Relative significance of soil and nitrogenous fertilizer in nitrogen nutrition and growth of wetland rice (Oryza sativa L.)

Summary A pot experiment was conducted to compare the yields from five commercially cultivated varieties (Bas-198, Bas-370, Bas-Pak, Bas-385, and IR-6) of rice (Oryza sativa L.) and to establish the relative significance of soil N and fertilizer N (¹⁵N-labelled ammonium sulphate) in affecting crop performance. Another aim was to study the interaction of fertilizer N and soil N as influenced by different rice varieties. Among the five varieties tested, Bas-Pak gave the maximum dry matter and N yield. The N-use efficiency (percentage of applied N taken up by the plants) of different varieties ranged between 33.7 and 43.7%, Bas-Pak being the most efficient. Significant losses of fertilizer N occurred from the soil-plant system. The maximum N loss (52.1% of applied N) was observed with IR-6 and the minimum loss (39.2%) with Bas-Pak. A substantial increase in the uptake of soil N following the application of fertilizer and an interaction between the two N sources were observed with all varieties except Bas-385. The interaction was attributed to greater root proliferation following the application of fertilizer. It was concluded that a varietey with greater potential to use soil N is likely to give a better yield and that, of the two N sources, the availability of soil N was more important in determining the yield performance of different varieties of rice.     

Fate and interaction with native soil N of ammonium N applied to wetland rice (Oryza sativa L.) grown under saline and non-saline conditions

Summary The effect of salts on the balance of fertilizer N applied as ¹⁵N-labelled ammonium sulphate and its interaction with native soil N was studied in a pot experiment using rice (Oryza sativa L.) as a test crop. The rice crop used 26%–40% of the applied N, the level of applied N and salts showing no significant bearing on the uptake of fertilizer N. Losses of fertilizer N ranged between 54% and 68% and only 5%–8% of the N was immobilized in soil organic matter. Neither the salts nor the rate of N application had any significant effect on fertilizer N immobilization. The effective use of fertilizer N (fertilizer N in grain/fertilizer N in whole plant) was, however, better in the non-saline soil. The uptake of unlabelled N (N mineralized from soil organic matter and that originating from biological N₂ fixation in thes rhizosphere) was inhibited in the presence of the salts. However, in fertilized soil, the uptake of unlabelled N was significantly enhanced, leading to increased A values [(1-% Ndff/% Ndff)x N fertilizer applied, where Ndff is N derived from fertilizer], an index of interaction with the added N. This added N interaction increased with increasing levels of added N. Since the extra unlabelled N taken up by fertilized plants was greater than the fertilizer N immobilized, and the root biomass increased with increasing levels of added N, a greater part of the added N interaction was considered to be real, any contribution by an apparent N interaction (pool substitution or isotopic displacement) to the total calculated N interaction being fairly small. Under saline conditions, for the same level of fertilizer N addition, the added N interaction was lower, and this was attributed to a lower level of microbial activity, including mineralization of native soil N, rootdriven immobilization of applied N, and N₂ fixation.     

Effect of green manuring,blue-green algae and neem-cake-coated urea on wetland rice (Oryza sativa L.)

Summary A field trial was set up to examine the effect of green manuring, blue-green algae, and neem-cake-coated urea on a rice crop. Summer green manuring using Sesbania aculeata increased the crop yield. Inoculation of blue-green algae increased the rice grain yield when 60 kg N ha^-1 was applied as prilled urea, but the increase in grain yield was greater when 60 kg N ha^-1 was applied as neem-cake-coated urea. The results of the present study show that applications of green manure, neem-cake-coated urea, and blue-green algae are complementary and that the three treatments can be used together in the rice ecosystem. The green manure and the fertilizer treatments had no effect on the algal flora of the soil.     

Crop growth and nitrogen transformations in wheat (Triticum aestivum L.) planted after wetland rice (Oryza sativa L.)

Summary A field study was undertaken to examine the effects of various management strategies on wheat (Triticum aestivum L.) performance and N cycling in an intensively cropped soil. Microplots receiving 100 kg N ha^–1 as¹⁵NH₄ ^{+ 15}NO₃ ^– at sowing, tillering or stem elongation were compared with unfertilized microplots. Stubble from the previous rice crop was either incorporated, burnt without tillage, burnt then tilled or retained on the surface of untilled soil. Wheat grain yield ranged from 1.5 to 5.1 t ha^– and was closely related to N uptake. Plant accumulation of soil N averaged 36 kg N ha^–1 (LSD 5% = 10) on stubble-incorporation plots and 54 kg N ha^–1 on stubble-retention plots. Fertilizer N accumulation averaged 18 kg N ha^–1 (LSD 51% = 6) on stubble-incorporation plots and 50 kg N ha^–1 on stubble-retention plots. Tillage had little effect on burnt plots. Delaying N application from sowing until stem elongation increased average fertilizer N uptake from 26 to 39 kg N ha^–1 (LSD 5% = 6), but reduced soil N uptake from 50 to 37 kg N ha^– (LSD 5% = 10).Immobilization and leaching did not vary greatly between treatments and approximately one-third of the fertilizer was immobilized. Less than 1% of the fertilizer was found below a depth of 300 mm. Incorporating 9 t ha^–1 of rice stubble 13 days before wheat sowing reduced net apparent mineralization of native soil N from 37 to 3 kg ha^–1 between tillering and maturity. It also increased apparent denitrification of fertilizer N from an average 34 to 53 kg N ha^–1 (LSD 5% = 6). N loss occurred over several months, suggesting that denitrification was maintained by continued release of metabolizable carbohydrate from the decaying rice stubble. The results demonstrate that no-till systems increase crop yield and use of both fertilizer and soil N in intensive rice-based rotations.     

Selection of efficient vesicular-arbuscular mycorrhizal fungi for wetland rice (Oryza sativa L.)

Summary We tested the response of the wetland rice cultivar Prakash to inoculation with ten vescular-arbucular mycorrhizal (VAM) fungi (three selected from the first screening and seven isolated from local paddy fields) in a pot experiment under flooded conditions in order to select the most efficient mycorrhizal fungi to inoculate the rice nursery. A sandy clay loam soil was used as the substrate, fertilized with the recommended N and K levels (100 kg N ha^–1 as ammonium sulphate and 50 kg K ha^–1 as muriate of potash) and half the recommended level of P (25 kg ha^–1 as super phosphate). The inoculation was made into dry nursery beds and the beds were flooded when the seedlings were about 25 cm high, in 15 days. Twenty-eight-day old seedlings were transferred to pots filled with well puddled soil flooded with 5 cm of standing water. Based on the increase in grain yield and total biomass, Glomus intraradices and Acaulospora sp. were considered efficient and suitable for inoculation into rice nurseries.     

Effect of methods of application on the efficiency of urea broadcast onto lowland rice (Oryza sativa L.)

Summary We evaluated the effect of different methods of application on the efficiency of urea broadcast at a rate of 100 kg N ha^-1 onto lowland rice (Oryza sativa L. var. SPR 60) in a field experiment conducted on a Phimai soil (Fluvic Tropaquepts) during the dry season of 1989. Analysis of the floodwater on the first day after the fertilizer application showed a high initial concentration of urea-N. Addition of the urease inhibitor N-(n-butyl) thiophosphoric triamide (nBTPT), broadcast with the urea into the floodwater, caused an apparent reduction in the rate of urea disappearance and a subsequent accumulation of NH₃–N in the floodwater; this ureas inhibitor also suppressed the rise in floodwater pH, with a resultant reduction in the partial pressure of ammonia (pNH₃) compared with the unamended urea application. The use of nBTPT did not decrease the N loss from broadcast urea not did it increase the grain yield. Among the different methods of applying broadcast urea that we tested, the broadcast application of granular urea onto drained soil shortly after removing floodwater followed by flooding 2 days later appeared to be a good N management practice, offering considerable potential for improving the efficiency of urea applied to lowland rice crops.     

Effects of inorganic fertilizers and micronutrients on methane production from wetland rice (Oryza sativa L.)

We compared the effects of adding different forms of nitrogenous fertilizers on the production of CH₄ in soil and on CH₄ emission from rice plants, Urea and diammonium phosphate gave the highest rates of CH₄ production from the soil and emission through rice plants, followed by (NH₄)₂SO₄. NaNO₃ was the least effective. The effects of micronutrients like Mo, Ni, or B were more prominent than those of Fe, Mn, Zn, V, or Co. It is concluded that CH₄ emission from rice paddies is influenced by both macro- and micronutrients, through effects on both microbial methanogenesis in soil and elimination through rice plants as a consequence of the effects on plant growth.     

N2 fixation in two Sesbania species and its transfer to rice (Oryza sativa L.) as revealed by 15N technology

Summary We used ¹⁵N technology to investigate N₂ fixation by Sesbania speciosa and Sesbania rostrata and its transfer to a lowland rice crop after incorporation of the Sesbania spp. into soil as green manure. During the first 50 days after establishment in November–December 1989, S. speciosa and S. rostrata produced 1126 and 923 kg dry matter ha^-1 respectively. They gathered 31 and 23 kg N ha^-1 respectively, of which 62%±5% and 55%±3% respectively, came from N₂ fixation. Both these species produced a greater biomass during September–October 1989, with S. rostrata producing more than S. speciosa. These results reflected differential responses by the plants to different day lengths at different times of the year. Furthermore, the dry matter yield and %N of ¹⁵N-labelled S. speciosa were smaller than those of the unlabelled plants, possibly due to inhibition of N₂ fixation in root nodules by the chemical N fertilizers added during labelling. These differences were not so pronounced in the stem-nodulated S. rostrata. The increased grain yield of rice fertilized with N in the form of chemical fertilizer or green manure was a result of an increased number of panicles per hill. The rice crop manured with S. speciosa produced a lower grain yield, with a lower grain weight than that manured with S. rostrata. This was due to a low uptake of soil N by rice manured with S. speciosa. Recovery of N from the green manure in rice straw with S. speciosa was significantly higher than from rice manured with S. rostrata, because of the higher applied N uptake by rice manured with the former.     

Availability of soil and fertilizer nitrogen to wetland rice following wheat straw amendment

Summary A pot experiment was conducted to study the availability of soil and fertilizer N to wetland rice as influenced by wheat straw amendment (organic amendment) and to establish the relative significance of the two sources in affecting crop yield. Straw was incorporated in soil at 0.1, 0.2, and 0.3% before transplanting rice. Inorganic N as ¹⁵N-ammonium sulphate was applied at 30, 60, and 90 g g^-1 soil either alone or together with wheat straw in different combinations. After harvesting the rice, the plant and soil samples were analyzed for total N and ¹⁵N. Straw incorporation significantly decreased the dry matter and N yield of rice, the decrease being greater with higher rates of straw. The reduction in crop yield following the straw incorporation was attributed mainly to a decrease in the uptake of soil N rather than fertilizer N. The harmful effects of organic matter amendment were mitigated by higher levels of mineral N addition. The uptake of applied N increased and its losses decreased due to the straw incorporation. Mineral N applied alone or together with organic amendment substantially increased the uptake of unlabelled soil N. The increase was attributed to a real added N interaction.     

Areal variation in the relationship between natural 15N abundances of rice (Oryza sativa L.) and soil

This study was performed to clarify whether areal variation exists in the relationship between natural ¹⁵N abundances (δ¹⁵N values) of rice (Oryza sativa L.) and soil without an applied nitrogen (N) source, and to explore possible reasons for any areal variation. We investigated the relationships between δ¹⁵N values of rice and those of unamended soil with no applied N source in two locations; Daisen and Ogata, in Akita Prefecture, Japan. The δ¹⁵N values of rice in Daisen were higher than those in Ogata from 2007 to 2009, irrespective of the cropping year. Results demonstrated areal variation in the relationship between δ¹⁵N values of rice and those of unamended soil. The variation might be attributed to variation in the δ¹⁵N of natural N input and to ammonia nitrification and subsequent denitrification. When the relationship between δ¹⁵N values of rice and those of unamended soil is used to discriminate between organic and conventional rice, the areal variation of the relationship in the target area should be taken into account, from the point of the δ¹⁵N value of natural N input and N transformation in the soil.     

中国亚热带红壤旱坡地花生系统尿酸氮素去向

A field experiment was conducted to study the outcome of N using isotope-labeled urea with a peanut cropping system in a Udic Ferrosol on a 7% slope at the Ecological Experimental Station of Red Soil, Chinese Academy of Sciences. The micro-plots were designed in two sets with three replicates and four N treatments. An iron frame with its edge 10 cm above the ground was used to control soil erosion and runoff in set A, but in set B the upper edge of the frame was level with the ground. Randomly positioned ^15N fertilized micro-plots were established along the contour. At harvest on August 13, 2002, which was four months after urea application, the peanuts had absorbed 30.6%-38.0% of the urea, transpired 13.1%-22.4% to the atmosphere, lost 6.4%-8.4% through soil erosion and surface runoff, and maintained 40.7%-48.9% in the 0-60 cm soil layer with 0.5%-5.0% below 60 cm.     

Translocation and recovery of 15N-labelled N derived from foliar uptake of 15NH3 by rice (Oryza sativa L.) cultivars

Foliar uptake of ¹⁵NH₃ applied at two growth stages (tillering and anthesis) and the subsequent ¹⁵N-labelled vegetative-N distribution in different plant components at maturity was investigated in three rice cultivars, IR-6, NIAB-6 and Bas-385. Rice plants absorbed 22–30% and 18–24% of the ¹⁵NH₃ applied at tillering and anthesis stages, respectively. Of the total ¹⁵NH₃ absorbed at tillering stage, IR-6 and Bas-385 showed higher recovery (71%) in different plant components at maturity as compared to NIAB-6 (48% recovery). At maturity, percent recovery of the ¹⁵NH₃ absorbed at anthesis stage was almost comparable in different cultivars, but it was lower (46–55%) than that absorbed at the tillering stage. Recovery of the absorbed ¹⁵NH₃-N in the soil was negligible and ranged from 0.3–1%. At maturity, the cultivars IR-6 and Bas-385 showed a higher loss (45–53%) of ¹⁵NH₃ absorbed at anthesis than at the tillering stage (29% loss), whereas for NIAB-6, the corresponding figures were comparable for the two growth stages (tillering, 51% loss; anthesis, 49% loss). Results indicated a variable potential of the tested rice cultivars for foliar uptake of atmospheric ¹⁵NH₃ and distribution of ¹⁵N-labelled vegetative-N in different plant components.     

Effects of inoculation with N2-fixing cyanobacteria on the nitrogenase activity in soil and rhizosphere of wetland rice (Oryza sativa L.)

Summary A greenhouse experiment was conducted with wetland rice (Oryza sativa cv. IR-50) in a clay-loam soil (Fluventic Eutrochrept) to study the effect of cyanobacterial inoculation a mixed culture of Aulosira fertilissima, Nostoc muscorum, Nostoc spp., and Anabaena spp., applied at the rate of 0.15 g (dry weight pot^-1 or 43 kg ha^-1) on acetylene reduction activity in soil and the root system (excised root), and the grain and straw yield. The effects of applying P (40 kg ha^-1), N (60 kg ha^-1), and P+N to the soil were also evaluated. Cyanbacterial inoculation significantly increased (more than 200% on average) photo-dependent acetylene reduction activity in soils, particularly where the indigenous activity was considerably low, i.e. under unfertilized and N-fertilized conditions. The effect of inoculation was prominent at the maximum tillering and grain formation stages of the crop. This inoculation benefit was, however, marginal in P-applied soils (P and P+N), where the indigenous activity was stimulated more than threefold. The inoculation led to a remarkable increase in root-associative acetylene reduction activity after the maximum tillering stage of the crop, particularly with applied N but for other treatments this inoculation effect was not significant. Cyanobacterial inoculation also increased the grain and straw yield of the crop when N was not applied. The grain and straw yield was significantly correlated with the acetylene reduction activity in flooded soils and in the root system during the tillering and maximum tillering stages of rice growth, respectively.     

Sources of nitrogen and yield advantages for monocropping and mixed cropping with cowpeas (Vigna unguiculata L.) and upland rice (Oryza sativa L.)

Summary Sources of N used by cropped and intercropped cowpeas and rice were determined by the ¹⁵N isotope-dilution technique. The biological efficiency of intercropping cowpeas and rice was assessed by calculating the land equivalent ratio of dry matter yield, total N, and uptake of N. A reduced N uptake by both cowpeas and rice during mixed cropping was attributed to mutual competition, with both crops competing effectively for the scarce environmental resources. The lack of a significant difference in the uptake of fertilizer and soil N by mixed-crop rice and cowpeas is an indication that the soil N was sufficient and that the mixed cropping did not create any imbalance in soil and fertilizer N uptake. The land equivalent ratio ranged between 120% and 180% for shoot dry matter and total N, showing that biological efficiency was increased by intercropping cowpeas with rice. The proportion of N derived from the atmosphere by mono- and intercropped cowpeas was not significantly different, showing that the potential of cowpeas to fix N₂ was independent of the cropping system, since the rice did not stimulate the cowpeas to fix more N₂.     

Shoot,root, soil and microbial nitrogen dynamics in two contrasting soils cropped to barley (Hordeum vulgare L.)

Summary Dynamics of barley N, mineral N, and organic N were compared at Ellerslie (Black Chernozem, Typic Cryoboroll) and Breton (Gray Luvisol, Typic Cryoboralf) in central Alberta, using ¹⁵N-urea. On average, shoot N and shoot ¹⁵N recoveries at Ellerslie (14.1 g m^–2, 36%) were greater than at Breton (4.5 g m^–2, 17%). Root N (g m^–2) did not significantly differ between sites (0–30 cm) but root ¹⁵N recovery was greater at Breton (3.4%) than Ellerslie (1.8%). Low levels of shoot N and shoot ¹⁵N at Breton were partly due to very wet soil conditions in July, which resulted in premature shoot senescence and low plant N uptake. Although the total ¹⁵N recoveries from the system (to 30 cm depth) at Ellerslie (63%) and Breton (56%) were similar, soil ¹⁵N was greater at Breton (35%) than at Ellerslie (26%). There were no differences in mineral N between sites but the average ¹⁵N recovery in the mineral-N pool was significantly greater at Ellerslie (3.3%) than at Breton (1.6%). There was no difference in ¹⁵N recovery in the microbial biomass (3%) between sites, although non-microbial organic ¹⁵N was greater at Breton (31 %) than at Ellerslie (20%). The two soils showed differences in the relative size of kinetically active N pools and in relative mineralization rates. Microbial N (0–30 cm) was greater at Ellerslie (13.3 g m^–2) than at Breton (9.9 g m^–2), but total microbial N made up a larger proportion of total soil N at Breton (1.6%) than at Ellerslie (0.9%). In the 0–10 cm interval, microbial N was 1.7-fold greater and non-microbial active N was 3-fold greater at Breton compared to Ellerslie, when expressed as a proportion of total soil N. Net N mineralization in a 10-day laboratory incubation was 1.4-fold greater in the Black Chernozem (0–10 cm interval) from Ellerslie, compared to the Gray Luvisol from Breton, when expressed per gram of soil. Net N mineralization in the soil from Breton was double that of the soil from Ellerslie, when expressed as a proportion of soil N. Although soil N (g m^–2) was 2.5-fold greater at Ellerslie compared to Breton, it was cycled more rapidly at Breton.     

The effect of Azospirillum brasilense Sp7 and Azotobacter chroococcum on nitrogen balance in soil under cropping with oats (Avena sativa L.)

Summary Pot experiments with oats were carried out to study the effect of Azospirillum brasilense Sp 7 and Azotobacter chroococcum 94K on the yield of plants, the N content of soil and the ¹⁴N balance. The plants were grown on gray forest soil under irrigation with deionized water and application of ¹⁵N-labelled fertilizer at a rate of 4 mg N 100 g^-1 soil. Inoculation of plants with Azospirillum spp. and Azotobacter spp. failed to increase the plant yield. However, the increase in total N in the soil at the end of the experiment and the positive ¹⁴N balance in the soil-plant system due to increased nitrogenase activity in the rhizosphere were statistically significant. The amount of N accumulated in the soil was comparable with the rate of N applied as fertilizer.     

Effects of 15N-labelled ammonium nitrate and urea on soil nitrogen during growth of wheat (Triticum aestivum L.) under field conditions

We studied the effects of ¹⁵N-labelled ammonium nitrate and urea on the yield and uptake of labelled and unlabelled N by wheat (Triticum aestivum L., cv. Mexi-Pak-65) in a field experiment. The dry matter and N yields were significantly increased with fertilizer N application compared to those from unfertilized soil. The wheat crop used 33.6–51.5 and 30.5–40.9% of the N from ammonium nitrate and urea, respectively. Splitting the fertilizer N application had a significant effect on the uptake of fertilizer N by the wheat. The fertilizer N uptake showed that ammonium nitrate was a more available source of N for wheat than urea. The effective use of fertilizer N (ratio of fertilizer N in grain to fertilizer N in whole plant) was statistically similar for the two N fertilizers. The application of fertilizer N increased the uptake of unlabelled soil N by wheat, a result attributed to a positive added N interaction, which varied according to the fertilizer N split; six split applications gave the highest added N interaction compared to a single application or two split applications for both fertilizers. Ammonium nitrate gave 90.5, 33.5, and 48.5% more added N interaction than urea with one, two, and six split N applications. A values were not significantly correlated with the added N interaction (r=0.557). The observed added N interaction may have been the result of pool substitution, whereby added labelled fertilizer N replaced unlabelled soil N.     

Availability of soil and fertilizer nitrogen to wheat (Triticum aestivum L.) following rice-straw amendment

Summary A pot experiment was conducted to study the N availability to wheat and the loss of ¹⁵N-labelled fertilizer N as affected by the rate of rice-straw applied. The availability of soil N was also studied. The straw was incorporated in the soil 2 or 4 weeks before a sowing of wheat and allowed to decompose at a moisture content of 60% or 200% of the water-holding capacity. The wheat plants were harvested at maturity and the roots, straw, and grains were analysed for total N and ¹⁵N. The soil was analysed for total N and ¹⁵N after the harvest to determine the recovery of fertilizer N in the soil-plant system and assess its loss. The dry matter and N yields of wheat were significantly retarded in the soil amended with rice straw. The availability of soil N to wheat was significantly reduced due to the straw application, particularly at high moisture levels during pre-incubation, and was assumed to cause a reduction in the dry matter and N yields of wheat. A significant correlation (r=0.89) was observed between the uptake of soil N and the dry matter yield of wheat with different treatments. In unamended soil 31.44% of the fertilizer N was taken up by the wheat plants while 41.08% of fertilizer N was lost. The plant recovery of fertilizer N from the amended soil averaged 30.78% and the losses averaged 45.55%     

Beneficial effects of blue-green algae and Azolla, excluding supplying nitrogen, on wetland rice fields: a review

 The role of blue-green algae (BGA) and Azolla in supplying N to rice fields is well documented. In addition, they also bring about, directly or indirectly, a number of changes in the physical, chemical and biological properties of the soil and soil-water interface in rice fields. For example, BGA liberate extracellular organic compounds and photosynthetic O₂ during their growth, while Azolla prevent a rise in the pH, reduce water temperature, curb NH₃ volatilisation and suppress weeds; and both of them contribute biomass. On decomposing, they influence the redox activity and result in the formation of different organic acids in soil. All such changes brought about by BGA and Azolla in soil may ultimately influence plant-available nutrients and also soil characteristics. An attempt has been made in this review to highlight these effects of BGA and Azolla in rice fields and discuss their possible implications relating to management and productivity of rice-field systems. Received: 4 February 1998     

Yield and nitrogen response of lowland rice (Oryza sativa L.) to Sesbania rostrata and Aeschynomene afraspera green manure in different marginally productive soils in the Philippines

Organic-N fertilizers in the form of flood-tolerant, leguminous, stem-nodulating Sesbania rostrata and Aeschynomene afraspera may be useful alternatives to resource-poor rice farmers if applied as green manure. Therefore, the accumulation of N by these green manure species and their effect on the performance and yield of wetland rice (IR 64) was examined at four different sites in Luzon, Philippines. Soils deficient in N, P, and K were selected and compared with the fertile Maahas clay of the International Rice Research Institute (IRRI) at Los Baños. The green manure plants were grown under flooded conditions for 49 days in the wet season of 1987, chopped, and then ploughed in before transplanting rice seedlings. In a second experiment, the effect of S. rostrata green manure was studied under rainfed conditions. All green manure treatments were compared to an urea treatment (60 kg N ha^–1) and an untreated control. Both legumes developed well, even on the marginally productive soils. S. rostrata accumulated up to 190 kg N ha^–1 and A. afraspera even accumulated 196 kg N ha^–1 in the shoots. In all treatments, green manure increased grain yield significantly (P=0.05) over the untreated control, by 1.3–1.7 Mg ha^–1. The yields were comparable to those obtained with 60 kg N ha^–1 of urea fertilizer. S. rostrata caused the highest grain yield, of 6.5 Mg ha^–1 on the Maahas clay soil of IRRI. The apparent release of exchangeable NH ₄ ⁺ -N in the soils after green manuring and the rice grain yield response showed that both green manure species may provide sufficient available N throughout the development of IR 64 in the wet season. In the rainfed marginal soil site, green manure with S. rostrata produced even higher rice grain yields than urea. Green manure therefore seems particularly attractive for poor farmers on marginally productive soils, at least as a temporary strategy to improve yield and yield sustainability.