C and N ¯uxes of decomposing
13
C and
15
N
Brassica napus
L.:
eects of residue composition and N content
I. Trinsoutrot
a, S. Recous
b, B. Mary
b, B. Nicolardot
a,*
aINRA - Unite d'agronomie de ChaÃlons-Reims, Centre de Recherche Agronomique, 2 esplanade Roland Garros, BP 224, 51686 Reims Cedex 2, France
b
INRA - Unite d'agronomie de Laon-PeÂronne, rue Fernand Christ, 02007 Laon Cedex, France
Accepted 2 May 2000
Abstract
The interactions occurring between biochemical composition and N content of crop residues while decomposing in soil, and the associated N dynamics were assessed by studying the kinetics of C and N biotransformations of dierent tissues ofBrassica napus L. (roots, stems and pod walls). These residues were obtained by growing a rapeseed crop under low and high N nutrition, in a labeling growth chamber with enriched 13CO2 atmosphere and a
15
N nutritive solution. The resulting crop residues in which the C-to-N ratio varied between 22 and 135 were homogeneously labeled with 13C and 15N. Paired labeled residues (13C15N labeled residues with unlabeled soil inorganic N;13C14N residues with15N labeled soil inorganic N) were used to determine net and gross ¯uxes of immobilization and mineralization. Decomposition was studied during laboratory incubations at 158C, the initial soil N availability being non-limiting with regard to the rate of C decomposition. The rate of13C mineralization from the residues was in¯uenced by the biochemical composition of the tissues and particularly by their soluble C content. The N content of the tissues did not signi®cantly aect the kinetics or the amount of C mineralized, except in the very short-term. Decomposition was rapid and after 168 days of incubation at 158C, 82% of the C from the stems and pod walls and 69% from the roots at both low and high N contents had disappeared from the soil coarse fraction. Residue decomposition ®rst resulted in net immobilization of soil mineral N for all the residues. The intensity and duration of this immobilization depended on the tissues and the N content of the residues. Compared to the control, the residues with low N content, still induced net N immobilization after 168 days (ÿ22 toÿ14 mg N gÿ1of added C) whereas the high N residues induced little net immobilization or mineralization, atÿ3 to + 4 mg N gÿ1of added C at the same date. The NCSOIL model was used as a tool to calculate, by
®tting simulation against the data, the gross N mineralization and immobilization ¯uxes and also to determine the total N ¯uxes involved over the 168 days of decomposition. Depending on the tissues and their N content, gross cumulative immobilization ranged from 71 to 113 mg N gÿ1 of added C and gross mineralization varied from 66 to 123 mg N gÿ1 of added C. The dierences in net mineralization, observed during decomposition of the tissues with low and high N contents, were well explained by the dierences between gross mineralization ¯uxes which were themselves attributable to the dierent quantities of N mineralized from the residues. The use of modeling to calculate the total gross N ¯uxes demonstrates that the total amount of N involved in the decomposition of crop residues is much higher than the resulting net ¯uxes quanti®ed either by N balance or by15N tracing.72000 Elsevier Science Ltd. All rights reserved.
Keywords: Brassica napusL; Labeled plant residues; Gross ¯uxes; N mineralization; N immobilization
1. Introduction
The return of crop residues to the soil represents an input of ``fresh'' organic matter of which the character-istics are highly variable. This variability results from the biochemical composition of the tissues being
0038-0717/00/$ - see front matter72000 Elsevier Science Ltd. All rights reserved. PII: S 0 0 3 8 - 0 7 1 7 ( 0 0 ) 0 0 0 9 0 - 0
www.elsevier.com/locate/soilbio
* Corresponding author. Tel.: 77-35-83; fax: +33-3-26-77-35-91.
returned to the soil (stems, leaves, roots, pod walls etc.), their physical characteristics, relative contribution to the total amount of dry matter returned and their location at the start of decomposition (i.e. left on the soil surface or incorporated). For a given crop, the nature and amount of residues returning to the soil are in¯uenced by its growing conditions and, particularly, by N nutrition. N fertilization not only modi®es dry matter production, but also the root-to-shoot ratio, and N content of the residues (Rahn et al., 1992).
Oilseed rape (Brassica napus L.) has been developed in Northern Europe for the production of biofuel (MesseÂan, 1996). This crop has the peculiarity of ac-cumulating large and variable quantities of N depend-ing on its conditions of fertilization, so that the N contents of crop residues remaining on the soil at har-vest can vary considerably (Reau et al., 1994; Trinsou-trot et al., 2000a). Field studies have shown that the N from such residues is relatively unavailable during the period between crops (Thomsen and Christensen, 1998; Trinsoutrot et al., 2000a). Understanding the in-teractions between biochemical composition and N content of the residues and their eects on C de-composition and the associated N dynamics in soil is therefore important, and rape residues would seem suitable for the investigation of such relationships. Bio-chemical composition, and particularly the soluble C content, determines the initial rate of residue decompo-sition (Heal et al., 1997) whereas the lignin content controls the medium to long-term fate of added C. The N content of the residues is also an important fac-tor, but its eect on C decomposition and N mineraliz-ation is complex. On the one hand, residue-N is a source of organic N that is subject to mineralization, microbial assimilation, and humi®cation. Quanti®-cation of these processes, even by stable isotopic methods, is only partially feasible due to the direct assimilation of organic N by soil microorganisms (Mary et al., 1993; Barraclough, 1997), to assimilation by microorganisms growing on the residue itself (Che-shire et al., 1999; Gaillard et al., 1999), or to substi-tution eects between residue-derived N and soil N during microbial assimilation or plant uptake (Hart et al., 1986; Jenkinson et al., 1985). On the other hand, residue-N is itself a component of the overall N avail-ability (soil mineral N+ residue-N) that may control the growth of heterotrophic microorganisms, and con-sequently alter the kinetics of C decomposition (Recous et al., 1995). Whether the initial content of residue-N might or might not control residue de-composition, it cannot be analyzed independently of the conditions of incubation, i.e. of the soil N avail-ability. It is therefore essential to distinguish the eect of the nature and availability of C substrates from the eect of N availability in understanding the eect of chemical characteristics of crop residues and their
modeling (Trinsoutrot et al., 2000a). In the work reported here, our main objective was to analyze the interactions between biochemical composition and N content of the residues, and their eects on decompo-sition and soil N dynamics, under conditions where the overall N availability was not limiting the mi-crobial demand. Various methods (N balance, 15N tra-cing and modeling) were used to quantify the N ¯uxes involved, and the contribution of dierent sources of N to these processes. Our study, carried out under controlled conditions, was a part of a larger program intended to characterize the dynamics of oilseed rape residue decomposition in the ®eld (Trinsoutrot et al., 2000a) and the resulting eects on N dynamics during the intercropping period (Justes et al., 1999).
2. Materials and methods
2.1. Soil
The soil used was a highly calcareous rendosol over-lying cryoturbed chalk (Ballif et al., 1995). The surface horizon (0±10 cm) of the soil was sampled and sieved (4 mm) to eliminate plant debris. The main character-istics of the soil were: clay = 9.9%, silt = 11% and sand = 0.8%; CaCO3= 78.3%; organic C = 1.81%;
organic N = 0.17%; pH = 8.3.
2.2. Labeled plant residues
Spring oilseed rape seeds (Brassica napus L., cv. Star) were germinated in the laboratory and allowed to emerge in a glasshouse. At the 2±4 true leaf stage, the rape plants were transferred to an airtight growth chamber where the growing conditions (air moisture, air temperature and photoperiod) were monitored to mimic spring conditions. Day-time light intensity was kept constant (600mmol sÿ1mÿ2) throughout the crop cycle. The plants were grown under hydroponic con-ditions using nutrient solutions as described by Devi-enne et al. (1994). The nutrient solutions were recycled through the nutrient troughs to ensure their constant mixing. The plants were continuously labeled with 13C by using the same 13CO2 bottles (13C isotopic excess
= 3.13%) throughout the entire growth cycle. Two sets of plants were grown: one set received unlabeled KNO3 and the other set KNO3labeled with 15N (15N
isotopic excess = 9.8%), to ensure that the resulting plants were similar but dierently labeled (13C15N or
13
C14N).
A ®rst crop was grown for 195 days to produce plants with a low-N content. The KNO3concentration
exhausted. A second crop was grown for 78 days using a nutrient solution with 3.0 mM KNO3 to produce
plants with a high-N content. The concentration was readjusted, until crop ¯owering, as soon as it dropped by more than 10%. From ¯owering to plant maturity, the N source was removed from the nutrient solution to avoid obtaining mature plants high in NO3ÿ.
The roots, stems and pod walls were separated at harvest, dried at 808C for 48 h, and ground to <1 mm. Leaves present at harvest were not used in this study. The characteristics of the dierent plant ma-terials (determined in duplicate) are shown in Table 1.
2.3. Soil incubations
The treatments studied and the amounts of the dierent plant materials incorporated into the soil are presented in Table 2. Soil samples equivalent to 125 g of dry soil were incubated with or without residues in 500 ml polyethylene pots (4 replicates per treatment) placed in 2 l glass jars (1 pot per jar) in the presence of a CO2trap (30 ml 1 M NaOH). Mineral N and
mi-crobial biomass were monitored on other soil samples equivalent to 25 g dry soil (4 replicates per treatment) incubated with or without residues in 50 ml poly-styrene pots placed in 2 l jars (7 pots per jar) in the presence of a CO2 trap (30 ml 1 M NaOH). To
pre-vent decomposition being limited by mineral N avail-ability (Recous et al., 1995), mineral N (60 g N kgÿ1 dry soil) was applied to all treatments in the form of
K15NO3(for unlabeled N residues) or KNO3(for 15N
labeled residues) (Table 2). The soil samples were incu-bated at 1520.58C. The CO2 traps were periodically
replaced and the soil moisture content maintained at a matrix potential of 0.05 MPa by weighing and read-justing if necessary by the addition of deionized water.
2.4. Analytical measurements
The plant residues studied were characterized as fol-lows: (i) the soluble fraction was determined by hot water extraction (1008C) for 30 min followed by extraction with a neutral detergent (1008C) for 60 min (LineÁres and Djakovitch, 1993); (ii) the hemicellulose, cellulose and lignin fractions in the residual samples were then determined using the method described by Van Soest (1963); (iii) the nitrate content of the resi-dues was determined in other samples by water extrac-tion for 30 min at 208C (material-to-extractant ratio 2/ 1000). The total C, N, 13C and 15N contents of plant residues and fractions were measured using an elemen-tal analyzer (NA 1500, Carlo Erba) coupled to a mass spectrometer (Fisons Isochrom).
The amounts of CO2 trapped in the NaOH were
measured by continuous ¯ow colorimetry (Chaussod et al., 1986) using an auto-analyzer (TRAACS 2000, Bran & Luebbe). Another aliquot of NaOH containing CO2 was precipitated in BaCO3 using BaCl2. The
BaCO3 precipitate was separated by vacuum ®ltration
(glass ®ber ®lter, porosity 1.2 mm) and dried at 808C
Table 1
Total-C, -N and nitrate contents, and C and N distribution among biochemical fractions for the labeled plant materials
Plant material Low-N residues High-N residues
Stems Pod walls Roots Stems Pod walls Roots
Labeling 13C 13C15N 13C 13C15N 13C 13C15N 13C 13C15N 13C 13C15N 13C 13C15N
Total-C(%)a 45.2 46.2 43.7 42.8 43.6 43.1 44.5 45.6 45.6 45.5 44.5 44.3 Total-N(%)a 0.35 0.33 0.38 0.38 1.10 1.25 1.43 1.10 1.90 1.38 1.99 1.73
C-to-N ratio 129 140 115 112 40 35 31 41 24 33 22 26
NO3-N(%)b 0 0 0 0 0 0.05 4.70 1.75 0.33 0 0.58 0.25
Soluble
C(%)c 24.9 23.2 28.1 30.6 27.2 22.8 37.1 35.7 36.6 35.1 31.2 31.1
C-to-N ratio 66 64 51 50 25 17 16 21 11 14 12 15
Hemicellulose
C(%)c 18.0 17.8 20.1 15.5 19.3 22.3 20.0 17.4 23.2 19.3 25.6 20.7
C-to-N ratio 100 171 190 > 200 32 34 31 37 41 58 26 23
Cellulose
C(%)c 42.1 43.4 39.5 42.9 32.3 32.7 34.5 37.9 30.8 37.1 25.7 30.5
C-to-N ratio > 200 > 200 > 200 > 200 > 200 > 200 > 200 > 200 > 200 > 200 > 200 194 Lignin
C(%)c 15.0 15.9 12.3 11.6 19.5 23.1 8.3 8.8 9.4 8.5 17.5 17.6
C-to-N ratio 97 96 78 83 28 28 60 61 50 65 22 21
a
% of dry matter.
b
% of residue total-N.
c
for 15 h. The isotopic excess value was measured by mass spectrometry, after elemental dissociation of the precipitate in the presence of a catalyst (PbO2).
The soil mineral N (NO3ÿ and NH4+) was extracted
with 1 M KCl (30 min shaking at 208C, soil-to-sol-ution ratio 1/4). Soil extracts obtained by centrifu-gation (20 min at 5800 g) were stored at ÿ208C until analysis. The centrifuged soil was recovered and washed three times with 1 M KCl to remove all traces of mineral N (Recous et al., 1988). The organic N and its isotopic excess were measured on the soil. The min-eral N in the soil and plant extracts was analyzed by continuous ¯ow colorimetry (TRAACS 2000, Bran & Luebbe). Nitrate was measured using an adaptation of the method proposed by Kamphake et al. (1967) and ammonium by a method derived from that published by Krom (1980). Isotopic excess of inorganic N was measured on the ammonium sulfate obtained after steam distillation of the soil extracts (Keeney and Nel-son, 1982).
The microbial biomass carbon was determined using the fumigation-extraction method proposed by Vance et al. (1987), using the modi®cations suggested by Chaussod et al. (1988): extraction of the soluble car-bon from samples, fumigated or not with chloroform, by 250 mM K2SO4(soil-to-extractant ratio 1/4, 30 min
shaking at 208C), followed by separation of the extracts by centrifugation (20 min, 5800 g) and storage at ÿ408C. The soluble carbon concentration of the extracts was measured with an auto-analyzer (1010, O.I. Analytical) using an oxidation method at 1008C in a persulfate medium, the CO2 produced being
measured by infrared spectrometry (Barcelona, 1984). The isotopic excess was measured using the procedure described by C. Aita (unpub. Ph.D. thesis, Universite Pierre et Marie Curie, Paris VI, 1996) with13C analysis of the lyophilized soil extract (GT2/S0 apparatus, Amsco/Finn-Aqua). The microbial biomass carbon was calculated by subtracting the soluble C in the fumigated soil from the soluble C of the unfumigated
soil, and dividing by a coecientKEC0:38 (Vance et
al., 1987).
The non-decomposed residues were extracted after dispersion of 125 g of soil with 200 ml 250 mM K2SO4
(30 min agitation at 208C). Sieving under water resulted in separation of a fraction >1 mm made up of small chalk particles, a 50±1000 mm fraction con-taining the residues and a ®ne fraction (<50 mm) obtained after continuous centrifugation of the sieving water. The fraction containing the residues was decar-bonated by drop addition of 2 M HCl (pH stabilized at 3±3.5), rinsed with demineralized water to remove all traces of acid, and then dried at 808C. The carbon and N contents of this fraction and their 13C and15N isotopic excesses were then determined. The determi-nations of the isotopic excess of mineral N, organic N, total N and of N in the residues extracted from the soil, those of total C, BaCO3 precipitate, microbial
biomass C and C in the residues extracted from the soil were all performed with an elemental analyzer (NA 1500, Carlo Erba) linked to a mass spectrometer (Fisons Isochrom).
2.5. Calculation of N ¯uxes
Apparent mineralization was calculated by
subtract-ing at a given time the amounts of inorganic N accu-mulated in the soil with residues from those in the soil without residues. It was expressed in mg N gÿ1 of added C to permit comparison of treatments with dierent amounts added.
Mineralization of the residue-derived N was
calcu-lated from the net accumulation of inorganic 15N in soil incubated with the 13C15N residues.
Net immobilization rates were calculated from the
results obtained with the 13C14N residues incubated in the presence of 15N labeled soil mineral N using the method described by Recous et al. (1995).
Gross immobilization and mineralization ¯uxes
occur-ring duoccur-ring degradation of unlabeled residues (13C14N) in the presence of labeled K15NO3 were calculated
Table 2
Dry matter, C and N inputs from residues and labeled and non-labeled inorganic-N inputs for the dierent treatments
Treatments Control Soil Soil + low-N residues Soil + high-N residues
Residues Stems Pod walls Roots Stems Pod walls Roots
with the NCSOIL model (Molina et al., 1983; Hadas et al., 1989). The model which has been intensively tested under various conditions of incubation and for various residues (Nicolardot et al., 1994a; Hadas et al., 1998; Corbeels et al., 1999) was used here to calculate the gross N biotransformations over time and to ident-ify the processes involved. The data set was not used to test the model. We, therefore, sought the best ®t between the simulated and the measured data from which the model calculated the N ¯uxes involved using the following procedure. The NCSOIL model con-siders ®ve organic pools (Fig. 1): residues, zymogenous microbial biomass (pool 0), autochthonous (native) mi-crobial biomass (pool I), humads (pool II) and stable organic matter (pool III). The parameter values assigned to pools I, II and III (decomposition, recy-cling and humi®cation rates) were those published by Nicolardot et al. (1994a). The initial size and C-to-N ratio of pool II were calculated by optimization and comparing the simulated and experimental data for the
kinetics of the CO2 released, the mineral N and
mi-crobial biomass in the soil without residues. The resi-dues were described using the three fractions (soluble fraction, hemicellulose + cellulose and lignin) obtained from the biochemical characterization described earlier. In order to minimize the number of parameters requiring optimization, the decomposition rate and assimilation yield of lignin were ®xed at 1.0
10ÿ5 dÿ1 and 0.60, respectively. The kinetics of total CO2release,13C±CO2, total mineral N,15N labeled N,
biomass C and 13C biomass obtained for each 13C labeled residue (mean of four replicates) were simu-lated by optimizing the following parameters for each residue: C-to-N ratio of pool 0, degradation rate and assimilation yield of the soluble and (hemicellulose + cellulose) fractions. Simulated and experimental data were compared using the approach proposed by Barak et al. (1990), whereby a low ®gure-of-merit w2
indi-cated a good ®t. The cumulated immobilization and mineralization ¯uxes calculated by the model for the entire incubation (168 days) were then calculated, the immobilization ¯uxes corresponding to all the N ¯uxes entering pools 0 (zymogenous biomass) and I (auto-chthonous biomass) and the mineralization ¯uxes cor-responding to the ¯uxes leaving the residue pool and pools 0 (zymogenous biomass), I (autochthonous bio-mass), II (humads) and III (stable organic matter).
3. Results
3.1. Characteristics of the crop residues: eect of N nutrition
The N nutrition of oilseed rape in¯uenced the bio-chemical composition of its residues (Table 1). The or-ganic N content of plants grown with a high amount of N was 3- to 4-fold higher than in plants grown with the low amount of N. In addition, the cellulose and lignin contents of high-N content residues were lower and the soluble compounds content higher than those of N-low residues. Some dierences between tissues could also be observed. The biochemical characteristics of the stems and pod walls were almost the same, but the N and lignin contents were higher and the cellulose content lower for roots. Residue labeling (13C14N or
13
C15N) had no eect on the characteristics of low-N residues. However, for the high-N residues, the N con-tents of tissues labeled with 13C15N were lower than corresponding plant tissues labeled with 13C14N. The observed dierences could probably be attributed to small dierences in growth conditions (composition of nutrient solutions, position of the plant in the growth chamber).
3.2.13C and15N recovery in soil
The 13C balance at each sampling date (except day 0) was only determined for treatments in which stems were added and was calculated as the sum of the labeled13C present in the non-fractionated soil and the
13
C±CO2 mineralized (Fig. 2). The mean 13C balance
for these treatments and all measuring dates, was 94.9
20.7%. The 15N balances were calculated only at the end of the experiment for all treatments as the sum of
15
N present in soil organic and inorganic N. The mean
15
N balance for all treatments was 95.922.1% at 168 days of incubation (individual data not shown). In consequence and taking into consideration experimen-tal errors, the 13C and 15N losses to unknown ¯uxes were considered to have been negligible during the in-cubation.
3.3. C ¯uxes
3.3.1.13C±CO2released by the soil
The kinetics of carbon mineralization by the stems, pod walls and roots were identical for the 13C15N and
13
C14N residues (results not shown), so that only the kinetics obtained with the 13C14N residues are shown (Fig. 3). Mineralization occurred rapidly after incor-poration of the residues, representing 17±22% of the added13C for the low-N residues and 21±28% for the high-N residues after 21 day. The kinetics of C miner-alization was in¯uenced both by the type of tissue and the N content of the residues. Mineralization rates were ranked in the following order: pod walls > stems
> roots and high-N content > low-N content. The dierences were maximum after 35 days and decreased later. At 168 day, only the dierences between tissues were signi®cant. The cumulative amounts of 13C mineralized were 29, 37 and 39% for the roots, stems and pod walls, respectively.
3.3.2.13C labeled microbial biomass
The maximum incorporation of labeled C into the microbial biomass was attained on 7 day after incor-poration of the residues (Fig. 4) for all tissues and N contents. Greater incorporation of 13C was found in residues with a high N content: 13.8±16.7% of the C added, compared with 11.4±13.5% of the C added for the low-N tissues. Later, the labeled carbon in the mi-crobial biomass decreased rapidly, particularly for those residues highest in N. After 168 day incubation, the amounts of 13C remaining in the microbial bio-mass, in relation to tissues and N content were not statistically dierent, the mean 13C-biomass represent-ing 6.520.8% of the added13C.
Fig. 3. Labeled CO2 released during the decomposition of 13C14N
residues (bars represent standard deviation values,n4). Fig. 2.13C balance for the high-N content13C stems (bars represent
3.3.3.13C present in the 50±1000mm fraction
At time 0, the quantities of labeled C found in the 50±1000 mm soil fraction, representing the remaining residues, were only 57±61% of the added 13C (Fig. 5). They decreased rapidly during incubation with 36± 39% of the 13C added in the low-N residues, and 25± 41% of the13C added in the high-N tissues present in this fraction on day 35. Those residues with the highest N content decomposed faster, except for roots which also decomposed much more slowly than the two other tissues. At day 168, 33% of the 13C from roots remained in the coarse fraction compared with120%
for stems and pod walls.
3.4. N ¯uxes
3.4.1. Soil mineral N and apparent mineralization
The evolution of the mineral N in the soil was simi-lar for the 13C14N and 13C15N residues, so only the results for the former are presented (Fig. 6). The dynamics of mineral N exhibited two distinct phases.
The ®rst phase (2±3 weeks) consisted of net immobiliz-ation of soil mineral N. Its durimmobiliz-ation and intensity depended on the type of residue, and particularly on its initial N content. The second phase was a net min-eralization phase that lasted until the end of the exper-iment. Net N immobilization was greater for the low-N residues than for the high-low-N residues. Maximum im-mobilization (43 mg N kgÿ1 soil) occurred with the low-N stems. For these low-N residues, the amount of mineral N accumulated in the soil remained below that of the control soil until the end of incubation. The high-N pod walls and roots caused a smaller net im-mobilization and led to a higher accumulation of min-eral N than in the control after 84 day. By the end of incubation, apparent mineralization of the residues was ÿ22 toÿ14 mg N gÿ1of added C for low-N resi-dues and ÿ3 to + 4 mg N gÿ1of added C for high N residues.
3.4.2. Net mineralization of residue-N
The treatments with the 13C15N residuesenabled the net release of 15N from the dierent tissues to be esti-mated (Fig. 7). The small quantities of 15N found in mineral form at time 0, corresponding to 0.5±2.4 mg residue-N kgÿ1 soil, were attributed to the soluble N initially present in the residues. 15N release was faster with the high-N residues: 4.2±8.0% of the 15N added in the high-N tissues was mineralized within 28 days, whereas the corresponding values for the low-N tissues were 1.9±3.7%. The fraction of 15N released was also greater at 168 day, except for the stems. The pro-portion of N from the stems accumulated in the soil was the same at 168 day, i.e. 23% of the N added, but this represented very dierent amounts, i.e. about 3 and 12 mg N kgÿ1 soil for low-N and high-N stems, respectively. The 15N released from pod walls and roots, represented 3.2 and 7.0 mg N kgÿ1soil for low-N pod walls and roots respectively, and 12.4 and 13.8 mg N kgÿ1 soil for the corresponding high N tissues. Less mineralization occurred in the roots than in other tissues.
3.4.3. Net immobilization of soil N
The treatments which had received 13C14N residues
and mineral 15Nwere used to calculate soil net N
im-mobilization (Fig. 8). N imim-mobilization was very low in the control soil (0.7±0.9 mg N kgÿ1 soil by day 168). Net immobilization in the treatments with added residues, reached a maximum at 49 day of incubation and represented 24±40 mg N kgÿ1 soil for the low-N tissues and 17±26 mg N kgÿ1soil for the high-N tis-sues. After 49 day, the amounts of N immobilized fell progressively, indicating re-mineralization of the recently immobilized N.
Fig. 4. Evolution of labeled microbial biomass C during the de-composition of 13C14N residues (bars represent standard deviation
3.5. Gross N ¯uxes simulated with the NCSOIL model
The optimized parameter values and associated w2
values for each simulated treatment are shown in Table 3. The w2 values indicate that the simulations
made with the NCSOIL model were not entirely satis-factory. Residue application, in fact had a considerable priming eect on mineralization of the non-labeled soil C and N (results not presented) which was not simu-lated by the model. A large part of this priming eect might be due to the amounts of CO2 and NO3ÿ
pro-duced during the decomposition of residues which in-¯uence carbonate dissolution in chalky soil (Salomons and Mook, 1976; Jacquin et al., 1980). However, the kinetics of the labeled C and N added to the soil were satisfactorily simulated and were then used to calculate the gross immobilization and mineralization ¯uxes (Table 4).
The gross mineralization rates cumulated over the period 0±168 day varied from 66 to 123 mg N gÿ1 of added C, depending on the treatment. Gross mineraliz-ation was greater for the high-N than the low-N stems and pod walls, but was similar for the low-N and high-N roots.
The cumulative gross N immobilization varied from 71 to 113 mg N gÿ1 of added C. The amounts of N immobilized with stems were the same regardless of N content. Gross N immobilization was higher for the high-N roots than the low-N roots. In the case of pod walls, the large dierence between gross immobiliz-ation for the low-N tissues (76 mg N gÿ1 of added C) and the high-N ones (113 mg N gÿ1of added C) might be an artifact resulting from the poor agreement between experimental values and the model values simulated for the high-N pod walls residue w20:51).
Fig. 5. Evolution of organic13C present in the soil fraction (50 to 1000mm) after incorporation of13C14N residues (bars represent standard devi-ation values,n4).
Fig. 6. Evolution of soil mineral N after addition of13C14N residues
The net mineralization simulated by the model was calculated as the dierence between the simulated ¯uxes of gross mineralization and immobilization. The simulations reproduced the observed dierences between residues. The decomposition of low-N resi-dues always resulted in net immobilization at day 168, whereas the high-N residues caused a small positive net eect. The simulated values were, however, always slightly higher than those measured (+5 to +8 mg N gÿ1C).
4. Discussion
4.1. Residue decomposition
The decomposition of plant residues was estimated in two ways: (i) by measuring13C remaining in the soil coarse fraction (50±1000mm) and (ii) by measuring the
13
C±CO2emissions.
The 13C±CO2 produced by the soil corresponds to
mineralization of C from the actual residue and from turnover of the soil organic matter fractions (e.g. mi-crobial biomass, metabolites) into which the labeled carbon from the residue had been incorporated. The
13
C present in the 50±1000 mm fraction quanti®es the added C which is still present in the form of residue. The amount of residue C which has been degraded is then calculated from the dierence. However, at day 0, all the13C added in residue form was not found in the 50±1000 mm soil fraction, mainly because of the pre-sence of water-soluble C in the residues which was then eliminated during the separation procedure. This is con®rmed by measurement of 13C between 0 and 7 days for the low-N residues. In these treatments, the amount of labeled C present in this soil fraction
Fig. 7. N released in the form of mineral N during the decomposition of13C15N residues (bars represent standard deviation values,n4).
Fig. 8. Labeled N immobilized in soil organic N during the de-composition of 13C14N residues using the method proposed by
Recous et al. (1995) (bars represent standard deviation values,
remained constant during this period, whereas a release of 13C±CO2 was measured, indicating
degra-dation of the applied residues.
Measurement of the C ¯uxes through the dierent compartments shows that the decomposition of oilseed rape residues in the soil was rapid. After 168 day, 82% of the C had disappeared from the residue (coarse fraction) for the stems and pod walls and 69% for the roots. At this time, 37% of the added 13C had been mineralized as 13CO2 for the stems and pod walls and
30% for the roots. Aita et al. (1997) and Trinsoutrot et al. (2000a) found similar rates of C decomposition during in situ experiments but over a longer period, 71 and 74% of the C being decomposed 421 and 594 day, respectively, after incorporation of wheat and rape straw at a reference temperature of 158C. The more rapid decomposition of rape residues observed in our laboratory study can be attributed to more favorable conditions of temperature and soil moisture content. This result may also be due to the non-limiting avail-ability of N and ®ner grinding of residues which opti-mized the contact between soil and residues; these
latter two factors often enhance decomposition (Recous et al., 1995, 1998).
The initial rate of C mineralisation was strongly dependent upon the amounts of soluble C initially pre-sent in the residues. The C mineralized at 7 day was correlated with the initial soluble C content of the resi-due at 208C r0:93, P0:01 and the size of the
soluble fraction de®ned by Van Soest r0:95,
P0:01). From 28 day onwards, C mineralization
was largely related to the lignin content r0:91,
P0:01). This result agrees with the observations of
Collins et al. (1990) and Trinsoutrot et al. (2000b). The roots, which are richer in lignin and N than the stems and pod walls, decompose more slowly than the stems and pod walls, this con®rming the work of Saini et al. (1984) with residues of mustard, rice and wheat, and Amato et al. (1984) with Medicago. According to Camire et al. (1991), the lower degradation rate of roots can be attributed to the formation of lignin-N derivatives resulting from the simultaneous presence of large amounts of N and lignin.
The decomposition of rape residues was
ac-Table 3
Optimized values of the parameters with associated chi-squared values w2obtained for the simulation of experimental data with13C residues
13
C labeled residues Low-N residues High-N residues
Parameters Stems Pod walls Roots Stems Pod walls Roots
Residues Soluble fraction
C decay rate constant (dÿ1) 0.64 0.77 0.65 0.92 0.60 0.85
Fraction of decomposed C incorporated into pool 0 0.60 0.60 0.55 0.47 0.36 0.46
(Hemicellulose + cellulose) fraction
C decay rate constant (dÿ1) 0.024 0.016 0.012 0.012 0.014 0.007
Fraction of decomposed C incorporated into pool 0 0.21 0.07 0.27 0.32 0.46 0.51
Zymogenous biomass (pool 0)
C-to-N ratio 3.4 4.6 4.0 4.1 3.1 5.7
w2 0.20 0.16 0.22 0.30 0.51 0.47
Table 4
C and N mineralization, gross and net N ¯uxes measured and simulated by the NCSOIL model accumulated during the 168 d for13C14N
resi-dues incubated with K15NO 3
Variables Units Low-N residues High-N residues
Stems Pod walls Roots Stems Pod walls Roots
Measured13C±CO
2 % of added13C 36 37 29 37 39 30
Simulated13C±CO
2 % of added13C 38 36 27 33 35 23
Measured soil inorganic15N mg N kgÿ1soil 24.4 36.8 38.0 39.8 42.1 46.4
Simulated soil inorganic15N mg N kgÿ1
soil 24.8 33.7 33.5 32.3 32.7 40.1
Simulated gross N immobilization mg N gÿ1of added C 103 76 84 98 113 71 Simulated gross N mineralization mg N gÿ1of added C 86 66 79 102 123 82 Simulated net N mineralizationa mg N gÿ1of added C ÿ17 ÿ10 ÿ5 +4 +10 +11 Measured net N mineralization mg N gÿ1of added C ÿ22 ÿ18 ÿ14 ÿ3 +4 +3
a
companied by C assimilation in the microbial biomass: after 7 days of incubation, about 15% of the C added was found in this compartment, regardless of the resi-due. This value is in good agreement with other studies showing that 15±20% of the added C was present in the microbial biomass shortly after the start of de-composition, e.g. by Ladd et al. (1995) for Medicago
residues, by Nicolardot et al. (1994b) for hollocellu-lose, by Witter and Dahlin (1995) for barley residues or Wu et al. (1993) for ryegrass. A slightly higher C assimilation yield was found for the N-rich residues. This was probably related to their higher soluble C content. In fact, the more easily accessible the added C, the greater seems to be the assimilation of the car-bon by the biomass (Bremer and Kessel, 1992). A rapid decrease of biomass C was observed between 7 and 28 day, particularly for high-N content residues, con®rming the rapid turnover of this pool associated with rapid exhaustion of the easily-degradable com-pounds, as shown by Bremer and Kessel (1992) and Aita (op. cit.). Consequently the residues did not dier to much extent either in the proportion of their C in-corporated into the microbial biomass or in the dynamics of the labeled biomass over time.
The N nutrition of oilseed rape resulted in a signi®-cant dierence in biochemical composition between the high- and low-N that translated into dierent initial rates of decomposition. Later on, dierences were erased as decomposition proceeded, the larger soluble C pool being more rapidly exhausted for high-N resi-dues, and this resulted in a similar cumulative mineral-isation for high- and low-N treatments at day 168. In reality, this lack of dierence in the medium-term between residues with very dierent N contents is mainly due to the fact that decomposition of the resi-dues was not limited by the availability of mineral N, and only determined by the biochemical composition of the tissues themselves. We, therefore, conclude that the N nutrition of a crop in¯uences the biochemical composition of the tissues and its N content, but the N content does not have a signi®cant eect on de-composition in the medium term. This was also observed for the oilseed rape leaves decomposing in the soil (Dejoux et al., 2000). Conversely, Recous et al. (1995) showed that the rate of decomposition of maize residues of similar biochemical composition signi®-cantly varied whether or not the initial amount of N available in soil was above the threshold that limits de-composition (i.e. at 40 mg N gÿ1 of added C in the conditions explored). In that situation, the observed kinetics of decomposition were no longer determined by the residue chemical quality (which was similar whatever the treatment) but depended on the overall availability of N. Furthermore the ratio of immobi-lized N: decomposed C did vary according to the availability of N, suggesting that the C±N relationships
were modi®ed by N availability. It is thus important to understand the eect of the biochemical quality of dierent residues on their decomposition, by dieren-tiating the speci®c role of the N contained in the resi-dues from that of general availability of N for decomposition, which depends on experimental con-ditions or soil type. Consequently, it is essential that the experimental conditions are chosen in such a way that the amount of mineral N in the soil is above the threshold beyond which the amount of N immobilized per unit of C decomposed no longer varies. On the other hand, models which aim at describing the bio-transformations of C and N in soil need to include the simulation of control of the decomposition of residues by the availability of N (Hadas et al., 1998).
4.2. N ¯uxes involved
Despite rather similar decomposition kinetics, the dierent tissues with low- and high-N contents caused very dierent patterns in the course of mineral N in soil. The use of net changes in 15N both with labeled residues and by labeling the soil N pool, only partially revealed the soil N processes involved. Only a small proportion (14±27%) of the N coming from the resi-dues was mineralized after 168 day, which is compar-able to ®gures published for similar crop residues (e.g. Wagger et al., 1985; Jensen, 1997). This demonstrates that most of the residue-N released during decompo-sition was assimilated by the decomposing organisms or remained in more recalcitrant soil organic fractions. Still this net mineralization revealed by 15N is higher than net mineralization calculated from the N balance between amended soil and control soil, showing that soil N was immobilized in the amended treatment in place of the residue-N, i.e. that substitution eects occurred between labeled and unlabeled sources as im-mobilization was enhanced by C addition (Hood et al., 1999). Consequently net mineralization determined by
15
N mineral accumulation in soil both underestimates the actual decomposition of the residue-N and overes-timates the actual availability of that mineralized N.
The maximum immobilization of soil N revealed by labeling the soil mineral N pool, represented 14±21 mg N gÿ1of added C for the low-N residues and 10±16 mg N gÿ1 of added C for the high-N residues, the dier-ence (4±5 mg N gÿ1 residue) being assumed to be the extra contribution of high-N residue to microbial im-mobilization. Jensen (1997) observed maximum immo-bilization of 12 and 18 mg N gÿ1of added C for peas
N%2:51and barley N%0:98), respectively.
assimilation of the residue-N by the micro¯ora to be estimated. This method failed here to give estimates of the ¯uxes because the amounts of N in the low-N resi-dues were too small to provide the necessary sensitivity for this calculation. Moreover, the calculation of gross ¯uxes by isotopic dilution and enrichment requires samples to be taken at very short time of intervals, or use of successive ``pulses'' of labeled mineral N. Finally, the integration of instantaneous ¯uxes, measured with this approach, during prolonged de-composition remains rather dubious (Recous et al., 1999).
This is why the approach ®nally chosen consisted of estimating the gross N ¯uxes with a model that simu-lated C±N biotransformations. The advantage of such an approach is to take into account all the ¯uxes com-ing from the residue N, soil mineral N, and N recycled by the microbial biomass, that is usually not con-sidered, and to cumulate them over time. The simu-lations showed that the amounts of N involved in the decomposition process were very large. The require-ments of the micro¯ora for decomposition of the stems, pod walls and roots of oilseed rape were on average over the 168 day decomposition period 91 (216) mg N gÿ1 of added C or 287 (241) mg N gÿ1of C mineralized. These values are higher than those cal-culated previously from short-term estimates of gross N ¯uxes with wheat straw decomposing either in con-trolled incubation or in ®eld conditions (Mary et al., 1998; Recous et al., 1999), which were at 174 and 126 mg N gÿ1 of C mineralized, respectively. The calcu-lation for net mineralization obtained with the NCSOIL model was in good agreement with the measured net N mineralization. This led us to assume that the model had correctly simulated the gross N ¯uxes and therefore gave a reliable picture of the con-tribution of the various processes to the net dynamics of soil N. By way of example, the dierence in net mineralization between stems of low- and high-N con-tent therefore appears to be simply due to a dierence in the intensity of the gross mineralization ¯ux which itself is explained by a dierence in the N mineraliz-ation of the residues (5.2 and 25.8 mg N gÿ1of added C for low- and high-N stems, respectively). These con-clusions contradict those of Barraclough et al. (1998) who suggested that the dierences in net mineraliz-ation ¯uxes between residues were mainly due to dierences between gross immobilization ¯uxes. This apparent contradiction may arise from the fact that the N availability was probably limiting of the mi-crobial needs in the work of Barraclough et al. (1998), therefore, modifying the relationship between N im-mobilized and C decomposed.
By considering the contribution of each tissue on the dynamics of soil mineral N, it is possible to calcu-late the eect that the incorporation of all the residues
obtained from two rape crops of low- or high-N con-tent might produce on soil N dynamics. The pro-portions of stems, pod walls and roots at harvest in the related ®eld study (Trinsoutrot et al., 2000a) were 49, 31 and 20%, respectively for a rape crop without N fertilization and 61, 31 and 8% for the crop which had previously received 270 kg N haÿ1. Using the reported values obtained for the various tissues and the NCSOIL model, the simulated net mineralization values for the whole residue would be ÿ15.7 mg N gÿ1 of added C for the low-N crop residue and +4.1 mg N gÿ1 of added C for the high-N crop residue after 111 days at 158C. These ®gures can be compared with those of Justes et al. (1999) who estimated the net min-eralization rate for whole rape residues (with N con-tents of 0.4 and 0.9%) to be ÿ1.1 and + 1.7 mg N gÿ1 of added C, respectively, in a ®eld experiment of the same time length (i.e. equivalent to 111 day at 158C). The dierences are particularly large for the low-N residue, with the laboratory incubation over-estimating the net N immobilization compared to the ®eld situation. This was also observed for rape leaves decomposing at the soil surface in the ®eld or incu-bated under controlled conditions by Dejoux et al. (2000). Despite the dierences in composition of the residues (they were not exactly the same and the ®eld experiment included the leaves which had fallen during crop growth), the dierences observed between labora-tory and ®eld experiments may arise from the N avail-ability in soil, which is likely to have limited N immobilization in the ®eld (Justes et al., 1999; Trinsou-trot et al., 2000a). The dierences in size of the residue fragments and their localization in the soil, which determine the closeness of contact between soil and residues, may have also exerted a substantial eect on decomposition and N cycling as shown by Jensen (1994) and Ambus and Jensen (1997). This demon-strates the importance of understanding and modeling the role of initial location and physical characteristics of plant residues in soil on the subsequent dynamics of C and N.
4.3. Conclusion
the N content of the residue itself. Conversely the in-itial residue-N content had a major eect both on the intensity of net immobilization and on the duration of the immobilization phase. In order to identify par-ameters to describe the relationship between residue quality and C decomposition, and C±N relationships during OM cycling, it appears, therefore, essential to avoid any confounding eect of the control of C dynamics and changes in C±N relationships resulting from the N availability of the soil + residue system. It also requires being able to model the eect of N short-age on the C and N biotransformations, because in most situations the N availability will control the de-composition, due to low soil mineral N content or the residue location.
The use of 15N tracing to label either the residue or the soil mineral N pool gave incomplete and biased pictures of the fate of residue-N, owing to substitution eects and assimilation of residue-N by decomposing microorganisms. But combined with the use of a C±N biotransformation model such as NCSOIL, it allowed access to the total N ¯uxes involved over time, and understanding of the processes by which residue de-composition in¯uence the net dynamics of N in soil.
Acknowledgements
We thank G. Alavoine, M. Boucher, O. Delfosse, E. GreÂhan, M.J. Herre, S. Millon for technical assistance. This work was ®nanced by Europol'Agro (Reims, France) and CETIOM (Centre Technique Interprofes-sionnel des OleÂagineux MeÂtropolitains, Paris, France).
References
Aita, C., Recous, S., Angers, D.A., 1997. Short-term kinetics of re-sidual wheat straw C and N under ®eld conditions: characteriz-ation by 13C15N tracing and soil particle size fractionation. European Journal of Soil Science 48, 283±294.
Ambus, P., Jensen, E.S., 1997. N mineralization and denitri®cation as in¯uenced by crop residue particle size. Plant and Soil 197, 261±270.
Amato, M., Jackson, R.B., Butler, J.H.A., Ladd, J.N., 1984. Decomposition of plant material in Australian soils. II-residual organic14C and15N from legume plant parts decomposing under
®eld and laboratory conditions. Australian Journal of Soil Research 22, 331±341.
Ballif, J.L., GueÂrin, H., Muller, J.C., 1995. EleÂments d'agronomie Champenoise Ð Connaissance des sols et de leur fonctionnement Ð Rendzines sur craie et sols associeÂs Ð Esquisse geÂomorpholo-gique. INRA Editions, Versailles.
Barak, P., Molina, J.A.E., Hadas, A., Clapp, C.E., 1990. Optimization of an ecological model with the Marquardt algor-ithm. Ecological Modelling 51, 251±263.
Barcelona, M.J., 1984. TOC determinations in ground water. Ground Water 22, 18±24.
Barraclough, D., 1997. The direct or MIT route for N
immobiliz-ation: a 15N mirror image study with leucine and glycine. Soil
Biology & Biochemistry 29, 101±108.
Barraclough, D., Gibbs, P., MacDonald, A., 1998. A new soil N and carbon cycle. In: Proceedings of the 16th World Congress of Soil Science, 20±26 August 1998, Montpellier, France, ISSS-AISS-IBG-SICS-AFES (CD-rom), 20±26.
Bremer, E., Kessel, C.V., 1992. Seasonal microbial biomass dynamics after addition of lentil and wheat residues. Soil Science Society of America Journal 56, 1141±1146.
CamireÂ, C., CoÃteÂ, B., Brulotte, S., 1991. Decomposition of roots of black alder and hybrid poplar in short-rotation plantings: N and lignin control. Plant and Soil 138, 123±132.
Chaussod, R., Houot, S., Guiraud, G., HeÂtier, J.M., 1988. Size and turnover of the microbial biomass in agricultural soils: laboratory and ®eld measurements. In: Jenkinson, D.S., Smith, K.A. (Eds.), Nitrogen in Agricultural Soils. Elsevier, London, pp. 312±326. Chaussod, R., Nicolardot, B., Catroux, G., 1986. Mesure en routine
de la biomasse microbienne des sols par la meÂthode de fumi-gation au chloroforme. Science du Sol 2, 201±211.
Cheshire, M.V., Bedrock, C.N., Williams, B.L., Chapman, S.J., Solntseva, I., Thomsen, I., 1999. The immobilization of nitrogen by straw decomposing in soil. European Journal of Soil Science 50, 329±341.
Collins, H.P., Elliott, L.F., Papendick, R.I., 1990. Wheat straw de-composition and changes in decomposability during ®eld ex-posure. Soil Science Society of American Journal 54, 1013±1016. Corbeels, M., Hofman, G., Van Cleemput, O., 1999. Simulation of
net N immobilisation and mineralisation in substrate-amended soils by the NCSOIL computer model. Biology and Fertility of Soils 28, 422±430.
Dejoux, J.F., Recous, S., Meynard, J.M., Trinsoutrot, I., Leterme, P., 2000. The fate of nitrogen from winter-frozen rapeseed leaves: mineralization, ¯uxes to the environment and uptake by rapeseed crop in spring. Plant and Soil 218, 257±272.
Devienne, F., Mary, B., Lamaze, T., 1994. Nitrate transport in intact wheat roots. II. Long-term eects of NO3
ÿ
concentration in the nutrient solution on NO3ÿunidirectional ¯uxes and distribution
within the tissues. Journal of Experimental Botany 45, 677±684. Gaillard, V., Chenu, C., Recous, S., Richard, G., 1999. Carbon,
nitrogen and microbial gradients induced by residue decompo-sition in soil. European Journal of Soil Science 50, 567±578. Hadas, A., Molina, J.A.E, Feigenbaum, S., Clapp, C.E., 1989.
Simulation of N-15 immobilization by the model NCSOIL. Soil Science Society of America Journal 51, 102±106.
Hadas, A., Parkin, T.B., Stahl, P.D., 1998. Reduced CO2 release
from decomposing wheat straw under N-limiting conditions: simulation of carbon turnover. European Journal of Soil Science 49, 487±494.
Hart, P.B.S., Rayner, J.H., Jenkinson, D.S., 1986. In¯uence of pool substitution on the interpretation of fertilizer experiments with
15N. Journal of Soil Science 37, 389±403.
Heal, O.W., Anderson, J.M., Swift, M.J., 1997. Plant litter quality and decomposition: an historical overview. In: Cadish, G., Giller, K.E. (Eds.), Driven by Nature, Plant Litter Quality and Decomposition. CAB International, Wallingford, pp. 47±66. Hood, R.C., N'Goran, K., Aigner, M., Hardason, G., 1999. A
com-parison of direct and indirect15N isotope techniques for estimat-ing crop N uptake from organic residues. Plant and Soil 208, 259±270.
Jacquin, F., Haidouti, C., Muller, J.C., 1980. Dynamique de la matieÁre organique en sols carbonateÂs cultiveÂs. Science du sol 1, 27±36.
Jenkinson, D.S., Fox, R.H., Rayner, J.H., 1985. Interactions between fertilizer nitrogen and soil nitrogen Ð the so-called `priming' eect. Journal of Soil Science 36, 425±444.
amended with low C:N ratio plant residues with dierent particle sizes. Soil Biology & Biochemistry 26, 519±521.
Jensen, E.S., 1997. N immobilization and mineralization during in-itial decomposition of 15N-labeled pea and barley residues. Biology and Fertility of Soils 24, 39±44.
Justes, E., Mary, B., Nicolardot, B., 1999. Comparison of the eec-tiveness of a drilled radish cover crop, oilseed rape volunteers and oilseed rape residues incorporation for reducing nitrate leach-ing durleach-ing the period between crops. Nutrient Cyclleach-ing in Agroecosystems 55, 207±220.
Kamphake, L.J., Hannah, S.A., Cohen, J.M., 1967. Automated analysis for nitrate by hydrazine reduction. Water Research 1, 205±216.
Keeney, D.R., Nelson, D.W., 1982. N inorganic forms. In: Page, A.L., Miller, R.H., Keeney, D.R. (Eds.), Methods of Soil Analysis. Soil Science Society of America, Madison, pp. 643±698. Krom, M.D., 1980. Spectrophotometric determination of ammonia:
a study of a modi®ed Berthelot reaction using salicylate and dichloroisocyanurate. The Analyst 105, 305±316.
Ladd, J.N., Amato, M., Grace, P.R., 1995. Simulation of14C turn-over through the microbial biomass in soils incubated with14 C-labelled plant residues. Soil Biology & Biochemistry 27, 777±783. LineÁres, M., Djackovitch, J.L., 1993. CaracteÂrisation de la stabiliteÂ
biologique des apports organiques par l'analyse biochimique. In: Decroux, J., Ignazi, J.C. (Eds.), MatieÁres organiques et agricul-tures Ð QuatrieÁme journeÂes de l'analyse de terre et cinquieÁme forum de la fertilisation raisonneÂe. GEMAS-COMIFER, Paris, pp. 59±168.
Mary, B., Fresneau, C., Morel, J.L., Mariotti, A., 1993. C and N cycling during decomposition of roots mucilage, root and glucose in soil. Soil Biology & Biochemistry 25, 1005±1014.
Mary, B., Recous, S., Robin, D., 1998. A model for calculating N ¯uxes in soil using15N tracing. Soil Biology & Biochemistry 30,
1963±1968.
MesseÂan, A., 1996. OleÂagineux et environnement : vers une extension du concept d'eÂcobilan. OleÂagineux Corps gras Lipides 3, 284±285. Molina, J.A.E., Clapp, C.E., Shaer, M.J., Chichester, F.W., Larson, W.E., 1983. NCSOIL, a model of N and carbon trans-formations in soil: description, calibration, and behavior. Soil Science Society of America Journal 43, 85±91.
Nicolardot, B., Molina, J.A.E., Allard, M.R., 1994a. C and N ¯uxes between pools of soil organic matter: model calibration with long-term incubation data. Soil Biology & Biochemistry 26, 235± 243.
Nicolardot, B., Fauvet, G., CheÁneby, D., 1994b. Carbon and N cycling through soil microbial biomass at various temperatures. Soil Biology & Biochemistry 26, 253±261.
Rahn, C.R., Vaidyanathan, L.V., Paterson, C.D., 1992. N residues fromBrassicacrops. Aspects of Applied Biology 30, 263±270. Reau, R., Wagner, D., Palleau, J.P., 1994. End of winter diagnosis:
winter rapeseed (Brassica napus) and N fertilization. In:
Proceedings of the 3rd Congress of European Society of Agronomy. Padova, Italy, 18±22 September, 220±221.
Recous, S., Machet, J.M., Mary, B., 1988. The fate of labeled15N urea and ammonium nitrate applied to winter wheat crop. Part II: Plant uptake and N eciency. Plant and Soil 112, 215±224. Recous, S., Robin, D., Darwis, D., Mary, B., 1995. Soil inorganic N
availability: eect on maize residue decomposition. Soil Biology & Biochemistry 27, 1529±1538.
Recous, S., Richard, G., Fruit, L., Chenu, C., Angers, D., 1998. Factors aecting the contact soil and incorporated crop residues: short-term eects on C evolution. In: Proceedings of the 16th World Congress of Soil Science, 20±26 August, Montpellier, France, ISSS-AISS-IBG-SICS-AFES (CD-rom).
Recous, S., Aita, C., Mary, B., 1999. In situ changes in gross N transformations in bare soil after addition of straw. Soil Biology & Biochemistry 31, 119±133.
Saini, R.C., Gupta, S.R., Rajvanshi, R., 1984. Chemical composition and decomposition of crop residues in soil. Pedobiologia 27, 323± 329.
Salomons, W., Mook, W.G., 1976. Isotope geochemistry of carbon-ate dissolution and reprecipitation in soils. Soil Science 122, 15±24. Thomsen, I.K., Christensen, B.T., 1998. Cropping system and resi-due management eects on nitrate leaching and crop yields. Agriculture, Ecosystems and Environment 68, 73±84.
Trinsoutrot, I., Nicolardot, B., Justes, E.S., 2000a. Decomposition in the ®eld of residues of oilseed rape grown at two levels of N ferti-lisation: eects on the dynamics of soil mineral N between succes-sive crops. Nutrient Cycling in Agroecosystems 56, 125±137. Trinsoutrot, I., Recous, S., Bentz, B., LineÁres, M., CheÁneby, D.,
Nicolardot, B., 2000b. Relationships between biochemical quality of crop residues and C and N mineralization in temperate agro-systems. Soil Science Society of American Journal 64, 918±926. van Soest, P.J., 1963. Use of detergent in the analysis of ®brous
feeds. I. Preparation of ®ber residues of low N content. Journal of the Association of Ocial Agricultural Chemists 46, 825±835. Vance, E.D., Brookes, P.C., Jenkinson, D.S., 1987. An extraction
method for measuring soil microbial biomass C. Soil Biology & Biochemistry 19, 703±707.
Wagger, M.G., Kissel, D.E., Smith, S.J., 1985. Mineralization of N from N-15 labeled crop residues under ®eld conditions. Soil Science Society of America Journal 49, 1220±1226.
Watkins, N., Barraclough, D., 1996. Gross rate of N mineralization associated with the decomposition of plant residues. Soil Biology & Biochemistry 28, 169±175.
Witter, E., Dahlin, S., 1995. Microbial utilization of 14C-labelled straw and 13C-labelled glucose in soils of contrasting pH and
metal status. Soil Biology & Biochemistry 27, 1507±1516. Wu, J., Brookes, P.C., Jenkinson, D.S., 1993. Formation and
