Oecologia (2003) 136:183–192
DOI 10.1007/s00442-003-1258-8
ECOPHYSIOLOGY
Eva Castells · Josep Peuelas
Is there a feedback between N availability in siliceous
and calcareous soils and Cistus albidus leaf chemical composition?
Received: 2 August 2002 / Accepted: 6 March 2003 / Published online: 15 April 2003
Springer-Verlag 2003
Abstract The effects of the Mediterranean shrub Cistus
albidus on N cycling were studied in two siliceous
(granitic-derived and schistic-derived) and one calcareous
soil differentiated by their texture and acidity. We aimed
to find out whether soils under C. albidus were affected
by the release of C compounds from the canopy, and
whether phenolic compound production in C. albidus
changed depending on the soil N availability. Calcareous
soils, with higher clay content and polyvalent cations, had
a higher organic matter content but lower net N mineralization rates than siliceous soils, and C. albidus growing
therein were characterized by lower foliar N and phenolic
compound concentrations. Under C. albidus, all types of
soils had higher phenolic compound concentrations and
polyphenol oxidase activity. C. albidus presence and
leachate addition decreased net N mineralization and
increased soil respiration in siliceous soils, and these
changes were related to a higher soil C/N ratio under the
canopy. In calcareous soils, however, no significant
effects of plant presence on N cycling were found. In
the studied plant-soil system it is not likely that higher
phenolic compound concentrations were selected during
evolution to enhance nutrient conservation in soil because
(1) higher phenolic compound concentrations were not
associated with lower soil fertilities, (2) C compounds
released from C. albidus accelerated N cycling by
increasing N immobilization and no evidence was found
for decreased gross N mineralization, and (3) soil organic
E. Castells ()) · J. Peuelas
Unitat d’Ecofisiologia CSIC-CEAB-CREAF,
CREAF (Centre de Recerca Ecolgica i Aplicacions Forestals),
Universitat Autnoma de Barcelona,
08193 Bellaterra Barcelona, Catalonia, Spain
e-mail: [email protected]
Tel.: +1-217-3331165
Fax: +1-217-2443499
Present address:
E. Castells, Department of Entomology,
University of Illinois at Urbana-Champaign,
320 Morrill Hall, 505 S. Goodwin Ave., Urbana, IL 61801, USA
N content was more related to soil chemical and physical
properties than to the effects of the C. albidus canopy.
Keywords Leachates · Mediterranean vegetation ·
N cycling · N mineralization · Phenolic compounds
Introduction
Presence of plant canopy can be a major determinant of
small-scale patterns of nutrient availability by changing
organic inputs and soil microbial activity (Aguilera et al.
1999; Hook et al. 1991; Inderjit and Mallik 1999; Vinton
and Burke 1995). One of the ways plant presence can
influence soil N dynamics is through rainfall leachates of
phenolic compounds from green leaves and decomposing
litter. Phenolic compounds may play a dominant role in
controlling many aspects of plant-soil interactions, including nutrient availability and cycling, and organic
matter dynamics by either increasing or decreasing
microbial activity (Kuiters 1990; Northup et al. 1998;
Schimel et al. 1996). Phenolic compounds have been
shown to decrease soil N availability through several
mechanisms. First, they can form complexes with
proteins, thus delaying organic matter decomposition
and mineralization (Httenschwiler and Vitousek 2000;
Horner et al. 1988; Nicolai 1988; Palm and Sanchez
1990). Second, they can increase microbial activity and N
immobilization (Blum 1998; Blum and Shafer 1988;
Schimel et al. 1996; Shafer and Blum 1991; Sparling et al.
1981; Sugai and Schimel 1993). Third, they can directly
inhibit fungal respiration (Boufalis and Pellissier 1994)
and nitrification (Rice 1984). The presence and relative
importance of each one of those processes depend on the
plant species and the molecular size, structure and
concentrations of phenolic compounds (Fierer et al.
2001; Httenschwiler and Vitousek 2000; Horner et al.
1988; Kuiters 1990; Schimel et al. 1996).
Soil chemical and physical properties, such as pH, clay
content and nutrient status, can also play an important role
in the fate of phenolic compounds. Phenolic compounds
184
require oxidation for most of their ecological activities,
and their oxidation state varies with the physicochemical
conditions of the environment such as soil redox (pE and
pH) (Appel 1993). Soil pH also determines the type and
stability of bonds between phenolic compounds and
organic matter, which likely affects their lability. Thus,
at pH >8 phenolic compounds can form irreversible
covalent bonds with organic matter while at lower pH
they tend to form hydrogen bonds which are characterized
by their reversibility and lower strength (Appel 1993).
Presence of Ca can reduce the reactivity of phenolic
compound functional groups by mediating the formation
of bonds between clays and organic matter (Oades 1988).
Biological degradation rates of phenolic compounds can
also be influenced by pH and soil texture. Activity of
polyphenol oxidases (PPO), a family of enzymes that
participates in phenolic compounds degradation, has been
shown to increase with soil pH (Bending and Read 1995;
Pind et al. 1994) and decrease with high clay content and
cation-exchange capacity (Claus and Filip 1990).
There is also evidence that soil nutrient availability
affects production of phenolic compounds. Resource
availability hypotheses, such as carbon-nutrient balance
or growth-differentiation balance (Bryant et al. 1983;
Herms and Mattson 1992) predict increases in carbonbased secondary compound (CBSC) concentrations, such
as phenolic compounds, under high C-to-N availability.
One of the most interesting hypotheses on the reciprocal interactions between plant phenolic compounds and
soil nutrient cycling was proposed by Northup et al.
(1998). Phenolic compounds would exert a conservative
pressure on nutrients by binding proteins and delaying
decomposition and mineralization and thus slowing N
losses. This organic N would be specifically available for
those species producing phenolics by the action of
mycorrhizae. If this process enhances N conservation in
the soil, higher phenolic compound concentrations would
be selected during evolution in those plants growing on an
N-limited soil compared to plants with no nutrient
restriction (Bending and Read 1995; Northup et al.
1995b, 1998). The ecological relevance of phenolic
compounds can be especially interesting in ecosystems
such as Mediterranean ecosystems, where slow growing,
N-limited species predominate. These species are expected to have higher concentrations of CBSCs such as
tannins compared to fast-growing high nutrient requirement species (Coley et al. 1985).
In order to study the reciprocal interactions between
plant soluble phenolic compounds and soil N cycling
dynamics, we selected a species (Cistus albidus L.) that
releases phenolic compounds by rainfall and decomposing litter, and can be found both in more acidic siliceous
soils and more basic calcareous soils. We studied N
transformations in mineral soils under C. albidus and in
nearby sites without C. albidus in two siliceous (granitic
and schistic-derived soil) and one calcareous-derived soil
under similar conditions of precipitation, temperature and
radiation in a Mediterranean area in NE Spain. Those
soils were selected for their differences in pH and
carbonate content (granitic schistic <calcareous) and
clay content (granitic <schistic <calcareous). The soils
were also amended with leachates of C. albidus leaves
and incubated in controlled conditions. We aimed to
answer (1) whether presence of C. albidus individual
plants, as well as leachate addition, changed N cycling
dynamics and N availability, and whether those changes
were affected by soil type, and (2) whether C. albidus
phenolic compound concentrations changed depending on
the soil N status. We therefore wanted to find out whether
a positive feedback between plant phenolic compound
production and N availability can be found in our
Mediterranean plant-soil system. Under the hypothesis
of Northup et al. (1995b) we expect plants with higher
phenolic compound production to grow on more infertile
soils as these phenolic compounds stabilize organic
matter decreasing mineralization rates and thus avoid
losses of N from the system.
Materials and methods
Site description
The study sites were located in Capafons (Prades Mountains, SW of
Barcelona, Spain) where different lithologies converge. The mean
annual precipitation of this region is about 614 mm and the mean
annual temperature is 13.9C. We selected soils derived from three
Table 1 Physical and chemical properties of the soils selected for this study
Granodiorite
Horizons
A
Depth (cm)
0–9/10
pH
7.1
Electrical conductivity (dS/m) (1:5)
0.08
CaCO3 (%)
1
Organic matter (%)
2.0
Organic N (%) (Kjeldahl)
0.11
Phosphorus [P] (mg/kg) (Olsen)
5
Potassium [K+] (mg/kg)
132
Sand (% of fine earth)
92.5
Silt (% of fine earth)
4.8
Clay (% of fine earth)
2.7
USDA classification
sand
Schist
B
A
9/10–18/20
0–14/19
7.2
6.3
0.04
0.12
1
1
0.8
2.5
0.09
0.04
1
6
43
398
95.1
58.3
3.1
31.8
1.8
9.9
sand
sandy loam
Calcareous
B
14/19–56/59
7.0
0.07
1
0.9
0.04
2
152
62.6
26.6
10.8
sandy loam
C
56/59–100
7.2
0.11
1
0.5
0.03
3
81
77.0
15.1
7.9
sandy loam
A1
0–17
7.8
0.28
5
10.2
0.39
8
273
36.9
38.9
24.2
loam
A2
17–33
8.0
0.25
5
4.2
0.20
2
154
36.4
38.4
25.2
loam
B
33–68
8.3
0.17
5
2.3
0.12
1
131
39.8
41.3
18.9
loam
185
Plant and soil sampling
Fig. 1 A Cistus albidus petiole glands containing phenolic compounds. High density of glands and trichomes is also found on both
surfaces of green leaves and leaf litter. B Detail of glands and
trichomes. (Scale bar in A, 1 mm)
different bedrocks (granodioritic—granitic from now on for the
sake of simplicity, schistic, and calcareous—limestone) that differ
in their physical and chemical properties (Table 1). In brief,
siliceous soils (granite- and schist-derived soils) had lower pH,
electric conductivity and carbonate content compared to calcareous
soil. Organic matter, organic N and P were also lower in siliceous
soils. The soils differed in texture, with higher sand content in the
granitic soil and higher clay content in the calcareous soil. Schistic
soils had intermediate texture properties. To minimize differences
in temperature and precipitation, all sites were established within
600 m of one another at the same elevation (ca. 800 m) on S or SW
aspects. Vegetation from the calcareous soil was naturally burned
in summer 1994. The vegetation community at all sites was a low,
open shrubland with Cistus albidus and Quercus ilex as dominant
species. Other species present in granitic soil were Cistus
salviifolius, Corylus avellana, Rosmarinus officinalis, Cistus laurifolius, Ulex parviflorus and Lavandula sp.; in schistic soil: Pinus
halepensis, Corylus avellana, Vitis vinifera, Helichisum stoechas
and Euphorbia characias; and in calcareous soil: Pinus halepensis,
Quercus coccifera, Ulex parviflorus, Rosmarinus officinalis, Juniperus oxycedrus, Erica multifora, Dorycnium pentaphyllum, Daphne gnidium and Brachypodium retusum.
Cistus albidus is a Mediterranean evergreen shrub up to 1 m high
with leaf longevity not longer than 1 year, and frequently found in
xeric or disturbed sites. This species has ectomycorrhizae and can
be found both on siliceous and calcareous soils. Preliminary
analyses showed that they can release high concentrations of
phenolic compounds from green leaves and litter compared to other
shrubs present in the same vegetation community, including other
Cistus species. It also has a high density of hairs and glands both on
green shoots and on leaf surfaces which exude phenolic compounds
(Fig. 1).
Two different samples (winter and summer) were taken for soils
and plants. Since plant phenolic compound concentrations have
been shown to vary depending on the season (Feeny 1970), changes
in precipitation through the year (with maximum in spring and
autumn in Mediterranean climates) can affect the amount of soluble
C compounds released from the plant. The temporal variability in
the input of C compounds to the soil could affect nutrient dynamics.
Three and four sites (sampled during March and July 2000,
respectively) were located within each type of soil at an average of
75 m apart. At each site, we selected five C. albidus plants and five
adjacent nearby control plots located no more than 3 m apart. When
selecting control plots we avoided the presence of C. albidus or
other shrubs that could released C compounds from green leaves or
litter by rainfall, but not other herbaceous vegetation if present. For
the same reason, control plots were located at the same elevation or
slightly higher on the side of each selected C. albidus plant. Control
sites in the granitic and calcareous soils were bare ground and in
schist soils were covered by herbaceous vegetation. Under-plant
soil samples were taken just under C. albidus on the downslope side
of the stem where litter was present. The top 15 cm of soil was
sampled using a 5 cm diameter core. Two cores, one in control soil
and one underneath C. albidus were sampled in each plot (ten
samples in total) and bulked together within site. Soils were sieved
with a 2 mm mesh and kept at 4C until the experiments started.
C. albidus leaves were sampled in each one of the 5 selected
plants within site (15 plant samples per soil type in March and 20 in
July). They were dried at 65C for 48 h. Standing leaf litter was
also sampled from the same plants in June 2000.
Since leaf density may be correlated with the amount of
leachates released from the canopy and density may vary among
plants growing in the different soil sites, we estimated the C.
albidus shoot biomass per area of soil in each of the 60 selected
plants. Biomass of C. albidus plants was estimated by measuring
the basal diameter of all shoots per plant and their average height.
Thereafter, the shoot volume was calculated considering each shoot
as a cone shape. Mastrantonio (2000) found that the sum of shoot
volumes within an individual plant of C. albidus was a good
allometric estimator for plant biomass (r2=0.89). The area occupied
by each plant was calculated by measuring the canopy maximum
length and its perpendicular axis. Canopy shape was considered an
ellipse with a radius equal to half of the canopy length. We
estimated plant biomass per soil area (plant density) dividing the
resulting estimation of plant biomass by the canopy area. C. albidus
plants growing on granitic and schistic soils had higher biomass and
occupied a larger area than plants growing on calcareous soil
(P=0.002, t-test). However, plant biomass per soil area was
homogeneous among soil types (P=0.93, t-test).
Plant analyses and leachate preparation
Green leaves and leaf litter were extracted with 70% methanol for
45 min at room temperature. Total phenolic compound concentrations were determined by the Folin-Ciocalteu method improved by
using a blank of polyvinylpolypyrrolidone (PVPP) (Marigo 1973).
PVPP removes phenolic substances from the solution and avoids an
overestimation of total phenolic compounds due to non-phenolic
Folin-Ciocalteu reactive substances. Gallic acid was used as
standard. Condensed tannins were analyzed by the proanthocyani-
186
din method (Waterman and Mole 1994) and cyanidine chloride was
used as a standard.
In order to know whether the effects of soil type on phenolic
compound concentrations were caused by differences in water
stress we estimated water use efficiency of C. albidus growing on
siliceous and calcareous soils by analyzing leaf d13C ratio. The leaf
d13C ratio was measured on sub-samples (about 1 mg) of ground
leaves by a dual inlet triple collector isotope ratio mass spectrometer (SIRA II, VG Isotech, Middlewich, UK), operated in direct
inlet continuous flow mode. Before analysis, the solid samples were
combusted in a Dumas-combustion elemental analyzer (model NA
1500, Carlo Erba, Milan, Italy) for organic C analyses. The
reference CO2, calibrated against Pee Dee Belemnite standard
(PDB), was obtained from Oztech (Dallas, Tex., USA). The
analytical precision of these measurements of d13C was €0.1‰.
A C. albidus leachate was obtained by shaking fresh green
leaves (25 g equivalent DW), collected from schist-derived soils, in
1 l of distilled water for 24 h at room temperature (Zackrisson and
Nilsson 1992). The resulting leachate was filtered through a
Whatman 42 filter paper and analyzed for total phenolic compounds, condensed tannins (see Materials and methods) and
dissolved organic carbon (DOC) (Shimadzu TOC-5000). We also
measured phenolic compound and condensed tannin concentrations
in C. albidus leaves before and after making the leachate. A
fractionation of leachates into non-phenolic and phenolic fractions
using a C-18 solid-phase extraction column (Extra-sep, LIDA) was
analyzed for DOC in order to know what percentage of leachate
DOC was contributed by phenolic compounds.
The PPO activity was assessed using the method of NicholsOrians (1991), which determines the conversion of catechol to pquinone, a compound that absorbs light at 420 nm. PPO participates
in the degradation of phenolic compounds by breaking bonds
between phenolic compounds and other organic compounds.
Enzyme was extracted from soil by shaking the fresh soil with
50 mM of cold acetic acid-acetate buffer pH 5 (1:6 w/v) for 30 min,
centrifuging for 5 min at 4,200 rpm and filtering through Whatman
GF/A glass fiber filter; 500 ml of the resulting extract was added to
500 ml of 0.1 M citrate-phosphate buffer, 100 ml of distilled H2O
and 100 ml of 3% (w/v) catechol. As PPO activity varies with pH
(Bending and Read 1995; Leake and Read 1990), we conducted the
analyses using citrate-phosphate buffer at pH 6, 7 and 8, choosing
conditions similar to the studied soils. After mixing, the absorbance
of the solution was measured at 420 nm. Tubes were placed in a
25C water bath and after 1 h absorbance at 420 nm was
remeasured. The PPO activity was calculated by subtracting the
initial absorbance from the final absorbance. Results were
expressed as absorbance units (AU) per gram of soil DM h1.
Statistical analyses
A two-way ANOVA, one-trial repeated measures ANOVA or twotrial repeated measures ANOVA were conducted depending on the
experiment (see figure legends for more details). No transformations were needed in order to fit a normal distribution in any case.
An estimate of plant density was used as a covariant in the soil
experiments. Analyses were performed using Statistica ’99 Edition
(Statsoft, Tulsa, USA).
Incubations and analyses of soils
Soils sampled under C. albidus and in the nearby control sites were
incubated for 28 days. Mineral soil (30 g) was placed in glass jars
fitted with septa for taking headspace samples, and 3.5 ml of C.
albidus leachate or distilled water (control) was added once to the
soils just before starting the incubation. Field capacity was
measured from a soil subsample using a modified procedure from
Tan (1996). Additional distilled water was added to the soils until
field capacity when necessary, and they were incubated at 25C. A
soil subsample per type of soil (15 g FW) was extracted with 75 ml
2 M KCl for 1 h and analyzed for initial ammonium and nitrate by
an auto-analyzer (Flow Injection Analyzers, FOSS, Hgans,
Sweden). The C mineralization was estimated by analyzing CO2
accumulated in jars over a 4–7 day period by a gas chromatograph
provided with a Porapaq QS column and a thermal conductivity
detector (Hewlett Packard 5890 Series II). After each measurement,
jars were opened and ventilated in order to reestablish ambient CO2
concentrations, and distilled water was added when necessary to
maintain field capacity. After 28 days, final ammonium and nitrate
concentrations were measured. The net N mineralization rates were
calculated subtracting initial concentrations from final concentrations, and were expressed per unit of soil dry weight and per unit of
organic C. The C mineralization was calculated from the total CO2
produced during the incubation.
A ground soil subsample was analyzed for organic C and N with
an elemental analyser (model NA 1500, Carlo Erba, Milan, Italy)
using the standard configuration for those determination. Organic C
in calcareous soils was analyzed by the Walkley-Black procedure
from the rapid dichromate oxidation techniques (Nelson and
Sommers 1982). Soil pH and potential pH were analyzed after
shaking fresh soils with distilled water or 0.1 M KCl (1:2.5 w/v) for
30 min.
The phenolic compounds of the soil were extracted with NaOH
following Whitehead et al. (1982). Alkaline extractions are
normally used to remove phenolic compounds reversibly bound
to the soil. Twenty grams of soil was placed in a test tube with 50 ml
NaOH 2 M (1:2 w/v) under N2 and shaken for 20 h at room
temperature followed by filtration through Whatman 44 filter
paper. Concentrations of phenolic compounds were determined
using the Folin-Ciocalteu method described earlier.
Results
Plant analyses
Phenolic compound concentrations in C. albidus green
leaves ranged from 66.5 to 95.9 mg gallic acid g1 DM.
They varied during the year and with soil type, with
maximum phenolic compound concentrations in winter
and in plants growing on granitic and schistic soils
(Fig. 2). Litter phenolic compounds sampled during
winter ranged from 27.0 mg gallic acid g1 DM in
calcareous soil to 56.7 mg gallic acid g1 DM in granitic
soil, following the same pattern as in green leaves, with
higher concentrations in granitic and schistic soils than in
calcareous soil, but with lower values than in green
leaves. Condensed tannins and leaf N concentrations had
higher values in plants growing on granitic and schistic
soils and in winter leaves (Fig. 2). There was a significant
interaction between soil type and sampling time for leaf N
concentrations, with a stronger decrease in summer plants
growing on calcareous soils than those on other soils.
Leaf C/N ratio was lower in plants growing on granitic
and schistic soils and in leaves sampled during summer
(Fig. 2). Post hoc comparisons for soil type showed that
plants growing on granitic and schistic soils were
significantly different from plants growing on calcareous
soil in terms of phenolic compounds, condensed tannins,
N concentrations and C/N ratio (Fig. 2). Leaf d13C
analyses showed no differences in C fractionations
depending on date of sampling or soil type (data not
shown).
187
1
C. albidus leachate had 329.6 mg l of total phenolic
compounds, 67.2 mg l1 of condensed tannins and
1,190 mg l1 of DOC. Total phenolic compounds
comprised 46 % of the total DOC present in the leachate.
The leachate removed 33.7% of the phenolic compound
concentrations and 29.2% of the condensed tannin
concentrations present in C. albidus leaves.
Soil analyses
Fig. 2 Phenolic compounds, condensed tannins, N and C/N ratio
from C. albidus growing on granitic, schistic or calcareous soils.
Values represent means and SE (n=15 for phenolic compounds and
condensed tannins in winter samples and n=20 for summer
samples; n=9 for N and C/N in winter and summer samples).
Significant effects of the repeated measures ANOVA for soil type
(SOIL) and sampling season (SEASON) are shown. Different letters
show significant differences between soil type using a Tukey HSD
test for post hoc comparisons
Table 2 Mean values and results of a repeated measures ANOVA
for the effects of soil type (SOIL) and Cistus albidus presence
(PLANT) on soil pH, potential pH, phenolic compound concentrations and polyphenol oxidase (PPO) activity. The repeated
Granite
Schist
Calcareous
SOIL
PLANT
SOIL PLANT
Control
Under plant
Control
Under plant
Control
Under plant
Soils sampled under C. albidus had higher potential pH
(KCl) than control soils, and differences were stronger in
granitic soil (Table 2). Concentrations of reversibly bound
phenolic compounds were lower in granitic soils and
higher in calcareous soils. Within soil type they were
always higher in soils sampled under C. albidus than in
soils from nearby control sites (Table 2). PPO showed a
stronger activity measured at pH 6 in calcareous soils and
in soils sampled under C. albidus (Table 2). No PPO
activity was found at pH 7 or 8 in any soil type (data not
shown).
Organic C and N concentrations differed depending on
the soil type and date of sampling, with higher values in
calcareous soils and in winter (Fig. 3). Organic C
concentrations in all soils were higher under C. albidus.
A significant interaction was found between soil type and
sampling date, since calcareous soils had a stronger
decrease in organic C during summer compared to
siliceous soils. Calcareous soils had a higher C/N ratio
than siliceous soils. The significant interaction between
soil type and plant presence indicated that whereas C/N
ratio was higher under C. albidus in siliceous soils, no
changes due to plant presence were found in calcareous
soils (Fig. 3).
Soil type, plant presence and sampling time also had
an effect on soil microbial activity. Thus, potential net N
mineralization expressed on a weight basis was higher in
schistic soils than in granitic and calcareous soils, and
lower under C. albidus plants and in winter. Leachate
addition had no effects on net N mineralization expressed
measures ANOVA for plant presence was conducted using
estimated plant biomass per soil area as a covariant (n=15).
P<0.05 are highlighted in bold
pH (H2O)
pH (KCl 0.1 M)
Total phenolics
(mg g-1 d.m.)
PPO activity
(AU g-1 d.m. h-1 1,000)
6.7
6.9
6.3
6.3
7.9
8.0
df
2
1
2
5.6
6.1
5.5
5.8
7.4
7.5
df
2
1
2
0.43
0.62
0.87
0.93
1.25
1.34
df
F
2
37.3
1
5.1
2
0.5
0.99
2.49
2.10
2.77
6.05
6.55
df
F
2
39.1
1
10.33
2
0.91
F
71.9
1.3
1.2
P
<0.001
0.27
0.33
F
99.9
6.6
3.2
P
<0.001
0.03
0.08
P
<0.001
0.048
0.58
P
<0.001
0.012
0.43
188
Fig. 3 Organic C and N concentrations and C/N ratio from soils
sampled in winter and summer (SEASON) under C. albidus and in a
nearby control site (PLANT) in granitic (GRA), schistic (SCH) and
calcareous (CAL) soils. Values represent means and SE (n=4).
Significant effects of the two-trial repeated measures ANOVA for
SEASON and PLANT are shown. *** P<0.001, **P<0.01, *P<0.05.
Biomass density was used as a covariant in the statistical analyses
per gram soil (data not shown). However, net N
mineralization per unit C was higher in granitic and
schistic compared to calcareous soil and it changed
depending on plant presence and sampling date. For the
summer and winter samples C. albidus presence decreased N mineralization (per unit C) in siliceous soils but
no changes were found in calcareous soils. Plant presence
effect on net N mineralization was not related to sampling
date. Additions of plant leachate slightly decreased net N
mineralization in all soils although statistical differences
were marginally significant (Fig. 4).
C mineralization in soils sampled in summer was
higher for granitic and schistic soils and under C. albidus
in all soils and incubation dates although the effect was
stronger early in the incubation (Fig. 4). Leachate
addition also affected C mineralization the first 2 weeks
of soil incubation, although no global effects were found
(Fig. 4). The ratio between net N mineralized and C
mineralized, which indicate the relative limitation of
these nutrients, was higher for granitic and schistic
Fig. 4 Net N mineralization rates per unit C, soil respiration and
ratio between N-to-C mineralization from winter and summer soils
(SEASON) sampled under C. albidus and in a nearby control site
(PLANT) at granitic (GRA), schistic (SCH) and calcareous (CAL)
soils, and amended with leachate or distilled water (TREATMENT).
Values represent means and SE (n=4). Two-way (SOIL and
TREATMENT) two-trial repeated measures ANOVA for PLANT
and SEASON or PLANT and TIME were conducted for net N
mineralization and soil respiration respectively. TIME variable
refers to the date of sampling in a 1-month incubation period for
soils sampled in summer. Ratio between N-to-C mineralization was
tested by a two-way ANOVA (SOIL and TREATMENT) with a
repeated measures for PLANT. Statistically significant effects are
shown. *** P<0.001, ** P<0.01, * P<0.05. Biomass density was
used as a covariant in all cases
compared to calcareous soils. Under C. albidus this ratio
was lower than in the control ground in siliceous soils but
it was higher in calcareous soils. Leachate addition
decreased it independently of the analyzed soils (Fig. 4).
Post hoc comparisons for soil type showed different
responses of calcareous soil compared to granitic and
schistic in net N mineralization per unit C, C mineralization and N-to-C mineralization (data not shown).
189
Discussion
Plant presence effects on organic matter and N cycling
The effects of C. albidus presence on soil N cycling
depended on the soil type. A higher soil C/N ratio, a
higher soil respiration and lower net N mineralization
(that is the balance between the amount of N mineralized
or gross N mineralization and the amount of N immobilized) and N-to-C mineralization were found under C.
albidus in siliceous soils (granite and schist) compared to
the control sites. Regardless of a lower net N mineralization in soils under C. albidus and amended with leachates
the microbial activity increased. This lower N-to-C
mineralization ratio could only be explained by an
increase of inorganic N immobilization when microbes
used the C compounds released from the plant as a C
source. In calcareous soils, however, plant presence had
no effect on soil C/N ratio or microbial activity.
Differences between siliceous and calcareous soils in
their response to plant presence may be partially determined by differences in the proportional inputs of above
and below ground organic matter from C. albidus
compared to the accumulated organic matter under plants.
Since calcareous soils have higher concentrations of
organic matter compared to siliceous soils they could be
less constrained by releases of organic matter from C.
albidus. The lower responsiveness of calcareous soils to
C. albidus presence compared to siliceous soils could also
be explained by higher N-limitation compared to siliceous
soils. A low N-to-C mineralization ratio, such as in
calcareous soils, indicates the presence of an N limited
community (Schimel et al. 1996). In these conditions, the
use by microorganisms of carbon compounds released
from the plant as C-source would be restricted (Blum and
Shafer 1988). Physical and chemical properties of
calcareous soils could also determine a lower effect of
plant presence compared to siliceous soils. Calcium
present in calcareous soils reduces the reactivity of the
phenolic compounds functional groups and other organic
compounds by forming bonds with clays. Since both clays
and organic matter (humus) are negatively charged, the
polyvalent cations bridge organic molecules to clay
particles forming aggregates which are physically, chemically and biologically stable (Oades 1988). Degradation
of the complexes between phenolic compounds and
proteins are also affected by soil PPO, a family of
enzymes synthesized by ectomycorrhizal fungi that
mediate the first stage of phenolic compound mineralization and the consequent release of organic N (Leake and
Read 1990). In calcareous soils no PPO activity was
detected at pH 8, its actual pH in the field, and phenolics
may have lower degradation rates than granitic or schistic
soils.
Changes under C. albidus may be partially explained
by the effect of C compounds leached from green leaves
and litter. Soils under C. albidus and soils amended with
leachates had similar effects on soil N mineralization and
C mineralization. Since phenolic compound concentra-
tions and PPO activity in soils under C. albidus were
higher compared to control soils, leachate effects could be
due to phenolic compound presence. However, phenolic
compounds represent only 46% of the total DOC of
leachates and further analyses need to be made in order to
know whether leachate effects were caused by phenolic
compounds or by other C-based compounds.
Effects of soil type on C. albidus
leaf phenolic compounds
The few studies available comparing phenolic compound
production in plants growing on a acidity gradient in
natural conditions have found higher phenolic concentrations in plants associated with lower pH (Gylphis and
Puttick 1989; Muller et al. 1987; Nicolai 1988; Northup et
al. 1995b). This correlation is generally associated with
lower N availability generally found in acidic soils
(Northup et al. 1995a) although some studies have found
direct effects of pH (Laukkanen et al. 1997). Results from
our study follow the same trend as previous studies
related to soil pH, with higher phenolic compound
concentrations in C. albidus growing on siliceous soils
compared to individuals growing on calcareous soils.
However, foliar phenolic compounds were positively
related with soil N availability and thus leaf N content.
Moreover, phenolic compound concentrations within soil
decreased in summer when plants were subjected to lower
N availability. Changes related to nutrient availability in
CBSCs, such as phenolic compounds, have been explained by the resource allocation hypotheses carbonnutrient balance (Bryant et al. 1983) and growth-differentiation balance (Herms and Mattson 1992), which
predict increases in phenolic compound concentrations
when N availability decreases or C availability increases.
However, the positive trend between phenolic concentrations and N availability found in C. albidus cannot be
explained by the resource allocation hypothesis. There are
other factors besides soil nutrient status that could explain
increases in phenolic compound concentrations, such as
water stress (Gershenzon 1984; Tang et al. 1995). The
analyses of foliar d13C, which is a good estimator of water
use efficiency (Farquhar et al. 1989), showed no differences in water availability among soils. Many other
studies from the literature failed to demonstrate the
resource allocation hypothesis (Koricheva et al. 1998).
Hamilton et al. (2001) suggested that species historical
background, and not only plant nutrient availability,
would be especially relevant in order to predict responses
of secondary metabolites.
Is there a positive feedback between soil N availability
and plant phenolic compounds?
There is evidence that plant phenolic compounds can
affect N availability, and N availability can, in turn,
change phenolic compound synthesis. Some authors have
190
hypothesized that high plant phenolic compound concentrations could be selected during evolution in those
ecosystems with low N availability in order to minimize
N losses by leaching or denitrification (Bending and Read
1995; Northup et al. 1995b, 1998). These authors suggest
that high phenolic concentrations in plants growing on
infertile soils could benefit plant productivity by sequestering N into a large unavailable pool of recalcitrant
organic matter. The complexes between phenolics and
organic N would be only available to the mycorrhizae
associated with the plant producing phenolics. Since
phenolic compounds can decrease N availability by
delaying the N cycling when they bind proteins and other
N organic compounds (Bending and Read 1995; Oades
1988), they would have a conservative function of N in
the ecosystem (Horner et al. 1988).
Our data suggest that a positive feedback between
phenolic compounds and N conservation in the soil
system is not likely between C. albidus and soils in the
studied Mediterranean ecosystem. First, although a negative relationship between phenolic compounds and N
availability is expected from the “positive-feedback
hypothesis”, C. albidus growing on an N-poor calcareous
soil had lower phenolic compound concentrations than
plants growing on the N-rich acidic soil (Fig. 5). Second,
this hypothesis is based upon the capacity of phenolic
compounds in binding organic matter and thus decreasing
gross N mineralization. However, decreases in net N
mineralization associated with C. albidus presence were
mainly due to increases in N immobilization since higher
C mineralization and lower N-to-C mineralization was
found in soils sampled under plants. Although the effect
of phenolic compounds by binding organic matter and
decreasing the pool and form of nutrients has been
proposed to be the major link between phenolic compounds and nutrient cycling (Httenschwiler and Vitousek
2000), some studies found that phenolics primarily
increased microbial activity when biota used phenolic
compounds as a C source (Blum 1998; Fierer et al. 2001;
Schimel et al. 1996; Schmidt et al. 1997; Sparling et al.
1981; Sugai and Schimel 1993). In our study we cannot
reject the presence of phenolic-protein complexes, but
their effect decreasing N mineralization seems to be
counterbalanced by the effects of labile C phenolics.
Third, conservation of nutrients in the studied soil
system was related to physical and chemical soil properties independently of phenolic compound presence. In
calcareous soil, the slower microbial turnover due to high
clay content enhanced organic matter retention and thus
conservation of nutrients in the soil, while in siliceous
soils, which had low clay content, this process was not so
relevant. Plant-soil interactions between C. albidus and
siliceous and calcareous soils are described in Fig. 5. In
calcareous soils, the high content of clays and polyvalent
cations stabilized organic matter and slowed N cycling
rates (a), resulting in an accumulation of organic matter
(b). Litter inputs from the plant represented a low
proportion of the total organic matter present in the soil.
Thus, plant effects (c) were weaker than physical and
Fig. 5 Differences in plant-soil interactions for C. albidus growing
on siliceous (granite and schist) or calcareous soils. Arrow width
and letter size represent the proportional strength of the described
process. On calcareous soils, soil N cycling is strongly affected by
the presence of clays and carbonates. Clays form stable aggregates
with organic matter slowing down N cycling rate. As a result, soil
organic matter accumulates. The release of C compounds from the
plant has a minor effect compared to soil physical and chemical
properties. In siliceous soils, where clays have a weak effect on
organic matter stabilization, N cycling is driven by inputs from the
plant. Under low clay conditions, decomposition and mineralization
are faster and the resulting higher N availability causes higher leaf
N content compared to plants growing in calcareous soils. For a
more detailed explanation see the text
chemical effects in the calcareous soils (a). The slow N
cycling produced a low N availability (d), which in turn
influenced leaf chemical composition: plants had lower
leaf N content and higher C/N ratio. Phenolic compound
synthesis was also reduced. In siliceous soils (granite and
schist), the low clays and polyvalent cations content had a
weak effect on N cycling rates (a0 ), so that N cycling rates
were higher than in calcareous soils and the organic
matter retention was lower (b0 ). Inputs of organic matter
from the plant were expected to be similar compared to
calcareous soils since no differences in plant biomass per
soil area was found, but they were proportionally higher
than soil organic matter content. Thus, in siliceous soils,
the lower effect of soil physical and chemical properties
on N cycling rate (a0 ) determined a higher proportional
effect of plant presence (c0 ) than in calcareous soils.
Decomposition and mineralization was faster in siliceous
soils and the resulting higher N availability (d0 ) caused
higher leaf N and lower C/N ratio than in plants growing
on calcareous soils. Within siliceous soils, the plant effect
was stronger in granite compared to schist since the latter
had a higher clay content.
In summary, in the studied Mediterranean ecosystem it
is not likely that higher plant phenolic compound
concentrations had been selected during evolution in
order to increase N retention in nutrient limited soils
because (1) leaf phenolic compound concentrations were
not negatively correlated with soil N availability, (2) the
strongest effect on N cycling of carbon compounds
released from C. albidus, which include high phenolic
compound concentrations, was an increased N immobilization but not a decreased gross N mineralization, and
(3) the retention of organic matter in soil was more
191
determined by soil physical and chemical properties than
by plant phenolic compounds. Although we found that the
release of C compounds from plants affected strongly N
cycling, plant-soil interactions seem not to be the main
evolutionary drivers for the presence and concentration of
phenolic compounds in the studied species.
Acknowledgements The authors would like to thank Marta
Mastrantonio, Jordi Martnez, Roger Puig, Silvia Artesero, Bernat
Castells and Francesc Cebri
for field assistance. We also thank
Josep Maria Alcaiz, Ramon Vallejo and David W. Valentine for
critically reading this manuscript, and Oriol Ortiz for his expertise
help in soil analyses. E.C. had a predoctoral fellowship (FPI) from
the Ministerio de Educacion y Cultura (Spain) in collaboration with
Carburos Metalicos SA. We are thankful for financial support from
CICYT grants REN-2001–0003 and REN-2001–0278 (Spanish
Government).
References
Aguilera LE, Gutirrez JR, Meserve PL (1999) Variation in soil
micro-organisms and nutrients underneath and outside the
canopy of Adesmia bedwellii (Papilionaceae) shrubs in arid
coastal Chile following drought and above average of rainfall. J
Arid Environ 42:61–70
Appel HM (1993) Phenolics in ecological interactions: the
importance of oxidation. J Chem Ecol 19:1521–1552
Bending GD, Read DJ (1995) The structure and function of the
vegetative mycelium of ectomycorrhizal plants. VI. Activities
of nutrient mobilizing enzymes in birch litter colonized by
Paxillus involutus (Fr.) Fr. New Phytol 130:411–417
Blum U (1998) Effects of microbial utilization of phenolic acids
and their phenolic acid breakdown products on allelopathic
interactions. J Chem Ecol 24:685–708
Blum U, Shafer SR (1988) Microbial populations and phenolic
acids in soil. Soil Biol Biochem 20:793–800
Boufalis A, Pellissier F (1994) Allelopathic effects of phenolic
mixtures on respiration of two spruce mycorrhizal fungi. J
Chem Ecol 20:2283–2289
Bryant JP, Chapin FSI, Klein DR (1983) Carbon/nutrient balance of
boreal plants in relation to vertebrate herbivory. Oikos 40:357–
368
Claus H, Filip Z (1990) Effects of clays and other solids on the
activity of phenoloxidases produced by some fungi and
actinomycetes. Soil Biol Biochem 22:483–488
Coley PD, Bryant JP, Chapin FS III (1985) Resource availability
and plant antiherbivore defense. Science 230:895–899
Farquhar GD, Ehleringer JR, Hubick KT (1989) Carbon isotope
discrimination and photosynthesis. Annu Rev Plant Physiol
Plant Mol Biol 40:503–537
Feeny PP (1970) Seasonal changes in oak leaf tannins and nutrients
as a cause of spring feeding by winter moth caterpillars.
Ecology 51:565–581
Fierer N, Schimel JP, Cates RG, Zou J (2001) Influence of balsam
poplar tannin fractions on carbon and nitrogen dynamics in
Alaskan taiga floodplain soils. Soil Biol Biochem 33:1827–
1839
Gershenzon J (1984) Changes in the levels of plant secondary
metabolites under water and nutrient stress. Rec Adv Phytochem 18:273–320
Glyphis JP, Puttick GM (1989) Phenolics, nutrition and insect
herbivory in some garrigue and maquis plant species. Oecologia 78:259–263
Hamilton JG, Zangerl AR, DeLucia EH, Berenbaum MR (2001)
The carbon-nutrient balance hypothesis: its rise and fall. Ecol
Lett 4:86–95
Httenschwiler S, Vitousek PM (2000) The role of polyphenols in
terrestrial ecosystem nutrient cycling. Trends Ecol Evol
15:238–243
Herms DA, Mattson WJ (1992) The dilemma of plants: to grow or
defend. Q Rev Biol 67:283–335
Hook PB, Burke IC, Lauenroth WK (1991) Heterogenity of soil and
plant N and C associated with individual plants and openings in
North American shortgrass steppe. Plant Soil 138:247–256
Horner JD, Gosz JR, Cates RG (1988) The role of carbon-based
plant secondary metabolites in decomposition in terrestrial
ecosystems. Am Nat 132:869–883
Inderjit, Mallik AU (1999) Nutrient status of black spruce (Picea
mariana [mill.] BSP) forest soils dominated by Kalmia
angustifolia L. Acta Oecol 20:87–92
Koricheva J, Larsson S, Haukioja E, Keinnen M (1998) Regulation of woody plant secondary metabolism by resource
availability: hypothesis testing by means of meta-analysis.
Oikos 83:212–226
Kuiters AT (1990) Role of phenolic substances from decomposing
forest litter in plant-soil interactions. Acta Bot Neerl 39:329–
348
Laukkanen H, Julkunen-Tiitto R, Hohtola A (1997) Effect of
different nitrogen nutrients on the viability, protein synthesis
and tannin production of Scot pine callus. Physiol Plant
100:982–988
Leake JR, Read DJ (1990) Proteinase activity in mycorrhizal fungi.
I. The effect of extracellular pH on the production and activity
of proteinase by ericoid endophytes from soils of contrasted
pH. New Phytol 115:243–250
Marigo G (1973) Sur une mthode de fractionnement et d’estimation des composs phnoliques chez les vgtaux. Analusis
2:106–110
Mastrantonio M (2000) Master Thesis. Universitat Autnoma de
Barcelona, Barcelona
Muller RN, Kalisz PJ, Kimmerer TW (1987) Intraspecific variation
in production of astringent phenolics over a vegetation-resource
availability gradient. Oecologia 72:211–215
Nelson DW, Sommers LE (1982) Total carbon, organic carbon and
organic matter. In: Page AL, Miller RH, Keeney DR (eds)
Methods in soil analysis. II. Chemical and microbiological
properties. Soil Science Society of America, Madison, Wis.,
USA, pp 539–580
Nichols-Orians CM (1991) Condensed tannins, attine ants and the
performance of a symbiotic fungus. J Chem Ecol 17:1177–1185
Nicolai V (1988) Phenolic and mineral content of leaves influences
decomposition in European forest ecosystems. Oecologia
75:575–579
Northup RR, Dahlgren RA, Yu Z (1995a) Intraspecific variation of
conifer phenolic concentration on a marine terrace soil acidity
gradient; a new interpretation. Plant Soil 171:255–262
Northup RR, Yu Z, Dahlgren RA, Vogt KA (1995b) Polyphenol
control of nitrogen release from pine litter. Nature 377:227–229
Northup RR, Dahlgren RA, McColl JG (1998) Polyphenols as
regulators of plant-litter-soil interactions in northern California’s pygmy forest: a positive feedback? Biogeochemistry
42:189–220
Oades JM (1988) The retention of organic matter in soils.
Biogeochemistry 5:35–70
Palm CA, Sanchez PA (1990) Decomposition and nutrient release
patterns of the leaves of three tropical legumes. Biotropica
22:330–338
Pind A, Freeman C, Lock MA (1994) Enzymic degradation of
phenolic materials in peatlands- measurement of phenol
oxidase activity. Plant Soil 159:227–231
Rice EL (1984) Allelopathy, 2nd edn. Academic Press, Orlando
Schimel JP, van Cleve K, Cates RG, Clausen TP, Reichardt PB
(1996) Effects of balsam poplar (Populus balsamifera) tannins
and low molecular weight phenolics on microbial activity in
taiga floodplain soil: implications for changes in N cycling
during succession. Can J Bot 74:84–90
Schmidt IK, Michelsen A, Jonasson S (1997) Effects of labile soil
carbon on nutrient partitioning between an artic graminoid and
microbes. Oecologia 112:557–565
192
Shafer SR, Blum U (1991) Influence of phenolic acids on microbial
populations in the rhizosphere of cucumber. J Chem Ecol
17:369–388
Sparling GP, Ord BG, Vaughan D (1981) Changes in microbial
biomass and activity in soils amended with phenolic acids. Soil
Biol Biochem 13:455–460
Sugai SF, Schimel JP (1993) Decomposition and biomass incorporation of 14C-labeled glucose and phenolics in taiga forestfloor: effect of substrate quality, successional state, and season.
Soil Biol Biochem 25:1379–1389
Tan KH (1996) Soil sampling, preparations and analysis. Dekker,
New York, USA
Tang C, Cai W, Kohl K, Nishimoto RK (1995) Plant stress and
allelopathy. In: Inderjit, Dakshini KMM, Einhellig FA (eds)
Allelopathy: organisms, processes and applications. American
Chemical Society, Washington, D.C., USA, pp 142–157
Vinton AM, Burke IC (1995) Interactions between individual plant
species and soil nutrient status in shortgrass steppe. Ecology
76:1116–1133
Waterman PG, Mole S (1994) Analysis of phenolic plant metabolites. Blackwell, Oxford
Whitehead DC, Dibb H, Hartley RD (1982) Phenolic compounds in
soil as influenced by the growth of different plant species. J
Appl Ecol 19:579–588
Zackrisson O, Nilsson MC (1992) Allelopathic effects by Empetrum hermaphroditum on seed germination of two boreal tree
species. Can J For Res 22:1310–1319