Productivity of sodic soils can be enhanced through the use of salt tolerant rice varieties and proper agronomic practices

Regaining the agricultural potential of sodic soils in the Indo-Gangetic plains necessitates the development of suitable salt tolerant rice varieties to provide an entry for other affordable agronomic and soil manipulation measures. Thus selection of high yielding rice varieties across a range of sodic soils is central. Evaluation of breeding lines through on-station and on-farm farmers’ participatory varietal selection (FPVS) resulted in the identification of a short duration (110–115 days), high yielding and disease resistant salt-tolerant rice genotype ‘CSR-89IR-8’, which was later released as ‘CSR43’ in 2011. Several agronomic traits coupled with good grain quality and market value contributed to commercialization and quick adoption of this variety in the sodic areas of the Indo-Gangetic plains of eastern India. Management practices required for rice production in salt affected soils are evidently different from those in normal soils and practices for a short duration salt tolerant variety differ from those for medium to long duration varieties. Experiments were conducted at the Indian Council of Agricultural Research-Central Soil Salinity Research Institute (ICAR-CSSRI), Regional Research Station, Lucknow, Uttar Pradesh, India during 2011 and 2013 wet seasons, to test the hypothesis that combining matching management practices (Mmp) with an improved genotype would enhance productivity and profitability of rice in sodic soils. Mmp were developed on-station by optimizing existing best management practices (Bmp) recommended for the region to match the requirements of CSR43. The results revealed that transplanting 4 seedlings hill−1 at a spacing of 15 × 20 cm produced significantly higher yield over other treatments. The highest additional net gain was US$ 3.3 at 90 kg ha−1 N, and the lowest was US$ 0.4 at 150 kg ha−1 N. Above 150 kg ha−1, the additional net gain became negative, indicating decreasing returns from additional N. Hence, 150 kg N ha−1 was considered the economic optimum N application rate for CSR43 in these sodic soils. Using 150–60–40–25 kg N–P2O5–K2O–ZnSO4·7H2O ha−1 in farmers’ fields grown to CSR43 produced an average of 5.5 t ha−1 grain. The results of on-farm evaluation trials of CSR43 showed that matching management practices (Mmp) increased yield by 8% over existing best management practices (Bmp) recommended by ICAR-CSSRI for sodic soils and by 16% over framers’ management practices; however, combining Mmp with CSR43 resulted in 35% higher yields over farmers’ current varieties and management. This approach of combining cost effective crop and nutrient management options and a salt-tolerant variety can maximize the productivity and profitability of sodic soils in the alluvial Indo-Gangetic plains and in neighboring salt-affected areas of the Ganges mega delta in South Asia.


Introduction
Salt affected lands are estimated at about 955 million ha worldwide (Szabolcs, 1994), afflicting 7% of the world's total arable land (Flowers et al., 1997). The Indo-Gangetic region in India (21 • 55 -32 • 39 N and 73 • 45 -80 • 25 E;Singh et al., 2010) has about 2.7 million ha of salt affected soils, consisting mostly of centuriesold barren sodic soils with no land use opportunities (NRSA and Associates, 1996). These soils have been regarded as unfit for agriculture due to high pH (>8.5) and concentrations of soluble salts that produce alkaline hydrolysis products such as Na 2 CO 3 and NaHCO 3 , together with sufficient exchangeable sodium to cause poor physical soil characteristics.
Grain yield of rice in salt affected soils is much lower because of its high sensitivity to salt stress (Flowers and Yeo, 1989;Gao et al., 2007;Ismail et al., 2007). Rice is exceptionally sensitive to salinity and sodicity at early seedling stage (Aslam et al., 1988(Aslam et al., , 1993 and high yield losses have been observed because of high mortality and poor crop establishment. Modern high yielding varieties require considerable investment to ameliorate these soils to ensure reasonable yields, but this investment is beyond the capabilities of the resource-limited small holder farmers living off these salt affected areas. Increasing and sustaining yields in these areas will require a system that integrates salt tolerant varieties with effective and affordable crop and nutrient management practices. Chowdhury et al. (1993) observed that the number of seedlings hill −1 and plant spacing were important factors determining plant population per unit area for optimum nutrient uptake and for accessing sufficient light for photosynthesis, which ultimately determine grain yield. Plant mortality in sodic soils is high when young seedlings are transplanted, adversely affecting plant establishment and growth (Dargan and Gaul, 1974). Older seedlings survive better and establish earlier, but they produce less tillers and exhibit poor growth (Mandal et al., 1994). The effects of seedling age become more noticeable in short duration varieties (Chandrakar and Chandravanshi, 1988).
Sodic soils are inherently low in organic matter (<0.1%), and available N, and are more responsive to N application. These soils are more prone to N losses due to higher N volatilization caused by high pH, further aggravating N deficiency. Microbial activity, which influences N mineralization, is restricted by salt stress (Abrol et al., 1988;Liu and Kang, 2014;Wong et al., 2010). Therefore, the requirement of added N in sodic soils is higher than in normal soils and salt tolerant varieties appear to respond better to higher N application than sensitive varieties (Dixit and Patro, 1994;Meena et al., 2003).
Apparently, the lack of high yielding salt tolerant rice varieties, together with good management strategies specific for sodic soils are the main reasons for low and unstable productivity. High yielding salt tolerant varieties, like CSR10, CSR13, CSR23, CSR27, CSR36, Narendra Usar dhan3, which can tolerate sodicity of up to pH 9.8, were developed through conventional plant breeding and have made good impacts in salt-affected areas of India (Mishra et al., 1992). However, these varieties are of medium to long duration (130-140 days) and often do not fit well in the rice-wheat system predominating the Indo-Gangetic plains. Integrating farmers' participatory varietal selection approach (FPVS) within the evaluation process during breeding strengthens and accelerates the selection of varieties with characteristics desired by farmers and expedites their adoption when released. Using this approach, field evaluation of salt-tolerant breeding lines was conducted through the network of the "Stress Tolerant Rice for Africa and South Asia (STRASA)" project for five years under varying sodic soil conditions in Uttar Pradesh, India, culminating in the release of the salt tolerant variety "CSR43" in 2011 (Singh et al., 2013(Singh et al., , 2014. Suitable management practices for salt affected soils are obviously different than in normal soils. Moreover, any management option for sodic soils developed using medium to long duration salttolerant varieties might not be suitable for a short duration variety like CSR43. This study reports on the evaluation and release of CSR43 and on developing suitable crop and nutrient management practices to ensure better establishment and higher productivity of this variety, with the hypothesis that this combination will help to improve productivity and profitability of rice in the sodic soils of the eastern Indo-Gangetic plains.

Germplasm evaluation through participatory varietal selection
A set of 126 geographically and genetically diverse rice genotypes was screened through researchers' managed on-station and on-farm trials during 2001. This set includes local genotypes, advanced salt tolerant breeding lines and salt tolerant high yielding varieties. From 2002to 2005, FPVS (Paris et al., 2011 was employed during evaluation, and a set of 6 genotypes was selected from on-farm trials (Table 1). These genotypes were further evaluated, taking into account traits desired by men and women farmers, including grain yield and quality. A scale of 1-10 (1 being least preferred, 10 most preferred) was used to score and rank important traits. Preference scores for each genotype were calculated by subtracting the total negative votes from total positive votes, then dividing by the total votes. Through this process, the genotype CSR-89IR-8 consistently ranked first based on both farmers' preference ranking and grain yield. From 2006 to 2008, the performance of this genotype was further evaluated in farmers' managed trials and compared with popular high yielding varieties, then released as CSR43 (Singh et al., 2013(Singh et al., , 2014 (Table 1).

Site characterization
Field experiments were conducted from 2011 to 2013 at the experimental farm of ICAR-CSSRI, Regional Research Station, Uttar Pradesh, India (26 • 47 58 N, 80 • 46 24 E, 120 m above MSL). The study site is representative of large areas of abandoned alkali soils in the Indo-Gangetic plains. The soil is Typic Natrustalfs, with surface soil (0-15 cm) pH (1:2 soil:water) of 8.9 and sub-surface (>15 cm) pH of >9.4. The soil presents physical and nutritional constraints to plant growth due to poor soil water and soil air characteristics caused by high bulk density (>1.5 g cm −3 ) and low infiltration rate (<2 mm day −1 ) (Sharma et al., 2006). Poor soil aeration in sodic soils restricts root development, respiration and soil microbial activities resulting in poor crop growth and productivity. Before the experiment, soil samples were collected from the 0-120 cm soil profile of the experimental field. The samples were air dried, ground in a Wiley mill and passed through a 2.0 mm sieve. The pH and EC of the soil were determined using digital pH and conductivity meters (Jackson, 1967). The pH and EC of surface soil (0-15 cm) were 9.2 and 0.61 dS m −1 , respectively. Organic carbon was determined using the rapid titration method of Walkley and Black, modified by Walkley (1947) and was 0.29%. The concentration of Ca 2+ + Mg 2+ , measured with an Inductivity Coupled Plasma (ICP) Analyzer (PerkinElmer), was 2.60 me l −1 . Na + and K + concentrations were determined using a Flame Photometer (Richards, 1954) and were 4.50 and 0.06 me l −1 . Carbonate and bicarbonate concentrations were determined by titration with 0.01 N H 2 SO 4 using phenolphthalein and methyl orange as indicators and chlo- ride was determined through titration with 0.01 N AgNO 3 solution using potassium chromate as indicator (Tandon, 1995;Richards, 1954). The respective concentrations were 1.0 me l −1 , 4.0 me l −1 and 2.0 me l −1 . The climate of the experimental site is subtropical monsoon, with average annual rainfall of 817 mm. The highest rainfall during the cropping period (June-October) was received in August (331.8 mm), followed by July (215.5 mm), September (208.2 mm), and then June (167 mm). Total rainfall during 2011-2013 cropping seasons were 1034, 990 and 795 mm, respectively. The average annual evapotranspiration was 1580 mm, which varies with air temperature. The mean maximum temperature during the cropping period (July-October) was 33.5 • C in July and mean minimum temperature was 18.8 • C in October.

Experimental details
Details of the two sets of experiments conducted during this study are presented below and summarized in Table 2. These experiments were conducted to develop matching management practices (Mmp) for CSR43 by optimizing existing management practices (Bmp) recommended for the Indo-Gangetic plains of Eastern Uttar Pradesh.

Optimizing number of seedlings hill −1 and spacing
To determine the optimum number of seedlings hill −1 and hill spacing for CSR43, a field experiment was conducted in 2011 with three replications. The main plots comprised two treatments, T 1 : 2 seedlings and T 2 : 4 seedlings per hill. Three spacing treatments were used as sub-plots; S 1 : 15 × 15 cm, S 2 : 15 × 20 cm, S 3 : 20 × 20 cm (Table 2). Thirty-day-old seedlings were transplanted in puddled and leveled field. Fertilizer rate at 120-60-40-25 kg N-P 2 O 5 -K 2 O-ZnSO 4 ·7H 2 O ha −1 was applied uniformly. Half N and full P 2 O 5 , K 2 O and ZnSO 4 ·7H 2 O were broadcasted as basal after puddling and the remaining N was applied in equal splits at active tillering (30 days after transplanting; DAT) and panicle initiation (60 DAT).

Optimizing nitrogen requirement in the field
To determine the optimum nitrogen application for transplanted CSR43 in these sodic soils, an experiment was conducted in 2013, with three replicates and six N treatments (N kg ha −1 ; N 1 : 0, N 2 : 100, N 3 : 125, N 4 : 150, N 5 : 175 and N 6 : 200) ( Table 2). Thirty-day-old seedlings were transplanted at 4 seedlings per hill spaced at 15 × 20 cm. Full P 2 O 5 , K 2 O and Zn (60-40-25 kg N-P 2 O 5 -K 2 O-ZnSO 4 ·7H 2 O ha −1 ) and half of N were applied as basal and the remaining N was applied in equal splits at active tillering (30 DAT) and panicle initiation (60 DAT).
In both experiments, seeds of CSR43 were treated with ceresin before sowing to control seed borne diseases. Nutrients at rates of 100-60-40 kg N-P 2 O 5 -K 2 O ha −1 along with 5 t FYM ha −1 were applied in the nursery, followed by seed sowing of 40 g m −2 . Weeds were controlled using Butachlore (1.5 l ha −1 ) applied 2-3 DAT. No major disease or pest infestation was observed during 2011-2013.

Observations and data collection
After transplanting, 10 plants were tagged in each treatment and used for assessing growth and yield attributes at maturity. Plant height was measured from the base of the stem to the tip of the longest panicle at maturity. Days to maturity were determined when 80% of the spikelets within each treatment reached physiological maturity. Leaf area was measured using a leaf area meter (LAI-2200C, LI-COR) at the flag leaf stage and calculated using the formulae of Watson (1958). Panicles were hand threshed, filled and unfilled grains were separated manually and the number of spikelets per panicle was counted. After initial sun drying, grains and straw were dried in an oven (60-70 • C) to a constant weight and weighed. Floret fertility percentage, 1000-grain weight and harvest index (HI) were computed from collected data (Ying et al., 1998). To determine grain yield, grains were harvested from the net plot area of 20 m 2 , sun-dried and weighed. Grain moisture content was determined with a digital moisture meter (DMC-700, Seedburo, Chicago, IL, USA) and grain weight was adjusted to moisture content of 14%.

On-farm evaluation of the tolerant variety and best management practices
To validate the effect of management practices developed for CSR43 in on-station trials, on-farm evaluation was conducted at five farmers' fields during the wet season of 2014 in Patwakhera village, Lucknow district, Eastern Uttar Pradesh, India (Table 2). Soil samples were collected and analyzed from 12 farmer's fields, then five fields were selected with soil pH ranging from 9.0 to 9.5. Each farmer was considered one replicate with plot size of 300-400 m 2 . Four treatments were used; T 1 : Farmers' varieties (Fv) with farmers' management (Fm); T 2 : Improved variety (Iv, CSR43) with Fm, T 3 : Iv with existing best management practices (Bmp) recommended for sodic soils of Eastern Uttar Pradesh by ICAR-CSSRI; T 4 : Iv with matching management practices (Mmp) developed through on-station trials during 2011-2013. Grain yield was determined from an area of 20 m 2 .

Statistical and economic analyses
All data were subjected to analysis of variance (ANOVA) (Gomez and Gomez, 1984) using WINDOSTAT. The least significant differ- ence (LSD) at 5% probability was used to compare treatment means (Steel and Torri, 1980). Data was presented as means across experiments for all parameters and main effects were reported as there are no significant interactions between plant spacing and number of seedlings hill −1 . Economic analysis of different treatments was based on total production cost (fixed and variable), gross return, gross margin and benefit/cost ratio. Fixed costs included the cost of inputs (seed, fertilizers and FYM) and labor (field preparation, bunding, irrigation, fertilizers, harvesting and threshing), whereas, variable cost included labor charges for uprooting and transplanting of seedlings based on treatments in 2011 and varying rates of N and costs of application in 2013. Gross returns were calculated by multiplying the weight of grain by their corresponding minimum support price for respective year (US$ 190/t in 2011 and US$ 218/t in 2013) and market price of straw at harvest (16.66 US$/t during both years). The gross margin was calculated by subtracting total variable cost from gross returns, and the benefit/cost ratio was calculated by dividing gross return by total variable cost. The values were calculated using Indian rupees (INR), and then converted into US$ using conversion rate of 1US$ = 60 INR. Grain yield response to nitrogen was determined using curve expert model 32 and the economically optimum dose of N was calculated on the basis of maximum economic yield produced per unit of N (Fageria and Baligar, 2003).

Evaluation and release of CSR43
After intensive evaluation and selection within the 126 genotypes from 2001 to 2004, only 6 breeding lines were selected (Table 1). Data on preference analysis and grain yield of these lines from a trial conducted in the 2005 wet season are presented in Table 3. Farmers ranked CSR-89IR-8, CSR36 and NDR359 as their first, second, and third preferred genotypes, respectively, with positive correlations between preference scores of male and female farmers (r = 0.91**) and between farmers and researchers (r = 0.74*). CSR-89IR-8 ranked first in all desired traits. CSR-2K-239 and CSR13 had the lowest scores for yield and other traits indicating they are not suitable for sodic soils. CSR36, and CSR-2K-262 showed high sodicity tolerance and grain yield but their duration is longer. Farmers selected CSR-89IR-8 as their preferred genotype because of several traits, including its good taste, aroma, color, non-cohesiveness when cooked and higher grain yield (Table 3). Moreover, the shorter duration of this variety saves 2-3 irrigations, allowed early establishment of rabi wheat; and its early maturity provided good market price (Singh et al., 2013(Singh et al., , 2014.

Management options
3.2.1. Optimizing number of seedlings hill −1 and spacing (Experiment I) Growth, grain yield and yield attributes (except % fertility and 1000-grain weight) were significantly higher when 4 seedlings were transplanted per hill compared with 2 seedlings per hill (Table 4). Productive tillers hill −1 , grain and straw yields with 4 seedlings hill −1 were, respectively, 15.9%, 22.7% and 30.3% higher than those with 2 seedlings hill −1 (Fig. 1a and b). Spacing had significant (P < 0.05) effect on plant height, fertility and grain yield. Using 15 × 20 cm spacing resulted in about 4-7% more spikelets panicle −1 over both 15 × 15 and 20 × 20 cm spacing though these differences were not significant. It also resulted in 11% higher grain yield and 16% higher straw yield over the values obtained with 20 × 20 cm spacing ( Fig. 1a and b). Planting of 4 seedlings hill −1 resulted in 23.4% and 42.6% higher gross and net returns, respectively, than planting of 2 seedlings hill −1 due to higher grain and straw yields. Transplanting of 4 seedlings hill −1 at a spacing of 15 × 20 cm significantly increased gross return, net return and BCR because of enhanced number of productive tillers and other yield attributes (Table 5). Fixed cost (329.5 US$ ha −1 ) incurred in this experiment is presented as Supplemental Table 1.

Optimizing N requirements for sodic soils (Experiment II)
Plant height and productive tillers hill −1 increased with increasing N application ( Table 6). Application of 200 kg N ha −1 produced 70%, 53.8%, 21.5%, 9.7% and 3.5% more productive tillers hill −1 than the application of 0, 100, 125, 150 and 175 kg N ha −1 , respectively. Maximum leaf area index (LAI) was attained with 200 and 175 kg N ha −1 and was lowest in the control treatment. Increasing nitrogen extended maturity duration because of prolonged vegetative growth ( Table 6).
Number of spikelets panicle −1 and 1000-grain weight increased significantly with increasing N to 200 and 175 kg ha −1 (Table 6). Applying 100 kg N ha −1 resulted in 125% increase in grain yield over the control (Fig. 2a). Grain yield was statistically similar when 175, 200 and 150 kg N ha −1 were used, but these yields were significantly higher than when 0, 100 and 125 kg of N were applied. Similar pattern was also observed for straw yield ( Fig. 2a and b). Trend analysis using the curve model "Expert 32" (Danial, 1995) indicated that grain yield progressively increased with increasing N application, peaking at 169 kg ha −1 . Afterwards, grain yield declined with further increase in N application (Fig. 3).

Grain yield of the salt tolerant variety CSR43 as affected by different management packages in on-farm trials
The set of best management practices established for CSR43 in the above trials was then validated in 5 farmers' fields during the wet season of 2014 (Table 2). The selected matching management practices (Mmp) included using 4 seedlings per hill at 15 × 20 cm spacing and application of fertilizers in the field at 150-60-40-25 kg N-P 2 O 5 -K 2 O-ZnSO 4 .7H 2 O ha −1 . Grain yield varied significantly between the four treatments (Fig. 5). The highest grain yield of CSR43 (Iv) was when using the new package of Mmp, which resulted in 8% higher yield than the treatment using existing best management practices (Bmp) recommended for the region (Table 2). Moreover, the combination of IvMmp treatment resulted in 35% and 16% higher grain yield over FvFm and IvFm, respectively. Ganga Kaveri, Moti and Narendra359 rice varieties were used to represent farmers' varieties (Fv), and neither produced higher yields mainly because of their relatively lower tolerance of salt stress and poor management practices followed by farmers when using FvFm (Table 2). However, under the same farmers' management (Fm), yield of CSR43 was 17% higher than the other varieties (IvFm vs FvFm) (Fig. 5).

Selection and release of CSR43 as salt tolerant variety through FPVS
The FPVS program conducted in sodic soils (Paris et al., 2011) has confirmed the relevance and utility of participatory approaches in these less favorable environments. On-station screening of a large number of genotypes helped to narrow the number of breeding lines suitable for sodic soils. Studies on farmers' preferences (Table 3) revealed that, farmers generally preferred salt tolerant, short duration varieties that are suited for the rice: wheat or rice: legume: wheat rotations. Moreover, the shorter stature of CSR-89IR-8 made it relatively tolerant to lodging at maturity, causing less yield losses. Based on farmers' ranking, CSR-89IR-8 was most preferred. These attributes together culminated in the release of this breeding line as CSR43 in 2011 (Singh et al., 2013). The early maturity of CSR43 further saves two irrigations compared with other medium to longer duration varieties (Singh et al., 2014); this saving of irrigation water increased BCR and early harvest can benefit subsequent rabi season crops. Singh et al. (2014) further observed that in the selected villages where FPVS was conducted, the adoption of CSR43 among farmers increased remarkably year after year. In contrast, the adoption rate of a popular variety 'Naren-dra359', which has been recommended for normal high yielding environments in the region, declined by one fourth over the previous year in sodic areas (Singh et al., 2014).

Developing matching management options
4.2.1. Effect of seedlings hill −1 and spacing during transplanting on crop growth and yield Grain yield was significantly enhanced with transplanting of four seedlings hill −1 under medium spacing (15 × 20 cm). Higher number of seedlings hill −1 resulted in significantly higher dry matter production and more productive tillers hill −1 , whereas medium spacing enhanced floret fertility resulting in higher grain yield over other treatments. The production of more tillers in widely spaced plants is usually explained by the availability of more nutrients, moisture and sunlight as compared with dense planting. This was also reported by Shieh (1979); Ayub et al. (1987);Miah et al. (1990) and Prasad et al. (1992). Karmakar et al. (2002) observed that panicles became gradually shorter with increasing number of seedlings hill −1 . These results indicate the need for optimizing crop management practices such as number of seedlings hill −1 and spacing between hills to attain optimum plant growth, canopy development and yield formation for specific soil conditions and variety.

Effect of nitrogen applied in the field on crop growth and yield of transplanted CSR43
N application is necessary to maintain plant growth and enhance grain yield. The highest dose of 200 kg N ha −1 produced the tallest plants and the highest number of productive tillers. Behra (1998) and Dwivedi et al. (2000) also reported significant increase in plant height and number of tillers with increasing N application. Similar beneficial effects of N application in similar environments were reported before (Mandal et al., 1994;Singh et al., 1990;Assy et al., 1992;Singh and Sharma, 1993;Chaturvedi, 2005;Manzoor et al., 2006;Nawaz, 2002). Grain yield of CSR43 responded positively to N application up to 169 kg ha −1 (Fig. 3), however, gross margin and BCR were highest with 150 kg N ha −1 (Fig. 4; Table 7). This N rate can therefore, be recommended for CSR43 in these sodic soils. This is 25% higher than the recommended dose of N (120 kg N ha −1 ) for farmers in this region. Responses of rice to higher N in sodic soils was also reported before (Singh and Sharma, 1984;Singh, 1987;Dixit and Patro, 1994;Meena et al., 2003).

On-farm evaluation of CSR43 using matching management options
CSR43 had on average, a yield advantage of 0.62 t ha −1 over farmers' varieties across locations (range of 0.39-0.82 t ha −1 ). However, combined effect of improved variety with matching management practices resulted in a 35% increase in grain yield over farmers' current practices. The average yield enhancement due to these best management practices across locations was 16%. Studies by Ismail et al. (2012) highlighted the benefits of combining good management practices with genetic tolerance for better crop establishment in less favorable areas, as tolerant genotypes respond better to these agronomic manipulations than sensitive ones. Purushottam et al. (2012) reported that grain yield of a rice-wheat system in on-farm trials did not increase significantly with the use of improved varieties alone, but a significant increase over farmers' practices was observed when improved varieties were combined with line sowing/transplanting. Sarangi et al. (2015) also reported higher yields when farmers use improved varieties with improved nursery management practices. Our study highlighted the importance of developing and validating management options for new salt tolerant varieties to maximize their attainable yields in farmers' fields.

Conclusions
CSR43 produced about 0.5 t ha −1 additional grain yields over farmers' current varieties across locations. This yield advantage, together with tolerance of sodicity, short duration, good grain quality and market value contributed to the selection and commercialization of this variety and to its faster adoption in the sodic soils of the Indo-Gangetic plains (Singh et al., 2014). Matching crop establishment and nitrogen rates were developed for this variety, and when combined, resulted in significantly higher yields and cost/benefit ratios compared with current varieties and agronomic practices used by local farmers. Suggested recommendations include transplanting of 4 seedlings hill −1 at a spacing of 15 × 20 cm and use of 150-60-40-25 kg N-P 2 O 5 -K 2 O-ZnSO 4 ·7H 2 O ha −1 in the field. On-farm validation trials of these practices using CSR43 showed that the matching management practices alone increased grain yield by 8% over existing recommendations and by 16% over management strategies being practiced by local farmers. This supports our hypothesis that matching crop and nutrient management practices are essential for harnessing the full potential of a new stress-tolerant variety in targeted conditions. These cost effective management options for sodic soils can further be refined to match other new tolerant varieties as they become available. The benefits can then be extended to larger areas including the salt-affected areas of the Ganges mega deltas in India and Bangladesh, in addition to the Indo-Gangetic plains of Eastern Uttar Pradesh and Bihar in India.
Agricultural Development (EC-IFAD) for supporting the development of agronomic practices for stress tolerant varieties developed and disseminated through the STRASA (Stress tolerant rice varieties for Africa and South Asia) project funded by the Bill and Melinda Gates Foundation. We are thankful to the Director of the ICAR-CSSRI, Karnal, Haryana, India for providing facilities to carry out these studies.