Nitrous oxide emissions from grazed grassland

Abstract. Grazing animals on managed pastures and rangelands have been identified recently as significant contributors to the global N₂O budget. This paper summarizes relevant literature data on N₂O emissions from dung, urine and grazed grassland, and provides an estimate of the contribution of grazing animals to the global N₂O budget.
The effects of grazing animals on N₂O emission are brought about by the concentration of herbage N in urine and dung patches, and by the compaction of the soil due to treading and trampling. The limited amount of experimental data indicates that 0.1 to 0.7% of the N in dung and 0.1 to 3.8% of the N in urine is emitted to the atmosphere as N₂O. There are no pertinent data about the effects of compaction by treading cattle on N₂O emission yet. Integral effects of grazing animals have been obtained by comparing grazed pastures with mown-only grassland. Grazing derived emissions, expressed as per cent of the amount of N excreted by grazing animals in dung and urine, range from 0.2 to 9.9%, with an overall mean of 2%. Using this emission factor and data statistics from FAO for numbers of animals, the global contribution of grazing animals was estimated at 1.55 Tg N₂O-N per year. This is slightly more than 10% of the global budget.     

Nitrous oxide emissions from cattle-impacted pasture soil amended with nitrate and glucose

There is little information concerning N₂O fluxes in the pasture soil that has received large amounts of nutrients, such as urine and dung, for several years. The aims of this study were to (1) experimentally quantify the relationship between mineral N input and N₂O emissions from denitrification, (2) describe the time course of N₂O fluxes resulting in N inputs, and (3) find whether there exists an upper limit of the amount of nitrogen escaping the soil in the form of N₂O. The study site was a grassland used as a cattle overwintering area. It was amended with KNO₃ and glucose corresponding to 10–1,500 kg N and C per hectare, covering the range of nutrient inputs occurring in real field conditions. Using manual permanent chambers, N₂O fluxes from the soil were monitored for several days after the amendments. The peak N₂O emissions were up to 94 mg N₂O–N m⁻² h⁻¹, 5–8 h after amendment. No upper limit of N₂O emissions was detected as the emissions were directly related to the dose of nutrients in the whole range of amendments used, but the fluxes reflected the soil and environmental conditions, too. Thus, in three different experiments performed during the season, the total cumulative losses of N₂O–N ranged from 0.2 to 5.6% of the applied 500kg ha⁻¹. Splitting of high nutrient doses lowered the rate of N₂O fluxes after the first amendment, but the effect of splitting on the total amount of N₂O–N released from the soil was insignificant, as the initial lower values of emissions in the split variants were compensated for by a longer duration of gas fluxes. The results suggest that the cattle-impacted soil has the potential to metabolize large inputs of mineral nitrogen over short periods (∼days). Also, the emission factors for did not exceed values reported in literature.     

Carbon sequestration in a limestone quarry mine soil amended with sewage sludge

To reclaim a limestone quarry, 200 and 400 Mg/ha of municipal sewage sludge were mixed with an infertile calcareous substrate and spread as mine soil in 1992. Soil samples were taken 1 week later and again after 17 yr of mine soil rehabilitation so as to assess changes in the amount and persistence of soil organic carbon (SOC). Sludge application increased SOC as a function of the sludge rate at both sampling times. Seventeen years after the sludge amendments, the nonhydrolysable carbon was increased in the 400 Mg/ha of sludge treatment. The recalcitrance of SOC was less in sludge‐amended soils than in the control treatment at the initial sampling, but 17 yr later this trend had reversed, showing qualitative changes in soil organic carbon. The CO₂‐C production had not differed between treatments, yet the percentage of mineralized SOC was less in the high sludge dose. When the size of active (C_active) and slow (C_slow) potentially mineralizable C pools was calculated by curve fitting of a double‐exponential equation, the proportion of C_active was observed to be smaller in the 400 Mg/ha sludge treatment. Soil aggregate stability, represented by the mean weight diameter of water‐stable soil aggregates, was significantly greater in mine soil treated with the high dose of sludge (18.5%) and SOC tended to be concentrated in macro‐aggregates (5–2 mm). Results suggest that SOC content in sludge‐amended plots was preserved due by (i) replacement of the labile organic carbon of sludge by more stable compounds and (ii) protection of SOC in aggregates.     

Runoff and losses by erosion in soils amended with sewage sludge

In order to promote the transformation of a burnt Mediterranean forest area into a dehesa system, 10 t ha⁻¹ of dry matter of the same sewage sludge in three different forms: fresh, composted and thermally‐dried, were added superficially to field plots of loam and sandy soils located on a 16 per cent slope. This application is equivalent to 13ċ8 t ha⁻¹ of composted sludge, 50 t ha⁻¹ of fresh sludge and 11ċ3 t ha⁻¹ of thermally‐dried sludge. The surface addition of a single application of thermally‐dried sludge resulted in a decrease in runoff and erosion in both kinds of soil. Runoff in thermally‐dried sludge plots was lower than in the control treatment (32 per cent for the loam soil and 26 per cent for the sandy soil). The addition of any type of sludge to both soil types also reduces sediment production. Significant differences between the control and sludge treatments indicate that the rapid development of plant cover and the direct protective effect of sludge on the soil are the main agents that influence soil erosion rates. Results suggest that the surface application of thermally‐dried sludge is the most efficient way to enhance soil infiltration. Copyright © 2003 John Wiley & Sons, Ltd.     

The effect of increased atmospheric carbon dioxide concentration on emissions of nitrous oxide, carbon dioxide and methane from a wheat field in a semi-arid environment in northern China

There are no reports on the effects of elevated carbon dioxide [CO₂] on the fluxes of N₂O, CO₂ and CH₄ from semi-arid wheat cropping systems. These three soil gas fluxes were measured using closed chambers under ambient (420 ± 18 μmol mol⁻¹) and elevated (565 ± 37 μmol mol⁻¹) at the Free-Air Carbon dioxide Enrichment experimental facility in northern China. Measurements were made over five weeks on a wheat crop (Triticum aestivum L. cv. Zhongmai 175). Elevated [CO₂] increased N₂O and CO₂ emission from soil by 60% and 15%, respectively, but had no significant effect on CH₄ flux. There was no significant interaction between [CO₂] and N application rate on these gas fluxes, probably because soil N was not limiting. At least 22% increase in C storage is required to offset the observed increase in greenhouse gas emissions under elevated [CO₂].     

Nitrous oxide and methane emissions from grassland treated with bark- or sawdust-containing manure at different rates

On the main Japanese island of Honshu, bark or sawdust is often added to cattle excreta as part of the composting process. Dairy farmers sometimes need to dispose of manure that is excess to their requirements by spreading it on their grasslands. We assessed the effect of application of bark- or sawdust-containing manure at different rates on annual nitrous oxide (N₂O) and methane (CH₄) emissions from a grassland soil. Nitrous oxide and CH₄ fluxes from an orchardgrass (Dactylis glomerata L.) grassland that received this manure at 0, 50, 100, 200, or 300?Mg?ha^?1?yr^?1 were measured over a two-year period by using closed chambers. Two-way analysis of variance (ANOVA) was employed to examine the effect of annual manure application rates and years on annual N₂O and CH₄ emissions. Annual N₂O emissions ranged from 0.47 to 3.03?kg?N?ha^?1?yr^?1 and increased with increasing manure application rate. Nitrous oxide emissions during the 140-day period following manure application increased with increasing manure application rate, with the total nitrogen concentration in the manure, and with cumulative precipitation during the 140-day period. However, manure application rate did not affect the N₂O emission factors of the manure. The overall average N₂O emission factor was 0.068%. Annual CH₄ emissions ranged from ?1.12 to 0.01?kg?C?ha^?1?yr^?1. The annual manure application rate did not affect annual CH₄ emissions.     

Nitrous oxide emissions from grassland and spring barley, following N fertiliser application with and without nitrification inhibitors

The aims of this study were to assess the effectiveness of the nitrification inhibitors dicyandiamide (DCD) and nitrapyrin on reducing emissions of nitrous oxide (N₂O) following application of NH₄ ⁺ or NH₄ ⁺-forming fertilisers to grassland and spring barley. DCD was applied to grassland with N fertiliser applications in April and August in 1992 and 1993, inhibiting N₂O emissions by varying amounts depending on the fertiliser form and the time of application. Over periods of up to 2 months following each application of DCD, emissions of N₂O were reduced by 58–78% when applied with urea (U) and 41–65% when applied with ammonium sulphate (AS). Annual emissions (April to March) of N₂O were reduced by up to 58% and 56% in 1992–1993 and 1993–1994, respectively. Applying DCD to ammonium nitrate (AN) fertilised grassland did not reduce emissions after the April 1993 fertilisation, but emissions following the August application were reduced. Nitrapyrin was only applied once, with the April fertiliser applications in 1992, reducing N₂O emissions over the following 12 months by up to 40% when applied with U. When N fertiliser was applied in June without DCD, the DCD applied in April was still partly effective; N₂O emissions were reduced 50%, 60% and 80% as effectively as the emissions following the April applications, for AS in 1993, U in 1992 and 1993, respectively. In 1992 the persistence of an inhibitory effect was greater for nitrapyrin than for DCD, increasing after the June fertiliser application as overall emissions from U increased. There was no apparent reduction in effectiveness following repeated applications of DCD over the 2 years. N₂O emissions from spring barley, measured only in 1993, were lower than from grassland. DCD reduced emissions from applied U by 40% but there was no reduction with AN. The results demonstrate considerable scope for reducing emissions by applying nitrification inhibitors with NH₄ ⁺ or NH₄ ⁺-forming fertilisers; this is especially so for crops such as intensively managed grass where there are several applications of fertiliser nitrogen per season, as the effect of inhibitors applied in April persists until after a second fertiliser application in June. Received: 30 August 1996     

Behaviour of cadmium and nickel in a soil amended with sewage sludge

Various urban and industrial sewage sludges were applied to a soil at two doses (50 and 100 t ha⁻¹ y⁻¹) during eight years in a field experiment. The soil was analysed at two depths (0–30 and 30–60 cm) for extractable cadmium and nickel. In general these trace metal increased with dosage. However, cadmium formed complexes with organic matter and nickel bound to iron and manganese oxides. Hence, the available fractions of these metals constituted a small proportion of the total content. The results obtained show a low risk of contamination due to the available fractions of these metals at sludges dosages of up to 100 t ha⁻¹. Copyright © 2001 John Wiley & Sons, Ltd.     

Heavy metals in mine soils amended with sewage sludge

In order to reclaim a clay quarry, a topsoil material was mixed with gravelly spoil at different ratios and with various rates of sewage sludge. The influence of three spoil/topsoil ratios (1:1, 2:1 and 3:1) and three sludge rates (40, 80 and 120 t ha⁻¹) on chemical properties of the resulting material was investigated, with emphasis on heavy metal (Fe, Cu, Mn, Ni and Zn) contents. The mixtures topsoil/spoil/sludge were water saturated and incubated for 15 or 30 days in a chamber under controlled conditions. The incubated samples were analysed for pH, total carbon and nitrogen, and total, available, exchangeable and soluble heavy metals. The addition of spoil to the topsoil increased the volume of material available, by utilizing an inert material unsuitable by itself to grow plants. The addition of sewage sludge repaired the disadvantages of the spoil, increasing the pH and the organic matter contents. The total heavy metal contents in the mixtures followed the sequence Fe>>Mn>>Zn, Cu>Ni. All except Cu were within the ranges allowed for agricultural lands. The available heavy metals constituted a small fraction of total contents and decrease with time due to complexation and immobilization processes. The exchangeable and soluble fractions were almost negligible; only small amounts of Mn, Zn and Cu were detected. Therefore, the risk of contamination by heavy metals is insignificant in the conditions investigated. Copyright © 1999 John Wiley & Sons, Ltd.     

Nitrous oxide emissions from boreal organic soil under different land-use

Nitrous oxide emissions were studied with a static chamber technique during 2 years from a drained organic soil in eastern Finland. After drainage, the soil was forested with birch (Betula pendula Roth) and 22 years later, part of the forest was felled and then used for cultivation of barley (Hordeum vulgare L.) and grass. The annual N₂O emissions from the cultivated soil (from 8.3 to 11.0 kg N₂O-N ha⁻¹ year⁻¹) were ca. twice the annual emission from the adjacent forest site (4.2 kg N₂O-N ha⁻¹ year⁻¹). The N₂O emissions from the soils without plants (kept bare by regular cutting or tilling) were also lower (from 6.5 to 7.1 kg N₂O-N ha⁻¹ year⁻¹) than those from the cultivated soil. There was a high seasonal variation in the fluxes with a maximum in spring and early summer. The N₂O fluxes during the winter period accounted for 15-60% of the total annual emissions. N₂O fluxes during the snow-free periods were related to the water table (WT) level, water-filled pore space, carbon mineralisation and the soil temperature. A linear regression model with CO₂ production, WT and soil temperature at the depth of 5 cm as independent variables explained 54% of the variation in the weekly mean N₂O fluxes during the snow-free periods. N₂O fluxes were associated with in situ net nitrification, which alone explained 58% of the variation in the mean N₂O fluxes during the snow-free period. The N₂O-N emissions were from 1.5 to 5% of net nitrification. The acetylene blockage technique indicated that most of the N₂O emitted in the snow-free period originated from denitrification.     

Nitrous oxide emissions from grassland in an intensive dairy farm in the Basque Country of Spain

Abstract. Intensively managed grasslands are potentially a large source of N₂O in the North Coast of Spain because of the large N input, the wet soil conditions and mild temperatures. To quantify the effect of fertilizer type and management practices carried out by farmers in this area, field N₂O losses were measured over a year using the closed chamber technique. Plots received two types of fertilizer: cattle slurry (536 kg N ha^–1) and calcium ammonium nitrate (140 kg N ha^–1). N₂O losses were less in the slurry treatment than after mineral fertilizer. This was probably due to high, short‐lived peaks of N₂O encountered immediately following mineral N addition. In contrast, the seasonal distribution of N₂O losses from the slurry amended plot was more uniform over the year. The greater N₂O losses in the mineral treatment might have been enhanced by the combined effect of mineral fertilizer and past organic residues present from previous organic amendments. Weak relationships were found between N₂O emission rates and soil nitrate, soil ammonium, soil water content and temperature. Better relationships were obtained in the mineral treatment than in the slurry plots, because of the wider range in soil mineral N. Water filled pore space (WFPS) was a key factor controlling N₂O emissions. In the > 90% WFPS range no relationships were found. The best regressions were found for the mineral treatment in the 40–65% WFPS range, 49% of the variance being explained by soil nitrate and ammonium content. In the 65–90% WFPS range, 43% of the variance was explained by nitrate only, but the inclusion of soil ammonium did not improve the model as it did in the 40–65% WFPS range. This fact indicates that nitrification is likely to be an important process involved in N₂O emissions at the 40–65% WFPS.     

Effects of sewage sludge stabilization on fertilizer value and greenhouse gas emissions after soil application

Application of sewage sludge on agricultural land becomes more and more common in many parts of the world in order to recycle the nutrients from the sludge. A range of sewage sludge stabilization techniques are available to make the sludge more stable prior to storage, transportation, and application. These stabilization techniques include dewatering, drying, anaerobic digestion, composting, and reed bed sludge treatment. However, very few studies have investigated the effect of these techniques after the sludge has been applied to agricultural land. The objective of the current study was therefore to investigate the effect of sewage sludge stabilization techniques on the C and N mineralization and gaseous emissions from soil. A soil incubation was conducted to determine the rate of C and N mineralization and N₂O and CH₄ emissions of sewage sludge stabilized using different techniques. Unstabilized sludge released up to 90% of their C content as CO₂, part of which could be caused by release of CO₂ from carbonates. Compared with this, sludge stabilization including anaerobic digestion and drying resulted in a reduction of the C mineralization rate of about 40%. Liming reduced C mineralization with around 29%, while treatment in a reed bed system reduced it by 74%. The current study thus clearly demonstrated that stabilization techniques resulted in sludge that was more stable once they were applied to agricultural land. Stabilization also reduced the N immobilization phase, potentially improving the value of the sludge as a fertilizer. Emissions of CH₄ were also reduced through sludge stabilization and mainly occurred after application of easily degradable sludge types, which is likely to have enhanced the creation of anaerobic microsites. The stabilization processes also decreased emissions of N₂O. The results for both CH₄ and N₂O indicate that the stabilization tends to reduce the chance of developing conditions where these gases could be produced.     

不同水稻、小麦品种对N2O排放的影响

Plant species of cropping systems may affect nitrous oxide (N2O) emissions. A field experiment was conducted to investigate dynamics of N2O emissions from rice-wheat fields from December 2006 to June 2007 and the relationship between soil and plant parameters with N2O emissions. The results indicated that N2O emissions from different wheat varieties ranged from 12 to 291 μg N2O-N m-2 h-1 and seasonal N2O emissions ranged from 312 to 385 mg N2O-N m-2. In the rice season, it was from 11 to 154 μg N2O-N m-2 h-1 with seasonal N2O emission of 190--216 mg N2O-N m-2. The seasonal integrated flux of N2O differed significantly among wheat and rice varieties. The wheat variety HUW 234 and rice variety Joymoti showed higher seasonal N2O emissions. In the wheat season, N2O emissions correlated with soil organic carbon (SOC), soil NO3--N, soil temperature, shoot dry weight, and root dry weight. Among the variables assessed, soil temperature followed by SOC and soil NO3--N were considered as the important variables influencing N2O emission. N2O emission in the rice season was significantly correlated with SOC, soil NO3--N, soil temperature, leaf area, shoot dry weight, and root dry weight. The main driving forces influencing N2O emission in the rice season were soil NO3--N, leaf area, and SOC.     

Potential contribution of Lumbricus terrestris L. to carbon dioxide, methane and nitrous oxide fluxes from a forest soil

 Potential effects of earthworms (Lumbricus terrestris L.) inoculated into soil on fluxes of CO₂, CH₄ and N₂O were investigated for an untreated and a limed soil under beech in open topsoil columns under field conditions for 120 days. Gas fluxes from L. terrestris, beech litter and mineral soil from soil columns were measured separately in jars at 17  °C. The inoculation with L. terrestris and the application of lime had no effect on cumulative CO₂ emissions from soil. During the first 3–4 weeks earthworms significantly (P<0.05) increased CO₂ emissions by 16% to 28%. In contrast, significantly lower (P<0.05) CO₂ emission rates were measured after 11 weeks. The data suggest that earthworm activity was high during the first weeks due to the creation of burrows and incorporation of beech litter into the mineral soil. Low cumulative CH₄ oxidation rates were found in all soil columns as a result of CH₄ production and oxidation processes. L. terrestris with fresh feces and the beech litter produced CH₄ during the laboratory incubation, whereas the mineral soil oxidised atmospheric CH₄. Inoculation with L. terrestris led to a significant reduction (P<0.02) in the CH₄ oxidation rate of soil, i.e. 53% reduction. Liming had no effect on cumulative CH₄ oxidation rates of soil columns and on CH₄ fluxes during the laboratory incubation. L. terrestris significantly increased (P<0.001) cumulative N₂O emissions of unlimed soil columns by 57%. The separate incubation of L. terrestris with fresh feces resulted in rather high N₂O emissions, but the rate strongly decreased from 54 to 2 μg N kg^–1 (dry weight) h^–1 during the 100 h of incubation. Liming had a marked effect on N₂O formation and significantly (P<0.001) reduced cumulative N₂O emissions by 34%. Although the interaction of liming and L. terrestris was not significant, N₂O emissions of limed soil columns with L. terrestris were 8% lower than those of the control. Received: 2 September 1999     

Nitrous oxide emissions from wetland soil amended with inorganic and organic fertilizers

Agricultural soils are a primary source of anthropogenic trace gas emissions, and the subtropics contribute greatly, particularly since 51% of world soils are in these climate zones. A field experiment was carried out in an ephemeral wetland in central Zimbabwe in order to determine the effect of cattle manure (1.36% N) and mineral N fertilizer (ammonium nitrate, 34.5% N) application on N₂O fluxes from soil. Combined applications of 0 kg N fertilizer + 0 Mg cattle manure ha^?1 (control), 100 kg N fertilizer + 15 Mg manure ha^?1 and 200 kg N fertilizer + 30 Mg manure ha^?1 constituted the three treatments arranged in a randomized complete block design with four replications. Tomato and rape crops were grown in rotation over a period of two seasons. Emissions of N₂O were sampled using the static chamber technique. Increasing N fertilizer and manure application rates from low to high rates increased the N₂O fluxes by 37–106%. When low and high rates were applied to the tomato and rape crops, 0.51%, 0.40%, and 0.93%, 0.64% of applied N was lost as N₂O, respectively. This implies that rape production has a greater N₂O emitting potential than the production of tomatoes in wetlands.     

Nitrous oxide emissions under different soil and land management conditions

Nitrous oxide (N₂O) emissions of three different soils – a rendzina on cryoturbed soil, a hydromorphic leached brown soil and a superficial soil on a calcareous plateau – were measured using the chamber method. Each site included four types of land management: bare soil, seeded unfertilized soil, a suboptimally fertilized rapeseed crop and an overfertilized rapeseed crop. Fluxes varied from –1g to 100g N₂O-nitrogen ha^–1 day^–1. The highest rates of N₂O emissions were measured during spring on the hydromorphic leached brown soil which had been fertilized with nitrogen (N); the total emissions during a 5-month period exceeded 3500gNha^–1. Significant fluxes were also observed during the summer. Very marked effects of soil type and management were observed. Two factors – the soil hydraulic behaviour and the ability of the microbial population to reduce N₂O – appear to be essential in determining emissions of N₂O by soils. In fact, the hydromorphic leached brown soil showed the highest emissions, despite having the lowest denitrification potential because of its water-filled pore space and low N₂O reductase activity. Soil management also appears to affect both soil nitrate content and N₂O emissions. Received: 4 April 1997     

Effects of disturbance and glucose addition on nitrous oxide and carbon dioxide emissions from a paddy soil

The effects of disturbance and glucose addition on N₂O and CO₂ emissions from a paddy soil at 45% WFPS (water-filled pore space) and at 25 °C were determined. During a 45-day incubation, disturbances with and without glucose addition were imposed 0, 1, 3, and 5 times. The total amount of glucose added to soil with 1, 3, and 5 disturbances was equal (0.6% of oven-dry soil basis). Strong nitrification occurred in the paddy soil during the incubation. Disturbance alone did not influence N₂O and CO₂ emissions significantly, but disturbance with glucose addition did (P < 0.01). A flush of N₂O as well as CO₂ was always observed following disturbance with glucose addition. The discrepancy in N₂O emission between disturbance alone and disturbance with glucose addition was ascribed to the different magnitude of denitrification and/or heterotrophic nitrification. Greater cumulative emission of N₂O was observed in the treatment of three disturbance times with glucose addition (4.3 mg N kg⁻¹ soil), compared with five disturbances with glucose addition (2.5 mg N kg⁻¹ soil) and one disturbance with glucose addition (2.5 mg N kg⁻¹ soil). Cumulative CO₂ emission was significant larger in one and three disturbances with glucose addition than that five disturbance with glucose addition. Supplies of available organic C appear to be a critical factor controlling denitrification and/or heterotrophic nitrification processes and N₂O emission under relatively low moisture conditions, i.e. 45% WFPS.     

Nitrous oxide emission from soils after incorporating crop residues

Abstract. Emissions of N₂O were measured from different agricultural systems in SE Scotland. N₂O emissions increased temporarily after fertilization of arable crops, cultivation of bare soil, ploughing up of grassland and incorporation of arable and horticultural crop residues, but the effect was short-lived. Most of the emission occurred during the first two weeks, returning to 'background' levels after 30–40 days. The highest flux was from N-rich lettuce residues, 1100 g N₂O-N ha⁻¹ being emitted over the first 14 days after incorporation by rotary tillage. The magnitude and pattern of emissions was strongly influenced by rainfall, soil mineral N, cultivation technique and C:N ratio of the residue. Comparatively large emissions were measured after incorporation of material with low C:N ratios. Management practices are recommended that would increase N-use efficiency and reduce N₂O emissions from agricultural soils.     

Initial organic matter transformation of soil amended with composted sewage sludge

 An incubation experiment with composted sewage sludge (CSS) just added to the soil was conducted to determine its initial effects on C decomposition, N nitrification and the transformation of organic matter. CSS was mixed with a sandy loam soil from uncultivated ochric epipedon of a Typic Haploxeralf at rates of 0, 40 and 80 t ha^–1 (dry weight). The data obtained showed that with regard to the unamended soil, both the 40 and the 80 t ha^–1 treatments produced the same result in decreasing respiratory activity, but the addition of increasing amounts of CSS progressively delayed C decomposition. The nitrification index (NI), defined as the relation between nitrate-N and nitrate-N + ammonium-N, increased in correlation with the C mineralization coefficient. Total organic matter decreased after incubation whereas the humic substances increased in relation to the total C mineralized. Received: 28 October 1999