Alleviating soil sickness caused by aerobic monocropping: Responses of aerobic rice to various nitrogen sources

Abstract

Yield decline resulting from continuous cropping of aerobic rice is a constraint to the widespread adoption of aerobic rice technology. Shifts in water management from flooded to aerobic conditions are known to influence the availability and form of N present in the soil and might require a different approach to N management in aerobic rice. The present study was conducted to determine the effects of different N sources on the plant growth and grain yield of aerobic rice. Four pot experiments were conducted in which rice was aerobically grown in soil that was taken from fields where aerobic rice has been cultivated for 11 consecutive seasons and an adjacent field where flooded rice has been grown continuously. Nitrogen was applied as ammonium sulfate, urea, ammonium chloride, ammonium nitrate and potassium nitrate at four N rates of 0.3, 0.6, 0.9 and 1.2 g N pot⁻¹. Two unfertilized controls consisting of soil that was either untreated or oven heated at 120°C for 12 h were also included. Plants were sampled during the vegetative stage or at maturity to measure plant growth, N uptake, grain yield and the yield components. Growth of aerobic rice in aerobic soil was generally better with the application of ammonium-N than nitrate-N. Potassium nitrate decreased plant growth and caused plant death at the high N rate. Ammonium sulfate was more effective in improving the vegetative plant growth, N nutrition and grain yield of aerobic rice than urea at the high N rates. The application of ammonium sulfate achieved the same and even better plant growth than the soil oven-heating treatment. These results suggest that there is a possibility of reversing the yield decline observed in the continuous aerobic rice system by using the right source of N fertilizer at the optimal rate.

Key words:

• aerobic rice
• ammonium sulfate
• nitrogen source
• soil sickness
• urea

INTRODUCTION

Increasing scarcity of freshwater resources for agriculture has prevented farmers from growing flooded rice in many areas (Tuong and Bouman 2003). Various water-saving technologies, such as aerobic rice, have been developed to help farmers cope with water scarcity in irrigated environments (Bouman et al. 2007). Aerobic rice is input-responsive and high-yielding rice that is grown under dryland conditions, much like other cereal crops (Bouman and Tuong 2001). Depending on the environment, aerobic rice can be grown with or without supplemental irrigation and aims to save water by limiting losses resulting from seepage, percolation and evaporation. Relative to flooded rice, aerobic rice generally requires 30–50% less water, but also tends to result in a yield penalty of 20–30% (Yang et al. 2005). Using currently available germplasm, grain yields of 5–6 t ha⁻¹ can be achieved with aerobic rice (George et al. 2002).

At present, aerobic rice is grown commercially on 80,000 ha in northern China (Wang et al. 2002); however, to improve the yield potential and the yield stability of aerobic rice, high-yielding aerobic varieties and sustainable crop management systems must be developed. These objectives are particularly relevant in the tropics, where several studies have reported that continuous monocropping of aerobic rice can result in rapid declines in the yield (George et al. 2002; Guimaraes and Stone 2000; Nishizawa et al. 1971; Peng et al. 2006; Ventura and Watanabe 1978). The yield decline of continuous aerobic rice is generally believed to be caused by soil sickness, which includes biotic factors, such as nematodes and soil pathogens, and abiotic factors, such as nutrient deficiency and toxicity (Lin et al. 2002; Nishio and Kusano 1975; Nishizawa et al. 1971; Ventura et al. 1981).

Table 1 Nitrogen sources, N rates, soil, varieties and sampling dates in the four pot experiments

Experiment	Nitrogen source	Nitrogen rate (g pot^−1)	Soil	Variety	Sampling stage
1	Ammonium sulfate, urea, ammonium chloride, ammonium nitrate and potassium nitrate	0.3, 0.6, 0.9 and 1.2	Aerobic soil^†	Apo	Vegetative
2	Ammonium sulfate and urea	0.3, 0.6, 0.9 and 1.2	Aerobic soil^†	Apo	Maturity
3	Ammonium sulfate and urea	1.2	Aerobic soil^† and flooded soil^‡	Apo	Vegetative
4	Ammonium sulfate	1.2	Aerobic soil^†	Apo, IR78877-208-B-1-2 and IR80508-B-57-3-B	Vegetative
^†Soil from a field where aerobic rice has been grown continuously for 11 seasons since 2001 in both dry and wet seasons.
^‡Soil from a field where flooded rice has been grown continuously.

Shifts in water management from flooded to aerobic conditions are known to influence the availability and form of N present in the soil and might require a different approach to N management for aerobic rice (Savant and De Datta 1982). Nie et al. (2008) conducted a series of pot experiments studying the individual effects of nutrients and reported that N application improved plant growth under continuous aerobic rice cropping, whereas P, K and micronutrients had little or no effect. These results suggested that N deficiency as a result of poor N availability and reduced plant N uptake might be important factors contributing to the yield decline of continuously cropped aerobic rice.

A number of researchers have reported different responses in plant growth to various N forms under upland or aerobic conditions. In general, NH⁺ ₄ is the dominant N form in paddy fields (Savant and De Datta 1982) and NO^- ₃ is only stable in the oxidized rhizosphere (Shen 1969). Lin et al. (2005) reported that aerobic rice produced higher shoot dry matter under a sole NO^- ₃ supply than a sole NH⁺ ₄ supply during the early growth stages and that co-provision of ammonium and nitrate could have a synergistic effect on the total N fluxes compared with the supply of either ammonium or nitrate alone. Qian et al. (2004) found that aerobic rice did not show a significant preference for NH⁺ ₄ and NO^- ₃ alone and that provision of both N forms improved the growth and N use efficiency.

The effectiveness of different N sources on alleviating the soil sickness caused by the monocropping of aerobic rice is not clear. In the present study, a series of pot experiments were conducted using soils collected from two adjacent fields at the International Rice Research Institute (IRRI) farm: an aerobic field where aerobic rice has been grown continuously for 11 seasons and a flooded field where flooded rice has been grown continuously. The objectives of the present study were: (1) to determine the effects of different N sources on the plant growth and grain yield of aerobic rice grown in the continuous aerobic rice soil, (2) to compare the response of plants in aerobic and flooded soils to the application of ammonium sulfate and urea, (3) to examine differences among rice varieties in their growth responses to ammonium sulfate application in continuous aerobic rice soil.

MATERIALS AND METHODS

Four pot experiments were conducted in the greenhouse at the IRRI using soil taken from the top 25 cm of the aerobic and flooded rice fields at the research farm of IRRI (Table 1). The aerobic soil for all four experiments was collected from a field where aerobic rice has been grown for 11 consecutive seasons since 2001 and where a gradual decline in yield has been observed (Peng et al. 2006). The flooded soil for experiment 3 was collected from an adjacent field where only flooded rice has been grown. The aerobic rice soil was Aquandic Epiaquoll with pH 7.1, 16.4 g kg⁻¹ organic C, 1.74 g kg⁻¹ total N, 29.7 mg kg⁻¹ Olsen P, 393 mg kg⁻¹ available K, 40.5 meq 100 g⁻¹ cation exchange capacity, 58% clay, 33% silt and 9% sand. The flooded rice soil was Aquandic Epiaquoll with pH 6.6, 17.6 g kg⁻¹ organic C, 1.89 g kg⁻¹ total N, 13.0 mg kg⁻¹ Olsen P, 455 mg kg⁻¹ available K, 40.3 meq 100 g⁻¹ cation exchange capacity, 57% clay, 33% silt and 10% sand.

The soil was air-dried, crushed into small pieces and mixed well for the experiments. Porcelain pots (4 L) filled with 3.0 kg of air-dried soil were used in all four pot experiments. One day before sowing, chemical fertilizers were applied to the pots and mixed well with the soil and then the soil was soaked with tap water. An improved upland rice variety, Apo, was used in all experiments because of its good performance under aerobic conditions (George et al. 2002; Lafitte et al. 2002).

In experiment 1, the treatments consisted of five N sources, ammonium sulfate, urea, ammonium chloride, ammonium nitrate and potassium nitrate, at four N rates, 0.3, 0.6, 0.9 and 1.2 g N pot⁻¹. In experiment 2, two N sources (ammonium sulfate and urea) were used at the same N rates as in experiment 1. In experiment 3, both aerobic and flooded soils received ammonium sulfate and urea at a N rate of 1.2 g pot⁻¹. In experiment 4, two newly developed rice varieties were compared with Apo under the application of ammonium sulfate at a rate of 1.2 g N pot⁻¹. The two new varieties were IR80508-B-57-3-B and IR78877-208-B-1-2, which were developed from crossing Apo with Aus257 and IR72, respectively. Two common treatments were included in all experiments: an untreated control and a soil oven-heating treatment. For the soil-heating treatment, pots with soil were placed inside an oven and heated at 120°C for 12 h. Both treatments did not receive nutrient inputs.

Each treatment was replicated five times with one pot per replicate. The distance between the pots was kept at 30 cm to avoid shading. Six pre-germinated seeds were sown into each pot on 28 May 2006, 8 February 2007, 25 July 2006 and 22 March 2007 for experiments 1–4, respectively. Seedlings were thinned 1 week after sowing to three uniform seedlings per pot. The pots were kept saturated for 1 week after sowing to promote good crop establishment after which the pots were kept under aerobic conditions. All pots were watered once every 1–3 days whenever drying of the soil surface was observed, which corresponded to a soil moisture tension at a depth of 15 cm of −15 to −25 kPa. The soil water content in the pots was not controlled rigorously, but frequent irrigation ensured that the plants did not experience drought stress and no standing water was kept in the pots throughout the experiment. Pesticides were sprayed on three to four occasions to control insect damage. Weeds were removed manually.

The plants were sampled at 37, 40 and 40 days after sowing for experiments 1, 3 and 4, respectively. Before plant sampling, the number of stems per pot was counted and the plant height from the plant base to the tallest leaf tip in each pot was measured. Plants were separated into leaves and stems including sheath. The leaf area was measured with a leaf area meter (LI-3100; Li-cor, Lincoln, NB, USA). Plants were harvested at maturity in experiment 2. Panicle number per pot was counted. Plants were separated into leaves, stems including sheath, rachis and filled and unfilled spikelets. The dry weights of the plant organs were determined after oven-drying at 70°C to a constant weight in all four experiments. The above-ground biomass was the sum of the leaf and stem dry weights in experiments 1, 3 and 4. In experiment 2, the above-ground biomass included the dry weights of the leaves, stems, rachis and filled and unfilled grains. The spikelets per panicle, grain-filling percentage (100 × filled spikelet number/total spikelet number) and harvest index (filled spikelet weight/above-ground total biomass) were calculated for experiment 2. Three chlorophyll meter (SPAD) readings were taken from one of the topmost fully expanded leaves per pot at 39 days after sowing in experiments 1 to 4.

The tissue N concentration was determined using the micro Kjeldahl digestion, distillation and titration (Bremner and Mulvaney 1982) to calculate the above-ground N uptake. The data were analyzed using an anova (SAS Institute 2003) and a least significant difference (LSD) test was used to compare the means between the treatments (P < 0.05).

RESULTS

In experiment 1, the application of ammonium sulfate and urea improved all measures of plant growth and N nutrition compared with the untreated control (Fig. 1). The application of ammonium chloride and ammonium nitrate increased the plant height and SPAD value compared with the control. For stem number, leaf area, above-ground biomass and above-ground N uptake, ammonium chloride was effective only at the highest N rate, whereas ammonium nitrate had little or no effect on these parameters compared with the control. The application of potassium nitrate decreased plant growth and caused plant death at the rate of 1.2 g N pot⁻¹. For the application of ammonium sulfate and urea, increasing N rates from 0.3 to 1.2 g N pot⁻¹ generally improved plant growth and N nutrition and the N response was greater with ammonium sulfate than with urea. Plants responded to the application of ammonium chloride only up to 0.9 g N pot⁻¹. The application of ammonium sulfate and ammonium chloride at higher rates resulted in similar SPAD values to the oven-heating treatment. For the other five parameters, only the application of ammonium sulfate at the highest rate was close to or greater than the oven-heating treatment.

In experiment 2, the application of ammonium sulfate at all N rates improved the yield traits and N nutrition compared with the untreated control (Fig. 2). However, the application of urea at 1.2 g N pot⁻¹ did not increase the grain filling percentage and reduced the harvest index compared with the control. Increasing the rate of ammonium sulfate from 0.3 to 1.2 g N pot⁻¹ increased the panicle number, spikelets per panicle, above-ground biomass, grain yield, SPAD value and above-ground N uptake. Spikelets per panicle and grain yield were highest at a rate of 0.9 g N pot⁻¹ when N was applied as urea. Plants did not respond to N rates of ammonium sulfate and urea consistently in grain filling percentage and harvest index. At rates of 0.3 and 0.6 g N pot⁻¹, ammonium sulfate and urea had similar effects on all parameters. At higher N rates, plant growth was better with ammonium sulfate than with urea and the differences were greatest in the 1.2 g N pot⁻¹ treatment. At a rate of 1.2 g N pot⁻¹, the grain yield with ammonium sulfate application was 41.9 g pot⁻¹ compared with 12.9 g pot⁻¹ with the application of urea. The above-ground N uptake with ammonium sulfate and urea was 783 and 543 mg pot⁻¹, respectively. Applications of ammonium sulfate at 0.6–1.2 g N pot⁻¹ and of urea at 0.9 g N pot⁻¹ resulted in higher grain yields than the oven-heating treatment. This was mainly because of the differences in above-ground biomass, N uptake, panicle number and spikelets per panicle between the treatments.

Figure 1  (a) Plant height, (b) stem number, (c) leaf area, (d) above-ground biomass, (e) SPAD value and (f) above-ground N uptake of Apo grown aerobically in soil under five N sources (ammonium sulfate, urea, ammonium chloride, ammonium nitrate and potassium nitrate) at four N rates (N1–N4 = 0.3, 0.6, 0.9 and 1.2 g N pot−1, respectively) and in an untreated control (CK) and oven-heated soil treatment in pot experiment 1. The soil was from an aerobic field where aerobic rice has been grown continuously for 11 seasons. Oven heating of the soil was done at 120°C for 12 h. Error bars represent the standard error.

Figure 2  (a) Panicle number per pot, (b) spikelets per panicle, (c) grain filling, (d) above-ground biomass, (e) harvest index, (f) grain yield, (g) SPAD value and (h) above-ground N uptake of Apo grown aerobically in soil with four rates of ammonium sulfate or urea application (N1–N4 = 0.3, 0.6, 0.9 and 1.2 g N pot−1, respectively) and in an untreated control (CK) and oven-heated soil treatment in pot experiment 2. The soil was from an aerobic field where aerobic rice has been grown continuously for 11 seasons. Oven heating of the soil was done at 120°C for 12 h. Error bars represent the standard error.

Table 2 Plant growth of Apo grown aerobically under the application of ammonium sulfate and urea and oven heating of the soil compared with the untreated control in pot experiment 3

Parameter	Control	Ammonium sulfate^†	Urea^†	Oven^‡
Aerobic soil
Plant height (cm)	45.2 c	77.5 a	58.7 b	80.8 a
Stem number per pot	4.7 d	19.7 b	8.8 c	22.2 a
Leaf area (cm^2 pot^−1)	105 d	951 b	300 c	1,423 a
Above-ground biomass (g pot^−1)	0.96 d	8.51 b	2.70 c	13.98 a
SPAD value	27.7 c	37.4 a	34.4 b	33.7 b
Above-ground N uptake (mg pot^−1)	21 c	266 a	81 b	249 a
Flooded soil
Plant height (cm)	68.8 c	75.7 ab	74.3 b	79.2 a
Stem number per pot	11.5 b	21.3 a	14.3 b	20.7 a
Leaf area (cm^2 pot^−1)	589 b	923 a	690 b	1,106 a
Above-ground biomass (g pot^−1)	5.44 c	8.09 ab	6.28 bc	9.21 a
SPAD value	34.6 b	37.9 a	37.3 a	37.7 a
Above-ground N uptake (mg pot^−1)	121 c	260 a	196 b	262 a
^†The rate of ammonium sulfate and urea was 1.2 g N per pot
^‡Oven heating of the soil was done at 120°C for 12 h. Within a row, means followed by different letters are significantly different at P < 0.05 according to a least significant difference test. The aerobic soil was from a field where aerobic rice has been grown continuously for 11 seasons. The flooded soil was from an adjacent field where flooded rice has been grown continuously.

In experiment 3, plants were aerobically grown in aerobic or flooded soils under the application of ammonium sulfate, urea and oven heating of the soil in comparison with the untreated control. The application of ammonium sulfate at 1.2 g N pot⁻¹ improved plant growth and N nutrition compared with the untreated control in both aerobic and flooded soils (Table 2). The application of urea at 1.2 g N pot⁻¹ improved plant growth and N nutrition compared with the control in the aerobic soil. In the flooded soil, urea application increased the plant height, SPAD value and above-ground N uptake of aerobic rice compared with the control. In the control treatment, plant growth was much poorer in the aerobic soil than in the flooded soil. For example, the above-ground biomass of the control plants grown in the flooded soil was 4.7-fold greater than that of the control plants grown in the aerobic soil at 40 days after sowing. The application of ammonium sulfate resulted in similar plant growth between the aerobic and flooded soils, but this was not the case with urea. Overall, plants grown aerobically in both aerobic and flooded soils responded more to the application of ammonium sulfate than to urea. The response of aerobic rice to N application was greater in the aerobic soil than in the flooded soil. Plants responded to oven heating of the soil more in the aerobic soil than in the flooded soil. In the aerobic soil, oven heating resulted in greater stem number, leaf area, above-ground biomass and N uptake than the application of ammonium sulfate. In the flooded soil, there was no difference between the oven-heating treatment and ammonium sulfate application in plant growth and N nutrition. Plants produced more leaf area and above-ground biomass in the aerobic soil than in the flooded soil when the soils were oven heated.

In experiment 4, the application of ammonium sulfate at 1.2 g N pot⁻¹ and oven heating of the soil consistently improved plant growth and N nutrition compared with the untreated control across all three varieties (Fig. 3). The effect of ammonium sulfate application on plant growth and N nutrition was greater than that of oven heating of the soil in all three varieties. IR80508-B-57-3-B had higher plant height and leaf area than IR78877-208-B-1-2 and Apo in all three treatments. The stem numbers of IR80508-B-57-3-B and IR78877-208-B-1-2 were greater than that of Apo in the control and ammonium sulfate application, whereas there was no difference in stem number among the three varieties in the soil oven-heating treatment. IR80508-B-57-3-B and IR78877-208-B-1-2 produced greater above-ground biomass than Apo in all three treatments. There were inconsistent differences among the three varieties with regard to the SPAD value across the treatments. In general, IR80508-B-57-3-B had the highest above-ground N uptake, followed by IR78877-208-B-1-2 and Apo.

Figure 3  (a) Plant height, (b) stem number, (c) leaf area, (d) aboveground biomass, (e) SPAD value and (f) above-ground N uptake of three rice varieties grown aerobically under untreated, ammonium-sulfate fertilized soil and oven-heated soil in pot experiment 4. The soil was from an aerobic field where aerobic rice has been grown continuously for 11 seasons. The three varieties were Apo, IR78877-208-B-1-2 and IR80508-B-57-3-B. The rate of ammonium sulfate application was 1.2 g N pot−1. Oven heating of the soil was done at 120°C for 12 h. Error bars represent the standard error.

Table 3 Apparent N recovery rate and root dry weight of Apo grown aerobically under the application of ammonium sulfate and urea in pot experiment 1 (vegetative stage) and 2 (maturity)

Nitrogen rate (g N pot^−1)	Vegetative stage		Maturity
	Ammonium sulfate	Urea	Ammonium sulfate	Urea
Apparent N recovery rate
0.3	0.11 a	0.11 a	0.45 a	0.42 a
0.6	0.22 a	0.11 b	0.43 a	0.41 a
0.9	0.18 a	0.07 b	0.53 a	0.42 b
1.2	0.24 a	0.07 b	0.56 a	0.36 b
Root dry weight (g pot^−1)
0.3	0.92 a	0.95 a	4.1 a	3.5 a
0.6	1.39 a	0.74 b	5.2 b	7.2 a
0.9	1.30 a	0.81 a	6.8 a	6.6 a
1.2	2.00 a	0.76 b	6.5 b	9.8 a
Within a row under each sampling stage, means followed by different letters are significantly different at at P < 0.05 according to a least significant difference test. The aerobic soil was from a field where aerobic rice has been grown continuously for 11 seasons.

Increasing the rates of N as ammonium sulfate significantly increased the apparent N recovery rate (ANR) in both the vegetative and mature stages, whereas plants did not show any increase in ANR as N rates increased when N was applied as urea (Table 3). At higher N rates, ANR was significantly higher with ammonium sulfate application than with urea at the same N rate. In general, root dry weight increased as the rates of N application increased, except for the treatment of urea at the vegetative stage. At the vegetative stage, root dry weight was generally higher with ammonium sulfate application than with urea at higher N rates, but the reverse was true at maturity.

DISCUSSION

The greater effect of oven heating on plant growth in the aerobic soil than in the flooded soil (Table 2) suggests that the aerobic soil used in the present study was sick compared with the flooded soil. This result is consistent with findings from our previous studies (Nie et al. 2007, 2008). Nie et al. (2008) reported that N application alleviated the soil sickness of aerobic soil where aerobic rice has been grown continuously, suggesting that N deficiency as a result of poor soil N availability or reduced plant N uptake was associated with the yield decline of monocropped aerobic rice. This is supported by the findings of Belder et al. (2005), who showed that plant ¹⁵N recoveries were lower in aerobic rice than in flooded rice at all times of urea-N application within the same field experiment. The present study demonstrated that the aerobic rice plants grown in “sick” soil responded differently to various N sources.

Among the five N sources, both ammonium sulfate and urea enhanced vegetative plant growth at all N rates in the aerobic soil compared with the control (Fig. 1). Ammonium chloride had a positive effect on plant growth only at high N rates, whereas ammonium nitrate had little or no effect on plant growth regardless of the N rate. We observed that potassium nitrate significantly reduced plant growth and caused plant death at a rate of 1.2 g N pot⁻¹. These results suggest that ammonium-N was more effective than nitrate-N in alleviating the soil sickness caused by the monocropping of aerobic rice. However, Lin et al. (2005) documented that aerobic rice produced significantly lower above-ground biomass under sole NH⁺ ₄-N supply than under sole NO^- ₃–N and mixed N supply.

In general, ammonium sulfate was much more effective in improving vegetative plant growth and N nutrition than urea in the aerobic soil (Fig. 1). Furthermore, ammonium sulfate had a greater positive effect on reproductive plant growth, reflected by grain yield and yield components, than urea in the aerobic soil (Fig. 2). The difference between ammonium sulfate and urea became greater as the N rates increased. These results were supported by the fact that ANR was significantly higher with ammonium sulfate than with urea at 0.6–1.2 g N pot⁻¹ in the vegetative stage and at 0.9–1.2 g N pot⁻¹ at maturity (Table 3). Stephen and Waid (1963) reported that the application of ammonium sulfate and urea at low rates gave similar yields in various upland crops, whereas at intermediate and high rates of N application, yields with urea were considerably lower than the yields with ammonium sulfate. The reason why ammonium sulfate was more effective than urea in alleviating the soil sickness caused by the monocropping of aerobic rice is not clear. The soil pH values of the control, urea, ammonium sulfate and oven-heating treatments measured at 2 weeks after the application of 1.2 g N per 3.0 kg air-dried soil and wet incubation without plant growth were 6.93 (± 0.01 standard deviation), 6.65 (± 0.02), 6.29 (± 0.04) and 7.29 (± 0.04), respectively. Therefore, the application of ammonium sulfate and urea to the aerobic soil reduced the soil pH and the reduction was greater for ammonium sulfate than for urea. However, oven heating of the soil increased the soil pH. van Asten et al. (2005) reported that rice N uptake and ANR were significantly higher on a pH-neutral soil than on an alkaline soil. Changes in the nutrient availability and microbial community caused by the acidification of soil as a result of the application of ammonium sulfate could be associated with its greater effect on plant growth and this hypothesis should be tested in future research. Both soil acidification and urea-induced ammonia toxicity (Bremner and Krogmeier 1989) might explain the large difference in ANR in the vegetative stage between the application of ammonium sulfate and urea at high rates. Our study suggests that ammonium sulfate is a better N source than urea for the basal N application of aerobic rice established with direct seeding.

The positive effect of ammonium sulfate application on vegetative plant growth was observed not only in Apo, but also in two other newly developed aerobic rice varieties (Fig. 3). These two new varieties produced significantly more above-ground biomass than Apo with the application of ammonium sulfate at 1.2 g N pot⁻¹. This was true even in the control without N input. We observed that the two new varieties had much larger root systems than Apo (data not shown).

The difference in vegetative plant biomass between the application of ammonium sulfate at 1.2 g N pot⁻¹ and the oven-heating treatment was inconsistent across the experiments (Figs 1, 3; Table 2). When the plants were grown until maturity, the application of ammonium sulfate as low as 0.6 g N pot⁻¹ resulted in greater above-ground biomass and grain yield compared with the oven-heating treatment (Fig. 2). This suggests that the N released from the aerobic soil after oven heating was not enough for the whole growth duration. In both experiments 1 and 2, plant growth was significantly better with the ammonium sulfate application at 1.2 g N pot⁻¹ than at 0.9 g N pot⁻¹ in the aerobic soil (Figs 1, 2). Therefore, we cannot rule out the possibility that plant growth in the aerobic soil can be further improved at a higher rate of ammonium sulfate input. The above-ground N uptake of Apo under the same treatment was quite different among the experiments. These differences might be associated with differences in radiation and temperature across the experiments.

Both biotic factors (pathogenic nematodes, fungi and bacteria) and abiotic factors (nutrient deficiency and toxicity) could cause soil sickness under continuous cropping of aerobic rice (Lin et al. 2002; Nishio and Kusano 1975; Nishizawa et al. 1971; Ventura et al. 1981). Oven heating of the soil removed some or all of the factors causing soil sickness. We observed that oven heating of continuous aerobic rice soil at 120°C for 12 h increased the release of NH⁺ ₄ by 62% without incubation compared with untreated soil (data not shown). However, the increase in NH⁺ ₄ release would not explain entirely the enhancement in plant growth and N uptake resulting from oven heating of the soil. The soil heating treatment might have also changed other soil physical and chemical properties that have an influence on rice plant growth or removed the biotic factors that limited the plant growth under the continuous monocropping of aerobic rice. The present study indicates that the application of ammonium sulfate could achieve the same and even better plant growth than the soil oven-heating treatment, suggesting that abiotic factors are more likely to cause the soil sickness associated with the continuous cropping of aerobic rice.

Conclusions

A series of pot experiments demonstrated that the plant growth of aerobic rice in the sick soil was generally better with the application of ammonium-N than nitrate-N. Furthermore, ammonium sulfate was more effective in improving the vegetative plant growth, N nutrition and grain yield of aerobic rice than urea at the high application rates in the aerobic soil. This was also true in the flooded soil, but the difference between ammonium sulfate and urea was smaller in the flooded soil than in the aerobic soil. In addition to Apo, a positive growth response of aerobic rice to the application of ammonium sulfate was observed in two other newly developed varieties. The application of ammonium sulfate could achieve the same and even better plant growth than the soil oven-heating treatment. These results suggest that there is a possibility of reversing the yield decline observed in continuous aerobic rice systems by using the right source of N fertilizer at the optimal rate.

ACKNOWLEDGMENTS

This work is part of the Consultative Group on International Agricultural Research Challenge Program on Water and Food through the project “Developing a System of Temperate and Tropical Aerobic Rice in Asia (STAR)”. The study was also supported by the National Natural Science Foundation of China (Project No. 30528005) and the 973 Project of the Ministry of Science and Technology in China (Project No. 2005CB120900).