No. 13

Improving productivity and livelihood for fragile environments 

The International Rice Research Institute (IRRI) was established in 1960 by the Ford and Rockefeller Foundations with the help and approval of the Government of the Philippines. Today, IRRI is one of the 15 nonprofit international research centers supported by the Consultative Group on International Agricultural Research (CGIAR – www.cgiar.org). IRRI receives support from several CGIAR members, including the World Bank, European Union, Asian Development Bank, International Fund for Agricultural Development, Rockefeller Foundation, Food and Agriculture Organization of the United Nations, and agencies of the following countries: Australia, Brazil, Canada, Denmark, France, Germany, India, Iran, Japan, Malaysia, Norway, People’s Republic of China, Republic of Korea, Republic of the Philippines, Sweden, Switzerland, Thailand, United Kingdom, United States, and Vietnam. The responsibility for this publication rests with the International Rice Research Institute.

This publication is copyrighted by the International Rice Research Institute (2008) and is licensed for use under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 License (Unported). Unless otherwise noted, users are free to copy, duplicate, or reproduce, and distribute, display, or transmit any of the articles or portions of the articles, and to make translations, adaptations, or other derivative works under the following conditions: Attribution: The work must be attributed, but not in any way that suggests endorsement by IRRI or the author(s). NonCommercial: This work may not be used for commercial purposes. ShareAlike: If this work is altered, transformed, or built upon, the resulting work must be distributed only under the same or similar license to this one. To view the full text of this license, visit http://creativecommons.org/licenses/by-nc-sa/3.0/.

Mailing address: DAPO Box 7777, Metro Manila, Philippines Phone: +63 (2) 580-5600 Fax: +63 (2) 580-5699 Email: irri@cgiar.org Web: www.irri.org. Rice Knowledge Bank: www.knowledgebank.irri.org Courier address: Suite 1009, Security Bank Center 6776 Ayala Avenue, Makati City, Philippines Tel. +63 (2) 891-1236, 891-1174, 891-1258, 891-1303 Suggested Citation: IRRI (International Rice Research Institute). 2008. Improving productivity and livelihood for fragile environments. IRRI Technical Bulletin No. 13. Los Baños (Philippines): IRRI. 54 p. Editing: Bill Hardy Layout and Design: Emmanuel Panisales

ISSN 0074-7807 


Contents....................................................................................................................III Preface......................................................................................................................IV Physiological basis of tolerance of flash flooding during germination and early seedling establishment in rice..........................................................................1
Abdelbagi M. smail, Evangelina S. Ella, Gina Vergara, Donna F. Holt-Stevens, Alvaro Pamplona, and Dave Mackill

Salinity tolerance in rice: physiological bases and implications for management strategies for better crop establishment.................................................8
Abdelbagi M. smail, Babita Thapa, and James Egdane

Opportunities for direct seeding and improved weed control in the Barind of Bangladesh...................................................................................................15
M.A. Mazid, C.R. Riches, A.M. Mortimer, and D.E. Johnson

Breeding for submergence tolerance.............................................................................20
D.J. Mackill, A.M. smail, S. Heuer, E. Septiningsih, A.M. Pamplona, R.M. Rodriguez, C.N. Neeraja, D. Sanchez, K. ftekhar, and G. Vergara

Participatory varietal selection of salinity-tolerant rice for the coastal wetlands of Bangladesh....................................................................................27
M.A. Salam, G.B. Gregorio, D.L. Adorada, and R.D. Mendoza

Increasing profits from rice production in Bangladesh: direct wet seeding of rice using a plastic drum seeder.....................................................32
M. Zainul Abedin

Characterizing and understanding the socioeconomic conditions of farming households in rainfed rice environments: a case in eastern Uttar Pradesh..............................................................................................36
T. Paris, A.D. Cueno, and A. Singh 


Improving Productivity and Livelihood for Fragile Environments is one of IRRI’s four programs and 11 projects stipulated in its Medium-Term Plan for 2006-2008. The Medium-Term Plan (MTP) for 2006-2008 reflects IRRI’s core agenda in addressing current and emerging problems in rice. It is guided by the broad framework of the strategic plan outlined in the document IRRI Toward 2020 published in 1996, and updated in November 2003. It also takes advantage of the scientific opportunities that assist the Institute in reaching its goals. Improving Productivity and Livelihood for Fragile Environments is Program 3 of the MTP for 2006-2008. It examines risk reduction in rice cultivation. The program also focuses on helping increase yield and farm income via the development of stress-tolerant better-yielding varieties using efficient crop management practices. In the past, the probability of success in research for building tolerance for abiotic stresses into better-yielding varieties was low, leading to inadequate allocation of research resources to solve these problems. But with the recent advances in molecular biology for tagging and characterizing genes and their transfer to other species, the probability of success in this area brightened. Since the environments are diverse and their domains vary across countries, the research is being done in partnership with the national agricultural research and extension systems, drawing on local scientific expertise and farmers’ indigenous knowledge. Program 3: Improving Productivity and Livelihood for Fragile Environments attempts to solve the abovementioned problems in three projects―Project 7: Genetic enhancement for improving productivity and human health in fragile environments; Project 8: Natural resource management for rainfed lowland and upland rice ecosystems; and Project 9: Consortium for Unfavorable Rice Environments (CURE). This Technical Bulletin showcases the studies and research results from each of the projects under Program 3. 


Physiological basis of tolerance of flash flooding during germination and early seedling establishment in rice
Abdelbagi M. smail, Evangelina S. Ella, Gina Vergara, Donna F. Holt-Stevens, Alvaro Pamplona, and Dave Mackill

More than 16 million ha of lowland and deepwater rice areas are unfavorably affected by flooding because of complete submergence. Rice is the only crop plant adapted to aquatic environments because of its well-developed aerenchyma tissue that facilitates oxygen diffusion through continuous air spaces from shoot to root and avoids anoxia development in roots. However, complete submergence due to frequent flooding can adversely affect plant growth and yield. Two types of flooding cause damage to rice: flash flooding, which results in rapid ascending of water levels with complete submergence for about 1–2 weeks, and deep water, in which water depth exceeds 100 cm and persists for longer periods of up to several months. Plants may become completely submerged for short periods if flooding is severe. Elongation ability of leaves and internodes is essential in deepwater conditions to keep pace with the rising water levels and to escape complete submergence. Traditional varieties adapted to these environments are low yielding because of their low-tillering ability, susceptibility to lodging, and poor grain quality. Flash flooding can occur any time during the growing season and usually occurs more than once. It is usually more damaging if it occurs early in the season, particularly during early crop establishment. This is because younger seedlings are more sensitive to flooding. In direct-seeded areas, even waterlogging can be devastating because of the high sensitivity of all crop plants to low oxygen during germination.

Tolerance of flooding during germination and early seedling establishment
The likelihood of occurrence of flash flooding at different stages of growth requires different tolerance strategies. In rainfed lowlands, direct seeding is becoming more popular because of the escalating expense and scarcity of labor, and farmers are also enthusiastic to adopt it because of its additional benefits such as shortening crop duration and the enhanced tolerance of some stresses, particularly drought. Breeding cultivars with tolerance of flooding during germination and early seedling establishment will help avoid crop failures commonly encountered when early flooding occurs in both rainfed and irrigated ecosystems. Moreover, maintaining a shallow water head after seeding under irrigated conditions can help suppress weed growth and this could provide an effective, cheap, and sustainable method for weed management. However, this practice is stalled by the unavailability of suitable germplasm. Previous studies showed that rice is capable of germination underwater, but this capability is limited to coleoptile elongation, with failure to develop further. In an attempt to discern rice germplasm with higher tolerance of submergence during germination, we screened more than 8,000 genebank accessions and breeding lines and a few tolerant lines were identified. Screening was conducted by direct dry seeding in plastic trays containing a shallow layer of soil. Seeds of individual lines were sown in rows and immediately submerged by adding water to a depth of 10 cm. This depth is maintained for 2–3 weeks, when the percentage of seeds that germinate and successfully emerge from water 

Table 1. Rice accessions most tolerant of flooding during germination. Cultivar Khaiyan Khao Hlan On Cody Dholamon 64-3 Liu-Tiao-Nuo Ma-Zhan (Red) Sossoka Kaolack Kalonchi Nanhi R6855-00---- R6855-00--3- Origin – Myanmar United States Bangladesh China China Guinea Guinea – ndia NPT-RR NPT-RR Percent survival (first screening) 90 75 70 80 70 90 85 85 90 80 75 55

is determined. Through this method, about 12 lines were identified as being tolerant (Table 1) and the tolerance of some of them was further confirmed in replicated trials (Fig. 1A,B).

Traits associated with tolerance of flooding during germination and early growth
Tolerant lines and a few intolerant checks were evaluated for some agronomic traits to test their association with tolerance. A subset of these is shown in Table 2. Tolerant cultivars vary substantially with respect to mature plant height, days to flowering, and average grain weight. This variation provides opportunities for breeders to select appropriate parental lines for crossing to incorporate this trait. Tolerant lines seem to emerge faster from the soil and water (by about 2–3 d and 2–4 d, respectively) when compared with intolerant lines. The strong negative correlation of survival with time of emergence from soil (R = –0.97) and water (R = –0.95) indicated that fast germination and growth under hypoxic conditions is crucial for survival. However, survival did not correlate with plant height at maturity, days to flowering, or average grain weight. Tolerant cultivars tend to grow faster underwater, produce taller seedlings with more leaves, and attain greater leaf area than intolerant lines (Fig. 2), traits that might also be useful for weed competitiveness early in the season.

Seedling survival (%) 00 80 60 40 0 0

A Not submerged Submerged

Khaiyan Khao Hlan On Tolerant

FR3A R4 ntolerant

Seedling survival (%) after dry seeding and submergence for 21 d 00 B 80


LSD 0.05

Germination under anoxia (absence of O2)
Six tolerant and two sensitive cultivars were evaluated for their ability to germinate under anoxia. Seeds were sterilized using sodium hypochloride and tween 20, and then placed in a 1% stagnant agar solution deoxygenated by continuous bubbling of nitrogen gas.




Cody Khaiyan Kalonch



Khao Hlan On







Fig. 1. Average survival of selected lines over two separate replicated experiments. Vertical bars indicate (A) S.E. and (B) LSD0.05. 

Table 2. Comparison between selected tolerant and sensitive lines for selected agronomic traits and their association with survival. Cultivar Tolerant lines Dholamon Liu-Tiao-Nuo Khaiyan Khao Hlan On Intolerant lines R64 FR3A R4 Correlation with survival Philippines ndia Philippines 65 70 73 –0.47 30 67 5 0.50 0.0 9.7 8.4 –0.00 7 7 7 –0.97 3 3 3 –0.95 7 0 9 Bangladesh China ? Myanmar 7 6 66 85 03 43 60 45 .5 7.0 8. 8.3 4 4 5 4 9 9  9 8 7 74 73 Origin Days to flowering Mature height (cm) 1,000-seed wt. Days to emergence Soil Water Survival (%)

Coleoptile length (cm) 8 6

Coleoptile length (cm) 3 



Seedling height (cm)

Root length (cm)

Leaf area (cm)


Aerated tap water

Anoxic agar

Fig. 2. Average shoot height, root length, and leaf area of five tolerant (shaded columns) and three intolerant (black columns) cultivars germinated underwater. Results are means of 3 replicates taken 21 d after seeding.

Fig. 3. Average coleoptile length of six tolerant (open columns) and two intolerant (black columns) cultivars under anoxia and aerated conditions. Data are from 7-d-old seedlings and vertical bars indicate ± S.E.

Evaluation of percent germination, length of coleoptiles, as well as appearance of roots was carried out after 7 d of continuous anoxia. Tolerant lines had greater coleoptile length under both anoxic and normal conditions (Fig. 3).

Physiological mechanisms associated with tolerance of flooding during germination
Total amylase activity The ability to break down starches into simple sugars is a major factor limiting germination under flooded conditions. This is because of the high sensitivity of key enzymes involved

in this process to low oxygen. Total amylase activity was measured using two tolerant and two sensitive lines germinated in flooded soils for 3 days. The activity of these enzymes increased substantially in germinating seeds of tolerant lines under flooding but did not change significantly in sensitive lines (Fig. 4). Higher activity of these enzymes in germinating seeds of tolerant lines under flooding indicated that these lines are able to break down stored starch into simple sugars better than sensitive lines, and this step is essential for germination and growth. Metabolic breakdown of simple sugars is much less sensitive to low oxygen than the


Amylase activity (units mg– protein) 50 A 40 Not submerged Submerged

Plant survival (%) 6 B



0  0

R = 0.9



Khao Hlan On Tolerant

FR3A ntolerant 




0 30 Amylase activity



Fig. 4. (A) Total amylase activity in germinating seeds of two tolerant and two intolerant rice cultivars sown in submerged soils for 3 days. Vertical bars are S.E. (B) Correlation between amylase activity and seedling survival.

processes involved in starch degradation. Total activities of these enzymes correlated positively with shoot and root length (R = 0.85, 0.83, respectively), and with plant survival (Fig. 4B), suggesting a strong relationship between the activity of these enzymes and the ability of seeds to germinate and seedlings to grow under submerged conditions. Plant hormones Germinating seeds of tolerant lines produced more ethylene than susceptible ones under submergence (Fig. 5). Ethylene is known to counteract ABA synthesis and increase synthesis of and sensitivity to GA. High levels of ABA in seeds are associated with the inhibition of seed germination and dormancy; however, high GA is associated with activation of amylase enzymes and breakdown of stored starches to be used for further growth of germinating seeds. Peroxidase activity Peroxidases are responsible for the assembly of lignins and proteins in the cell wall and for the binding of ferulic acid to cell walls by the formation of diferuloyl cross-links to

Ethylene content (mg seedling –) 30 5 0 5 0 5 0 Khao Hlan On (N) Khao Hlan On (S) IR42 (N) IR42 (S)


5 7 8 Days of incubation in soil 


Fig. 5. Ethylene content of germinating seeds of one tolerant and one sensitive line under submerged conditions. N = normal, S = stress.

matrix polysaccharides, both of which are associated with reduced cell wall extensibility. Higher peroxidase activity is closely associated with reduced growth in other plants such as in mung bean and peanut. We measured peroxidase activity in germinating seeds of two tolerant (Khaiyan and Khao Hlan On) and two sensitive (R64 and IR22) cultivars to evaluate its involvement in growth under low-oxygen conditions. Tolerant genotypes showed much


less peroxidase activity than intolerant ones (Fig. 6A). A strong negative correlation was observed between peroxidase activity and shoot growth (R = –0.69) and seedling survival (Fig. 6B). The ability of seedlings to elongate underwater and emerge quickly is critical for survival. This may also suggest that lower peroxidase activity in seedlings germinating under anoxia could be used as an indicator for ability to germinate and grow under flooded conditions. Nonstructural carbohydrate content and expression of key enzymes involved in carbohydrate catabolism Nonstructural carbohydrate content. Seed starch and soluble sugars are the main sources of energy for embryo growth in germinating seeds and, under anaerobic conditions, breakdown of starch into sugars decreases greatly because of reduced activity or expression of the enzymes involved in starch catabolism under low-oxygen conditions. Comparisons of total starch levels were made between tolerant Khaiyan and intolerant IR42 from 0 to 5 d of hypoxia (0.03 mol O2 m–3). We found that total starch was always higher for the tolerant cultivar for up to 5 d of hypoxia and did not change significantly between time points. The availability of total soluble sugars (TSS) was compared between tolerant Khaiyan and intolerant IR42 at 0 d, 1 d, 3 d, and 5 d of hypoxia. TSS levels at 0 d (25 to 32 µg per seed dry wt.) were not significantly different between tolerant and intolerant rice cultivars. However, after 1 d of hypoxic treatment, availability of TSS declined rapidly, by 60%, and then increased at 3 d in germinating embryos. We found significant differences in TSS between tolerant Khaiyan and intolerant IR42 starting at 3 d and up to 5 d of hypoxia (Fig. 7). The increased levels of available TSS may stem from the higher carbohydrate breakdown brought about by the higher amylase activity found in Khaiyan, which in turn contributes to the greater ability of Khaiyan to survive flooding and low-oxygen stress during germination.

Peroxidase activity (µg α-naphthylamine consumed g – DW h –) 5 A  LSD 0.05

9 6

3 0

Khaiyan Khao Hlan On Tolerant 


R ntolerant

% survival 60 B

40 R = –0.7 





6 Peroxidase activity


Fig. 6. (A) Peroxidase activity in seedlings incubated in anoxic agar solution for 5 d and (B) correlation between peroxidase activity and seedling survival. Vertical bar in A indicates LSD at P = 0.05.

Alpha-amylases. Starches are a major energy source for developing rice embryos and the total amylases needed to break them down were generally found to be higher for tolerant cultivars under low-oxygen or submerged conditions than for intolerant lines. Analyses of transcriptional activity of rice α-amylases (RAmy1A, RAmy2A, RAmy3C, RAmy3D, and RAmy3E) were performed using RT-PCR of RNA extracted from germinating embryos of tolerant (Khaiyan) and intolerant (IR42) genotypes subjected to hypoxia: air (0.25 mol O2 m–3) or anoxia (<0.003 mol O2 m–3) treatments for

Total soluble sugars (µg per seed dry wt.) 35 30 5 0 5 0 5 0 0  3 Treatment days 5 Khaiyan R4

Fig. 7. Total soluble sugars of Khaiyan and IR42 at different time points after hypoxia treatment.

0 h, 12 h, 24 h, 48 h, and 72 h. Rice α-amylases (RAmy1A, RAmy3C, RAmy3D, and RAmy3E) generally showed up-regulation starting at 12 h following treatment (Fig. 8). RAmy2A was not expressed. During hypoxia, RAmy3D, which is important for oligosacharride degradation, showed relatively higher expression in the tolerant Khaiyan than in the intolerant IR42. RAmy1A, which breaks down soluble starch, was less induced during anoxia but not during hypoxia in tolerant cultivars. No differences in expression levels were found for RAmy3C and RAmy3E in the three O2 regimes. Using RT-PCR analyses, we found that rice αamylases were rapidly induced in germinating embryos. RAmy3D, a key and major enzyme for carbohydrate use for germinating embryos, was expressed at higher levels in tolerant cultivars during hypoxia and this probably helps provide and sustain energy during germination. Sucrose synthases. Sucrose is another major energy source for germinating embryos. It can be hydrolyzed via the sucrose synthase (Sus) pathway or the invertase pathway. Under low-O2 stress, Sus transcript levels increase while those of invertase decrease, suggesting that sucrose synthase is the principal enzyme that converts sucrose to phosphorylated hexose under low oxygen. Comparison of

transcriptional activity of sucrose synthase patterns in tolerant and intolerant rice embryos showed no differences between treatments under aerated conditions or during hypoxia or anoxia. We showed the induction of Sus1 during the first 72 h of treatment, whereas Sus3 activity was completely shut down after 24 h of treatment (Fig. 8). Overall, the result suggests that Sus1, but not Sus3, appeared to contribute to sucrose degradation during germination. Pyruvate decarboxylases and alcohol dehydrogenases. Rice seedlings under low-oxygen stress are capable of survival for a limited time by shifting to the fermentative pathway to generate ATP for cellular metabolism. Using RT-PCR analyses, we showed that Pdc1, Adh1, and Adh2 transcripts were induced during the first 72 h of rice seedling growth whether in air or under hypoxic or anoxic conditions (Fig. 8). Although oxidative phosphorylation provides major ATP needs for rapid growth under aerobic conditions, our results showed that germinating rice embryos do experience anaerobic conditions and partly use the fermentative pathway to generate energy. PDC and ADH activity. For total PDC and ADH enzyme activity levels, marked differences were observed between tolerant and intolerant cultivars under hypoxic compared with aerobic conditions (Fig. 9). PDC activity was 1.6x higher immediately after 12 h of
Khaiyan H o u rs RAmy1A RAmy3C RAmy3D RAmy3E Sus1 Sus3 Pdc1 Adh1 Adh2 Gapdh
0 12 24 48 72

IR42 Khaiyan
24 48 72

IR42 Khaiyan
24 48 72

2 4 48 7 2

0 12 24 48 72

24 48 72

Fig 8. RT-PCR of key enzymes involved in carbohydrate metabolism.


PDC levels (µ min – mg – protein) . A .0 0.8 0.6 Air Khaiyan Air R4 Hypoxia Khai Hypoxia R4

ADH levels (µ min – mg – protein) .4 . .0 0.8 0.6 B

0.4 0. 0.0

0.4 0. 0.0 0h  h 4 h 48 h 7 h 0h  h 4 h 48 h 7 h

Fig 9. (A) PDC and (B) ADH activities. Means with the same letter are statistically not different at each time point. The asterisks indicate significant differences between genotypes and treatment at particular time point (P<0.01). Vertical bars indicate ± S.E.

hypoxia in tolerant Khaiyan and 1.48x higher for intolerant IR42. At the end of 72 h, PDC levels increased 4x for Khaiyan and only 2x for IR42 under hypoxic vs. aerobic conditions. PDC levels were significantly higher in Khaiyan than in IR42 germinating seeds from 12 h to 72 h. ADH activity was also higher in Khaiyan than in IR42 after 24 h of hypoxia treatment. The tolerant cultivar Khaiyan showed as much as 3x greater ADH activity compared with only a 1.5x increase in the intolerant IR42 after 72 h of hypoxia compared with that under aerated conditions. The higher PDC and ADH activities suggest the involvement of the alcohol fermentation pathway in the tolerant cultivars during hypoxia. Higher PDC and ADH enzyme activity probably supports the faster growth and development of tolerant lines than intolerant ones during hypoxia.

greater soluble sugars, particularly a few days after imbibition under hypoxia, and maintain higher activity of the key enzymes involved in anaerobic respiration (ADH and PDC). Further studies are ongoing to establish the full array of physiological and biochemical mechanisms associated with tolerance as well as to unravel their genetic bases. Backcross populations were developed and are being used to map this trait. Crosses were also made and are being selected and advanced to incorporate tolerance into modern varieties, for both irrigated and rainfed ecosystems.


These studies demonstrated substantial genetic variation in rice in ability to germinate and establish underwater; however, the trait is relatively rare. Tolerance is associated with faster root and shoot growth and leaf area development, faster coleoptile growth under hypoxia and anoxia, enhanced activity of total amylases, increased ethylene levels in germinating seeds, and lower peroxidase activity. Tolerant lines also seem to maintain

Salinity tolerance in rice: physiological bases and implications for management strategies for better crop establishment
Abdelbagi M. smail, Babita Thapa, and James Egdane

In most coastal areas, salinity is high in both soil and water during the dry season, mostly between December and July, and with the peak during May-June. Salinity then decreases progressively with time after the onset of the monsoon season. This poses great challenges for farming during the dry season, when most of these areas are often left barren or are grown to limited short-maturing crops when freshwater resources are available. For wetseason rice, the main problems are encountered during crop establishment in June-July, when salinity is still high at the beginning of the rainy season. This is particularly important because rice is highly sensitive to salt stress during early seedling growth, which is a major obstacle because of high seedling mortality and difficulty in establishing a sufficient crop stand. Varieties with reasonable genetic tolerance as well as proper management strategies during crop establishment are needed to ensure a good stand to significantly enhance and stabilize productivity in these coastal areas. For transplanted rice, damage to young seedlings is further provoked by the fact that more salt is absorbed passively through the injured roots, coupled with the fact that seedling growth and uptake of nutrients are greatly hindered during the first week after transplanting and until the seedlings are well established. Osmotic stress of the soil solution together with a massive uptake of salts and low nutrient uptake will make nutritional deficiencies even higher for some minerals and toxicities of others, causing higher seedling mortality and poor stand establishment. Based on our understanding of the mechanisms associated with tolerance of salt stress, we attempt to develop management strategies

that can reduce seedling mortality after transplanting and enhance crop establishment. Another objective is to compare rice cultivars contrasting in tolerance of salt stress for their responsiveness to such stress-mitigating options.

Nursery management for enhanced seedling survival
The effects of seedling N status, seedling age, and handling at transplanting on seedling survival after transplanting in saline soils were investigated. Two rice lines were used, a moderately tolerant variety, IR64, and a tolerant breeding line, IR651-4B-10-3 (referred to as IR651 henceforth). The two cultivars were grown in grid plastic trays, with one seedling per partition and half of the seedlings supplied with extra N at planting. Twenty-d-old and 40d-old seedlings with and without N treatments were transplanted either after washing their roots or with soil left attached to their root system (unwashed). The trial was repeated once with 4 replications each time. Average standard evaluation system (SES) scores of 6.4 and 1 were observed for plants grown under salt stress and normal conditions, respectively (Table 1). Average scores were significantly lower for the tolerant cultivar (5.1) than for the moderately tolerant cultivar (7.8), suggesting that IR651 experienced less salt injury and had better growth under salt stress than IR64. Likewise, transplanting older seedlings and seedlings with soil attached to their roots resulted in substantially better overall growth as evidenced from visual observations. The artificially salinized plot seems ideal for evaluating plants for salt stress


Table 1. Standard evaluation system (SES) scores and percentage survival of rice seedlings as affected by cultivar, age of seedlings in the nursery, and seedling handling under salt stress and normal conditions. Measurements were taken 20 d after transplanting. Variables Cultivar R65 R64 Seedling age 0 days old 40 days old Seedling handling Roots washed Roots not washed Salinity EC 8 dS m– Normal Significance Cultivar Age Seedling handling Salinity LSD0.05 (others) LSD0.05 for salinity

Table 2. Correlations of seedling survival with dry weight and carbohydrate content at transplanting. Trait Seedling dry weight Green leaf area Root dry weight Shoot dry weight Total carbohydrates Stem starch Leaf starch Shoot starch Root starch

R-valuea 0.65** 0.60* 0.68** 0.6* 0.6** 0.60* 0.68** 0.5* 0.40 ns

SES scores 5.06 7.75 7.56 5.0 6.93 5.87 6.40 .00 ***a *** ** *** 0.47 0.3

% survival 55.9 37.7 35.4 58. 38.6 55.6 46.9 00.0 ** *** *** *** 3.3 7.0

ns = not significant; *, ** = significant at P< 0.05 and 0.0, respectively.

*,**, *** = significant at P< 0.05, 0.0, and 0.00, respectively.

tolerance where a drastic reduction (53%) in plant establishment was observed compared with the normal plot (100%). Using older seedlings and seedlings transplanted with soil intact on roots showed survival of 58% and 56% compared with 35% and 39% for younger seedlings and seedlings with their roots washed before transplanting, respectively. The salt-tolerant line IR651 maintained significantly higher seedling survival than IR64. These results showed that survival of seedlings transplanted in saline soils could be enhanced substantially if older seedlings of tolerant varieties were used, particularly if their roots were protected during transplanting. Older seedlings have higher shoot and root dry weights and green leaf area; they also maintain higher starch content in the shoot, all of which are positively associated with survival

after transplanting (Table 2). High nonstructural carbohydrates in shoot and root could probably act as a reserve source of energy at the time of transplanting to overcome the period of slow growth until the transplanted seedlings retain their full capacity for photosynthetic carbon fixation. The older seedlings were taller and experienced less reduction in plant height under salt stress than the younger ones. IR651 had more tillers per plant under both saline and normal conditions and showed significantly lower reduction in tiller number under salinity. Younger seedlings produced more tillers per plant under normal conditions whereas the older seedlings produced more tillers under saline conditions and transplanted younger seedlings showed a higher percent reduction in tiller number (61%) than the older seedlings (41%) under salt stress. Salt stress delayed maturity by about 11 d, with greater delays in younger seedlings when roots are washed and in sensitive cultivars (Table 3). This study demonstrated that salinity stress at transplanting can reduce seedling survival by >50% and further decrease grain yield of surviving plants by about 37%, mainly because of a reduction in number of panicles per plant as determined by the number of surviving tillers and number of spikelets per panicle. Grain yield under salt stress also correlated positively with the number of panicles per plant


Table 3. Plant height, tiller number at 20 d after transplanting, and days to maturity as affected by cultivar, seedling age, and handling under salt stress and normal conditions. Values are means of 4 replications. tem Cultivar R659 R64 Age of seedlings 0 days old 40 days old Seedling handling Roots washed Roots not washed Salinity Cultivar Age Seedling handling Salinity Cultivar*salinity

Plant height (cm) Normal 50.5 48. 4.4 5.5 5.7 48. 49. ns ** ns *** ns EC 8 39.5 44. 33.4 43. 4.9 40. 4.0 % change 7.8 8. 9.3 6. 6.9 6.7 6.6 LSD0.05 – 3.7 – 3. – Normal 0.37 9.8 0.7 9.48 9.30 0.80 0.08

Tiller number EC 8 5.5 3.9 4.4 5.58 5.0 5.03 5.03 ** * ns *** *** % change 46.77 60.04 6.44 4.3 46.0 53.4 50.0 LSD0.05 0.49 0.50 – 0.5 0.69

Days to maturity Normal 4.6 98.0 05.4 07.4 06.9 05.9 06.4 *** ns *** ** ** EC 8 0.9 4.6 8.8 6.9 .8 3.8 7.9 % change 6.3 6.6 3.4 9.5 5.9 7.9 .7 LSD0.05 . – .04 .06 1.98




ns = not significant; *, **, *** = significant at P < 0.05, 0.0, 0.00, respectively.

and spikelets per panicle and to a lesser extent with individual grain weight but not with fertility. This is probably because salinity stress declined gradually 30 d after transplanting to simulate the actual conditions of coastal saline areas, and number of panicles as well as spikelets was determined earlier. Fertility, on the other hand, is mainly affected by pollen grain development and pollination, both of which occur at a time when salinity stress is low. The negative effects on individual grain weight are probably because of the effect of early stress on photosynthetic leaf area and stored assimilates. Older seedlings had higher dry matter and starch content that correlated positively with seedling survival under salt stress. In summary, seedlings transplanted with protected roots had higher shoot dry matter, less salt uptake, and earlier maturity. The positive consequences of using older seedlings and root protection are more evident in the salt-tolerant IR651 than in the moderately tolerant IR64. This enhanced responsiveness of the salt-tolerant cultivar to stress-mitigating 

amendments suggests that combining salt tolerance with proper nursery and seedling handling options can substantially improve crop establishment and early growth in saltaffected areas, which can later be reflected in higher grain yield.

Nutrient management to enhance seedling establishment under direct seeding
Despite the well-documented beneficial effects of Ca+ in mitigating salt stress in different plant species, contrasting observations were noted in rice. This is probably because in most of the studies conducted so far, only one or a few sensitive lines were used. Salt stress often coexists with P deficiency because affected soils are often either acidic (saline) or alkaline (sodic), and both conditions promote P fixation in forms that are poorly available for plants. Thus, the induced P deficiency in these soils could further worsen the detrimental effects of salt stress with consequent high seedling mortality. Salt tolerance is also associated with restricted toxic ion absorption and adequate uptake of essential inorganic nutrients, such as K+, to overcome the

nutritional imbalances caused by salinity stress. Here, we attempt to investigate the effects of calcium and phosphorus in enhancing tolerance of salt stress and in reducing seedling mortality. Furthermore, we used rice cultivars contrasting in salinity tolerance to test whether differential responses to additional Ca2+ and P could be observed in these genotypes. Two experiments were conducted using culture solutions; in each experiment, two levels of P (2 and 10 µL L–1) and three levels of Ca2+ (20, 40, and 60 µL L–1) were used under normal or salt stress (12 dS m–1) conditions. Three rice cultivars were used, one tolerant (IR651), one moderately tolerant (IR64), and one sensitive (IR29). The tolerant line had higher leaf area, higher root and shoot dry weight, and lower SES scores when evaluated 3 weeks after the start of the stress treatment (Table 4). The addition of higher levels of P and Ca2+ enhanced leaf area; however, P enhanced shoot growth while

Ca2+ enhanced root growth under salt stress. Higher levels of both nutrients resulted in better tolerance of salt stress across cultivars as shown by the significantly lower SES scores. Increasing P from 2 µL L–1 to 10 µL L–1 in the nutrient solution resulted in about a 145% increase in P concentration in shoots. It also enhanced the uptake of Mg2+ and Ca2+ and significantly reduced Na+ concentration in plant tissue under salt stress. However, the effect of higher P was greater in IR64, followed by IR651, and with IR29 showing an opposite effect, in which Na+ concentration in plant tissue increased at higher P. This result suggests that tolerant and moderately tolerant cultivars are more responsive to higher P in a nutrient medium. Increasing calcium concentration in the culture solution significantly reduced sodium concentration in shoots of all three cultivars and with a relatively greater reduction

Table 4. Leaf area (cm2 seedling–1), shoot and root dry weights (g), and SES scores of rice cultivars under normal conditions and salt stress (12 dS m–1) measured 21 d after the start of a salt-stress treatment. tem Leaf area Normal Cultivar R9 R64 R65 Phosphorus  µL L– 0 µL L– Calcium 0 µL L– 40 µL L– 60 µL L– Mean Significancea Salinity (S) Cultivar (C) Phosphorus (P) Calcium (Ca) LSD0.05 Salinity Cultivar P Ca 4. 3.6 .9 3.6 0.056 0.048 0.039 – 0.00 0.08 – 0.08 0.5 0.46 0. 0.46 *** *** *** ** *** *** *** ns *** *** ns *** *** *** *** *** 5. 53.4 55.5 53.7 4.9 7.7 36.7 9.8 0.45 0.454 0.444 0.437 0.83 0.94 0.97 0.9 0.4 0.6 0.80 0.37 0.066 0.08 0.089 0.079 .00 .9 .9 .94 8.4 7.78 7. 7.67 50.6 56.8 6.3 33.3 0.408 0.466 0.70 0.33 0.65 0.60 0.08 0.077 .90 .00 7.83 7.5 43.6 54.7 6.7 9.8 33. 36.4 0.407 0.460 0.444 0.5 0.3 0.348 0.46 0.36 0.30 0.054 0.079 0.03 .00 .9 .9 8.83 7.83 6.36 Stress Shoot dry wt. Normal Stress Root dry wt. Normal Stress SES scores Normal Stress 

Na+ total uptake in roots (µg SDW–) 8 4 3 0 8 6 4  0 K+ total uptake in roots (µg SDW–) 8 B 6 A

Na+ uptake in shoots (µg SDW–) 70 D 60

Ca 0 ppm Ca 40 ppm Ca 60 ppm

50 40 30 0 0 0 K+ uptake in shoots (µg SDW–) 70 60 50 40 E

4 30 3 0 0 0 N /K ratio in roots .5
+ +

0 N+/K+ ratio in shoots 3.5 F 3.0 .5

C .0 .5 

.0 .5 .0 


0.5 0.0

0.5 0.0 


R64 Cultivar 



R64 Cultivar 


Fig. 1. Uptake of Na+ and K+ and Na+/K+ ratio in roots (A, B, C) and shoots (D, E, F) of three contrasting rice cultivars as affected by different levels of Ca2+ in nutrient solution. SDW = shoot dry weight, RDW = root dry weight. Vertical bars indicate standard error. 

with increasing genetic tolerance. Sodium concentration decreased by 14%, 24%, and 35% in IR29, IR64, and IR651, respectively, with increasing Ca2+ concentration from 20 to 60 µL L–1. This also suggests that the response to higher calcium under salt stress is strongly dependent on the level of tolerance, with salttolerant cultivars being substantially more responsive. This is also clearly reflected in the total uptake of sodium in the shoot of salttolerant cultivars, in which total uptake was substantially lower in IR651 under higher calcium in the nutrient solution (Fig. 1D). Potassium concentration in plant tissue as well as total uptake into shoots (Fig. 1E) increased with increasing calcium concentration. Consequently, Na+/K+ ratio in shoots decreased significantly with increasing concentrations of calcium in the growth medium (Fig. 1F). Under salt stress, increasing Ca2+ levels also increased Ca2+ concentrations in all cultivars, but with a greater increase in IR651 and IR64, in which it increased by about 19% and 17%, respectively, compared with only 8% in IR29. Higher Ca2+ concentrations also reduced Na+/Ca2+ ratios in all cultivars. Root total Na+ uptake increased with increasing Ca2+ concentration in the culture solution in all cultivars (Fig. 1A), which contrasts with the trend observed for total sodium uptake into shoots, particularly in IR651 (Fig. 1D). Salt-tolerant cultivar IR651 showed the highest total sodium uptake into roots, whereas salt-sensitive cultivar IR29 showed the lowest Na+ uptake into roots. This greater compartmentation of sodium into roots of IR651 may partially explain its greater tolerance of salt stress, which seems to be further enhanced by supplementary Ca2+. Higher Na+ in roots could also act as an osmoticum to allow water uptake into roots of tolerant lines without much detriment to shoot growth. Total uptake of K+ into roots also increased with increasing calcium concentration in the nutrient solution and with the uptake being relatively higher in IR651, followed by IR64. Thus, Na+/K+ ratio in roots was highest at lower Ca2+ in IR64 and IR651.

The accumulation of Na+ in rice tissues under salt stress seems to influence the overall nutrient balance by changing the internal ion concentrations in shoots. Na+ concentration in plant tissue was negatively associated with K+ (R = –0.61**), Ca2+ (R = –0.30**), and Mg2+ (R = –0.28**), but positively correlated with Na+/K+ ratio (R = 0.89**) under salt stress, suggesting that the optimum balance of these essential nutrients could be deleteriously affected with increasing sodium uptake. Phosphorus showed a positive correlation with Mg2+ under both normal and salt-stress conditions. However, Ca2+ concentration in shoots under salt stress showed a strong positive correlation with Mg2+ (R = 0.72**), P (R = 0.38**), and K+ (R = 0.29**), and a negative correlation with Na+. Based on these findings, it seems that an addition of more calcium under saline conditions is beneficial because it helps reduce Na+ uptake while enhancing the uptake of other essential nutrients such as magnesium, phosphorus, and potassium, whose uptake was negatively affected under higher salt stress. The synergistic effects of the use of genetic tolerance combined with the mitigating effects of P and Ca+ are further summarized in Figure 2, in which a lower ratio of Na+/K+ in the shoot is taken as an indicator of lower salt injury. The ratio is very high in IR29 but decreased with the addition of Ca2+ as well as when both Ca2+ and P were combined. IR64 with its intermediate level of tolerance showed a better response to both P and Ca2+ when applied separately, and a
Na+/K+ ratio in shoots .0 .6 . 0.8 0.4 0.0 R9 R64 Cultivar R65 Cultivar (C) C+P C + Ca C + P + Ca

Fig. 2. Na+/K+ ratio in shoots as affected by genotype and P and Ca2+ in three rice cultivars contrasting in their response to salt stress. 


greater response when the two nutrients were combined. IR651, on the other hand, showed a response similar to that of IR29 but with a strong genetic effect, particularly when the two nutrients were added together. A genotype effect is also obvious. Na+/K+ ratio in plant tissue drops to <0.4 when genetic tolerance is combined with higher Ca2+ and P. So, these amendments seem to be effective only when combined with intermediate or high levels of genetic tolerance. In summary, our studies demonstrated a few nursery and nutrient management strategies that can effectively enhance crop establishment

in salt-affected areas. However, these strategies may be effective only if they are applied using salt-tolerant cultivars. This is because the incremental effects of genetic tolerance seem to act additively to the effects of these mitigating strategies. Combining salt tolerance with proper nursery and nutrient management options can therefore substantially improve crop establishment and early growth in saltaffected coastal areas, which can subsequently enhance and stabilize rice productivity in these ecosystems. Further studies are needed to validate these strategies under farmers’ field conditions. 


Opportunities for direct seeding and improved weed control in the Barind of Bangladesh
M.A. Mazid, C.R. Riches, A.M. Mortimer, and D.E. Johnson

In the High Barind Tract, in northwest Bangladesh, a single crop of transplanted rainfed rice (TPR), grown in the monsoon aman season from June to October, provides a major component of rural livelihoods. Aman rice is vulnerable to late-season drought during grain filling in October and in the rabi (winter) season much of the land lies fallow. Cultivation intensity in much of the Barind is considerably less than in districts where irrigation allows two or three rice crops to be grown each year. Farmers’ lands are typically distributed over a shallow sloping landscape or toposequence. Two challenges of agricultural improvement in such areas are to simultaneously improve the reliability and yield of aman rice while increasing total system productivity. Research in recent years has demonstrated that these objectives can be achieved through the introduction of dry-seeded rice (DSR) and the planting of short-duration rabi crops (e.g., mustard or chickpea) on residual moisture immediately after the rice harvest. Late onset of the monsoon or low rainfall can delay rice transplanting as a minimum of 600 mm of cumulative rainfall is needed to complete land preparation and transplanting. Dry seeding, on the other hand, can be completed after land preparation by a power tiller after much less rainfall and the earlier planted DSR crop matures 1–2 weeks before TPR, thus reducing the risk of terminal drought, and allows earlier planting of a following nonrice crop. TPR requires less labor and draft power for rice establishment than DSR, but the high costs associated with weed control in DSR are a major constraint to its adoption. Monitoring of farmer-managed transplanted aman rice crops in the Barind revealed that labor availability constrains the timeliness of first weeding for

many households and, with current practices, 34% of farmers lose more than 0.5 t ha–1 of the attainable yield because of weed competition (Mazid et al 2001). Farmers in the Barind have a strong preference for the late-maturing rice cultivar Swarna. Use of this cultivar, however, reduces the opportunity for establishing chickpea or other rabi crops on residual moisture, whereas growing earlier maturing modern cultivars may contribute to an earlier harvest. This report summarizes selected findings (after Mazid et al 2003) from a longterm field experiment in the Barind designed to explore the contribution of rice establishment method, rice cultivar duration, and weed control practices to aman rice performance and the likely long-term impact on the composition of the rice weed flora.

Rice establishment, nutrient management, and weeding practices have been investigated on farmland at Rajabari, Rajshahi, in northwest Bangladesh from 2001 in an ongoing long-term trial as described in greater detail in Mazid et al (2001). The results for rice crops in 2000, 2001, and 2002 are reported, comparing rice crop establishment methods and weed management practices. Treatments were (1) transplanted rice (TPR)—soil is puddled prior to transplanting; the crop is hand-weeded twice at 30 and 45 days after transplanting (DAT); (2) direct-seeded rice (DSR)—soil is plowed prior to seeding in rows by hand, with hand weeding at 21, 33, and 45 days after sowing (DAS); (3) direct-seeded rice with chemical weed control (DSRH)—as for DSR but with oxadiazon (375 g a.i. ha–1) applied 2–4 days after seeding, followed by one hand weeding at 33 DAS. Plots of these treatments were sown to the cultivars Swarna (maturity 

140–145 days) and BRRI dhan39 (maturity 120–125 days). Rice was harvested in 5-m2 plots. Biomass of individual weed species was recorded in two unweeded quadrats per plot at 28 days DAS/DAT and total weed biomass at 45 DAS/DAT and again at harvest.

Crop establishment method With the exception of BRRI dhan39 in 2000, yields from direct seeding of this and cv. Swarna were as good as or better than from transplanting, the usual method of rice culture in the district (Fig. 1). Early-season weed control by preemergence application of herbicide resulted in the highest yields, except for BRRI dhan39 in 2000.
Grain yield (t ha –) 5 A 4 000 3  



5 B 4 3  

Establishment method Fig. 1. Effect of establishment and weed control practices on the yield (mean ±S.E.M) of rice cultivars BRRI dhan39 (A) and Swarna (B). DSR = direct-seeded, hand-weeded; DSRH = direct-seeded + herbicide; TPR = transplanted rice + hand-weeded.

Weed species shifts The weed flora of rainfed rice in the Barind is diverse and exhibits high interseasonal variability depending on water regimes at rice establishment and soil moisture status of toposequence position. Weed species present in the experiment included Alternanthera sessilis, Ammania baccifera, Cyanotis axillaris, Cynodon dactylon, Cyperus difformis, Cyperus iria, Cyperus rotundus, Cyperus tenuispica, Echinochloa 

colona, Eclipta prostrata, Eriocaulon cinereum, Fimbristylis dichotoma, Fimbristylis miliacea, Hedyotis corymbosa, Lindernia ciliata, Ludwigia sp., Monochoria vaginalis, Paspalum distichum, and Sphaeranthus indicus. At harvest, there were significantly higher densities of weeds in DSR (228 m–2) than in TPR (75 m–2; P ≤ 0.023). As expected, however, at 45 DAS/DAT, the least weed density and biomass were recorded in DSRH. The range of responses by individual weed species over three consecutive seasons to crop establishment and weed management practices is shown in Figure 2. An increase in abundance (biomass at 28 DAS/DAT) of the broadleaf species Alternanthera sessilis, Eclipta prostrata, Lindernia ciliata, and Ludwigia sp. and the sedges Cyperus difformis and Fimbristylis miliacea was noticeable in DSR. Conversely, the biomass of Monochoria vaginalis was decreased by direct seeding. The most noticeable increase in abundance was seen in the perennial grass Paspalum distichum. A long-term trial has demonstrated that, although rice yield can be maintained with the switch from transplanting to direct seeding, farmers will face a greater weed problem early in the crop season. Not only is there an increased burden of weeds in direct-seeded rice but the change in establishment practice also leads to a shift in the relative abundance of important species. Previous findings indicate that direct seeding was associated with higher labor inputs for first weeding than is the case for transplanting. With the labor constraint, late first weeding, and a significant yield gap due to weeds on many farms in transplanted rice with current weed control practices (Mazid et al 2001), it is clear that the adoption of direct seeding will need to be associated with the use of chemical weed control. Studies suggest that herbicides may find a ready market in Rajshahi District because (1) weeding is done almost exclusively by hired labor and (2) the supply of hired labor is local, with very little weeding done by seasonal migrant labor. Together, these factors will create intense competition for labor, especially on larger farms, which would be intensified by the adoption of direct seeding.





















Mean dry biomass g m- at 8 DAS/DAT 7 9
Alternanthera sessilis

Cyperus difformis 

0 5 0 5

Cyperus iria


7 5



Eclipta prostrata

Fimbristylis miliacea

Ludwigia sp.

6 3 30 4  0 

Lindernia ciliata 

Monochoria vaginalis 

Paspalum distichum

7 0 5 3  5 

50 00 50
















Establishment method Fig. 2. Effect of establishment and weed control practices on the biomass of nine rice weeds at 28 days after planting in unweeded plots. See text for details.


Direct seeding advanced the rice harvest by 7 to 10 days. Earlier harvest reduces the problem of terminal drought in rice when rains end abruptly in October and will ensure that postrice crops are sown while seedbeds are moist. Farmers with the least land under cultivation (0.6 ha for the lowest quartile of households) on average plant 43% to postrice crops, often on the least favorable land where moisture is limiting (Mazid et al 2003). Our trials suggest that this group can maximize rice yield and achieve timely planting of a highvalue chickpea crop by direct seeding rice. On larger farms (> 2.5 ha for the upper quartile), a lower proportion of land is planted after rice, using more favorable soils. Adoption of DSR by this group could increase the area planted to chickpea.

A single rice crop, combined with land pressure and a high level of sharecropping in the Barind Tract, leads farmers to place a premium on optimizing rice yield and household food security. This study demonstrates that, although the widely grown rice cultivar Swarna performs well under direct seeding, the shorter duration BR39 is not well adapted for this planting practice. This is because of the high levels of sterility in this cultivar associated with flowering during wet periods. An earlier maturing variety with yields to match those of Swarna could contribute further to avoiding late-season drought during grain filling and also allow earlier planting of postrice crops. This challenge requires a broadbased approach combining rice breeding with agronomy and weed science. 




Information availability With a shift to direct seeding, farmers will be increasingly dependent on information from outside sources. Indeed, successful adoption and correct decision making would be likely only if farmers had a greater availability of current knowledge. Direct seeding and associated weed management comprise not single recommendations but a wide range of options that will be dependent on toposequence, seasonal effects/rainfall pattern, the weeds present, and farmers’ resources. Decision trees can provide farm-level information in the form of structured questions that enable answers “in the form of options” to be chosen (Johnson and Mortimer 2004). These trees specifically focus on technical issues related to the adoption of a particular system and they may be useful tools to help refine technology options for researchers and extension staff. For a farmer in the favorable

rainfed environment, such as the Barind, the question “What are my options for rice establishment?” might initiate a tree involving several steps and covering the range of directseeding options (see Fig. 3 for an example). In this example, the decision process recognizes that a primary consideration will be the ability of a farmer to drain his field as only if this is possible should direct seeding be recommended because of the risk of early flooding. A second decision level will involve the choice between wet and dry cultivation and this will depend on the ability to dry-cultivate and/or the presence of perennial weeds. Further levels indicate choices for row or broadcast seeding and weed control options. Such a diagram allows a structured approach to the range of options available, and a more complex decision tree may comprise weed control options for individual or groups of weeds.



Yes DRY SEEDING into a seedbed

No WET SEEDING sowing onto puddled saturated soil





Fig. 3. Illustrative decision tree for the adoption of direct seeding with respect to favorable rainfed lowland rice. 


The transition to direct seeding and the management of challenging weed problems will require substantial information to enable farmers to judge objectively what the best technology options are. Gaining access to such information may be a major obstacle for potential adopters. The challenge for researchers is to adequately address the variability of the rice-farming systems for which they are making recommendations and to synthesize the results in ways that will make the conclusions available to those who will use them. Studies on direct seeding, weed management, and decision tools are ongoing.

Johnson DE, Mortimer AM. 2004. Issues for integrated weed management and decision support in direct-seeded rice. In: Rice is life: scientific perspectives for the 21st century. Proceedings of the World Rice Research Conference held in Tokyo and Tsuukuba, Japan, 4-7 November 2004. Los Baños (Philippines): International Rice Research Institute, and Tsukuba (Japan): Japan International Research Center for Agricultural Science. CD. p 211-214. Mazid MA, Jabber MA, Riches CR, Robinson EJZ, Mortimer M, Wade LJ. 2001. Weed management implications of introducing dry-seeded rice in the Barind Tract of Bangladesh. Proceedings of the BCPC Conference – Weeds 2001. 1:211-216. Mazid M, Jabber MA, Mortimer M, Wade L, Riches CR, Orr AW. 2003. Improving rice-based cropping systems in north-west Bangladesh: diversification and weed management. Proceedings of the BCPC International Congress on Crop Science and Technology — 2003, SECC, Glasgow, UK. p 1029-1034.

The research in Rajshahi was conducted by the Bangladesh Rice Research Institute, Natural Resources Institute (UK), and University of Liverpool in association with the Consortium for Unfavorable Rice Environments (CURE), and was partially funded under the Crop Protection Programme by the Department for International Development, UK. 


Breeding for submergence tolerance
D.J. Mackill, A.M. smail, S. Heuer, E. Septiningsih, A.M. Pamplona, R.M. Rodriguez, C.N. Neeraja, D. Sanchez, K. ftekhar, and G. Vergara

This project aims to develop submergencetolerant cultivars that will improve the livelihood and food security of farmers and rice consumers in submergence-prone areas. The major objective is to convert widely grown varieties in rainfed lowland areas into submergence-tolerant varieties through incorporation of the Sub1 QTL on chromosome 9.

Some 11 million hectares of shallow rainfed lowland rice in South and Southeast Asia are submergence-prone, and another 5 million ha of medium-deep area experience stagnant flooding of up to 50 cm. Submergence also affects some of the areas classified as deepwater (sustained water depths above 50 cm) on around 4 million ha in Asia. Some estimates indicate that submergence stress causes annual losses of around US$1 billion in Asia (Dey and Upadhyaya 1996, Herdt 1991). The “normalized” yield loss (an index taking into account several parameters) in submergence-prone areas was estimated at about 80 kg ha–1, causing a production loss of about 3.2 million tons per year, with a value of about $384 million. About 140 million people are at risk from flooding damage in Bangladesh and five states of eastern India. With an average poverty ratio of 45%, approximately 74 million poor people stand to benefit significantly from improved submergence-tolerant rice cultivars. Modern high-yielding rice cultivars are seriously damaged if they are completely submerged for a few days; however, a few tolerant landraces were identified that can withstand inundation for up to 2 weeks. In these tolerant landraces, the Sub1 major 

QTL accounts for most of the variation in the trait. DNA markers linked to the Sub1 locus have facilitated its transfer into widely grown cultivars that are locally adapted and possess the quality aspects preferred by local consumers. Those tolerant Sub1 cultivars have considerably higher survival and yield under submergence than susceptible cultivars. Rice production can therefore be improved and stabilized in submergence-prone areas of South and Southeast Asia predominantly inhabited by resource-poor farmers. The value of submergence tolerance could be further enhanced by combining it with tolerance for medium-deep stagnant flooded conditions and tolerance for submergence during germination. Yield of these tolerant cultivars can be further increased and stabilized through proper management strategies.

Previous studies on submergence tolerance
The effects of flooding on rice as well as the physiological bases of tolerance were recently reviewed (Ram et al 2002, Jackson and Ram 2003). Plant survival in flooded areas depends on various aspects of floodwater environments, particularly the limitation of gas diffusion, irradiance level, and water temperature. Among the important plant traits associated with tolerance are high nonstructural carbohydrate content before submergence, slower underwater shoot extension, optimum alcoholic fermentation when O2 is low, an efficient protective system upon air entry after exposure to low O2, and limited leaf chlorosis (Setter et al 1997, Ram et al 2002, Jackson and Ram 2003, Ella et al 2003). Carbohydrates remaining after submergence are necessary for recovery growth and are correlated better with survival than

carbohydrate level before submergence (Das et al 2005). A rapid regeneration growth following submergence is essential under frequent or prolonged flooding as this can ensure early recovery and the production of sufficient biomass for high yield. Rice plants that exhibit only limited elongation during submergence are more tolerant of complete flooding and a strong association between limited underwater shoot growth and survival is commonly observed (Jackson and Ram 2003, Das et al 2005). Evidence is also accumulating for the role of postsubmergence events in tolerance for submergence. Tolerant cultivars acquired a more efficient protective system to suppress the level of active oxygen species and to lower the extent of lipid peroxidation upon exposure to air (Kawano et al 2002). Ethylene, a plant hormone, accumulates in plant tissue during submergence because of both enhanced synthesis and entrapment, and this promotes underwater leaf senescence. This effect is suppressed in tolerant cultivar FR13A. Ella et al (2003) found that blocking ethylene enhanced chlorophyll retention and carbohydrate content in the plant tissue and improved survival of intolerant cultivar IR42. Early flooding can cause poor crop establishment in direct-seeded rice areas. Landraces tolerant of these conditions emerge faster from the soil during flooding, produce taller seedlings with more leaves, and attain greater leaf area than intolerant lines. These cultivars also maintain higher activity of enzymes involved in the breakdown of starches, for example, total amylases, under flooding and this correlated positively with survival. Other studies showed increased activity of some of the enzymes involved in the fermentative pathways during anaerobic conditions (Xie and Wu 1989, Hossin et al 1996). Phosphofructokinase (Fukao et al 2003) and pyruvate orthophosphate dikinase (Huang et al 2005) were found to be induced under low-oxygen stress, both of which could enable substrate cycle operation of adenosine tri-phosphates (ATPs) necessary for energy production during germination. Breeding improved submergence-tolerant cultivars has been ongoing for more than

three decades (Mackill 1986, Mohanty and Chaudhary 1986, Singh and Dwivedi 1996). The initial work focused on transferring the trait from traditional landraces into semidwarf breeding lines. However, these lines were low-yielding and had many undesirable traits. Additional crosses resulted in the development of tolerant breeding lines with improved agronomic characteristics (Mackill et al 1993, Mackill and Xu 1996, Mohanty et al 2000). Although the lines with the highest levels of tolerance had some yield penalty, some breeding lines, such as IR49830-7, had a yield equivalent to that of the irrigated checks. These improved lines have been used for further crosses. New breeding lines with submergence tolerance were developed through the Eastern Indian Rainfed Lowland Shuttle Breeding Network (Singh et al 1998, Mallik et al 2002). Some of these lines have been released or recommended for release in India, such as Kishori, Satyam, OR1234-12-1, CN 1035-61 (Bhudev), CRLC 899 (Varshadhan), TTB 238-338-3 (Prafulla), NDR 8002, CR 2003-2, CR 20033, CR 978-8-2, and IR54112-B2-1-6-2-2-2-CR2-1. These lines together with Sub1 introgression lines would serve as a basis for studies on crop management and farmer participatory research. Genetically, submergence tolerance is largely governed by a single major QTL, designated Sub1, located on rice chromosome 9 (Xu and Mackill 1996). Almost all strongly tolerant cultivars possess the Sub1 QTL (Setter et al 1997); however, additional QTLs of smaller effect appear to give increasing amounts of tolerance (Nandi et al 1997). Through positional cloning, a cluster of three putative ethylene response factor (ERF) genes has been identified in the Sub1 locus. ERFs are transcription factors unique to plants, in which they constitute a large multigene family related to Apetala2 (AP2) and dehydration-responsive element binding (DREB) factors (McGrath et al 2005). ERF genes are induced in response to several biotic and abiotic stresses, and by ethylene and other plant hormones, and might be involved in cross-talk between the different pathways (for review, see Gutterson and Reuber 2004). In addition, it was shown in rice and tomato 

that phosphorylation of the proteins enhances DNA binding affinity (Cheong et al 2003, Gu et al 2000) and that ERFs act as transcriptional activators and repressors, respectively (Fujimoto et al 2000, Ohta et al 2001). An allelic survey of the Sub1 ERF genes in a range of tolerant and intolerant germplasm revealed tolerant-specific alleles for Sub1A (Sub1A-1) and Sub1C (Sub1C-1). The tolerant and intolerant alleles (Sub1A-2; Sub1C-2 to 8) show differences in several putative phosphorylation sites and are differentially expressed during submergence. Whereas Sub1A-1 is highly expressed in tolerant lines and expressed at a very low level in intolerant accessions, Sub1C shows the opposite expression pattern. Overexpression of Sub1A-1 in an intolerant variety conferred submergence tolerance, suggesting that Sub1A-1 is the major determinant of tolerance (Xu et al 2006). The sequence information of these three genes facilitated the development of ideal markers suitable for backcrossing this locus into widely grown varieties. The use of markerassisted backcrossing has been shown to be effective in transferring Sub1 into a widely grown Thai cultivar (Siangliw et al 2003). In a previous project, the Sub1 genes were introduced into widely grown varieties in the rainfed areas of Asia. The sequences are also being used to screen new landraces for the presence of this gene. Preliminary data indicate that some tolerant varieties such as FARO 27 do not have the typical Sub1 gene and might possess a different mechanism of tolerance. Preliminary evaluation of Sub1 introgression lines showed that Sub1 can be effective in conferring tolerance for 10–14 days, depending on floodwater conditions. However, submergence sometimes occurs for a longer duration or more than once. Identification of new sources of tolerance and genes additive to Sub1 will therefore be highly desirable.

Recent studies at IRRI
The physiological basis of tolerance for flash flooding is now reasonably well understood. Two main factors were identified as being 

important in contributing to injury when rice is completely inundated: limited gas exchange and reduced illumination. The consequences of these are reduced underwater photosynthesis, accelerated stem and leaf extension, and enhanced chlorosis and leaf senescence, resulting in a shortage in energy supply for maintenance of metabolism. The mechanisms by which tolerant cultivars depress the damaging effects of submergence are becoming more evident. Identification of the traits associated with submergence tolerance will help in designing efficient evaluation methods to pyramid component traits and in gene discovery. In addition, suitable management strategies could be efficiently designed and tuned toward exploiting the potential of the traits that are important for survival and recovery. Initial work at IRRI indicated that submergence tolerance in the most tolerant cultivars is mainly controlled by Sub1 and that the most tolerant cultivars such as FR13A, Goda Heenati, and Kurkaruppan all possess this locus (Xu and Mackill 1996, Setter et al 1997, Xu et al 2006). However, data suggest that other genes are needed to gain higher submergence tolerance. We have identified a tolerant-specific allele of the Sub1C gene and have shown differential expression of this gene in tolerant and intolerant accessions upon submergence. It is currently unclear whether the relative abundance of Sub1A and Sub1C gene products is important for tolerance or whether Sub1 modulating factors are absent from intolerant varieties. Such factors might be present in additional minor QTLs controlling submergence tolerance (Nandi et al 1997, Kamolsukyunyong et al 2001) and could explain why some tolerant breeding lines do not have as high a tolerance as FR13A. Crosses were made with these lines for further genetic and mechanistic analysis. Breeders have been using the sources of submergence tolerance such as FR13A and Kurkaruppan to develop highly tolerant cultivars with a high-yielding plant type. The initial semidwarf breeding lines with submergence tolerance similar to that of FR13A

were developed in the late 1970s (described in Mackill and Xu 1996). Breeding lines with submergence tolerance from FR13A and high yield potential were developed in the late 1980s (Mackill et al 1993). More recently, breeding lines developed from these sources have shown promise in eastern India and a few of them were released as new varieties.

Summary of recent achievements
• Development of Swarna with submergence tolerance: A submergence-tolerant version of Swarna was produced in 2 years by marker-assisted backcrossing (MAB): only the small fragment carrying the QTL Sub1 was introduced. A large number of BC1F1 seeds were produced. Based on five informative markers from the Sub1 region, 54% of the plants were detected to be Sub1+, out of which 3% (21 plants) had desired recombination for the Sub1 region. Background selection was performed for the 12 chromosomes with 58 markers (approx. 5 markers per chromosome). From the progeny of 320 plants from the second backcross, four plants were selected based on foreground and background selection. Out of 937 BC2F2 progeny from the selected plants, one plant was selected for its maximum recipient genome and its target locus. • Providing NARES partners with multiplied seeds of Swarna-Sub1: The seeds of Swarna-Sub1 have been multiplied and used by CRRI, BRRI, and NDUAT for phenotypic evaluation. In preliminary yield trials at Rangpur, Bangladesh, and Cuttack, India, Swarna-Sub1 was evaluated against other elite breeding lines and checks. Swarna-Sub1 showed higher survival (25% for Swarna vs. 100% for SwarnaSub1), a lower level of elongation under submergence, and earlier maturity and it did not lodge after 10 days of submergence. Data on yield and other attributes are being analyzed. Swarna-Sub1 was evaluated for adaptation to local conditions by growing it

in normal farmers’ fields at Cuttack, Orrissa, India, and Rangpur, Bangladesh. In India, both Swarna and Swarna-Sub1 produced similar yields in farmers’ fields (approx. 5.5 t ha–1), whereas, at Rangpur, SwarnaSub1 slightly outyielded Swarna (3.9 t ha–1) and the local check Red Swarna (3.5 t ha–1). Swarna-Sub1 also had good plant height and panicle number. Seeds of Swarna-Sub1 were multiplied at two sites in both India and Bangladesh to produce sufficient seeds for large-scale testing in multiple farmers’ fields in 2007 through participatory varietal selection trials. • Identification of diagnostic markers for precise MAB of Sub1: Previously, two cleaved amplified polymorphic markers (CAPs) were designed targeting a silent single nucleotide polymorphism (SNP) located in the Sub1A gene and a unique phosphorylation site in the Sub1C gene. To more precisely measure the introgression region of the Sub1 locus and to obtain markers specific for functional genes underlying the QTL, additional allelespecific markers were developed. A second SNP where the IR40931 allele causes an amino acid change in the Sub1A protein (CCG in Teqing encoding for proline and TCG in IR40931-33 encoding for serine) was targeted for marker design. Since no restriction enzyme sites were located at the SNP locus that could be used to develop a CAP marker, a dominant STS marker was developed by designing a PCR primer with the SNP at the 3′ end. In addition, several CAPs and indel (insertion/deletion) markers were also designed in the promoter region of Sub1A and Sub1C. Several of these gene-based markers have been used in our MAB program, and have been found very useful as alternative foreground markers. The Sub1A marker has already been used in CRRI, India, to confirm the presence of the gene in some of their landraces and improved varieties. 


• Identification of additional microsatellite markers for foreground/recombinant selection of the Sub1 locus: Additional highly polymorphic SSR primers from the International Rice Genome Sequencing Project (IRGSP) were identified in the region of Sub1. Two of them were tightly linked and located upstream of the Sub1 locus, that is, RM23865 (6.2 Mb) and RM23869 (6.3 Mb). These two SSR markers have been very useful in minimizing the size of introgression of the Sub1 locus since previously there were no tightly linked polymorphic SSR markers located upstream of the Sub1 locus. • Development of Samba Mahsuri and IR64 with submergence tolerance: Previously, it was reported that Samba Mahsuri-Sub1 and IR64-Sub1 were produced within two rounds of backcrossing and one generation of self-pollination by MAB. Only the small fragment around the tip of chromosome 9, where the QTL Sub1 is located, was introduced into Samba Mahsuri, a popular variety from India, and IR64. In addition, another version of Samba Mahsuri-Sub1 with a smaller introgression of the Sub1 region was selected among 48 BC3F2 progenies derived from a double-side crossover type of BC3F1 plant. A similar process was conducted for IR64. Seven hundred BC3F2 plants derived from three selected BC3F1 plants have been genotyped; however, no double-side crossover has been found. Nevertheless, several promising recombinants having a smaller region of Sub1 have been identified. We will genotype the BC3F3 progenies of these recombinants to identify the best plants having the smallest introgression region of the Sub1 locus. This plant can then be multiplied and used as an alternative seed source of IR64-Sub1. Both versions of Samba Mahsuri-Sub1 (BC2F2 and BC3F2) and IR64-Sub1 (BC2F2 and BC2F3) have been used for seed multiplication. A preliminary submergence field test has shown that Samba Mahsuri-Sub1 and IR64Sub1 were comparable with the tolerant check. 

• Development of TDK1 with submergence tolerance: Because of the small size of this population from the earlier backcross and several markers that were biased toward the tolerant parent, there were no optimal plants to be selected out of the BC2F2 population. However, several BC3F1 with very small Sub1 introgression regions having no or one background introgression have been identified. In addition, plenty of BC3F2-derived BC3F1 seeds are available to maximize the probability of finding the best version of TDK1-Sub1. These BC3F2 plants have been planted and will soon be ready for genotyping. • Development of BR11 with submergence tolerance: A special case study was done for BR11. Four selection strategies associated with a marker-assisted backcross breeding program were compared for validating suitable selection strategies for an efficient and effective MAB breeding program. The four treatments related to the development of BR11-Sub1 were as follows: a. Foreground, recombinant, and background selection b. Foreground and phenotypic selection c. Foreground, phenotypic, and background selection d. Foreground, recombinant, and phenotypic selection As many as 1,430 BC1F1 plants were grown from 44 F1 plants, and the best BC2F1 progenies had been selected. Selection activities are ongoing in the BC2F2 generation for treatment “b” and in the BC3F1 generation for treatment “c.” However, we could not advance the selection processes a further generation for treatment “d” because of the unavailability of double-recombinant-type plants segregating in the BC2F1 generation. Preliminary results showed that phenotypic selection cannot be used as an alternative for a marker-based background in the MAB scheme.

• Development of CR1009 with submergence tolerance: DNA of CR1009 BC2F2 and BC3F1 plants has been isolated and is ready for genotyping.

Cheong YH, Moon BC, Kim JK, Kim CY, Kim MC, Kim IH, Park CY, Kim JC, Park BO, Koo SC, Yoon HW, Chung WS, Lim CO, Lee SY, Cho MJ. 2003. BWMK1, a rice nitrogen-activated protein kinase, locates in the nucleus and mediates pathogenesis-related gene expression by activation of a transcription factor. Plant Physiol. 132(4):1961-1972. Das KK, Sarkar RK, Ismail AM. 2005. Elongation ability and non-structural carbohydrate levels in relation to submergence tolerance in rice. Plant Sci. 168:131-136. Dey MM, Upadhyaya HK. 1996. Yield loss due to drought, cold and submergence tolerance. In: Evenson RE, Herdt RW, Hossain M, editors. Rice research in Asia: progress and priorities. CAB International, Wallingford, UK. p 291-303. Ella ES, Kawano N, Yamauchi Y, Tanaka K, Ismail AM. il 2003. Blocking ethylene perception enhances ��ooding tolerance in rice seedlings. Func. Plant Biol. 30:813-819. Fujimoto SY, Ohta M, Usui A, Shinshi H, Ohme-Takagi M. 2000. Arabidopsis ethylene-responsive element binding factors act as transcriptional activators or repressors of GCC box-mediated gene expression. Plant Cell 12:393404. Fukao TR, Kennedy A, Yamasue Y, Rumpho ME. 2003. Genetic and biochemical analysis of anaerobicallyinduced enzymes during seed germination of Echinochloa crus-galli varieties tolerant and intolerant of anoxia. J. Exp. Bot. 54:1421-1429. Gu YQ, Yang C, Thara VK, Zhou J, Martin GB. 2000. Pti4 is induced by ethylene and salicylic acid, and its product is phosphorylated by the Pto kinase. Plant Cell 12:771786. Gutterson N, Reuber TL. 2004. Regulation of disease resistance pathways by AP2/ERF transcription factors. Curr. Opin. Plant Biol. 7:465-471. Herdt RW. 1991. Research priorities for rice biotechnology. In: Khush GS, Toenniessen GH, editors. Rice biotechnology. Oxon (UK): CAB International. p 19-54. Hossin MA, Huq E, Growver A, Dennis ES, Peacock WJ, Hodeg TK. 1996. Characterization of pyruvate decarboxylase genes from rice. Plant Mol. Biol. 31:761770. Huang S, Greenway H, Colmer TD, Millar AH. 2005. Protein synthesis by rice coleoptiles during prolonged anoxia: implications for glycolysis, growth and energy utilization. Ann. Bot. 96:703-715. Jackson MB, Ram PC. 2003. Physiological and molecular basis of susceptibility and tolerance of rice plants to complete submergence. Ann. Bot. 91:227-241. Kawano N, Ella E, Ito O, Yamauchi Y, Tanaka K. 2002. Comparison of adaptability to ��ash ��ood between rice cultivars differing in ��ash ��ood tolerance. Soil Sci. Plant Nutr. 48:659-665.

Kamolsukyunyong W, Ruanjaichon V, Siangliw M, Kawasaki S, Sasaki T, Vanavichit A, Tragoonrung S. 2001. Mapping of quantitative trait locus related to submergence tolerance in rice with aid of chromosome walking. DNA Res. 8:163-171. Mackill DJ. 1986. Rainfed lowland rice improvement in South and Southeast Asia: results of a survey. In: Progress in rainfed lowland rice. Los Baños (Philippines): International Rice Research Institute. p 115-144. Mackill DJ, Amante MM, Vergara BS, Sarkarung S. 1993. Improved semidwarf rice lines with tolerance to submergence of seedlings. Crop Sci. 33:749-753. Mackill DJ, Xu K. 1996. Genetics of seedling-stage submergence tolerance in rice. In: Khush GS, editor. Rice genetics III. Manila (Philippines): International Rice Research Institute. p 607-612. Mallik S, Mandal BK, Sen SN, Sarkarung S. 2002. Shuttlebreeding: an effective tool for rice varietal improvement in rainfed lowland ecosystem in eastern India. Curr. Sci. 83:1097-1102. McGrath KC, Dombrecht B, Manners JM, Schenk PM, Edgar CI, Maclean DJ, Scheible WR, Udvardi MK, Kazan K. 2005. Repressor- and activator-type ethylene response factors functioning in jasmonate signaling and disease resistance identified via a genome-wide screen of Arabidopsis transcription factor gene expression. Plant Physiol. 139:949-959. Mohanty HK, Chaudhary RC. 1986. Breeding for submergence tolerance in rice in India. In: Progress in rainfed lowland rice. Los Baños (Philippines): International Rice Research Institute. p 191-200. Mohanty HK, Mallik S, Grover A. 2000. Prospects of improving ��ooding tolerance in lowland rice varieties by conventional breeding and genetic engineering. Curr. Sci. 78:132-137. Nandi S, Subudhi PK, Senadhira D, Manigbas NL, Sen-Mandi S, Huang N. 1997. Mapping QTLs for submergence tolerance in rice by AFLP analysis and selective genotyping. Mol. Gen. Genet. 255:1-8. Ohta M, Matsui K, Hiratsu K, Shinshi H, Ohme-Takagi M. 2001. Repression domains of class II ERF transcriptional repressors share an essential motif for active repression. Plant Cell 13:1959-1968. Ram PC, Singh BB, Singh AK, Ram P, Singh PN, Singh HP, Boamfa EI, Harren FJM, Santosa E, Jackson MB. 2002. Physiological basis of submergence tolerance in rainfed lowland rice: prospects of germplasm improvement through marker aided breeding. Field Crops Res. 76:131-152. Setter TL, Ellis M, Laureles EV, Ella ES, Senadhira D, Mishra SB, Sarkarung S, Datta S. 1997. Physiology and genetics of submergence tolerance in rice. Ann. Bot. 79:67-77. Siangliw M, Toojinda T, Tragoonrung S, Vanavichit A. 2003. Thai jasmine rice carrying QTL ch9 (SubQTL) is submergence tolerant. Ann. Bot. 91:255-261. Singh RK, Dwivedi JL. 1996. Rice improvement for rainfed lowland ecosystem: breeding methods and practices in Eastern India. In: Fukai S et al, editors. Breeding strategies for rainfed lowland rice in drought-prone environments. Proc. Intl. Workshop, Ubon, Thailand, 5-8 Nov. 1996. ACIAR Proceedings 77:50-57. 


Singh S, Singh ON, Singh RK. 1998. A shuttle breeding approach to rice improvement for rainfed lowland ecosystem in eastern India. In: Sustainable agriculture for food, energy and industry. James & James (Science Publishers) Ltd., UK. p 105-115. Xie Y, Wu R. 1989. Rice alcohol dehydrogenase genes: anaerobic induction, organ specific expression and characterization of cDNA clones. Plant Mol. Biol. 13:5368.

Xu K, Mackill DJ. 1996. A major locus for submergence tolerance mapped on rice chromosome 9. Mol. Breed. 2:219-224. Xu K, Xia X, Fukao T, Canlas P, Maghirang-Rodriguez R, Heuer S, Ismail AI, Bailey-Serres J, Ronald PC, Mackill DJ. 2006. Sub1A is an ethylene response factor-like gene that confers submergence tolerance to rice. Nature 442:705-708. 


Participatory varietal selection of salinity-tolerant rice for the coastal wetlands of Bangladesh
M.A. Salam, G.B. Gregorio, D.L. Adorada, and R.D. Mendoza

Soil salinity is one of the major constraints to rice production on about 1 million ha of the coastal areas of Bangladesh. To reduce salinity, major engineering infrastructure and expensive soil amendments are needed. However, these amendments require large investments to tailor rice plants to adapt to salt stress. In the past, it was difficult to release high-yielding varieties in the coastal areas because no single rice variety could be adapted to the varied soil and climatic conditions of those areas. The urgent need for the development of saline-tolerant varieties in the coastal areas of Bangladesh was therefore prioritized as a researchable issue in a stakeholder analysis. The Bangladesh Rice Research Institute (BRRI) and International Rice Research Institute (IRRI) developed advanced lines that can tolerate salt stress of 8–10 dS m–1 and with a large range of plant characters. All these genotypes were grown on the saline-prone farms of BRRI research stations at Sonagazi and Satkhira. Three sets composed of 72 lines for the wet season (T. aman), 76 for the irrigated (boro) season, and 86 for uplands were selected for evaluation in “mother and baby” trials of participatory varietal selection (PVS). PVS trials were conducted at five sites in the saline-prone coastal districts. Table 1 shows the soil salinity status of the mother trial plots across the three crop seasons. All the trial plots had initial soil salinity above the critical level (>4.0 dS m–1) except for Noakhali in the T. aman season. The mean salinity level was highest in boro, followed by upland and T. aman plots. The mother trial for T. aman was conducted at four sites, Char-Sonapur (Sonagazi), CharLawrence (Laxmipur), Char-Jabbar (Noakhali), and Benerpota (Satkhira), in 2001, followed by

baby trials in 2002 and evaluation in farm-walk farmers’ fields at three villages in 2003. A mother trial of the upland ecosystem was carried out at the Sonagazi BRRI farm in 2002 and at Upakul Villa (Char-Lawrence, Laxmipur) in 2003. Similarly, a mother trial of the boro season was carried out at Benerpota (Satkhira) and Gobindopur (Kaliganj) sites in 2000-01 and also at Benerpota in 2001-02, followed by baby trials in 2001-02 and 2002-03 crop seasons. Farmwalks of farmers were organized for baby trials at seven villages in this season. Nonreplicated mother trials and researchermanaged on-farm experiments were done in cooperation with the participating farmers. PVS was launched at crop maturity. The Department of Agriculture and Extension (DAE) and five NGOs (Gurpukur, Uttaran, Sushilan, Proshika, and BRAC) organized 25–40 RPFs, including 0– 15 women for the mother trials, across the sites and crop seasons. A briefing was given on the importance of involving farmers in the progress of variety development through PVS. The participating farmers were divided into three to four groups (8–10 farmers per group), and a researcher led each group. All farmers were given a simple PVS evaluation sheet asking them to choose one or two genotypes to grow
Table 1. Soil salinity of mother trial plots at five sites across rice crop seasons. Site Sonagazi Laxmipur Noakhali Satkhira Kaligonj Mean Plot location Sonapur Char-Lawrence Char-Jabbar Benerpota Gobindopur 5.9 8.3 3.6 6.9 – 6. Soil EC in crop seasons T. aman Boro – – – 6.9 0.5 8.7 Upland 6. 0.0 – – – 8. 


as baby trials on their land under their own management. After a field visit, PVS sheets were collected and two or three farmers from each group were called immediately on a random basis to express their views and opinions on the selection made. All the PVS farmers were supplied with 500 g of seed of the chosen genotypes to be grown in their fields as baby trials. On the other hand, the farm-walk farmers selected seeds from baby-trial farmers’ fields. Baby trials and the farm-walk farmers’ plots were monitored using a household-level questionnaire (HLQ).

Varieties selected through PVS
PVS is the selection of fixed lines done by farmers in a target environment using their own selection criteria. In PVS, access to and selection of breeding lines are decentralized, with farmers’ participation. Plant materials, which have all the combinations of characteristics, are selected by the farmers for the appropriate plant type and yielding ability on their respective salt-stressed farms instead of researchers themselves selecting for the farmers in the coastal areas of Bangladesh. In three years, 245 salt-tolerant genotypes from IRRI and BRRI were placed in PVS mother trials at five coastal sites of Bangladesh. PVS in T. aman season In 2001, 87 farmers took part in PVS at four mothertrial sites. The farmers selected 10 of the 72 lines and these genotypes were designated as PVS-T1 to

PVS-T10 representing T. aman (Table 2). Variation was great in the choice of PVS-T genotypes by the farmers across locations. The advanced lines had site-specific preference. BRRI dhan40 (PVS-T9) and BRRI dhan41 (PVS-T10) were chosen by a majority of the farmers, followed by PVS-T7 at all the sites. The farmers’ preferences for choosing the genotypes were the following: lodging tolerance, high yield, freeness from disease, uniformity in maturity, and short growth duration. The choice of varieties by farmers also depended on the location. This confirms the location specificity of varieties for complex environments such as coastal areas. There is considerable variation in yield and other agronomic characteristics of the 10 PVS-T genotypes (Table 3). Most of the farmers’ preferred varieties had growth duration of 140 ± 5 days except for PVS-T4, with early growth duration, and they were chosen at Sonagazi and Satkhira for highlands and crop diversification. The PVS-T farmers grew their selected genotypes as baby trials in 2002, under their own management practices. Data from HLQs showed that 50% of the farmers were discouraged by the results because of water stagnation in the field, which ranged from 30 to 50 cm from transplanting to maximum tillering stage. The farm walks for baby trials in which 96 farmers partipated were conducted at three villages where there was no high water stagnation. Five PVS-T genotypes showed promising results and were selected by farmers (Table 4). Two sister lines (PVST5 and PVS-T7) were selected at Sonapur, whereas two sister lines (PVS-T1 and PVS-T2) and PVS-T4 were chosen at Satkhira and Kaliganj.

Table 2. Selected genotypes from mother trials and the number of PVS farmers across locations, T. aman, 2001-02. Entry PVS-T PVS-T PVS-T3 PVS-T4 PVS-T5 PVS-T6 PVS-T7 PVS-T8 PVS-T9 PVS-T0 Designation Sonagazi BR5778-56--3-HR BR5778-56--3-HR4 BR5778-56--3-HR5 R6640-B-4-- BR5999-8-3--HR BR5999-8-3--HR0 BR5999-8-3--HR6 BR5333-34-4-6 BRR dhan40 BRR dhan4 5 5 6 3 8 0 6 4 6 8 Number of farmers Noakhali  8 4 – 3 3 3 – 9 8 Laxmipur – – –  4 3 8 – 9 8 Satkhira 6 8 6 9 3 5 8 – 9  Total 3 5 36 3 38 3 55 4 63 86 


Table 3. Agronomic performance of PVS genotypes in mother trials across locations, T. aman, 2001-02. Entry PVS-T PVS-T PVS-T3 PVS-T4 PVS-T5 PVS-T6 PVS-T7 PVS-T8 PVS-T9 PVS-T0 Plant height (cm) 6 0 4 06 30 30 30  36 38 Yield (t ha–) Sonagazi .4 . . .0 3.6 3.5 . .9 .7 .8 Noakhali 5.9 4.7 5.4 .9 6.4 5.4 4.4 5.4 4.9 4.4 Laxmipur . . .3 .5 5. 4.5 3.8 4. 4.5 4. Satkhira 3.8 3.7 4.3 3.0 4.7 4.7 3.8 – 3.8 3.7 Mean 3.3 3.0 3.3 .3 5.0 4.5 3.5 4. 3.7 3.7 Duration (days) 35 35 36 6 35 36 35 46 38 44

Table 4. Performance of PVS-T genotypes in baby trials and number of farm-walk farmers that selected each genotype, T. aman, 2002-03. Genotype Plant height (cm) 3 33 9 40 39 Growth duration (days) 6 6  4 40 Number of preferring farmers 7 8 34 9 7 Yield (t ha–) La 4.5 5.0 4.5 – – L 5.0 5. 3.4 – – L3 – – – 5.0 5.0 X 4.8 5. 4.0 5.0 5.0


L1 = Parkumira, Tala; L2 = Gobindapur, Kaliganj; L3 = Sonapur, Feni.

All the farm-walk farmers grew their individual selection under their own management practices in T. aman 2003. PVS-T4 showed hightemperature sensitivity at the reproductive phase and consequently heterogeneity of heading and sterile panicles were observed. PVS-T4 is a shortduration genotype; thus, its flowering coincided with the high-temperature spell (>35 °C) during early to mid-October. On the other hand, PVS-T1, PVS-T2, PVS-T5, and PVS-T7 were relatively tall in stature and prone to lodging at wind speed of 30–50 km h–1, a usual phenomenon in the coastal areas in early November. Thus, PVS in T. aman was terminated in the subsequent season. PVS in saline-prone upland culture In 2002, a mother trial for 85 short-duration (120–130 d) IR genotypes was conducted at a BRRI farm in Sonagazi. The prevailing crop growing conditions in upland culture in coastal wetlands are the following:

1. Dry direct seeding using the dibbling method, 2. A drought spell plus salinity (6–8 dS m–1) stress from seedling to tillering stage, 3. Water stagnation of 20–50 cm at tillering through maturity stage depending on elevation, and 4. Harvesting of the upland rice crop at or before mid-August for the transplanting of the main rice crop (T. aman). Considering the above-mentioned diversified and difficult conditions, a mother trial was conducted at the Sonagazi station. Crop establishment was done using dry direct seeding with the dibbling method. The crop received drought after seeding and also stagnation of water at 20–30 cm at maximum tillering up to maturity. PVS activity was conducted in collaboration with DAE and two NGOs (Proshika and BRAC) of Sonagazi at crop maturity in August 2002. Twenty-nine farmers, including two women, joined the PVS activity and 12 salt-tolerant genotypes with good yielding ability were selected. The performance of the farmers’ selected varieties demonstrated higher yield than BRRI dhan27 and varieties with shorter plant height. Farmers finally selected four PVS-A genotypes out of the 85 short-duration genotypes introduced. Seeds of the selected PVS-A genotypes were distributed to farmers for adaptability tests in their own fields in aus 2004. 


A mother trial composed of 22 genotypes, including 12 PVS-A, was conducted at Upakul Villa under Char Alexgender of Laxmipur in 2003. Proshika, BRAC, and DAE organized 31 farmers, including five women, to join the PVS activity. The male farmers visited the trial and selected four PVS-A genotypes (Table 5). Seeds (500 g) of the PVS genotypes were supplied to the PVS farmers’ baby trials to be conducted on their own farms for the 2004 crop season. PVS in boro season Researcher-managed on-farm mother trials composed of 76 salt-tolerant (8–10 dS m–1) genotypes were conducted at Satkhira and Kaligonj sites in 2000-01. Thirty-seven farmers took part in the PVS activity at crop maturity in cooperation with the Department of Agricultural Extension (DAE) and three NGOs, Gurpukur, Uttaran, and Sushilan. Government officials, block supervisors of DAE, field-level workers of the NGOs, and local leaders (union council members and the chairman) participated. The 24 genotypes selected by participating farmers were designated as PVS-B1 to PVS-B24, representing the boro season. Since handling such a large number of genotypes was difficult, we distributed seeds of PVS-B1 to PVS-B8 and decided to conduct another PVS with the remaining PVS-B genotypes in the 2001-02 crop seasons. In 2001-02, at Benerpota, 42 farmers, including 15 women from Kaliganj, Tala, and Satkhira, participated in the PVS activity. The PVS farmers selected 12 genotypes as shown in Table 6. The performance of the PVS-B genotypes was highly accepted by farmers, especially for yielding ability and shorter growth duration. The farmers’ criteria for the selection of PVS genotypes were short growth duration and yield-contributing characters (high number of panicles per plant, high number of grains per panicle, and low sterility percentage). The fields with initial soil salinity of 8–10 dS m–1 had 2–3 dS m–1 at transplanting. Salnity gradually increased and exceeded the critical level (4.0 dS m–1) from late April. Thus, farmers preferred short-duration genotypes for saline conditions because they noted that, with longer

Table 5. Performance of selected genotypes in mother trials and number of farmers who chose PVS genotypes at Char Alexgender, Laxmipur, upland rice, 2002-03. Entry Designation Plant height (cm) 85 03 9 3 Growth duration (days) 7 0 04 00 Yield (t ha–) 3.0 3. .7 . No. of farmers 7  6 0


R7046-B-R-6-3- R63307-4B-4-3 R6449-3B-4-3 BRR dhan7

Table 6. Performance of genotypes selected from mother trials and number of participating farmers at Benerpota, Satkhira, boro season, 2001-02. PVS no. Designation Number of farmers 4 7  3 5 3 3 7 5 5 7 0 Plant height (cm) 99  94 78 90 93 93 04 97 93 96 96 Yield (t ha–) 6.7 6. 5.6 6.3 4.9 6.7 6.7 5.7 7.0 6.4 6. 6.3


BR5778-56--3-HR R63307-4B-4-3 R6449-3B-4-3 R659-4B-4- AT309--GAZ R006-9-- BR5777---4--HR R6375-B---3-3- R7046-B-R--3- R7046-B-R-6-- R6640-B-4-- R60483-B-7---

growth duration, the crops were damaged by high soil and water salinity at the later growth phases, which are the reproductive stage. The PVS farmers were supplied with 500 g of seeds of their selected PVS-B genotypes for growing as baby trials under their own management in the 2002-03 boro crop season. Farm walks were conducted at seven villages and 211 farmer participants identified five genotypes for diffusion for placement in the variety release system (Table 7). Ten PVS farmers were involved in growing the five PVS-boro genotypes together with a check variety to obtain comparative data to serve as the basis for nomination for varietal release in the coastal wetlands of Bangladesh. Seed samples of these five PVS-B genotypes were submitted to the Seed Certification Agency for completing a District, Uniformity, and

Table 7. Performance of PVS-B genotypes in baby trials and the number of farm-walk farmers who selected each genotype, boro season, 2002-03. Genotype Plant height (cm) 96 9 97 93 99 Growth duration (days) 49 53 54 45 5 Number of farmers Yield (t ha–) L – 6.0 6. – 6.5

L – 6. 6.4 – 6.

L3 – 6.0 – – 6.8

L4 – 6.0 6.5 – 6.8

L5 – – – 6.0 –

L6 6.0 – – – –

L7 – – 6.4 – –

X 6.00 6.05 6.38 6.00 6.48


5 36 57 7 08

L = Satkhira Sadar, L = Ashashuni, L3 = Kaliganj, L4 = Sakdah, L5 = Uttaran, L6 = Benerpota, L7 = Kashipur.

Stability (DUSt) evaluation as a requirement for national varietal release. With PVS in resource-poor farmers’ fields and involvement of relevant NGOs, local government, and extension scientists, varieties were evaluated, selected, and distributed to farmers in the coastal wetlands of Bangladesh. Five genotypes each for T. aman and boro seasons identified by farmers through PVS

showed good potential for diffusion, and are now included in the national varietal release system of Bangladesh. After six seasons of PVS, at least 500 resource-poor farmers are now using farmers’ selected varieties for the boro season. NGOs such as Gorpukor at Sathkira are conducting their own PVS in a techno-demo using the five PVS-B materials in three villages involving 53 farmers.


Increasing profits from rice production in Bangladesh: direct wet seeding of rice using a plastic drum seeder
M. Zainul Abedin

In Bangladesh, there is a need to increase income from rice production for a sustainable livelihood of rice farmers, particularly for those who derive their main income from rice production. Transplanting is an age-old practice alongside direct seeding. The introduction of high-yielding varieties (HYVs) under irrigation has increased the area under transplanted rice, which requires lots of labor. Because of economic development in other sectors, there is a labor shortage in agriculture, which often delays transplanting, thus causing a reduction in yield. Consequently, the cost of labor is increasing. The cost of other inputs is also increasing, but the rice price is not increasing at the same rate, thus reducing the net benefits from rice production. Direct seeding was found to reduce costs, increase net benefits, and offer other advantages. The use of a plastic drum seeder was found to be more acceptable than broadcasting. The drum seeder was originally developed by IRRI but has been improved by researchers and entrepreneurs in Vietnam. Farmers in Vietnam use a drum seeder on a large scale, basically to reduce seed rates. The plastic 12–16-row drum seeder has the following characteristics: • Light weight, 6 to 7 kg, and requires low power to pull. • Very cheap at US$25 (investment cost per ha is about $0.60). • With 6 drums, each pass can cover 12 rows, with 2.4 m width; can accommodate 8 drums to cover 3.2 m width. • Each drum can hold about 2 kg of seeds. • Options exists for variable seed rates.

• One person can seed about 1 ha in a day at about 2 km per hour. To use pregerminated seeds, soak them for 12–24 hours, incubate them for 48–72 hours, and air-dry the pregerminated seeds just before seeding. For land preparation, use dry tillage, puddle and level the land, and drain out excess water. Water management is critical during the first week. Do not irrigate during the first 3–4 days, just to keep soil moist. Start irrigating after about 3–4 days to keep the soil moist. After about 10 days, increase the depth of water as the seedlings grow. Subsequently, follow normal water management as in transplanted rice. The following trials were conducted in Bangladesh: Wet season (aman), 2003: Dry season (boro), 2003-04: At BRRI HQ and with 5 farmers At 7 BRRI stations At 1 BSMRAU, Gazipur At 56 on-farm locations At 4 BRRI stations At > 80 onfarm research locations Reach about 10,000 farmers

Wet season, 2004:

Goal for dry season, 2004-05:

The growth vigor of boro rice plants was affected by establishment method (Table 1). The number of tillers at harvest in boro rice was influenced by stand establishment method

(Table 2). Direct-wet-seeded boro rice was established using a plastic drum seeder in 2004 (Table 3). Variety BRRI dhan29 was also established using a plastic drum seeder in 2004 (Table 4). Yield of boro rice was affected by method of stand establishment (Fig. 1). Direct-wet-seeded T. aman rice was evaluated in 2003 using a drum seeder (Table 5). Direct-wet-seeded boro rice was also evaluated with a drum seeder in 2004 (Table 6). The effect of date of seeding on rice yield was evaluated in 2004 (Fig. 2).
Table 1. Growth vigor of boro (dry-season) rice plants as affected by method of establishment, Gazipur, Bangladesh, 2004. Treatment Seedling dry weight at time of transplanting (mg seedling–) 0.4 75.3 Dry weight at 40 DAT (g plant–) 70.8 68.7 Dry weight at harvest (g plant–) ,08.5 898.5

Early harvest of rice in Bangladesh may have the following advantages: • May save the T. aman crop from drought and the boro crop from early flood. • May help early establishment of winter crops after T. aman. • May allow growing of another shortduration crop in the cropping pattern. • May help early establishment of deepwater rice after boro rice. Farmers made several observations regarding the early harvest: • Better and sturdier plant growth leads to - Reduced insect infestation - Need for less pesticide application • Lodging in some locations—was it due to variety BRRI dhan28? • Ability to escape flood and/or tolerate flooding Direct-wet-seeded rice was evaluated economically in 2003 (Table 7). Direct wet seeding can have several constraints. Weeds can be controlled by mechanical weeding, water management, and manual weeding. Two weedings are usually enough. Herbicides are the least preferred control option.

Direct wet-seeded rice Transplanted rice

Table 2. Number of tillers at harvest in boro rice as influenced by stand establishment method, Bangladesh, 2004. Method Effective Direct dry-seeded rice Transplanted rice 389 3 No. of tillers m– neffective 89 68 Total 678 8

Table 3. Performance of direct-wet-seeded boro rice (BR4828-54-4-1-4-9) established with a plastic drum seeder, Barisal, 2004. Treatment Direct drum seeding in single thin row Direct drum seeding in single thick row Direct drum seeding in both thin and thick row Direct hand seeding Transplanting Level of significance % CV ± SE Root length (cm) 3.3 Tillers m– 865 Panicles m– 498 Grains panicle– (no.) 7. Sterility (%) 5.6 Straw yield (t ha–) 8.69 Grain yield (t ha–) 7.04 















.7 .0 ns 7.45 0.993

959 45 *** . 57.4003

474 75 ** 9.9 5.0474

64.9 7.7 ns 8.53 3.446 

3.5 3.0 ns .6 .9806

9.4 6.5 ** 9.44 0.4703

6.99 5.57 *** 3.4 0.8


Table 4. Performance of direct-wet-seeded boro rice established with a plastic drum seeder, Barisal, 2004. Treatment Root length (cm) 0.57 Tillers m– 758 Panicles m– 46 Grains panicle– (no.) 0 Sterility (%) 3. Straw yield (t ha–) 7.4 Grain yield (t ha–) 8.3

Direct drum seeding in single thin row Direct drum seeding in single thick row Direct drum seeding in both thin and thick row Direct hand seeding Transplanting Level of significancea % CV ± SE














.87 . ns

743 33 *** .04 44.49

48 87 *** 6.9 4.589 

0 ns 5.56 6.989 

0.3 5.8 ns 5.53 .74

6.5 6.4 ** 4.04 0.57

7.89 7.05 * 5. 0.34

9.8 .96

ns = nonsignificant.

Table 5. Performance of direct-wet-seeded T. aman rice using a drum seeder in Bangladesh, 2003 (average of 6 locations). Variety BR 60-0-- Method Direct-wet-seeded rice Transplanted rice Yield (t ha–) 4.8 (+18%) 4. Duration (days)  3

Yield (t ha-) 7 6.4 6 5 4 3 Both rows Thick row   0 Transplanting 8 Dec Broadcasting 5.76 5.84 5.57 5.57

Variety BRR dhan8 BRR dhan9 BRR dhan36


Yield (t ha–) 6.0 (20%) 5.0 7.1 (18%) 6.0 6.5 (19%) 5.5

Duration (d) 30 4 5 6 9 4

Fig. 1. Yield of dry (boro) rice as affected by method of stand establishment, Rajshahi, Bangladesh, 2004.

DWSR = direct-wet-seeded rice, TPR = transplanted rice.

Grain yield (t ha-) 8 6

Table 7. Economic analysis of direct-wet-seeded aman rice, 2003, Bangladesh (6 locations). ndicators Cost of production Gross return Gross margin Benefit-cost ratio Direct-wetseeded rice 9,79 3,803 3,084 .66 Transplanted rice 0,955 7,086 6,3 .9 +/– % –5.9 . 3.4

4  0

BRR dhan8 BRR dhan36 BRR dhan9

8 Dec

Thin row

Table 6. Performance of direct-wet-seeded boro rice using a drum seeder in Bangladesh, 2004 (average of 56 on-farm research locations).

Establishment method 

8 Dec Date of seeding

Fig. 2. Effect of date of seeding on yield due to direct wet seeding by a drum seeder, boro, 2004.


In some places, labor displacement could be an issue. As part of our social responsibility, a critical ex ante analysis is needed for each situation/state or country. Experience suggests that economic development in rural areas over time may absorb the labor displaced. In June 2004, a farmer participatory workshop was held at BRRI headquarters. Farmers identified the strengths and weaknesses of direct wet seeding. Recommendations were made for suitable areas, cropping systems, and land and soil types for use of a drum seeder for direct seeding.

• In many circumstances, the improved, cheap ($25), lightweight (6–7 kg) plastic drum seeder will enable farmers to deal with labor shortages and the short period of time available for planting, and thus may enable them to plant on time. • Apart from economic and yield advantages, earlier harvest may allow further crop intensification and save crops from drought or early floods, thus enhancing food security.

• A reduction in pesticide use is expected to have a positive effect on the environment and on households and the national economy. • A community participatory approach is essential for the technology’s success due to its uniqueness. • A farming systems perspective is needed to identify where the technology will bring the desired benefits. • Several research needs exists: - Weed and water management - N management - Adjustment with cropping systems for improving system productivity - Varieties - Lodging behavior - Insect and disease interaction - Can this technology be successful in favorable upland conditions? - A socioeconomic analysis of labor displacement


Characterizing and understanding the socioeconomic conditions of farming households in rainfed rice environments: a case in eastern Uttar Pradesh
T. Paris, A.D. Cueno, and A. Singh

Rice (Oryza sativa L.) is the staple food for more than three billion people, more than half the world’s population. It provides 27% of dietary energy and 20% of dietary protein in the developing world. Rice is cultivated in at least 114 mostly developing countries and is the primary source of income and employment for more than 100 million households in Asia and Africa (FAO 2004). Of the 840 million people suffering from chronic hunger, more than 50% live in areas dependent on rice production. About 80% of the world’s rice is produced on small farms, primarily to meet family needs, and poor rural farmers account for 80% of all rice producers (FAO 2004). In Asia and subSaharan Africa, almost all rice is grown on small farms of 0.5 to 3 ha. Fully two-thirds of the world’s poor live in Asia and almost all of these eat rice. Rice is also becoming a staple food in sub-Saharan Africa, where urban dwellers who only a few decades ago rarely ate rice now consume it daily. Per capita consumption has doubled since 1970 to 27 kg. In the mega-cities of Asia, the poorest of the poor may spend up to 50% of their total income on rice, not other food, but rice! So, anything that lowers the price of rice will directly benefit hundreds of millions of poor consumers and anything that increases rice-farming productivity will benefit millions of rice farmers and their families (Zeigler 2006). Because rice is critical for food security in so many of the poorest countries, investments in the rice sector should be designed to alleviate poverty and meet the food demands of still-growing—and increasingly urbanized— populations. To meet the dual challenge of producing enough food and alleviating poverty, more rice needs to be produced at a low unit

cost so that producers can be ensured of reasonable profits, poor consumers can have the benefit of low prices, and the environment and ecosystem services can be safeguarded (Bouman et al 2007). In a symposium, Zeigler (2006) mentioned that, if we are going to focus on poverty, we have to invest relatively more resources and establish a greater presence in rainfed areas. Rainfed rice accounts for 50% of the rice-growing area. If productivity can be improved in these areas, it will not only improve people’s lives but will also contribute to food security. Rainfed means that the crop depends exclusively on rainfall, and in some cases unpredictable floods, for water. Because rainfall can be so variable, rice in rainfed areas typically is prone to drought and catastrophic flooding, sometimes in the same year, and, since the environments are so difficult and yields so unreliable, farmers rarely apply fertilizer and tend to not grow improved varieties. Thus, yields are very low (1–2 t ha–1). Aside from drought and flooding, the productive capacity of rice environments is also being threatened by salinity and climate change. Although these conditions cause generally low productivity, about one billion people depend on rainfed lowland rice in South and Southeast Asia. Because of the low and unstable productivity, poverty is severe in communities largely dependent on rainfed rice (e.g., most of eastern India). In drought years, food consumption decreases, indebtedness increases, assets are sold, and household members migrate. Therefore, droughts can have long-term destabilizing effects on poor communities and reduce productivity even in nondrought years because farmers, fearing

crop loss, avoid investing (Pandey et al 2001). Although the importance of rainfed areas and the growing dependency of farm families are evident, research efforts in this area thus far have been limited and less successful compared with irrigated ecosystems that have been the focus of rice research because of their leading role in rice production. However, advances in genomics and molecular biology of rice, enabled by the sequencing of its genome (the first of the crop species), and improved analytical approaches have allowed rice scientists— breeders, geneticists, and physiologists—to make dramatic progress in developing rice lines that tolerate complete submergence, drought, and salinity (Zeigler 2006). In this paper, the farms and farm households under various types of stress environments were characterized to develop a systematic understanding of the livelihood of poor rural households in drought-prone, submergence-prone, and salt-affected areas of eastern Uttar Pradesh, India. This serves as a basis for exploring opportunities to improve land and water productivity of rice-based crop production systems through innovative interventions that integrate genetic improvement and management strategies, which are environmentally sustainable and socially acceptable to resource-poor farmers. This paper also investigates the input-output (i.e., land, labor, and capital) markets in rice production and their effects on income distribution. It also identifies the constraints in crop production under different types of ecosystems and presents plausible opportunities to improve farmers’ livelihood.


The International Rice Research Institute (IRRI), in cooperation with our collaborators in Narendra Deva University of Agriculture and Technology (NDUAT), began conducting benchmark surveys in 2004-05 on droughtprone, submergence-prone, and salt-affected areas in India. The principal investigatorscollaborators facilitated the work, especially in the selection of the project sites, sample respondents, and, consequently, in the

management of the household surveys and data collection. The study covered five villages in Faizabad, Sultanpur, and Siddharthnagar districts located in the eastern part of Uttar Pradesh, India, popularly known as “Purvanchal” (Fig. 1). Rainfed agriculture is still a major activity in the region. Rice is the dominant crop and accounts for 60% of the total rice area of the state. The agroecological conditions are diverse. Crops are affected by frequent flood, drought, salinity, and sodicity, which result in low productivity. Two villages, Bhola Pandey ka Purwa and Sinhani in Faizabad and Sultanpur districts, respectively, were selected to represent the area affected by salinity/sodicity. A participatory rural appraisal (PRA) was first conducted in these areas to identify the villages appropriate for the purpose of the study of the Challenge Program on Water and Food (CPWF) project. The villages were selected based on (1) the vastness of area affected by salinity/sodicity, (2) physiographical conditions, and (3) rice is the major crop grown in the area. On the other hand, three villages with rainfed lowland environments were purposively selected in Faizabad and Siddharthnagar to carry out activities of a Consortium for Unfavorable Rice Environments (CURE) project. The villages Zivpur in Faizabad and Karaideeh and Pipehrawa in Siddharthnagar were selected representing two different major rice-growing environments. Zivpur characterizes the drought environment, whereas the latter villages represent submergence-prone areas due to flash flooding during the wet season. These villages are the same sites used in on-farm trials for submergence-tolerant varieties under CURE BMZ and PVS trials on component technologies. This includes the rice varietal trials that were conducted by NDUAT and IRRI under the ICAR-IRRI Collaborative Rainfed Rice Program for validation of technologies in farmers’ fields. A total of 125 household respondents were interviewed in the five villages classified into three different ecosystems. Twenty-five samples were randomly chosen to represent droughtprone areas and 50 samples each to characterize

Fig. 1. Map of selected sites in Uttar Pradesh, India, 2005.

submergence-prone and salt-affected areas. Primary household data collection was carried out with the use of a well-structured pretested field schedule. The information collected was socio-demographic, economic, and biophysical characteristics of the farming households; farm characteristics; and household economic conditions. Other qualitative and quantitative information was also gathered through focus group discussions with key informants. Descriptive statistical tools were essentially used to analyze the survey data. Costs and returns analysis was carried out to determine the proportion of production inputs to total costs, as well as to derive the income received by farmers in rice production in various stress environments.

Other factors that significantly constrained rice production were determined and problems associated with these constraints were then prioritized to identify alternative management options that can alleviate the adverse effects.

Results and discussion

Household characteristics The majority of respondents in all types of environment belong to the lower caste with an average household size of 7 in both droughtprone and sodic environments, and 8 family members in submergence-prone areas (Table 1). The number of unpaid farming laborers can be determined by household size. The bigger the household size, the more unpaid family labor is available to do farming activities, thus lowering the production cost. Male farm operators, as

Table 1. Socio-demographic characteristics of sample households by type of environment. ndicator Total number of sample households Total number in households Proportion of households by caste Upper Backward Scheduled Muslim Total Average household size by caste Upper Backward Scheduled Muslim All castes (av) Average age Male operator Wife Average years in school Male operator Wife Literacy rate (in %) Adults Male Female Children Male Female 4 3 5 4 6 6  9  9 9 4 5  57 4  60 5  65 45 40 45 4 48 4 6 9 4 8 7 – 7 – 8 8 8 8 6 – 7 44 40 8 8 00 – 50 – 50 00 0 54 6 – 00 Drought 5 75 Submergence 50 375 Sodic conditions 50 367

Source: Baseline survey on rural households, CURE projects, 004-05.

well as their wives, are by and large in their forties. The wives, however, are younger than their husbands by 3 to 6 years. The average years in school of male operators is 5 years, which is 3 years higher than the educational level of the wives. A higher literacy rate can be observed among males than among females, especially in the salt-affected environment. In theory, education is expected to improve productivity in all spheres of activities, including agriculture. Education will increase farmers’ accessibility to useful information,

which may shift their production possibility curve upward. Educated farmers are more likely to adopt modern farm inputs and prefer risky production technologies. However, in some studies, a negative impact of education on productivity arises, probably because of the cross-country variation in the nature of the technology underlying agricultural production. Among the young population (5–15 years old), the literacy rate was also high, suggesting that farmers, irrespective of their socioeconomic status, are now more aware of the advantages of sending their children to school. A high literacy rate is also due to awareness of many programs launched by the government to educate the children. The government provides subsidies such as a mid-day meal, free education at the primary level, and free education for girls up to the college level. A majority of the respondents are male heads of the household. In most cases when the husbands have to leave the village to find seasonal employment during the slack periods on the farm, the wife may act as the de facto head of the family in both the household and on the farm. Agriculture is the main source of income of farming families in the drought-prone (45%) and salt-affected areas (31%). In submergence-prone areas, household members are mostly migrants and are engaged in nonfarm activities (33%) such as tailor, salesman, driver, and a skilled worker in the private sector. A significant proportion (22%), though, worked in agriculture-related activities either in crop and animal farming or as an agricultural laborer. The average of 43% nonworking population in all types of environment consists of students, family members who belong to the working population but do not work at all, and housewives who also played an active role in rice farming but classified themselves as housewives (Table 2). They do most of the farming activities after land preparation, while the husbands migrate and engage in nonfarm activities for better income. For well-off families, female members are confined in their home and just do household activities. Others also do livestock raising.

Biophysical and farm characteristics Socioeconomic groups by farm size. In all environments, an average of 71% of the farm families were very poor with less than 1 ha of landholding (marginal category), 17% had 1 to 2 ha of cultivated land (small category), whereas the rest (12%) had more than 2 ha (medium and large category) (Fig. 2). A majority of the
Table 2. Proportion of adult household members by occupation and by type of environment. Occupation Agriculture Nonagriculture Student No occupation Total Drought 45 5 8 3 00 Submergence  33  44 00 Sodic conditions 3 6 7 36 00

farmers with very small landholdings can be observed in the drought-prone (84%) and saltaffected (70%) areas, with an average farm size of 0.84 ha and 1.0 ha, respectively (Fig. 2). In the submergence-prone areas, average landholding is 1.4 ha, which shows that the farmers are also operating in small-scale rice farming. This indicates that, in all types of environment, it is difficult for farmers to ensure household food (rice) security only from their own production. In addition, rice is grown only during the kharif season and on limited lands. With these conditions, marginal and small farmers either have to increase rice productivity through improved technologies or seek nonfarm employment and income for family survival. However, according to Ramasamy and Selvaraj

Source: Baseline Survey on Rural Households, CURE and CPWF Projects, 004-05.

Medium and large (>2 ha) 8%

Marginal (1 ha and below) 84%

Small (1.01to 2 ha) 8% DROUGHT 0.84 ha

Medium and large (>2 ha) %

Marginal (1 ha and below) 58%

Small (1.01 to 2 ha) 30% Medium and large (>2 ha) 6% Marginal (1 ha and below) 70%


Small (1.01 to 2 ha) 4% SODC Fig. 2. Proportion of sample households by size of landholding and by type of environment.


(2006), even though technologies are available to restore the productivity of degraded lands, adoption of technologies by farmers is poor. Hence, the development and commercialization of technologies for drought and salt tolerance for water-limiting and saline and alkaline soils hold great promise for the prosperity of agriculture, particularly for rice production. Cropping pattern. The crop year is divided into two growing seasons: the kharif or monsoon season (June-September) and the rabi or winter season (October-May). In the lowlands, rice-wheat is the predominant cropping pattern (Fig. 3). Rice is planted sometime in June-July and harvested in November, depending on the growth duration of the varieties used. Farmers who used short- and medium-duration varieties (normally 110–135 days) harvested their produce in September, especially those in the midlands and uplands. After harvesting rice, farmers broadcast wheat immediately to benefit from the moisture left in the soil. In most cases, farmers also broadcast mustard seeds after sowing wheat. Wheat is harvested from the last week of March until mid-April, whereas mustard is usually harvested in early March. If rice is harvested late, wheat is also sown late. In this case, wheat is no longer mixed with mustard. Rice-potato, rice-pea, rice-linseed, and ricelentil are other cropping sequences. All these rabi crops are planted in October after rice except for lentil, which is grown in November. Potato is harvested in January, whereas the rest are harvested in March. Basically, these crops are grown for home consumption. Farmers who raise livestock and have access to supplementary irrigation also grow berseem (a fodder crop) during this period to supplement animal feed requirements. In the uplands, some farmers grow sugarcane and pigeon pea throughout the year. We can thus see how farmers maximize the use of their limited resources to augment their income as well as meet their daily needs. Rice is the main crop grown during the wet season in all types of environment, particularly in submergence-prone areas, covering 93% of the total cultivated land, whereas the other two stress environments cover a little less than

60% (Table 3). Other crops such as sugarcane, pigeon pea, pulses, vegetables, and fodder crops are also grown in upland areas during this season. An average of 17% of the total cultivable land in all types of environment is kept fallow because farmers do not have a suitable variety or management option for this type of land, notably farms in the salt-affected areas with 32% barren land. This is because of the distinct characteristic of the soil that hardly any single crop could grow under high sodicity and the absence of effective crop and water management options to mitigate the adverse effect of this type of soil. In the dry season, there is more diversity in the crops grown after rice (Table 4). Wheat is largely cultivated during this season, either alone or mixed with other rabi crops. Land type. Rice in the drought-prone and sodic areas is mostly grown in uplands, whereas, in submergence-prone areas, rice is grown in both uplands (48%) and lowlands (52%) (Table 5). In India, out of the 7 million hectares of upland rice, 6 million hectares are concentrated in eastern India, comprising eastern Uttar Pradesh (the study area), Chhattisgarh, Bihar, Orissa, West Bengal, and Assam. Upland areas of these states together constitute about 13.5% of the total area under rice in the country. The productivity of upland rice is very poor. In comparison with the present national average productivity of about 1.9 t ha–1, the average yield of rice in upland areas in the country is only 0.90 t ha–1. This is because rice cultivation under upland conditions faces different degrees of moisture stress, which affects plant growth and brings about less tillering capacity, sterility, delayed flowering, and lower harvest ratio (grain-straw ratio). These factors or a combination of these contribute to lower grain yield in upland rice areas. Aside from soil moisture stress, most studies suggest that the low productivity of rice in upland areas is due to drought, lack of resistance to/tolerance of diseases and pests, soil acidity/toxicity, inadequate plant population, and low nutrient status of soils. Lowland rice area, on the other hand, is about 14.4 million ha, which accounts for 32.4% of the total area under

Fig. 3. Cropping pattern of selected villages.


Table 3. Crops grown during the kharif season by type of environment. Crop Drought Total area (ha) Rice Sugarcane Pigeon pea Maize Curbi Pulses Spices Vegetables Urd Fallow Total . 0.5 – – – 0.9 0.0 5. – .7 .3 % of total area 56.9 .3 – – – 4. 0. 3.8 – .7 00.0 Submergence Total area (ha) 45.5 – – – – 0.3 – – 0.3 .7 48.8 % of total area 93.3 – – – – 0.5 – – 0.6 5.5 00.0 Sodic conditions Total area (ha) 3.7 .7 .4 .0 .0 – – – 0. 8.6 58.6 % of total area 55.8 .9 4. .7 3.4 – – – 0.3 3.7 00.0

Source: Baseline survey on rural households, CURE and CPWF projects, 004-05.

Table 4. Crops grown during the rabi season by type of environment. Crop Drought Total area (ha) Wheat Sugarcane Pulses Oil seed Potato Onion Burseem Spices Vegetables Mixed rabi crops Pea Lentil Linseed Fallow Total 0. 0.5 0.9 – .0 0.4 – 0.0 3.5 0. 0.8 – – 4. .5 % of total area 47. .3 4. – 4.4 .9 – 0. 6. 0.5 3.9 – – 9.6 00.0 Submergence Total area (ha) 6.9 – 0.4 – . 0.3 – 0.3 0. 8.6 .8 5.6 – 3.3 48.5 % of total area 55.6 – 0.8 – .5 0.6 – 0.5 0. 7.9 3.6 .5 – 6.8 00.0 Sodic conditions Total area (ha) 7.5 .3 – 0.5 .9 – 0.3 – – 5.4 4.6 3. 0.3 .5 58.5 % of total area 47. . – 0.9 5.0 – 0.6 – – 9. 7.9 5.4 0.5 .4 00.0

Source: Baseline survey on rural households, CURE and CPWF projects, 004-05.

rice in the country. The average productivity of rice in lowland areas ranges from 1.0 to 1.2 t ha–1 vis-à-vis national average productivity of 1.9 t ha–1. Still, productivity is low because rice farms in lowland areas are subjected to flooding during the southwest monsoon period. The intensity of floods differs from year to year due to variation in factors such as nature and frequency of flooding, water depth, turbidity, and silt and vegetation from place to place (Diwakar 2006).

Soil type. The soil in the major rice-growing belt was sandy-loam type, particularly in drought-prone and sodic areas (also in Table 5). In this type of soil, plant nutrient applied through fertilizer is lost rapidly and investment in fertilizers becomes risky, especially when crops are grown in high-elevation rainfed areas. Rainwater in this area is lost rapidly through deep percolation because of the upland location and loose texture of the soil. This causes low water retention capacity by the soil due to

high permeability that brings in moisturestress conditions quickly after the cessation of rain (Diwakar 2006). In submergence-prone areas, soil type was mostly clay/black soil (61%), as well as loamy (18%). The clay type of soil is a deep bluish black soil with high moisture retention capacity. It is well suited for rabi crops, particularly wheat. However, submergence is predominant, especially in low-lying areas. The loamy type, on the other hand, is intermediate in soil moisture retention between clayey and sandy-loamy. This is more or less an all-purpose soil, which is suitable for rice production. Type of irrigation. Irrigation has proved to be the most effective drought-proofing mechanism and single biggest factor in bringing about a large measure of stability in agricultural production (Indian Ministry of Water Resources 2007). The water requirement of the rice crop is higher than that of any other crop. An assured and timely supply of irrigation water has a significant influence on crop yield. Obviously, drought-prone and salt-affected areas require a considerable amount of water for their crops to survive; thus, access to supplementary irrigation is of paramount importance (see also Table 5). Sources of irrigation are diesel pumps, electric motors, and government-owned electric tubewells. Many households own diesel pumps and electric tubewells, but the majority of farmers rent them, which is very expensive. Diesel pumps are the most reliable source of irrigation and they cost Rs 60 per hour. In times of oil price hikes, farmers rarely irrigate their crops and depend on rainwater for irrigation.

Table 5. Biophysical characteristics of farms by type of environment. tem Soil type Loamy Sandy-loam Clay/black soil Clay-loam Sandy Saline-alkaline Total Land type Upland Lowland Total Type of irrigation Own diesel pump Rent diesel pump Government tubewell No irrigation Total 67 33 00 4 0 7  7 00 – 78 00 48 5 00 9 3 75 5 00 34 8 9 –  00 7 59 6 – 8 – 00 8 9 6  – – 00 9 5 3  – 7 00 Drought Submergence Sodic conditions

management and farming practices adopted by farmer-respondents in rice cultivation and the technologies adopted to improve their productivity.

Crop production management and farming practices
Rice varieties. Farmers in drought-prone and sodic areas basically used improved varieties in rice farming (i.e., Sarjoo 52, Sambha Mahsuri, Swarna, Pant 10, Pant 12, Mahsuri, Usar Dhan3, Basmati, NDR 97, NDR 359, and Hybrid). In sodic areas, Sarjoo 52 and Pant 10 are most commonly used for the following reasons: medium duration, good straw, medium plant height, bold grains, responsiveness to fertilizer, and good yield. Aside from these two varieties, Sambha, Swarna, NDR 359, and Mahsuri are also high-yielding; thus, many farmers also plant these types in both environments, especially in low-lying areas (Table 6A). Usar Dhan-3, grown by a few marginal and small farmers, is also observed to be suitable for sodic lands. It yields around 2.63–3.06 t ha–1

Current farming practices adopted by farmers in rice cultivation and technologies being


To achieve sustainability in agriculture, effective farming practices and efficient crop production management are necessary to uphold and increase the viability of agricultural production and safeguard or enhance the natural resource base and other ecosystems that are affected by agricultural activities. This section presents various production

compared with other varieties that yielded no more than 1 t ha–1 or yield nil in worse conditions (Table 6B). In eastern Uttar Pradesh, one characteristic looked for in the selection of varieties to produce is plant duration since this area is affected by irrigation and soil quality problems. In submergence-prone areas, farmers mostly grow traditional varieties in both lowland and upland areas. Sarya is grown in uplands while Jarethwa, Bhaislot, and Kalanamak are grown in lowlands. Sambha Mahsuri is the only improved variety grown by farmers in shallow lowlands because of the many positive traits possessed regarding quality, duration, yield, good straw, medium plant height, fine grain size, suitability to land type, and a high price for paddy. Crop establishment. In drought-prone and sodic areas, farmers normally apply farmyard manure (FYM) and plow the field 2 to 3 times for seedbed preparation. Sowing is usually done in mid-May to mid-June. Thereafter, land is plowed 3–4 times by tractor. When seedlings reach 22–30 days of age, transplanting is done. Harvesting and threshing are usually carried out manually and are mostly done by females. Farmers basically depend on rainwater for irrigation because of the high cost of diesel, although, sometimes, if farmers have the means, they use supplementary irrigation by either diesel or electric pump. In submergenceprone areas, sowing, land preparation, planting, and harvesting are carried out as in other types of environment except that, in seedbed preparation, farmers basically use direct seeding and do not apply FYM. Threshing is done mostly by tractor, followed by manual threshing. Fertilizer application and plant protection. Farmers in all types of environment generally do not apply chemical fertilizer, although a few of them apply N fertilizer after establishing seedlings. Doses vary according to land type and variety used per season. Before planting, a few farmers use DAP as basal. No control measures are used against insect pests and diseases unless necessary because of health hazards, especially during the dry season.

Manual weeding is done once or twice during crop growth, though some farmers do not use any weed control measures, especially in submergence-prone areas. Adoption of technology NDUAT scientists introduced the participating farmers to technologies that will enhance their productivity. Among these are the use of Sesbania for green manuring, pressmud for integrated nutrient management, salttolerant varieties adaptable to soil with high pH (i.e., Usar Dhan-3, CSR 30, NDRK 5089), submergence-tolerant lines in affected areas, and seed health training. In the selection of varieties, farmers appreciated the use of Usar Dhan-3 and CSR 30 in salt-affected areas because of their remarkable qualities compared with those of other varieties. Farmers mentioned that Usar Dhan, besides its good taste, gives good yield and has medium planting duration, whereas CSR 30 has fine grain, is easy to thresh, and does not lodge. Regarding fertilizer, the use of Sesbania fixes nitrogen in the soil, thereby reducing the use of fertilizer containing nitrogen, thus reducing the cost of fertilizer input, which is favorable to farmers. In addition, the crop following rice also benefits from a residual effect. This method reclaims sodic soil, can be easily adopted by farmers, and is environment-friendly. This is why farmers by far use Sesbania before planting rice. Alternatively, pressmud, which consists of sugarcane wastes, is also found to be good in sodic soils and it gives good yield, but the cost of transporting it inhibits farmers from adopting this technology because of insufficient capital.

Labor use and costs and returns
Gender division of labor in rice production Rice production is a labor-intensive activity, especially in areas where irrigation is a problem. Both men and women are actively involved in rice farming either as family labor or hired labor. Men basically are tasked to do land preparation, using either animals or machines, fertilizer and pesticide application,

Table 6A. Total and average rice area and yield by land type, by variety, and by type of environment. Village Number Drought Saryu-5 Swarna Mahsuri Sambha Mahsuri Hybrid Tall Mahsuri Total Submergence Saryu-5 Swarna Mahsuri Sambha Mahsuri Hybrid PVS line Gorakhnath NDR 97 Kalanamak Sarya Rashmi Tall Mahsuri Sugandhwala Padhnii Radha-4 Bengalia Farmbagri Jarethwa Bhaislot Lalsengar Total Sodic conditions Saryu-5 Swarna Mahsuri Sambha Mahsuri Hybrid Pant 0 Pant  Mahsuri Usar Dhan-3 Basmati NDR 97 NDR 359 Total 45 –   54  – 7 6 6 4 6 0.38 – 0.50 0.40 9.8 0.0 – .3 0.60 0.90 0. 3.54 4.9 – 5.9 5.9 3.96 3.65 – .8 3.64 .85 5.00 4.0 4 9 4 –  – 5 – – – – 43 .70 .03 .78 – 0.30 – .35 – – – – 9.6 4.0 4.6 4.5 – 4.7 – – – 3.04 – – 3.88 49 9 6  55  5 7 6 6 4 69 .08 .03 3.8 0.40 9.48 0.0 .35 .5 0.60 0.90 0. 3.68 4.8 4.6 4.59 5.9 3.97 3.65 3.65 .8 3.64 .85 5.00 3.97  – –  –  0  9 7 –   3 3 3 – – – 7 0.5 – – .63 – 0.38 .69 0.06 5.63 .9 – 0.3 0.06 0.88 0.38 .3 – – – 3.4 .40 – – 3.7 – 3.50 .9 .50 .60 .5 – .50 3.00 .90 .8 .00 – – – .09 –   – 3  – 0  – 0  – 0 – – 3 7  90 – 0.5 .9 – 0.5 0.5 – .06 0.88 – .94 – – – – – .06 .88 0.38 3.4 – .50 .59 – .43 3.00 – .6 .5 – .99 – – – – – 0.7 0.89 0.38 .5     3  0  3 7 0   3 3 3 3 7  6 0.5 0.5 .9 .63 0.5 0.63 .69 .3 6.50 .9 .94 0.3 0.06 0.88 0.38 .3 .06 .88 0.38 45.55 .40 .50 .59 3.7 .43 3.5 .9 .8 .65 .5 .99 .50 3.00 .90 .8 .00 0.7 0.89 0.38 .77 5  – 4   0.75 0.0 – 0.30 0.40 .65 .78 3.63 – 4.3 .96 3.30 – 5 0 3 8 6 – .08 4.90 0.80 .70 0.48 – 3.95 3.70 4.89 .54 3.53 5 7 0 7 9 38 0.75 .8 4.90 .0 3.0 .3 .78 3.86 3.70 4.45 .48 3.46 Upland Total area (ha) Average yield (t ha–) Number Lowland Total area (ha) Average yield (t ha–) Number Total Total area (ha) Average yield (t ha–)

Source: Baseline survey on rural households, CURE and CPWF projects, 004-05.


Table 6B. Total rice area and average yield by farm size, by variety, and by type of environment. Village Number Marginal Total area (ha) 0.75 0.98 .90 .0 .89 7.53 Average yield (t ha–) .78 .46 3.55 4.45 .35 3.3 Number Small Total area (ha) – 0.30 – – 0.30 0.60 Average yield (t ha–) – 4.7 – – 3.50 3.84 Medium and large Number Total area (ha) – .00 3.00 – – 4.00 Average yield (t ha–) – 5.50 4.03 – – 4.40 Number Total Total area (ha) Average yield (t ha–) .78 3.86 3.70 4.45 .48 3.46

Drought Saryu-5 Swarna Mahsuri Sambha Mahsuri Hybrid Tall Mahsuri Total Submergence Saryu-5 Swarna Mahsuri Sambha Mahsuri Hybrid PVS line Gorakhnath NDR 97 Kalanamak Sarya Rashmi Tall Mahsuri Sugandhwala Padhnii Radha-4 Bengalia Farmbagri Jarethwa Bhaislot Lalsengar Total Sodic conditions Saryu-5 Swarna Mahsuri Sambha Mahsuri Hybrid Pant 0 Pant  Mahsuri Usar Dhan-3 Basmati NDR 97 NDR 359 Total 8 4 6 – 36  3 4 5 3 4 04 5.08 0.63 0.78 0.00 4.08 0.0 .55 0.35 0.50 0.30 0. 3.58 4.0 3.7 4.44 – 4. 3.65 3.00 .63 4.07 .75 5.00 3.93 0 5 5  3 –  3 0 3 – 3 .80 .40 .50 0.40 0.40 0.00 0.50 0.90 0.00 0.60 0.00 8.50 4.48 4.5 4.56 5.9 3.77 – 3.30 3.06 – .95 – 4.3  – 5 – 6 0  –  – 0 34 4.0 0.00 .00 0.00 5.00 0.00 0.30 0.00 0.0 0.00 0.00 0.60 4.9 – 4.8 – 3.66 – 3.30 – .50 – – 3.96 49 9 6  55  5 7 6 6 4 69 .08 .03 3.8 0.40 9.48 0.0 .35 .5 0.60 0.90 0. 3.68 4.8 4.6 4.59 5.9 3.97 3.65 3.04 .8 3.64 .85 5.00 3.97 – – 9  –  6 3 7 – 3   – –   3  68 – – 3.00 0.3 – 0.5 .00 0.9 .44 – 0.44 0.3 0.06 – – 0.3 5.06 .3 0.3 4.9 – – .84 5.00 – 3.00 .59 .53 .45 – 3. .50 3.00 – – .50 0.63 . 0.00 .54   9 6 3  4 7 0 7 7 – –  3  6   70 0.5 0.5 4.56 .00 0.5 0.38 0.69 0.69 .44 .9 .50 – – 0.38 0.38 0.50 .00 0.50 0.5 7. .40 .50 .60 .93 .43 3.50 .40 .33 .79 .5 .46 – – .60 .8 .50 . .30 0.75 .96 – – 4 4 – – –  4 – – – –  –  5 3 – 3 – – 4.63 0.50 – – – .5 .63 – – – – 0.50 – 0.50 4.00 .5 – 4.6 – – .0 4.55 – – – 0.0 .3 – – – – .50 – .00 0.45 0.53 – .88     3  0  3 7 0   3 3 3 3 7  6 0.5 0.5 .9 .63 0.5 0.63 .69 .3 6.50 .9 .94 0.3 0.06 0.88 0.38 .3 .06 .88 0.38 45.55 .40 .50 .59 3.7 .43 3.5 .9 .8 .65 .5 .99 .50 3.00 .90 .8 .00 0.7 0.89 0.38 .77 5 5 7 7 8 3 –  – –   –  3 – – 4 5 7 0 7 9 38 0.75 .8 4.90 .0 3.0 .3

Source: Baseline survey on rural households, CURE Project and CPWF projects, 004-05.


and irrigation. On the other hand, women are responsible for transplanting, weeding, harvesting, and threshing, but men sometimes help in performing these activities. To do all the rice cultivation activities, a total average of 131 person-days is needed to complete one crop. Farmers in salt-affected areas work as many as 193 person-days, whereas in drought-prone and submergenceprone areas, farmers use only 126 and 75 person-days per hectare, respectively. Women notably contribute the days in rice cultivation (Table 7). Table 8 clearly shows that activities mostly done by women require more labor. An average of 92% of the total labor-days is needed to carry out transplanting, weeding, harvesting, and threshing activities in all environments. These results indicate that female labor participation, regardless of the type of ecosystem, dominates activities done on the farm. This is because of economic pressure that pushes members of farm households, who normally work on the farm, to seek off-farm work, leaving the partner to look after the farm (normally the women). Given this condition, there is an enormous need to provide women with technical knowledge and expertise to enable them to make sound farm management decisions since they are the ones left on the farm. This strategy is expected to contribute toward improvements in labor productivity, increase income, reduce hunger, and improve the well-being of the entire family. Costs and returns analysis of rice in different rainfed environments To determine the profitability of rice production on different rainfed environments in eastern Uttar Pradesh, a costs and returns analysis was done (Table 9). Values for cost and return items were expressed on a per hectare basis. The structure of the cost accounts in the analysis includes noncash costs. Costs for family labor and use of own resources were given imputed values using the existing wage rates and rental fees. In doing so, net income should therefore be seen as an indicator of a farmer’s ability to recover both cash and noncash costs of rice

production instead of the typical notion of profit. The average parcel size cultivated during the wet season in all types of environment is 0.32 ha with an average rice yield of 3.03 t ha–1. The high yield of rice in salt-affected areas (4.11 t ha–1) and drought-prone areas (3.47 t ha–1) could be attributed to the intensive use of highyielding varieties accompanied by suitable crop and natural resource management practices. In submergence-prone areas, the low yield (1.51 t ha–1) of rice could be accounted for by continued use of local rice varieties using low inputs because of several abiotic constraints faced by farmers. Traditional varieties according to farmers are more tolerant of both submergenceand drought-prone environments. In addition to the nonavailability of location-specific high-yielding varieties, farmers are forced to continue to use local traditional varieties due to a lack of awareness about high-yielding varieties. In a study by Sarkarung (1996), as cited by Paris et al (2000), he mentioned that the majority of improved rice cultivars developed on-station failed to perform under farmers’ conditions. This implies that the newly released cultivars could not compete with native and traditional cultivars under adverse conditions in which water and fertility are not controlled. The total cost spent on rice cultivation was highest in sodic (Rs 10,100 ha–1) and droughtprone (Rs 7,701 ha–1) areas. This was caused by high labor input use as well as fertilizer consumption (Rs 2,320 in sodic areas and Rs 1,257 in drought-prone areas), as these areas use modern varieties in rice farming that require meticulous crop and nutrient management (see also Table 9). Among the rice input requirements, the highest proportion of expenditures was on labor use, both family and hired labor, that is a little more than 50% of the total cost (Fig. 4). A low proportion was spent on irrigation from nil in the submergence-prone areas to 5% in sodic areas. In addition, the cost of seed and seedling establishment (10%) also represented the lowest share in total rice spending. The total gross returns obtained from rice cultivation were also highest in drought-prone

Table 7. Labor input (person-days per hectare) in rice production by caste and type of environment. tem Upper Labor days Drought Family labor Male Female Total family labor Hired labor Male Female Total hired labor Labor days Male Female Submergence Family labor Male Female Total family labor Hired labor Male Female Total hired labor – Labor days Male Female Total labor days Sodic conditions Family labor Male Female Total family labor Hired labor Male Female Total hired labor Labor days Male Female Total labor days 5 38 63 40 60 40 37 77 5 48 00 60 37 97 6 38 – – – – – – 4 37 79 53 47 00 0 38 58 3 60 9 3 9  4 5 9 3 6 9 3 6 9 – – – – – – 7 9 6 9 4 33 5 – 5 8 – 8 37 8 55 48 3 7 57 3 88 59 3 9 – – – – – – 35 8 53 44 3 67 – – – – – – 4 8 4 57 43 00 – – – – – – 9  5 56 44 00 6 0 46 57 43 00 – 4  – –   8 7 – – – –   6 5 – – – – 5 7 9 3 4 4 8 9 – – – –  6 37 5 38 89 – – – – 4 6 40 47 3 78 3 6 39 49 34 83 6 8 37 63 35 30 54 46 33 4 44 56 40 37 5 48 8 3 47 53 0 7 37 3 6 85 – – – – – – 3 39 6 3 5 83 – – – – – – 6 4 0 0 4 34 6  7 4  5 35 30 65 54 46 00 0 3 3 4 3 7 40 37 77 5 48 00  7 39 37 9 66 Percent Backward Labor days Percent Scheduled Labor days Percent Muslim Labor days Percent Total labor days Percent

Source: Baseline survey on rural households, CURE and CPWF projects, 004-05.


Percent 70 60 50 40 30 0 0 0 Labor Seed Fertilizer 7 7 3 6 9  0 Animal/ machine 5 rrigation 3 6 5 5 5 58 53 Drought Submergence Sodic

Fig. 4. Proportion of farm expenditure items to total farm expenditure by type of environment.

Table 8. Labor input (person-days ha–1) used in kharif and rabi seasons, by agricultural operation and by village. Activity Land preparation Transplanting Weeding rrigation Fertilizer application Harvesting Threshing Total Drought Labor days 5 44 4 –  43 30 7 Percent 4 34 3 –  34 3 00 Submergence Labor days 6  3 –  30 5 75 Percent 8 7 4 –  40 9 00 Sodic conditions Labor days 0 58 4 3  4 37 93 Percent 5 30     9 00

Source: Baseline survey on rural households, CURE and CPWF projects, 004-05.

areas (Rs 20,820 ha–1) and sodic areas (Rs 17,341 ha–1) where 91% and 99%, respectively, were mainly obtained from rice grain and the rest from its by-products such as rice straw.

Source of income of farm families
Despite the importance of rice as a staple food and the dependence of most resourcepoor farmers on rice farming as a source of income, rice contributed only a small (11–19%) proportion of total income. Hence, farming families resort to different sources of livelihood to spread risks (Table 10). Among the sources of income, cash from nonfarm activities had the most share, especially in submergenceprone areas (75%) and sodic areas (53%). This includes remittances received from male family members who migrated and worked

as a skilled laborer (tailor, driver, mason, and plumber), construction worker, salaried worker, businessman, and salesman in nearby cities, specifically Faizabad City. In the droughtprone areas, income from agriculture (71%) was substantially higher, which could be attributed to the cultivation of crops other than rice such as vegetables and sugarcane. Wheat production also contributed 13% of total income.

Perception on household economic condition and self-sufficiency
A majority (90%) of the farmer-respondents in all types of stress environments revealed that their economic conditions had improved. Only 6% from salt-affected areas mentioned that their situation has worsened, whereas other farmers expressed that no change had taken place for

Table 9. Costs and returns of rice production by caste and by type of environment. Drought Submergence Sodic conditions Total Upper Backward 4.6 0.6 7,58 38 7,899 ,47 5,407 34 40 405 6 – 6,97 ,84 4 6,553 5 03 3 – 4,845 4,97 97 43 ,5 ,9 64 96 0,773 6,5 50 ,407 4,864 440 ,5 90 09 707 9,698 8,0 89 .5 4.0 0.35 7,85 00 7,85 0.38 8,385 657 9,04 ,35 ,83 Scheduled 4.00 0. 6,6 – 6,6 ,43 6,54 38 ,035 ,07 6 377 0,45 5,8 34 Total 4. 0.4 7,49 9 7,34 ,559 5,334 348 ,36 ,00 8 539 0,00 7,4 67 .00 0.4 ,773 ,077 ,850 ,34 3,86 375 43 Muslim Total Backward Muslim .0 0.35 4,998 37 5,35 ,36 ,80 309 48 – – – 3,394 3.47 0.36 8,989 ,83 0,80 ,967 3,96 50 835 6 60 6 7,70 3,8 30 .97 0.9 7,050 ,500 8,550 ,50 4,69 433 708 75 – – 8,58 0,9 37 3.64 0.35 8,43 ,775 0,88 ,50 4,30 45 ,000 63 – 35 8,743 ,445 63 Upper Backward Scheduled 3.68 0.7 0,000 ,03 ,04 ,667 4, 340 908 8 – – 7,355 4,658 337 3.35 0.46 8,57 ,735 0,6 ,36 3,466 673 76 45 364 80 7,76 ,537 88 


Rice yield (t ha–)

Area (ha)

Returns (Rs)

Value of production

Value of by-products

Total returns

Costs (Rs)

Animal and machine labor

Human labor




Chemical fertilizer 


Total cost

Net returns (Rs)

Net returns ($)a

Exchange rate (1-1-05): $1 = Rs 43.47. Source: Baseline survey on rural households, CURE and CPWF projects, 004-05.



Table 10. Average income by source and by type of environment. Source Rice Wheat Other crops Other agricultural income Total agricultural income Nonfarm income Remittances Other nonfarm income Total nonfarm income Total Drought Percent 9 4 33 6  8 5 5 8 00 Average 8,877 6,473 5,67 ,768 33,385 8,30 ,50 ,560 3,400 46,785 Submergence Percent  0 5 0 6 53  0 75 00 Average 7,08 6,3 3,3 00 6,573 34,30 4,360 0 48,680 65,53 Sodic conditions Percent 3 5 – 9 47 – 5  53 00 Average ,76 3,73 – 7,56 4,73 – 47,74 ,300 49,574 9,305

Source: Baseline survey on rural households, CURE and CPWF projects, 004-05.

Table 11. Condition of farming households by type of environment. Condition Household economic condition Has improved No change Has worsened Total Better off More self-sufficient Poorer No response Total 3  – 5 7 5   5 9 8 – 00 8 60 8 4 00 49  – 50 3 36  – 50 98  – 00 6 7  – 00 40 7 3 50 9 8 3 – 50 80 4 6 00 38 36 6 – 00 Drought Number Percent Submergence Number Percent Sodic conditions Number Percent

Household condition compared to other households

them. In addition, more than three-quarters of the farmers felt that they had become betteroff and more self-sufficient now than before. The rest believed that they had become poorer (Table 11). The primary reasons for this improvement are high crop productivity due to good-quality inputs, remittances received from family members, as well as income from nonfarm sources. Others reasons cited are the presence of irrigation facilities, the use of improved varieties, income from livestock by-products,

land and tractor rental, additional area for cultivation, and small household size, which is contrary to some farmers who said that their large family size helped to improve their economic conditions because more labor is available to do farming activities at no cost. In addition, more efficient and productive male and female laborers in the family could contribute to increasing household income. Conversely, farmers mentioned that the lower income (67%) they received than before was the indicator that made them realize that

their economic condition had worsened. This is due to a lack of a regular income source, which pushes farmers toward poverty. Family problems, large family size and more often than not an extended family, small landholding, and drug addiction are also some of the reasons stated for worsening household conditions.

Constraints in crop production and research opportunities
Problems identified by farmers Constraints to crop productivity in various rice stress environments are related to hydrology, soil and nutrient management, the availability of improved varieties suitable to the area, goodquality seed, and the high cost of inputs. These constraints follow: 1. Moisture stress due to unpredictable and often inadequate rainfall, poor soil health, and nonavailability of supplementary irrigation facilities at the right time. 2. In early drought (June-July) periods, most of the areas are fallowed because of the absence of good irrigation facilities. When the heavy rain comes, flash floods and waterlogging/submergence occur because of poor drainage. High salt content in the soil is also a serious problem, especially in the dry season. 3. The nonavailability of stress-tolerant varieties and farmers’ lack of awareness about the use of modern varieties caused an incessant use of traditional local cultivars. 4. High costs of inputs, for example, high rental costs of tractors, and irrigation pumps, the increasing price of diesel (about Rs 36 per liter), and high costs of fertilizer. 5. An inadequate supply of electricity to run electric motors for irrigation. 6. A lack of awareness of suitable technology and technical knowledge on how to efficiently use the technology. Potential opportunities to improve farmers’ livelihood Given these constraints, scientists, field workers of the government, as well as farmers

recommended and noted some opportunities available to improve the livelihood of farm families. The following potential opportunities can be regarded as tools for enhancing the rural livelihood of farmers at the target sites: 1. The introduction of suitable drought-, submergence-, and salt-tolerant varieties. 2. Awareness of efficient crop and nutrient management regarding the use of these types of cultivars and ensuring their availability. 3. The introduction of 4–5 new lines in different ecosystems (shallow lowland, semideep, and flash-flood areas). 4. The introduction of better crop establishment methods such as the use of pressmud (waste from sugarcane factories) and Sesbania (as green manure) in rice fields before planting in sodic areas. 5. The introduction of integrated nutrient management in seedbed preparation. 6. Improvement in the level of technical knowledge of farmers by need-based training about recommended new agricultural technologies and appropriate cropping pattern according to the land situation. 7. Dissemination of technologies and evaluation of more lines of drought-, submergence-, and salt-tolerant varieties in the form of adaptive trials in farmers’ fields.

The productivity of rice is strongly influenced by the hydrological conditions prevailing in the area, particularly in the growing season. These hydrological influences, according to Borkakati et al (2000), can be mitigated considerably if crop management strategies are developed on the basis of knowledge of environmental conditions. Likewise, other appropriate technologies can be developed using such knowledge and can easily be applied by farmers in various rainfed situations. Farmers showed positive feedback on participatory experiments in different types of stress environments and associated crop and water management technologies conducted in

their fields. However, with the presence of these technologies to address the problems faced by farmers, the following concerns still require further attention. 1. A need for strong interaction among biological and social scientists in problemoriented research, as well as among nongovernment and private organizations in cooperation with various research institutions to develop more effective approaches of impact-oriented agricultural research to help farmers explore the maximum possibilities to increase their productivity. 2. The use of community participatory approaches from design, planning, validation of technologies, and dissemination such that farmers will be actively involved from the very beginning to intensify their receptivity for adoption. 3. A need to enhance the capacities of both men and women in rice production and improve seed management. 4. A need to anticipate and address constraints to the widespread adoption of technologies such as increasing government support for public investment in infrastructure (i.e., good roads, good-quality drainage and irrigation, and construction of small ditches) and increased and regular extension services on farms to encourage farmers to continue their production and become responsive to technology developments.

Borkakati K, Singh VP, Singh AN, Singh RK, Sastri ASRAS, Mohanty SK. 2000. Planning and managing rice farming through environmental analysis. In: Tuong TP, Kam SP, Wade L, Pandey S, Bouman BAM, Hardy B, editors. Characterizing and understanding rainfed environments. Proceedings of the International Workshop on Characterizing and Understanding Rainfed Environments, 5-9 December 1999, Bali, Indonesia. Los Baños (Philippines): International Rice Research Institute. p 191-213. Bouman B, Barker R, Humphreys E, Tuong TP. 2007. Rice: feeding the billions. In: Tuong TP, Kam SP, Wade L, Pandey S, Bouman BAM, Hardy B, editors. Water for food, water for life: a comprehensive assessment of water management in agriculture. London (UK):

Earthscan, and Colombo (Sri Lanka): International Water Management Institute. p 515-549. Diwakar MC. 2006. Productivity status of rice during ninth plan. India: Ministry of Agriculture. 67 p. Food and Agriculture Organization (FAO) 2004. Rice is life. Italy: FAO. www.fao.org/newsroom/en/ focus/200436887/index.html. ICAR signs workplan agreement with IRRI for collaboration in rice research genetic enhancement of rice major thrust area. (n.d.). Retrieved September 25, 2008, from www.parinda.com/news_archives/june2005/icar-signs-workplan-agreement-with-irri-forcollaboration-in-rice-research---genetic-enhancementof-rice-major-thrust-area.shtml Pandey S, Behura D, Villano R, Naik D, 2001. Drought risk, farmers’ coping mechanisms, and poverty: a study of the rainfed rice system in eastern India. In: Peng S, Hardy B, editors. Rice research for food security and poverty alleviation. IRRI, Los Baños, Philippines. p 267-275. Paris T, Singh A, Hossain M, Luis J. 2000. Using Gender Analysis in Characterizing and Understanding Farm-Household Systems in Rainfed Lowland Rice Environments. In: Tuong TP, Kam SP, Wade L, Pandey S, Bouman BAM, Hardy B, editors. Characterizing and understanding rainfed environments. Proceedings of the International Workshop on Characterizing and Understanding Rainfed Environments, 5-9 December 1999, Bali, Indonesia. Los Baños (Philippines): International Rice Research Institute. p 339-369. Ramasamy and Selvaraj, 2006. Do drought tolerant rice and salinity tolerant rice offer a panacea for Indian farmers farming in fragile environments? ABSP II Newsletter. 1(2):6. Sarkarung S. 1996. Breeding rice cultivars suitable for rainfed lowland environments: a farmer participatory approach in Eastern India. In: Eyzaguire P, Iwanaga M, editors. Participatory plant breeding. Proceedings of a Workshop on Participatory Plant Breeding, 26-29 July 1995, Wageningen, The Netherlands. Rome (Italy): International Plant Genetic Resources Institute. Status of irrigation in drought prone areas. n.d. Retrieved September 25, 2008, from http://wrmin.nic.in/index3. asp?sslid=439&subsublinkid=364&langid=1 Zeigler RS. 2006. Rice research for poverty alleviation and environmental sustainability in Asia. Los Baños (Philippines): International Rice Research Institute. www.niaes.affrc.go.jp/sinfo/sympo/h18/20061212/ pdf/s3_zeigler.pdf.

Authors’ addresses: T. Paris and A.D. Cueno, Social Sciences Division, International Rice Research Institute (IRRI), DAPO Box 7777, Metro Manila, Philippines; A. Singh, Narendra Deva University of Agriculture and Technology (NDUAT), Faizabad, Uttar Pradesh, India.