Positive climate feedbacks of soil microbial communities in a semi-arid grassland

Abstract

Soil microbial communities may be able to rapidly respond to changing environments in ways that change community structure and functioning, which could affect climate-carbon feedbacks. However, detecting microbial feedbacks to elevated CO(2) (eCO(2) ) or warming is hampered by concurrent changes in substrate availability and plant responses. Whether microbial communities can persistently feed back to climate change is still unknown. We overcame this problem by collecting microbial inocula at subfreezing conditions under eCO(2) and warming treatments in a semi-arid grassland field experiment. The inoculant was incubated in a sterilised soil medium at constant conditions for 30 days. Microbes from eCO(2) exhibited an increased ability to decompose soil organic matter (SOM) compared with those from ambient CO(2) plots, and microbes from warmed plots exhibited increased thermal sensitivity for respiration. Microbes from the combined eCO(2) and warming plots had consistently enhanced microbial decomposition activity and thermal sensitivity. These persistent positive feedbacks of soil microbial communities to eCO(2) and warming may therefore stimulate soil C loss.

Full text

LETTER Positive climate feedbacks of soil microbial communities in a
semi-arid grassland
Ming Nie,
1
* Elise Pendall,
1
Colin
Bell,
2
Caley K. Gasch,
3
Swastika
Raut,
1
Shanker Tamang
1
and
Matthew D. Wallenstein
2
Abstract
Soil microbial communities may be able to rapidly respond to changing environments in ways that change
community structure and functioning, which could affect climate–carbon feedbacks. However, detecting
microbial feedbacks to elevated CO
2
(eCO
2
) or warming is hampered by concurrent changes in substrate
availability and plant responses. Whether microbial communities can persistently feed back to climate
change is still unknown. We overcame this problem by collecting microbial inocula at subfreezing condi-
tions under eCO
2
and warming treatments in a semi-arid grassland field experiment. The inoculant was
incubated in a sterilised soil medium at constant conditions for 30 days. Microbes from eCO
2
exhibited an
increased ability to decompose soil organic matter (SOM) compared with those from ambient CO
2
plots,
and microbes from warmed plots exhibited increased thermal sensitivity for respiration. Microbes from the
combined eCO
2
and warming plots had consistently enhanced microbial decomposition activity and ther-
mal sensitivity. These persistent positive feedbacks of soil microbial communities to eCO
2
and warming
may therefore stimulate soil C loss.
Keywords
Climate change, climate–carbon feedback, common garden experiment, elevated CO
2
, enzyme stoichiome-
try, microbial community, soil decomposition, temperature sensitivity.
Ecology Letters (2012)
INTRODUCTION
Feedbacks between terrestrial C cycling and climate change deter-
mine the capacity for terrestrial ecosystem C storage (Pendall et al.
2004; Luo 2007). For example, warming can stimulate microbial
decomposition of litter and SOM, which accounts for two-thirds of
soil C losses in terrestrial ecosystems (Luo & Zhou 2006). However,
microorganisms may be able to rapidly respond to changing envi-
ronments in ways that alter community structure and functioning,
which could affect climate–carbon feedbacks (Luo 2007; Bardgett
et al. 2008; Wallenstein & Hall 2012). Numerous studies have
reported that climate change significantly affects microbial composi-
tion and biomass (Carney et al. 2007; He et al. 2010), enzyme activi-
ties (Carney et al. 2007; Zhou et al. 2012) and physiological profiles
(Zhou et al. 2012). Nevertheless, our understanding of microbial
feedbacks to climate change is still limited, especially for combined
effects of eCO
2
and warming, despite their importance for soil C
dynamics (Bardgett et al. 2008; Dieleman et al. 2012).
Microbial mediation of soil C cycling can be affected by climate
change primarily through two mechanisms. First, heterotrophic
microbial metabolism is regulated by the quantity and quality of
substrates, which can be affected by climate change (Hu et al. 2001;
He et al. 2010). eCO
2
often results in increased labile C inputs to
soils through plant root exudation and litter production (Phillips
et al. 2011). Because of diverse plant species and soil properties,
microbial metabolism responses to increased substrate availability at
eCO
2
have been observed to exhibit highly variable patterns, rang-
ing from positive (Carney et al. 2007), neutral (Zak et al. 2000; Aus-
tin et al. 2009) to negative feedbacks (Hu et al. 2001). Increased
labile C inputs resulting from eCO
2
can also stimulate SOM decom-
position in a process known as priming (Kuzyakov et al. 2000). For
example, eCO
2
increased the relative abundances of fungi and soil
enzyme activity in a scrub–oak ecosystem, and the stimulation of
SOM decomposition offset more than 50% of the stimulation in
plant belowground productivity, leading to a net decrease in soil C
(Carney et al. 2007). However, in other ecosystems belowground C
storage was enhanced by eCO
2
(Jastrow et al. 2000).
Second, warming may induce changes in the temperature sensi-
tivity of microbially mediated processes (Wallenstein & Hall
2012). Warming typically accelerates soil microbial respiration rates
due to increased soil enzyme activities, which drive decomposition
(Wallenstein et al. 2009, 2011). Primarily as a result of this
response, a 2 °C global average temperature increase is predicted
to stimulate soil C loss by 10 Pg year
1
(Pendall et al. 2004).
However, this stimulation of soil C loss could be entirely miti-
gated by a decline in the temperature sensitivity of microbial
activity (Allison et al. 2010). For example, one field experiment
showed that temperature sensitivity of soil respiration decreased at
high temperatures, possibly reducing the potential for C losses in
tallgrass prairie (Luo et al. 2001). Similarly, microbial temperature
sensitivity for respiration decreased in response to long-term soil
warming in a hardwood forest, even when incubated under non-
1
Department of Botany and Program in Ecology, University of Wyoming, Lar-
amie, WY, USA
2
Natural Resource Ecology Laboratory, Colorado State University, Fort Collins,
CO, USA
3
Department of Ecosystem Science and Management and Program in Ecology,
University of Wyoming, Laramie, WY, USA
*Correspondence: E-mail: mnie@uwyo.edu
©2012 Blackwell Publishing Ltd/CNRS
Ecology Letters, (2012) doi: 10.1111/ele.12034

limiting substrate conditions (Bradford et al. 2008). By removing
plant-mediated labile C availability in Arctic soil, Hartley et al.
(2008) used experimental cooling to avoid the confounding factor
of warming-induced substrate depletion and reported no declines
in microbial community temperature sensitivity. Microbial commu-
nities have also been observed to respond to variable precipitation
regimes (Evans & Wallenstein 2012) and redox fluctuations
(DeAngelis et al. 2010), suggesting other environmental drivers
which can initiate microbial feedbacks besides those associated
with atmospheric CO
2
.
Another challenge in assessing microbial feedbacks to climate
change is that responses may vary seasonally or interannually. For
example, SOM decomposition driven by microbial activity can shift
in response to temporal precipitation patterns (Parton et al. 2007;
Carrillo et al. 2011). The inherent complexity and diversity of micro-
bial communities and the many ways that they can be affected by
climate change and other environmental factors hamper our ability
to understand microbial mediation of soil C cycling (Bardgett et al.
2008). Determining microbial feedbacks to climate change requires
the separation of such changes from substrate availability and other
environmental factors (Bradford et al. 2008; Wallenstein & Hall
2012). Soil microorganisms can shift seasonally from dormant states
in frozen soils to a resuscitated state with high activity during the
plant growing season (Panikov 2009; Lennon & Jones 2011). For
microbial feedbacks to climate change to be important for ecosys-
tem response, they should be persistent. For example, dormant
microorganisms which can recover from winter and perform eco-
system processes during the plant growing season may represent a
persistent change in soil microbial activity in response to the previ-
ous years’ climate conditions (Panikov 2009; Lennon & Jones
2011).
Here, we collected soils in cold winter conditions from the long-
term Prairie Heating and CO
2
Enrichment (PHACE) experiment to
use as dormant microbial inocula in a sterile-soil incubation experi-
ment. This study site is unique because it is one of only a few field
studies in the world where both atmospheric CO
2
and temperature
are manipulated. We designed this experiment to address whether
any persistent soil microbial feedbacks exist in response to long-
term continuous climate change treatments. Our goal was to focus
on changes in microbial community structure and functioning by
excluding differences in plant activity, SOM properties and other
environmental factors by collecting off-plot soil as the substrate for
an incubation experiment. We measured differences in SOM
decomposition associated with C turnover and microbial tempera-
ture sensitivity (Q
10
) by assessing thermal responses for SOM
decomposition. Our previous work suggests that eCO
2
and com-
bined eCO
2
and warming gradually decrease soil organic C pool rel-
ative to expectations from high plant belowground biomass and C
inputs (Parton et al. 2007; Morgan et al. 2011; Carrillo et al. 2012),
and warming could increase soil respiration and decrease SOM con-
centrations over years (Parton et al. 2007; Carrillo et al. 2011).
Therefore, we hypothesised that (1) soil microbial communities
exposed to eCO
2
exhibit increased soil microbial respiration rates as
a result of increased microbial enzyme activity and subsequent
SOM decomposition, (2) soil microbial communities exposed to
warming increase their metabolic temperature sensitivity and (3) the
combined effects of eCO
2
and warming on microbial activities will
produce positive feedbacks with possible additive or synergistic
effects.
MATERIALS AND METHODS
Experimental design
Soil microbial inocula were prepared from the PHACE experiment,
which is located at the US Department of Agriculture Agricultural
Research Service (USDA-ARS) High Plains Grasslands Research
Station, Wyoming, USA (41°11′N, 104°54′W). The PHACE eco-
system is a northern mixed-grass prairie dominated by C
4
grass
Bouteloua gracilis (H.B.K) Lag and C
3
grass Pascopyrum smithii (Rydb.).
Mean annual precipitation (2006–2011) was 352.5 32.6 mm
(mean standard error) and mean air temperature (2006–2011)
was 20.2 0.6 °C in July and 2.9 1.0 °C in January. The soil
is a fine-loamy, mixed, mesic Aridic Argiustoll with pH of 7.9 (Dijk-
stra et al. 2010; Morgan et al. 2011). The experiment has imposed a
factorial combination of two levels of CO
2
(ambient and elevated
600 ppmv) since 2006, and two temperature regimes [ambient and
elevated (1.5/3.0 °C warmer day/night)] since 2007, with five repli-
cate plots (3.4 m diameter) of each treatment combination (ct,
ambient CO
2
and ambient temperature; Ct, elevated CO
2
and ambi-
ent temperature; cT, ambient CO
2
and elevated temperature; CT,
elevated CO
2
and elevated temperature). Warming is induced using
infrared heaters placed above the plant canopy (Dijkstra et al. 2010;
Morgan et al. 2011).
Three 5-cm-depth soil cores were collected from each treatment
plot on Dec 13 2011 using a 2.5-cm-diameter auger and then mixed.
In the field, soil temperatures were monitored to assess when they
reached 0°C before sampling (which occurred on Dec 3 2011;
Figure S1). PHACE soils were handpicked to remove rocks and
plant residual roots. Twenty soil–water slurry subsamples for each
experimental plot were prepared using 20 g soil and 20 mL sterile
DI water (van de Voorde et al. 2012). Off-plot soil was collected
from 0- to 5-cm depth near the PHACE plots to use as a common
growth substrate for the incubation. Off-plot soil was passed
through a 2-mm sieve to remove stones and plant residues, and then
homogenised to best produce an unbiased composite sample. After
sterilisation by autoclaving (121 °C, 45 min) twice in succession and
again 24 h later, the off-plot soil was again homogenised and then
pre-incubated at room temperature for a week in sterilised jars with
frequent ventilation by 0.45-lm filtered CO
2
-free air. We measured
respiration rates in the soil to evaluate sterilisation efficiency prior to
inoculating the microbial inocula. A very low amount of activity
(c. 0.2 lgCg
1
day
1
) was detected among all jars, which could be
due to abiotic CO
2
production, extracellular enzyme activities or
remnant microbial populations. However, this remnant CO
2
flux
contributed only c. 1.2% of the respiration in inoculated samples.
Each soil slurry (i.e. soil community) was inoculated by adding
1.5 mL of soil slurry solution to each specimen cup containing 30 g
sterile off-plot soil (c. 60% of WHC) in a 500-mL incubation jar.
The 1.5 mL of inoculum contained <5.7 lgCg
1
, compared with
23.6 mg C g
1
in soil originally present.
Each of the 20 PHACE soil communities were inoculated into
four replicate specimen cups for two different incubation experi-
ments: (1) Microbial community C dynamics: three jars were incu-
bated at 25 °C for 30 days. Headspace gas samples (30 mL) of one
jar were collected at 1, 2, 3, 5, 7, 10, 14, 20 and 30 days of incuba-
tion for measuring microbial respiration, and the other two jars
were harvested and combined after 10 days of incubation for mea-
suring microbial community structure and functioning. This was jus-
©2012 Blackwell Publishing Ltd/CNRS
2M. Nie et al. Letter

tified because microbial respiration rates on day 10 fell between the
stages of slow and rapid respiration decay (Figure S2); thus incuba-
tion samples on day 10 captured the major responses to the treat-
ments while avoiding longer term feedbacks that may have occurred
and (2) Temperature sensitivity: sample jars were incubated at 35 °
C and 25 °C for 30 days and measured on the same days as in
Experiment 1 for microbial respiration to evaluate Q
10
of microbial
SOM decomposition. We chose the 25 °C incubation temperature
based on the average soil ambient temperature during the main per-
iod of vegetative growth from June to August (22.9 0.6 °C); and
35 °C based on the highest daytime temperature during this stage.
All jars were thoroughly flushed with ambient air immediately after
gas sampling. Both incubation experiments had five laboratory repli-
cates, following the PHACE field replication. Uninoculated blank
jars (i.e. additions of 1.5 mL sterile DI water to sterile soil) were
incubated under the same conditions for each experiment.
Respiration
On each sampling date, a headspace sample of 30 mL was collected
by a syringe. Gas samples were directly analysed for CO
2
concentra-
tion using an infrared gas analyser (Li-Cor 820; LICOR Inc., Lin-
coln, NE, USA) calibrated with standard gases (Carrillo et al. 2011).
Temperature sensitivity (Q
10
) of SOM decomposition was deter-
mined using eqn 1:
Q10 ¼ðR2=R1Þð10=DTÞð1Þ
where R
2
and R
1
are the mean respiration rates at 35 °C and 25 °C
within 30 days of incubation (Balser & Wixon 2009) respectively.
Thus, DT(the difference in incubation temperatures) is 10 °Cin
this study.
Physiological profiles
Physiological profiles of microbial communities were assessed on
day 10 using the MicroResp system (Macaulay Institute, Aberdeen,
UK), which allows substrate-induced CO
2
measurements on whole
soil rather than dilutions in microbial growth media. A total of 15
C sources were chosen to cover a range of rhizodeposits, including
sugars, amino acids and carboxylic acids (Figure S3). Each well of a
96-well deep-well microplate was inoculated with 300 mg of the soil
and a solution of pre-dispensed C sources (Campbell et al. 2003).
The deep-well plate was then immediately sealed with the gasket
and a detection plate by a metal clamp. Colour development on the
detection plate was measured immediately before and after 6 h of
incubation at 25 °C using a Biotek microplate spectrophotometer
(BioTek Instruments Inc., Winooski, VT, USA) at 590 nm, and
converted to CO
2
evolution based on a standard calibration curve
prepared with standard gas mixtures (Campbell et al. 2003).
Enzyme assays
On day 10, activities of seven extracellular enzymes [b-Glucosidase
(BG; EC: 3.2.1.21), b-D-Cellubiosidase (CB; EC: 3.2.1.91), N-acetyl-
b-Glucosaminidase (NAG; EC: 3.2.1.14), Phosphatase (PHOS; EC:
3.1.3.1), b-Xylosidase (XYL; EC: 3.2.1.37), a-Glucosidase (AG; EC:
3.2.1.20) and leucine amino peptidase (LAP; EC: 3.4.11.1)] were
measured using 4-methylumbelliferyl (MUB) or 4-methylcoumarin
hydrochloride-linked (MUC) substrates yielding the highly fluores-
cent cleavage products MUB or MUC upon hydrolysis (Wallenstein
et al. 2009). All the enzyme assays were set up in 96-well micro-
plates. Briefly, 2.75 g of fresh soil was added into blender, and ho-
mogenised with 91 mL of 50 mMacetate buffer. Then 800 lLof
sample slurry was added into wells, which contained 200 lLof
200 lMsubstrate for enzyme activity measurement. Standard curves
were created with 200 mL of MUB or MUC solution and 800 lL
of sample slurry. Twelve replicate wells were set up for each sample
and each standard concentration. The assay plates were incubated in
the dark at 25 °C for 3 h. Fluorescence was measured using a
Tecan infinite M200 microplate fluorometer (Gr€odig, Austria) with
365 nm excitation and 460 nm emission filters. The activities were
expressed in units of nmol h
1
g
1
dry soil.
Microbial community structure
Microbial community structure was assessed by analysing the compo-
sition of extractable ester-linked phospholipid fatty acids (PLFAs) on
day 10 of the incubation. Lipids were extracted from 5 g of lyophi-
lised soil in a chloroform–methanol–phosphate buffer mixture
(1 : 2 : 0.8), and the phospholipids were separated from other lipids
on a solid-phase silica column (Agilent Technologies, Palo Alto, CA,
USA). The phospholipids were subjected to mild-alkaline methanoly-
sis, dissolved in chloroform and purified using a solid-phase amino
column (Agilent Technologies). The resulting fatty acid methyl esters
were dissolved in 0.2 mL 1 : 1 hexane:methyl t-butyl ether contain-
ing 0.25 mg 20 : 0 ethyl ester mL
1
, separated using an Agilent 6890
gas chromatograph with an Agilent Ultra 2 column (Agilent Technol-
ogies, and identified according to the MIDI eukaryotic method with
Sherlock software (MIDI Inc., Newark, DE, USA). The dominant
PLFAs were classified as Gram-positive bacteria (i14:0, i15:0, a15:0,
i16:0, i17:0 and a17:0), Gram-negative bacteria (16:1x9c, cy17:0,
18:1x9c and cy19:0), saprotrophic fungi (18:2x6c) and arbuscular
mycorrhizal fungi (AMF) (16:1x5c) (Zak et al. 2000). Each PLFA
and the sum of all PLFAs are expressed as lg PLFA g
1
dry soil.
Statistical analyses
To determine the effects of eCO
2
and warming on cumulative
microbial respiration, physiological profiles, enzyme activities and
community structure, we used a two-way ANOVA with eCO
2
and
warming as fixed effects using SPSS 13.0 (SPSS Inc., Chicago, IL,
USA). Data not meeting assumptions of normality and homogeneity
of variance were log-transformed before statistical testing. Power
regression and Pearson correlation analysis were performed to eval-
uate relationships underlying microbial processes using SPSS. Power
regression (y=ax
b
) was used to fit cumulative respiration (y) against
incubation day (x). In the power-regression equation, the coefficient
value (a) and exponent value (b) stand for initial respiration efflux
and cumulative respiration efficiency respectively. Canonical corre-
spondence analysis (CCA) was performed to determine how micro-
bial community structure was related to microbial function variables
using Canoco v4.5 (Microcomputer Power, NY, USA). Overall
CCA plot and their axes were significantly constructed (all
P<0.05) by a Monte Carlo permutation test in Canoco. Permuta-
tional multivariate analysis of variance (PERMANOVA) was used
to test whether eCO
2
and warming affected overall community-level
©2012 Blackwell Publishing Ltd/CNRS
Letter Microbial feedbacks to climate change 3

physiological profiles using PRIMER v6.1 (PRIMER-E Ltd., Ply-
mouth, UK). The Shannon–Weaver index (H’) and evenness index
(J) were used to determine whether eCO
2
and warming affected
physiological specificity for different C sources based on Micr-
oResp: H’=ΣPi (ln Pi), where Pi is the ratio of the activity of a
particular substrate and the sum of activities of all substrates;
J=H’/H’
max
, where H’
max
is the maximum level of physiological
diversity (Zak et al. 1994).
RESULTS
Respiration
After 30 days of incubation at 25 °C, microbes inoculated from
eCO
2
plots had significantly greater cumulative CO
2
efflux
(F
1, 19
=9.92; P<0.01; Fig. 1a) than communities from ambient
plots when averaged across warming treatments, whereas warming
treatment had no effect on cumulative CO
2
efflux when averaged
across CO
2
treatment (F
1, 19
=0.52; P>0.05; Fig. 1a). The cumula-
tive CO
2
efflux from CT plots marginally increased compared with
those from ct plots (P=0.10; Fig. 1a). The exponent values (b)
from power regressions were significantly higher in soils inoculated
from Ct and CT than those from ct plots (all P<0.05; Table 1),
suggesting that microbial communities from Ct and CT plots were
able to decompose the standardised SOM more rapidly than com-
munities from ct plots. Soils inoculated with microbes from warmed
Table 1 Power-regression coefficients (Y=aX
b
) of cumulative respiration after
30 days incubation at 25 °C
Treatment a b Curve-Fitting P
ct 2.22 0.17 a 0.57 0.02 a all <0.0001
Ct 1.94 0.16 a 0.64 0.02 b all <0.0001
cT 1.86 0.12 a 0.60 0.02 ab all <0.0001
CT 1.91 0.15 a 0.63 0.01 b all <0.0001
ct: ambient CO
2
and ambient temperature; Ct: elevated CO
2
and ambient tem-
perature; cT: ambient CO
2
and elevated temperature; CT: elevated CO
2
and ele-
vated temperature. Values are given as mean standard error (n=5). Values in
each column followed by the same letter are not significantly different at
P<0.05.
0.0
0.5
1.0
1.5
2.0
2.5
ct Ct cT CT
Temp. P < 0.01
Q10 value
a
a
b
b
(b)
(a)
0
6
12
18
ct Ct cT CT
Cumulative CO2 efflux
(mg C-CO2)
CO2 P < 0.01
ab
c
a
bc
Figure 1 Cumulative respiration by microbial communities for the entire jars
after 30-days incubation at 25 °C (a), and temperature sensitivity (Q
10
)of
decomposition (b). Cumulative CO
2
effluxes were adjusted by the organic C in
the 1.5 mL inoculants (C
cumulative efflux
C
inoculant
). ct: ambient CO
2
and
ambient temperature; Ct: elevated CO
2
and ambient temperature; cT: ambient
CO
2
and elevated temperature; CT: elevated CO
2
and elevated temperature.
Error bars show standard error of the mean (n=5). The same letters denote
non-significant differences between treatments (P>0.05).
0
1
2
3
4
ct Ct cT CT
Total PLFAs
(µg g−1 dry soil)
(a)
aaaa
0
2
4
6
ct Ct cT CT
Microbial respiration
/total PLFAs ratio
(b)
CO2 P < 0.05
ab
a
ab
Figure 2 Total PLFAs for soil microbial biomass (a), and microbial respiration
rate on a per-unit-PLFA basis (b). ct: ambient CO
2
and ambient temperature; Ct:
elevated CO
2
and ambient temperature; cT: ambient CO
2
and elevated
temperature; CT: elevated CO
2
and elevated temperature. Error bars show
standard error of the mean (n=5). The same letters denote non-significant
differences between treatments (P>0.05).
©2012 Blackwell Publishing Ltd/CNRS
4M. Nie et al. Letter

plots had a greater Q
10
of respiration measured in laboratory incuba-
tions than those inoculated from unwarmed plots (F
1, 19
=10.32;
P<0.01; Fig. 1b).
Physiological profiles
Treatments did not change the overall substrate use by PERMA-
NOVA (all P>0.05). Microbial physiological diversity (H’) and
evenness (J) were also not changed by any treatment (all P>0.05;
Table S1). However, microbes inoculated from eCO
2
plots were sig-
nificantly increased by 43.8% in total substrate respiration rate from
Ct than those from ct plots (P<0.05; Figure S3).
Enzyme assays
eCO
2
alone significantly increased total enzyme activity, CB, NAG
and PHOS; warming alone significantly increased PHOS; and com-
bined eCO
2
and warming significantly increased total enzyme activity
and NAG (Table S2; Figure S4). Cumulative CO
2
efflux after
30 days of incubation at 25 °C significantly increased with increasing
NAG (r=0.45; P<0.05), PHOS (r=0.39; P<0.05) and total of
seven enzyme activities (r=0.40; P<0.05) across the treatments.
Microbial community structure
Although total PLFA biomass of inoculated soils was not changed
by any treatments (all P>0.05; Fig. 2a), eCO
2
significantly
increased microbial respiration per unit of PLFAs (F
1, 19
=6.69;
P<0.05; Fig. 2b), which was calculated from total PLFA biomass
on day 10 and cumulative CO
2
efflux after 30 days of incubation.
Microbial communities represented by PLFA’s were separated by
treatment in canonical correlation analysis (Fig. 3). Axis 1 and axis
2 can be interpreted as the microbial activity gradient because of
their positive correlations with measured microbial functions (Fig. 3;
Table S3). Along axis 1 and axis 2, microbial communities from Ct
and CT plots appeared to have higher activity than those from ct
and cT plots (Fig. 3). Moreover, eCO
2
alone significantly decreased
16:1x9c (Gram-negative bacteria), warming alone significantly
increased a17:0 (Gram-positive bacteria), 18:1x9c (Gram-negative
bacteria), cy19:0 (Gram-negative bacteria) and 18:2x6c (saprotrophic
fungi) and combined eCO
2
and warming significantly increased
18:2x6c (Figure S5; Table S4).
DISCUSSION
Our findings add to a growing body of evidence that microbial com-
munities respond to climate change through functional changes that
affect the rates of ecosystem processes (Bardgett et al. 2008; Wallen-
stein & Hall 2012). Most previous studies of microbial responses to cli-
mate change were performed using individual climate factor studies;
for example eCO
2
(e.g. Hu et al. 2001; Carney et al. 2007; He et al.
2010; Phillips et al. 2011) or warming (e.g. Frey et al. 2008; Hartley
et al. 2008; Bradford et al. 2010; Zhou et al. 2012). The multiple inter-
acting effects of eCO
2
and warming on soil microbial communities
and their feedbacks to climate change are not well known (Bardgett
et al. 2008; Dieleman et al. 2012). In part, our current lack of under-
standing is hampered by concurrent climate interactions with changes
in substrate availability and/or plant responses (Bardgett et al. 2008).
In this study, we were able to test for potential microbial feedbacks
under eCO
2
and warming while controlling for confounding effects of
Total MicroResp
™
respiration efficiency
Total enzyme activity
Cumulative CO
2
efflux
Q
10
Low activity High activity
High activity
36
3
6
–3
–3
ct
cT
Ct
CT
Axis 1 (63.9%)
Axis 2 (30.5%)
Figure 3 Canonical correspondence analysis (CCA) ordination of microbial communities based on PLFAs. Arrow length indicates the importance of each microbial
function variables and their relationships with microbial communities. The community-function correlations were 0.73 and 0.53 for axis 1 and axis 2 respectively. ct:
ambient CO
2
and ambient temperature; Ct: elevated CO
2
and ambient temperature; cT: ambient CO
2
and elevated temperature; CT: elevated CO
2
and elevated
temperature. The data are means standard errors (n=5).
©2012 Blackwell Publishing Ltd/CNRS
Letter Microbial feedbacks to climate change 5

soil substrate heterogeneity and root activity (van de Voorde et al.
2012). Our results demonstrated positive feedbacks of soil microbial
communities to eCO
2
and warming, which may accelerate soil C losses
under future climate conditions. These results can explain, at least par-
tially, the lower rates of soil C accumulation relative to expectations
from high plant C inputs to soils or even a decline in the soil C pool
under climate treatments in our field observations and model predic-
tions (Parton et al. 2007; Carrillo et al. 2011, 2012; Morgan et al. 2011).
Increased microbial activities at eCO
2
Results from this experiment are consistent with model predictions
of increased soil respiration under eCO
2
at the PHACE experiment
(Parton et al. 2007), but suggest that these effects can be attributed
to microbial responses that are independent of changes in substrate
availability. eCO
2
has a significant influence on plant biomass accu-
mulation, and promotes increased C allocation to fine roots at the
PHACE experimental site (Morgan et al. 2011). Greater plant
belowground C allocation to soil through root exudation and root
litter would increase labile C availability and thus increase microbial
biomass and activity in the PHACE soils (Dijkstra et al. 2010; Car-
rillo et al. 2011, 2012). Previous studies suggest that eCO
2
could
increase the diversity of microbial genes related to decomposition
of C-rich substrates, such as cellulose (He et al. 2010; Weber et al.
2011). Increased cumulative CO
2
efflux in laboratory incubations
from Ct and CT plots (Fig. 1a; Table 1), despite similar microbial
biomass, suggests that these eCO
2
observations represent greater
microbial community metabolism, indicating community-level feed-
backs to eCO
2
.
In our experiment, autoclaving may have increased the pool of
labile C through the production of solubilised substrates (Trevors
1996), and thus overestimated potential SOM decomposition rates.
However, our observations followed a similar temporal pattern to
direct laboratory incubations of PHACE field soils under the same
incubation conditions, suggesting that the contribution of these sol-
ubilised substrates to respiration was transient (Figure S2) (Carrillo
et al. 2010). The average range of cumulative CO
2
efflux from the
inoculated soils was about half of that from PHACE soils (Carrillo
et al. 2010). This difference is probably explained by lower microbial
biomass in inoculated soils than field soils.
We suggest that the mechanism accounting for increased micro-
bial respiration by communities exposed to eCO
2
was increased res-
piration per unit biomass (Hu et al. 2001); because we observed
enhanced eCO
2
respiration on a per-unit-PLFA basis with no
change in total PLFA biomass (Fig. 2). Thus, these findings suggest
that eCO
2
may decrease microbial C utilisation efficiency and
increase respiration loss in this prairie ecosystem. Our study demon-
strates respiration per unit biomass in microcosms. Further assess-
ment is needed to extend these findings to a field experiment.
Microbial taxa are known to specialise in the degradation and
metabolism of specific substrates, and substrate availability may
drive community composition through species sorting (Hanson et al.
2008). Thus, microbial communities have the potential to alter res-
piration rates for specific C sources in response to altered soil C
availability under climate change. We observed greater CO
2
emis-
sion rates by Ct communities compared with the other treatment
communities for 15 of the substrates assessed (Figure S3). However,
soil microbial communities within the cT plots did not show a
higher ability to decompose C than those from Ct plots based on
CO
2
respiration rates. This suggests that microbial communities in
the eCO
2
plots have a higher capacity to decompose SOM than
microbial communities from warming plots, possibly because eCO
2
induced microbial communities to prefer the environment of high
quantity of plant-derived C inputs with low substrate quality (high
C : N) (van Veen et al. 1991). We also observed that experimental
climate change did not alter microbial community-level C substrate
specificity, as no differences in microbial physiological diversity (H’)
or evenness (J) were observed among treatment plots (Table S1).
This suggests that the differences in physiological profiles between
microbial communities were mainly due to community-level changes
in respiration efficiencies rather than their specificities for different
C sources. Previous studies suggested that microbial functional
genes involved in labile C decomposition and enzyme activities were
increased by eCO
2
(Carney et al. 2007; He et al. 2010). Therefore,
increased community-level decomposition rates may be due to ele-
vated microbial metabolic activity occurring within the eCO
2
plots.
The extracellular enzyme data showed that the microbial commu-
nities from Ct and CT plots had higher enzyme activities pertinent
to SOM decomposition, including CB, NAG, PHOS and AG, than
ct and cT communities (Figure S4). Our results are consistent with
previous studies which found that eCO
2
stimulated soil enzyme
activities and thus accelerated SOM decomposition (Carney et al.
2007; Phillips et al. 2011). Out of the seven enzyme activities mea-
sured, PHOS had the highest activity across all treatments (Figure
S4), probably due to P limitation in this prairie ecosystem (Dijkstra
et al. 2012). As an indicator of potential N : P acquisition activity
(Sinsabaugh et al. 2008), the ratios of (NAG +LAP)/(PHOS) were
lower in microbial communities from Ct (P<0.05) and CT
(P=0.06) plots compared with ct communities, suggesting micro-
bial N/P acquisition activities could be mediated by eCO
2
. Our pre-
vious study found that soil P availability increased relative to N in
eCO
2
(Dijkstra et al. 2012), which could be due to deceased ratios
of (NAG +LAP)/(PHOS).
Increased microbial temperature sensitivity under warming
We compared the abilities of microbial communities to decompose
SOM at two different incubation temperatures. The highest Q
10
was
observed in soil inoculums from the increased temperature plots (cT
and CT plots) (Fig. 1b). This implies that warming can amplify the
microbial community’s temperature sensitivity, which could result in
a further increase in microbial respiration with rising temperature.
Warming could increase C-decomposition genes, such as cellulose-
and chitin-degradation genes (Yergeau et al. 2012; Zhou et al. 2012),
which may account for a high temperature sensitivity of microbial
respiration. The current study measured the temperature sensitivity
of microbial respiration rates based on the data taken at 25 °C and
35 °C. The temperature sensitivity of microbial respiration rates at
lower temperature range may have to be further studied.
Previous studies showed that microbial temperature sensitivity had
a negative response to warming (Luo et al. 2001; Bradford et al.
2008). However, this phenomenon is difficult to parse from changes
in substrate availability in field studies or laboratory incubations. For
example, Belay-Tedla et al. (2009) demonstrated that warming signifi-
cantly increased labile C/N fractions through plant biomass input,
which increased soil microbial biomass. In our study, we were able
to avoid influences from soil initial heterogeneity (e.g. initial sub-
strate availability) and root activity. In laboratory incubations, sub-
©2012 Blackwell Publishing Ltd/CNRS
6M. Nie et al. Letter

strates become depleted faster at higher temperatures, which may
lead to an apparent decline in temperature sensitivity (Bradford et al.
2010; Tucker et al. 2012). Therefore, we compared microbial respira-
tion during the initial 1 or 2 days to calculate Q
10
, to test whether
microbial temperature sensitivity changed with minimum variation in
substrate availability during incubation (Steinweg et al. 2008). Like-
wise, we found that microbes inoculated from warmed plots exhib-
ited a significantly greater Q
10
of respiration measured within initial
1–2 days than those from unwarmed plots (all P<0.05; Figure S6).
We determined that microbial temperature sensitivity was not
affected by incubation duration (which may have affected substrate
availability) (three-factor ANOVA,CO
2
and warming as fixed effects
and incubation day as random effect; all P>0.05; Figure S6). In
fact, warming (F
1, 159
=106.49; P<0.0001; Figure S6) and com-
bined CO
2
and warming effect (F
1, 159
=6.54; P<0.05; Figure S6)
significantly increased microbial temperature sensitivity during all
30 days of incubation. These results suggest that warming increases
microbial temperature sensitivity regardless of changes in substrate
availability during the incubation.
Increased temperature sensitivity in response to warming suggests
that there are little to no trade-offs associated with community-
physiological constraints under changing temperature (Allison et al.
2010). If trade-offs occur, microbial respiration rates will negatively
respond to a warmer environment (Allison et al. 2010). Although
warming may decrease microbial respiration due to substrate deple-
tion, microbial baseline respiration can be stimulated by a high tem-
perature under substrate saturation (Tucker et al. 2012). Similar to
our observations, Hartley et al. (2007) demonstrated that microbial
respiration from warmed plots was higher than those from control
plots, when measurements were made at a common temperature;
and any positive microbial response to warming can be maintained
through cold winter season. Balser & Wixon (2009) found that the
temperature sensitivity of microbial respiration increased with site
mean annual temperature in a latitudinal gradient study. Our study
demonstrates temperature sensitivity between 25 and 35 °C. Further
assessment is needed to extend these findings to a wider range of
temperatures.
Combined effects of eCO
2
and warming
A positive effect of combined eCO
2
and warming on cumulative
respiration efficiency was observed in this study (Table 1). The
combined mechanisms of increased respiration efficiency with eCO
2
and increased Q
10
with warming may potentially interact to amplify
individual effects on soil communities and their feedback to climate
change. In addition, the CT microbial communities had the highest
Q
10
among all treatments (Fig. 1b). It seems that combined effects
of eCO
2
and warming on microbial activities will produce positive
feedbacks with non-additive effects, which could create a positive
feedback on C loss from soils.
Shifts in turnover time of the microbial biomass or possible
changes of composition and functioning of the microbial community
may cause higher rates of SOM decomposition estimated from the
respiration rate on a per-unit-PLFA basis of Ct and CT communities
(Fig. 2b) (Kandeler et al. 2008). The CCA biplot supports that eCO
2
and warming led to microbial community structural and functional
shifts of inoculated soils (Fig. 3). Microbial communities from Ct
and CT plots were more similar to one another than to other com-
munities, and had higher activity than those from ct and cT plots
(Fig. 3). Therefore, our results provide support for a recent meta-
analysis (Dieleman et al. 2012), which suggests that soil process
responses to combined eCO
2
and warming are more similar to those
in individual eCO
2
treatment than individual warming treatment.
CONCLUSIONS
This research demonstrated that microbial community-level responses
to climate change affected the ecosystem function of SOM decompo-
sition. Our results highlighted two major mechanisms by which the
microbial community can mediate terrestrial C cycling feedback to cli-
mate change. First, microbial ability to decompose SOM increased
with elevated CO
2
at ambient and warmed temperatures due to
increased substrate respiration rates or/and enzyme activities. Second,
microbial temperature sensitivity increased in response to warming at
ambient and elevated CO
2
, suggesting a positive feedback between
microbial activities related to SOM decomposition and climate warm-
ing. Overall, our study indicates that positive microbial community
feedbacks in response to eCO
2
and warming can accelerate microbial
decomposition and potentially lead to soil C losses.
ACKNOWLEDGEMENTS
We thank Dan LeCain for overall project management and Jack
Morgan for leading the PHACE experiment. We appreciate support
from the University of Wyoming Biodiversity Institute. This
research was supported by the US Department of Agriculture, US
Department of Energy’s Office of Science (BER) and by the
National Science Foundation (DEB# 1021559).
AUTHORSHIP
MN, EP and MW designed the study, MN, EP, CB, CG, SR and
ST performed the experiment, MN, EP, CB and MW analysed the
data and MN, EP, CB and MW wrote the paper.
REFERENCES
Allison, S.D., Wallenstein, M.D. & Bradford, M.A. (2010). Soil-carbon response
to warming dependent on microbial physiology. Nat. Geosci., 3, 336–340.
Austin, E.E., Castro, H.F., Sides, K.E., Schadt, C.W. & Classen, A.T. (2009).
Assessment of 10 years of CO
2
fumigation on soil microbial communities and
function in a sweetgum plantation. Soil Biol. Biochem., 41, 514–520.
Balser, T.C. & Wixon, D.L. (2009). Investigating biological control over soil
carbon temperature sensitivity. Glob. Change Biol., 15, 2935–2949.
Bardgett, R.D., Freeman, C. & Ostle, N.J. (2008). Microbial contributions to
climate change through carbon cycle feedbacks. ISME J., 2, 805–814.
Belay-Tedla, A., Zhou, X.H., Su, B., Wan, S.Q. & Luo, Y.Q. (2009). Labile,
recalcitrant, and microbial carbon and nitrogen pools of a tallgrass prairie soil
in the US Great Plains subjected to experimental warming and clipping. Soil
Biol. Biochem., 41, 110–116.
Bradford, M.A., Davies, C.A., Frey, S.D., Maddox, T.R., Melillo, J.M., Mohan, J.
E. et al. (2008). Thermal adaptation of soil microbial respiration to elevated
temperature. Ecol. Lett., 11, 1316–1327.
Bradford, M.A., Watts, B.W. & Davies, C.A. (2010). Thermal adaptation of
heterotrophic soil respiration in laboratory microcosms. Glob. Change Biol., 16,
1576–1588.
Campbell, C.D., Chapman, S.J., Cameron, C.M., Davidson, M.S. & Potts, J.M.
(2003). A rapid microtiter plate method to measure carbon dioxide evolved
from carbon substrate amendments so as to determine the physiological
profiles of soil microbial communities by using whole soil. Appl. Environ.
Microbiol., 69, 3593–3599.
©2012 Blackwell Publishing Ltd/CNRS
Letter Microbial feedbacks to climate change 7

Carney, K.M., Hungate, B.A., Drake, B.G. & Megonigal, J.P. (2007). Altered soil
microbial community at elevated CO
2
leads to loss of soil carbon. Proc. Natl
Acad. Sci. USA., 104, 4990–4995.
Carrillo, Y., Pendall, E., Dijkstra, F.A., Morgan, J.A. & Newcomb, J.M. (2010).
Carbon input control over soil organic matter dynamics in a temperate grassland
exposed to elevated CO
2
and warming. Biogeosciences Discuss., 7, 1575–1602.
Carrillo, Y., Pendall, E., Dijkstra, F.A., Morgan, J.A. & Newcomb, J.M. (2011).
Response of soil organic matter pools to elevated CO
2
and warming in a
semi-arid grassland. Plant Soil, 347, 339–350.
Carrillo, Y., Dijkstra, F., Pendall, E., Morgan, J. & Blumenthal, D. (2012).
Controls over soil nitrogen pools in a semiarid grassland under elevated CO
2
and warming. Ecosystems, 15, 761–774.
DeAngelis, K.M., Silver, W.L., Thompson, A.W. & Firestone, M.K. (2010).
Microbial communities acclimate to recurring changes in soil redox potential
status. Environ. Microbiol., 12, 3137–3149.
Dieleman, W.I., Vicca, S., Dijkstra, F.A., Hagedorn, F., Hovenden, M.J., Larsen,
K.S. et al. (2012). Simple additive effects are rare: a quantitative review of
plant biomass and soil process responses to combined manipulations of CO
2
and temperature. Glob. Change Biol., 18, 2681–2693.
Dijkstra, F.A., Blumenthal, D., Morgan, J.A., Pendall, E., Carrillo, Y. & Follett,
R.F. (2010). Contrasting effects of elevated CO
2
and warming on nitrogen
cycling in a semiarid grassland. New Phytol., 187, 426–437.
Dijkstra, F.A., Pendall, E., Morgan, J.A., Blumenthal, D., Carrillo, Y., LeCain, D.
et al. (2012). Climate change alters stoichiometry of phosphorus and nitrogen
in a semiarid grassland. New Phytol., 196, 807–815.
Evans, S.E. & Wallenstein, M.D. (2012). Soil microbial community response to
drying and rewetting stress: does historical precipitation regime matter?
Biogeochemistry, 109, 101–116.
Frey, S.D., Drijber, R., Smith, H. & Melillo, J. (2008). Microbial biomass,
functional capacity, and community structure after 12 years of soil warming.
Soil Biol. Biochem., 40, 2904–2907.
Hanson, C.A., Allison, S.D., Bradford, M.A., Wallenstein, M.D. & Treseder, K.
K. (2008). Fungal taxa target different carbon sources in forest soil. Ecosystems,
11, 1157–1167.
Hartley, I.P., Heinemeyer, A., Evans, S.P. & Ineson, P. (2007). The effect of soil
warming on bulk soil vs. rhizosphere respiration. Glob. Change Biol., 13, 2654–
2667.
Hartley, I.P., Hopkins, D.W., Garnett, M.H., Sommerkorn, M. & Wookey, P.A.
(2008). Soil microbial respiration in arctic soil does not acclimate to
temperature. Ecol. Lett., 11, 1092–1100.
He, Z.L., Xu, M.Y., Deng, Y., Kang, S.H., Kellogg, L., Wu, L.Y. et al. (2010).
Metagenomic analysis reveals a marked divergence in the structure of
belowground microbial communities at elevated CO
2
.Ecol. Lett., 13, 564–575.
Hu, S., Chapin, F.S., Firestone, M.K., Field, C.B. & Chiariello, N.R. (2001).
Nitrogen limitation of microbial decomposition in a grassland under elevated
CO
2
.Nature, 409, 188–191.
Jastrow, J.D., Miller, R.M. & Owensby, C.E. (2000). Long-term effects of
elevated atmospheric CO
2
on below-ground biomass and transformations to
soil organic matter in grassland. Plant Soil, 224, 85–97.
Kandeler, E., Mosier, A.R., Morgan, J.A., Milchunas, D.G., King, J.Y., Rudolph, S.
et al. (2008). Transient elevation of carbon dioxide modifies the microbial
community composition in a semi-arid grassland. Soil Biol. Biochem., 40, 162–171.
Kuzyakov, Y., Friedel, J.K. & Stahr, K. (2000). Review of mechanisms and
quantification of priming effects. Soil Biol. Biochem., 32, 1485–1498.
Lennon, J.T. & Jones, S.E. (2011). Microbial seed banks: the ecological and
evolutionary implications of dormancy. Nat. Rev. Microbiol., 9, 119–130.
Luo, Y.Q. (2007). Terrestrial carbon-cycle feedback to climate warming. Annu.
Rev. Ecol. Evol. Syst., 38, 683–712.
Luo, Y. & Zhou, X. (2006). Soil Respiration and the Environment. Elsevier, San
Diego.
Luo, Y.Q., Wan, S.Q., Hui, D.F. & Wallace, L.L. (2001). Acclimatization of soil
respiration to warming in a tall grass prairie. Nature, 413, 622–625.
Morgan, J.A., LeCain, D.R., Pendall, E., Blumenthal, D.M., Kimball, B.A.,
Carrillo, Y. et al. (2011). C
4
grasses prosper as carbon dioxide eliminates
desiccation in warmed semi-arid grassland. Nature, 476, 202–205.
Panikov, N.S. (2009). Microbial activity in frozen soils. In: Permafrost Soils (ed
Margesin, R.). Springer, Berlin, pp. 119–147.
Parton, W.J., Morgan, J.A., Wang, G.M. & Del Grosso, S. (2007). Projected
ecosystem impact of the prairie heating and CO
2
enrichment experiment. New
Phytol., 174, 823–834.
Pendall, E., Bridgham, S., Hanson, P.J., Hungate, B., Kicklighter, D.W., Johnson,
D.W. et al. (2004). Below-ground process responses to elevated CO
2
and
temperature: a discussion of observations, measurement methods, and models.
New Phytol., 162, 311–322.
Phillips, R.P., Finzi, A.C. & Bernhardt, E.S. (2011). Enhanced root exudation
induces microbial feedbacks to N cycling in a pine forest under long-term
CO
2
fumigation. Ecol. Lett., 14, 187–194.
Sinsabaugh, R.L., Lauber, C.L., Weintraub, M.N., Ahmed, B., Allison, S.D.,
Crenshaw, C. et al. (2008). Stoichiometry of soil enzyme activity at global
scale. Ecol. Lett., 11, 1252–1264.
Steinweg, J.M., Plante, A.F., Conant, R.T., Paul, E.A. & Tanaka, D.L. (2008).
Patterns of substrate utilization during long-term incubations at different
temperatures. Soil Biol. Biochem., 40, 2722–2728.
Trevors, J.T. (1996). Sterilization and inhibition of microbial activity in soil.
J. Microbiol. Methods., 26, 53–59.
Tucker, C.L., Bell, J., Pendall, E. & Ogle, K. (2012). Does declining carbon-use
efficiency explain thermal acclimation of soil respiration with warming? Glob.
Change Biol., DOI: 10.1111/gcb.12036.
van Veen, J.A., Liljeroth, E., Lekkerkerk, L.J.A. & Vandegeijn, S.C. (1991).
Carbon fluxes in plant-soil systems at elevated atmospheric CO
2
levels. Ecol.
Appl., 1, 175–181.
van de Voorde, T.F.J, van der Putten, W.H. & Bezemer, T.M. (2012). Soil
inoculation method determines the strength of plant–soil interactions. Soil Biol.
Biochem., 55, 1–6.
Wallenstein, M.D. & Hall, E.K. (2012). A trait-based framework for predicting
when and where microbial adaptation to climate change will affect ecosystem
functioning. Biogeochemistry, 109, 35–47.
Wallenstein, M.D., McMahon, S.K. & Schimel, J.P. (2009). Seasonal variation in
enzyme activities and temperature sensitivities in Arctic tundra soils. Glob.
Change Biol., 15, 1631–1639.
Wallenstein, M., Allison, S.D., Ernakovich, J., Steinweg, J.M. & Sinsabaugh, R.
(2011). Controls on the temperature sensitivity of soil enzymes: a key driver
of in situ enzyme activity rates. In: Soil Enzymology (eds Shukla, G.C. & Varma,
A.). Springer, Berlin, pp. 245–257.
Weber, C.F., Zak, D.R., Hungate, B.A., Jackson, R.B., Vilgalys, R., Evans, R.D.
et al. (2011). Responses of soil cellulolytic fungal communities to elevated
atmospheric CO
2
are complex and variable across five ecosystems. Environ.
Microbiol., 13, 2778–2793.
Yergeau, E., Bokhorst, S., Kang, S., Zhou, J.Z., Greer, C.W., Aerts, R. et al.
(2012). Shifts in soil microorganisms in response to warming are consistent
across a range of Antarctic environments. ISME J., 6, 692–702.
Zak, J.C., Willig, M.R., Moorhead, D.L. & Wildman, H.G. (1994). Functional
diversity of microbial communities: a quantitative approach. Soil Biol. Biochem.,
26, 1101–1108.
Zak, D.R., Pregitzer, K.S., Curtis, P.S. & Holmes, W.E. (2000). Atmospheric
CO
2
and the composition and function of soil microbial communities. Ecol.
Appl., 10, 47–59.
Zhou, J., Xue, K., Xie, J., Deng, Y., Liyou, W., Xiaoli, C. et al. (2012). Microbial
mediation of carbon-cycle feedbacks to climate warming. Nat. Clim. Chang.,2,
106–110.
SUPPORTING INFORMATION
Additional Supporting Information may be downloaded via the online
version of this article at Wiley Online Library (www.ecologyletters.com).
Editor, Peter Thrall
Manuscript received 2 August 2012
First decision made 5 September 2012
Second decision made 10 October 2012
Manuscript accepted 19 October 2012
©2012 Blackwell Publishing Ltd/CNRS
8M. Nie et al. Letter