Geoderma 214–215 (2014) 19–24

Contents lists available at ScienceDirect

Geoderma

j ourna l homepage: www.e lsev ie r .com/ locate /geoderma

Exogenous nutrient manipulations alter endogenous extractability ofcarbohydrates in decomposing foliar litters under a typical mixedforest of subtropics

Fu-Sheng Chen a,⁎, David S. Duncan b, Xiao-Fei Hu c, Chao Liang d

a College of Forestry, Jiangxi Agricultural University, Nanchang 330045, PR Chinab Department of Agronomy, University of Wisconsin, Madison, WI 53706, United Statesc College of Life Sciences, Nanchang University, Nanchang 330031, PR Chinad Department of Soil Science, University of Wisconsin, Madison, WI 53706, United States

⁎ Corresponding author at:No. 1101, ZhimindaRoad,NanDevelopment Area, Nanchang 330045, PR China. Tel./fax: +

E-mail address: chenfush@yahoo.com (F.-S. Chen).

0016-7061/$ – see front matter © 2013 Elsevier B.V. All rihttp://dx.doi.org/10.1016/j.geoderma.2013.10.009

a b s t r a c t

a r t i c l e i n f o

Article history:Received 26 July 2013Received in revised form 6 October 2013Accepted 8 October 2013Available online 25 October 2013

Keywords:Carbohydrate extractabilityGlobal environmental changeLitter decompositionNutrient dynamicsSubtropical forest

Litter decomposition is a major driver of nutrient availability and carbon (C) cycling in forest ecosystems, andmay be susceptible to perturbation by exogenous nutrient inputs from anthropogenic activity. The role ofresource limitation in this process is unclear, although some models suggest the macronutrients nitrogen (N)and phosphorus (P) stimulate early-stage decomposition while availability of labile C restricts later-stagedecomposition. We studied this interplay in a subtropical mixed forest in southern China through in situincubations of litter samples over a 540-day period. Litter sampleswere amendedwith labile C (+C, as sucrose),N (+N, as NH4NO3), P (+P, as NaH2PO4), all three inputs (+CNP), or no inputs (CK). Litter mass and nutrientcontent were measured at 90-day intervals, while extractable carbohydrate profiles of the litter samples wereassayed at 90, 270, and 450days of incubation. The+P and+CNP treatments showed greater reductions in littermass, as well as a larger P pool throughout the incubation. The +N treatment had a larger N pool, but otherwisedid not differ from the no input control CK. The concentration of accessible carbohydrate fractions remainedconstant or increased from 90 to 450 days of incubation, while less accessible fractions thought to belignocellulose and hemi-cellulose decreased during that period. Total extractable carbohydrates decreased inthe +P treatment, but was not significantly different among other treatments. Our results suggest that mixedforest systems in southern China are likely to beminimally perturbed in the short term by exogenous N addition,and that decomposition activity is not regulated by labile C availability over the time period studied. Moreover,increased C inputs due to climate change-induced changes in litter deposition and root exudation will likelyhave a smaller impact on subtropical forestmanagement than anthropogenic disturbances such as P fertilization.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

Forests are a key element of global biogeochemical cycles, func-tioning as major terrestrial carbon (C) reservoirs (Pan et al., 2011) andplaying a critical role in many nutrient cycles (Attiwill and Adams,1993). Litter decomposition is one of the key mechanisms governingnutrient availability in many forest systems (Aber and Melillo, 2001)and helps determine long-term C balances (Berg and McClaugherty,2007). Systems such as the subtropical mixed forests in southern andeastern China face growing human activity (Bruelheide et al., 2011),creating a potential for substantial alteration of nutrient cycling anddynamics. Nitrogen (N) and phosphorus (P) in particular have becomemuch more available in many environments due to anthropogenicinfluences on their biogeochemical cycles (Falkowski et al., 2000;Vitousek et al., 1997). Improving our understanding of the processesthat drive litter decomposition dynamics will augment our knowledge

chang Economic&Technological86 791 83813243.

ghts reserved.

of the regulation of forest ecosystem function (Drewnik, 2006), and willbe crucial for informing responses to future environmental perturbations(Cornwell et al., 2008; Knorr et al., 2005).

One of the key dynamics governing the fate of carbohydrates duringdecomposition is whether decomposers are limited by nutrients orenergy. Decomposer biomass requires a relatively strict stoichiometryamong C, N, and P, and the availability of specific nutrients can alterboth the rate of decomposition and eventual composition of residualbiomass (Aber andMelillo, 2001; Cherif and Loreau, 2007). For instance,changes to N availability in N-limited systems can influence decom-position of recalcitrant C, even if more accessible energy sources arepresent (Craine et al., 2007). These effects may be highly variable,however. Various studies have found the effects of N and P availabilityon litter decomposition to be neutral (Bridgham and Richardson,1992; Coulson and Butterfield, 1978; Prescott, 1995), positive (Bergand Tamm, 1994; Hobbie, 2000), negative (Mcclaugherty and Berg,1987; Mo et al., 2008), or specific to sites or species (Chen et al.,2012). Other studies have found the availability of energy, rather thannutrients, primarily limits decomposer activity (Hattenschwiler andJorgensen, 2010), and that decomposer communities respond to the

Table 1Nutrient addition treatments and their associated decay constants.

Treatment Nutrient level Decay constant (k)

Control (CK) −9.4 × 10−4 A aCarbon (+C) 75 g-C m−2 year−1 −9.7 × 10−4 AB(as glucose) 150 g-C m−2 year−1 −10.3× 10−4 AB

Combined −10.0× 10−4 abNitrogen (+N) 7.5 g-Nm−2 year−1 −9.9 × 10−4 AB(as NH4NO3) 15 g-Nm−2 year−1 −10.8× 10−4 AB

Combined −10.4× 10−4 abPhosphorus (+P) 3.75 g-P m−2 year−1 −10.3× 10−4 AB(as NaH2PO4) 7.5 g-P m−2 year−1 −11.2× 10−4 AB

Combined −10.7× 10−4 bCombined nutrients Low −11.4× 10−4 B(+CNP) High −11.5× 10−4 B

Combined −11.5× 10−4 b

Treatments sharing a letter did not have significantly different decay constants (P b 0.05,based on Tukey's multiple comparison correction). Uppercase letters correspond totreatments distinguishing among nutrient levels, lowercase letters correspond tocombined addition levels. Low and High levels for +CNP correspond to the low andhigh levels for individual nutrients.

20 F.-S. Chen et al. / Geoderma 214–215 (2014) 19–24

identity and accessibility of plant-derived C sources (Fontaine et al.,2007; Orwin et al., 2006). Manipulating the relative availabilities ofkey nutrients and energy sources can reveal system-specific nutrientcontrols on decomposition (Krashevska et al., 2010).

Soluble litter carbohydrates are the primary source of energy fordecomposers, making the quality of litter composition a potential deter-minant of energy availability. Efforts to characterize litter compositionquality, as defined by its influence on decomposition processes, arehampered by complicated measurement and multifarious features(Coq et al., 2010; McLeod et al., 2007). One challenge may be thattypical quantifications of litter components such as lignin and carbo-hydrate polymers do not capture the properties of litter that are mostimportant for decomposition (Newsham et al., 2001). One potentiallyeffective alternative approach is the sequential extraction of structuralcomponents, which characterizes litter composition based on thecross-linkage of various constituents of the cell wall (Fry, 1986;Selvendran and Oneill, 1987). This method is effective at quantifyingthe relative availability of accessible energy-rich carbon compounds,and can reflect decomposition-relevant characteristics of decaying litter(McLeod et al., 2007). By tracking the abundance of functionally-defined carbohydrate fractions under a range of conditions, it may bepossible to improve our understanding of decomposition constraintsand dynamics in forest ecosystems.

In this study, we explored the effects of exogenous addition ofdifferent stoichiometries of labile C, N, and P on the in situ decompositionof leaf litter in the context of a subtropical mixed forest in China. Weemployed sequential extraction of plant structural components to trackthe changing composition of decomposing litter over a 540-day period.The aims of our studywere to: 1) evaluate the effect of stoichiometricallybalanced and unbalanced exogenous nutrient addition on litter decom-position rates and nutrient status; 2) track changes in extractablecarbohydrate fractions during decomposition; and 3) evaluate theeffects of exogenously applied nutrients on changes litter carbohydratecomposition during decomposition. Our study improves our under-standing of litter decomposition dynamics in an important andecologically sensitive ecosystem, while also demonstrating the utilityof an established but little-used method for functionally characterizingdecomposing litter.

2. Materials and methods

2.1. Site description

Our study was conducted in Xinjian County (115°27′–116°35′E,28°09–29°11′N), Jiangxi Province in southern China, described in detailin our recent publication (Chen et al., 2012). The area has a subtropicalwarm humid climate, with an annual mean air temperature of 17.5 °C,and mean annual precipitation of 1700 mm. Over 60% of the studyarea is covered by forest vegetation, with Pinus massoniana andLiquidambar formosana as the dominant coniferous and deciduousspecies respectively. Ultisols are the dominant soil order, with pH 4.4,13.5 g organic C kg−1 soil, 1.15 g total N kg−1 soil, and 0.25 g total Pkg−1 soil, on average (Zhan et al., 2009).

2.2. Experimental design and treatments

To track decomposition over time, we employed the litter nylonnet bag method (Berg and Staaf, 1980; Chen et al., 2012). Briefly, wecollected P. massoniana and L. formosana foliar litter from three20× 20m plots, composited equal dry weights from both species, andplaced 5 g of the litter mixture in 15 × 15 cm nylon bags with a 1mmmesh. In each plot from which we sampled the litter, we arbitrarilyplaced 9 subplots measuring 2 × 2 m and separated by N5 m. In eachsubplot, we placed 12 bags on the soil surface below the litter layer,for 324 bags in total. Nutrient addition was carried out at the subplotlevel, with each subplot receiving one of nine treatments including: 75

or 150 g sucrose-C m−2 year−1, 7.5 or 15 g NH4NO3-N m−2 year−1,3.75 or 7.50 g NaH2PO4-P m−2 year−1, all three nutrients together attheir respective high or low levels, and a no nutrient control (Table 1).All nutrients were delivered as an aqueous solution at 30-day intervals,with the control receiving an equal volume of water. The lower level ofN addition was matched to observed rates of N deposition in south-central China (Lü andTian, 2007). Carbon and P rates selected to achievethe C:N and N:P ratios of 10 and 2 commonly found in soils (Zhan et al.,2009), so that treatments receiving all nutrients would be stoichio-metrically balanced with standard soil conditions.

2.3. Sampling and chemical analysis

At 90-day intervals for 6 time points, 2 bags were sampled fromeach subplot, cleaned of roots and attached soil, washed with deionizedwater, dried at 45°C, thenweighed, composited and ground to 0.25mmfor chemical analysis. Litter organic C content was determined bythe Walkley–Black wet oxidation method following acid removal ofcarbonates (Allen, 1989). Nitrogen and phosphorus concentrationswere determined by the Kjeldahl method and phosphomolybdic acidblue color method after the samples were digested with 18.4M H2SO4

solution (Allen, 1989).We quantified extractable carbohydrate fractions at days 90,

270, and 450 using a seven-step serial fractionation that classifiedcarbohydrates based on their linkages within cell wall architecture(McLeod et al., 2007): (1) free sugars and polysaccharides wereextracted with 10% (v/v) formic acid, which also denatured enzymesand scavenged hydroxyl radicals; (2) weakly-bound polysaccharidesand pectins were extracted with phosphate buffer (200mM NaH2PO4,0.5% w/v chlorobutanol, 10mM Na2S2O3, final pH adjusted to 7.0 withNaOH); (3) strongly-bound polysaccharides and pectinswere extractedby removal of calciumand solubilization byCDTA (50mMw/v trans-1,2-diaminocyclohexane-N,N,N′, N′-tetraacetic acid, 0.5%w/v chlorobutanolat pH7.5); (4) bonds between hemicellulose and cellulose were brokenby extraction with urea (8 M H2NCONH2, 50 mM N-(2-hydroxyethyl)piperazine-N′-(2-ethanesulfonic acid at pH7.5) leading to extraction ofsome hemicelluloses; (5) otherwise inaccessible sugars were extractedby low temperature ester hydrolysis with mildly alkaline sodiumcarbonate (200 mM Na2CO3, at 5 °C); (6) remaining hemicelluloseswere extracted with strongly alkaline sodium hydroxide with NaBH4

to protect carbohydrates from oxidation (6 M NaOH, 1% w/v NaBH4

at 37 °C); (7) lignocellulose released by removal of previous fractionswas solubilized in 5% (v/v) formic acid. Carbohydrates remainingafter the seventh extraction were assumed to be structural cellulose(Fry, 2000). Extractions consisted of 3 replicates of 0.3 g ground litter

21F.-S. Chen et al. / Geoderma 214–215 (2014) 19–24

material placed into 50ml centrifuge tubes, with solvents added in 6mlaliquots and incubated for 24h, followed by removal of extractant usingcentrifugation and addition of the next solvent in the sequence. Totalsoluble and insoluble carbohydrates were measured with a phenol/sulfuric acid assay (Fry, 1986). Samples were quantified by 485 nmabsorbance calibrated with dextran standards following sonication todissipate CO2 bubbles.

2.4. Statistical analysis

Statistical analysis of mass and nutrient loss was carried out usingthe nlme package (v3.1, Pinheiro et al., 2012) in the R statisticalenvironment (v2.14.2, R Core Development Team, 2012). In all models,variance differences among treatment groups (weights = varIdent)were compared to a model assuming homogenous variance by log-likelihood ratio testing with a significance cutoff of Pb0.05. Litter massas a proportion of starting mass was modeled as an exponential decayprocess where the fraction remaining was a function of incubationtime and a decay constant (k). Nitrogen and P contents were analyzedas proportions of their values in the original litter mixture andmodeledas second-order polynomials of decomposition time. The polynomialmodels were allowed to set an intercept (i.e. initial nutrient content),rather than constrained to the known content of the starting litter toaccommodate the possibility of rapid leaching in the early stages ofthe incubation. Carbohydrate fractions were analyzed independentlyof each other, with incubation time treated as a factor. Significancetests were corrected using the Tukey–Kramer method in the lsmeanspackage (v1.06, Lenth, 2013).

Fig. 1. Change in litter mass and nutrient pools remaining during decomposition. A) Dry mass,line indicates quadratic fit calculated for all samples due to lack of differences in regression coenutrient treatment. Modeled lines are extrapolated to day 0. Treatments are additions of glucoscontrol (CK), combining high and low levels of nutrient addition.

3. Results

Modeling separate litter mass decomposition decay constants (k)for high and low nutrient addition levels did not improve the modelfit significantly over estimating k values for combined levels of nutrienttreatments based on log-likelihood ratios (4 d.f. difference, ratio=5.56,P= 0.22). We found that litter without exogenous nutrient additions(CK) had a significantly lower decay rate than litter receiving eitherP alone (+P) or in conjunction with other nutrients (+CNP) (Table 1,Fig. 1A).

The second order polynomial model fit to remaining litter N contentwas not improved by estimating separate regression coefficients forthe nutrient treatments (4 d.f. difference, ratio = 6.77, P = 0.15).The model indicated a slight increase in N content in the litter duringthe first two sampling periods, even as remaining mass had visiblydecreased (Fig. 1B).

The second order polynomial fit to remaining litter P content wassignificantly improved by estimating separate linear and quadraticcoefficients for nutrient addition combinations, but not by separatingthese into high and low levels of addition. In addition, intercept termswere not significantly different among nutrient treatments, althoughthe modeled intercept term implied the starting phosphorus contentwas only 70% of the known initial content (Fig. 1C). For all nutrienttreatments, P content increased initially in the early stages ofincubation, then decreased in the later stages. Both the +P and +CNPtreatments differed significantly from all other treatments in their linearand quadratic coefficients, demonstrating a greater initial accumulationof P and a faster subsequent decline (Fig. 1C).

lines indicate exponential fits calculated for each nutrient treatment. B) Nitrogen, doublefficients among treatments. C) Phosphorus, lines indicate quadratic fits calculated for eache (+C), nitrogen (+N), phosphate (+P), all three nutrients (+CNP), and a no-treatment

Fig. 2. Relative concentrations of carbohydrates recovered by sequential fractionation from decomposing foliar litter. Panels A–G depict mean concentrations of seven sequential fractionsin drymass (note different axis scales). Panel H depicts the average concentration summed over all fractions. Bar textures indicate the following nutrient addition treatments: no nutrientcontrol (CK, solid), low and high labile carbon supplied as glucose (+C, horizontal bars), low and high inorganic nitrogen (+N, lines sloping left), low and high inorganic phosphorus (+P,lines sloping right), and low and high levels of all three nutrients (+CNP, brick pattern). Groups sharing a letter within an incubation time were not significantly different (equivalent toP N 0.05 using Tukey's multiple comparison correction), when no letters are present no significant differences existed among nutrient treatments. Letters above horizontal bars indicatesignificant differences among incubation periods. Error bars indicate +1 standard error.

22 F.-S. Chen et al. / Geoderma 214–215 (2014) 19–24

23F.-S. Chen et al. / Geoderma 214–215 (2014) 19–24

The fraction of total litter biomass that was extractable increasedfrom day 90 to day 270, then decreased from day 270 to day 450(Fig. 2A). All fractions we analyzed changed significantly during theincubation period. Fractions 1, 2, 3 and 5 broadly became moreabundant after day 90 (Fig. 2, Table 2). In contrast, fractions 4 and 7generally became less abundant during incubation. Fraction 6 wasboth the largest and the most temporally dynamic, growing from58.5 g kg−1 in day 90 to 71.1 g kg−1 in day 270 then dropping to45.2 gkg−1 by day 450 (Fig. 2G).

In contrast to the strong effect of incubation time on carbohydratefraction abundances, the effects of nutrient addition treatments wererelatively minor. The effect of nutrient addition was distinct for eachfraction with few generally discernible patterns (Fig. 2). Fractions 2, 3,and 6 responded to differences in nutrient addition levels (Table 2).Fractions 2 and 6 were the only individual fractions to have an inter-action between nutrient treatments and incubation time (Table 2).Both of these fractions were reduced in the low level of the +Ptreatment, but not the higher level of this treatment (Fig. 2D, G).

The concentrations of the first five carbohydrate fractions weresignificantly and negatively correlated to litter P concentration, butonly at day 450 (Table S1). These fractions were positively correlatedwith C/N, C/P, and N/P ratios at day 450, although not all correlationswere statistically significant (Table S1).

4. Discussion

We found that exogenous P addition, whether alone or in stoichio-metric balance with labile C and N, was the only nutrient treatmentthat significantly impacted foliar litter decomposition rates relative toa control receiving no additional nutrients. Litter receiving exogenousP accumulated at least as much P as was present in the initial litterby day 180. However, nutrient treatments that received P achieveda similar accumulation, if on a much smaller scale. Given the decom-position and nutrient dynamic responses, it is evident that our systemis P limited, which is in line with previous studies of comparable mixedforests in China (Zhan et al., 2009). It is unclear how all treatmentswere able to increase their P content. The mechanisms and scope ofenvironmental P addition may be a fruitful area for further research.

Contrary to previous studies on comparable systems, we found nosignificant effects of N addition, either alone or in combination with P.Mo et al. (2008) found that N addition inhibited mass loss, althoughtheir study considered only P. massoniana foliar litter rather than the

Table 2Extractable carbohydrate fraction interpretations and the effects of exogenous nutrientaddition and incubation time.

Extraction step Interpretation Two-way ANOVA F-values

Nutrientaddition

Incubationtime

Interaction

1. Formic acid (10%) Free sugars 1.60 n.s. 4.59⁎ 1.39 n.s.2. Phosphate buffer Weakly bound pectins/

polysaccharides2.56⁎ 25.80⁎⁎⁎ 1.70 n.s.

3. CDTA Strongly bound pectins/polysaccharides

7.17⁎⁎⁎ 38.02⁎⁎⁎ 2.64⁎⁎

4. Urea Weakly boundhemicellulose

0.68 n.s. 40.15⁎⁎⁎ 1.12 n.s.

5. Sodium carbonate Inaccessible sugars 5.25⁎⁎ 13.66⁎⁎⁎ 1.51 n.s.6. Sodium hydroxide Strongly bound

hemicellulose2.93⁎⁎ 147.78⁎⁎ 3.16⁎⁎⁎

7. Formic acid (5%) Lignocellulose 0.23 n.s. 19.38⁎⁎⁎ 0.97 n.s.Total 2.06 n.s. 86.91⁎⁎⁎ 2.4⁎⁎

Note: Different variances were modeled for each time period for these fractions. All otherfractions had no significant departures from equal variance based onmodel log-likelihoodcomparisons (P N 0.05). Carbohydrate fraction interpretations derived from Fahey et al.(2011) and personal communication from Joseph Yavitt. n.s., P N 0.05.⁎ P b 0.05.⁎⁎ P b 0.01.⁎⁎⁎ P b 0.001.

mixed species litter we employed. Fanin et al. (2012) reported thateven though their system was P limited, there was still an observableeffect from the combination of N and P addition. The lack of N-relatedeffects, as well as the slight accumulation of N observed acrosstreatments, may be due to the high rates of environmental N depositionin southern China (Lü and Tian, 2007). Our low and high N treatmentseffectively doubled and tripled estimated rates of atmospheric Ndeposition, suggesting that decomposition processes in this systemmay not be ecologically sensitive to increased N inputs on this scale.At the ecosystem level, however, N inputsmay influence decompositionby the availability of accessible C sources that previous work hasidentified as a key determinant of decomposition rates and dynamics(Hattenschwiler and Jorgensen, 2010). Based on work in other systems(Aber and Melillo, 2001; Berg and Matzner, 1997), we expected tosee changes in the chemical accessibility of carbohydrate fractionsduring the course of incubation. In particular, based on the nitrogenmining model put forth by Craine et al. (2007), we anticipated moreaccessible carbohydrate fractions would decrease resulting in anincrease in the relative amounts of more recalcitrant fractions. Contraryto our expectations, we found a broad increase from day 90 to day 450in fractions 1–3 and 5. These fractions have been interpreted as freesugars, weakly- and strongly-bound pectins and polysaccharides, andinaccessible sugars (Fahey et al., 2011). By contrast, the fractions asso-ciated with hemicelluloses and lignocelluloses had broadly decreasedby day 450. The concentrations we found for fractions 2–7 were verysimilar to values reported by McLeod et al. (2007), but fraction 1concentrations, and consequently the concentrations of extractablecarbohydrates, were approximately 100 mg g−1 higher than in ourstudy. Meanwhile, the initial N and P concentrations (11.05 mg g−1

and 0.41 mg g−1, respectively) of the mixed litter were generallylower than those of Quercus robur leaf litter studied by McLeod et al.(2007). Given that this earlier study considered fresh litter, it seemslikely that much of this difference could be due to leaching or otherprocesses that occurred early during decomposition.

Extracellular enzymes can cleave bonds, making structural carbo-hydrates soluble (Colpaert and vanTichelen, 1996), so it is plausiblethat the relative increase of fractions 2 and 3 may be due to thebreakdown of fraction 4 hemicelluloses due to microbial activity.Determining the precise chemical composition of these carbohydratefractions, and whether this changes during decomposition, couldprove to be a fruitful area of research in the future.

Despite the clear effects of nutrient, specifically P, addition ondecom-position rates and nutrient dynamics, nutrient addition treatmentsgenerated few systematic changes to carbohydrate concentrationprofiles. The most consistent trend was the reduced concentration ofmost carbohydrate fractions in the low-level P addition treatment (+PLow) at day 450 and to a lesser extent at day270. This reduction indicatesa greater proportion of litter biomass in this treatment is comprisedof unextractable structural cellulose. We hypothesize that this level ofP addition increases resources enough to stimulate microbial activity(Allison and Vitousek, 2004), while not removing the need for microbesto degrade more stable carbohydrates to meet their resource needs.Higher P addition levels, by contrast, may entirely satisfy P needs anddisincentivize degradation of less chemically accessible carbohydrates.

5. Conclusions

Litter decomposition is a major driver of nutrient availability and Ccycling in forest ecosystems. We found that litter decomposition wasinfluenced by exogenous additions of P, but not by additions of Nabove current rates of atmospheric deposition. This suggests litterdecomposition processes may not be directly impacted by increasedrates of N deposition. Contrary to our expectations, more chemicallyaccessible carbohydrate fractions tended to become relatively moreabundant duringdecomposition,while less accessible fractions declinedin abundance. A better understanding of the chemical composition of

24 F.-S. Chen et al. / Geoderma 214–215 (2014) 19–24

extractable carbohydrate fractionswould improve our understanding ofdecomposition processes and dynamics.

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.geoderma.2013.10.009.

Acknowledgments

This study was supported by grants from the National BasicResearch Program of China (973 Program, 2012CB416903) and theNational Natural Science Foundation of China (31160107 & 30960311)and Jiangxi Provincial Department of Science and Technology(20122BCB23005). We greatly appreciate Xiao-Jing Gong, Xue Fengand Wen Ren for their help in the sampling and analyses, and thankDr. Joseph Yavitt for his patience in explaining the interpretation ofcarbohydrate extractability fractions.

References

Aber, J.D., Melillo, M.M., 2001. Terrestrial Ecosystems. Academic Press, San Diego.Allen, S.E., 1989. Chemical Analysis of EcologicalMaterials. Blackwell Scientific Publications,

Oxford.Allison, S.D., Vitousek, P.M., 2004. Rapid nutrient cycling in leaf litter from invasive plants

in Hawaii. Oecologia 141, 612–619.Attiwill, P.M., Adams, M.A., 1993. Tansley review No. 50. Nutrient cycling in forests. New

Phytol. 561–582.Berg, B., Matzner, E., 1997. Effect of N deposition on decomposition of plant litter and soil

organic matter in forest systems. Environ. Rev. 5, 1–25.Berg, B., McClaugherty, C., 2007. Plant Litter: Decomposition, Humus formation, Carbon

Sequestration. Springer Verlag, New York.Berg, B., Staaf, H., 1980. Decomposition rate and chemical changes of Scots pine needle

litter. II. Influence of chemical composition. Ecol. Bull. 32, 373–390.Berg, B., Tamm, C.O., 1994. Decomposition and nutrient dynamics of litter in long-term

optimum nutrition experiments.2. Nutrient concentrations in decomposing Picea-Abies needle litter. Scand. J. For. Res. 9, 99–105.

Bridgham, S.D., Richardson, C.J., 1992. Mechanisms controlling soil respiration (CO2 andCH4) in southern peatlands. Soil Biol. Biochem. 24, 1089–1099.

Bruelheide, H., Böhnke,M., Both, S., Fang, T., Assmann, T., Baruffol, M., Bauhus, J., Buscot, F.,Chen, X.Y., Ding, B.Y., 2011. Community assembly during secondary forest successionin a Chinese subtropical forest. Ecol. Monogr. 81, 25–41.

Chen, F.S., Feng, X., Liang, C., 2012. Endogenous versus exogenous nutrient affects C, N,and P dynamics in decomposing litters in mid-subtropical forests of China. Ecol.Res. 27, 923–932.

Cherif, M., Loreau, M., 2007. Stoichiometric constraints on resource use, competitiveinteractions, and elemental cycling inmicrobial decomposers. Am. Nat. 169, 709–724.

Colpaert, J.V., vanTichelen, K.K., 1996. Decomposition, nitrogen and phosphorusmineralization from beech leaf litter colonized by ectomycorrhizal or litter-decomposing basidiomycetes. New Phytol. 134, 123–132.

Coq, S., Souquet, J.-M., Meudec, E., Cheynier, V., Hattenschwiler, S., 2010. Interspecificvariation in leaf litter tannins drives decomposition in a tropical rain forest of FrenchGuiana. Ecology 91, 2080–2091.

Cornwell, W.K., Cornelissen, J.H.C., Amatangelo, K., Dorrepaal, E., Eviner, V.T., Godoy,O., Hobbie, S.E., Hoorens, B., Kurokawa, H., Perez-Harguindeguy, N., Quested, H.M.,Santiago, L.S., Wardle, D.A., Wright, I.J., Aerts, R., Allison, S.D., van Bodegom, P.,Brovkin, V., Chatain, A., Callaghan, T.V., Diaz, S., Garnier, E., Gurvich, D.E., Kazakou, E.,Klein, J.A., Read, J., Reich, P.B., Soudzilovskaia, N.A., Vaieretti, M.V., Westoby, M.,2008. Plant species traits are the predominant control on litter decomposition rateswithin biomes worldwide. Ecol. Lett. 11, 1065–1071.

Coulson, J., Butterfield, J., 1978. An investigation of the biotic factors determining the ratesof plant decomposition on blanket bog. J. Ecol. 66, 631–650.

Craine, J.M., Morrow, C., Fierer, N., 2007. Microbial nitrogen limitation increasesdecomposition. Ecology 88, 2105–2113.

Drewnik, M., 2006. The effect of environmental conditions on the decomposition rate ofcellulose in mountain soils. Geoderma 132, 116–130.

Fahey, T.J., Yavitt, J.B., Sherman, R.E., Groffman, P.M., Fisk, M.C., Maerz, J.C., 2011. Transportof carbon and nitrogen between litter and soil organicmatter in a northern hardwoodforest. Ecosystems 14, 326–340.

Falkowski, P., Scholes, R.J., Boyle, E., Canadell, J., Canfield, D., Elser, J., Gruber, N., Hibbard,K., Hogberg, P., Linder, S., Mackenzie, F.T., Moore, B., Pedersen, T., Rosenthal, Y.,Seitzinger, S., Smetacek, V., Steffen, W., 2000. The global carbon cycle: a test of ourknowledge of earth as a system. Science 290, 291–296.

Fanin, N., Barantal, S., Fromin, N., Schimann, H., Schevin, P., Hattenschwiler, S., 2012.Distinct microbial limitations in litter and underlying soil revealed by carbon andnutrient fertilization in a tropical rainforest. PLoS ONE 7 (e49990-e49990).

Fontaine, S., Barot, S., Barre, P., Bdioui, N., Mary, B., Rumpel, C., 2007. Stability of organiccarbon in deep soil layers controlled by fresh carbon supply. Nature 450, 277-U210.

Fry, S.C., 1986. Cross-linking of matrix polymers in the growing cell-walls of angiosperms.Annu. Rev. Plant Physiol. 37, 165–186.

Fry, S.C., 2000. The Growing Plant Cell Wall: Chemical and Metabolic Analysis, Reprinted.The Blackburn Press, Caldwell, NJ (333 pp.).

Hattenschwiler, S., Jorgensen, H.B., 2010. Carbon quality rather than stoichiometrycontrols litter decomposition in a tropical rain forest. J. Ecol. 98, 754–763.

Hobbie, S.E., 2000. Interactions between litter lignin and soil nitrogen availability duringleaf litter decomposition in a Hawaiian Montane forest. Ecosystems 3, 484–494.

Knorr, M., Frey, S.D., Curtis, P.S., 2005. Nitrogen additions and litter decomposition:a meta-analysis. Ecology 86, 3252–3257.

Krashevska, V., Maraun, M., Ruess, L., Scheu, S., 2010. Carbon and nutrient limitation ofsoil microorganisms and microbial grazers in a tropical montane rain forest. Oikos119, 1020–1028.

Lenth, R.V., 2013. lsmeans: Least-squares means. R package version 1.06–05. http://CRAN.R-project.org/package=lsmeans.

Lü, C., Tian, H., 2007. Spatial and temporal patterns of nitrogen deposition in China:synthesis of observational data. J. Geophys. Res. 112, D22S05.

Mcclaugherty, C., Berg, B., 1987. Cellulose, lignin and nitrogen concentrations as rate regu-lating factors in late stages of forest litter decomposition. Pedobiologia 30, 101–112.

McLeod, A.R., Newsham, K.K., Fry, S.C., 2007. Elevated UV-B radiation modifies theextractability of carbohydrates from leaf litter of Quercus robur. Soil Biol. Biochem.39, 116–126.

Mo, J., Fang, H., Zhu, W., Zhou, G., Lu, X., Fang, Y., 2008. Decomposition responses of pine(Pinus massoniana) needles with two different nutrient-status to N deposition in atropical pine plantation in southern China. Ann. For. Sci. 65, 405. http://dx.doi.org/10.1051/forest:2008017.

Newsham, K.K., Anderson, J.M., Sparks, T.H., Splatt, P., Woods, C., McLeod, A.R., 2001. UV-Beffect on Quercus robur leaf litter decomposition persists over four years. Glob. Chang.Biol. 7, 479–483.

Orwin, K.H., Wardle, D.A., Greenfield, L.G., 2006. Ecological consequences of carbonsubstrate identity and diversity in a laboratory study. Ecology 87, 580–593.

Pan, Y., Birdsey, R.A., Fang, J., Houghton, R., Kauppi, P.E., Kurz, W.A., Phillips, O.L.,Shvidenko, A., Lewis, S.L., Canadell, J.G., 2011. A large and persistent carbon sink inthe world's forests. Science 333, 988–993.

Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D., The R Development Core Team, 2012. nlme:Linear and Nonlinear Mixed Effects Models. R Package Version 3. 1–103.

Prescott, C.E., 1995. Does nitrogen availability control rates of litter decomposition inforests. Plant Soil 168, 83–88.

RDevelopment Core Team, 2012. R: A Language andEnvironment for Statistical Computing.R Foundation for Statistical Computing, Vienna, Austria (ISBN 3-900051-07-0, URLhttp://www.R-project.org/).

Selvendran, R.R., Oneill, M.A., 1987. Isolation and analysis of cell walls from plantmaterial.Methods Biochem. Anal. 32, 25–153.

Vitousek, P.M., Aber, J.D., Howarth, R.W., Likens, G.E., Matson, P.A., Schindler, D.W.,Schlesinger, W.H., Tilman, D., 1997. Human alteration of the global nitrogen cycle:sources and consequences. Ecol. Appl. 7, 737–750.

Zhan, S.X., Chen, F.S., Hu, X.F., Gan, L., 2009. Nitrogen and phosphorus availability undertypical stages of forest succession at hilly red soil in middle subtropical region. ActaEcol. Sin. 29, 4673–4680.