance of
resistance to blast, R917 was crossed with the susceptible cultivar Nonghu 6 and its
two parents. It had a single dominant gene with resistance to races ZB13, ZC15, and
ZE3, nonallelic to those of Chengte 232 and Xiushui 37. After being inoculated with
seven Japanese differential strains, R917 showed the same reaction pattern as
Toride 1 and differed from the other differential varieties, although R917 was
different from Toride 1 in its reaction to seven Chinese races. The allelism test
indicated that the resistance genes between R917 and Toride 1 were nonallelic to
Chinese races ZC15 and ZE3. R917 was a semidwarf with a strong stem, narrow and
erect leaf, tight plant type, and monogenic broad spectrum resistance to blast.
Several lines newly bred using R917 as a donor of blast resistance had the same
broad spectrum of resistance to blast as R917 and desirable agronomic traits,
indicating that there is no linkage drag between the resistance gene and other
important agronomic genes in R917. Recently, R917 has been used as a donor for
many rice breeding programs in China.

Blast infection and yield performance of blast-resistant multilines of rice

Y.G. Kim, H.C. Choi, H.C. Hong, H.G. Hwang, S.N. Ahn, H.P. Moon, S.S. Han, and
J.D. Ryu

To ensure durable resistance to blast disease, we developed multilines of

Chucheongbyeo and Suwon345. Field trials were performed over a 4-year period,
1996-99, at four locations in Korea. Two types of rice blast-suppressing effects of
multilines were observed. Compared with the Chucheongbyeo multilines, the
Suwon345 multilines showed a broader spectrum of race-specific resistance, but
showed more severe neck blast infection and a lower suppression effect in both leaf
and neck blast. These differences in suppressive effect on blast infection between
multilines seemed to be due to the different field resistance between the two
recurrent parents, suggesting the presence of different genetic components related
to neck blast resistance. This study suggests that we should consider the genetic
background of the recurrent parents when the multiline approach is used.
Whitebacked planthopper resistance in Chinese japonica rice
Kazushige Sogawa and Liu Guanjie

Along with the widespread planting of hybrid rice in China in the 1980s, the
whitebacked planthopper (WBPH) Sogatella furcifera has become a major pest of rice
not only in hybrid rice area but also in japonica rice area in China and Japan because
of its massive windborne migrations. We found a Chinese japonica rice, Chunjiang 06
(CJ-06), with a resistance mechanism against WBPH, by which WBPH eggs were
quickly killed in the watery lesions induced at oviposition sites. In addition, CJ-06
expressed strong antixenosis and antibiosis resistance by suppressing WBPH sucking.
These dual mechanisms of varietal resistance on CJ-06 may give it durable field
resistance to WBPH.

Biological and genetic analysis of rice germplasm with resistance to brown

planthopper in the collection of Indira Gandhi Agricultural University
D.J. Pophaly, A.K. Gupta, D.K. Rana, and V.K. Verma

The rice germplasm of the Indira Gandhi Agricultural University, Raipur (Madhya
Pradesh), India, is a unique germplasm collection. This is the world's second largest
collection of rice germplasm, which was collected from 1971 to 1988, but systematic
studies for locating genes with resistance to various insect pests started only
recently. From 1992 to 1998, about 7,500 rice accessions were evaluated against
brown planthopper (BPH, Nilaparvata lugens) in the glasshouse where this insect is
being mass-reared throughout the year on potted TN1 rice plants. So far, 327 rice
accessions were identified with resistance to BPH on a 0-9 internationally accepted
plant damage score scale (Entomology Annual Progress Report 1992 to 1998).
However, 28 cultivars exhibited plant damage scores between 0 and 1. These are
being considered as potential donors for genetic improvement programs. Other
biological parameters, such as a feeding test on resistant cultivars and probing
marks, clearly indicated that BPH behavior on these resistant cultivars differs,
indicating variability in the genetic makeup of these resistant lines. Studies on wild
rice for resistance showed that six lines of Oryza spontanea, O. nivara, and O.
rufipogon were resistant to BPH. Studies on the genetics of BPH resistance indicated
that the resistance is controlled by single dominant genes in cultivars OR1334-8,
CRK7-2-8, TTB148-174-3-2-1, RP1579-1662-72-52, RP1976-18-6-4-2, RP1579-
1627-32-220, and B6932-MR 25-1, whereas a single recessive gene was responsible
in cultivars Basangi (B:2682), Bakada (B:2699), Barhi I (B:1253-I), Barhi II
(B:1253-II), Lal Basant (L:289-I), and Budhiya Banko (B:57-II). In the cross
combinations Lal Basant/Budhiya Banko, Lal Basant/Basangi, Barhi II/Budhiya
Banko, Barhi I/Lal Basant, and Barhi II/Basangi, resistance was governed by two
independent recessive genes. Duplicate dominant genes determine BPH resistance in
the cross OR1334-8/TTB148-174-3-2-1. Identical alleles for BPH resistance were
found in the crosses RP1579-1662-72-52/RP1976-18-6-4-2, B6932-MR25-1/CRK7-2-
8, RP1579-1627-32-220/RP1976-18-6-4-2, and Bakada/Barhi I.

Session 5:
Integrated Nutrient and Pest Management
The quest for connections: developing a research agenda for integrated pest
and nutrient management
Gary C. Jahn, Peter G. Cox, Elsa Rubia-Sanchez, and Michael Cohen

As we manipulate the rice ecosystem to improve production, the main components

that we usually change are water availability, rice cultivars, and fertilizers. Biotic
components of the rice ecosystem (i.e., microbes, flora, and fauna) change in
response to these intentional alterations. Rice intensification, associated with
increased fertilizer use, may increase pest problems. Conversely, some rice pest
management tactics, such as burning rice stubble or adjusting water levels, could
affect soil fertility and reduce the yields of certain cultivars. A deeper understanding
of the interactions of varieties, nutrients, pests, yields, and production costs will
allow us to integrate nutrient and pest management techniques for maximum benefit
to rice producers and consumers. New cultivars (e.g., hybrid rice and low-tillering
rice) and existing cultivars may interact with pests in different ways. If changes in
cultivar and soil nutrient levels create new or more severe pest problems, then the
effects of cultivar and fertilizers on natural enemies (of pests) must be considered as
a possible cause of the changes in pest diversity. Results from greenhouse and field-
plot experiments ultimately must be tested at the field and village levels. Depending
on soil properties, water availability, and climate, it may be possible to put ecological
theories into practice and manage some pest problems by adjusting soil nutrient
levels. The challenge of understanding soil and cultivar interactions with pests and
yield (SCIPY) can be approached from different directions. One research strategy, for
example, would be to overlay maps of soil types, water availability, cultivar
distribution, and pest distribution to characterize the ecosystems. Then, to use
factorial combinations of fertilizer rates and cultivars in the different ecosystems to
assess the effects on pest damage and yield. Another strategy would be to first
identify cases where changes in cultivar and fertilizer interactions have specific
effects on the biotic constraints to crops at a particular place and time. Then, to work
back, to see how widespread the effect is, and to determine the underlying
mechanisms. The advantage of this second approach is that it more rapidly leads to
discoveries that can be applied to actual field problems through integrated nutrient
and pest management. A third approach might not emphasize characterization or
causation, but attempt to solve specific problems on a case-by-case basis through
participatory means, that is, with farmers and scientists designing and conducting
the research together. A fourth approach might combine deductive and inductive
aspects with action research to simultaneously solve problems with farmers, build
models, and investigate causation. This chapter describes the need to investigate
SCIPY, the status of this research, and the advantages and disadvantages of various
research approaches to this issue.

Pest profiles at varying nutrient input levels

P.C. Sta. Cruz, G.C. Simbahan, J.E. Hill, A. Dobermann, R.S. Zeigler, Pham Van Du,
F.A. dela Peña, K. Samiayyan, Suparyono, Nguyen Van Tuat, and Zheng Zhong

Pests were monitored over three seasons at six Irrigated Rice Research Consortium
(IRRC) sites in China (Jinhua), India (Thanjavur), Indonesia (Sukamandi), the
Philippines (Muñoz), and Vietnam (Cuu Long and Hanoi) to characterize pest damage
profiles as affected by nutrient input levels based on the farmers' fertilizer practice
(FFP) and site-specific (improved) nutrient management (SSNM). The research was
superimposed on 137 monitoring farms of the Reversing Trends of Declining
Productivity (RTDP) Project. Biophysical data were taken by the RTDP team and pest
survey information was gathered by the pest impact assessment team at each site.
For the first two monitoring seasons, sheath blight incidence was observed at all
sites. Grain discoloration due to pathogens was recorded at five sites (except Jinhua)
and was dominant at four sites (Cuu Long, Muñoz, Sukamandi, and Thanjavur). Red
stripe was observed in Cuu Long and Sukamandi for both the dry and wet seasons.
Other pests observed at each site were (1) brown spot, narrow brown spot, thrips,
leaffolder, and rice bug in Cuu Long; (2) bacterial blight, brown spot, leaffolder, and
stem borer in Hanoi; (3) stem borer in Jinhua; (4) kernel smut, narrow brown spot,
rice bug, and whorl maggot in Muñoz; (5) narrow brown spot, stem rot, gall midge,
stem borer, and rat in Sukamandi; and (6) bacterial blight, brown spot, gall midge,
brown planthopper, leaffolder, stem borer, and whorl maggot in Thanjavur. Pest
incidence in the FFP and SSNM usually differed. Damage-based assessments of
sheath blight, grain discoloration, stem borer, leaffolder, and rat were higher in FFP
than in SSNM plots. On the other hand, incidence of brown spot, red stripe, and
whorl maggot did not significantly differ by nutrient management practice. Some
pest- and site-specific cases showed the tendency to favor either the FFP or SSNM
treatment; thus, further analysis of the data at specific sites is needed. Initially, the
results showed that SSNM is effective in increasing yields from 5% to 20%, while
pest incidence of five out of eight major pests observed was lower (compared with
the farmers' fertilizer practice) in the improved nutrient management treatments at
high pest incidence. However, this does not negate the possibility of differential
trends at the site level.

Effects of nitrogen, calcium, and silicon nutrients on bacterial blight

resistance in rice
Su-jein Chang, Dean Der-syh Tzeng, and Cheng-chang Li

This study aims to investigate whether nutrient management can reduce epidemics
of bacterial blight disease caused by Xanthomonas oryzae pv. oryzae in rice by
determining the effects of nitrogen, calcium, and silicon on disease prevalence. Rice
varieties with different levels of resistance to the disease were cultured under
nutrient culture conditions. Different concentrations of nitrogen, calcium, and silicon
were added to the nutrient solution to form different treatments. The X. oryzae pv.
oryzae pathogen was artificially inoculated on the leaves of rice plants at the tillering
and booting stages. When the disease appeared, the degree of disease infection was
observed in terms of lesion length. The length of disease lesion on the leaves of
resistant line TSWY7 increased with an increase in nitrogen levels, especially the
plants grown under the high nitrogen level of 120 ppm pot-1. Under conditions of
higher or lower level of calcium application, the disease lesion increased its length
considerably at the tillering stage. The effectiveness of silicon in reducing bacterial
blight disease epidemics was quite obvious in this experiment. This phenomenon was
observed not only in the disease-resistant line TSWY7 but also in the susceptible
variety TN1. A highly negative correlation (r = -0.998) was observed between silicon
content (in leaves) and the length of disease lesions. A negative correlation was also
observed between soluble sugar content in leaves and disease resistance. The results
indicated that epidemics of bacterial blight can be reduced to some extent by
applying silicon as a supplement to fertilizers.

Recent advances in research on rice migrant insect pests in China

Zhang Xiao-xi and Cheng Xia-nian

The migratory characteristics and pathway of northward and southward (backward)

migration of the rice pests brown planthopper (BPH) Nilaparvata lugens, whitebacked
planthopper (WBPH) Sogatella furcifera, and rice leaffolder (RLF) Cnaphalocrocis
medinalis were investigated in China from 1976 to 1982. The methods used were
alpine net catch, aircraft and sea catches, mark-recapture, and dissection of ovaries.
Physiological and ecological mechanisms of flight or migration, including oogenesis-
flight syndrome, flight rhythm and energy resources, and the weather conditions that
affect migration, were also studied during the 1980s. This chapter summarizes four
aspects of recent advances during the 1990s, focusing on radar detection for
migrants at high altitude and forecasting for rice migrant pests. Two types of radar,
the X-band and Q-band, with centimeter and millimeter short wave, respectively,
and an air net suspended from a "kytoon" (balloon) were used to detect migratory
behavior during high-altitude flight of BPH and RLR. The "layering" and "non-
orientation" behaviors and the aerial density and flight speed during autumn
southward migration were described. Based on the parameterization of the migratory
behavior detected by radar, trajectory analysis was used to explore the mid-scale
emigratory sources of these two species, thereby substantially improving the
accuracy of forecasting pest immigration. Various methods for long-, mid-, and
short-term series forecasting for RLR and BPH were introduced. Indices for the long-
term prediction of the relationship between the occurrence of an El Niño southern
oscillation event and the outbreak of BPH in China were explored. Based on a
geographic information systems platform, an attempt was made to establish a
prewarning system for RLR and BPH.

Soil biochemical and microbiological clues to the sustainability of intensive

rice-cropping systems in Southeast Asia
Wolfgang Reichardt, Ruthchelle de Jesus, Luu Hong Man, and Laddawan Kunnoot

 Observations of declining grain yields in intensive tropical rice cropping systems

triggered a search for potential biochemical clues to soil N supply. Derived from a
simplified "N pump model" describing the dynamics of biomass-linked labile soil
organic N pools, the analytical targets comprised total soil protein including microbial
biomass, soluble phenols, respiratory electron transport system (ETS) activity, and
functional profiles of cultivable soil bacteria. Despite the absence of soil moisture
fluctuation in a continuous intensive rice-cropping system, total soil protein
concentrations reflected annual pool size changes of almost an order of magnitude.
N-sink functions of total protein pools prevailed significantly in N-omission plots
during the low-yielding wet season, whereas net immobilization was significant over
all seasons. Intraseasonal variation of soil protein pool sizes was positively correlated
with grain yields. The capacity of the soil to remineralize organically bound nutrients
as indicated by electron transport system (ETS) activities was negatively correlated
with soluble phenol concentrations. ETS activities as well as functional profiles of soil
bacteria were altered due to interactive effects of nutrient and pest management.
Representing biochemically defined soil constituents and processes, the parameters
monitored are amenable to mechanistic models in soil biochemistry and, most
importantly, can also be assessed at national agricultural research and extension
systems using basic analytical laboratory equipment.
Balanced nutrient management and beyond
C. Witt, A. Dobermann, G.C. Simbahan, and H.C. Gines

Highly variable soil nutrient supplies within small localities, low efficiencies of applied
fertilizer nitrogen, and unbalanced nutrition are major constraints to increasing
productivity in irrigated lowland rice systems of South and Southeast Asia. A new
concept for site-specific nutrient management (SSNM) was developed to overcome
these constraints. Using this approach, fertilizer recommendations can be worked out
on a field-specific basis to account for the variability in indigenous nutrient supply.
Data from a case study in the Philippines are presented showing an average increase
in grain yield of about 11% because of SSNM in comparison with the farmer's
fertilizer practice (FFP) under current management practices. SSNM performed better
than FFP regardless of yield constraints related to general crop management or
pests. Results also indicated a substantial scope for further increases in productivity,
if different aspects of crop management were improved simultaneously (such as pest
and disease control and nutrient management). We highlight the importance of
strategic on-farm research to collect primary data for model development at the farm
level, covering a much wider range of environmental and socioeconomic conditions
than possible through on-station research only. A conceptual model is presented for
integrating crop and pest management into a nutrient decision support system for
irrigated rice.

Reducing the gap between attainable and potential yield in double rice-
cropping systems of Zhejiang Province, China
Wang Guanghuo, A. Dobermann, C. Witt, Q. Sun, R. Fu, G. Simbahan, and M.A.A.
Adviento

The climate-adjusted, genetic yield potential of modern rice varieties in Zhejiang

Province of China is currently about 10-12 t ha-1. Attainable yields in farmers' fields,
however, have been maintained at only about 5.5-6.0 t ha-1 in both the early and
late cropping seasons since 1985 despite the large-scale use of hybrid and other
modern rice varieties and increasing fertilizer use. On-farm experiments involving 21
rice-growing farmer families began in 1997 in the central part of Zhejiang Province
to assess the status of soil fertility and productivity under intensive rice-rice
cropping. Low and variable recovery efficiencies of applied inorganic fertilizer
nitrogen were identified as a major constraint to increasing plant nutrient uptake and
grain yield in farmers' fields. A new, site-specific nutrient management strategy was
developed to increase productivity and break the apparent attainable yield barrier of
6 t ha-1. Field- and season-specific fertilizer recommendations were calculated
considering indigenous soil nutrient supply, reasonable grain yield targets, and
corresponding nutrient demand, nutrient balance, and nutrient-use efficiencies. The
agronomic performance of SSNM was tested on the 21 farms against the farmer's
fertilizer practice (FFP) in 1998 and 1999. Across seasons and years, grain yield was
8% greater with SSNM than with FFP (6.35 vs 5.90 t ha-1), whereas plant N, P, and
K uptake increased by 8-14%. The gross return over fertilizer cost was about 10%
greater with SSNM than with FFP (US$998 vs $905 ha-1 crop-1). In general, yields
were about 1 t ha-1 or 19% greater in late rice (July to October, mostly hybrid
cultivars) than in early rice (April to July, mostly inbred cultivars), but SSNM
performed equally better than FFP in both seasons. Since fertilizer N, P, and K rates
were lower in SSNM, especially in 1999, increases in grain yield and nutrient uptake
were most likely due to improved timing and splitting of fertilizer N, which increased
the average recovery efficiency of applied fertilizer N from 0.19 kg kg-1 in FFP to
0.29 kg kg-1 in SSNM. The average agronomic N-use efficiency (grain yield increase
per kg fertilizer applied) was 80% greater with SSNM than with FFP (11.4 vs 6.3 kg
kg-1). However, N-use efficiencies were generally low, thus indicating further scope
for improving N management strategies.

Session 6:
Water and Weed Management in Direct-Seeded Rice

Water and weed management in direct-seeded rice: Are we headed in the

right direction?
J.E. Hill, A.M. Mortimer, O.S. Namuco, and J.D. Janiya

Water management is commonly recognized as the single most important cultural

practice determining the success or failure of weed control in rice. Transplanted rice
evolved in large part to allow the use of early flooding as a means of weed control.
However, as the cost of labor rises, South and Southeast Asian farmers have moved
increasingly to direct seeding; yet, paradoxically, direct-seeded rice in the tropics is
difficult to establish under early or continuous flooding. Early or extended field
drainage to facilitate root growth and anchorage of the rice seedling almost always
results in moderate to dense weed infestations. Furthermore, the time and extent of
drainage greatly influence the weed species composition and abundance. Drainage
and shallow flooding early in the life of the crop generally promote germination and
growth of grass weed species such as Echinochloa and Leptochloa, whereas
continuous flooding encourages the aquatic broadleaf weeds such as Monochoria
vaginalis. Thus, the conversion from transplanted to direct-seeded rice results in a
more competitive weed flora and increases reliance on herbicides for their control.
The problem of delayed flooding or draining to establish direct-seeded rice is further
compounded by poor field leveling and water control systems throughout much of
Asia. Even where fields are flooded, the difference of only a few cm in water depth
can substantially change the weed flora. Poorly leveled fields and the inability to
move water quickly on and off contribute to variability in both rice stand
establishment and in weed species. We review the available evidence and discuss
research directions and priorities. We conclude that research and technology
development at the interface of weed science and water management should focus
on (1) a better understanding of the relationship of water depth and drainage to
weed recruitment, (2) mechanisms to improve rice seedling survival and emergence
under flooding, (3) improvement of infrastructure at the field scale to allow for
pinpoint flooding and uniform water depth, and (4) knowledge-based decision
support systems for farmers to fully integrate weed management practices.

Economical use of water for direct-seeded rice

R. Balasubramanian, J. Krishnarajan, and S. Ramesh

The need for greater food production at prices affordable by consumers and
profitable to farmers with emphasis on natural resources and the environment has
been of prime concern to all people. Rice is the staple food of half the world. Rice
production must increase from 500 to 800 million t in the next 25 years to meet
projected world rice demand. It is estimated that per capita availability of water,
about 85% of which is used in agriculture, has declined by 60% from 1950 to 1990.
A water crisis for rice is fast approaching. India, Pakistan, and the Philippines are
particularly expected to suffer a sharp decline in per capita water availability over the
next two decades. Rice production systems are rapidly changing due to the declining
availability of water resources for agriculture. Direct seeding is being practiced
increasingly to cope with the reduced availability of irrigation water. The water-use
efficiency of rice is much lower than that of other crops. On average, more than
5,000 liters of water are used to produce 1 kg of rice. In irrigated wet-seeded rice
culture, water-use efficiency on the farm can be increased by applying only the
amount of water needed. This needs more research to economize on water use. It is
in this context that studies were begun on the irrigation requirement of direct-seeded
rice versus transplanted rice. This study focused on reducing the water requirement
and improving its use efficiency in wet-seeded puddled lowland rice. Field
experiments were conducted at Tamil Nadu Agricultural University, Coimbatore,
India, during the kharif and rabi seasons of 1997 in a randomized block design with
nine levels of irrigation replicated three times. Irrigation levels significantly
influenced the weed population and biomass in both seasons. Grain yield was the
highest with irrigation at 5-cm depth 1 d after the disappearance of ponded water in
direct-seeded rice, transplanted rice, and continuous submergence of 2.5 cm. Water
use was the maximum with transplanted rice due to extended land preparation and
nursery raising. Continuous submergence of 2.5 cm on wet-seeded rice recorded the
highest water productivity and saved 25% and 24% versus transplanted rice (TPR) in
the kharif and rabi seasons, respectively, without impairing productivity and net
returns. Higher water productivity and considerable net returns and benefit:cost ratio
clearly shows the scope for economizing on irrigation water by continuous
submergence of 2.5 cm in wet-seeded rice. Hence, continuous submergence of 2.5
cm may be recommended for wet-seeded rice in lieu of 5 cm, which resulted in a
25% savings in irrigation water versus transplanted rice.

Effect of tillage intensity, water control, and planting method on seedling

establishment and growth of direct-seeded rice
M. Kondo, D.V. Aragones, P.P. Pablico, M. Hagiwara, T.P. Tuong, and M. Yamauchi

Anaerobic seeding, wherein seeds are sown in puddled soil, has been developed to
stabilize seedling establishment of direct-seeded rice. We carried out field and
laboratory experiments at IRRI to investigate the factors affecting seedling
establishment, agronomic performance, and yield in rice in broadcast and row-drilled
anaerobic seeding and compared these methods with surface-broadcast direct
seeding and transplanting. Early flooding severely lowered seedling establishment,
which was further decreased by intensive tillage. Intensive tillage, early flooding,
increased flooding depths, and incorporation of straw lowered seedling
establishment. The reduced seedling establishment was related to the increasingly
reduced soil redox potentials and probably to the increased concomitant Fe++
formation. High plant density in all direct-seeding methods led to vigorous vegetative
growth in terms of dry matter production, tiller number, and N uptake. However, dry
matter production and N uptake in direct seeding became inferior to those of
transplanting after the reproductive stage, which resulted in comparable or lower
total spikelet number and yield in direct seeding. Minimizing the decrease in spikelet
number per panicle associated with increased panicle number and low N uptake after
the reproductive stage is a critical factor in improving yield. Anaerobic broadcast
seeding showed growth patterns similar to those of surface broadcast seeding.
Anaerobic drill seeding showed comparable yield, higher N uptake, and slower leaf
senescence after the ripening stage compared with anaerobic broadcast seeding.

Establishment of irrigated rice under zero and conventional tillage systems

in the Philippines
C.M. Piggin, C.O. Garcia, J.D. Janiya, M.A. Bell, E.C. Castro, Jr., E.B. Razote, and J.
Hill

Zero-till establishment is used widely for many crops around the world. There has
been some work on rice, but this was mostly before the development of broad-
spectrum, nonresidual herbicides and direct-drill seeders. The technology has
potential to allow savings in time, energy, water, and labor during crop
establishment. This is important in Asia as labor and water shortages are becoming
more important constraints in rice cropping. A long-term field trial to develop and
evaluate a methodology for zero-till establishment of irrigated lowland rice was
established in 1998 at the International Rice Research Institute (IRRI). Zero-till crops
were established with small hand-tractor-mounted seeders, which placed the seed at
3-5 cm in either a narrow slit opened by a tyne or a narrow furrow opened by a
rotovator. Zero-till treatments were compared with conventional treatments of direct
dry seeding on cultivated soil or direct wet seeding with pregerminated seed on
puddled soil. Various weed control strategies using preemergence (butachlor,
oxadiazon, pretilachlor) and postemergence (bentazon, cyhalofop, 2,4-D) herbicides
were evaluated under the different crop establishment treatments and compared
with hand weeding and no weed control. Regular observations are being made on
crop and weed growth, soil parameters, and water use. Some results on crop and
weed growth for the first three seasons are discussed. Crop establishment and grain
yield were comparable from zero-till and cultivated treatments, especially in the third
season, when crop establishment and management technologies were better
developed and yields of 5 t ha-1 were achieved. Hand weeding and various herbicide
combinations were effective in controlling weeds, but weeds were abundant and
yields dropped markedly to 0.1-1 t ha-1 where there was no weed control. There
were 3-4 dominant weeds in a suite of 10-12 species in most treatments. Generally,
Fimbristylis miliacea and Leptochloa chinensis were major weeds under all
establishment treatments, Cyperus difformis and Scirpus maritimus were common
under wet seeding, and Echinochloa colona and Ludwigia octovalvis were common on
zero-till and dry-seeded plots. Results suggest that zero-tillage technology has great
promise for the establishment of lowland irrigated rice. Further research is required
to refine the machinery and technology for weed control, crop establishment, and
crop management.

Wet-seeded rice cultivation technology in Korea

Je-Kyu Kim, Sang-Soo Kim, Sung-Tae Park, Yang-Soon Kang, Soon-Chul Kim, Moon-
Hee Lee, and Shaobing Peng

Since direct-seeded rice was first introduced in 1991, the area for direct seeding in
1995 increased to 117,500 ha, which is 11.1% of the total paddy field in Korea.
There are two methods of direct seeding in rice: dry and wet seeding based on the
physical condition of the field and seeds. In Korea, wet seeding is further divided into
two types: wet drill seeding and water seeding. Water seeding is a technology in
which pregerminated seeds are sown in standing water by a motorized seed sprayer.
Wet drill-seeding technology was developed to solve the problems of water seeding
such as unstable seedling stand and lodging. A wet drill seeder having six furrow
openers with the operator riding on it was designed to make furrows by pressing the
hardening paddy. The optimal seeding time is 1 May to 5 June, depending on the
temperature at seeding time. For both drill and water seeding, the optimum seeding
rate is 30-50 kg ha-1 and the optimum seedling stand is 80-120 seedlings m-2. The
optimal nitrogen (N) application is 110-130 kg m-2. Recently, a furrow drill-seeding
method was developed that is a combination of the dry and wet direct-seeding
methods. To control weeds in wet-seeded rice, a one-shot application of a
sulfonylurea mixture is recommended after good root anchor. Additional application
of herbicide depends on the succeeding growth of weeds. Direct seeding removes the
need for the cultural practices of raising seedlings and transplanting; thus, it is
expected to decrease labor compared with transplanting. The labor hour of direct
seeding is about 30% lower than that of machine transplanting.

Seedling growth and control of propanil-resistant barnyardgrass in lowland

direct-seeded rice fields of Sri Lanka
B. Marambe and L. Amarasinghe

Development of herbicide resistance to propanil in Echinochloa crus-galli (L.) Beauv.,

a major grass weed in lowland rice cultivation in Sri Lanka that was introduced in the
early 1960s, has been reported recently. Experiments were conducted to
characterize the seedling growth of resistant (R) and sensitive (S) biotypes and to
suggest alternative control measures for propanil-resistant populations of E. crus-
galli. Chemical analysis revealed that there was no adulteration of propanil
formulations at the retailer level. Field studies showed that retail propanil
formulations (36% ai) gave less than 30% control of E. crus-galli found in several
locations of the low-country dry zone of Sri Lanka. However, in pot experiments,
68% mortality in R populations occurred under spraying. Dose-response experiments
showed that these plants were three times more resistant to propanil than the
sensitive biotypes. Growth studies carried out in controlled environmental conditions
indicated that the germination % and seedling growth of both R and S populations
were similar at respective day/night temperature regimes. However, plants grown
under a 34/31 °C (day/night) regime had a 27-29% higher germination percentage
than those grown in a 28/24 °C regime. Both R and S biotypes reached the 2-3-leaf
stage 5 days earlier and 4-5-leaf stage 7 days earlier at the higher temperature.
Quinclorac (25% ai), when applied at the recommended rate (200 g ai ha-1),
successfully controlled propanil-resistant populations of E. crus-galli. Contrastingly,
thiobencarb + propanil (40% + 20% ai) and oxadiazone + propanil (8% + 23% ai)
did not result in satisfactory control.

Splash application with pyriftalid/cinosulfuron improves efficiency and user

safety in wet-sown rice in Southeast Asia
H. Lojo, N. Saavedra, J. Sribuddhachart, P. Wongyala, F.W. Lim, Mohd. Hezri Sairi,
Iskandar Zulkarnain, Harris Burhan, H. Ohtomo, K.F. Kon, and J.L. Allard

Weeds are one of the major constraints in wet-sown rice in Southeast Asia. The
aging of the farming population and shortage of farm labor are causes of poor weed
control in rice. This study investigated the splash application technique, using
pyriftalid/cinosulfuron WG 40.8, as a simple and easy application technique for
herbicides. Splash application of pyriftalid/cinosulfuron at 113-150 g ai ha-1 at 6-8
days after sowing controlled more than 90% of grasses including Echinochloa crus-
galli, Leptochloa chinensis, and Ischaemum rugosum. Cinosulfuron in this ready-mix
formulation broadened the weed control to cover many key sedges and broadleaf
weeds such as Cyperus difformis, Cyperus iria, Fimbristylis miliacea, Monochoria
vaginalis, Sphenoclea zeylanica, Ludwigia spp., Limnocharis flava, and Marsilea
crenata. The product, applied by "splash," showed excellent safety for rice. The crop
stand was high and exceeded that of the check plots treated with herbicide in the
spray form. The splash technique required 56% less time to apply the herbicide than
the current standard method of using mist-blowers. With the splash method, farmers
carry only 0.5 L of water compared with 15-20 L of water in addition to the 10-kg
mist-blower on their back. The splash technique is simple, easy to use, and fast. It
requires no spray equipment and improves farmers' efficiency and applicator safety.
The significant productivity improvement indicates that pyriftalid/cinosulfuron fits
well in wet-sown rice in Southeast Asia.

Session 7:
Impact of Technologies on Food Security and Poverty Alleviatio

Development of modern rice varieties: impacts on food security and poverty

Howarth E. Bouis

It is well documented that the development of modern varieties of rice has

contributed importantly to higher growth rates of agricultural output, lower cereal
prices, and increased demand for farm labor, all of which have stimulated faster
growth in the nonfarm economy. These developments have probably reduced
poverty significantly, although rigorous studies of the counterfactual situation had
rice yields remained low are a complex and difficult undertaking. Some earlier
investigations did find that higher-income farmers with relatively large landholdings
were the first to adopt modern varieties. However, later studies concluded that
poorer farmers were equally likely to adopt this scale-neutral technology, although
with some lag compared with larger farmers, when policies did not distort input
markets (e.g., subsidization of credit and mechanization). For food security, there is
unambiguous evidence that lower cereal prices have increased the energy intakes of
the poor, which are typically equated with adequate food security in the economics
literature. In the past decade, however, new research findings by the human
nutrition community have shifted attention away from protein-energy malnutrition to
the importance of vitamins and minerals (micronutrients) in maintaining good health,
adequate growth, and even acceptable levels of cognitive ability. Food staples such
as rice are not rich sources of bioavailable micronutrients, which typically must be
provided by vegetables, fruits, pulses, and fish and animal products if recommended
daily allowances for a range of vitamins and minerals are to be met from dietary
sources. The diets of the poor are, of course, dominated by staple food consumption,
such as rice in Asia. Micronutrient malnutrition occurs precisely because the poor
must give first priority to meeting energy needs by consuming relatively inexpensive
food staples to keep from going hungry. They are too poor to buy enough of the
nonstaple foods that they require and desire, but cannot afford. Some have argued
that the Green Revolution has worsened dietary quality (defined here as
consumption of a range of micronutrients) by raising the prices of nonstaple foods,
as more and more land has been converted into rice production and out of
production of nonstaple crops. However, expenditures for rice are by far the largest
component of the food budgets of the poor in Asia, who spend a high proportion of
total income on food. Lower rice prices have the potential for freeing up significant
resources for consumption of more nonstaple foods by the poor. An analysis of food
consumption data from Bangladesh suggests that lower rice prices have contributed
importantly to maintaining dietary quality in Bangladesh, although rising nonstaple
food prices (brought on in large part by high population growth and the land
constraint) severely constrain improvements to dietary quality for the poor. Because
nonstaple food consumption is so low among the poor, and even though rice is a
poor (nondense) source of trace minerals, rice may contribute 50% or more of trace
minerals consumed. For the most part, past breeding efforts have ignored the
nutritional qualities of new releases, although this attitude is changing. A key
unanswered question is whether modern varieties are richer, poorer, or about equal
sources of trace minerals compared with traditional varieties. Because rice is eaten in
such large amounts every day by the poor, it is an excellent vehicle for fortification
through plant breeding. Plant breeding is more cost-effective than conventional
interventions to reduce micronutrient malnutrition. Loading more trace minerals into
seeds improves plant nutrition and may well result in improved agricultural
productivity.

Technology impact on distribution of rural household income and poverty:

insights from a repeat village study representing different rice ecosystems
in the Philippines
Esther B. Marciano, Fe B. Gascon, Ellanie Cabrera, and Mahabub Hossain

The adoption of modern rice varieties and crop management practices has
substantially increased rice yield in the Philippines. But it is reported in various
circles that little improvement has been made in the poverty situation of the Filipino
rice farmer. The point of controversy is whether the diffusion of modern rice varieties
has caused growing inequality of rural incomes and accentuation of poverty. As a
case study, this chapter will explore the issue of changes in the structure of rural
household income and its distribution over the past decade with primary data
generated through in-depth household surveys in four villages representing the
irrigated, favorable rainfed, and upland rice ecosystems in the Philippines for 1985
and 1997. The study uses a Gini decomposition analysis to assess changes in the
magnitude of income inequality and the contribution of different sources of income,
including rice, to such changes. The study also measures the incidence, intensity,
and severity of poverty using the FGT index and assesses changes in poverty
situations for different rice ecosystems. Factors contributing to these changes are
discussed. An important finding is that rice accounts for a small and declining
fraction of the total household income, and that rice income is relatively less
unequally distributed than income from nonfarm sources. The inequality in the
distribution of education and landholdings is a more important factor behind growing
income inequality than the diffusion of modern rice technologies.

Impact of modern rice varieties on food security and cultivar diversity: the
Bangladesh case
Manik Lal Bose, Mohammad Abu Isa, Abdul Bayes, Binayak Sen, and Mahabub
Hossain
Bangladesh, an extremely land-scarce country in Asia, has done fairly well in
meeting the food needs of its fast-growing population. Modern rice varieties now
cover about 60% of the land. Technological progress has contributed to a reduction
in the unit cost of production, a factor that helped to keep food prices within
affordable limits of the urban and rural poor. But has the diffusion of modern rice
varieties led to an erosion of cultivar diversity? This chapter examines this question
with data generated on the adoption of specific varieties at the district, village, and
household level by measuring the varietal diversity index. The data show low cultivar
diversity at the household level but substantial diversity at the village and district
level. A multivariate regression analysis does not support the hypothesis that the
diffusion of modern varieties has contributed to an erosion of cultivar diversity.
Farmers use a diverse set of varieties to suit different land elevations and soil types
and to preserve other traits such as good grain quality, appropriate growth duration,
and pest resistance. Cultivar diversity is higher at the district and village levels
because of the higher complexity of agroecological conditions at larger geographical
scales.

Indigenous technologies in the rice-based cropping system among tribals of

eastern Madhya Pradesh, India
M.L. Sharma, P.N. Sharma, M.A. Khan, R.K. Tiwari, and M. Hossain

The indigenous techniques of a farming population are derived from their past
farming experiences. Farmers worldwide have adopted their own systems of farming
and this is within the framework of local possibilities and limitations of available
resources. These ancient agro-technologies are still carried on, sometimes with slight
modifications in many parts of the country. The eastern part of the largest province
in India, Madhya Pradesh, is obviously known for its ethnic values and varying
culture. Among various domiciles of this region, scheduled tribes contribute more
than 40%. This population is still far away from the modern world; even today, these
people depend on agriculture. Traditional agro-practices are still followed and
preserved in their natural form. In view of these facts, our study was designed to
study the knowledge and adoption pattern of indigenous techniques by tribal farmers
under the rainfed rice-based cropping system. The findings mainly relied on primary
data collected from 150 randomly selected tribal farm families of 10 villages from the
two most tribal-dominated districts, Bastar and Surguja. Semistructured interview
schedules in the local dialect were used for data collection. The findings revealed that
about 40% of the respondents were illiterate; they belonged to a family with 5-10
members. Agriculture was the main occupation, but others engaged in forest produce
collection and animal husbandry. Agriculture was the major source of income,
followed by labor work. The study also revealed that traditional varieties occupied
more than 60% of cultivated land. Similarly, a majority of the farmers also used
several indigenous technologies from seed selection to storage of produce. Some of
these technologies such as the use of locally made implements; sun drying; use of
salt solution for selection of healthy seeds; bushening operation for weed control;
seed germination test; the use of mahua cake as biofertilizer, spray of kerosene or
soil solution, and use of herbs/plants to control rice caseworm; the use of mud, cow
dung, and rice husks to make storage bins; and the use of neem leaves as insect
repellant were many popular among the tribals. Therefore, looking at the many
indigenous technologies in the area, we suggest that efforts be made to refine and
consider these indigenous technologies when planning, developing, and
recommending location-specific modern technologies for sustainable development in
the tribal areas.

Impact of improved technologies for rice on poverty alleviation: microlevel

evidence from Orissa, India
Parshuram Samal and Surendra Rawlo

Rice is the main source of income and employment for 23 million people in Orissa,
where 91% of the farmers are small and marginal and 58% of the rural people are
below the poverty line. Rice accounts for 47% of the total cropped area and 71% of
the total food-grain production. It is mainly grown during the wet season, which
accounts for 95% of the total rice area and 91% of the total rice production in the
state. The rice crop in the wet season is grown under a variety of situations. Data
from farmers' fields for the period 1990-91 to 1992-93 comprising 702 observations
were analyzed by season and by rainfed/irrigated plot for modern and local varieties
to find out the increase in yield and income of small and large farmers because of
the adoption of improved technologies in rice cultivation. The analysis revealed that
yield increased by 46% and income by 34% because of the adoption of modern
technologies. Small farmers increased their yield and income per hectare by 49%
and 37%, respectively. Season-based analysis of data revealed that small farmers
increased income per hectare by 36% while large farmers gained by 24% during
winter, the main rice-growing season, because of the adoption of improved
technologies. The average yield advantage of modern over local varieties was 8.9
quintals per hectare during winter and 8.8 quintals during autumn (40% and 59%,
respectively). The maximum yield and income advantage per hectare was enjoyed by
large farmers during autumn, the short rice season. The average gain in yield and
income was 38% and 28% from irrigated rice and 36% and 26% from rainfed rice,
respectively. The per hectare gain in yield and income by small farmers was more
than for large farmers under both irrigated and rainfed conditions, thus improving
their purchasing power and food security. Two policy suggestions emerged from the
study: (1) measures should be taken to develop and spread improved technologies
to new areas/ecosystems to increase the income of farmers, which will lead to food
security of the poor in rural areas; (2) there is a need to invest more in irrigation
facilities in the state, which will not only increase yield but also create a favorable
environment for intensification of rice cultivation to increase employment and income
of the rural masses.

Effects of groundwater markets on rice productivity and poverty alleviation:

empirical evidence from the Tarai region of Nepal
Hum Nath Bhandari and Sushil Pandey

Lack of irrigation and heavy reliance on monsoon are the major constraints to
agricultural production and productivity growth in Nepal. The increasing costs, long
gestation period, and potential negative environmental effects of large-scale surface
irrigation schemes have limited the opportunity for further expansion of conventional
surface irrigation. Groundwater (GW) irrigation has been considered as the best
alternative source of irrigation that can harness the potential benefits of available
resources at a minimum possible cost and within a short period of time. The
increased use of GW has been a major source of agricultural growth in many parts of
eastern India with a plentiful reserve of GW. GW markets provide the institutional
mechanisms for increasing the access of resource-poor farmers to GW irrigation. This
chapter presents the findings of a study on the adoption of shallow tubewells,
demand and supply of GW, and effects of GW markets on rice productivity and rural
poverty alleviation in the Tarai region of Nepal. This study is based on farm-level
data collected from 162 owners of STW and 162 nonowners sampled from two Tarai
districts (Sarlahi and Banke). The result of the study shows that access to GW
irrigation has significant positive effects on rice productivity, resource-use efficiency,
cropping intensity, farm income, and employment. While medium and large farmers
were able to obtain groundwater through direct investments in tubewells, small and
marginal farmers relied mostly on purchased water. Farmers having access to
groundwater irrigation through water markets had a significantly higher rice yield
and farm income than those having no access to GW irrigation. In addition to
improving farmer income, groundwater markets also reduced the inefficiency
associated with a high degree of land fragmentation. Groundwater markets,
however, were highly imperfect with the selling price being substantially higher than
the opportunity cost of water. A relatively price-elastic demand for GW indicates that
institutional and policy reforms that help reduce the GW price will be critical to
improve farm productivity and reduce poverty in rainfed areas where the social cost
of groundwater extraction is low.

Land-use dynamics and changes in rice production in the Mekong River

Delta in the 1990s: an econometric analysis using longitudinal household
survey data
Christopher M. Edmonds, Suan Pheng Kam, Ho Cao Viet, and Lorie Villano

The increase in rice production in Vietnam during the 1990s represents one of the
great success stories of Asian agricultural development in that decade. The increase
in national production has taken the country from having a large deficit between rice
demand and supply to becoming the third-largest rice exporter worldwide. This
expansion has contributed to Vietnam's high rate of GNP growth by providing urban
areas with cheap food and generating foreign exchange earnings. Increases in rice
production in the Mekong River Delta, which supplies about half of Vietnam's total
rice production, have averaged about 6.3% y-1 during the 1990s according to
government statistics. Although national statistics on rice production in Vietnam are
widely known, there have been few studies of farm-level changes in rice production
techniques and land-use changes that have led to production increases. This study
helps provide insights into the household-level changes in agricultural production,
which, when aggregated, caused the production increases. Farm-level changes in
rice output with production, land-use, and supply estimates are examined by using
data from a longitudinal household survey that solicited farm production information
for each year between 1994 and 1998. The survey covered about 150 farms from
eight villages in the Mekong River Delta and two villages from the river basin areas
in Dong Nai Province. The chapter begins with an overview of provincial-level
statistics and describes the major changes in policies, institutions, and infrastructure
relevant to rice agriculture during the 1990s. Then, the neoclassical production
framework, which drives estimates, is outlined. The model assumes that households
seek to maximize their revenue from agricultural production, given the resources
available within the household and the price of inputs that can be purchased from
the market. We develop a more complex model of land use and agricultural
production that incorporates the effects of the spatial location of the household and
the transaction costs it faces in determining the farming activities to pursue.
Indicators of the accessibility of farms to input and output markets make use of
recent methodologies developed using geographic information systems software.
Estimable forms of the production, land-use, and revenue functions implied by the
analytical model are derived. Econometric models make use of robust panel data
estimation procedures that control for the effect of variables that cannot be observed
but are known to have a strong effect on estimated outcomes (e.g., household
preferences and motivation). Instrumental variable and system of equation
estimators are also used to eliminate endogeneity bias in estimates of the effect of
variables that are simultaneously determined with the outcomes of interest (e.g.,
cropping intensity and rice production). Finally, results of the estimates are discussed
and a simulation model is developed to highlight the implications of our findings for
policy aiming to increase rice production or change land use in the Mekong Delta in
the future.

Impact of technologies on food security and poverty alleviation in

Cambodia: designing research processes
Peter Cox, Mak Solieng, Gary Jahn, and Mot Sana

Food security is a national priorit