SCOPE 48 - Sulphur Cycling on the Continents

11.1 INTRODUCTION

Most Asian countries have adopted intensified rice-based cropping systems to increase food production for a rapidly growing population. Crop intensification and increased yield have significantly increased the output of nutrients and, where there has been an imbalance between outputs and inputs, has resulted in declining soil fertility and an increase in the incidence of deficiencies of certain plant nutrients, including sulphur.

To understand the cycling of sulphur in rice wetlands, it is essential that the geographical distribution and characteristics of these wetlands be appreciated, along with the important biological, physical, and chemical changes which occur as a result of flooding and which influence the transformation and thus the cycling of sulphur. Many of the characteristics of rice wetlands and the biological, physical, and chemical processes which occur in them are similar to those of natural wetlands (Giblin and Wieder, this volume) . However, there are some significant differences which largely result from their distribution and the management practices employed. For instance, land preparation for rice involves 'puddling' , a management practice aimed at establishing an impermeable layer at lower soil depths while thoroughly mixing the upper soil layers to provide a uniform seed bed and sufficient rooting depth. In many areas, the heavy texture of the soil makes cultivation difficult during the dry season, and puddling starts after the onset of flooding.

11.2 RICE WETLAND DISTRIBUTION

Wetlands have been defined as areas where saturation with water is the dominant factor determining the nature of soil development and the types of plant and animal communities (Guthrie, 1985). On this basis the US Department of Agriculture has classified about 10% of the world's total land area as wetlands. These range from frozen wetlands in Alaska, Canada, and the USSR, to the coastal and alluvial plains of the temperate and tropical zones.

In their undrained state, these wetlands support only a limited number of agriculturally important species. Sago (Metroxylon spp.) and taro (Colocasia spp.) are two species which are of local importance in limited areas, but rice (Oryza sativa) is the only major food crop of wetland soils. It has been estimated that rice constitutes nearly 20% of the world's food grain production (Stangel, 1979) and that 90% of this production takes place in monsoonal Asia. Kyuma (1988) suggested that the combination of physiography and monsoonal climate of the region has resulted in the large areas which are suitable for rice cultivation. These areas either receive sufficient rainfall during the monsoonal periods or, in the case of the alluvial plains and deltas of the major continental rivers and the broad coastal lowlands of the islands, receive enough water in addition to rainfall as a result of their extended catchment areas.

Rice is grown under a range of hydrological regimes from the unflooded rainfed, through the rainfed and irrigated flooded, to the deepwater and floating cultivation systems. Neue and Bloom (1989) outlined a comprehensive classification of rice environments by applying water source and floodwater depth as diagnostic criteria. Further differentiation is made by modifiers related to climate, landforms, field forms, floodwater, and soil and cropping systems. Greenland and De Datta (1985) estimated that of the 136 X 10⁶ ha planted with rice in 1980, just over 50% was irrigated (Table 11.1). A further 11% was rainfed and considered hydrologically favourable for rice production, while the remaining 38% consisted of rainfed submerged, dryland, deepwater, or floating rice cultivation, which were susceptible to hydrological problems of drought or excessive submergence. Man has significantly altered the hydrology of many of these rice wetlands by draining, terracing, embanking, and irrigating.

It is clear that the wetlands used for rice production differ from a large number of the natural wetlands, not only in their geographical distribution with respect to latitude and altitude, but also in their hydrology. Most soils used for rice cultivation go through at least one natural or management imposed annual wetting and drying cycle, with the dry period ranging from several weeks to eight months. Even those rice soils which are only drained for a short period, and thus may remain saturated, normally receive regular inundation with oxygenated surface water. In contrast, many natural wetlands are continuously flooded, or have relatively short dry periods (Giblin and Wieder, this volume) .Those natural wetlands which are not continuously flooded, such as inter-tidal salt marshes, differ from most rice wetlands in characteristics other than their hydrology, for instance in their high organic matter content.

Table 11.1. Areas of world rice production on the basis of hydrological regime, climatic zone, and cropping system (from Greenland and De Datta, 1985)

11.3 SOIL CHARACTERISTICS

11.3.1 SOIL TYPES 

Within the various geographical and hydrological constraints discussed above, rice is grown on a wide range of soils, from all the 10 Soil Taxonomy Orders (USDA, 1975). Those of major importance are the Alfisols, Entisols, Inceptisols, and Ultisols (Moormann, 1978). Soils which are highly weathered and high sorbing, very light textured or organic/peaty soils are seldom of major importance. In contrast many natural wetland soils have organic-C contents of from 5 to 80% .

11.3.2 CHANGES WITH FLOODING 

The flooding of a soil sets in motion a series of physical, chemical, and biological processes that have a very marked effect on the properties of soils, and therefore on the ability to support plant growth. Many of the changes that take place during flooding have already been dealt with in a previous chapter (Giblin and Wieder, this volume) and need not be dealt with in great detail here. The limited oxygen diffusion into the soil, the supply of biodegradable C, the level of other oxidants, and the biological activity in the soil, particularly in the root rhizosphere, are the key to most of the biochemical and chemical processes in flooded soils.

Early laboratory work suggested that the reduction of substances in flooded soils should proceed stepwise in a sequence based on thermodynamic considerations (Ponnamperuma, 1972; Patrick and Reddy, 1978), but such thermodynamic considerations are only partly applicable under field conditions (Neue, 1988). Since most oxidation ^_ reduction reactions in soils are mediated by microorganisms, their presence, population dynamics, and physiological activity are further controllers of the rate and order of the reactions. A thermodynamically possible reaction may not occur due to substrate or pH effects on the mediating microorganism. In addition, an immediate effect of flooding and puddling a soil is to disrupt soil aggregates, disperse soil colloids, and, depending on the floodwater being used, dilute the soil solution. These changes increase the area of reactive surfaces, and thus change the ion exchange and pH buffering capacities (Samosir, Blair and Lefroy, 1988).

The combined effect of pH, oxidation ^_ reduction potential, microbial activity, and reactant and product concentrations is that the sequence of oxidation ^_ reduction reactions does not entirely follow thermodynamic consideration but is set by the kinetics of the reactions. Consequently, within a soil a range of reactions can be occurring at the same bulk soil oxidation ^_ reduction potential. Further, measured E_h potentials correspond only loosely to actual in situ oxidation ^_ reduction potentials (Neue, 1988).

11.3.3 HETEROGENEITY WITHIN RICE WETLANDS 

The heterogeneity of a flooded soil, in terms of its redox potential and the concomitant chemical characteristics, is a result of initial heterogeneity in its mineral phase, differential diffusion and mixing of the various redox reaction substrates and products and differences in the distribution of soil organisms, which will affect the chemistry and physical mixing of the soil. One of the most obvious and important aspects of this heterogeneity is the distribution of oxygen. While diffusion of oxygen in water is much lower than in air , significant diffusion into the surface soil does occur; this is further enhanced by the activity of soil microfauna and macrofauna (Roger and Kurihara, 1988). The result is that in a flooded rice field, there is a large range in the oxygen tension of different parts of the system,  from an often highly oxidized floodwater and soil surface, to highly reduced subsoil layers. These properties are not fixed but have large diurnal and seasonal fluctuations.

11.3.4 THE EFFECT OF THE RICE PLANT 

A further level of heterogeneity is introduced by the plant. Rice roots have well developed aerenchyma which have been shown to enlarge in response to reduced aeration (Katayama, 1961; Armstrong, 1971). The aerenchymatous nature of rice roots enables them to transfer oxygen to the roots from the shoots (van Raalte, 1940; Barber, Ebert and Evans, 1962; John, Limpinuntana and Greenway, 1974). This capacity for rapid diffusion of oxygen through the extracellular spaces of shoots and roots is shared by many plants which are adapted to natural wetlands (Armstrong, 1964). Transfer of oxygen to the roots allows the more efficient oxidative phosphorylation to continue in root cells and also results in the oxidation of the root rhizosphere (Armstrong, 1971). Yu (1985) has reported higher oxygen contents and higher E_h at, or near, the root surface than in soil away from the roots. The spatial difference in E_h as a result of rice root distribution is also reflected in lower levels of reducing substances in the rhizosphere. Yu (1985) cites work by Lui and Yu (1962) which showed an inverse correlation between planting density and concentrations of ferrous iron, manganous manganese, and reduced organic substances at a range of growth stages.

The root also provides substrates for microorganisms in the form of root exudates, changes rhizosphere pH as a result of proton pump activity, and alters soil solution concentrations as a result of nutrient uptake. The result is that within the vertical stratification of oxygenation in a flooded soil, from the oxidized floodwater and soil surface to the reduced subsoil, roots provide a horizontal and vertical discontinuity in oxygen tension and other factors. The nature of these rhizosphere conditions will not only stimulate microbial activity, but also will change the population dynamics of soil microorganisms by providing a niche for aerobes and a change in the physiology of facultative anaerobes.

11.4 SULPHUR CYCLING IN RICE WETLAND SOILS 

For many nutrients, the balance between ionic forms, as affected by redox potential and pH, changes the solubility and strength of adsorption/exchange and therefore the amount of that nutrient available for plant uptake. In the case of nitrogen, the ionic form also changes the uptake mechanism. Sulphur , however, is taken up by the roots of most plants in the oxidized sulphate form. There is evidence that some wetland plant species, such as the saltmarsh grass Spartina alterniflora can also assimilate sulphide (Carlson and Forrest, 1982; Peterson, Howarth and Garritt, 1986). It is unknown whether or not rice also can assimilate sulphide, but we suspect it is unlikely to be a major source of sulphur to the plant since rice is sensitive to low levels of sulphide in its tissues. The hydrology, chemistry , biology, and physical characteristics of rice wetlands, as discussed above, are all important in sulphur cycling in flooded rice systems as they affect the form of and availability of S to the rice plant.

The level of total S and sulphate-S in wetland rice soils varies widely. The mean concentrations reported for total S in rice wetland soils of a number of different countries range from 83 to 1176 ĩg (S) g^-l dry weight (2.6 to 37 ĩmol g^-l) , while the mean sulphate-S levels range from 2 to 210 ĩg (S) g^-l dry weight (0.06 to 6.6 ĩmol g^-l) (Table 11.2). Generally, the S content of rice wetlands is related to soil texture, with light textured soils containing less S than clay soils (Parkpian et al., 1986). The total S content is also correlated to total organic matter content, as has been reported for a large number of Philippine soils (Reyes et al., 1985), with a significant fraction of the total S in the soil being present as organic sulphur (Blair and Lefroy, 1987). The same is generally true of natural freshwater wetlands; their normally higher organic matter content being correlated with higher total sulphur content (Giblin and Wieder, this volume).

Organic sulphur compounds in soil have been grouped on the basis of their chemical bonding or reactivity with certain reducing agents (Tabatabai, 1982). Very few analyses of the groups of organic S in rice wetland soils have been reported, although some measurements of the amount of organic S reduced to H₂S by hydriodic acid have been made; this form of organic S is not bonded directly to C and is believed to be largely in the form of ester sulphates. Furusaka and Freney (cited by Freney, Jacq and Baldensperger, 1982) reported that the mean HI-reducible S in some lowland soils in Japan was 50% of total organic S, while the mean of 12 wetland rice soils in the Philippines was 36% of total organic S (Gaddi, 1988; Cacnio and Mamaril unpub. data). These values are within the range reported for HI-reducible organic S in upland soils of between 30 and 70% of organic S (Freney, Jacq and Baldensperger, 1982). The form of S in flooded soils is dependent on the balance between the various S transformations that can occur in flooded soils. Freney et al. have divided these transformations into five categories, namely:

Table 11.2. Total sulphur and sulphate sulphur contents of wetland rice soils from different countries

(1) Immobilization or assimilation of S into organic compounds
(2) Mineralization or decomposition of organic S compounds
(3) Sulphide production
(4) Production of volatile S compounds
(5) Oxidation of S and inorganic S compounds.

Reactions (1), (2), and (4) are prevalent in both oxic and anoxic zones of flooded soils, reaction (3) occurs principally in anoxic zones and reaction (5) occurs principally in oxic zones (but see Luther and Church, this volume, for a discussion of exceptions). Measurements of sulphate-S in the soil solution under field conditions would therefore vary with time, distance from the root, or root density, and the depth of sampling.

The significant fraction of soil S present as organic sulphur has to be mineralized to sulphate (or perhaps sulphide, if rice behaves like saltmarsh grasses) before it is available to plants. This mineralization depends not only on microbial activity but also on the nature of the organic matter. The mineralization of organic matter in wetland rice soils is enhanced by the annual wetting and drying cycles, generally high soil temperatures, and generally high and balanced nutrient supply. In contrast to many natural wetlands of the temperate zone, there is likely to be less build-up of organic matter that is resistant to decomposition. In addition, after each rice crop there is a substantial input of predominantly readily mineralizable organic matter from floodwater flora and fauna, rice roots and, depending on residue management practices, rice shoots. As production increases, as a result of management practices, so the mineralizable organic matter and therefore sulphur returned to the system is likely to increase.

The decomposition of organic matter that occurs after flooding and puddling is accompanied by a rapid fall in redox potential and changes in pH. Once oxygen and other more energetically favourable electron acceptors have been consumed, at least in a restricted zone, sulphate will be reduced to sulphide. The delay in sulphide formation and its subsequent rate of formation is controlled not only by the amount of oxygen in the system but also by other oxidants such as nitrate and manganese and iron oxides, which are generally thought to be reduced earlier in the soil reduction sequence. Productive rice wetland soils are usually relatively rich in iron oxides and in some cases MnO₂. The redox potential in the bulk soil is consequently poised by the Fe(III) ^_ Fe(II) and MnO₂ ^_ Mn(II) couples, well above the potential where rice scientists have generally thought sulphate reduction can occur (Ponnamperuma, 1972; Patrick and Reddy, 1978). Because of this assumption, few workers have actually studied sulphate reduction in flooded rice soils. However, E_h measurements can be a poor indication of whether or not particular microbially mediated reactions occur (Neue, 1988). In salt-marsh soils, high rates of sulphate reduction are found in soils with measured redox potentials as high as +350 mV (Howarth and Teal, 1979; Howes et al.,1981). The measured redox potential remains high because sulphides are rapidly precipitated as pyrite (FeS₂), and the low abundance of more energetically favourable electron acceptors, rather than the measured redox potential, regulates sulphate reduction in these natural wetlands (Howarth and Teal, 1979; Howarth, 1984). High rates of sulphate reduction have also been measured in the surface layers of subtidal marine sediments where measured redox potentials of +400 mV occur (Howarth and Jørgensen, 1984).

When sulphate is reduced, hydrogen sulphide, which is toxic to plants and microorganisms, is formed. If not precipitated as metal sulphides or oxidized in the surface soil and water layers, S can be lost from the system as H₂S gas. It is thought that the presence of significant amounts of H₂S is only likely to occur in organic or very sandy soils which are very low in reducible iron or when very large quantities of organic matter have been incubated in flooded soils; these are conditions rare in productive rice soils. There is very little evidence for significant losses of sulphur as H₂S from rice wetlands (Freney, Jacq and Baldensperger, 1982). Fluxes of H₂S are generally not significant losses of sulphur from natural wetlands as well (Andreae and Jaeschke, this volume; Giblin and Wieder, this volume). However, very few measurements of the flux or controls of H₂S production have been made in rice soils. More effort should be put into actual measurements, especially as there are several reasons to believe that sulphate reduction rates may be greater than once thought.

It can be speculated that losses of H₂S from productive rice soils may be greater than currently thought due to increased losses at night as a result of the nocturnal decline in the degree of oxidation of the rhizosphere, surface soil, and floodwater. In addition, as rice cropping is intensified and larger areas are brought under controlled irrigation, many rice wetland soils are being maintained in anoxic states for longer periods. This potentially increases the degree of soil reduction and thus gaseous S emissions. The loss of S may be further increased by the use of waste water which is high in carbon to irrigate rice fields, as is occurring near some large cities, particularly in China. The planned or incidental application of sulphate may also increase gaseous S losses, especially where more reducing management practices are used.

11.5 THE FATE OF SULPHATE

Sulphate in the floodwater and soil solution of the oxidized layers may be taken up by plants, immobilized into organic matter, sorbed on to anion exchange sites, removed by leaching and runoff or reduced to sulphide. With an active floodwater and surface soil biota, some sulphate will be rapidly incorporated into organic matter, but with the relatively short generation time of many of these organisms the turnover should be rapid. Anything that affects the activity of this wetland biota, such as temperature or floodwater/soil fertility, will clearly affect the rate of sulphate immobilization.

Sorption of sulphate from the floodwater and oxidized surface soil layer will be affected by changes in the anion exchange capacity, the sulphate concentration, and changes in the concentration of competing anions (Mitchell, David and Harrison, this volume). Sulphate sorption capacity declines as pH increases in upland (Kamprath, Nelson and Fitts, 1956; Barrow, 1970; Gebhart and Coleman, 1974; Couto, Lathwell and Bouldin, 1979) and low-land soils (Samosir, Blair and Lefroy, 1988). Thus on acid soils, the pH increase that occurs with flooding would decrease the sorption capacity, while flooding of alkaline soils would increase sorption capacity. Changes in sorption capacity with changes in redox will depend on the nature of the major sorption sites. The reduction of ferric iron compounds to the soluble ferrous forms, along with the effects of puddling can weaken soil structure, thus increasing the anion adsorption capacity by exposing more exchange sites. Similarly, the formation of colloidal or amorphous iron compounds from the solubilized ferrous iron compounds would increase sorption capacity. Increases in colloidal iron, as measured in an ammonium acetate extract, have been found to occur across a range of pH conditions, as redox potential declined (Samosir, Blair and Lefroy, 1988). Exchange capacity can also be increased by the interaction of organic matter breakdown products, such as humic and fulvic acid, with iron oxides (Parfitt, Fraser and Farmer, 1977; Tipping, 1981). Alternatively, the solubilizing of ferric oxide coatings may remove all the associated positive charges and reveal negatively charged clay surfaces, thus reducing anion adsorption. The net effect on exchange capacity will be a balance between these processes.

Whatever the change in anion exchange capacity, the net effect on sulphate sorption will also depend on changes in the concentration of potentially competing anions. As the redox potential falls and ferric ions are reduced, the levels of iron and phosphate in solution increase dramatically (Patrick, Gotoh and Williams, 1973; Chang, 1976; Samosir, Blair and Lefroy, 1988). The increased phosphate in solution is likely to increase the sulphate in solution by competition with sulphate for adsorption sites (Parfitt, 1980; Nagarajah, Posner and Quirk, 1970).

At the onset of flooding, CO₂ and bicarbonate accumulate from organic matter decomposition. The bicarbonate ion will also compete with sulphate for sorption sites and so is likely to increase the sulphate in solution. The organic acids that result from organic matter decomposition will also increase competition for adsorption sites (Parfitt, 1980; Nagarajah, Posner and Quirk, 1970). The effect of flooding on solution sulphate concentrations will be a balance between these changes in anion exchange capacity and the concentration of competing ions (Samosir, Blair and Lefroy, 1988). The competitive effects of the increased concentrations of phosphate, organic acids, and bicarbonate are most likely to result in an increase in sulphate in solution in the oxidized layers unless the anion exchange capacity is increased enormously. An increase in sulphate in solution after flooding has been observed by Ponnamperuma, Attanandana and Beye (1973) and Patnaik (1978).

The loss of sulphate in floodwater will depend on the water management system. If significant amounts of floodwater are flowing through the rice fields, significant losses of sulphate and other nutrients may occur. However, if water is added when required to maintain floodwater depth, then losses would be reduced.

Movement of sulphate in the soil solution will largely be controlled by soil structure, although influenced by the degree of puddling and soil macrofauna activity. Movement of sulphate down the profile will generally result in its reduction and subsequent precipitation as sulphide. In some very light textured soils, losses of sulphur may occur as a result of lateral and vertical soil solution movement.

The precipitation of relatively immobile sulphide in the reduced soil layers renders the sulphur temporarily unavailable for plants. As a root grows into a zone where metal sulphides have been precipitated, the oxygenated rhizosphere allows bacteria to oxidize sulphide into the plant available sulphate form. If the sulphide is present deep in the profile it is unlikely to be available during the growing season as rice does not have a deep root system; in fact, after maximum tillering a large portion of the rice roots form a mat on the soil surface (Alberda, 1954). Sulphides deposited deep in the soil profile can be oxidized to sulphate under the unsaturated conditions which occur during fallow periods. This sulphate may then be available for the rice plants at the beginning of the next cropping season, however it may also be leached from the soil as the profile becomes saturated.

11.6 SULPHUR CYCLING IN RICE CROPPING SYSTEMS

11.6.1 SULPHUR UPTAKE AND USE BY RICE

Sulphur is important in rice nutrition for the synthesis of amino acids and proteins, which account for approximately 90% of organic S in the plant. The sulphur requirement of rice varies according to the nitrogen supply. When S becomes limiting, addition of N does not change the yield or protein level of plants (Dijkshoorn and van Wijk, 1967). Sulphur is required early in the growth of rice plants. If it is limiting during early growth, then tiller number and therefore final yield will be reduced (Blair, Momuat and Mamaril, 1979). Although the concentration of sulphur in rice grain is higher than in rice straw, there is less re-mobilization of S to the grain than for phosphorus. The uptake of S and its distribution within the plant for rice grown in northeast Thailand can be seen in the sulphur balance sheet in Table 11.3.

Table 11.3. Sulphur balance sheet for rice grown on an Aeric Paleaquult at Ubon Ratchatani, Thailand (from Lefroy, Blair and Blair, 1988)

If it is assumed that all of the sulphur from the atmosphere is available to the rice crop and that no leaching losses occur then, with no added sulphur fertilizer, the system shows a net loss of 60 mg (S) m^-2 (0.6 kg ha^-1, or 2 mmol m^-2) if no residues are returned. Less than half of the sulphur in the above ground part of the plant was in the grain; consequently, the return of all residue to the cropping area reduces the sulphur loss to the point where there was a net gain of 210 mg (S) m^-2 (6.6 mmol m^-2). The addition of 800 mg (S) m^-2 (8 kg ha^-1, or 25 mmol m^-2) as gypsum resulted in a net gain of S irrespective of residue management. In the following season, when no further S fertilizer was applied and the previous season's residue was incorporated, a response in grain and straw yield was seen. This suggests the S balances in Table 11.3 are an overestimate, presumably because not all the S from the atmosphere can be used by the rice crop and leaching losses were not taken into account. On a slightly more fertile site in Thailand, yielding about 2.5 t ha^-1 (250 g m^-2) of grain, the total S uptake when no sulphur fertilizer was applied was 770 mg (S) m^-2 (24 mmol m^-2). This site received 340 mg (S) m^-2 (11 mmol m^-2) during the year from the atmosphere. In this case, even when the stubble was returned, there was a net loss of 200 mg (S) m^-2 (6 mmol m^-2) .This site showed no response to sulphur and had been cropped for many years; it would therefore appear that either there was a very large soil sulphur pool which was being depleted, or there was an additional input which kept the system in a positive or near-neutral sulphur balance. The most likely additional input is in irrigation water.

11.6.2 INPUTS OF SULPHUR IN IRRIGATION WATER AND RAINFALL 

There has been relatively little work that has accurately assessed the role of irrigation water in rice nutrition. Amarasiri and Latheef (1982) estimated that the sulphur input in irrigation water across four sites in Sri Lanka ranged from 1.4 to 4.3 g (S) m^-2 (44 to 130 mmol m^-2) per season and that this was from 2.6 to 17 times as much as received in rainfall. Freney, Jacq and Baldensperger (1982) report from various sources a range of 0.2 to 20.2 mg (S) l^-1 (6 to 630ĩM) in irrigation and river water.

The question of what concentration of sulphur is required in irrigation water is more complex, as it is necessary to consider the plant's uptake system and various site-specific factors, such as the adsorption capacity of the soil and the residence time of the floodwater in the rice field. Thus, there is no consensus among research reports that try to relate sulphur concentration in irrigation waters to crop requirements and production (Yoshida and Chaudhry , 1979; Wang, 1979; Ishizuka and Tanaka, 1959; Blair, Momuat and Mamaril, 1979).

There is also little work on the variation in amounts of sulphur in rainfall or its effectiveness to rice plants. The S in rainfall is often assumed to be largely in the sulphate form and is generally higher in industrialized countries and near industrial centres, volcanoes, swamps, and seas (Blair, Mamaril and Momuat, 1978). Measurements in Korea have shown that the concentration of sulphate in rainfall is inversely related to the amount of precipitation, with very high concentrations in the first 5 mm and lower concentrations with increasing amounts of rainfall (Shin, 1987).

A wide range of sulphur deposition in rainfall has been measured in rice growing countries of Asia (Table 11.4). Total sulphur inputs to these rice systems from the atmosphere may be larger than these values indicate depending on the amount of dry deposition from the atmosphere and how much dry deposition has been included in these measurements. There are virtually no separate measurements of wet and dry deposition of sulphur in the major rice growing areas of the world.

Recent measurements of S in rainfall at approximately 60 sites across Malaysia and Thailand show large differences in the amount of S received and its temporal distribution (Lefroy, unpub. data). As expected, the S inputs were found to vary with proximity to industrial and population centres, the amount of rainfall, proximity to the coast, and the season. While there were large local inputs from industrial and population centres, the major influence was the marine input during monsoonal periods, particularly near the east coast of peninsular Malaysia during the northeast monsoons. On light textured soils in this region, large responses of rice to sulphur applications were seen at the beginning of the northeast monsoon, a month after transplanting. By maturity these differences had disappeared with the extra sulphur input coming in monsoonal rain estimated to be between 500 mg and 1 g (S) m^-2 (15 to 31 mmol m^-2), an amount sufficient to explain the disappearance of S deficiency.

Table 11.4. Rates of sulphur deposition in rainfall in rice growing countries

Sulphur in rainfall declines along a transect from the coast as shown in northern Queensland (Probert, 1976) and in northern New South Wales (Lefroy, unpub. data). Monthly measurements of rainfall and sulphur accession across a 700 km transect from the coast have shown, after 12 consecutive collections, a decline from a total of more than 2 g (S) m^-2 (63 mmol m^-2) at the coast to less than 200 mg (S) m^-2 (6.3 mmol m^-2) 150 km from the coast. At 13 km from the coast, despite similar rainfall to the coastal site, the sulphur accession was down to less than 700 mg (S) m^-2 (22 mmol m^-2).

11.6.3 RESIDUE MANAGEMENT 

The balance sheet approach used above indicates the relatively large quantity of sulphur that remains in rice stubble and panicle straw and thus emphasizes the importance of residue management. The management of crop residues is influenced by agronomy, socio-economics, and tradition. In many rice-growing regions, rice residues are an important source of animal fodder. In Indonesia it is estimated that 50% of rice crop residue produced during the dry season is used for animal fodder, but during the wet season only 2% is used, since more forage is available (Ranjhan, 1986). The use of rice residue as animal feed may not represent a total loss of sulphur from the rice wetlands if some animal manure is returned. Other uses, such as for paper production, for mushroom cultivation or for fuel, generally represent more complete losses from the system.

In some areas all the rice residues are incorporated (i.e. added back to the soil). However, there are significant problems associated with incorporating large quantities of residue, especially when cropping intensities are high and there is only a short fallow period between crops. In these and other situations the residues are often burnt. Attempts to simulate the slow burn that typifies the normal practice have shown that between 40 and 60% of the sulphur is lost (Chaitep, 1990). By using the ³⁵S-labelled residues, it was possible to show that similar proportions of the sulphur applied in burnt or unburnt residue were recovered by the plant. However, when the amount of sulphur lost during burning was taken into account, the proportion of sulphur in the original unburnt residue that was recovered by the plant was much lower when the residue was burnt; a larger amount of sulphur remained in the soil after the crop when the unburnt residue was applied (Table 11.5). It appears that any small gains in efficiency of use of S from burnt residue are offset by the substantial losses of S when straw is burnt, assuming that sulphur lost during burning is lost from the system and not returned as ash particles or in rainfall.

Many traditional cropping systems include an upland crop in rotation with rice, especially when there is insufficient water for an additional rice crop. These upland crops are grown for many and often multiple uses, including human and animal feed, fuel, fibre, and green manure for improved physical and chemical soil fertility. With intensification of rice cultivation, often associated with improved irrigation, many of these traditional multiple cropping systems have been replaced by multiple rice crop systems. Recently there has been more interest in using green manure crops in rice culture (IRRI, 1988), especially as legume germplasm shows greater prospects for multiple use in more sustainable farming systems. The case for green manuring must take into account the effect on rice production, in terms of a possible reduction in cropping intensity and the benefits of improved soil fertility, as well as the changes in nutrient demands of these multiple cropping systems. The sulphur requirements of many of these upland crops, especially the legumes, are higher than for rice, but the sulphur dynamics of these systems with regard to their residue and soil, particularly hydrological, management is not well known and needs further clarification.

Table 11.5. The proportion of S in straw recovered, either in the rice plant or soil, or lost in burning

11.6.4 SULPHUR FERTILIZERS 

Responses of flooded rice to S application in Bangladesh, Burma, Brazil, India, Indonesia, and Taiwan are cited by Freney, Jacq and Baldensperger (1982), with yield responses as high as 278% in Indonesia and 330% in Brazil. Blair, Momuat and Mamaril (1979) reported an average grain yield response of 18.6% at 28 sites in South Sulawesi, Indonesia.

Sulphur is applied in fertilizers as either sulphate or elemental sulphur. Rice plants take up S from solution as sulphate (as mentioned above, there is some evidence that some marsh plants can take up sulphide, but if this is true of rice it is most likely a minor pathway). Consequently, elemental sulphur must be converted to sulphate before it is available; the inhibition of oxidation of elemental sulphur to sulphate will decrease the availability of fertilizer sulphur to the plant. Also, the reduction of sulphate to sulphide will reduce the availability of the sulphur, particularly if the sulphides are precipitated. To be effective, S-containing fertilizers must be placed in a zone of high oxidation-reduction potential such as the oxidized soil layer at the soil-water interface or within the root rhizosphere.

Samosir (1989) found in a pot experiment using ³⁵S-labelled fertilizers, that over 80% of the rice plant's S was derived from a surface application of 4 g (S) m^-2 (13 mmol m^-2), applied as potassium sulphate, yet when the same application was incorporated 7 to 21 cm below the surface, only 34% of the plant's S came from the fertilizer. The lower recovery from deep placement was most likely due to reduction to sulphide and precipitation of the sulphide. However, a lower biomass of roots at depth may also have contributed to this finding. The surface application of 4 g (S) m^-2 as elemental sulphur was effective as a fertilizer, but somewhat less so than the surface application of sulphate, with 50% of the plant's S being derived from this treatment. The inhibition of oxidation of elemental sulphur when it was placed deep within the soil greatly reduced its effectiveness as a fertilizer, providing only 8% of plant sulphur. Similar reductions in the effectiveness of sulphate and elemental fertilizers when deep placed in pots have been found by Friesen and Chien (1986) and Chien, Hellums and Henao (1987).

In field experiments in Sulawesi, a comparison of sulphur fertilizers showed that sulphate and elemental sulphur were equally effective when surface applied (Blair, Momuat and Mamaril, 1979). Glasshouse trials comparing urea-S, elemental S, ammonium sulphate, gypsum, and S bentonite on 11 wetland soils from the Philippines showed that urea-S produced significantly higher yields than elemental sulphur, ammonium sulphate, or gypsum and that S bentonite was not immediately effective, although it showed a residual effect on succeeding crops (Mamaril and Gonzales, 1987).

The effectiveness of applied S on rice grain yield depends on timing application to coincide with plant requirement as well as placement of fertilizer sulphur in the vicinity of roots. Sulphur is best applied between transplanting or seeding and maximum plant tillering. Application after maximum tillering does not have an appreciable effect on yield since S has a large affect on tiller formation and therefore grain yield via panicle number. In a field experiment in the Philippines using two varieties of different growth duration (110-111 and 111-120 days), S application at 15 days after planting or later significantly reduced the yield of the short duration variety, when compared to an application at transplanting, but did not affect the longer duration variety (Mamaril and Gonzales unpub. data). Application at 30 days after transplanting significantly affected the yield of both varieties.

The work of Blair, Momuat and Mamaril (1979) also showed the importance of timing of sulphur applications. The application of elemental sulphur 20 days before transplanting produced significantly lower yields than the same application at transplanting. This was presumably because of the oxidation of elemental sulphur and its subsequent loss by leaching as sulphate from the paddy or immobilization as sulphides in the soil. Clearly the management of sulphur fertilizer, in terms of source, placement and timing, as well as rate, is of significance to the rice crop and to the environment at large.

11.7 CONCLUSIONS: SULPHUR MANAGEMENT OF RICE 

There are many aspects of the sulphur cycle in wetland rice systems, as outlined above, which are influenced by management practices. Such practices not only influence the yield of rice, but also the efficiency of use of sulphur which ultimately affects the economics of production and the environmental consequences of the wetland rice cropping system.

The flooding and puddling of a soil can generally be said to produce an improved medium for the cultivation of rice in terms of its hydrology, nutrition, and weed control. This is particularly true if, as is most often the case with rice, the soil undergoes wetting and drying cycles between crops and substantial amounts of readily decomposable organic matter are incorporated. This general case is true for sulphur as long as highly reduced conditions do not dominate, particularly in soils low in reducible iron. In these cases not only will sulphur availability be reduced by sulphide formation but also damage can be done to the crop as a result of sulphide toxicity, and in extreme cases losses of sulphur to the atmosphere may occur. Other losses of sulphur from rice wetland systems arise from outflow in floodwater and leaching losses in light textured soils. These leaching losses may occur in the saturated state or when sulphides deposited deep in the profile are oxidized and leached during unsaturated fallow periods.

The management of sulphur relies on balancing input with output in crop harvests. In recent years, rice production has been increased in many areas by greater use of fertilizers, the use of high yielding rice varieties, the improvement of management practices, such as irrigation, and the intensification of cropping (Blair and Lefroy, 1987). This increased production has resulted in greater sulphur output. The use of high yielding varieties, with their improved harvest index, as well as changes in residue management, with more use of residues as forage and more burning of residues to facilitate intensive cropping, has resulted in a lower proportion of plant sulphur being returned to the soil. Similarly, the increased use of low S or S-free fertilizers, such as urea and triple superphosphate, has resulted in less sulphur being applied in fertilizers.

In some areas the reduction of industrial sulphur emissions may have reduced sulphur input in rainfall, and further reductions are likely in some areas. However, on the industrial side of agriculture, the increasing demand for fertilizers, both sulphur fertilizers and phosphate fertilizers produced by sulphuric acid processes, may lead to greater emission of SO₂ during elemental sulphur production from sour gas sources. Sales of elemental sulphur are increasing yearly, and the manufacture of fertilizer is the major use.

There is increasing evidence that the result of these changes in the balance between sulphur input and sulphur output is an increased deficiency in sulphur, reducing the yields of rice, and other crops, in certain areas. In terms of the whole area of rice production, the areas currently showing or having the potential to develop sulphur deficiency are probably not large, although on the local scale they may be very important. They are likely to be on lighter textured soils, which are intensively cropped, with little residue return and with low inputs of sulphur in rainfall or irrigation _¯ in other words, not too near the coast where brackish irrigation water or high sulphur in rainfall increases the sulphur inputs. It is important that appropriate management practices be developed for these soils. First, the return of as much plant residue as possible should be encouraged. This may involve the application of ash or animal manure, particularly where the demand for rice residue for fuel, in industry , or for animal forage is high, or the quantity and timing of application of organic matter may result in a very rapid decline in the redox potential of the soil.

Second, appropriate fertilizers and fertilizer application methods should be developed. The use of elemental sulphur in a fertilizer management scheme has the attraction of being a high analysis form of sulphur, and the oxidation rate is dependent on temperature and moisture, as is plant demand. Elemental sulphur can be applied alone, as a powder or in a pelleted form, or with other fertilizers as a coating or a mixture. The application of elemental sulphur in a S ^_ N fertilizer is probably not appropriate since for maximum efficiency nitrogen fertilizers should be deep placed, while deep placed elemental sulphur will be oxidized only slowly. The application of elemental sulphur combined with a phosphate fertilizer, such as triple superphosphate, seems most appropriate, as both are most effective in the more oxidized surface layers and both need to be applied early in crop growth.

Balancing the demands of the wetland rice crop, both at particular stages during the crop and for the crop as a whole, with the inputs of sulphur is important for maximizing yield and the economics of production as well as for achieving a sustainable agriculture system with minimal adverse environmental consequences. Adequate understanding of the sulphur cycle is essential for development of management practices which will achieve these aims. At present there are many unknowns, not the least of these being an adequate knowledge of the soil sulphur transformations and fluxes on a daily and seasonal basis under various cropping systems and management conditions. It appears that agriculturalists now need to adopt some of the methods used by geochemists and ecologists in natural wetlands to understand more fully the dynamics of the sulphur cycle.

REFERENCES

Alberda, T. (1954). Growth and root development of lowland rice and its relation to oxygen supply. Plant Soil, 5, 1-28.

Amarasiri, S. L. and Latheef, M. A. (1982). The sulphur deposition by rainfall in some agriculturally important areas of Sri Lanka. Trop. Agric., 138, 33-43.

Andreae, M. O. and Jaeschke, W. A. This volume.

Armstrong, W. (1964). oxygen diffusion from the roots of some British bog species. Nature, 204, 801-2.

Armstrong, W. (1971). Radial oxygen losses from intact rice plants as affected by distance from the apex, respiration and waterlogging. Physiol. Plant., 25, 192-7.

Barber, D. A., Ebert, M. and Evans, N. T. S. (1962). The movement of ¹⁵o through barley and rice plants. J. Exp. Bot. , 13, 397-403.

Barrow, N. J. (1970). Comparison of the adsorption of molybdate, sulfate and phosphate by soils. Soil Sci., 109, 282-8.

Blair, G. J. and Lefroy, R. D. B. (1987). Sulphur cycling in tropical soils and the agronomic impact of increasing use of S free fertilizers, increased crop production and burning of crop residue. In: Proceedings of the Symposium on Fertilizer Sulphur Requirements and Sources in Developing Countries of Asia and the Pacific. FADI- NAP, FAO, TSI and ACIAR Bangkok, Thailand, pp. 12-17.

Blair, G. J., Mamaril, C. P. and Momuat, E. O. (1978). Sulfur nutrition of wetland rice. IRRI Res. Pap. Ser., 21, 29.

Blair, G. J., Momuat, E. O. and Mamaril, C. P. (1979). Sulfur nutrition of rice. II. Effect of source and rate of S on growth and yield under flooded conditions. Agron. J. , 71, 477-80.

Carlson, P. R. and Forrest, J. (1982). Uptake of dissolved sulfide by Spartina alternifltora: evidence from natural sulfur isotope abundance ratios. Science, 216, 633-5.

Chang, S. C. (1976). Phosphorus in submerged soils and phosphorus nutrition and fertilization of rice. In: The Fertility of Paddy Soils and Fertilizer Applications for Rice. Food and Fertilizer Technology Center, ASPAC, Taipei, Taiwan, pp. 93-116.

Chaitep, W. (1990). The use of fertilisers and crop residues as sources of sulfur in flooded and unflooded rice cropping systems. Ph.D. Thesis, University of New England.

Chien, S. H., Hellums, D. T. and Henao, J. (1987). Greenhouse evaluation of elemental sulfur and gypsum for flooded rice. Soil Sci. Soc. Am. J. , 51, 120-3.

Couto, W., Lathwell, D. J. and Bouldin, D. R. (1979). Sulfate sorption by two oxisols and an alfisol of the tropics. Soil Sci., 127,108-16.

Dijkshoorn, W. and van Wijk, A. L. (1967). The sulphur requirements of plants as evidenced by the sulphur-nitrogen ratio in the organic matter. A review of published data. Plant Soil, 26, 129-57.

Freney, J. R., Jacq, V. A. and Baldensperger, J. F. (1982). The significance of the biological sulfur cycle in rice production. In: Dommergues, Y. R. and Diem, H. G. (Eds) .Microbiology of Tropical Soils and Plant Productivity. Developments in Plant and Soil Science 5. Martinus Nijhoff, The Hague, pp. 271-317.

Friesen, D. K. and Chien, S. H. (1986). Sulphur fertilizers and their application: General review and S source evaluation at IFDC. In: Sulphur in Agricultural Soils. BARC and TSI, Dhaka, Bangladesh, 1986, pp. 407-31.

Gaddi, V. Q. (1988). Soil sulphur fractions as indices of plant-available sulphur and the influence of presubmergence on sulphur availability in selected rice soils in the Philippines. MS Thesis, University of the Philippines, Los Baņos.

Gebhart, H. and Coleman, N. T. (1974). Anion adsorption by allophanic tropical soils. II. Sulfate adsorption. Soil Sci. Soc. Am. Proc., 38, 259-62.

Giblin, A. E. and Wieder, R. K. Chapter 5 in this volume.

Greenland, D. J. and De Datta, S. K. (1985). Constraints to rice production and wetland soil characteristics. In: Wetland Soils: Characterization, Classification and Utilization. International Rice Research Institute, Los Baņos, Philippines, pp. 23- 36.

Guthrie, R. L. (1985). Characterizing and classifying wetland soils in relation to food production. In: Wetland Soils: Characterization, Classification and Utilization. International Rice Research Institute, Los Baņos, Philippines, pp. 11-20.

Hegde, T. M., Rao, N. N. and Biddappa, C. C. (1980). Evaluation of sulphur status of different rice soils ^_ I. Mysore J. Agric. Sci., 14, 171-6.

Howarth, R. W. (1984). The ecological significance of sulfur in the energy dynamics of salt marsh and coastal marine sediments. Biogeochemistry, 1, 5-27.

Howarth, R. W. and Jørgensen, B. B. (1984). Formation of ³⁵S-labelled elemental sulfur and pyrite in coastal marine sediments (Limfjorden and Kysing Fjord, Denmark) during short-term ³⁵S0₄^2- reduction measurements. Geochim. Cos- mochim. Acta, 48, 1807-18.

Howarth, R. W. and Teal, J. M. (1979). Sulfate reduction in a New England salt marsh. Limnol. Oceanogr., 24, 999-1013.

Howes, B. J., Howarth, R. W., Teal, J. M. and Valiela, I. (1981). Oxidation ^_ reduction potentials in a salt marsh: Spatial patterns and interactions with primary production. Limnol. Oceanogr., 26, 350-60.

IRRI (1988). Green manure in rice farming. Proceedings of a Symposium on Sustainable Agriculture. International Rice Research Institute, Los Baņos, Philippines, 379pp.

Ishizuka, Y. and Tanaka, A. (1959). Inorganic nutrition of rice plants. IV. Effect of calcium, manganese and sulfur levels in culture solution on yields and chemical composition of the plant. J. Sci. Soil Man., Jpn, 30,414-16 (in Japanese). Cited in Freney, Jacq and Baldensperger (1982).

Islam, M. M. and Ponnamperuma, F. N. (1982). Soil and plant tests for available sulphur in wetland soils. Plant Soil, 68, 97-113.

Ismunadji, M. and Zulkarnaini, I. (1978). Sulphur deficiency of lowland rice in Indonesia. Sulphur Agric., 2, 17-19, 22.

John, C. D., Limpinuntana, V. and Greenway, H. (1974). Adaptation of rice to anaerobiosis. Aust. I. Plant Physiol., 1, 513-20.

Kamprath, E. J., Nelson, W. L. and Fitts, J. W. (1956). The effect of pH, sulfate and phosphate concentrations on the adsorption of sulfate by soils. Soil Sci. Soc. Am. Proc., 20, 463-6.

Katayama, T. (1961 ). Studies on the intercellular spaces in rice I. Proc. Crop Sci. Soc. Jpn., 29, 229-33.

Kyuma, K. (1988). Paddy soils in Japan in comparison with those in tropical Asia. In: First International Symposium on Paddy Soil Fertility. Department of Land Development, Bankghen, Thailand, pp. 5-19.

Lefroy, R. D. B., Blair, S. and Blair, G. J. (1988). Phosphorus and sulfur efficiency in tropical cropping systems. ACIAR Project 8328 Final Report. University of New England, Armidale, NSW 2351, Australia.

Liu, C. Q. (1987). Sulphur in the agriculture of China. In: Proceedings of the Symposium on Fertilizer Sulphur Requirements and Sources in Developing Countries of Asia and the Pacific. FADINAP, FAO TSI and ACIAR, Bangkok, Thailand, pp. 26-30.

Liu, C. Q., Chen, G. A. and Cao, S. Q. (1980). Sulphur content and distribution on paddy soils of South China. In: Proceedings of Symposium on Paddy Soils, pp. 628- 34.

Luther, G. W. and Church, T. M. Chapter 6 in this volume.

Mamaril, C. P. and Gonzales, P. B. (1987). Sulphur status and requirements of the Philippines. In: Proceedings of the Symposium on Fertilizer Sulphur Requirements and Sources in Developing Countries of Asia and the Pacific. FADINAP, FAO, TSI and ACIAR, Bangkok, Thailand, pp. 67-75.

Mitchell, M. J., David, M. B. and Harrison, R. B. This volume.

Moorman, F. R. (1978). Morphology and classification of soils on which rice is grown. In: Soils and Rice. International Rice Research Institute, Los Baņos, Philippines, pp. 255-72.

Nagarajah, S., Posner, A. M. and Quirk, J. P. (1970). Competitive adsorption of phosphate with polygalacturonate and other organic anions on kaolinite and oxide surfaces. Nature (Lond)., 228, 83-5.

Neue, H. U. (1988). Wholistic view of chemistry of flooded soil. In: First International Symposium on Paddy Soil Fertility. Department of Land Development, Bangkhen, Thailand, pp. 21-53.

Neue, H. U. and Bloom, P. R. (1989). Nutrient kinetics and availability in flooded soils. In: Proceedings of the International Rice Research Conference 1987. International Rice Research Institute, Los Baņos, Philippines.

Parfitt, R. L. (1980). Chemical properties of variable charge soils. In: Theng, B. K. G. (Ed.). Soils with Variable Charge. New Zealand Society of Soil Science, pp. 167-94.

Parfitt, R. L., Fraser, A. R. and Farmer, V. C. (1977). Adsorption on hydrous oxides. III. Fulvic acid and humic acid on goethite, gibbsite and imogolite. J. Soil Sci., 28, 289-96.

Parkpian, P., Pongsakul, P., Sangtong, P., Cholitkul, W. and Ratanarat, S. (1986). Indigenous soil properties influencing the availability of sulphur in selected Thai soils. In: Sulphur in Agricultural Soils. BARC and TSI, Dhaka, Bangladesh, pp. 469-87.

Patnaik, S. (1978). Natural sources of nutrients in rice soils. In: Soils and Rice. The International Rice Research Institute, Los Baņos, Philippines, pp. 501-19.

Patrick, W. H. Jr, Gotoh, S. and Williams, B. G. (1973). Strengite dissolution in flooded soils and sediments. Science, 179, 564-5.

Patrick, W. H. Jr and Reddy, C. N. (1978),. Chemical changes in soils. In: Soils and Rice. The International Rice Research Institute, Los Baftos, Philippines, pp. 361-37.

Peterson, B. I., Howarth, R. W. and Garritt, R. H. (1986). Sulfur and carbon isotopes as tracers of salt-marsh organic matter flow. Ecology, 67, 865-74.

Ponnamperuma, F. N. (1972). The chemistry of submerged soils. Adv. Agron., 24, 29-96.

Ponnamperuma, F. N., Attanandana, T. and Beye, G. (1973). Amelioration of three sulphate soils for lowland rice. In: Dost, H. (Ed.). Acid Sulphate Soils. International Institute for Land Reclamation and Improvement, Wageningen, The Netherlands, pp. 391-405.

Probert, M. E. (1976). The composition of rainwater at two sites near Townsville, Qld., Australia. J. Soil Res., 14, 397-402.

Ranjhan, S. K. (1986). Sources of feed for ruminant production in Southeast Asia. In: Blair, G. I., Ivory, D. A. and Evans, T. R. (Eds). Forages in Southeast Asian and South Pacific Agriculture. ACIAR Proceedings Series No.12, Canberra, Australia, pp. 24-8.

Reyes, I. C., Villapando, R. R., Rosales, R. T., Gongales, P. B. and Mamaril, C. P. (1985). Sulphur studies in some Philippine wetland rice soils. Paper presented at Saturday Seminar, International Rice Research Institute, Los Baņos, Philippines.

Roger, P. A. and Kurihara, Y. (1988). Floodwater biology of tropical wetlands. In: First International Symposium on Paddy Soil Fertility .Department of Land Development, Bangkhen, Thailand, pp. 275-300.

Samosir, S. S. R. (1989). Dynamics of sulfur in flooded soils and the sulfur nutrition of rice plants. Ph.D. Thesis, University of New England.

Samosir, S. S. R., Blair, G. J. and Lefroy, R. D. B .(1988). Sulfur transformations in soils undergoing flooding. In: First International Symposium on Paddy Soil Fertility. Department of Land Development, Bangkhen, Thailand, pp. 587-600.

Shin, J. S. (1987). Sulphur in Korean agriculture. In: Proceedings of the Symposium on Fertilizer Sulphur Requirements and Sources in Developing Countries of Asia and the Pacific. FADINAP, FAO, TSI and ACIAR, Bangkok, Thailand, pp. 76-82.

Stangel, P. J. (1979). Nitrogen requirement and adequacy of supply for rice production. In: Nitrogen and Rice. International Rice Research Institute, Los Baņos, Philippines, pp. 45-69.

Tabatabai, M. A. (1982). Sulfur. In: Method of Soil Analysis, Part 2. Chemical and Microbiological Properties. American Society of Agronomy and Soil Science, Madison, Wisconsin, pp. 501-38.

Tipping, E. (1981). The adsorption of aquatic humic substances by iron oxides. Geochim. Cosmochim. Acta, 45, 191-9.

USDA Soil Conservation Service, Soil Survey Staff (1975). Soil Taxonomy. A Basic System for Soil Classification for Making and Interpreting Soil Surveys. USDA Agricultural Handbook 436. US Department of Agriculture, US Government Printing Office, Washington DC.

van Raalte, M. H. (1940). On the oxygen supply of rice roots. Ann. Jardin Bot. Buitenzorg, 50, 99-114.

Wang, C. H. (1979). Sulphur fertilization of rice ^_ diagnostic techniques. Sulphur Agric., 3, 12-15.

Wang, C. H., Liem, T. H. and Mikkelsen, D. S. (1976). Sulphur deficiency ^_ a limiting factor in rice production in the lower Amazon Basin. I. Development of sulphur deficiency as limiting factor for rice production. IRI Res. Inst. Bull. No.47.

Yoshida, S. and Chaudhry, M. R. (1979). Sulfur nutrition of rice. Soil Sci. Plant Nutr., 25, 121-34.

Yu, T. R. (1985). Soil and plants. In: Yu, T. R. (Ed.). Physical Chemistry of Paddy Soils. Science Press, Beijing, pp. 195-217.

Zahari, A. B., Tham, K. C., Wong, N. C. and Kerridge, P. C. (1983). Sulfur deficiencies in the agriculture of Malaysia. In: Blair, G. J. and Till, A. R. (Eds). Sulfur in South East Asian and South Pacific Agriculture. UNE, Indonesia, pp. 155- 64.