Effect of N fertilizer top-dressing at various reproductive stages on growth, N2 fixation and yield of three soybean (Glycine max (L.) Merr.) genotypes

Yinbo Gana, , , Ineke Stulena, Herman van Keulenb, Pieter J.C Kuipera

a Laboratory of Plant Physiology, University of Groningen, P.O. Box 14, 9750 AA Haren, The Netherlands

b Agrosystems Department, Plant Research International, Wageningen University and Research center, P.O. Box 16, 6700 AA Wageningen, The Netherlands

Received 24 April 2002; revised 5 September 2002; Accepted 22 September 2002. Available online 25 October 2002.

Abstract

Soybean (Glycine max (L.) Merr.) is one of the most important food and cash crops in China and a key protein source for the farmers in northern China. Previous experiments in both the field and greenhouse have shown that N2 fixation alone cannot meet the N requirement for maximizing soybean yield, and that N top-dressing at the flowering stage was more efficient than N top-dressing at the vegetative stages. However, the effect of N fertilizer application at other reproductive stages of soybean is unknown. Thus, a field experiment was conducted to study the effects of N applications at various reproductive stages on growth, N2 fixation and yield of three soybean genotypes. The results showed that starter N at 25 kg ha−1 resulted in minimum yield, total N accumulation and total amount of N2 fixed in all three genotypes. N top-dressing at 50 kg ha−1 at either the V2 or R1 stages, significantly increased N accumulation, yield and total amount of N2 fixed in all three genotypes. However, N top-dressing at the same rate at either the R3 or R5 stage did not show this positive effect in any of the three genotypes. Thus, the best timing for N top-dressing during reproduction is at the flowering stage, which increased seed yield by 21% for Wuyin 9, 27% for You 91-19, and 26% for Jufeng, respectively, compared to the treatment without N top-dressing.

Keywords: N fertilizer application; N2 fixation; Nodulation; Reproductive stage; Seed yield; Soybean

Article Outline

1. Introduction

2. Materials and methods

2.1. Field survey

2.2. Field experiment

2.3. Plant sampling and analysis

2.4. Plant N derived from N2 fixation (Pfix) and N balance

2.5. Statistics

3. Results

3.1. Total biomass accumulation

3.2. Nodule dry weight

3.3. Relative ureide content in the xylem sap

3.4. Total N accumulation, total N2 fixed and N balance

3.5. Seed yield and yield components

4. Discussion

4.1. Biomass and seed yield

4.2. Nodulation, total N uptake, N2 fixation and N balance

Acknowledgements

References

1. Introduction

Soybean (Glycine max (L.) Merr.) is an economically important crop worldwide and the most important cultivated legume in China (Gan, 1999). Although the N requirement of soybean can be met by both mineral N assimilation and symbiotic N2 fixation (Harper, 1974), soybean has a relatively high N requirement, especially at the later growth stages ( [Sinclair and de Wit, 1976] and Watanabe et al., 1986). Approximately 300 kg N is needed to produce 3 t ha−1 of soybean, whereas rice only requires 100 kg N to produce 5 t ha−1 (Hanway and Weber, 1971). Hanway and Weber (1971) reported that about half the N in mature soybean seed has been translocated from other parts of the plant, the remainder being derived from soil and nodules. Thus, soil mineral N supply may be a critical factor for soybean during the reproductive stages (George and Singleton, 1992). This high N demand during the ripening stage might be met by applying supplemental fertilizer N after flowering. To maximize yield and N2 fixation by soybean, a better understanding is required of the interactions between mineral N supply, N requirements and N uptake at the reproductive stage.

In central China, soybean is a component of major cropping systems, such as rice–soybean, soybean–rice and vegetable–soybean–vegetable. To maximize yields, farmers apply high nitrogen doses as a single dressing, before flowering, which seriously reduces symbiotic N2 fixation. The consequences of these practices are inefficient use of nitrogen fertilizer and high costs of soybean production (Gan et al., 2002c). Earlier experiments have shown that N top-dressing at the R1 stage (Fehr et al., 1971) was more efficient at vegetative stages such as V4 (farmers’ practice) and/or V6. However, limited information is available on effects of N top-dressing at other reproductive stages.

The objective of this field experiment therefore was to study effects of N top-dressing at various reproductive stages on yield, nodulation and N2 fixation in three different genotypes.

2. Materials and methods

2.1. Field survey

Informal and formal surveys were conducted at three villages of Laohekou City, Hubei Province, China before the field experiment, to describe and analyze current farmers’ practices with respect to N fertilizer application, plant density, spacing arrangement and variety selection. This information was used to design of the field experiment.

2.2. Field experiment

The field experiment was conducted at the Oil Crop Institute, Chinese Academy of Agricultural Sciences, Hubei Province, China, from April to August, 1997. The soil (pH 7.8) contained 1.8% organic matter. The experiment was irrigated and total water supply during this period was 1040 mm. Average daily temperature was 20–30 °C. Just prior to the start of the experiment, available N (mineral N, 73 ppm), P (exchangeable, 24 ppm) and K (exchangeable, 86 ppm) in the top 20 cm were measured using the procedures described by Li et al. (1984). Before sowing, the field was inoculated with Bradyrhizobium japonicum 113-2. Basal fertilizer was applied at 45 kg P2O5 ha−1, as triple super phosphate (45% P2O5) and 25 kg N ha−1 starter N as urea (46% N). The experimental design was a randomized complete block with four replicates for each treatment. Five N fertilizer (urea) treatments were applied as top-dressing: F1: 0 kg ha−1; F2: 50 kg ha−1 at the V2 stage (best timing at the vegetative stage, Gan et al., 2002c); F3: 50 kg ha−1 at the R1 stage; F4: 50 kg ha−1 at the R3 stage and F5: 50 kg ha−1 at the R5 stage (see Fehr et al., 1971 for explanation of developmental stages), plot size was m with two beds per plot, each with four rows of soybeans. Three different genotypes were included, the determinate Wuyin 9, the dominant early maturity genotype in the spring season; the semi-determinate You 91-19 and the indeterminate Jufeng, the dominant late maturity genotype in central China. Plants were arranged in a cm lattice with two plants per hill for Wuyin 9 and one plant per hill for You 91-19 and Jufeng (current farmers’ practice in central China). The experiment thus comprised 60 (five treatments×three genotypes×four replications) sub-plots.

2.3. Plant sampling and analysis

Plants were sampled six times during the growing season, starting at the vegetative stage V4, with a final sampling at physiological maturity, R7. The field was irrigated 1 week before each sampling, so that the xylem sap could be easily collected and the roots easily dug out. At each sampling, at least 12 plants per sub-plot were sampled by first removing the shoots at the cotyledon node, leaving the root stumps. Xylem sap was collected as bleeding sap from each root stump (Peoples et al., 1989) and placed immediately under −10 °C for later analysis. The roots were dug up as a 20 cm

(width)×40 cm (depth) volume. Individual roots with attached soil were placed in plastic bags for later nodule determination in the laboratory, and disconnected nodules were collected and added. In the laboratory, soil was carefully removed and individual roots washed. Nodules were then collected from the soil and the root. All plant parts were dried at 80 °C, and analyzed for total N (Bergersen, 1980). At the R7 stage, total seed yield was determined from at least

24 plants from a harvested area of 2 m2. Yield components, viz. number of nodes/plant, number of branches/node, number of pods/branch, number of seeds/pod and 1000-seed weight were determined on a randomly selected sub-sample of 10 plants.

2.4. Plant N derived from N2 fixation (Pfix) and N balance

Concentrations of ureide (allantoid and allantoic acid) in the bleeding sap were determined after Young and Conway (1942), and sap nitrate by the salicylic acid technique (Cataldo et al., 1975). Total amino acid N content of the bleeding sap was determined colorimetrically with ninhydrin, using an asparagine/glutamine standard in a 1:1 ratio (Peoples et al., 1989). The relative abundance of ureide-N in root-bleeding xylem sap was calculated after Peoples et al. (1989). The proportion of plant N derived from N2 fixation (Pfix) was calculated with the regression equations given by Peoples et al. (1989).

The potential contribution of N2 fixed to the soil N-store depends on the N balance, calculated as

, at harvest. Soybean cropping contributes to soil N enrichment only if total N2 fixation during crop growth exceeds the amount of N removed in seed (assuming straw is left in the field; Peoples et al., 1995b).

2.5. Statistics

The results were tested by analysis of variance, using the Statistix program, Analytical Software, St. Paul, MN, USA, version 3.5.

3. Results

3.1. Total biomass accumulation

The treatment with starter N at 25 kg N ha−1 yielded minimum biomass at all samplings in all three genotypes (Table 1). N top-dressing at 50 kg ha−1 at the V2 stage significantly increased crop biomass at the R1 stage, with 9, 9 and 10% increase for determinate Wuyin 9, semi-determinate You 91-19, and indeterminate Jufeng, respectively. N top-dressing at either the V2 or R1 stage, significantly increased crop biomass of You 91-19 and Jufeng at both the R5 and R6.5 stages, compared to the absence of top-dressing. However, N top-dressing at either the R3 or R5 stage, did not significantly increase biomass in any genotype. N top-dressing at the R1 stage resulted in biomass increases at final harvest of 10, 16 and 16%, for Wuyin 9, You 91-19 and Jufeng, respectively. However, N top-dressing at the R3 or R5 stage had no significant effect in either genotype.

Table 1. Effect of N fertilizer application on crop biomass (g m−2) at different growth stages of three soybean genotypes grown in Hubei, China

Growth stage treatmenta V4 R1 R3 R5 R6.5 R7

Wuyin 9

F1 20 34 81 193 246 268

F2 22 37 93 219 275 284

F3 19 34 95 210 285 294

F4 20 34 82 200 256 273

F5 20 35 81 194 264 282

You 91-19

F1 19 104 252 351 394 406

F2 20 113 268 405 443 457

F3 19 104 269 408 446 469

F4 19 104 253 378 395 408

F5 19 105 252 352 401 424

Jufeng

F1 23 143 279 378 473 488

F2 24 158 298 424 546 562

F3 23 143 302 434 551 568

F4 23 144 280 441 504 520

F5 23 143 280 378 485 530

Analysis of varianceb

P(G) <0.01 <0.01 <0.01 <0.01 <0.01 <0.01

P(N) NS <0.05 <0.01 <0.01 <0.01 <0.01

P(N × G) NS NS NS NS NS NS

LSD0.05 2 6 20 47 60 62

Full-size table

a F1: 0kgNha−1; F2: 50kgNha−1 at the V2 stage; F3: 50kgNha−1 at the R1 stage; F4: 50kgNha−1 at the R3 stage and F5: 50kgNha−1 at the R5 stage.

b G: genotype; N: nitrogen fertilizer treatment; NS: not significant.

View Within Article

3.2. Nodule dry weight

N management affected nodule growth similarly for all three genotypes during the whole growing period (Table 2). Nodule mass was greater for You 91-19 and Jufeng than for Wuyin 9 from R1 until the final harvest stage, R7. Nodule mass in all three genotypes reached its peak at the R5 stage and declined sharply at both R6.5 and R7. N top-dressing at the V2 stage resulted in maximum nodule dry weight at all samplings, being 23 and 16% higher for Jufeng and You 91-19 at the R1 stage, than without N top-dressing. On the other hand, N top-dressing at either the R1 or R3 stage resulted in significantly lower nodule dry weight at the R5 stage than without N top-dressing: 17 and 11% for Wuyin 9, 27 and 21% for You 91-19 and 26 and 20% for Jufeng, respectively. However, this difference had disappeared at final harvest.

Table 2. Effect of N fertilizer application on nodule dry weight (mg plant−1) at different growth stages of three soybean genotypes grown in Hubei, China

Growth stage treatmenta V4 R1 R3 R5 R6.5 R7

Wuyin 9

F1 49 51 75 106 82 32

F2 52 55 81 116 89 33

F3 47 50 63 88 68 31

F4 48 52 77 94 70 32

F5 45 49 73 105 70 30

You 91-19

F1 41 130 175 280 102 53

F2 44 151 180 295 104 52

F3 42 134 150 205 92 50

F4 43 128 178 220 82 53

F5 39 133 173 276 96 51

Jufeng

F1 35 53 115 270 144 75

F2 50 65 130 283 158 76

F3 34 50 106 200 134 74

F4 37 55 112 217 130 70

F5 38 51 117 268 138 72

Analysis of varianceb

P(G) NS <0.01 <0.01 <0.01 <0.01 <0.01

P(N) <0.05 <0.01 <0.01 <0.01 <0.01 NS

P(N × G) NS NS NS NS NS NS

LSD0.05 7 9 22 51 11 6

Full-size table

a See Table 1 for treatments.

b G: genotype; N: nitrogen fertilizer treatment; NS: not significant.

View Within Article

3.3. Relative ureide content in the xylem sap

Relative ureide content in the xylem sap in all three genotypes started at about 20% at V4, reached a maximum of about 80% at R5 and dropped to about 40% at final harvest, but was consistently somewhat higher in You 91-19 and Jufeng than in Wuyin 9 (Table 3). It responded similarly to N top-dressing in all three genotypes. Although N top-dressing at the V2 stage resulted in the maximum relative ureide content in the xylem sap at all sampling stages, there was no significant difference at any stage, compared to the treatment without top-dressing. Top-dressing at either the R1 or R3 stages significantly suppressed relative ureide content in the xylem at the R5 stage: by 9.5 and 11.3% for Wuyin 9, 13.6 and 15.7% for You 91-19 and 13.1 and 14.3% for Jufeng, respectively, compared to the treatment without N top-dressing.

Table 3. Effect of N fertilizer application on relative ureide content (%) in xylem sap at different growth stages for three soybean genotypes grown in Hubei, China

Growth stage treatmenta V4 R1 R3 R5 R6.5 R7

Wuyin 9

F1 17.7 30.9 55.7 77.5 70.0 43.8

F2 18.9 32.1 57.2 78.6 71.5 44.0

F3 17.8 31.1 51.0 70.1 63.4 43.6

F4 17.5 30.7 55.9 68.7 63.2 41.1

F5 17.4 31.2 55.4 77.8 66.2 42.5

You 91-19

F1 18.7 32.7 62.4 80.4 71.8 43.5

F2 20.4 35.2 64.2 81.5 72.8 44.6

F3 18.9 32.9 56.6 69.5 63.6 41.1

F4 18.4 32.2 62.7 67.8 62.2 40.6

F5 18.2 32.5 62.1 80.2 65.6 42.0

Jufeng

F1 23.0 35.2 65.5 81.7 72.3 45.1

F2 24.2 38.1 67.9 82.9 735 46.3

F3 23.2 35.8 59.3 71.0 66.8 44.0

F4 22.8 35.0 65.8 70.0 64.3 44.8

F5 23.1 34.9 65.4 81.1 67.6 44.3

Analysis of varianceb

P(G) <0.01 <0.01 <0.01 <0.05 NS NS

P(N) NS NS <0.01 <0.01 <0.01 NS

P(N × G) NS NS NS NS NS NS

LSD0.05 2.7 4.1 4.6 5.5 3.7 –

Full-size table

a See Table 1 for treatments.

b G: genotype; N: nitrogen fertilizer treatment; NS: not significant.

View Within Article

3.4. Total N accumulation, total N2 fixed and N balance

Total N accumulation and total N2 fixed in all three genotypes was lowest in the treatment with starter N at 25 kg N ha−1 (Table 4). N top-dressing at either the V2 or R1 stage, significantly increased total N accumulation: by 16 and 20% for Wuyin 9, 23 and 27% for You 91-19 and 20 and 27% for Jufeng, respectively. However, top-dressing at either the R3 or R5 stage, did not yield a similar response.

Table 4. The effect of N fertilizer application on total N accumulation, total N2 fixed, proportion of plant N derived from N2 fixation (Pfix) and final N balance in three soybean genotypes grown in Hubei, China

Treatmenta Total N (kg ha−1)

Total N2 fixed (kg ha−1)

Pfix (%)

Seed N (kg ha−1)

N balance

Nfix-seed N (kg ha−1)

Nfix-shoot N (kg ha−1)

Wuyin 9

F1 70.4 50.8 72.2 53.6 −2.8 −16.7

F2 81.6 58.5 71.7 65.9 −7.4 −21.3

F3 84.2 55.7 66.2 69.6 −13.9 −27.6

F4 76.6 51.6 67.4 58.9 −7.3 −23.2

F5 79.9 55.0 68.8 64.7 −9.7 −23.1

You 91-19

F1 122.4 78.8 64.4 95.3 −16.5 −39.1

F2 150.1 93.6 62.3 126.3 −24.7 −52.8

F3 155.3 89.0 57.3 130.8 −41.8 −66.6

F4 139.3 83.5 59.9 105.1 −21.6 −44.4

F5 145.9 89.6 61.4 109.9 −20.3 −45.0

Jufeng

F1 179.9 119.1 66.2 133.3 −14.1 −46.5

F2 215.0 135.6 63.1 150.0 −14.4 −52.7

F3 228.1 130.3 57.1 154.4 −24.1 −60.5

F4 192.3 118.4 61.6 133.2 −14.8 −51.3

F5 194.4 120.9 62.2 138.7 −17.8 −54.4

Analysis of varianceb

P(G) <0.01 <0.05 <0.01 <0.01 <0.01 <0.01

P(N) <0.01 <0.01 <0.05 <0.01 <0.05 <0.05

P(N × G) NS NS NS NS NS NS

LSD0.05 24.0 15.0 6.8 13.5 8.8 12.7

Full-size table

a See Table 1 for treatments.

b G: genotype; N: nitrogen fertilizer treatment; NS: not significant.

View Within Article

Total N2 fixed is the resultant of the proportion of plant N fixed (Pfix) and total N accumulated in the crop. Although the treatment without N top-dressing showed the highest Pfix in all three genotypes, total N2 fixed was lowest for all three. Total N fixed was highest, for all three genotypes, in the treatment with top-dressing at the V2 stage. N2 fixation capacity differed among genotypes: Wuyin 9 fixed 49–59 kg ha−1, You 91-19 84–94 and Jufeng 119–136, depending on the timing of N top-dressing. Surprisingly, even if only the grain was harvested and the residues left in the field as green manure, the crop could not contribute to soil N stocks. If total above-ground material was removed from the field, nitrogen balances would be negative for all three genotypes, particularly for You 91-19 and Jufeng with values from −39 to −66 kg ha−1, depending on the timing of N top-dressing.

3.5. Seed yield and yield components

For all three genotypes, a single 25 kg N ha−1 starter treatment gave the lowest seed yield (Table 5). N top-dressing at either the V2 or R1 stage increased seed yield significantly: by 18 and 21% for Wuyin 9, 24 and 27% for You 91-19, and 23 and 26% for Jufeng, respectively. Top-dressing at either the R3 or R5 stage, did not significantly increase seed yield.

Table 5. Effect of N fertilizer application on seed yield (g m−2) and yield components (unit of all yield components is number per plant, except for height (cm) and 100-seed weight (g)) in three soybean genotypes grown in Hubei, China

Treatmenta Yield Height Branch Node Pod Seed Seed per pod 100-Seed weight

Wuyin 9

F1 100 33 0.8 10.5 19.1 34.7 1.82 15.5

F2 118 36 0.8 10.9 22.2 40.4 1.82 15.1

F3 121 33 0.9 10.5 22.7 41.4 1.82 15.5

F4 105 35 0.7 10.3 19.5 36.6 1.88 15.1

F5 109 36 0.7 10.9 21.7 40.7 1.88 15.1

You 91-19

F1 152 43 5.6 13.5 73.7 113.6 1.54 13.4

F2 188 49 5.9 13.6 89.3 139.7 1.56 14.7

F3 193 47 5.6 13.2 92.3 145.0 1.57 13.8

F4 165 46 5.1 13.6 78.4 124.2 1.58 14.4

F5 168 48 5.5 13.6 82.5 127.8 1.55 14.5

Jufeng

F1 176 82 5.6 19.7 61.3 98.9 1.61 17.9

F2 217 84 6.4 21.4 72.4 121.1 1.67 17.7

F3 222 84 6.6 20.3 74.7 125.4 1.68 17.9

F4 184 80 5.9 20.0 65.5 106.8 1.63 18.3

F5 191 78 5.9 20.1 66.0 111.0 1.68 18.3

Analysis of varianceb

P(G) <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01 <0.01

P(N) <0.01 NS NS NS <0.05 <0.01 NS NS

P(N×G) NS NS NS NS NS NS NS NS

LSD0.05 17 5.0 7.5 1.4 8.9 17.2 0.31 1.1

Full-size table

a See Table 1 for treatments.

b G: genotype; N: nitrogen fertilizer treatment; NS: not significant.

View Within Article

Plant height at the R1 stage, branch and node number per plant, seed number per pod and 1000-seed weight did not differ significantly among the nitrogen treatments in any of the three genotypes. However, pod number and seed number per plant significantly varied in all three genotypes. N top-dressing at either the V2 or R1 stage significantly increased seed number per plant: by 16 and 19% for Wuyin 9, 23 and 28% for You 91-19 and 22 and 27% for Jufeng, respectively, compared to the treatment without top-dressing.

4. Discussion

4.1. Biomass and seed yield

Positive soybean seed yield responses to N fertilizer application have been widely reported (Watanabe et al., 1986, [Kuwahara et al., 1986] , [Tancogne et al., 1991] , [Takahashi et al., 1992] , [Sharma, 1992] , [Purcell and King, 1996] , [Yinbo et al., 1997] , [Botha, 1997] and [Gan et al., 2002c] ), as well as their absence ( [Hardarson et al., 1984] , [Herridge and Brockwell, 1988] , [Ying et al., 1992] , [Wood et al., 1993] , [Koutroubas et al., 1998] and [Seneviratne et al., 2000] ). Reasons for this variation in response may be manifold. Growth conditions, planting date, rainfall, management practices, initial level of soil fertility, level of nodulation, absence of B. japonicum in the soil, and timing of N application may all influence soybean responses ( [Gault et al., 1984] and [Wood et al., 1993] ; [Peoples et al., 1995a] and [Peoples et al., 1995b] ; [Gan et al., 2002a] and [Gan et al., 2002b] ).

In our field experiment, N top-dressing, at either the V2 or R1 stage, significantly increased both total biomass and grain yield (Table 1 and Table 5), while top-dressing at either R3 or R5 did not yield a significantly positive response. Thus, the timing of N application was critical in determining the

yield response, in agreement with previous field and greenhouse experiments with the same genotypes ( [Gan et al., 2002b] and [Gan et al., 2002c] ). Those experiments showed that soybean yield responded significantly to N top-dressing at both the V2 and R1 stages, but not at either the V4 or V6 stages. Watanabe et al. (1986) suggested that the high soybean N demand during the grain filling stage, leading to the risk of ‘self-destruction’ (Sinclair and de Wit, 1976), might best be met by applying supplemental N after flowering. This practice results in increased leaf area duration, and hence increased carbon assimilation by extending leaf longevity. It might also increase nitrogen assimilation by N2 fixation. In an evaluation of the photosynthate and N requirements for seed production by 24 crops, Sinclair and de Wit (1975) showed that soybean was unique in terms of seed composition and potential limitations to yield, because it had the highest N demand, the highest seed protein level, and one of the lowest values of seed production per unit of photosynthate. They concluded that in soybean, the length of the seed filling period is inherently limited, because seed-fill is dependent on active vegetative tissue for both photosynthate and N. The high demand for N of the seeds is partly met by translocation from the vegetative tissue, leading to accelerated senescence. A reduction in symbiotic N assimilation during seed-fill was hypothesized to lead to greater demand for remobilized N, resulting in shortening of the seed-fill period and finally lower seed yield. Our current experimental results show that N top-dressing at both the V2 and R1 stages, significantly stimulated biomass and total N accumulation during the seed filling stage R5, leading to higher seed yield. Top-dressing at either the R3 or R5 stages, had no effect on either biomass accumulation or total N accumulation at the R5 stage, and did not lead to increased seed yield. Symbiotic N2 fixation begins only after nodule formation, which is preceded by colonization of the rhizosphere and infection of legume roots by rhizobia. N2 fixation by soybean generally reaches a peak at early pod fill and then declines during the late reproductive stage ( [Imsande, 1988] and [George and Singleton, 1992] ). This may well explain the positive seed yield response to N top-dressing at both the V2 and R1 stages. Hence, N top-dressing at either R1 or V2 is the most efficient fertilizer strategy for maximizing seed yield. Although all three genotypes responded to the N top-dressing in a similar way, You 91-19 and Jufeng responded better, with 27% yield increase, compared to Wuyin 9 with 20% increase.

4.2. Nodulation, total N uptake, N2 fixation and N balance

Our results show that N top-dressing at different stages differentially affected nodulation. Top-dressing at either the R1 or R3 stages adversely affected nodulation in all three genotypes, whereas top-dressing at the V2 stage had a positive effect, leading to maximum nodulation at all samplings, particularly for You 91-19 and Jufeng. This response was in agreement with an earlier greenhouse experiment (Gan et al., 2002c) and with results reported by DeMooy and Sutherland (1979). Marschner (1986) concluded that during the first few weeks of crop growth, combined nitrogen (soil and fertilizer nitrogen) is essential for optimal plant growth, which in turn stimulates nodule growth. Mineral N is also necessary for nodule formation and development. This is why starter N is usually recommended for soybean in China and other parts of Asia, such as Thailand.

Total N2 fixed by soybean can be expressed as total N accumulated times the proportion of plant N derived from N2 fixation (Pfix) (Peoples et al., 1995b). Therefore, strategies that influence either Pfix or crop N accumulation affect gross input of fixed N. Our results showed that although N top-dressing inhibited nodulation, which in turn resulted in lower Pfix, total N2 fixed was compensated by greater total N accumulation. Although absence of N top-dressing resulted in maximum Pfix, it

strongly reduced total N accumulation, resulting in minimum total N2 fixed, in agreement with previous field experiments (Yinbo et al., 1997; [Gan et al., 2002b] and [Gan et al., 2002c] ). The present study revealed the great variability among genotypes in their ability to fix N2: determinate Wuyin 9 fixed up to 60 kg N ha−1, whereas semi-determinate You 91-19 and indeterminate Jufeng fixed up to 90 and 140 kg ha−1, respectively.

Total N uptake from the soil mineral N-store and the 25 kg starter N varied in a similar way, i. e. from about 20 kg ha−1 for Wuyin 9, 44 for You 91-19–60 for Jufeng (Table 4). It is impossible to distinguish between the two sources, hence no further conclusions can be drawn with respect to the uptake from the mineral N-store. The value reported for mineral N before the start of the experiment (73 ppm) is impossible to interpret, as the samples were stored at ambient temperature for at least a month before analysis. During that period, appreciable mineralization may have taken place (Peoples et al., 1989). Unfortunately that was also the case for the samples taken in the course of the growing season.

The efficacy of the top-dressing can also be derived from the ‘apparent recovery’ of the nitrogen from this dressing (calculated as the additional uptake from the top-dressing above F1), that varied between 0.07 and 0.37 for F2, 0.18 and 0.75 for F3, 0.04 and 0.26 for F4 and 0.11 and 0.25 for F5, for the three varieties. Hence, top-dressing at R1 was by far most effective, followed by V2, in terms of nitrogen uptake. The reason for the much lower recovery in F4 and F5 is not immediately evident. The higher N accumulation is directly translated into seed yield, as the relation between total N accumulation and seed yield is linear up to the highest accumulation level. This suggests that indeed N-availability was the yield-limiting factor.

The actual contribution of N2 fixation to soil fertility depends on the N balance at final harvest (Peoples et al., 1995b). Net input of N into the system only occurs if total N2 fixed during the whole crop growth cycle exceeds the amount of N removed in the seed at harvest. Thus, in terms of nitrogen supply, soybean is not beneficial for the subsequent crop if seed N content exceeds total N2 fixed and the decline in soil N fertility is even stronger, if all above-ground material is removed from the field ( [Peoples et al., 1995b] and [Yinbo et al., 1997] ). In our field experiment, soybean did not make any contribution to soil fertility, as more N was removed in seed than total N2 fixed (Table 4).

Acknowledgements

We would like to express our thanks to NWO-WOTRO, The Netherlands Foundation for the Advancement of Tropical Research, for financial support of the Ph.D. project of Dr. Yinbo Gan, and Hubei Academy of Agricultural Sciences for financial support of the field experiment in China. We are very grateful to Prof. Xuejiang Zhang, Oil Crops Institute, Chinese Academy of Agriculture Sciences, who kindly provided the Rhizobium strain and the supervision on this project.

References

Bergersen, 1980 Bergersen, F.J. (Ed.), 1980. Measurement of nitrogen fixation by direct means. In: Methods for Evaluating Biological Nitrogen Fixation. Wiley, New York, pp. 65–110..

Botha, 1997 A.D.P Botha, Effect of -N, applied during four growth stages at four levels, on N-composition and yield of soya-beans (Glycine max (L.) Merrill). S. Afr. J. Plant Soil, 14 (1997), pp. 31–37. | View Record in Scopus | | Cited By in Scopus (4)

Cataldo et al., 1975 D.A Cataldo, M Haroon, L.E Schrader and V.L Youngs, Rapid colorimetric determination of nitrate in plant tissues by titration of salicylic acid. Commun. Soil Sci. Plant Anal., 6 (1975), pp. 71–80.

DeMooy and Sutherland, 1979 DeMooy, C.J., Sutherland P.L., 1979. Soil-fertility requirement of soybeans with reference to irrigation. In: Hudy, W.H., Jackobs, J.A. (Eds.), Irrigated Soybean Production in Arid and Semi-Arid Regions. TNTSOY Series No. 20. International Agriculture Publications, University of Illinois, Urbana, Champaign, pp. 276–352..

Fehr et al., 1971 W.R Fehr, S Canvins, D.T Burmood and J.S Pennington, Stage of development descriptions for soybean Glycine max (L.) Merrill. Crop Sci., 11 (1971), pp. 929–931.

Gan, 1999 Gan, Y., 1999. Towards improved N2 fixation and yield in soybean. Thesis, University of Groningen, The Netherlands..

Gan et al., 2002a Y Gan, I Stulen, H van Keulen and P.J.C Kuiper, Physiological response of soybean

genotypes to plant density. Field Crops Res., 74 (2002), pp. 231–241. Article | PDF (299 K) | | View Record in Scopus | | Cited By in Scopus (7)

Gan et al., 2002b Y.B Gan, I Stulen, H van Keulen and P.J.C Kuiper, Physiological changes in soybean (Glycine max) Wuyin9 in response to N and P nutrition. Ann. Appl. Biol., 140 (2002), pp. 319–329. | View Record in Scopus | | Cited By in Scopus (5)

Gan et al., 2002c Y.B Gan, I Stulen, F Posthumus, H van Keulen and P.J.C Kuiper, Effects of N management on growth, N2 fixation and yield of soybean. Nutr. Cycl. Agroecosyst., 62 (2002), pp. 163–174. | View Record in Scopus | | Cited By in Scopus (13)

Gault et al., 1984 R.R Gault, D.L Chase, L.W Banks and L Brockwell, Remedial measures to salvage unnodulated soybean crops. J. Aust. Inst. Agric. Sci., 50 (1984), pp. 244–246.

George and Singleton, 1992 T George and P.W Singleton, Nitrogen assimilation traits and dinitrogen fixation in soybean and common bean. Agron. J., 84 (1992), pp. 1020–1028.

Hanway and Weber, 1971 J.J Hanway and C.R Weber, Accumulation of N, P and K by soybean (Glycine max (L.) Merrill). Agron. J., 63 (1971), pp. 406–408.

Hardarson et al., 1984 G Hardarson, F Zapata and S.K.A Danso, Effect of plant genotype and nitrogen fertilizer on symbiotic nitrogen fixation by soybean cultivars. Plant Soil, 82 (1984), pp. 397–405.

Harper, 1974 J.E Harper, Soil and symbiotic nitrogen requirements for optimum soybean production. Crop Sci., 14 (1974), pp. 255–260.

Herridge and Brockwell, 1988 D.F Herridge and J Brockwell, Contribution of fixed nitrogen and soil nitrate to the nitrogen economy of irrigated soybean. Soil Biol. Biochem., 20 (1988), pp. 211–217.

Imsande, 1988 J Imsande, Interrelationship between plant developmental stage, plant growth rate, nitrate utilization and nitrogen fixation in hydroponically grown soybean. J. Exp. Bot., 39 (1988), pp. 775–785.

Koutroubas et al., 1998 S.D Koutroubas, D.K Papakosta and A.A Gagianas, The importance of early dry matter and nitrogen accumulation in soybean yield. Eur. J. Agron., 9 (1998), pp. 1–10.

Kuwahara et al., 1986 M Kuwahara, S Hoshio and T Yoshida, Soil management and nitrogen fertilizer for increasing the yield of soybean in Japan. ASPAC/FFTC Tech. Bull., 98 (1986), pp. 1–9.

Li et al., 1984 Li, Y., Jiang, B., Yuan, K., Wang, Z., Zhang, S., Zhang, C., Zhao, T., Bao, S., 1984. Conventional Methods of Soil and Agrochemistry Analysis. Scientific Publishing House, Beijing..

Marschner, 1986 Marschner, H., 1986. Mineral Nutrition of Higher Plants. Academic Press, London..

Peoples et al., 1989 Peoples, M.B., Faizah, A.W., Rerkasem, B., Herridge, D.F., 1989. Methods for Evaluating Nitrogen Fixation by Nodulated Legumes in the Field. ACIAR, Canberra..

Peoples et al., 1995a M.B Peoples, D.F Herridge and J.K Ladha, Biological nitrogen fixation: an efficient source of nitrogen for sustainable agricultural production?. Plant Soil, 174 (1995), pp. 2–28.

Peoples et al., 1995b M.B Peoples, J.K Ladha and D.F Herridge, Enhancing legume N2 fixation through plant and soil management. Plant Soil, 174 (1995), pp. 83–101.

Purcell and King, 1996 L.C Purcell and C.A King, Drought and nitrogen source effects on nitrogen nutrition, seed growth and yield in soybean. J. Plant Nutr., 19 (1996), pp. 969–993.

Seneviratne et al., 2000 G Seneviratne, L.H.J Van Holm and E.M.H.G.S Ekanayake, Agronomic benefits of rhizobial inoculant use over nitrogen fertilizer application in tropical soybean. Field Crops Res., 68 (2000), pp. 199–203.

Sharma, 1992 R.A Sharma, Efficient water use and sustainable production of rainfed soybean and safflower through conjunctive use of organics and fertilizer. Crop Res. (HISAR), 5 (1992), pp. 181–194.

Sinclair and de Wit, 1975 T.R Sinclair and C.T de Wit, Photosynthate and nitrogen requirements for seed production by various crops. Science, 189 (1975), pp. 565–567.

Sinclair and de Wit, 1976 T.R Sinclair and C.T de Wit, Analysis of the carbon and nitrogen limitations to soybean yield. Agron. J., 68 (1976), pp. 319–324.

Takahashi et al., 1992 Y Takahashi, T Chinushi, T Nakano and T Ohyama, Evaluation of N2 fixation and N absorption activity by relative ureide method in field-grown soybean plants with deep placement of coated urea. Soil Sci. Plant Nutr., 38 (1992), pp. 699–708.

Tancogne et al., 1991 M Tancogne, A Bouniols, S.U Wallace and R Blanchet, Effect of nitrogen fertilization on yield component distribution and assimilate translocation of determinate and indeterminate soybean lines. J. Plant Nutr., 14 (1991), pp. 963–973.

Watanabe et al., 1986 Watanabe, I., Tabuchi, K., Nakano, H., 1986. Response of soybean to supplemental nitrogen after flowering. In: Shanmugasundaram, S., Sulzberger, E.W., Mclean, B.T.

(Eds.), Proceedings of the Symposium on Soybean in Tropical and Subtropical Cropping Systems, Tsukuba, 26 September–1 October 1983, pp. 301–308..

Wood et al., 1993 C.W Wood, H.A Torbert and D.B Weaver, Nitrogen fertilizer effects on soybean growth, yield, and seed composition. J. Prod. Agric., 6 (1993), pp. 354–360.

Yinbo et al., 1997 G Yinbo, M.B Peoples and B Rerkasem, The effect of N fertilizer strategy on N2 fixation, growth and yield of vegetable soybean. Field Crops Res., 51 (1997), pp. 221–229. Article |

PDF (719 K) | | View Record in Scopus | | Cited By in Scopus (26)

Ying et al., 1992 J Ying, D.F Herridge, M.B Peoples and B Rerkasem, Effect of N fertilization on N2 fixation and N balances of soybean grown after lowland rice. Plant Soil, 147 (1992), pp. 235–242.

Young and Conway, 1942 E.G Young and C.F Conway, On the estimation of allantoin by the Rimini-Schryverreaction. J. Biol. Chem., 142 (1942), pp. 839–853.