Journal of Experimental Botany, Vol. 54, No. 386, pp. 1461±1469, May 2003
DOI: 10.1093/jxb/erg153
RESEARCH PAPER
The speed of soil carbon throughput in an upland
grassland is increased by liming
Philip L. Staddon1,4, Nick Ostle2, Lorna A. Dawson3 and Alastair H. Fitter1
1
Department of Biology, University of York, PO Box 373, York YO10 5YW, UK
2
Centre for Ecology and Hydrology, Merlewood Research Station, Grange-over-Sands, Cumbria LA11 6JU, UK
Macaulay Institute, Craigiebuckler, Aberdeen AB15 8QH, UK
3
Abstract
Introduction
In situ 13C pulse labelling was used to measure the
temporal and spatial carbon ¯ow through an upland
grassland. The label was delivered as 13C-CO2 to
vegetation in three replicate plots in each of two
treatments: control and lime addition. Harvests
occurred over a two month period and samples were
taken along transects away from the label delivery
area. The 13C concentration of shoot, root, bulk soil,
and soil-respired CO2 was measured. There was no
difference in the biomass and 13C concentration of
shoot and root material for the control and lime treatments meaning that the amount of 13C-CO2 assimilated by the vegetation and translocated below
ground was the same in both treatments. The 13C
concentration of the bulk soil was lower in the lime
treatment than in the control and, conversely, the 13C
concentration of the soil-respired CO2 was higher in
the lime. Unlike the difference in bulk soil 13C concentration between treatments, the difference in the
13
C concentration of the soil-respired CO2 was obvious only at the delivery site and primarily within 1 d
after labelling. An observed increase in the abundance of mycorrhizal fungi in the lime treatment was
a possible cause for this faster carbon throughput.
The potential key role of mycorrhizas in the soil carbon cycle is discussed. The importance of a better
understanding of soil processes, especially biological ones, in relation to the global carbon cycle
and environmental change is highlighted.
Predicting the impact of environmental change on terrestrial ecosystems is severely limited by a lack of understanding of basic soil processes (Lloyd, 1999; White et al.,
1999; Swift et al., 1998). In many, if not all, models of
terrestrial ecosystem response to global environmental
change, soil is simply treated as a black box (Kirschbaum,
1999). One question which frequently arises in connection
with global environmental change (GEC) and rising
atmospheric CO2 concentration in particular, concerns
how the terrestrial carbon cycle will be affected, i.e. will
GEC lead to increased C sequestration in soil or faster C
cycling (Grace and Rayment, 2000)? However, this is
rather dif®cult to answer, as it is ®rst necessary to have
some idea of how a perturbation to the soil biota affects the
soil carbon cycle. To answer questions of this nature, the
use of an in situ C tracer is a necessity. Stable isotopes
offer a safe and manageable means to do this. Stable
isotopes, in this case 13C, are now routinely used to answer
process-driven ecological questions in the ®eld (Ostle
et al., 2000; Stewart and Metherell, 1999). Due to the small
mass difference, 13C is considered to provide a more
representative biological tracer of the 12C dominated
terrestrial C cycle (Schimel, 1993; Simard et al., 1997)
than 14C (Gregory and Atwell, 1991; Meharg and Killham,
1995; Nguyen et al., 1999) or 11C (Freckman et al., 1991;
Minchin and McNaughton, 1984).
The NERC Soil Biodiversity Programme ®eld site at
Sourhope, Scotland includes various treatments, which
were imposed on a Festuca ovina±Agrostis capillaris
grassland on C-rich, base-poor mineral soil, to alter soil
biodiversity (www.nmw.ac.uk/soilbio). This study concentrated on one treatment, lime addition, which is known
Key words: 13C-CO2, extraradical mycorrhizal hyphae, pulse
labelling, root length colonized, soil carbon storage.
4
Present address and to whom correspondence should be sent: Risù National Laboratory, Plant Research Department, Building 313, Postbox 49, Roskilde,
DK-4000, Denmark. Fax: +45 4677 4122. E-mail: [email protected]
Journal of Experimental Botany, Vol. 54, No. 386, ã Society for Experimental Biology 2003; all rights reserved
Downloaded from http://jxb.oxfordjournals.org/ at Pennsylvania State University on February 27, 2014
Received 12 December 2002; Accepted 17 February 2003
1462 Staddon et al.
13
C labelling
The vegetation was labelled with 13C-CO2 using a mobile stable
isotope delivery system (SID) (for a full description see Ostle et al.,
2000). SID works by ®rst removing the CO2 from the air and then
replacing it in a mixing tank with 13C-CO2 (99 atom% 13C in this
experiment) to 370 ppm (controlled by an IRGA, Gascard II plus,
Edinburgh Sensors Ltd., Livingston, UK). The labelled air is then
supplied to individual labelling chambers via gas hoses. The air ¯ow,
c. 2.0 l min±1, to each of the chambers is adjusted to maintain similar
chamber CO2 concentrations, which are monitored by SID's CPUcontrolled IRGA (ADC Bioscienti®c Ltd., Hodersdon, UK) (the
chamber air is sucked back to SID in PTFE tubing). The chambers
used for labelling were made of plastic cylinders 20 cm in diameter
and 15 cm in height, ®tted with a transparent acrylic lid (3 mm
thick). The rapid turnover time of air in the labelling chamber meant
that the dilution of the 13C-CO2 label by soil and canopy respiration
was kept to a minimum (Ostle et al., 2000). There was a 1.6 cm
diameter port for the attachment of the in¯ow hose, and an exit port
of similar size through which the PTFE tubing, for individual
chamber CO2 monitoring, could be inserted. The chambers were
placed on the soil surface, taking care to minimize any vegetation
being trapped under the sides. After labelling, the chamber tops were
removed, but the chamber cylinders were left in place as markers for
the 13C labelled areas. The 13C labelling occurred on 28 September
2000, one month after the last cutting event (30 August 2000), and
lasted for 4 h during the brightest part of the day (from 12.00 h to
16.00 h) (see Fig. 1 for the climatic conditions during the
experiment). In total, there were three 13C-labelled replicates each
of the control and lime treatments located in three separate blocks
(plots 1B, 1F, 2B, 2C, 3B, and 3D at the site: www.nmw.ac.uk/
soilbio/Sourhope_Design.htm).
Materials and methods
Site, vegetation and treatment
The ®eld site is an upland ancient grassland located at the Sourhope
Research Station, Roxburghshire, Scotland. The site was previously
grazed for at least 40 years until livestock was excluded in 1998. The
plant community was dominated by the grasses Agrostis capillaris,
Festuca rubra, Nardus stricta, Anthoxanthum odoratum, and Poa
pratensis and closely matched the National Vegetation Classi®cation
community U4d (Cooper, 1997). This experiment formed part of the
NERC Soil Biodiversity Programme and used the control and lime
plots at the site (www.nmw.ac.uk/soilbio/Sourhope_Design.htm).
The lime was applied annually (May 1999 and 2000) at 600 g m±2 as
CaCO3 (powdered lime, 39.4% calcium ash insoluble in HCl
0.5%). Soil pH (measured in distilled water) was 4.5±5.0 in the
control plots and 5.5±6.0 in the limed plots at the time of this
experiment. During the summer (June±September 1999, May±
August 2000), the vegetation was cut to about 6 cm every 3±4 weeks
and the cuttings were removed. The following environmental
variables were continuously logged throughout the experimental
period using a Type DL2 automatic weather station (AWS; type
WS01, Delta-T, Cambridge, UK): solar radiation (kW m±2), rainfall
(mm), atmospheric temperature (°C) and soil moisture (m3 m±3).
Fig. 1. Environmental variables as measured during the course of the
experiment. (a) Solar radiation (line) and rainfall (bars). (b) Air
temperature (line) and volumetric soil moisture (bars).
Downloaded from http://jxb.oxfordjournals.org/ at Pennsylvania State University on February 27, 2014
to affect soil biota in terms of their community structure
and activity (Chagnon et al., 2001). As an initial study, the
primary objective was to determine whether there were
any differences in the fate of plant-®xed C between the
control and lime treatments. Also, lime addition, by
increasing pH, has been shown to stimulate mycorrhizal
fungal abundance (Wang et al., 1985). Mycorrhizal fungi
form symbiotic associations with the roots of most plant
species (Smith and Read, 1997). Arbuscular mycorrhizal
associations (the most common type of mycorrhizas) can
account for a relatively large proportion of plant photosynthate, in some cases up to 20% (Jakobsen and
Rosendahl, 1990). Their extensive hyphal networks in
the soil enables potentially a substantial amount of C to
enter the soil via mycorrhizal hyphae (Finlay and
SoÈderstroÈm, 1992; Staddon et al., 1999b). Mycorrhizas
could therefore be playing a key role in soil carbon cycling,
particularly in grasslands where they are abundant (Miller,
1987). This is particularly relevant to GEC as upland
grasslands hold a considerable stock of C and have, in the
past, been considered as net sinks of atmospheric C
(Goudriaan, 1995). Any environmental change that affects
soil microbial communities and alters mycorrhizal abundance is likely, therefore, to have a signi®cant effect on the
net ecosystem C ¯ux.
In the current study, using a ®eld-based 13C pulse
labelling experiment, in which the 13C label was followed
on both a temporal and spatial scale, an attempt was made
to test the following hypotheses: (1) lime addition results
in a faster throughput of carbon from plant to soil and back
to the atmosphere; and (2) lime addition results in less
carbon being sequestered in the soil.
Soil carbon cycling and mycorrhizas 1463
13
Plant biomass data
On 3 October 2000, duplicate cores, 3.5 cm internal diameter and
14 cm deep, were taken in each of the control and lime plots used for
the 13C labelling experiment (P and R subplots). Shoots were
collected and dried at 70 °C for 48 h and shoot biomass density was
expressed in mg cm±2 soil surface. Each soil core was washed over a
0.5 mm wire sieve to collect the roots. The roots from the 1±8 cm
depth were spread out onto a glass tray, ¯oated in water and
examined by eye using a magnifying lens. Dead and decomposing
root material was separated from the live roots on the basis of colour,
texture and breaking strength. Root samples were dried as above and
root biomass density was expressed as mg cm±3 soil volume.
Mycorrhizal parameters
Prior to the setting up of this experiment, mycorrhizal colonization
and extraradical hyphal density were assessed in the control and lime
treatments. Two cores, 2 cm diameter and 10 cm deep, were taken
randomly from the control and lime treatments in all ®ve of the
blocks at the site. Each core was processed for measurements of
extraradical mycorrhizal hyphal (EMH) density and root length
colonized (RLC) as follows. A subsample of soil was weighed fresh
and then placed in 500 ml of water. This was then mixed using a
magnetic stirrer and diluted 3-fold. From the ®nal solution, duplicate
5 ml samples where taken and passed through 0.45 mm ®lters under
vacuum for extraradical mycorrhizal hyphae collection (for a full
description of the procedure, see Staddon et al., 1999a). A small
random subsample of roots from each core was stained with acid
fuchsin for internal mycorrhizal assessment. The timing of the
staining procedure was 2±3 min in KOH at 80 °C, 1 min in HCl at
room temperature, 25 min in acid fuchsin at 80 °C, followed by
destaining in lactoglycerol (for a full description of the procedure,
see Staddon et al., 1998). Both the ®lters containing the extraradical
mycorrhizal hyphae and the stained roots were mounted in
lactoglycerol onto microscope slides. To allow for conversion of
soil fresh weight (FW) to soil dry weight (DW), the remaining soil
from each core was weighed fresh and then oven dried at 80 °C for
72 h.
Extraradical mycorrhizal hyphal density was assessed with a
compound microscope (Zeiss Jenamed 2) ®tted with a 131 cm 100grid graticule (Graticules Ltd, UK) at 3250; where a minimum of 40
grids per ®lter were observed. The grid-line intercept method
(Tennant, 1975) was used to obtain a length of hyphae per ®lter. The
mean value of the four ®lters (two ®lters per duplicate) per replicate
was used. Hyphal density is ®nally expressed in m hyphae g±1 soil
DW. Hyphae were counted as mycorrhizal if they showed the
following typical characteristics: dichotomous branching, angular
projections, and absence of septa (Nicolson, 1959). This standard
method (Kabir et al., 1997; Miller et al., 1995; Rillig et al., 2002;
Schweiger and Jakobsen, 1999) may slightly underestimate EMH
density by excluding some hyphae which may actually be
mycorrhizal. Internal mycorrhizal colonization was assessed with a
compound microscope (Nikon EFD-3 Optiphot-2) ®tted with a
cross-hair graticule at 3200 using epi¯uorescence (Merryweather
and Fitter, 1991) with a minimum of 100 intersections assessed per
core. Scoring followed McGonigle et al. (1990). The mean value for
the two cores per replicate was used. Further details on the methods
used for collection of mycorrhizal data are available in Staddon et al.
(1998, 1999a).
Statistical analysis
All statistical analyses were performed in SPSS 10.0. 13C label
concentration data was analysed both as a function of time after and
distance from source of labelling. The F-ratio method of ®tted line
comparison (Mead and Curnow, 1983; Sokal and Rohlf, 1995) was
used to test for differences between the control and lime treatments
for the concentration of 13C label in soil-respired CO2 and in the
shoots. Brie¯y, the F-ratio method compares the residual sum of
squares when two regression lines are ®tted to two data sets to that
when a single regression line (of the same order) is ®tted to a
combination of both data sets: i.e. the null hypothesis is that there is
no difference between the two data sets and that therefore there
should be no signi®cant decrease in variance of the data when two
lines are ®tted instead of one. To satisfy normality, the shoot data
were log10 transformed for the analysis. For both soil-respired CO2
and shoot 13C label concentration, non-linear regression was
performed to ®t asymptotic lines to the data (for P <0.05, R2 ranged
from 0.33 to 0.68). Due to the large number of zeros and general
variability in the root 13C and soil 13C label concentration data, the
Downloaded from http://jxb.oxfordjournals.org/ at Pennsylvania State University on February 27, 2014
Sampling and sample processing for C concentration
Sampling was performed on both a temporal and a spatial scale. The
sampling dates were 28, 29 September, 3, 10, 17, 24 October, and 1
and 29 November, which are equivalent to 2 h, 1, 5, 12, 19, 26, 34,
and 62 d after 13CO2 labelling. Samples were taken within the
labelled area and on random transects away from the centre of
the labelled area, at the following distances from the outer edge of
the chamber: 5, 10, 20, 35, and 50 cm. For each date and for each
plot, one sample was taken per distance. Samples for background
(natural abundance) 13C levels were taken a minimum of 5 m away
from the labelled areas (labelling occurred in the `P' subplots,
background levels were obtained from the `R' subplots at the site:
www.nmw.ac.uk/soilbio/Sourhope_Design.htm).
At each sampling point, a 10 ml Exetainer tube (Labco Ltd., High
Wycombe, UK) was placed upside down touching the soil surface.
Great care was taken to ¯ush the tubes with ambient air so they all
had the same starting point and vegetation was pulled back so as not
to include shoot respiration within the trapped soil-respired CO2.
The tubes were left in place for c. 1 h 45 min. On collection, each
tube was carefully lifted from the soil surface and capped. The soil
CO2 ef¯ux samples were stored at ambient temperature until
subsequent 13C/12C analysis.
After the collection of the gas sample, a shoot sample was
obtained as close as possible to where the Exetainer tube was located
and always at the same distance from the labelled area (i.e. if a little
movement was required, this was always perpendicular to the
transects). A soil core, 2 cm diameter and 10 cm deep, was then
taken at the same location as the gas sample. If this proved
impossible due to the rather stony nature of the site, then as for the
shoot sampling, any movement occurred perpendicular to the
transect. Shoot samples and soil cores were frozen until analysis.
Soil cores were subsequently processed as follows. Firstly, they were
defrosted at room temperature. Roots were extracted from each core
and a small random subsample of root (for 13C determination) was
placed inside an Eppendorf tube and then frozen. Great care was
taken to exclude all soil particles from the root samples. More timeconsuming was the exclusion of all root material (especially very
®ne dead fragments) from the (bulk) soil samples. Once collected,
the soil samples were also placed in Eppendorf tubes and frozen.
Prior to analysis, plant and soil samples were frozen at ±20 °C and
then freeze-dried using a `Spex' liquid nitrogen-cooled mill (Glen
Creston, Crewe, UK). Soil samples were pretreated with excess
0.1 M orthophosphoric acid to remove inorganic C salts (e.g. the
CaCO3). Tests con®rmed that this was suf®cient to remove CaCO3
as CO2 (data not shown). All samples were analysed for 13C
concentration by continuous ¯ow isotope ratio mass spectrometry at
the NERC Stable Isotope Facility at ITE Merlewood. Plant and soil
samples were analysed using an elemental analyser (Carlo Erba/
Fisons, UK) linked to a modi®ed isotope ratio mass spectrometer
(IRMS) (Dennis Leigh, UK) and soil-respired CO2 samples using a
Trace Gas preconcentration unit coupled to a Isoprime IRMS
(Micromass, Manchester, UK).
1464 Staddon et al.
Wilcoxon (paired) signed rank test (a non-parametric test) was used
for these two parameters. Data are presented as 13C atom% above
background (ABG) for the four parameters as a function of either
time or distance. Only a few of the most revealing graphs are
presented. In most cases, graphs have been truncated due to the
measured 13C atom% ABG being no different from zero, especially
at distances greater than 20 cm and more than 2±4 weeks after
labelling. Mycorrhizal data was analysed using a paired sample ttest. Percentage RLC was arcsine square root transformed and EMH
density was ln transformed to satisfy conditions of normality.
Results
13
C label concentration of the pools measured
Fig. 2. The 13C label content of shoots expressed as 13C atom% above
background (ABG) for (a) shoots located within the 13C labelled area
as a function of time after labelling and (b) shoots 1 d after labelling
as a function of distance from the labelled area. Treatments are control
(®lled diamonds, solid line) and lime (®lled squares, dotted line).
Error bars represent standard errors. For day 1 within the labelled
area, the equivalent d13C values range from ±28.0½ for the unlabelled
background to over +2100½ for the peak average.
Fig. 3. The 13C label content of roots expressed as 13C atom% above
background (ABG) for (a) roots located within the 13C labelled area as
a function of time after labelling, (b) roots located 5 cm from the 13C
labelled area as a function of time after labelling and (c) roots 1 d
after labelling as a function of distance from the labelled area.
Treatments are control (®lled diamonds, solid line) and lime (®lled
squares, dotted line). Error bars represent standard errors. For day 1
within the labelled area, the equivalent d13C values range from
±28.2½ for the unlabelled background to ±20.7½.
Downloaded from http://jxb.oxfordjournals.org/ at Pennsylvania State University on February 27, 2014
The shoots had the greatest concentration of 13C when
compared with the root and soil carbon pools. There was
no difference between the control and lime treatments for
the 13C label concentration of the shoots (F-ratio method,
P >0.05, see Fig. 2). Their 13C concentration peaked on the
®rst day after labelling (Fig. 2a). The vast majority of
the 13C label in the shoots was con®ned to those within the
labelled area, hardly any label was detected outside the
labelled area (Fig. 2b). The shoots within the labelled area
had a generally stable 13C concentration from 5 d to 62 d
after labelling of around 0.5 13C atom% above background
(Fig. 2a). Similarly to the shoots, there was no signi®cant
effect of the lime treatment on the 13C label concentration
of the roots (Wilcoxon test, P >0.05; Fig. 3), but the 13C
label concentration of the roots was two orders of
magnitude less than for the shoots. Contrary to the shoots,
there was no obvious peak in the 13C concentration of the
roots (Fig. 3a, b). Most of the 13C label in the roots was
located within the labelled area and up to a distance of
5 cm away (Fig. 3c), i.e. presumably only within roots
directly attached to the labelled shoots.
Contrary to there being no treatment effect on the plant
carbon pools, there was less 13C label in the bulk soil in the
lime treatment compared to the control (Table 1; Fig. 4).
For all data combined, this was highly signi®cant
(P <0.001, Wilcoxon signed ranks test). This effect of
less 13C in the bulk soil under lime was particularly evident
Soil carbon cycling and mycorrhizas 1465
Table 1. The signi®cance values (P) for the effect of the lime
treatment on the amount of 13C label located in the bulk soil
as obtained by Wilcoxon (paired) signed rank tests
Results are presented both as a function of time after and of distance
from labelling, for individual distances and dates, respectively. In
other words, tests were performed on all data from either a single
time point (all distances combined) or a single distance (all harvests
combined). NS, not signi®cant; *, P <0.05; ±, not available.
Versus time
0 cm
5 cm
10 cm
20 cm
NS (0.086)
0.036 *
NS
NS
Versus distance
2
1
5
12
h
d
d
d
NS (0.071)
±
0.026 *
0.028 *
Plant biomass and mycorrhizas
Standing plant biomass was unaffected by the lime
treatment (P >0.05). Shoot biomass was 18.7 (66.6 SE)
and 18.4 (67.9 SE) mg cm±2 for control and lime,
respectively, and root density was 2.22 (60.15 SE) and
2.40 (60.31 SE) mg cm±3 for control and lime, respectively.
However, root length colonized by arbuscular mycorrhizal
fungi was consistently and signi®cantly (P=0.007) less in
the control (33%) compared with the lime treatment (39%)
(Fig. 6). Although the overall difference in extraradical
mycorrhizal hyphal density between the control and lime
treatments (5.85 m 60.73 SE and 7.46 m 61.92 SE hyphae
g±1 soil dry wt, respectively) mirrored that for RLC (Fig. 6),
it was not signi®cant due to the vary large variation in
densities between cores. Lime increased RLC and EMH by
Fig. 4. The 13C label content of the bulk soil expressed as 13C atom%
above background (ABG) for (a) bulk soil located within the 13C
labelled area as a function of time after labelling and (b) bulk soil 5 d
after labelling as a function of distance from the labelled area.
Treatments are control (®lled diamonds, solid line) and lime (®lled
squares, dotted line). Error bars represent standard errors. For day 5
within the labelled area, the equivalent d13C values range from
±26.9½ for the unlabelled background to ±25.4½ (61.3 SE) and
±26.3½ (60.5 SE) for the control and lime treatments, respectively.
20% and 28% respectively (Fig. 6), the latter not being
signi®cant.
Discussion
Root
13
C concentration
Unlike these data, Ostle et al. (2000) reported a similar
magnitude of 13C concentration in the shoots compared to
the roots. However, they collected young roots visibly
attached to non-senescing shoots, whereas a random root
sample from a soil core was extracted in this work. There
are therefore several reasons why the 13C concentration of
the roots was so much lower than in the shoots in this
study: (a) the root sample included dead or senescing roots,
evident from the examination of the mixed samples of
roots, (b) roots attached to non-photosynthesizing shoots
were included and (c) a proportion of the sampled roots
were attached to shoots outside the labelled area (and could
therefore not directly have received any 13C label). During
root collection (for estimation of root biomass) the amount
of obviously dead roots was recorded as around 8%,
although the amount of dead or senescing roots can be as
high as 70% (Clark and Campion, 1976). In addition, there
Downloaded from http://jxb.oxfordjournals.org/ at Pennsylvania State University on February 27, 2014
close to the area of label delivery at 5±12 d after labelling
(Table 1). Similarly to the roots, there was no obvious peak
of 13C label in the bulk soil (Fig. 4a). The bulk soil 13C
label concentration gradually decreased with distance from
the labelled area (Fig. 4b).
The 13C label concentration of soil-respired CO2 was
also affected by the lime treatment (F-ratio method, P
<0.05; Table 2), but the effect was in the opposite
direction, i.e. there was more 13C label in the soil-respired
CO2 in the lime treatment when compared to the control
(Fig. 5). By contrast to the effect of lime on the 13C label
concentration of the bulk soil, the differences between the
lime and control treatments in the 13C concentration of the
soil-respired CO2 were primarily only seen at the 13C
delivery site (Fig. 5a; Table 2), and only immediately after
labelling (Fig. 5b, c; Table 2). The peak in 13C label in the
soil-respired CO2 occurred within a few hours of delivery
(Fig. 5a) and most of the released 13C label occurred within
5 cm of the labelled area (Fig. 5b). The maximum
concentration of 13C label in the soil-respired CO2 was of a
similar order of magnitude to that measured in the shoots,
i.e. it was 2 and 3 orders of magnitude higher than in the
roots and bulk soil, respectively.
1466 Staddon et al.
Table 2. The signi®cance values (P) for the effect of the lime
treatment on the amount of 13C label in the soil-respired CO2
as obtained by the F-ratio method of ®tted line comparisons
Results are presented both as a function of time after and of distance
from labelling, for individual distances and dates, respectively. In
other words, tests were performed on all data from either a single
time point (all distances combined) or a single distance (all harvests
combined). NS, not signi®cant; *, P <0.05; NA, not applicable (due
to large difference in variance).
Versus time
0 cm
5 cm
10 cm
20 cm
< 0.025 *
NS
NA
all 0
Versus distance
2
1
5
12
h
d
d
d
<0.05 *
NS
NS
all 0
Carbon ¯ow through the plant±soil system
Nearly all the energy that enters the soil does so via plants.
Roots are the primary producers in soil ecosystems
(Coleman and Crossley, 1996). When comparing the
carbon throughput in soils under different treatments, it is
critical to know if the total amount of carbon entering the
soil is comparable between treatments before assessing
carbon allocation to different pools. Here it is shown that
there was no difference in the amount of carbon ®xed by
the vegetation under the control and lime addition
conditions in this grassland. The value of the 13C
concentration peak measured in the shoots was identical
between control and lime and there was no statistical
difference in the standing shoot biomass. Also, root 13C
concentration and biomass was similar between the two
treatments, suggesting that there was a comparable amount
of photosynthate allocated below ground in the control and
lime treatments. The fact that the total amount of carbon
entering the soil in these two treatments was very similar
simpli®es the interpretation of the results. Consequently
other pools can be directly compared between treatments.
There were two main differences between the control
and lime treatments in terms of where and when the 13C
label was detected: a far greater amount of 13C was
immediately Returned to the atmosphere in the lime
compared to the control treatment and more 13C remained
in the bulk soil in the control compared to the lime
treatment. Previous tests (N Ostle, unpublished results)
showed that when labelling with 99 atom% 13C-CO2 for
5 h occurred in the dark (i.e. in darkened chambers) the
13
C-CO2 coming back from the soil and vegetation, as a
result of diffusion processes during the labelling, was
Fig. 5. The 13C label content of soil-respired CO2 expressed as 13C
atom% above background (ABG) for (a) soil-respired CO2 from
within the 13C labelled area as a function of time after labelling, (b)
soil-respired CO2 2 h after labelling as a function of distance from the
labelled area and (c) soil-respired CO2 1 d after labelling as a function
of distance from the labelled area. Treatments are control (®lled
diamonds, solid line) and lime (®lled squares, dotted line). Error bars
represent standard errors. For day 1 within the labelled area, the
equivalent d13C values range from ±20.8½ for the unlabelled
background to +662½ (6306 SE) and +1273½ (61286 SE) for the
control and lime treatments, respectively.
negligible by 1 h after the end of labelling. This means that
most of the 13C-CO2 evolving from the soil surface due to
diffusion processes would have occurred prior to the ®rst
sampling point (2 h after the end of labelling). Therefore,
the difference in the amount of rapidly returned 13C-CO2
(i.e. within hours), was most likely linked to proportional
differences in the components of below-ground respiration. Unless root respiration was signi®cantly affected by
the lime treatment, microbial respiration, including that of
mycorrhizal fungal, was the probable source of this. This
faster return of CO2 to the atmosphere in the lime
treatment resulted in less carbon being sequestered in the
soil itself, at least over the short term in this study.
Mycorrhizas, for example, have been shown to be
responsible for a large proportion of below-ground
Downloaded from http://jxb.oxfordjournals.org/ at Pennsylvania State University on February 27, 2014
would have been a considerable amount of root attached to
non-photosynthesizing shoots as the 13C labelling occurred
near the end of the growing season. Also, due to lateral root
spread (Weaver, 1926) and stoloniferous growth (Grime,
1979) of many grass species, not all roots within the
labelled areas were linked to the shoots within that area.
Soil carbon cycling and mycorrhizas 1467
weeks of plant growth in soil in their experiment), which
was maximal after 24 h. A similar timescale for the
partitioning of labelled carbon into microbial biomass was
observed by Warembourg and Estelrich (2000). Therefore,
in addition to the mycorrhizal fungi, bacteria and other soil
micro-organisms were likely to have had some role in the
observed rapid throughput of the labelled carbon in this
study.
Conclusions
respiration (Rygiewicz and Andersen, 1994). Lime addition, by increasing soil pH, generally leads to an increased
abundance of mycorrhizas (Wang et al., 1985; Coughlan
et al., 2000). Indeed, a 20% increase was found in the
amount of arbuscular mycorrhizas in the lime treatment
compared to the control, particularly within the roots. As
arbuscular mycorrhizal fungi can account for 10±20% of
plant photosynthate (Tinker et al., 1994; Jakobsen and
Rosendahl, 1990), it is possible that the increased abundance of mycorrhizas in the lime treatment was responsible, at least in part, for the faster short-term cycling of the
carbon back to the atmosphere. Also, there is the
possibility that mycorrhizal turnover and/or vitality could
have been increased by liming, further stimulating
mycorrhizal respiration, or that the higher pH in the
lime-treated plots favoured faster-growing mycorrhizal
fungal species.
However, an increase in soil pH, for example, by liming,
has many other effects on soil biota, such as a stimulation
of microbial respiration triggered by an increase in root
exudation (Meharg and Killham, 1990). It should be noted
that the conclusions of Meharg and Killham (1990) were
based on soil pH increasing from 4.3 to 6.5 and their data
for a soil pH increase similar to the one reported here
would lead to rather different conclusions (they reported a
lower soil respiration at pH 5.7 compared with pH 4.7). It
was not possible in this ®eld labelling experiment to
distinguish between the various components of soil
respiration, i.e. between root, mycorrhizal fungal, or
bacterial respiration. The authors are aware that some of
the increase in soil-respired 13C-CO2 could have been due
to the response of other biota apart from mycorrhizal fungi.
Rattray et al. (1995) reported a rapid partitioning of
labelled carbon in the rhizosphere, presumably to bacteria
(mycorrhizal fungi could not have established within the 2
As is often the case with ®eld data, the results raise many
more questions than they answer. For example, the
potential critical role of soil micro-organisms, and
mycorrhizal fungi in particular, in soil carbon cycling is
particularly pertinent to the global carbon cycle. If an
increase in abundance of soil micro-organisms simply
speeds up soil carbon cycling then many of the current
ideas on the role of soil biota in the terrestrial carbon cycle
must be re-evaluated. The key role of soil carbon cycling
and storage within the terrestrial carbon is undeniable, but
the processes affecting the fate of carbon entering the soil
are not well understood. These processes are both physicochemical and biological, with the latter being extremely
complex (Killham, 1994). Without a better understanding
of the role of soil biota in carbon cycling, the soil will
unfortunately continue to be treated as a black box, for
example, in terms of global carbon cycle modelling.
This research showed that lime addition resulted in
faster carbon cycling through the soil system. Both an
increase in soil respiration and a decrease in soil carbon
storage was observed. It was postulated that this was due to
the increase in mycorrhizal fungal colonization and
extraradical hyphal density as measured in the limed
plots. This is not to imply that other changes in the soil
biota did not occur, which undoubtedly they did.
Nevertheless, as far as is known, mycorrhizal fungi with
their direct access to plant photosynthate and their
relatively large biomass in relation to other soil organisms
(Coleman, 1994) were the most likely candidates to
account for the observed changes in carbon throughput.
This is particularly so, as the increase in soil-respired 13CCO2 took place within 1 d after labelling. Despite their
importance to the terrestrial carbon cycle, little is known
about the role of mycorrhizas in soil carbon storage. This is
an area requiring further research, especially with a view to
predicting consequences of global environmental change.
Acknowledgements
This research forms part of the Soil Biodiversity NERC Thematic
Programme. We thank Dr A Stott, H Grant and D Sleep at CEH
Stable Isotope Facility for analyses and N McNamara for assistance
in the ®eld. Also thanks to P Ineson for early discussions on the
planned work.
Downloaded from http://jxb.oxfordjournals.org/ at Pennsylvania State University on February 27, 2014
Fig. 6. Mycorrhizal root length colonized (RLC) and extraradical
mycorrhizal hyphal (EMH) density. Treatments are control (grey bars)
and lime (striped bars). Error bars represent standard errors. Data are
expressed relative to the mean of the control treatment (assigned a
baseline value of 100). Actual means for percentage RLC are 33% and
39% for control and lime, respectively; actual means for EMH density
are 5.85 and 7.46 mm g±1 soil dry wt for control and lime,
respectively.
1468 Staddon et al.
References
Downloaded from http://jxb.oxfordjournals.org/ at Pennsylvania State University on February 27, 2014
Chagnon M, Pare D, Herbert C, Camire C. 2001. Effects of
experimental liming on collembolan communities and soil
microbial biomass in a southern Quebec sugar maple (Acer
saccharum Marsh.) stand. Applied Soil Ecology 17, 81±90.
Clark FE, Campion M. 1976. Distribution of nitrogen in root
materials of blue grama. Journal of Range Management 29, 256±
258.
Coleman DC. 1994. Compositional analysis of microbial
communities: is there room in the middle? In: Ritz K, Dighton
J, Giller KE, eds. Beyond the biomass. Chichester: John Wiley
and Sons, 201±220.
Coleman DC, Crossley DA. 1996. Fundamentals of soil ecology.
San Diego, CA: Academic Press.
Cooper EA. 1997. Summary descriptions of National Vegetation
Classi®cation of grassland and montane communities.
Peterborough: UK Nature Conservation.
Coughlan AP, Dalpe Y, Lapointe L, Piche Y. 2000. Soil pHinduced changes in root colonization, diversity and reproduction
of symbiotic arbuscular mycorrhizal fungi from healthy and
declining maple forests. Canadian Journal of Forest Research 30,
1543±1554.
Finlay R, SoÈderstroÈm B. 1992. Mycorrhiza and carbon ¯ow to the
soil. In: Allen MF, ed. Mycorrhizal functioning. New York:
Chapman and Hall, 134±160.
Freckman DW, Barker KR, Coleman DC, Acra M, Dyer MI,
Strain BR, McNaughton SJ. 1991. The use of the C-11
technique to measure plant responses to herbivorous soil
nematodes. Functional Ecology 5, 810±818.
Goudriaan J. 1995. Global carbon cycle and carbon sequestration.
In: Beran MA, ed. Carbon sequestration in the biosphere. Berlin:
Springer-Verlag, 3±18.
Grace J, Rayment M. 2000. Respiration in the balance. Nature
404, 819±820.
Gregory PJ, Atwell BJ. 1991. The fate of carbon in pulse-labelled
crops of barley and wheat. Plant and Soil 136, 205±213.
Grime JP. 1979. Plant strategies and vegetation processes.
Chichester: John Wiley and Sons.
Jakobsen I, Rosendahl L. 1990. Carbon ¯ow into soil and external
hyphae from roots of mycorrhizal cucumber plants. New
Phytologist 115, 77±83.
Kabir Z, O'Halloran IP, Fyles JW, Hamel C. 1997. Seasonal
changes of arbuscular mycorrhizal fungi as affected by tillage
practices and fertilization: hyphal density and mycorrhizal root
colonization. Plant and Soil 192, 285±293.
Killham K. 1994. Soil ecology. Cambridge: Cambridge University
Press.
Kirschbaum MUF. 1999. Modelling forest growth and carbon
storage in response to increasing CO2 and temperature. Tellus
series B ± Chemical and Physical Meteorology 51, 871±888.
Lloyd J. 1999. Current perspectives on the terrestrial carbon cycle.
Tellus series B ± Chemical and Physical Meteorology 51, 336±
342.
McGonigle TP, Miller MH, Evans DG, Fairchild GL, Swan JA.
1990. A new method which gives an objective measure of
colonization of roots by vesicular±arbuscular mycorrhizal fungi.
New Phytologist 15, 495±501.
Mead R, Curnow RN. 1983. Statistical methods in agriculture and
experimental biology. London: Chapman and Hall.
Meharg AA, Killham K. 1990. The effect of soil-pH on
rhizosphere carbon ¯ow of Lolium perenne. Plant and Soil 123,
1±7.
Meharg AA, Killham K. 1995. Loss of exudates from the roots of
perennial ryegrass inoculated with a range of microorganisms.
Plant and Soil 170, 345±349.
Merryweather JW, Fitter AH. 1991. A modi®ed method for
elucidating the structure of the fungal partner in vesicular±
arbuscular mycorrhiza. Mycological Research 95, 1435±1437.
Miller RM. 1987. The ecology of vesicular±arbuscular
mycorrhizae in grass- and shrublands. In: Sa®r GR, ed.
Ecophysiology of VA mycorrhizal plants. Boca Raton, FL: CRC
Press.
Miller RM, Reinhardt DR, Jastrow JD. 1995. External hyphal
production of vesicular±arbuscular mycorrhizal fungi in pasture
and tallgrass prairie communities. Oecologia 103, 17±23.
Minchin PEH, McNaughton GS. 1984. Exudation of recently
®xed carbon by non-sterile roots. Journal of Experimental Botany
35, 74±82.
Nguyen C, Todorovic C, Robin C, Christophe A, Guckert A.
1999. Continuous monitoring of rhizosphere respiration after
labelling of plant shoots with (CO2)-C-14. Plant and Soil 212,
191±201.
Nicolson TH. 1959. Mycorrhiza in the Gramineae. I. Vesicular±
arbuscular endophytes, with special reference to the external
phase. Transactions of the British Mycological Society 42, 421±
438.
Ostle N, Ineson P, Benham D, Sleep D. 2000. Carbon assimilation
and turnover in grassland vegetation using an in situ 13CO2 pulse
labelling system. Rapid Communications in Mass Spectrometry
14, 1345±1350.
Rattray EAS, Paterson E, Killham K. 1995. Characterization of
the dynamics of C-partitioning within Lolium perenne and to the
rhizosphere microbial biomass using 14C pulse chase. Biology
and Fertility of Soils 19, 280±286.
Rillig MC, Wright SF, Shaw MR, Field CB. 2002. Arti®cial
climate warming positively affects arbuscular mycorrhizae but
decreases soil aggregate water stability in an annual grassland.
Oikos 97, 52±58.
Rygiewicz PT, Andersen CP. 1994. Mycorrhizae alter quality and
quantity of carbon allocated below ground. Nature 369, 58±60.
Schimel DS. 1993. Theory and application of tracers. San Diego,
CA: Academic Press.
Schweiger PF, Jakobsen I. 1999. Direct measurement of
arbuscular mycorrhizal phosphorus uptake into ®eld-grown
winter wheat. Agronomy Journal 91, 998±1002.
Simard SW, Durall DM, Jones MD. 1997. Carbon allocation and
carbon transfer between Betula papyrifera and Pseudotsuga
menziesii seedlings using a C-13 pulse-labelling method. Plant
and Soil 191, 41±55.
Smith SE, Read DJ. 1997. Mycorrhizal symbiosis, 2nd edn. San
Diego, CA: Academic Press.
Sokal RR, Rohlf FJ. 1995. Biometry, 3rd edn. San Fransico, CA:
W.H. Freeman and Company.
Staddon PL, Fitter AH, Graves JD. 1999a. Effect of elevated
atmospheric CO2 on mycorrhizal colonization, external hyphal
production and phosphorus in¯ow in Plantago lanceolata and
Trifolium repens in association with the arbuscular
mycorrhizal fungus Glomus mosseae. Global Change Biology
5, 347±358.
Staddon PL, Graves JD, Fitter AH. 1998. Effect of enhanced
atmospheric CO2 on mycorrhizal colonization by Glomus
mosseae in Plantago lanceolata and Trifolium repens. New
Phytologist 139, 571±580.
Staddon PL, Robinson D, Graves JD, Fitter AH. 1999b. The d13C
signature of the external phase of a Glomus mycorrhizal fungus:
determination and implications. Soil Biology and Biochemistry
31, 1067±1070.
Stewart DPC, Metherell AK. 1999. Carbon (C-13) uptake and
allocation in pasture plants following ®eld pulse-labelling. Plant
and Soil 210, 61±73.
Swift MJ, Andren O, Brussard L, Briones M, Couteaux M-M,
Soil carbon cycling and mycorrhizas 1469
Ekschmitt K, Kjoller A, Loiseau P, Smith P. 1998. Global
change, soil biodiversity, and nitrogen cycling in terrestrial
ecosystems: three case studies. Global Change Biology 4, 729±
743.
Tennant D. 1975. A test of a modi®ed line intersect method of
estimating root length. Journal of Ecology 63, 995±1001.
Tinker PB, Durall DM, Jones MD. 1994. Carbon use ef®ciency in
mycorrhizas: theory and sample calculations. New Phytologist
128, 115±122.
Wang GM, Stribley DP, Tinker PB, Walker C. 1985. Soil pH and
vesicular±arbuscular mycorrhizas. In: Fitter AH, Atkinson D,
Read DJ, Usher MB, eds. Ecological interactions in soil. Oxford:
Blackwell Scienti®c Publications, 219±224.
Warembourg FR, Estelrich HD. 2000. Towards a better
understanding of carbon ¯ow in the rhizosphere: a timedependent approach using carbon-14. Biology and Fertility of
Soils 30, 528±534.
Weaver JE. 1926. Root development of ®eld crops. New York:
McGraw-Hill Book Company.
White A, Cannell MGR, Friend AD. 1999. Climate change
impacts on ecosystems and the terrestrial carbon sink: a new
assessment. Global Environmental Change 9, S21-S30.
Downloaded from http://jxb.oxfordjournals.org/ at Pennsylvania State University on February 27, 2014