Glycine mineralization in situ closely correlates with soilcarbon availability across six North American forestecosystems

Jack W. McFarland . Roger W. Ruess .Knut Kielland . Kurt Pregitzer . Ronald Hendrick

Abstract Free amino acids (FAA) constitute asignificant fraction of dissolved organic nitrogen (N)in forest soils and play an important role in the Ncycle of these ecosystems. However, comparativelylittle attention has been given to their role as labile

carbon (C) substrates that might influence themetabolic status of resident microbial populations.We hypothesized that the residence time of simpleC substrates, such as FAA, are mechanisticallylinked to the turnover of endogenous soil C pools.We tested this hypothesis across a latitudinalgradient of forested ecosystems that differ sharplywith regard to climate, overstory taxon, and edaphicproperties. Using a combined laboratory and fieldapproach, we compared the turnover of isotopicallylabeled glycine in situ to the turnover of mineral-izable soil C (Cmin) at each site. The turnover ofglycine was rapid (residence times <2 h) regardlessof soil type. However, across all ecosystems glycineturnover rates were strongly correlated with indicesof soil organic matter quality. For example, CNratios for the upper soil horizons explained ~ 80%of the variability observed in glycine turnover, andthere was a strong positive correlation between insitu glycine-C turnover and Cmin measured in thelaboratory. The turnover of glycine in situ wasbetter explained by changes in soil C availabilitythan cross-ecosystem variation in soil temperature orconcentrations of dissolved inorganic N and FAA-N.This suggests the consumption of these low-molec-ular-weight substrates by soil microorganisms maybe governed as much by the overall decomposabilityof soil C as by N limitation to microbial growth.

Keywords Soil free amino acid· Glycine · 13C .Soil C and N . Mineralizable C . Forest soils

Introduction

Free amino acids (FAA) represent a labile pool of soilnitrogen (N) for plant and microbial uptake and playa key role in the N economy of terrestrial ecosystems(Atkin 1996; Kaye and Hart 1997; Kielland 2001;Lipson and Nasholm 2001). This recognition has ledto a number of studies focusing on the turnoverdynamics of amino acids in soils spanning multiplecontinents (Kuzyakov and Demin 1998; Jones andShannon 1999; Lipson et al. 1999; Vinolas et al.2001a, b; Jones and Kielland 2002; Bardgett et al.2003; Jones et al. 2004; Finzi and Berthrong 2005;Kielland et al. 2007). These studies have yieldedinsights into the factors regulating the turnover ofFAA in soils. In most instances, the residence timesof soil amino acids are less than a few hours;however, they may persist for longer periods byadsorption to humic and mineral components of soil(Gonod et al. 2005), or chemical inclusion into humicsubstances (Kuzyakov and Galitsa 1993). Plants inmany ecosystems take up amino acids readily (Schi-mel and Chapin 1996; Schmidt and Stewart 1999;Nasholm et al. 1998; Nasholm et al. 2000), but themajority of FAA turnover is the direct result ofmicrobial uptake and assimilation (Jones 1999;Nordin et al. 2004). Once assimilated by microor-ganisms, amino acids can be channeled into growth,cell maintenance, or energy production (Kuzyakovand Demin 1998; Vinolas et al. 2001 a, b) ortransferred to plants via fungal symbionts.

While past research on FAA mineralization hasfocused on these compounds in the context of soil-Ncycling (Jones and Kielland 2002; Berthrong andFinzi 2006), the importance of FAA as a carbon (C)source for microorganisms has been given lessattention. There is increasing evidence that hetero-trophic growth in many soil environments is C- ratherthan N-limited (Morita 1988; Zak et al. 1994; DeNobli et al. 2001; Mondini et al. 2006; Treseder2008). For most terrestrial ecosystems, the bulk ofsoil C consists of complex polymers that are resistantto decomposition. Despite the low bioavailability of alarge percentage of soil organic matter, the microbialbiomass in many soils maintains a high level ofendogenous energy with adenyl ate charge ratiosapproaching those of microorganisms grown in pureculture studies (Brookes et al. 1987). The likelyreason for sustaining such a high metabolic rate in an

energy-poor environment stems from the need tocapitalize on pulses of labile substrate. The metab-olism of soil microorganisms depends heavily on theavailability of low-molecular-weight substrates suchas mono- and di-saccharides, peptides, and freeamino acids. A large proportion of this labile C poolderives from plant exudates and fine root turnoverthat serves to prime decomposition processes (Ku-zyakov et al. 2000). Activation of the microbialcommunity by pulses of labile C can be nearlyinstantaneous (Jones and Murphy 2007). However,evidence for this rapid response comes largely fromlaboratory incubation studies of processed soilswhere plant components are removed and microbialcommunity structure is severely disrupted. Samplehandling prior to laboratory measurements couldaffect biodegradation rates for FAA rendering thosedata inadequate for field simulations or modeling (Diet al. 1998).

Here we present a study using a nondestructivemethod (McFarland et al. 2002; Kielland et al. 2007)to assess the cycling dynamics of a soil amino acid insitu across a wide gradient of North American forestecosystems that differed with regard to climate, planttaxa, dominant mycorrhizal association (ectomycor-rhizal versus arbuscular mycorrhizal), and edaphicproperties. Numerous studies have examined turnoverdynamics of soil FAA (Kuzyakov and Galitsa 1993;Kuzyakov and Demin 1998; Jones and Kielland2002; Jones et al. 2005; Berthrong and Finzi 2006;Jones and Murphy 2007, Kielland et al. 2007).However, these studies focused within a particularecosystem or landscape in which many of the statefactors influencing soil development (climate, parentmaterial, vegetation, or time) were held constant. Toour knowledge, this is the first attempt to use acommon experimental approach to develop estimatesof amino acid turnover in situ across multiple biomeswhere C availability was expected to vary widely.

Our study was designed to test the idea thatturnover dynamics of FAA in forest soils areregulated more by C limitations than N limitationsto microbial growth (Jones and Murphy 2007). Wehad two objectives. First, without mechanicallydisturbing the soil profile, we estimated the turnoverrate of uniformly labeled 13C-glycine using 13CO2release as a proxy for residence time. Secondly, werelated the in situ turnover dynamics of glycine toseveral indices of C availability determined in the

this continent. Study sites selected within eachbiome, including the dominant forest ecosystemtypes were: Bonanza Creek Long-Term EcologicalResearch (LTER) site, AK, white spruce (Piceaglauca) and balsam poplar (Populus balsamifera);Ford Forestry Center, MI, sugar maple (Acer sac-charum); and Houghton, MI, red pine (Pinus resin-osa); Coweeta LTER, NC, tulip poplar (Liriodendrontulipifera); B.F. Grant Experimental Forest, GA,white oak (Quercus alba). Relevant features forforest types are discussed below while specific standcharacteristics are presented in Table 1.

Balsam poplar and white spruce are mid- to late-successional stands, respectively, in a primary suc-cessional sequence along the Tanana River floodplainin interior Alaska. These stands are predominantlyectomycorrhizal (EM). The soils of this chronose-quence are classified as Typic Cryofluvents (Vierecket al. 1993) and are predominantly silt-textured. Soilsalong older terraces are overlain with well-developedorganic horizons extending to 10 cm or more indepth. The climate is strongly continental, and forestsare exposed to sub-freezing conditions for much of

laboratory. We predicted that rate constants forglycine mineralization would vary inversely withthe overall decomposability of soil C across foresttypes (Kielland et al. 2007). Due to a high microbialdemand for labile substrate, we anticipated that foresttypes with high-quality litter containing low concen-trations of lignin, tannins, and other recalcitrantcompounds (temperate deciduous) would exhibitslower rates of consumption of added FAA thanforest types where plant litter chemistry and reducedsoil temperatures (boreal) act as a constraint to Ccycling.

Materials and methods

Study sites

Study sites were located across three North Americanbiomes: boreal, northern temperate and southerntemperate. We chose these sites in order to encom-pass the wide range of vegetation types, andenvironmental conditions represented by forests of

the year. Though classified as semi-arid, precipitationoften exceeds evapotranspiration due to low temper-atures and a restricted growing season. Rates of net Nmineralization are low compared to temperate forestecosystems, so putatively the availability of labile Nfor plant uptake is reduced. Lower N availabilityreduces plant litter quality as the plant communitycomposition changes through succession. It has beensuggested that the shift to plant detritus with higherC:N ultimately reduces the overall decomposabilityof soil organic matter in late successional communi-ties, thus decreasing C turnover. Consequently, soilmicroorganisms in these stands are believed tobecome increasingly C-limited (Flanagan and VanCleve 1983) during the transition from deciduoustree-dominated stands to conifers.

Sugar maple is a common deciduous species in theGreat Lakes and Acadian forest regions. As a habitatgeneralist, it is often found in mixed stands; however,our study area is located in a relatively pure stand ofsugar maple that was previously managed under aselective cutting regime. The entire area was cut over100 years ago, and most of the large overstory treeswere about 95-100 years old. A second harvestoccurred about 25 years prior to our study, at whichtime approximately 2/3 of the basal area was leftintact. Overstory taxa in this stand are predominantlyarbuscular mycorrhizal (AM). Understory vegetationis relatively sparse and consists primarily of perennialherbaceous plants and sugar maple seedlings andsaplings. Soils in this stand are well drained TypicHaplorthods, consisting of cobbly, silt and sandyloams with 2-12% clay content.

The red pine site (EM) is located at the WilliamPayne La Croix plantation established in 1950 nearHoughton, Michigan. This stand consists of evenlyspaced (1.8 m x 1.8 m square) mature trees with nounderstory, so red pine accounts for 100% of thebasal area. The overall terrain is relatively flat togently sloping. Soils are sandy loams, classified asEntic Haplorthods, with a thin organic horizon at thesurface consisting almost entirely of pine litter invarious states of decomposition.

The tulip poplar stand (AM) is situated inWatershed 3 of the Coweeta LTER research sitenear Franklin, North Carolina. The terrain of thisdeciduous hardwood cove is steep (>30% slope) withdeep ( ~ 1 m) well drained Humic Hapludults derivedfrom folded schist and gneiss. Natural reforestation

began ~ 50 years ago following agricultural aban-donment. The oldest trees in this stand date from thatperiod; however, there was some underplanting oftulip poplar seedlings in the 1970s in an effort toincrease stand density. The understory is sparse withscattered dogwood (Cornus florida) and stripedmaple (Acer pennslyvanicum) trees. The forest flooris rarely thicker than 2-3 cm except immediatelyfollowing leaf fall; there is no Oe or Oa horizon. Thehigh quality litter (C:N = 40; Lignin:N = 15)derived from this stand rapidly decomposes underthe. mesic conditions characteristic of this forest type.

The oak site (EM) is located within the B.F. GrantMemorial Forest, a 5000 ha mixed-use researchforest managed by the Warnell School of Forestryand Natural Resources at the University of Georgia,Athens. There is a long history of disturbance atthis site beginning with Native American encamp-ments, followed by slash and bum agriculture, andeventually cotton production. Soils, classified asTypic-Rhodic Hapludults, are well-drained, clayey,kaolinitic, and range in color from dark red to yellow-brown. The understory consists primarily of oak(Quercus spp.) and hickory (Carya spp.) with somemaple (Acer spp.), beech (Fagus grandifolia), anddogwood (Cornus florida). The forest floor is rela-tively thick in places (12 cm) given the climate andconsists primarily of an Oi layer with weaklydeveloped Oe and Oa horizons. Overall the terrainfor our study area is gently sloping with slopes of2-6%.

Soil incubation experiment

To evaluate C availability, we performed gas fluxmeasurements (net C mineralization) on soils fromeach of our research sites. In August 1998, wecollected twelve 5.5 cm diameter soil cores fromeach of the six locations and randomly paired them toproduce six laboratory replicates (Lancaster andKeller-McNulty 1998). The paired cores were storedfrozen for 3 months, thawed and partitioned intoupper (U) and lower (L) soil horizons. For the mullforest floor of red pine, sugar maple, white oak, andtulip poplar, U = 0-7 cm and L = 7-20 cm belowthe litter layer, while the mor forest floor of the borealstands was separated into relatively pure organic andmineral horizons that approximated the samplingdepths defined for the other stands. We define the

Joseph, Michigan, USA). Due to the acidity (Table 2)of these soils, carbonate removal was not necessaryprior to analysis.

In order to balance soil microbial activity againstgaseous N loss, soil moisture content was adjusted to55% water-holding capacity (Nunan et al. 2000) withdistilled water after gravimetric determinations weremade for each soil type. Approximately 100 g freshweight of soil was placed into 980 ml Mason jarswhose caps were fitted with butyl rubber septa. Thejars were sealed and preincubated for 3 days in thedark before gas measurements were initiated. Incu-bation temperatures varied according to the regionfrom which the soils were collected (for tulip poplarand white oak, 19.8°C; balsam poplar and whitespruce, 9°C; sugar maple and red pine, 17.0°C).These temperatures were representative of dailyaverage soil temperatures recorded at 7 cm depthfor each site 10 days prior to and following corecollection in July 1998.

We sampled the headspace of each jar weekly for16 weeks and determined the CO2 concentrationwithin using gas chromatography on a ShimadzuGC-8A fitted with a 200-cm Poropak column and a

core divisions in white spruce and balsam poplar asapproximate because compression within the organichorizons during sample collection likely resulted incore depths that were slightly deeper than 20 cm,Moreover, spatial heterogeneity in depth of the forestfloor along with buried organic layers among whitespruce and balsam poplar stands contributed differingamounts of organic matter to the mineral cores. Wechose these divisions to coordinate our laboratoryincubation experiment with temperature-dependentmodels of soil respiration developed from fieldmeasurements taken throughout the growing seasonat each site (unpublished data).

Subsequent to partitioning, we homogenized eachhorizon by removing all obvious woody debris, roots,and rocks and hand-mixing the remaining soil toobtain a relatively uniform substrate. Two subsam-pIes were taken from each homogenized core for soilmoisture and other chemical measurements. Subsam-ples for C and N analysis were dried to a constantweight at 60°C and powdered in a modified rollermill. Soil organic C (Ctotal) and total N weredetermined for each horizon using combustion anal-ysis on a LECO 2000 CNS autoanalyzer (LECO, St.

thermal conductivity detector (Shimadzu Corpora-tion, Japan). In order to prevent inhibition ofrespiration due to excessive concentrations of CO2

in the headspace, jars were capped for only 24 h priorto each measurement and then aerated before beingreturned to the incubation chamber. Between sam-pling periods, each jar was covered with 0.8 milpolyethylene sheeting secured with a rubber band toprevent excessive moisture loss while still permittinggas exchange (Gordon et al. 1987). Following gassampling, the water content of each jar was main-tained at 55% WHC by adding deionized water tocompensate for the measured weight loss. First orderrate constants for microbial respiration were calcu-lated using the following equation:

Ct = Cmin (1 - e-kt)

where Ct is the cumulative carbon mineralization upto time, t (days), Cmin is the potentially mineralizablepool of soil carbon, and k is the instantaneous rateconstant describing the daily release of C from thatpool (Kielland et al. 1997). Due to differences in Ccontent among soil types we normalized Cmin by totalsoil organic C (Ctotal), so that instantaneous rateconstants (Kc) reflect site to site variation in Ctotal.Additionally, we tested a double exponential modelwhich considers two soil organic matter pools withdiffering susceptibility to decomposition (Alavarezand Alavarez 2000). However, due to the relativelyshort duration of our soil incubation experiment, wefound that a single pool C mineralization modeladequately described our results.

In situ glycine C mineralization experiment

The field component for this experiment was con-ducted during July 1999 for white spruce, July 2005for balsam poplar, and from June to July 2000 for theremaining forest types. Randomly, we established 2soil injection grid locations within a 9 m2 plot. Eachplot was replicated six times along a transect withineach forest type. Injection grids were constructed of3.2 mm lexan sheets that were flexible enough tomould to the surface of the forest floor. Grids wereheld in position by four steel pins buried to a depth of20 cm, which made it easy to return periodically andprecisely align our gas sampling chamber over thehead space above each injection core (see McFarlandet al. 2002 for visual depiction of grid design). Within

each injection grid, we administered either U-[l3C2]-glycine (glycine treatment) or distilled water (con-trol) to a depth of 10 cm. As part of a companionexperiment examining plant-microbial competitionfor N, we added (l5NH4

+)2SO4 to the 13C-glycinetreatment and established a second set of injectioncores in which glycine was labeled with 15N.Following gas measurements for glycine-derivedl3CO2, we harvested a 5.5 x 12 cm (w x d) core totrack the fate of our N additions from the(15NH4 +)2SO4 and 15N-glycine treatments. Thesecores were harvested in the manner of McFarlandet al. (2002) which permitted us to reasonablyestimate total recovery of NH4

+ - and glycine-derivedN in dissolved inorganic N (DIN), dissolved organicN, (DON), microbial N, (MBN) and fine root Npools. Results from the 15N tracer study are brieflydiscussed in the context of microbial C:N balancebelow and in greater detail elsewhere (McFarlandet al. in press). Total injection volume was 37 ml(~1 ml cm-2), which delivered 0.39 g l3C m-2 andensured that cross-site differences in soil moisturewere minimized.

We collected the l3CO2 efflux above each injec-tion core using a capped segment of 10.2 cm ABSpipe fitted with a #10 rubber stopper. Inserted intoeach stopper was a short segment of polyethylenetubing connected to a 30 ml syringe via an air-tightstop cock. We used high-pressure vacuum grease toestablish an airtight seal for the luer fitting betweenthe stopcock and syringe as well as the connectionbetween the polyethylene tubing and stopcock. Thelitter layer above each injection point was removed toreveal the partially decomposed horizons below.While eliminating the portion of the microbialcommunity associated with litter, removing the litterlayer dramatically improved our ability to seal thesampling chambers against the forest floor. Samplingchambers were pressed against the soil surface for3 min at which time a 15 ml sample was collected.We exercised caution in sealing the chamber to thesoil so as to avoid depressing the surface andreleasing excess CO2. Gas in the sampling chamberwas thoroughly mixed by slowly pumping the syringe10 times prior to sample collection. Gas samples weretransferred over-pressurized to 10 ml exetainers(Labco. Ltd., UK) pre-evacuated to 0.007 kPa. Allsoils, with the exception of balsam poplar, weresampled at 6 periods (0.75, 2, 12,24, 168, and 336 h)

following injection. Balsam poplar was sampled onlyat the first four sampling periods. We monitored soiltemperature at a depth of 7 cm continuously through-out the experiment using HOBO temperature loggers(Onset Computer Corporation, Massachusetts, USA).

Gas samples were analyzed for l3C-CO2 using aEuropa Scientific continuous flow mass spectrometer(SPEC-PDZ Europa Inc., UK). We report the data ascumulative 13C atom percent excess (APE) ofrespired CO2. APE was determined by subtractingthe atom % l3C of control samples from the atom %13C of samples treated with labeled glycine. Controlvalues were averaged within a site prior to use inestimating enrichment. Data were fitted to the samesingle exponential model used for the soil incubationstudy, except that all fitted curves were forcedthrough the origin based on the assumption that 13Cexcess was zero prior to injection.

Soil amino acid-N and DIN extraction andquantification

We randomly collected 15 soil cores to a depth of12 cm along our transect using a 5.5 cm (I.D.) steelcorer combusted at 450°C for 6 h prior to use.Sampling depth was chosen to coincide with coresharvested to trace the fate of our 15N additions. Allcores were handled with nitrile gloves and stored inclean polyethylene bags on ice during transport to thelaboratory. Within 4 h, each core was gently hand-mixed and sieved (2.5 mm mesh) to remove rocks,large roots and as many small roots as possible. Wetook two subsamples from each homogenized core.One subsample was dried at 70°C to determinegravimetric moisture content. The other subsample(15 g wet weight) was extracted with 75 ml distilledwater (15 min at 150 rev min-1) and vacuum-filteredthrough a 0.2 um cellulose acetate filter (Corning Inc,Coming, New York, USA.). Soil extracts were storedfrozen in 2 ml sterile polyethylene vials until analysis.

We analyzed soil extracts for total FAA usingfluorometries (Jones et al. 2002). Briefly, 20 ul ofsample, standard, or blank was pipetted to a 96 wellmicroplate. We used a Precision 2000 automatedpipetting system (Bio- Tek Instruments, Inc., Winoo-ski, VT, USA) to add 100 ul of a working reagentconsisting of a borate buffer, o-phthaldialdehyde, andB-mercaptoethanol to each well. Following derivati-zation (=2 min), the fluorescence in each well was

measured using a Biotek FL600 Fluorescence andAbsorbance Microplate Reader (Bio- Tek Instruments,Inc., Winooski, VT, USA) with excitation and emis-sion wavelengths set to 340 and 450 nm, respectively.Each sample was run in quadruplicate and the resultsare reported as ug amino acid-N per g dry soil.

Statistical analyses

We fitted the first order rate equations to CO2

production from the microcosm study (n = 6) aswell as cumulative APE l3C from the glycinemineralization experiment (n = 6) for each foresttype using PROC NLIN in SAS. Tukey's multiplesample t tests were applied to all pairwise compar-isons of kglycine, 13CAPEcum, and residence time forglycine in situ (1/kglycine); the rate constant kc;estimates for pool size, Cmin; and all measured soilvariables from our laboratory incubations, includingsoil organic C and total N. We tested the assumptionof normality for all the aforementioned parametersusing PROC UNIVARIATE in SAS prior to con-ducting one and two-way ANOVAs to test forsignificant differences among forest types. Whennecessary, variables were log-transformed to meet theconditions for normality and constancy of variance.When log transformed values also failed to meet theassumptions for ANOVA we performed analyses onranked values. Simple linear regression analyses wereused to relate kglycine and 1/kglycine with soil temper-ature, soil C:N, and CO2 production for the upper soilhorizon. All inferences regarding pool dynamics aremade at the stand level, and significance for allstatistical analyses was accepted at a = 0.05. Resultsfor kglycine,13CAPEcum, and residence time for glycine(l/kglycine), did not meet these criteria for statisticalsignificance. However, since these parameters werederived from measurements made in an open systemas opposed to a highly controlled laboratory micro-cosm, we discuss these results as a non-significanttrend where P < 0.10;

Results

Soil C, N, and mineralizable C

Soil organic C and total N varied significantlybetween the boreal and temperate stands, particularly

temperate stands. Average values ranged from1.6 g N kg-l in white oak to 12.4 g N kg-1 inbalsam poplar. Soil N concentrations in the lowerhorizon were less variable, with the highest valuesmeasured in tulip polar and sugar maple(2.9 g N kg-l) and the lowest in red pine(0.8 g N kg-1),

Total C respired in the upper soil horizon washighest in balsam poplar and white spruce soils, leastin tulip poplar and white oak soils, and intermediatein sugar maple and red pine soils (Fig. la). Carbon

for the upper soil horizon. Soil organic C in the upperhorizon averaged 169 and 233 g C kg-l for whitespruce and balsam poplar, respectively, whereasvalues for the temperate stands fell within a narrowerrange, averaging between 26 and 67 g C kg-1

(Table 2). Carbon content in the lower horizon wasmore similar among all forest types, ranging from45 g C kg-l in tulip poplar to 16 g C kg-1 in redpine. Total soil N also differed significantly betweenthe boreal and temperate forests with larger concen-trations of N in the upper horizon for boreal than

across ecosystems (F5,30 = 36,58, P < 0.0001),Since our intention was to estimate net C mineraliza-tion rates under near-natural environmental condi-tions, we remind the reader that cores from differentsites were incubated at different temperatures, There-fore, net C mineralization rates reflect differences inboth temperature and microbial C utilization. Thehighest C turnover rates were observed in tulip poplarsoils (0.0228 ± 0.0009 days-1) and the lowest inwhite spruce (0.0061 ± 0,0012 days-1; Table 3).Though rate constants appeared to increase along thenorth-south gradient, we observed two distinct group-ings between the temperate deciduous stands and theconiferous stands plus balsam poplar. These differ-ences were reflected in the mean residence time (1/k)for Cmin, where pool turnover was, on average, twiceas rapid for the temperate deciduous stands comparedwith the coniferous stands plus balsam poplar. Withina stand, the residence time for Cmin was very similarbetween upper and lower soil horizons for all standsexcept red pine (Table 3). Mean residence time forCmin in red pine was 92.3 days in the lower soilhorizon versus 184.0 days in the upper soil horizon.Similarly, average Cmin/Ctotal was almost 2-fold higher(P < 0.05; n = 6) for the upper soil horizon suggest-ing that substrate quality could be more verticallystratified in red pine than the other forest types.

Soil FAA-N and DIN concentrations

Soil FAA concentrations differed significantly amongforest types (FS,84 = 15.2, P < 0.0001). Most of the

dioxide accumulation curves constructed for each sitereveal that net mineralizable C (Cmin) ranged from1.51 g C kg soil-1 in white oak to 11.89 g C kgsoil-1 in white spruce, and was positively correlatedwith Ctotal (r2 = 0.65, P < 0.001; n = 36). ThoughCmin increased significantly with increasing latitude(P < 0.001), this trend was no longer apparent whenrespired C was adjusted for Ctotal (Fig. 1b). Normal-ized for C content, Cmin ranged from 3.4% of Ctotal inthe southern deciduous tulip poplar stand to 15.7% inthe northern red pine plantation (Table 3). Overall redpine soils had the highest C efflux per unit soil C in theupper horizon (P < 0,05), indicating that a higherproportion of SOM in red pine was metabolizable tosoil microbes. Alternatively, differences in Cmin couldindicate that microbial C-use efficiency differedamong forest types.

In the lower soil horizon, Cmin also varied relativeto Ctotal across all samples (r2 = 0.41, P < 0,001;n = 36; Fig. 2a). Soils with larger C stocks in thelower horizons generally had greater Cmin, with theexception of red pine which had significantly smallerCtotal than tulip poplar (Table 2), but similar valuesfor Cmin. When adjusted for Ctotal (Fig. 2b), differ-ences among forest types followed the same generalpattern observed for the upper horizon, again sug-gesting differences in C-use efficiency among foresttypes, or that a larger fraction of soil C in thetemperate deciduous stands was resistant to microbialdegradation compared to red pine.

Rate constants for C mineralization derived fromour single exponential model varied significantly

observed variation was attributed to soils fromAlaskan stands, which in some instances had overl0-fold higher concentrations of FAA than soils fromthe southern deciduous stands (Fig. 3). Averagevalues ranged from just over 0.5 mg AA-N kg-1 soilin tulip poplar to just over 8 mg AA-N kg-1 inbalsam poplar (Table 2). FAA-N in red pine soils(3.25 ± 0.69 mg N kg-1 were significantly greaterthan those of sugar maple (0.97 ± 0.14 mg N kg-1

P < 0.05; n = 15) even though these forest typesshare a similar climate and other factors influencing

soil development, e.g. topography and parent mate-rial. Conversely we observed little variation indissolved inorganic N (DIN) across all sites withthe exception of red pine stands which had signifi-cantly lower concentrations of inorganic N(P < 0.05) than any of the other forest types.

In situ glycine C mineralization

In situ turnover rates for glycine were rapid regard-less of soil type (Table 4); nevertheless, rate

constants for glycine turnover (kg1y) determined fromour single exponential model were statisticallydifferent across all forest types (F5,30 = 3.34,P = 0.02; Table 4). Mean residence times (kgly)did not vary systematically with latitude, but therewas clustering among forest types that somewhatparalleled results from the laboratory incubation(Fig. 4). Contrary to predictions, mean in situ turn-over rates for glycine were significantly faster in tulippoplar soils than in white spruce soils (P < 0.05)while turnover rates for the remaining forest typesshowed no statistical differences. However, it isworth noting that mean in situ glycine turnover rates

were nearly identical for southern white oak andboreal balsam poplar (0.70 and 0.69 h-1, respec-tively), whereas northern sugar maple demonstratednon-significantly higher turnover rates for glycine(0.88 h-1; Table 4), suggesting climatic effects alonecannot explain cross-ecosystem variation in theturnover of soil FAA.

In situ glycine turnover was correlated withindices of substrate (labile C) availability. Forexample, soil C:N ratio explained over 80% of thevariability observed in glycine turnover rate (kgly)among sites (Fig. 5; r2 = 0.82, P = 0.01). Thehighest values for kg1y were observed in soils with

during the field experiment (r2 = 0.14, P = 0.47).Thus it appears that among forest types, substratequality (Cmin/Ctotal) had a more dramatic impact onthe turnover of glycine than either temperature or soilFAA concentrations.

Discussion

We found that in situ rates of glycine turnover wererapid across all biomes, and that there was strongsupport for our hypothesis that consumption ofsoluble amino acids on a continental scale is linked

the lowest C:N ratio and vice versa. In contrast to ouroriginal predictions concerning substrate quality, alow CN ratio may be indicative of highly processedsoil C that presents a relatively C poor substrate formicrobial growth (see "Discussion"). Additionally,we found a strong positive correlation between rateconstants for in situ glycine turnover (kgly) and Cmin

(kc) determined from laboratory incubations (Fig. 6;r2 = 0.67, P < 0.05). However, there was no signif-icant relationship between Cminor FAA-N concen-trations and kgly across all ecosystems. Similarly, weobserved no significant correlation between kg1y andsoil temperature measured continuously at each site

to the turnover of soil C pools. In general, mineral-ization of our glycine additions increased withdecreasing pools of labile C, suggesting a microbialresponse to C limitation. Although soil microorganismsin all forest types rapidly mineralized glycine, neitherthe magnitude of response to our glycine addition northe size of the labile fraction of SOM varied predictablyalong our latitudinal gradient. We discuss these resultsin the context of relevant studies exploring the factorsregulating the turnover of soil FAA.

Recent studies indicate that low-molecular-weightorganic compounds, including FAA, play an impor-tant role in sustaining the short-term energy balanceof microorganisms involved in the decomposition ofSOM (De Nobli et al. 2001; Mondini et al. 2006). Themore recalcitrant the SOM, the less likely SOM aloneprovides sufficient substrate for the basal metabolismand growth of soil microorganisms. The rapidity withwhich our labeled substrate appeared in soil CO2

efflux provides further evidence that FAA repre-sented a labile source of soil C that was rapidlymetabolized by soil microbes. However, our initialideas concerning the variability of soil C availabilityacross forest types-increasing SOM quality with

decreasing latitude-were not as straightforward aspredicted.

Results from our laboratory incubations show thatboreal forest soils yielded substantially larger pools ofrespired C than mid-latitude soils (Fig. la) in the upperhorizon, and soil C stocks explained most of theobserved differences in mineralizable C (Cmin) amongforest types. This was not surprising given thatdecomposition is constrained in part by temperature(Hart and Perry 1999; Garten and Hanson 2006).Stands that had a lower mean annual temperature(MAT), also had significantly higher stocks of soil Cand N and thus larger pools of Cmin. However, despite astrong correlation between latitude and total C respired,the proportion of soil C that was readily mineralizableat each site did not necessarily conform to predictionsconcerning MAT or litter quality (Fig. 1b) particularlyin the temperate regions. This suggests that neitherMAT nor litter quality alone provided an adequateexplanation for FAA turnover and Cmin.

Forests dominated by species producing highquality (low lignin:N; Table 5) aboveground (AG)litter, e.g. tulip polar and sugar maple, would beexpected to have higher rates of litter decomposition

and thus proportionately larger pools of labile C thanforests dominated by species producing more recal-citrant litter, e.g. red pine (Moorehead et al. 1999).Contrary, Cmin accounted for a larger proportion oftotal C in stands producing lower quality AG litter,e.g. red pine, despite that turnover rates (kc) for soil Cwere slower for red pine than for sugar maple or tulippoplar. Organic matter from both soil horizonscontained 2-4 times more labile C under red pineversus sugar maple or tulip poplar (Figs. lb, 2b). Thisapparent discrepancy might reflect differences inearly-stage versus late-stage decomposition of thesedissimilar litter types. Field studies of litter decom-position have demonstrated a limit value for massloss beyond which decomposition either ceases orproceeds very slowly as the remaining mass becomespart of soil humus. This limit value is highlycorrelated with the initial N concentration of freshlitter inputs. The higher the N levels of a litter, thefaster the initial rates of decomposition, but morerecalcitrant mass remains during the late stages ofdecomposition (Berg and Ekbohm 1991; Berg andMeentemeyer 2002; but see Hobbie 2000).

In our study we observed a strong relationshipbetween soil CN and kgly whereby soils with lowerC:N trended towards higher turnover rates for glycine(Fig. 5). Assuming a C-limited soil environment and ahigh metabolic demand in soils with lower C:N couldexplain why microbes in tulip poplar and sugar mapleresponded to our glycine additions with faster turnoverrates than observed in oak or red pine where Cmin/Ctotal

is higher. In many soils, the microbial biomassmaintains a high state of metabolic readiness (Brookeset al. 1987), even though substrate availability isusually scarce. The rationale for sustaining such a highmetabolic status in an energy-poor environment stemsfrom the need to compete effectively for temporallyand spatially infrequent pulses of labile substrate, e.g.rhizo-deposition (Mondini et al. 2006). Still, glycinerepresents a source of labile N as well as C, and to linkthe turnover of glycine to C-limitation in stands withlow C:N, we must first reject the argument that theresidence time for glycine was not linked as well to N-limitation in those stands.

In our companion experiment (McFarland et al. inpress) we observed several interesting trends in thefate of our NH4

+, and glycine-derived 15

N tracers.First, microbial immobilization represented the larg-est short-term biotic sink for both N forms regardless

of forest type. Second, long-term measurements ofMBN turnover across our latitudinal gradient indi-cated that the majority of 15N for both N treatmentswas ultimately transferred to stable soil N pools.These results suggest that microbes do indeed targetboth substrates for N and thus may be N-limited.However, there were deviations in the cyclingpatterns for glycine and NH4

+ for some forest typesthat warrant further consideration. First, in sugarmaple and tulip poplar, microbial immobilization was41-56% higher for glycine than NH4

+ at the firstsampling period. In contrast, immobilization wassimilar for the two N forms for the other forest types.Second, the mineralization of glycine-N to dissolvedinorganic N (DIN) was higher than the conversion ofNH4

+ to dissolved organic N (DON) under sugarmaple and tulip poplar. Third, the incorporation ofNH4

+ into MBN was significantly lower than theincorporation of glycine-N under sugar maple andrelatively low for both N forms under tulip poplar.Together, these results indicate that soil microorgan-isms in the two AM-dominated stands were likelyutilizing glycine primarily as a C source. Unfortu-nately, since the cycling dynamics of NH4

+ andglycine-N varied more or less in concert for the otherforest types, we cannot rule out the possibility that theturnover of glycine is linked to microbial N-limita-tion in those soils.

Processes affecting the biodegradation of glycinereflect a complex of interacting factors that we didnot study including, among others, microbial com-munity composition. Community structure of soilmicroorganisms can strongly influence the incorpo-ration of plant litter into SOM and thus theavailability of labile C (Elliot et al. 1993). The sizeof the microbial biomass may be less important thanthe activity of the community in predicting decom-position rates, particularly if metabolic activities ofmicrobes are adapted to available substrates (Elliotet al. 1993; Waldrop and Firestone 2004; but seeMcClaugherty et al. 1985). Data from our companionstudy, (McFarland et al. in press), indicate nocorrelation between MBN and glycine turnover rates.Using MBN as a surrogate for microbial biomass, thissuggests that patterns in glycine turnover are likelydriven by differences in heterotrophic consumptionrather than the size of the microbial biomass per se.Recent evidence from low fertility black oak/whiteoak and high fertility sugar maple/basswood

ecosystems indicates that standing pools of freeamino acids are higher when mineral N availability islow despite that amino acid production does not varysignificantly between these two forest types (Roth-stein 2009). We propose that differences in aminoacid consumption, related to microbial C rather thanN balance, may explain this discrepancy betweenproduction and pool size.

In our study, forest types are characterized aspredominantly arbuscular (AM) or ectomycorrhizal(EM) with respect to the overstory taxa (Pregitzer et al.2002; Lansing, "unpublished data") and we suspectthat these fungal associations represent a significantportion of the soil microbial biomass. Mycorrhizalfungi can influence SOM quality by regulating bothavailability and turnover of soil C, but the effects ofAM and EM on rhizosphere processes appear to differ(Langley and Hungate 2003). Ectomycorrhizal fungihave the potential to reduce both the size and activityof bacterial biomass in the mycorrhizosphere bychanneling plant C into recalcitrant EM structuresrather than labile exudates (Olsson et al. 1996a). Incontrast some AM roots have a capacity to promote theactivity of rhizobacteria through root or fungal exu-dation into mycorrhizospheric soil (Olsson et al.1996b; Andrade et al. 1997). During the degradationof organic matter, calorimetric values can be signif-icantly higher for bacteria than fungi (Critter et al.2004), though the relative importance of bacteria andfungi to total metabolism may be soil and substratedependent. Elliot et al. (1993) found that fungalcontributions to microbial metabolism within theforest floor increased significantly with substraterecalcitrance across a range of forest types. We didnot assess the ratio of fungi:bacteria in soils from anyof our sites in relation to mycorrhizal association;however, it is worth noting in our study that in situturnover rates for glycine were highest in the standsdominated by AM species (sugar maple and tulippoplar), and lowest in the EM-dominated white spruceand red pine stands. If AM associations tend to enrichbacterial flora while EM associations render therhizopshere and surrounding soil less hospitable tobacterial growth, the predominance of one mycorrhi-zal type over the other among our experimental forestscould partially explain differences in glycine turnover(given its lability) among stand types.

There are several caveats to our interpretationsregarding FAA turnover in the field given the

limitations of our experimental design. First, microbialpreference for a particular substrate as well as differ-ential partitioning of that substrate into anabolic versuscatabolic pathways can influence the turnover of thatsubstrate in microbial pools. We concede that themagnitude of the mineralization response we measuredmight be amino acid specific. In our study, comparingturnover dynamics of a common substrate (glycine)versus a site-specific amino acid cocktail as determinedfrom pool constituency most certainly would have beenmore informative. However, we feel the overall patternof response among our sites using glycine alone is stilluseful in evaluating microbial response to thesesubstrates in intact systems with varying soil proper-ties. Second, we added the same quantity of glycine tosoils for each stand type rather than a fixed percent ofbackground pools. Prior to our experiment, values forsoil concentrations of FAA across our latitudinalgradient were not available and a priori characteriza-tion of soil FAA pool size for each stand type was notlogistically feasible. Glycine amendments representeda 2-5-fold increase in FAA-N concentrations for ourdeciduous stands, but less than a 2-fold increase inFAA-N for the coniferous stands plus balsam poplar.Our experimental design reflects a trade-off betweendetecting the l3C-tracer against background pools in anopen system, and potentially initiating a priming effect.Therefore, it is worth noting that the response (e.g.,kg1y) we measured in the field could be driven in part byan imbalance in nutrient to C ratios particularly amongstands with narrow soil C:N. Finally, other factors,some of which we measured (soil temperature) andsome of which we didn't (microbial C) could also haveinfluenced FAA turnover among forest types. How-ever, regardless of their contribution to the cyclingdynamics of FAA, these effects were not strong enoughto disrupt the robust relationship between the cyclingrates of FAA and labile fractions of SOM acrossdisparate forest types.

Conclusion

Amino acids represent a significant fraction ofdissolved organic N in forest soils and a number ofexperiments have elucidated factors controlling theproduction and/or turnover of these compounds.However, the primary motivation behind much thisresearch effort has centered on issues pertaining to

plant nutrition or the overall N economy of soils, nottheir role as a C substrate that influences the metabolicstatus of the soil microbial community. This studyillustrates, (1) that FAA are an important substrate forsoil microbial metabolism in many terrestrial forestcommunities, and (2) patterns of amino acid turnoverin situ across ecosystems are closely linked to indicesof SOM quality. We found that across large spatialscales, consumption of glycine by soil microorganismsis better explained by changes in soil C availabilitythan cross-ecosystem variation in soil temperature orstanding pools of FAA This suggests the overalldecomposability of native C and patterns of hetero-trophic consumption of soil C influence turnover ratesfor low-molecular-weight organic substrates such asamino acids.

Acknowledgements Greg Mauer, Jennifer Mitchell, andJason Schneider provided invaluable assistance in the field.We also wish to thank Lee Ogden of the Warnell School ofForest Resources at the University of Georgia, and Dr, AndrewBurton of the School of Forest Resources & EnvironmentalSciences at Michigan Technological University, for theirextraordinary logistical support during our field efforts. Dr.Adrien C. Finzi, Dr. Stephen C. Hart and Dr. F. Stuart 'Terry'Chapin III and two anonymous reviewers greatly improved thequality of this manuscript with their comments. This researchwas supported through grants from the University of Alaska'sCenter for Global Change, the National Science Foundation'sEcosystem program, and the Bonanza Creek LTER program.

References

Alavarez R, Alavarez CR (2000) Soil organic matter pools andtheir association with carbon mineralization kinetics. SoilSci Soc Am J 64:184-189

Andrade G, Mihara KL, Linderman RO, Bethlenfalvay GJ(1997) Bacteria from rhizosphere and hyphosphere soilsof different arbuscular mycorrhizal fungi. Plant Soil192:71-79

Atkin OK (1996) Reassessing the nitrogen relations of Arcticplants: a mini-review. Plant Cell Environ 19:695-704

Bardgett RD, Streeter TC, Bol R (2003) Soil microbes competeeffectively with plants for organic-nitrogen inputs totemperate grasslands. Ecology 84: 1277-1287

Berg B, Ekbohm G (1991) Litter mass-loss rates and decom-position patterns in some needle and leaf litter types.Long-term decomposition in a Scots pine forest. VII. CanJ Bot 69:1449-1456

Berg B, Meentemeyer V (2002) Litter quality in a northEuropean transect versus carbon storage potentiaL PlantSoil 242:83-92

Berthrong ST, Finzi AC (2006) Amino acid cycling in threecold-temperate forests of the Northeastern USiL Soil BiolBiochem 38:861-869

Brookes PC, Newcomb AD, Jenkinson DS (1987) Adenylateenergy charge measurements in soil. Soil Biol Biochem19:211-217

Critter SAM, Freitas 55, Airoldi CC (2004) Microcalorimetricmeasurements of the metabolic activity by bacteria andfungi in some Brazilian soils amended with differentorganic matter. Thermochimica Acta 417:275-281

De Nobli M, Contin M, Mondini C, Brookes PC (2001) Soilmicrobial biomass is triggered into activity by traceamounts of substrate. Soil Bioi Biochem 33: 1163-1170

Di HJ, Aylmore LAG, Kookana RS (1998) Degradation ratesof eight pesticides in surface and subsurface soils underlaboratory and field conditions. Soil Sci 163:404-411

Elliot WM, Elliot NB, Wyman RL (1993) Relative effect oflitter and forest type on rate of decomposition. Am MidlNat 129:87-95

Finzi AC, Berthrong ST (2005) The uptake of amino acids bymicrobes and trees in three cold-temperate forests. Ecol-ogy 86:3345-3353

Flanagan PW, Van Cleve K (1983) Nutrient cycling in relationto decomposition and organic-matter quality in taigaecosystems. Can J Forest Resour 13:795-817

Garten CT, Hanson PJ (2006) Measured forest soil C stocksand estimated turnover along an elevation gradient.Geoderma 136:342-352

Gonod LV, Jones DL, Chenu C (2005). Sorption regulates thefate of the amino acids lysine and leucine in soil aggre-gates. Eur J Soil Sci 57:320-329

Gordon AM, Tallas M, Van Cleve K (1987) Soil incubations inpolyethylene bags: effect of bag thickness and tempera-ture on nitrogen transformations and CO2 permeabilityCan J Soil Sci 67:65-75

Hart SC, Perry DA (1999) Transferring soils from high- to low-elevation forests increases nitrogen cycling rates: climatechange implications. Glob Change Biol 5:23-32

Hobbie SE (2000) Interactions between litter lignin and soil Navailability during leaf litter decomposition in a Hawaiianmontane forest Ecosystems 3:484-494

Jones DL (1999) Amino acid biodegradation and its potentialeffects on organic nitrogen capture by plants. Soil BiolBiochem 31:613-622

Jones DL, Kielland K (2002) Soil amino acid turnover domi-nates the nitrogen fiux in permafrost-dominated taigaforest soils. Soil Biol Biochem 34:209-219

Jones DL, Murphy DV (2007) Microbial response time to sugarand amino acid additions to soil. Soil Biol Biochem39:2178-2182

Jones DL, Shannon D (1999) Mineralization of amino acidsapplied to soils: impact of soil sieving, storage and inor-ganic nitrogen additions. Soil Sci Soc Am J 63: 1199-1206

Jones DL, Owen AG, Farrar JL (2002) Simple method toenable the high resolution determination of total freeamino acids in soil solutions and soil extracts. Soil BioiBiochem 34: 1893-1902

Jones DL, Farrar JL, Newsham KK (2004) Rapid amino acidcycling in Arctic and Anarctic soils. Water Air Soil PollutFocus 4:169-175

Jones DL, Kemmitt 51, Wright D, Cuttle SP, Bol R, EdwardsAC (2005) Rapid intrinsic rates of amino acid biodegra-dation in soils are unaffected by agricultural managementstrategy. Soil Biol Biochem 37:1267-1275

Kaye JP, Hart SC (1997) Competition for nitrogen betweenplants and soil microorganisms. Trends Ecol Evol12:139-143

Kielland K (2001) Short-circuiting the nitrogen cycle: strate-gies of nitrogen uptake in plants from marginal ecosys-tems. In: Ae N. Arihara J, Okada K, Srinivasan A (eds)Plant nutrient acquisition: new perspectives. Springer-Verlag, Berlin, pp 376-398

Kielland K, Bryant JP, Ruess RW (1997) Moose herbivory andcarbon turnover of early successional stands in interiorAlaska. Oikos 80:25-30

Kielland K, Mcfarland JM, Ruess RW, Olson K (2007) Rapidcycling of organic nitrogen in taiga forest ecosystems.Ecosystems 10:360-368

Kuzyakov Y, Demin V (1998) CO2 efflux by rapid decompo-sition of low molecular organic substances in soils. SciSoils 3:11-22

Kuzyakov YV, Galitsa SV (1993) Kinetics of [2-14C] glycinedecomposition and its incorporation into humus fractionsof Sierozem. Eurasion Soil Sci 25:39-53

Kuzyakov Y, Friedel JK. Stahr K (2000) Review of mecha-nisms and quantification of priming effects. Soil BiolBiochem 32: 1485-1498

Lancaster VA, Keller-McNulty S (1998) A review of com-posite sampling methods. J Am Stat Assoc 93:1216-1230

Langley JA, Hungate BA (2003) Mycorrhizal controls onbelowground litter quality. Ecology 84:2302-2312

Lipson D. Nasholm T (2001) The unexpected versatility ofplants: organic nitrogen use and availability in terrestrialecosystems. Oecologia 128:305-316

Lipson DA, Raab TK, Schmidt SK, Monson RK (1999) Vari-ation in competitive abilities of plants and microbes forspecific amino acids. Biol Fertil Soils 29:257-261

McClaugherty CA, Pastor J, Abel JD, Mellilo JM (1985) Forestlitter decomposition in relation to soil N dynamics andlitter quality. Ecology 66:266-275

McFarland JW, Ruess RW, Kielland K, Doyle AP (2002)Cycling dynamics of NH4

+ and amino acid nitrogen insoils of a deciduous boreal forest ecosystem. Ecosystems5:775-788

McFarland JW. Ruess R, Kielland K, Pregitzer K, Hendrick R,Allen M (in press) Cross-ecosystem comparisons of in situplant uptake of amino acid-N and NH4

+. EcosystemsMondini C, Cayuela ML, Sanchez-Monedero MA, Roig A,

Brookes PC (2006) Soil microbial biomass activation bytrace amounts of readily available substrate. Biol FertilSoils 42:542-549

Moorehead DL, Currie WS, Rastetter EB, Parton WJ, HarmonME (1999) Climate and litter quality controls on decom-position: an analysis of modeling approaches. GlobalBiogeochern Cycles 13:575-589

Morita RY (1988) Bioavailability of energy and its relationshipto growth and starvation survival in nature. Can JMicrobiol 34:436-441

Nasholm T, Ekblad A, Hogberg P (1998) Boreal forest plantstake up organic nitrogen. Nature 392:914-917

Nasholm T, Huss-Danell K, Hogberg P (2000) Uptake oforganic nitrogen by four agriculturally important plantspecies. Ecology 8l:l155-1161

Nordin A, Schmidt IK, Shaver GR (2004) Nitrogen uptake byArctic soil microbes and plants in relation to soil nitrogensupply. Ecology 85:955-962

Nunan N, Morgan MA, Scott J, Herlihy M (2000) Temporalchanges in nitrogen mineralization. microbial biomass,respiration and protease activity in a clay loam soil underambient temperature. Biol Environ Proc R Ir Acad100B:107-114

Olsson PA, Baath Chalot M, RD Soderstrom B (1996a) Ecto-mycorrhizal mycelia reduce bacterial activity in a sandysoil. FEMS Microb Lett 21 :77-86

Olsson PA, Baath E, Jakobsen I, Soderstrom B (1996b) Soilbacteria respond to presence of roots but not to myceliumof arbuscular mycorrhizal fungi. Soil Biol Biochem28:463-470

Pregitzer KS, Deforest JA, Burton AJ, Allen MF, Ruess RW,Hendrick RL (2002) Fine root architecture of nine NorthAmerican trees. Ecol Monogr 72:293-309

Rothstein DE (2009) Soil amino acid availability across atemperate-forest fertility gradient. Biogeochemistry92:201-215

Schimel JP, Chapin FS III (1996) Tundra plant uptake of aminoacid and NH4

+ nitrogen in situ: plants compete well foramino acid N. Ecology 77:2142-2147

Schmidt S, Stewart GR (1999) Glycine metabolism by plantroots and its occurrence in Australian plant communities.Aust J Plant Physiol 26:253-264

Treseder KK (2008) Nitrogen additions and microbial biomass:a meta-analysis of ecosystem studies. Ecol Lett ll: 1111-1120

Viereck LA, Dyrness CT. Foote MJ (1993) An overview of thevegetation and soils of the floodplain ecosystems of theTanana River. interior Alaska. Can J For Res 23:889-898

Vinolas LC, Vallejo VR, Jones DL (2001a) Control of aminoacid mineralization and microbial metabolism by tem-perature. Soil BioI Bicchern 33:1137-1140

Vinolas LC, Healey JR, Jones DL (200lb) Kinetics of soilmicrobial uptake of free amino acids. Biol Fertil Soils33:67-74

Waldrop MP, Firestone MK (2004) Microbial communityutilization of recalcitrant and simple carbon compounds:impact of oak-woodland plant communities. Oecologia138:275-284

Zak DR, Tilman D, Parmenter RR, Rice CW, Fisher FM, VaseJ, Milchunas D, Martin CW (1994) Plant production andsoil microorganisms in late-successional ecosystems: acontinental-scale study. Ecology 75:2333-2347