Fate of
15
N labelled nitrate and ammonium in a fertilized forest
soil
GoÈran Bengtsson*, Christer Bergwall
Department of Ecology, University of Lund, SoÈlvegatan 37, 223 62 Lund, Sweden Accepted 27 September 1999
Abstract
The possibility that long-term atmospheric nitrogen pollution and fertilization of forest soil may serve as a basis for adaptation for enhanced transformation rates of NO3ÿ and NH4+ in soil bacteria was elucidated in a laboratory bioassay.
Bacteria extracted from soils that had been fertilized at various rates for the last 30 yr were characterised with respect to their capability to reduce or oxidise dierent nitrogen sources. The same soils were used under oxic or anoxic conditions to quantify denitri®cation, dissimilatory nitrate reduction to ammonium (DNRA) and nitri®cation. 15NO3ÿ and 15NH4+ were added as
tracers to the soils, which were incubated in bottles for 3 to 5 d. Concentrations of15N2O in headspace and 15
NO3 ÿ
and15NH4 +
in soil extracts were determined by gas chromatography±mass spectrometry. Total numbers of bacteria were similar in all soils and ranged from 3 to 4108cells gÿ1d.wt. of soil. Between 50 and 70% of the isolated strains were capable of reducing nitrate and the majority of them reduced nitrate to ammonium. About 0.01- of all isolates were classi®ed as nitri®ers. Both nitrate reducers and nitri®ers were more common in fertilized soils than in the unfertilized control soil. The foremost fate of added
15NO 3
ÿand15NH 4
+in all soils was immobilisation. More than 85% was immobilised in anoxic soils and between 64 and 97% in
the oxic soils, with the lowest quantities in fertilized soils. As regards the remaining, non-immobilised N, DNRA dominated over denitri®cation, as could be expected from the higher frequency of ammonifying bacteria compared with denitri®ers. There was no obvious relationship between NH4+produced and the amount of fertilizer applied, whereas denitri®cation was negatively
correlated with the amount of fertilizer applied. Nitrifying activity was low in all soils with no obvious relationship between NO3ÿ produced and fertilizer applied. Hence, no correlation was found between the relative abundance of N transforming
bacteria and the transformation activity. The N ¯ux followed essentially the same pattern as that seen for product formation. The DNRA ¯ux was higher than that of both denitri®cation and nitri®cation. DNRA and denitri®cation ¯uxes were highest in the control soil, whereas the nitri®cation ¯ux was low in all soils. The absence of evidence for adaptation to enhanced rates of transformation of NO3ÿand NH4+in soil bacteria exposed to long-term N fertilization is re¯ected by the low concentration of
extractable inorganic N in the fertilized soils. As a consequence of the quantitative importance of immobilization of added N, dierences in physiological capacity evolved in soil bacteria to immobilize and mobilise N may determine the rates by which inorganic N is available for plant growth or lost to groundwater and air.72000 Elsevier Science Ltd. All rights reserved.
Keywords:Nitrogen; Mobilization; Immobilization; Denitri®cation; Nitri®cation; Soil bacteria; Gas chromatography; Mass spectrometry; Spruce
1. Introduction
Even if primary production in temperate forest soils is often N limited (Tamm, 1991), increased provision by atmospheric N pollution and by fertilization may
have adverse eects on forest vegetation and promote soil acidi®cation and leaching of nitrate and base cat-ions to surface water and groundwater (Aber et al., 1989; Murdoch and Stoddard, 1992). The extent to which vegetation and water quality is aected depends on the retention and transformation of N in the soil. Larger amounts of deposited and added N tend to be immobilized in the soil, notably by microbial
assimila-0038-0717/00/$ - see front matter72000 Elsevier Science Ltd. All rights reserved. PII: S 0 0 3 8 - 0 7 1 7 ( 9 9 ) 0 0 1 8 3 - 2
www.elsevier.com/locate/soilbio
* Corresponding author. Fax: +46-46-2223-790.
tion, than taken up by plants (NoÈmmik, 1966; Nadel-hoer et al., 1995; Buchmann et al., 1996) and remain unavailable to plants for years (Cheng and Kurtz, 1963; Hart et al., 1993). Mobilization and transform-ations may ensure that N is present in the soil as
NO3ÿ, NO2ÿ and NH4+ or removed by leaching and
denitri®cation to N2and N2O. Whereas the availability
of inorganic N tends to control transformation rates in forest soils, other factors, such as the concentration of
organic carbon (Corg) and O2and pH, may be
respon-sible for the discrimination between speci®c transform-ations. Hence, it has been proposed that the ratio of available C to electron acceptor concentration deter-mines whether NO3ÿis reduced in a dissimilatory
path-way to NH4+ or denitri®ed (Tiedje et al., 1982). In
environments with high C and low NO3ÿ
concen-trations, such as most forest soils, NH4+is predicted to
be the main product of NO3ÿ reduction, supplying
autotrophic nitri®ers with an energy source and plants and heterotrophs with a favoured nutrient. Obser-vations useful for testing such predictions are scarce and quanti®cation of the processes regulating the in-ternal cycling of N in forests still have a good deal of uncertainty (Davidson et al., 1992; Groman et al., 1993; Hart et al., 1993; Buchmann et al., 1996).
Although the fate of N imported to forest ecosys-tems by atmospheric N deposition and fertilization is of major concern in environmental management, the inability to foresee chemical and biological reactions in the soil precludes eorts to accurately estimate long-term eects on forest ecosystems. Data accumulated from long-term fertilizer experiments in forest ecosys-tems may be helpful in the assessment of the fate of long-term atmospheric N deposition. Priha and Smo-lander (1995) used plots in acid spruce forests that had been fertilized with 900 kg N haÿ1for 30 yr to indicate long-term eects of N deposition on forest soils. NH4
+
was the main inorganic N product formed during
incu-bation of soil samples: 1.5±2 times as much NH4+was
produced in the fertilized plots than in the controls,
both with the same and marginal NO3ÿconcentrations.
However, the immobilisation of N by microorganisms was reduced in the fertilized soils.
One possible explanation for enhanced transform-ation of mineral N in soils that have been amended or polluted with N for many years is selection for enzymes with higher transformation activities. Such a phenomenon has been demonstrated in bacteria in estuarine sediments (King and Nedwell, 1987) and in shallow groundwater aquifers (Bengtsson and Berg-wall, 1995) and is likely to occur in fertilized forest soils with high concentrations of organic carbon, in which increased capacity of the electron transport sys-tem should promote survival and ®tness of the cells.
The objectives of this work were to determine the eect of fertilization on transformation rates and fate
of inorganic N in a forest soil as a basis for an evalu-ation of potential adaptevalu-ations to higher denitrifying, ammonifying (DNRA) and nitrifying activities among soil bacteria. This is of long-term consequences of at-mospheric N deposition on N turnover in forest soils. Three hypotheses were tested:
1. The frequency of denitrifying, DNRA and nitrifying strains was expected to be higher in the fertilized forest soils compared with a control soil receiving no nitrogen fertilizer. This would be the conse-quence of the release of inorganic nitrogen as the major abiotic constraint for growth and, hence, den-sity, of those groups in forest soils.
2. Denitrifying, DNRA and nitrifying activities were assumed to be higher in the fertilized forest soils compared to the control in response to the increased availability and concentration of inorganic nitrogen. The long-term nitrogen amendment of the fertilized soils may constitute a basis for selection of nitrogen transformers with altered enzyme activities. Since organic carbon is present in surplus, a selection of ecient nitrogen transforming enzymes may result in increased nitrogen transformation activity. 3. The C/N ratio determines the fate of the added
nitrogen as per the discussion above. DNRA ac-tivity may dominate over denitri®cation as the major electron transport system, since the forest soils studied were characterized as electron donor rich and electron acceptor poor (e.g. high C/N ratio).
2. Materials and methods
2.1. Site description and soil sampling
The soil samples were taken from a Norway spruce (Picea abies Karst.) forest at Lake StraÊsan, central Sweden. The soil is a glacial till dominated by medium
and ®ne sand. The mean annual temperature is 3.18C
and the mean annual precipitation is 745 mm. The trees were planted in 1958 and the site was used for optimum nutrition experiments since 1967, when the average tree height was 1.3 m (Tamm et al., 1974). A
total of 56 plots (3030 m) were treated in dierent
ways: four rates of N, three rates of P, two rates of K+Mg+micronutrients treatments and combinations of these. Each treatment was replicated in two plots, except for the control, which was replicated ®ve times.
NH4NO3 has been added yearly from 1967 at rates
varying between 30 and 180 kg haÿ1 yrÿ1. The total additions of fertilizer N haÿ1 until 1994 were 0, 930, 1620 and 2610 kg for the N0, N1, N2 and N3 treat-ments, respectively. The N2 treatments were suspended
in 1990, the N3 treatments in 1992, whereas the N1 treatments are still in progress. Total N concentrations
of needles had increased from01.3 to01.7% in the
N2 and N3 treatments in 1982 and the annual growth
rate in the N1±N3 plots was 12±13 m3haÿ1
yrÿ1
com-pared with 5 m3 haÿ1 yrÿ1 in the N0 plots (Tamm,
1985).
Soil samples were taken in October 1995 from four of the plots: 13 (N0), 14 (N1), 52 (N2) and 33 (N3). A sterile planting trowel was used to collect soil from 30
30 cm squares of the O horizon, randomly located
within the plot, into autoclaved 500 ml brown glass jars. Six jars were ®lled with soil from each plot and shipped to the laboratory overnight in a refrigerated box. The jars were opened in a sterile hood and visible roots removed aseptically. Approximately 50% of the volume collected consisted of roots. The soil was stored for approximately 2 weeks at 48C prior to iso-lation of bacteria, physiological characterization and denitri®cation and nitri®cation bioassays.
2.2. Isolation and characterization of bacteria
Soil samples (2 g wet weight) were suspended in 20 ml of sterile 0.2% Calgon (sodiumhexametaphosphate) and 0.1% peptone solution and shaken for 30 min at 160 rpm by use of a wrist action shaker (Grin and George). Aliquots of 0.1 ml of dilutions of the samples
(10, 100 and 1000 using sterile Calgon and
pep-tone solution) were inoculated on the surface of three dierent agar media, one separate aliquot for each, by means of L-shaped glass rods. For heterotrophic
bac-teria (potential denitri®ers) peptone±yeast±glucose
(PYG) agar was used. The PYG agar contained 1.25 g peptone, 1.25 g yeast extract, 1.25 g glucose, 0.15 g MgSO47H2O, 0.015 g CaCl2 and 15 g agar lÿ1
ster-ile distilled water. The plates were incubated at 178C. After 7 d of growth on PYG agar, colonies were picked from plates, streaked several times on duplicate PYG agar plates for puri®cation and incubated at
178C. The purity of strains was examined by colony
check with a stereo microscope (10 and 30). Pure
cultures were stored in cryo tubes supplied with PYG
and 20% sterile glycerol at ÿ808C. The isolated
het-erotrophic bacteria were tested for nitrate reducing, denitrifying and DNRA capability by means of micro-titer plate tests described by KoÈlbel-Boelke et al. (1988) and Daniels et al. (1994). Physiological tests included nitrate reduction (NO3ÿ4NO2ÿ),
denitri®ca-tion (NO3ÿ4N2) and DNRA (NO3ÿ4NH4+,
ammo-nia by Nessler reaction). For chemolitotrophic bacteria (nitri®ers), media for ammonia oxidizers and nitrite oxidizers were used. Ammonia oxidizers were
enumer-ated on 1.0 g K2HPO4, 0.2 g MgSO47H2O, 50 mg
FeSO47H2O, 20 mg CaCl2,2H2O 2 mg MnCl2
4H2O, 1 mg Na2MoO4 2H2O, 5 g CaCO3, 1.5 g
NH4Cl and 15 g agar (Sigma puri®ed) lÿ1 of distilled
water. The nitrite oxidizers were incubated on 1.0 g
K2HPO4, 0.2 g MgSO47H2O, 5 mg FeSO47H2O,
2 mg CaCl2, 2H2O, 2 mg MnCl2 4H2O, 1 mg
Na2MoO4 2H2O, 2.0 g NaNO2 and 15 g agar
(Sigma puri®ed) lÿ1of distilled water. The plates were
incubated at 278C for 3±4 weeks. The colonies were
counted with the aid of a stereo microscope (10 and
30) and then transferred to PYG plates to test for heterotrophic contaminants.
2.3. Total numbers of bacteria
Soil samples of 2.0 g (wet weight) from each site were suspended in 50 ml of sterile 0.2% Calgon (sodiumhexametaphosphate) and 0.1% peptone sol-ution and homogenized for 2 min in an Omni mixer homogenizer (Omni Int.). Acridine orange solution (50
ml of a 0.1% solution) was added after dilution (10).
The mixtures were left for 2 min at room temperature and then slowly ®ltered through a black polycarbonate
0.22 mm membrane (Poretics Corp.). The number of
¯uorescent cells was counted in a Zeiss Axioskop 20 equipped for ¯uorescence observations (Carl Zeiss, Germany).
2.4. Bioassays
Soil samples (5 g for the denitri®cation assay and 3 g for the nitri®cation assay) from the four plots were transferred to 25 ml glass ¯asks. The soils were amended with 5 mg of Na15NO3ÿ-N gÿ1 soil (99 at% 15
N) and 1 mg of 15NH4+Cl-N gÿ1 soil (98 at% 15N),
respectively, both obtained from Cambridge Isotope Laboratories, Andover, MA. Flasks for the denitri®ca-tion assay were ¯ushed with He for 5 min to remove O2and C2H2added to the gas phase to a ®nal
concen-tration of 10%. The ¯asks were sealed with rubber stoppers and incubated for 3 d at 158C. Three replicate ¯asks per soil sample were used. Headspace gas (0.5 ml) was withdrawn with a gastight syringe once every 24 h and analysed for N2O. After the last injection on
the GC/MS, the soil was extracted with 0.5 M K2SO4
for analysis of NO3ÿand NH4+.
Flasks for the nitri®cation assay were sealed with
cotton plugs and incubated for 5 d at 158C. Three
replicates were also used. After incubation the soil was directly extracted with 0.5 M K2SO4.
2.5. Analytical procedures
2.5.1. Sample preparations
NO3ÿ and NH4+ were extracted from the soil with
0.5 M K2SO4 at 5:1 extractant to soil ratios. The soil
soil residues in the 25 ml bioassay ¯asks were washed
with 210 ml 0.5 M K2SO4 and transferred to the
test tubes. The extraction was performed on a rotary shaker for 6 h at 178C. Following extraction, the test tubes were centrifuged at 2000 rpm for 30 min and 7 ml aliquots of solution transferred to two 12 ml test tubes. The solution was evaporated in a Savant Speed Vac (Savant Instruments), redissolved in 1.5 ml of deionized water and used for the derivatization of NO3ÿ and NH4+. In addition, soil samples from each
plot were directly extracted to obtain the initial con-centrations of NO3ÿand NH4+.
2.5.2. Derivatization of NO3ÿand NH4+
We used convenient and sensitive methods to
deter-mine concentrations of 15NO3ÿ and 15NH4+ by GC±
MS. Nitrate was determined by modi®cation of a method developed by Tanaka et al. (1985). The sample
was derivatized using 2-sec-butylphenol (purity >
98%, Aldrich) and N,O-bis(trimethylsilyl)acetamide (BSA, purity > 97%, Fluka) to yield the trimethylsilyl ester of 4-NO2 2-sec-butylphenol. The aqueous sample
(1 ml) was transferred to an 8-ml test tube and cooled
in ice. Then 0.2 ml of Ag2SO4 solution (5% in
deio-nized water) and 1.4 ml of concentrated H2SO4 were
added followed by 100 ml 2-SBP (5% in 99.5%
etha-nol). The test tubes were now transferred to a heating block for 15 min at 308C with shaking every ®fth min-ute. The aqueous phase was extracted with 1 ml of toluene by the use of a wrist action shaker for 5 min. The water was removed after centrifugation (3500 rpm) and the organic phase washed twice with 2.0 ml of deionized water and dried with anhydrous Na2SO4.
30 ml of BSA was added to the test tubes which were
allowed to stand in room temperature for 10 min. Finally, the toluene was evaporated under a stream of N2and the residue redissolved in 200ml ethyl acetate.
Ammonium was derivatized using penta¯uoroben-zoylchloride (PFBOCl, purity >99%, Fluka) to yield penta¯uorobenzamide. The procedure was based on modi®cations of a method originally used by Fujihara et al. (1986) to measure NH4+ production in putrescine
oxidation by human plasma. The aqueous sample (1 ml) was transferred to a test tube containing 1 ml
NaHCO3 (5% in deionized water), 15 ml PFBOCl and
spiked with 2mg of methyl amine as internal standard. The test tube was shaken for 3 min using a wrist action shaker. The aqueous phase was extracted with 1 ml ethyl acetate for 5 min. The water was removed after centrifugation (3500 rpm) and anhydrous Na2SO4
was added to the organic phase. The ethyl acetate was evaporated at 308C under a stream of N2and the
resi-due redissolved in 0.5 ml diethyl ether:hexane (15:85% v/v). The silicic acid columns (clean up step, 100 mg Bond Elute, Analytichem Int.) were conditioned with 3 ml of hexane. The sample was then added to the
col-umn which was subsequently washed with 3 ml diethyl
ether:hexane (15:85% v/v). The PFBONH2 derivative
was eluted with 2 ml diethyl ether:hexane (50:50% v/
v). The eluate was evaporated to dryness under N2
and the residue redissolved in 1 ml ethyl acetate.
2.5.3. GC/MS
The mass spectrometer was a Hewlett-Packard 5889A MS-Engine coupled to Hewlett-Packard 5890
Series II gas chromatograph. For N2O, a 25-m plot
fused silica column (Poraplot Q, ®lm thickness 10 mm,
i.d. 0.32 mm, Chrompack, The Netherlands) was used with He as the carrier gas at a ¯ow rate of 2.9 ml minÿ1. The injector and interface temperatures of the
GC were 175 and 2708C, respectively. The oven
tem-perature was 408C. For derivatized NO3ÿ, a DB-17
15-m fused silica capillary colu15-mn (®l15-m thickness 0.25
mm, i.d. 0.25 mm, J&W Scienti®c, Folsom, CA) was
used with He as the carrier gas at a ¯ow rate of 1.5 ml
minÿ1 and the split was opened 40 s after injection.
The injector and interface temperatures of the GC
were 130 and 2708C, respectively. The oven
tempera-ture was initially held at 1008C for 1 min and then
increased by 208C minÿ1 to 2308C. For derivatized
NH4+, a DB-Wax 30-m fused silica capillary column
(®lm thickness 0.25 mm, i.d. 0.25 mm, J&W Scienti®c, Folsom, CA) with He as the carrier gas at a ¯ow rate
of 1.2 ml minÿ1 and the split was opened 40 s after
injection. The injector and interface temperatures of
the GC were 250 and 2608C, respectively. The oven
temperature was initially held at 1008C for 1 min and
then increased by 208C minÿ1 to 2008C.
Electron-impact (EI) mass spectra were obtained using 70 eV at
an ion source temperature of 2008C and at a
quadro-pole temperature of 1008C. The mass spectrometer was
operated in the selected ion monitoring mode (SIM).
Measured ions were of m/z 44 [14N14NO], m/z 45
[14N15NO],m/z46 [15N15NO],m/z238,1 [TMS ester of 4-14NO2 2-SBP], m/z 239.1 [TMS ester of 4-15NO2
2-SBP], m/z 211.0 [PFBO14NH2], m/z 212.0
[PFBO15NH2] and m/z225.0 [PFBONHCH3]. The
iso-tope-enrichment calculations for NO3ÿ and NH4+ were
made by using the m/z ratio 239.1/238.1 and 212.0/
211.0 of calibration solutions, with a natural isotopic
distribution, subtracting them from the measured m/z
239.1/238.1 and 212.0/211.0 ratios of samples. The
iso-tope-enrichment calculation for N2O is described in
Section 2.5.4.
2.5.4. Determination of the N2O ¯ux
The production of N2O by denitri®cation was
calcu-lated from the abundance of the ions m/z 44, m/z 45
and m/z 46, which correspond to dierent
combi-nations of the isotopes: 14N14N, 14N15N and 15N15N,
respectively. We assumed that 14N and 15N were in
isotopic equilibrium in the samples so that their
tions could be described by a binomial distribution:
44N
2O 1ÿ15a2, 45N2O 215a 1ÿ15a, 46
N2O 15a2, where 44[N2O], etc. are the fractions
of N2O molecules with molecular weights 44, etc. and
15
a is the fraction of 15N in the sample. The value for
15
a was determined as (46[N2O]+1245[N2O])/
(44[N2O]+45[N2O]+46[N2O]). Since the molecular
frac-tions as m/z 45 and 46 may be contaminated by the
heavy oxygen isotopes 17O and 18O and traces of
oxy-gen and nitrooxy-gen may react in the ion source to
pro-duce NO2, it was necessary to correct the abundances
of ions measured. We considered 44[N2O] to be
cor-rectly measured since more than 99.2% of its
abun-dance consists of only 14N14N16O and used its
abundance to calculate the abundance of N2O
mol-ecules without heavy oxygen isotopes: N2O
44
N2O= 1ÿ15a2: The correct abundances of the 45
N2O and 46N2O molecules were then calculated by
multiplying their binomial isotope distributions with
the abundance of N2O without heavy oxygen isotopes.
A corrected value for15a was then calculated from the correct abundances of the three N2O molecules.
A standard curve for unlabeled N2O was obtained
by injecting between 200 and 1000 ml lÿ1 in triplicate in the mass spectrometer. The dose±response curve was nonlinear but accurately described by a power function. This equation was used to calculate the con-centration of N2O in the samples, which all had total
and isotopic abundances in the range covered by the
standards. The total concentration of N2O in the
closed ¯ask (y) was calculated from an expression of
Henry's law:y=axsoil water volume/gas volume,
wherea is the solubility of N2O in cm2cmÿ2of water
and x is mg N2O in the gas phase (Moraghan and
Buresh, 1977). The correct abundances at m/z 44, 45
and 46 of the gas standard were also used to calculate
the natural abundance of 15a. It was found to be
0.003633, regardless of the concentration of N2O. This
value was subtracted from the calculated values for15a
in the samples to obtain the atom excess of 15N in
N2O.
Finally, the ¯ux of N2O derived from denitri®cation
of added15NO3ÿ was calculated:15N-N2ON2O/M
T15N-NO3ÿ, where 15N-N2O is atom excess of 15N,
N2O is the total corrected concentration of N2O
(44N2O+45N2O+46N2O) in the ¯ask, M is the weight
of soil in grams, T is incubation time in hours and
15
N-NO3ÿ is the atom excess of 15N-NO3ÿ in the soil.
Denitri®cation coupled to nitri®cation was calculated in a similar way but with reference to atom excess of
15
N-NH4+in the soil.
2.5.5. Other calculations and statistics
The relative concentration of the added 15NO3ÿ that
was immobilized was calculated as the mean percen-tage of the tracer that was not recovered as 15N2O or
extractable 15NH4+ and 15NO3ÿ. Likewise, the relative
concentration of immobilized added 15NH4+was
calcu-lated as the mean percentage of an added tracer that
was not recovered as extractable 15NH4
+
and 15NO3ÿ.
The contribution of added 15NO3ÿ to denitri®cation
was calculated as the ratio between the observed release of 15N2O-N and the product of observed N2
O-N and the initial added 15NO3ÿ. The contribution of
added 15NO3ÿ to the production of NH4+ was
calcu-lated as the production of 15NH4+-N divided by the
production of NH4+-N. Nitri®cation from the added
15
NH4+ source was calculated as the ratio of produced
15
NO3ÿ and produced NO3ÿ. We assumed that isotope
discrimination, if present, was undetectable by the quadrupole mass spectrometer due to the high enrich-ment. Analysis of variance was used to determine the
Table 1
Fate of15N incubated in soil samples with dierent history of fertilization. Fertilizer was added yearly to ®eld plots from 1967. Mean and standard deviation (S.D.) are given for measurements made at the end of laboratory incubations of soil sampled in 1995
Fertilizer applied (kg haÿ1)
Treatment suspended in year
15
N-NO3, initial (mg
kgÿ1)
15
N-NO3after incubation
(mg kgÿ1)
1 15
N-NH4, initial (mg
kgÿ1)
15
N-NH4after incubation (mg
kgÿ1)
15
N-N2O after incubation (mg
kgÿ1)
mean S.D. mean S.D. mean S.D. mean S.D. mean S.D.
Nitrate reduction
0 ± 4.953 3.15Eÿ06 0.187 0.0101 0.014 0.00025 0.205 0.0033 0.0254 0.00046
930 ongoing 4.953 1.94Eÿ06 0.138 0.0006 0.169 0.0026 0.176 0.0015 0.0150 0.00021
1620 1990 4.952 1.57Eÿ05 0.421 0.0041 0.060 0.0033 0.240 0.0018 0.0125 0.00013
2610 1992 4.952 1.06Eÿ05 0.630 0.0124 0.178 0.00012 0.205 0.0029 0.0097 3.66Eÿ05
15N-NH
4, initial (mg
kgÿ1)
15N-NH
4, after incubation
(mg kgÿ1)
115N-NO3, initial (mg
kgÿ1)
15N-NO
3, after incubation (mg
kgÿ1)
mean S.D. mean S.D. mean S.D. mean S.D.
Nitri®cation
0 0.000242 3.36Eÿ06 0.000511 6.28Eÿ06 0.994 0.384 0.034 0.0094
930 ongoing 0.000263 2.06Eÿ06 0.000577 6.90Eÿ06 1.149 0.099 0.328 0.0061 1620 1990 0.000322 1.67Eÿ05 0.000778 3.20Eÿ05 1.040 0.107 0.299 0.0021 2610 1992 0.000315 1.13Eÿ05 0.000655 1.55Eÿ05 1.158 0.683 0.417 0.0097
G.
Bengtss
on,
C.
Bergwall
/
Soil
Biology
&
Biochemis
try
32
(2000)
545±557
eects of treatments on N transfer and to examine dierences between means of dierent turnover
pro-ducts, e.g. between NH4+ and N2O (post-hoc tests
using Schee's F). Analysis of variance and
calcu-lations of linear correcalcu-lations were made in StatView 4.5 (Abacus Concepts).
3. Results
The dominant fate of15NO3ÿand15NH4+in the soils
was immobilization, regardless of the amounts of ferti-lizer added during the past 30 yr. Less than 13% of added 15NO3ÿ was extractable after 3 d of incubation
(Table 1). More than 85% of it was immobilized, whereas less than 5% was reduced in a dissimilatory pathway (Fig. 1). The immobilization was related to the intensity of fertilization, in as much as 15NO3ÿ
added to soil from the two plots receiving the largest doses of fertilizer was less immobilized than 15NO3ÿ
added to the control plot soil. Between 64 and 97% of
added 15NH4+ was immobilized and less than 0.5
-was nitri®ed (Fig. 2 and Table 1). Again, 15NH4+
added to soil from the two most fertilized plots was
less immobilized than 15NH4+ added to the control
soil.
Most of the dissimilatory reduced NO3ÿended up as
NH4+ (Fig. 1 and Table 1, P< 0.0001, Schee's F),
but there was no obvious pattern for the relationship
between NH4+ produced and the amount of fertilizer
applied or the concentration of extractable NO3ÿin the
soil. However, the relationship between the reduction of15NO3ÿto15NH4+and the concentration of
extracta-ble NH4
+
was slightly negative r ÿ0:105). Of the
15
NO3ÿ added to the soil, 0.5% or less was recovered
as 15N2O. Denitri®cation was negatively correlated
with the amount of fertilizer added r ÿ0:690, P<
0.0001) and the concentration of extractable NO3ÿ
r ÿ0:640, P< 0.0001). Only traces of the nitri®ca-tion detected, R0.22%, were attributed to oxidation of
extractable NH4+(Table 2). Less than 1% of the
deni-tri®cation could be traced to extractable NO3ÿand less
than 4% of NH4+ produced during nitrate reduction
resulted from extractable NO3ÿ. Most of the tracer
de-rived nitrate reduction and nitri®cation was found in the control soil (Table 2). The observations indicate that the main N source for nitrate reduction and nitri-®cation was organic or non-extractable inorganic.
Years of fertilization had a marginal in¯uence on
the concentration of extractable NO3ÿ in the soil
(Table 3). The NO3ÿ concentration had increased by
40% in the plot receiving the largest amount of fertili-zer compared with the control, which is order of mag-nitude less than the corresponding dierence in the dose of fertilizer applied. The concentration of extrac-table NH4+in the soil had increased by 3- to 10-fold in
the fertilized plots compared with the control (Table 3). The highest concentrations were found in the plots still fertilized (14) or most recently excluded from fertiliza-tion (33), but even the plot in which fertilizafertiliza-tion was discontinued 5 yr ago (52) had about 4-fold more extractable NH4+ concentrations than the control plot,
suggesting that the eects of fertilization on extracta-ble NH4+concentrations have lasted for at least 5 yr.
Fig. 3. Fluxes of denitri®cation and nitri®cation (left axis) and of NH4+(right axis) in soils from plots treated with dierent doses of fertilizers. The ¯uxes were calculated using isotope ratios as described in the methods.
The N ¯ux, calculated as the concentration of a pro-duct formed per atom % excess precursor added, fol-lowed essentially the same pattern of variation among the plots as described for the product formation
(Fig. 3). The ¯ux of NH4+ was three orders of
magni-tude faster than the ¯ux of denitri®cation and nitri®ca-tion and negatively correlated with the concentranitri®ca-tion of extractable NH4+ r ÿ0:429,P0:0004). The
denitri®cation ¯ux was faster in the control soil than in the fertilized soils (P< 0.0001, Schee'sF), whereas the opposite was true for the nitri®cation ¯ux P0:0015, Schee'sF).
The total number of bacteria extracted from the soils were 3.0, 3.3, 4.0 and 3.7108 gÿ1 in the soils treated with 0, 930, 1620 and 2610 kg haÿ1 fertilizer, respectively. More than 50% of the colonies appearing after incubation of soil extracts on PYG agar were
capable of reducing NO3ÿ in one or the other way
(Fig. 4). A majority of them reduced NO3ÿ to NH4+
and they, as well as the denitri®ers, were more numer-ous in the fertilized soils than in the control soil. The density of nitri®ers, calculated as the number of colo-nies observed on the minimal medium agar after incu-bation of aliquots of soil extract for 3±4 weeks, was higher in the fertilized plots than in the control (Fig. 5)
and NH4+ oxidizers were more common than NO2ÿ
oxidisers. The frequency of nitri®ers observed within the total count of bacteria on plates was 0.01-, with at most a factor of two in dierence between the con-trol and fertilized plot (52).
In summary, we found no direct evidence for
enhanced rates of NO3ÿ reduction in soils fertilized
with up to 2 600 kg haÿ1 of NH4NO3 for tens of
years. It is possible that the fertilized soils have selected for a greater number of strains with high rates of ammoni®cation and denitri®cation, whose charac-teristics are hidden by the low concentrations of extractable NO3ÿ. Even if that is true, their
contri-bution to the turnover of nitrogen in the StraÊsan soils seems to be subordinate to that of strains immobilising inorganic nitrogen and mobilising nitrogen bound by microorganisms, clays and organic material. Strains controlling immobilisation and mobilisation may be adapted to higher rates by fertilization, and the fate of nitrogen added to the StraÊsan soil determined by their activity.
4. Discussion
The weak relationship revealed by the experiments between long-term forest soil fertilization and mi-crobial transformation rates of inorganic nitrogen may be attributed to several causes, of which the low con-centration of extractable NO3ÿ following fertilization is
suggested to be the most important. The ®nding of <1 mg NO3ÿ-N kgÿ1in the experimental soils regardless of
the amount of fertilizer applied (Table 3) re¯ects the common view about soils in coniferous forests as being low in NO3ÿ (Robertson, 1982; Vitousek et al., 1982),
suggesting a combination of insigni®cant net nitri®ca-tion and signi®cant leaching and immobilizanitri®ca-tion.
The dominance of immobilization of added15N over
inorganic transformations suggests that fertilization of the StraÊsan soils has not supported any build up of
suciently high concentrations of NH4+ and NO3ÿ to
Table 2
Percentage of nitrate reduction and nitri®cation in experimental soils attributed to the15N labelled pool of extractable NO3ÿand NH4+ , respect-ively. Calculations are described in Section 2
Fertilizer applied (kg haÿ1)
Added15NO3ÿas a source of denitri®cation (%)
Added15NO3ÿas a source of produced NH4+(%)
Added15NH4+as a source of nitri®cation (%)
0 0.96 3.52 0.22
930 0.81 0.95 0.10
1620 0.76 2.68 0.11
2610 0.70 0.89 0.15
Table 3
Concentrations of extractable NO3ÿand NH4+in soil samples collected from the four plots examined at StraÊsan
Fertilizer applied (kg haÿ1) Plot Extractable NO3ÿ-N (mg kgÿ1) Extractable NH4+-N (mg kgÿ1)
mean S.D. mean S.D.
0 13 0.48 0.13 3.47 1.56
930 14 0.51 0.19 29.48 8.94
1620 52 0.70 0.78 13.30 8.21
2610 33 0.66 0.59 31.12 0.43
select for enhanced rates of their transformations. The N pool of the soil (litter layer, mor and mineral soil) at the N2 site is estimated to 3.5 t haÿ1(SjoÈgren and Persson, Swedish University of Agricultural Sciences, personal communication), of which extractable
inor-ganic N represents about 0.01-. The N pool of the
N2 site is 1.4 t haÿ1 larger than that of the control site, which corresponds reasonably well with the total
dose of N fertilizer added to N2, 1.6 t haÿ1. The
increase of organic N is con®ned to the litter and FH layer: the C-to-N ratio drops from 45 in the control site litter to 23 and in the N2 site and from 32.6 to 23.6, respectively, in the FH layer. The organic N con-centration of the mineral soil does not seem to have been in¯uenced by fertilization, the C-to-N ratio at 10±30 cm depth is higher at the N2 site (27.8) than at the control site (25.4). Immobilization in the soil or-ganic matter pool has also been recognized as the dominant fate of NH4+and NO3ÿadded to other forest
soils (Davidson et al., 1992; Groman et al., 1993; Hart et al., 1993; McKenney et al., 1995). Obser-vations by Davidson et al. (1992), Hart et al. (1993) and Nadelhoer et al. (1995) suggest that most of the added and retained inorganic N in a soil is assimilated by bacteria and fungi during the growth season, result-ing in small amounts of added N beresult-ing nitri®ed and denitri®ed, at least in well-drained soils (Groman et al., 1993). Soils can ®x NH4+-N by the interlayer of 2:1
clay minerals (NoÈmmik, 1957; Moore, 1965), especially when the ®xation sites are not blocked by soil organic matter (Hinnman, 1966). The mean residence time of an N atom may be a few days in the inorganic pools, 1 or 2 months in the microbial biomass pool (David-son et al., 1992) and probably much longer in the clay mineral and organic matter pools.
The observation that fertilized StraÊsan soils immobi-lize less of the added inorganic N than the unfertiimmobi-lized control soil is consistent with results obtained for mi-crobial N by Smolander et al. (1994) and Priha and Smolander (1995) with soils from fertilized and unlized Norway spruce forests. These forests were ferti-lized for 30 yr and the decreased immobilization in microbial biomass coincided with an increased net for-mation of mineral N. However, the relationship between fertilization and net formation of mineral N is not consistently positive but seems to be in¯uenced by such variables as the source of fertilizer and soil. Priha and Smolander (1995) made their observations in soils treated with ammonium sulphate, urea and ammonium
nitrate in succession. Martikainen et al. (1989)
observed no signi®cant eects on net formation of N in pine forest soils treated with urea or ammonium nitrate, whereas our own observations of the spruce forest soils at StraÊsan indicated a net 8-fold increase of
(NH4+-N + NO3ÿ-N) after 25 yr of addition of
am-monium nitrate to the N3 plots.
Although fertilization had increased the concen-trations of NH4
+
relative to the control (Table 3), they were probably too low to have anything but a mar-ginal in¯uence on the nitri®cation rates (Fig. 3). The
largest mean concentration of mineral N (NH4+
-Fig. 4. The frequency of NO3ÿreducers among colonies of bacteria derived from soil extracts from plots with dierent fertilizer doses and growing on PYG agar for seven days at 178C. The mean fre-quencies of denitri®ers as well as ammoni®ers were correlated with the amount of fertilizer applied r0:954andr0:985, respectively).
N+NO3ÿ-N) we found in the StraÊsan soils was 032
mg kgÿ1 (Table 3), which is below the threshold for
mineral N (60±90 mg kgÿ1) in forest ¯oor material at which nitri®cation ceases (Vitousek et al., 1982). Blew and Parkinson (1993) referred the low nitri®cation rate (10 ng N gÿ1
hÿ1
) that they measured after 4 weeks of incubation of samples from the F±H horizon of a 120-yr-old spruce forest to the low concentration of min-eral N (45 mg kgÿ1). Yet, that rate was two orders of magnitude higher than the rates calculated from 72 h of incubation of StraÊsan soils with 15N. The dierent rates are partly due to dierent methods of calculation.
When calculated rates are based on extracted NO3ÿ,
regardless of the source (15N) or concentration of
extractable NH4+, the nitri®cation rates at StraÊsan are
only 2±5 times lower than at the Canadian site. This latter dierence may be attributed to dierences in in-cubation time, dierences in pH and the contribution of mobilized NH4+to nitri®cation.
ThepHH2Oof the StraÊsan soil was low (between 3.60
and 3.87 in the O-horizon of the plots) compared with the Canadian soil (5.9 on average in the FH-horizon). Both acid-sensitive and acid-tolerant strains occur among nitri®ers (de Boer et al., 1990), but Persson and WireÂn (1995) were unable to detect net nitri®cation at pH values lower than 4.0 in Scandinavian spruce forest soils. Similarly, very low net nitri®cation was found by Johnsrud (1978) in spruce forests at pH below 4.3. Although the relative importance of autotrophic and heterotrophic nitri®cation in forest soils has been addressed in few studies, it seems as if mobilized NH4+
can be a signi®cant source of NO3ÿin some soils.
Schi-mel et al. (1984) estimated, for instance, that less than
5% of the nitri®cation rates, 45±90 ng N gÿ1 hÿ1,
measured in slurries of mixed coniferous forest soils incubated for 6 h was due to autotrophic oxidation of
NH4
+
. The measured nitri®cation in the StraÊsan soils was net rates, which may underestimate the gross rates if some of the added 15NH4+ was oxidized to 15NO3ÿ
and subsequently immobilized (Stark and Hart, 1997). Nitri®ers may produce N2O, either from various
in-termediates during NH4+ oxidation (Ritchie and
Nicholas, 1972) or from the reduction of NO2ÿcoupled
to the oxidation of NH4+ (Poth and Focht, 1985). The
latter is con®ned to soils with limited aeration, e.g. due to slow O2diusion in interstitial pore water, whereas
the former may be substantial in pH neutral, aerobic agricultural soils and be responsible for anything less
than 10% to more than 50% of the N2O ¯ux,
depend-ing on for example, soil moisture conditions (Bremner and Blackmer, 1979; Freney et al., 1979; Skiba et al., 1993; Webster and Hopkins, 1996; Stevens et al.,
1997). The nitri®er production of N2O in forest soil
seems to be more variable and a soil with pH 3.8
pro-duced virtually no nitri®er derived N2O (Robertson
and Tiedje, 1987). A preliminary 6 h bioassay with
StraÊsan soil samples initially ¯ushed with air and then sealed in the incubation ¯asks with rubber stopper
showed no detectable N2O peak and it was decided to
assay nitri®cation with continuous aeration without stoppers, with the loss of data on traces of N2O from
the nitri®cation.
The concentrations of NO3ÿ in the StraÊsan soils are
also considerably lower than those found in aquifers (>20 mg lÿ1), in which denitri®ers have evolved for enhanced rates of nitrate reduction (Bengtsson and Bergwall, 1995). However, even concentrations of that magnitude in the forest soil may have been insucient
to select for enhanced NO3ÿ reduction because of the
lower accessibility of NO3ÿions to cells on soil particles
compared with pore water. Murray et al. (1989) found,
for example, no correlation between bulk soil NO3ÿ
concentration and denitri®cation rates in agricultural
soils that were fertilized and where NO3ÿ-N
concen-trations occasionally were as high as 350 mg kgÿ1. The relative contribution of denitrifying and ammo-nifying bacteria to NO3ÿ reduction in a soil or a
sedi-ment is dicult to predict and the generally
acknowledged qualitative characteristics that tend to distinguish the two processes are diluted by exceptions. Some observations suggest that dissimilatory reduction of NO3ÿ to NH4+ (ammoni®cation) is quantitatively as
important in soil and sediment as denitri®cation
(Sùr-enson, 1978), whereas others ®nd that NO3ÿ
ammoni®-cation is insigni®cant (Binnerup et al., 1992). Based on
the high capacity of electron acceptors per NO3ÿ
reduced to NH4+ (8eÿper N atom), Tiedje et al. (1982)
suggested that DNRA bacteria should be competitively successful in electron donor rich and electron acceptor poor habitats, where selection should favour cells with a maximum electron acceptor capacity. Organic C con-centrations are known for the N0 and N2 treatments (43.8 and 48.4% in the FH layer, respectively (SjoÈgren, Persson, Swedish University of Agricultural Science, personal communication)) and those values as well as the corresponding C-to-N ratios (32.9 and 23.2, re-spectively) are suciently high to qualify the soils as electron donor rich and potentially favourable for ammoni®cation.
Denitri®cation in soils is often limited by high oxy-gen concentrations and low nitrate and carbon concen-trations (Tiedje, 1988). Some observations suggest that
the denitri®cation rate depends on the NO3ÿ
concen-tration (Murray et al., 1989; Weier et al., 1991) and others suggest the opposite (Burford and Bremner, 1975; Myrold and Tiedje, 1985). Vermes and Myrold (1992) found that in situ denitri®cation rates in soil of pine and Douglas-®r were better correlated with soil
NO3ÿ concentrations than with the availability of C,
measured by soil respiration rates, or with soil moist-ure. This is in contrast to conditions in agricultural soils (Myrold and Tiedje, 1985; Drury et al., 1991) and
is possibly a consequence of a lack of organic C for
denitrifying microorganisms in agricultural soils
(McCarty and Bremner, 1992). The negative
corre-lation between extractable NO3ÿ concentrations and
denitri®cation in the StraÊsan soils and the inverse re-lationship between soil organic C concentrations and denitri®cation suggests that other factors, such as low pH, that tend to depress denitri®cation (MuÈller et al., 1980; NaÈgele and Conrad, 1990) are more important in controlling rates. It is also possible that the NO3ÿ
concentrations are too low at all four sites to allow anything but marginal variations in the denitri®cation rates.
Another possibility is that denitri®cation rates are underestimated. A potential source of error may be the reduced eciency of the acetylene blockage technique
in soils with low concentrations of NO3ÿ (Slater and
Capone, 1989; Rudolph et al., 1991). The presence of acetylene in bioassays of denitri®cation can also reduce the production of NO and potentially cause an
under-estimate of N2O production (Davidson, 1992;
McKen-ney et al., 1996). However, in acid soils such as those at StraÊsan, where N2O tends to be the dominating end
product of denitri®cation, acetylene has not been observed to have any inhibitory eect on the N2O
pro-duction (Knowles, 1982; Nielsen et al., 1994).
The larger activity of ammoni®ers compared with denitri®ers coincides with dierences in frequency of the two groups of soil bacteria visible on selective media after incubation (Fig. 4), possibly suggesting that, on average, activity and abundance are related to each other. In contrast to the general correspondence between dierences in activity and in abundance of ammoni®ers and denitri®ers, little agreement was observed between variations in abundance of ammoni-®ers or denitriammoni-®ers on one hand and the corresponding ammoni®cation and denitri®cation on the other. The frequency of nitrate reducers was correlated with the fertilizer dosage (Fig. 4) and, consequently, unrelated to the variations in nitrate reduction rates at the sites. The same kind of lack of correlation between denitri®-cation rates and denitri®er counts were observed by Volz et al. (1975), suggesting that the abundance of denitri®ers, many of which may be capable of using
other electron acceptors than NO3ÿ and probably a
variety of electron donors, may be underestimated by the lack of appropriate nutritional media for their counting. Parsons et al. (1991) also found weak corre-lations between denitri®cation rate and denitri®er bio-mass r20:1±0:3 when all observations, re¯ecting
spatial and temporal variations, were considered. The suggested explanation was that denitri®er growth is primarily controlled by the availability of soil carbon under aerobic conditions and that the cells start deni-trifying under anaerobic conditions by activating or synthesizing denitrifying enzymes.
Denitri®ers are not necessarily limited in population sizes by the concentration of O2 and NO3ÿ, both
deni-tri®ers and ammoni®ers are probably capable of aerobic respiration and may prevail in high numbers and also metabolize carbon under oxygenated con-ditions. This would obviously make attempts to corre-late abundance or frequency of denitri®ers, identi®ed
by their production of N2 on selective media, with in
situ denitri®cation rates dicult in soils with transient O2 conditions, such as well drained forest soils.
How-ever, a large number of strains capable of aerobic
deni-tri®cation have been observed (Robertson and
Kuenen, 1984; Anderson and Levine, 1986; Lloyd et al., 1987; Bengtsson and Annadotter, 1989; Robertson et al., 1995) and some aerobic denitri®ers might also be heterotrophic nitri®ers (Robertson et al., 1989). Oxygen can be used simultaneously with nitrate as an electron acceptor, thus promoting the growth rate of the strain and potentially stimulating the activity of synthesized nitrate reductases (Bonin and Gilewicz, 1991; Patureau et al., 1994). How common aerobic denitri®ers are in soils compared with denitri®ers that respire O2is unknown.
Acknowledgements
Dr. Dan Berggren at Swedish University of Agricul-tural Sciences kindly provided the soil samples and the pH data from StraÊsan. Dr. Tryggve Persson and Mr. Michael SjoÈberg at the same university are acknowl-edged for sharing their unpublished data on C and N concentrations in the StraÊsan soils. The work was sup-ported by the National Swedish Environment Protec-tion Board and Crafoordska Stiftelsen.
References
Aber, J.D., Nadelhoer, K.J., Steudler, P.A., Melillo, J.M., 1989. Nitrogen saturation in forest ecosystems. BioScience 39, 378±386. Anderson, I.C., Levine, J.S., 1986. Relative rates of NO and N2O
production by nitri®ers, denitri®ers and nitrate respirers. Applied and Environmental Microbiology 51, 938±945.
Bengtsson, G., Annadotter, H., 1989. Nitrate reduction in a ground-water microcosm, determined by 15N gas chromatography±mass spectrometry. Applied and Environmental Microbiology 55, 2861±2870.
Bengtsson, G., Bergwall, C., 1995. Heterotrophic denitri®cation po-tential as an adaptive response in groundwater bacteria. FEMS Microbiology Ecology 16, 307±318.
Binnerup, S.J., Jensen, K., Revsbech, N.P., Jensen, M.H., Sùrensen, J., 1992. Denitri®cation, dissimilatory reduction of nitrate to am-monium, and nitri®cation in a bioturbated estuarine sediment as measured with 15N and microsensor techniques. Applied and Environmental Microbiology 58, 303±313.
Bonin, P., Gilewicz, M., 1991. A direct demonstration of `co-respir-ation' of oxygen and nitrogen oxides by Pseudomonas nautica: some spectral and kinetic properties of the respiratory com-ponents. FEMS Microbiology Letters 80, 183±188.
Bremner, J.M., Blackmer, A.M., 1979. Eects of acetylene and soil water content on emission of nitrous oxide from soils. Nature 280, 380±381.
Buchmann, N., Gebauer, G., Schulze, E-D., 1996. Partitioning of 15
N-labeled ammonium and nitrate among soil, litter, below- and above-ground biomass of trees and understory in a 15-year-old Picea abiesplantation. Biogeochemistry 33, 1±23.
Burford, J.R., Bremner, J.M., 1975. Relationships between the deni-tri®cation capacities of soil and the total, water-soluble and readily decomposable organic matter. Soil Biology & Biochemistry 7, 389±394.
Cheng, H.H., Kurtz, L.T., 1963. Chemical distribution of added nitrogen in soils. Soil Science Society of America Proceedings 27, 312±316.
Daniels, L., Hanson, R.S., Philips, J.A., 1994. Chemical analyses. In: Gerhardt, G., Murray, R.G.E., Wood, W.A., Krieg, N.R. (Eds.), Methods for General and Molecular Bacteriology. American Society for Microbiology, Washington DC, pp. 514±554. Davidson, E.A., 1992. Sources of nitric oxide and nitrous oxide
fol-lowing wetting of dry soil. Soil Science Society of America Journal 56, 95±102.
Davidson, E.A., Hart, S.C., Firestone, M.K., 1992. Internal cycling of nitrate in soils of a mature coniferous forest. Ecology 73, 1148±1156.
De Boer, W., Klein Gunnewiek, P.J.A., Troelstra, S.R., 1990. Nitri®cation in Dutch heathland soils. II. Characteristcs of nitrate production. Plant and Soil 127, 193±200.
Drury, C.F., McKenney, D.J., Findlay, W.I., 1991. Relationships between denitri®cation, microbial biomass and indigenous soil properties. Soil Biology & Biochemistry 23, 751±755.
Freney, J.R., Denmead, O.T., Simpson, J.R., 1979. Nitrous oxide emission from soils at low moisture contents. Soil Biology & Biochemistry 11, 167±173.
Fujihara, S., Nakashima, T., Kurogochi, Y., 1986. Determination of 15NH3 by gas chromatography±mass spectrometry. Application to the measurement of putrescine oxidation by human plasma. Journal of Chromatography 383, 271±280.
Groman, P.M., Zak, D.R., Christensen, S., Mosier, A., Tiedje, J.M., 1993. Early spring nitrogen dynamics in a temperate forest landscape. Ecology 74, 1579±1585.
Hart, S.C., Firestone, M.K., Paul, E.A., Smith, J.L., 1993. Flow and fate of soil nitrogen in an annual grassland and a young mixed-conifer forest. Soil Biology & Biochemistry 25, 431±442.
Hinnman, W.C., 1966. Ammonium ®xation in relation to exchange-able K and organic matter content of two Saskatchewan soils. Canadian Journal of Soil Science 46, 223±225.
Johnsrud, S.C., 1978. Heterotrophic nitri®cation in acid forest soils. Holarctic Ecology 1, 27±30.
King, D., Nedwell, D.B., 1987. The adaption of nitrate-reducing communities in estaurine sediments in response to overlying nitrate load. FEMS Microbiology Ecology 45, 15±20.
Knowles, R., 1982. Denitri®cation. Microbiological Reviews 46, 43± 70.
KoÈlbel-Boelke, J., Tienken, B., Nehrkorn, A., 1988. Microbial com-munities in the saturated groundwater. Environment I: methods of isolation and characterization of heterotrophic bacteria. Microbial Ecology 16, 17±29.
Lloyd, D., Boddy, L., Davies, K.J.P., 1987. Persistence of bacterial denitri®cation capacity under aerobic conditions the rule rather than the exception. FEMS Microbiology Ecology 45, 185±190. Martikainen, P.J., Aarnio, T., Taavitsainen, V.-M., PaÈivinen, L.,
Salonen, K., 1989. Mineralization of carbon and nitrogen in soil
samples taken from three fertilized pine stands: long-term eects. Plant and Soil 114, 99±106.
McCarty, G.W., Bremner, J.M., 1992. Availability of organic carbon for denitri®cation of nitrate in subsoils. Biology and Fertility of Soils 14, 219±222.
McKenney, D.J., Wang, S.-W., Drury, C.F., Findlay, W.I., 1995. Denitri®cation, immobilization and mineralization in nitrate lim-ited and nonlimlim-ited residue-amended soil. Soil Science Society of America Journal 59, 118±124.
McKenney, D.J., Drury, C.F., Wang, S.W., 1996. Eect of acetylene on nitric oxide production in soil under denitrifying conditions. Soil Science Society of America Journal 60, 811±820.
Moore, A.W., 1965. Fixed ammonium in some Alberta soils. Canadian Journal of Soil Science 45, 112±115.
Moraghan, J.T., Buresh, R., 1977. Correction for dissolved nitrous oxide in nitrogen studies. Soil Science Society of America Journal 41, 1201±1202.
Murdoch, P.S., Stoddard, J.L., 1992. The role of nitrate in the acidi-®cation of streams in the Catskill Mountains of New York. Water Resources Research 28, 2707±2720.
Murray, R.E., Parsons, L.L., Smith, M.S., 1989. Kinetics of nitrate utilization by mixed populations of denitrifying bacteria. Applied and Environmental Microbiology 55, 717±721.
MuÈller, M.M., Sundman, V., Skujins, J., 1980. Denitri®cation in low pH spodsols and peats determined with the acetylene inhibition method. Applied and Environmental Microbiology 40, 235±239. Myrold, D.D., Tiedje, J.M., 1985. Establishment of denitri®cation
capacity in soil: eects of carbon, nitrate and moisture content. Soil Biology & Biochemistry 17, 819±822.
Nadelhoer, K.J., Downs, M.R., Fry, B., Aber, J.D., Magill, A.H., Melillo, J.M., 1995. The fate of15N-labelled nitrate additions to a northern hardwood forest in eastern Maine, USA. Oecologia 103, 292±301.
Nielsen, T., Kjùller, A., Struwe, S., 1994. Potential denitri®cation and N2O formation in dry European coniferous forest soils. Forest Ecology and Management 68, 101±106.
NaÈgele, W., Conrad, R., 1990. In¯uence of soil pH on the nitrate-reducing microbial populations and their potential to reduce nitrate to NO and N2O. FEMS Microbiology Ecology 74, 49±58. NoÈmmik, H., 1957. Fixation and de®xation of ammonium in soils.
Acta Agriculture Scandinavia 7, 395±436.
NoÈmmik, H., 1966. The uptake and translocation of fertilizer15N in young trees of Scots Pine and Norway Spruce. Studia Forestalia Suecica 35, 18 pp.
Parsons, L.L., Murray, R.E., Smith, M.S., 1991. Soil denitri®cation dynamics: spatial and temporal variations of enzyme activity, populations and nitrogen gas loss. Soil Science Society of America Journal 55, 90±95.
Patureau, D., Davison, J., Bernet, N., Moletta, R., 1994. Denitri®cation under various aeration conditions in Comamonas sp. strain SGLY2. FEMS Microbiology Ecology 14, 71±78. Persson, T., WireÂn, A., 1995. Nitrogen mineralization and potential
nitri®cation at dierent depths in acid forest soils. Plant and Soil 168±169, 55±65.
Poth, M., Focht, D.D., 1985.15N kinetic analysis of N2O production byNitrosomonas europaea: an examination of nitri®er denitri®ca-tion. Applied and Environmental Microbiology 49, 1134±1141. Priha, O., Smolander, A., 1995. Nitri®cation, dentri®cation and
mi-crobial biomass N in soil from two N-fertilized and limed Norway spruce forests. Soil Biology & Biochemistry 27, 305±310. Ritchie, G.A.F., Nicholas, D.J.D., 1972. Identi®cation of the sources
of nitrous oxide produced by oxidative and reductive processes in Nitrosomonas europaea. Biochemical Journal 126, 1181±1191. Robertson, L.A., Kuenen, J.G., 1984. Aerobic denitri®cation: a
con-troversy revived. Archiv fuÈr Microbiologie 139, 351±354. Robertson, G.P., Tiedje, J.M., 1987. Nitrous oxide sources in aerobic G. Bengtsson, C. Bergwall / Soil Biology & Biochemistry 32 (2000) 545±557
soils: nitri®cation, denitri®cation and other biological processes. Soil Biology & Biochemistry 19, 187±193.
Robertson, L.A., Cornelisse, R., de Vos, P., Hadioetomo, R., Kuenen, J.G., 1989. Aerobic denitri®cation in various hetero-trophic nitri®ers. Antonie van Leeuwenhoek 56, 289±300. Robertson, L.A., Dalsgaard, T., Revsbech, N.-P., Kuenen, J.G.,
1995. Con®rmation of `aerobic denitri®cation' in batch cultures, using gas chromatography and 15N mass spectrometry. FEMS Microbiology Ecology 18, 113±120.
Robertson, G.P., 1982. Nitri®cation in forested ecosystems. Philosophical Transactions of the Royal Society of London Bulletin 296, 445±457.
Rudolph, J., Frenzel, P., Pfennig, N., 1991. Acetylene inhibition technique underestimates in situ denitri®cation rates in intact cores of freshwater sediment. FEMS Microbiology Ecology 85, 101±106.
Schimel, J.P., Firestone, M.K., Killham, K.S., 1984. Identi®cation of heterotrophic nitri®cation in a Sierran forest soil. Applied and Environmental Microbiology 48, 802±806.
Skiba, U., Smith, K.A., Fowler, D., 1993. Nitri®cation and denitri®-cation as sources of nitric oxide and nitrous oxide in a sandy loam soil. Soil Biology & Biochemistry 25, 1527±1536.
Slater, J.M., Capone, D.G., 1989. Nitrate requirement for acetylene inhibition of nitrous oxide reduction in marine sediments. Microbial Ecology 17, 143±157.
Smolander, A., Kurka, A., Kitunen, V., MaÈlkoÈnen, E., 1994. Microbial biomass C and N, and respiratory activity in soil of repeatedly limed, and N- and P-fertilized Norway spruce stands. Soil Biology & Biochemistry 26, 957±962.
Sùrenson, J., 1978. Capacity for denitri®cation and reduction of nitrate to ammonia in a coastal marine sediment. Applied and Environmental Microbiology 35, 301±305.
Stark, J.M., Hart, S.C., 1997. High rates of nitri®cation and nitrate turnover in undisturbed coniferous forests. Nature 385, 61±64. Stevens, R.J., Laughlin, R.J., Burns, L.C., Arah, J.R.M., Hood,
R.C., 1997. Measuring the contributions of nitri®cation and deni-tri®cation to the ¯ux of nitrous oxide from soil. Soil Biology & Biochemistry 29, 139±151.
Tamm, C.O., Aronsson, A., Burgtorf, H., 1974. The optimum
nutri-tion experiment StraÊsan: a brief descripnutri-tion of an experiment in a young stand of Norway spruce (Picea albies Karst.). Research Note No. 17. Department of Plant Ecology and Forest Soils, Royal College of Forestry, Stockholm Sweden, 29 pp.
Tamm, C.O., 1985. The Swedish optimum nutrition experiments in forest stands: aim, methods, yield, results. Journal of the Royal Swedish Academy of Agriculture and Forestry Suppl. 17, 9±29. Tamm, C.O., 1991. Nitrogen in the terrestrial environment.
Questions of productivity, vegetational changes and ecosystem stability. In: Ecological Studies, 81. Springer Verlag, Berlin (116 pp.).
Tanaka, A., Nose, N., Fujimoto, Y., 1985. Determination of 15 N-labelled nitrate and nitrite by mass fragmentography and its ap-plication to plant material. Agricultural and Biological Chemistry 49, 291±299.
Tiedje, J.M., 1988. Ecology of denitri®cation and dissimilatory nitrate reduction to ammonium. In: Zehnder, A.J.B. (Ed.), Biology of Anaerobic Microorganisms. John Wiley, New York, pp. 179±244.
Tiedje, J.M., Sexstone, A.J., Myrold, D.D., Robinson, J.A., 1982. Denitri®cation: ecological niches, competition and survival. Antonie van Leeuwenhoek 48, 569±583.
Vermes, J.-F., Myrold, D.D., 1992. Denitri®cation in forest soils of Oregon. Canadian Journal of Forest Research 22, 504±512. Vitousek, P.M., Gosz, J.R., Grier, C.C., Melillo, J.M., Reiners,
W.A., 1982. A comparative analysis of potential nitri®cation and nitrate mobility in forest ecosystems. Ecological Monographs 52, 155±177.
Volz, M.G., Belser, L.W., Ardakani, M.S., McLaren, A.D., 1975. Nitrate reduction and associated microbial populations in a ponded Hanford sandy loam. Journal of Environmental Quality 4, 99±102.
Webster, E.A., Hopkins, D.W., 1996. Contributions from dierent microbial processes to N2O emission from soil under dierent moisture regimes. Biology and Fertility of Soils 22, 331±335. Weier, K.L., MacRae, I.C., Myers, R.J.K., 1991. Seasonal variation