12
Plant and Soil 240: 275–286, 2002. © 2002 Kluwer Academic Publishers. Printed in the Netherlands. 275 Decomposability of C3 and C4 grass litter sampled under different concentrations of atmospheric carbon dioxide at a natural CO 2 spring D. J. Ross 1,3 , K. R. Tate 1 , P. C. D. Newton 2 & H. Clark 2 1 Landcare Research, Private Bag 11052, Palmerston North, New Zealand. 2 AgResearch Grasslands, Private Bag 11008, Palmerston North, New Zealand. 3 Corresponding author Received 12 December 2000. Accepted in revised form 24 January 2002 Key words: carbon mineralization, elevated CO 2 , gley soil, litter decomposition, nitrogen mineralization Abstract Elevated concentrations of atmospheric CO 2 can influence the relative proportions, biomass and chemical com- position of plant species in an ecosystem and, thereby, the input of litter nutrients to soil. Plant growth under elevated CO 2 appears to have no consistent effect on rates of litter decomposition; decomposition can, however, differ in C3 and C4 plant material from the same CO 2 environment. We here describe the decomposability of leaf litter of two grass species – the C3 Holcus lanatus L. (Yorkshire fog) and C4 Pennisetum clandestinum Hochst. (kikuyu) - from an unfertilized, ungrazed grassland at a cold CO 2 spring in Northland, New Zealand. Decomposability was measured by net CO 2 –C production from litter incubated for 56 days at 25 C in a gley soil from the site; net mineral-N production from litter was also determined. Both litter and soils were sampled under ‘low’ and ‘high’ concentrations of atmospheric CO 2 . Decomposition of H. lanatus litter was greater than that of P. clandestinum litter throughout the 56-day incubation. Decomposition tended to be greater in ‘high-CO 2 ’ than in ‘low-CO 2 H. lanatus litter, but lower in ‘high-CO 2 ’ than ‘low-CO 2 P. clandestinum litter; differences were, however, non-significant after 28 days. Overall, litter decomposition was greater in the ‘low-CO 2 ’ than ‘high-CO 2 soil. Differences in decomposition rates were related negatively to litter N concentrations and positively to C:N ratios, but were not predictable from lignin:total N ratios. Net mineral-N production from litter decomposition did not differ significantly in ‘high-CO 2 ’ and ‘low-CO 2 ’ samples incubated in ‘low-CO 2 ’ soil; in ‘high-CO 2 ’ soil some net immobilization was observed. Overall, results indicate the likely complexity of litter decomposition in the field but, nevertheless, strongly suggest that rates of decomposition will not necessarily decline in a ‘high-CO 2 environment. Introduction The continuing increase in atmospheric CO 2 concen- trations (IGBP Terrestrial Carbon Working Group, 1998) is likely to have a marked influence on terrestrial ecosystems, in which soils contain about 70% of the total C pool (Schimel, 1995). The effects of elevated CO 2 in these ecosystems can be varied, and include changes in plant species composition or dominance ( ˚ Agren et al., 1991; Newton et al., 1995), growth above and below ground (Newton et al., 1995; Rogers FAX No: +6463559230. E-mail: [email protected] et al., 1994), and associated effects on organic matter inputs to soil (Tate and Ross, 1997). These, in turn, can affect soil biological populations (Sadowsky and Schortemeyer, 1997), biochemical processes (Hungate et al., 1997; Paterson et al., 1997), and eventually C storage (Canadell et al., 1996; McGuire et al., 1997). Changes in the amounts and quality of litter from plants under elevated CO 2 are key factors influencing decomposition rates and the input of plant material to soil (Ball, 1997). Decomposition of litter from dif- ferent atmospheric CO 2 environments can vary, with increases (Taylor and Ball, 1994), decreases (Cotrufo et al., 1998; Lutze et al., 2000) or no change (Hirschel et al., 1997; Norby and Cotrufo, 1998) under elev-

Decomposability of C3 and C4 grass litter sampled under different concentrations of atmospheric carbon dioxide at a natural CO2 spring

Embed Size (px)

Citation preview

Plant and Soil 240: 275–286, 2002.© 2002 Kluwer Academic Publishers. Printed in the Netherlands.

275

Decomposability of C3 and C4 grass litter sampled under differentconcentrations of atmospheric carbon dioxide at a natural CO2 spring

D. J. Ross1,3, K. R. Tate1, P. C. D. Newton2 & H. Clark2

1Landcare Research, Private Bag 11052, Palmerston North, New Zealand. 2AgResearch Grasslands, Private Bag11008, Palmerston North, New Zealand. 3Corresponding author∗

Received 12 December 2000. Accepted in revised form 24 January 2002

Key words: carbon mineralization, elevated CO2, gley soil, litter decomposition, nitrogen mineralization

Abstract

Elevated concentrations of atmospheric CO2 can influence the relative proportions, biomass and chemical com-position of plant species in an ecosystem and, thereby, the input of litter nutrients to soil. Plant growth underelevated CO2 appears to have no consistent effect on rates of litter decomposition; decomposition can, however,differ in C3 and C4 plant material from the same CO2 environment. We here describe the decomposability ofleaf litter of two grass species – the C3 Holcus lanatus L. (Yorkshire fog) and C4 Pennisetum clandestinumHochst. (kikuyu) - from an unfertilized, ungrazed grassland at a cold CO2 spring in Northland, New Zealand.Decomposability was measured by net CO2–C production from litter incubated for 56 days at 25 ◦C in a gley soilfrom the site; net mineral-N production from litter was also determined. Both litter and soils were sampled under‘low’ and ‘high’ concentrations of atmospheric CO2. Decomposition of H. lanatus litter was greater than that ofP. clandestinum litter throughout the 56-day incubation. Decomposition tended to be greater in ‘high-CO2’ thanin ‘low-CO2’ H. lanatus litter, but lower in ‘high-CO2’ than ‘low-CO2’ P. clandestinum litter; differences were,however, non-significant after 28 days. Overall, litter decomposition was greater in the ‘low-CO2’ than ‘high-CO2’soil. Differences in decomposition rates were related negatively to litter N concentrations and positively to C:Nratios, but were not predictable from lignin:total N ratios. Net mineral-N production from litter decompositiondid not differ significantly in ‘high-CO2’ and ‘low-CO2’ samples incubated in ‘low-CO2’ soil; in ‘high-CO2’ soilsome net immobilization was observed. Overall, results indicate the likely complexity of litter decomposition inthe field but, nevertheless, strongly suggest that rates of decomposition will not necessarily decline in a ‘high-CO2’environment.

Introduction

The continuing increase in atmospheric CO2 concen-trations (IGBP Terrestrial Carbon Working Group,1998) is likely to have a marked influence on terrestrialecosystems, in which soils contain about 70% of thetotal C pool (Schimel, 1995). The effects of elevatedCO2 in these ecosystems can be varied, and includechanges in plant species composition or dominance(Agren et al., 1991; Newton et al., 1995), growthabove and below ground (Newton et al., 1995; Rogers

∗ FAX No: +6463559230.E-mail: [email protected]

et al., 1994), and associated effects on organic matterinputs to soil (Tate and Ross, 1997). These, in turn,can affect soil biological populations (Sadowsky andSchortemeyer, 1997), biochemical processes (Hungateet al., 1997; Paterson et al., 1997), and eventually Cstorage (Canadell et al., 1996; McGuire et al., 1997).

Changes in the amounts and quality of litter fromplants under elevated CO2 are key factors influencingdecomposition rates and the input of plant material tosoil (Ball, 1997). Decomposition of litter from dif-ferent atmospheric CO2 environments can vary, withincreases (Taylor and Ball, 1994), decreases (Cotrufoet al., 1998; Lutze et al., 2000) or no change (Hirschelet al., 1997; Norby and Cotrufo, 1998) under elev-

276

ated CO2 having been reported. Decomposition ratesof litter from C3 and C4 plants under the same CO2concentrations can also vary and be more rapid in theC3 material (Ball and Drake, 1997; Kemp et al., 1994).

In many experiments, the same ‘soil’ is used forcomparing the decomposition of plant material fromambient and elevated-CO2 environments. However,the effects of elevated CO2 on plant decompositioncould be influenced by soil properties, which can differappreciably under different long-term concentrationsof atmospheric CO2 (Newton et al., 2001; Norby et al.,2001; Ross et al., 2000). This would apply particularlyto the decomposer communities which, if adapted to aparticular quality of litter, may require time to adaptto a changed litter quality (Arp et al., 1997; Sowerbyet al., 2000b). An effect of soil from different CO2environments on decomposition rates might, therefore,be expected (Norby et al., 2001). This has been ob-served with Quercus cerris L. litter which decomposedmore slowly in the F layer of the forest floor under a‘high-CO2’ than ‘low-CO2’ environment (Gahrooee,1998a).

Generally, decomposition experiments are under-taken with material from plants that have been grownunder elevated CO2 for, at most, a few years. Re-cent studies have, however, included litter from plantsgrowing in the vicinity of natural CO2 springs, whereexposure to high concentrations of atmospheric CO2can be for at least several decades. Litter quality anddecomposition rates of Smilax aspera L., a Mediter-ranean C3 vine, and decomposition of Q. cerris litterwere unaffected by long-term exposure to elevatedCO2 at Laiatico mineral spring (Gahrooee, 1998b).In contrast, decomposition of litter of the C3 grassHolcus lanatus L. and, to a lesser extent, C4 grass Pen-nisetum clandestinum Hochst. Ex Chiov. was fasterin samples taken under elevated than under ‘normal’concentrations of atmospheric CO2 at a New ZealandCO2 spring (Sowerby et al., 2000a). We here extendthe work of Sowerby et al. (2000a), using additionalsamples of leaf litter and soil taken under ‘low’ and‘high’ concentrations of atmospheric CO2 from thesame New Zealand spring.

In light of the above studies, we hypothesised thatlitter decomposition would be: (1) greater in H. lan-atus than in P. clandestinum samples, (2) greater insamples from ‘high-CO2’ than ‘low-CO2’ environ-ments, and (3) lower in ‘high-CO2’ than in ‘low-CO2’soil. We also hypothesised that decomposability wouldbe positively related to the N concentration of the litter.Decomposability was assessed by net C mineralization

in systems incubated in the laboratory under standard-ized conditions; net N mineralization was determinedsimilarly.

Materials and methods

Site and soils

Descriptions of the site with a naturally occurring coldCO2 spring (Hakanoa Springs), located at 35◦ 40

′S

and 174◦ 16′

E, have been given by Newton et al.(1996) and Ross et al. (2000). Mean annual rainfallis about 1500 mm and mean annual temperature 15.5◦C. The almost flat site was undrained, ungrazed andunfertilized. The vegetation comprised a variety ofgrasses, including the C3 grass H. lanatus (Yorkshirefog) and C4 grass P. clandestinum (kikuyu grass), to-gether with some rushes (Juncus spp.) and C3 herbs(Newton et al., 1996).

Both a gley soil (Fluvaquent) and an organic soil(Histosol) (Soil Survey Staff, 1996) occurred at thesite; their distribution is shown by Ross et al. (2000).The organic soil was closest to the main CO2 vents;atmospheric CO2 concentrations above this soil wereconsequently mainly very high and well above pro-jected climate-change values (IPPC, 1996). Only datafor the gley soil are presented here. Mean atmo-spheric CO2 concentrations, measured ca. 20 cmabove ground level on five occasions above the gleysoil, ranged from 372 to 670 µL L−1 at 17 locations,and from 870 to 1490 µL L−1 at two others. Al-though coefficients of variation at the different meas-urement times averaged 26%, our measurements ofatmospheric CO2 gave a biologically meaningful es-timate of concentrations experienced by the plants atthe different sampling locations (Ross et al., 2000).H2S was generally not detected by smell, but verylow atmospheric concentrations have previously beenreported (Ross et al. (2000).

Sampling

Samples of the gley soil were taken in autumn (March1997) as 5–8 multiple cores (25 mm diam, 0–50 mmdepth) around a marker peg at each of several locationsin the transects described by Ross et al. (2000). Thesesamples were then pooled into two groupings, namely‘low-CO2’ and ‘high-CO2’ soil, to test the effects ofsoil environment on decomposition rates. Four loca-tions were used for the ‘low-CO2’ soil, and eight forthe ‘high-CO2’ soil; atmospheric CO2 concentrations

277

Table 1. Some properties of the pooled autumn samples of the gley soilgroupings occurring under different concentrations of atmospheric CO2;average values were 400 (range 372–448) µL L−1 for the ‘low-CO2’ soiland 644 (range 543–670) µL L−1 for the ‘high-CO2’ soil. Data are means,with SDs in parentheses, of duplicate analyses for moisture, organic C andtotal N and of 3 or 4 replicates for the other properties; for each property,significant differences between the soil groupings are indicated by ∗ = P<

0.05

Property ‘Low-CO2’ soil ‘High-CO2’ soil

Moisture (g kg−1) 830 (6) 840 (25)

pH 5.4 (0.1)∗ 5.1 (0.1)

Organic C (g kg−1) 125 (2) 207 (4)∗Total N (g kg−1) 8.8 (0.1) 14.9 (0.1)∗C/N 14 (0) 14 (0)

Microbial C (g kg−1) 1.44 (0.05) 5.48 (0.24)∗Microbial N (g kg−1) 0.24 (0.02) 0.76 (0.06)∗Min-N (µg g−1) 78 (1)∗ 51 (1)

CO2–C (g kg−1: 0–56 days) 2.81 (0.12) 12.0 (1.9)∗� Min-N (µg g−1: 0–56 days) 195 (15) 414 (53)∗0–120 days 349 (30) 1720 (530)∗

above these locations are given in Table 1. Additionalsamples were taken similarly in spring 1997 and 1998.At each sampling time, the pooled samples for eachsoil grouping were sieved (<5.6 mm), mixed thor-oughly and stored at 4 ◦C for 2–4 weeks before usein the incubation experiments. Sub-samples were air-dried and ground (< 0.25 mm) for measurements oforganic C and total N.

Depending upon its availability, leaf litter wascollected at the above times as visually similar stand-ing dead material from plants of H. lanatus and ofP. clandestinum at one or two locations under ‘low’and ‘high’ concentrations of atmospheric CO2, as in-dicated in Table 2. The samples of fully senescedleaves at each sampling time were air dried and ground(<1 mm mesh) in a Casella mill (Gallenkamp, U. K.)or Cyclone Sample Mill (UDY Corporation, U. S. A.).

Analytical methods

Results are expressed on the basis of oven-dry(105◦C) weight of material and are means of duplicate ortriplicate measurements.

Procedures used for determining the soil proper-ties shown in Table 1 are described by Ross et al.(2000). Litter organic C and total N concentrationswere measured by combustion in a Leco FP 2000 ana-lyser. ‘Acid-detergent’ lignin was determined by themethod of Rowland and Roberts (1994).

Litter decomposition was determined in triplicatewith samples of litter and soil collected at the sametime. Litter (in the proportions of 2.0% H. lanatusand 4.0% P. clandestinum on a litter:soil oven-dry(105 ◦C) weight basis) was mixed with 5.0 g soiladjusted to 60% of water-holding capacity, and incub-ated in biometer flasks at 25 ◦ C; soil without addedlitter was incubated under the same conditions. Pre-liminary experiments with ‘low-CO2’ soil showed nosignificant difference in net CO2–C production, on adry weight basis, in this system with either 2.0 or4.0% of P. clandestinum litter; the higher amount wastherefore subsequently used to increase the accuracyof measuring net CO2–C production from this relat-ively slowly decomposing litter. CO2–C productionwas measured at intervals, as described by Barthaand Pramer (1965), with flasks being aerated, withoutadjustment of moisture content, at about 2-weeklyintervals. Measurements were terminated at 56 daysbecause of increasing variability among replicates inseveral of the systems, and decreasing accuracy inmeasuring the differences between the litter-amendedand unamended soil samples. Net CO2–C productionfrom litter decomposition at each sampling time wascalculated as the difference between CO2–C producedby soil plus litter and by soil.

Net mineral N (min-N = NH4+–N + NO3

−–N)production (� min-N) was measured using the abovequantities of litter and moistened soil incubated at 25

278

Table 2. Total C and N and lignin concentrations of litter under different concentrations of atmospheric CO2 at differentsampling times; data are means, with SDs in parentheses, of duplicate analyses. For each litter and sampling time,significant differences between ‘low‘ and ‘high‘ CO2 samples are indicated by ∗ = P< 0.05

Litter Sampling Atmospheric CO2 Total C Total N C:N Lignin Lignin/

time (mean µL L−1) (g kg−1) (g kg−1) (g kg−1) total N

H.lanatus Autumn ‘Low‘ (480) 464 (6) 13.5 (0.1)∗ 34 (1) ND ND

’97 ‘High‘ (1090) 467 (4) 11.8 (0.1) 40 (0)∗ ND ND

Spring ‘Low‘ (404) 401 (4) 19.0 (0.4∗) 21 (1) 67† 3.5†

’97 ‘High‘ (1530) 451 (1)∗ 13.1 (0) 34 (0)∗ 84† 6.4†

P. clandestinum Autumn ‘Low‘ (426) 465 (5) 15.8 (0.4) 29 (0)∗ 70 (8) 4.4 (0.5)

’97 ‘High‘ (763) 476 (11) 18.8 (0.4)∗ 25 (1) 100 (8) 5.3 (0.4)

Spring ‘Low‘ (350) 493 (13) 25.9 (1.0) 19 (0) 89† 3.4†

’98 ‘High‘ (700) 500 (10) 27.1 (0.6) 18 (0) 90† 3.3†

ND, not determined.† Analyses not replicated.

◦C in 125 mL polypropylene containers; soil withoutadded litter was incubated similarly. The containerswere covered with polyethylene and placed in plastictrays containing water, that were then enclosed inlarge polythene bags to maintain high humidity. In-cubation was for 120 days to increase the likelihoodof producing net N mineralization, rather than net Nimmobilization, in the added litter. Mineral-N was ex-tracted with 2 M KCl at 0 and 120 days and determinedas described by Ross et al. (2000). Net N mineralizedfrom litter decomposition after 120 days was calcu-lated as the difference between net N mineralized inthe systems with and without added litter.

Statistics

The significance of differences in initial propertiesof the ‘high-CO2’ and ‘low-CO2’ soil groupings wasdetermined by t-tests (Table 1).

Litter decomposition data are presented for the au-tumn samples as cumulative net CO2–C productionvalues, with individual points at the different incub-ation times being the means of triplicate samples.Because sequential points in the cumulative curveswere derived from the same initial samples, standardrepeated measures ANOVA would be misleading foranalysing time trends in these data (Taylor and Parkin-son, 1988). We therefore selected the approach usedby Gorrisen and Cotrufo (2000) and relied on graphsto describe time trends informally. The effects of plantspecies, litter status (‘low-CO2’ and ‘high-CO2’) andsoil (‘low-CO2’ and ‘high-CO2’) on decompositionrates were assessed by a 3-factor ANOVA, using theGLM Estimate Model procedure (SYSTAT, 1996),

with separate analyses made at the different incuba-tion times. Log transformations were used to ensurethat normality and constant variances held. Least sig-nificant differences (LSD) were determined for eachincubation time from the ANOVAs; these were re-transformed to proportional changes from which theplotted LSDs were calculated.

Net N mineralization in the different litter samplesafter an incubation period of 120 days in the differentsoil groupings was also analyzed by ANOVA.

For each litter species and soil grouping, rela-tionships of CO2–C and of net min-N productionfrom litter decomposition with litter properties weredetermined by linear regression analysis (SYSTAT,1996). The data used were mean values from foursets of samples, viz. autumn and spring samples takenunder both ‘low-CO2’ and ‘high-CO2’ environments.

Results

Soil and litter properties

Some properties of the autumn samples of the soilgroupings used for the decomposition experimentswith the autumn litter are shown in Table 1; sim-ilar values were found in the spring samples (datanot shown). Except for pH and moisture and min-N concentrations, property values were higher (P<

0.05) in the ‘high-CO2’ than ‘low-CO2’ soil, espe-cially for CO2–C (0–56 days) and net min-N (0–120days) production.

Concentrations of total N were lower in ‘high-CO2’ than in ‘low-CO2’ H. lanatus litter, but higherin ‘high-CO2’ than ‘low-CO2’ P. clandestinum litter

279

(Table 2). The litter C:N ratios in these two speciesconsequently differed in the ‘high-CO2’ and ‘low-CO2’ environments. Most of the differences in C andN concentrations and C:N ratios between samplingtimes were also significant for each species (P<

0.05; data not shown). Lignin concentrations andlignin:total N ratios were similar in the ‘high-CO2’and ‘low-CO2’ litter samples of each species.

Effects of plant species and ‘low-CO2’ and‘high-CO2’ litter and soils on litter decomposition

Rates of CO2–C production by unamended soil pre-dictably declined with time. For example, in theautumn samples after 1, 7 and 56 days incubation,they averaged 4.2, 3.2 and 1.5 µg CO2–C g−1 h−1

in the ‘low-CO2’ soil and 16.2, 11.5 and 7.6 µg CO2–C g−1 h−1 , respectively, in the ‘high-CO2’ soil. Insoil plus added litter, the corresponding values weremarkedly higher, averaging 14.8. 10.4 and 3.2 µgCO2–C g−1 h−1 in the ‘low-CO2’ soil systems and26.6, 15.9 and 9.0 µg CO2–C g−1 h−1 in the ‘high-CO2’ soil systems. The difference between CO2–Cproduction in the systems with and without added lit-ter was attributed to CO2–C originating from litterdecomposition.

Cumulative CO2–C production values from de-composition of the autumn samples of litter duringthe 56-day incubation are shown in Figure 1, and thestatistical effects of the different variables are summar-ised in Table 3. Plant species had the most consistentinfluence, with decomposition invariably greater inH. lanatus than in P. clandestinum litter. Differencesbetween litter from ‘low-CO2’ and ‘high-CO2’ envir-onments were significant during the early stages ofthe incubation, but were non-significant after 28 days.The interactions between plant species and litter statuswere, however, significant throughout the incubation,with decomposition tending to be greater in ‘high-CO2’ than in ‘low-CO2’ H. lanatus litter but lower in‘high-CO2’ than in ‘low-CO2’ P. clandestinum litter(Figure 1). Overall, the soil used also had a signific-ant influence over most incubation periods (Table 3),with decomposition greater in the ‘low-CO2’ thanin the ‘high-CO2’ soil (Figure 1); in ‘high-CO2’ P.clandestinum litter, however, none of the between-soil differences was significant (data not shown). Eachplant species responded similarly in the different soils,with only one of the interactions being significant(Table 3).

Figure 1. Influence of autumn-litter status (viz. ‘low-CO2’ and‘high-CO2’ litter) on litter decomposition in ‘low-CO2’ soil and in‘high-CO2’ soil, as measured by net cumulative CO2–C production.Data are means of triplicate samples; LSD values (at P = 0.05) aregiven by the vertical bars to indicate the significance of differencesbetween the ‘low-CO2’ and ‘high-CO2’ litter decomposition valuesfor each plant species and soil.

In the spring (September 1997) samples of H. lan-atus litter, patterns of decomposition were similarto those in Figure 1; in the spring (October 1998)samples of P. clandestinum litter, ‘high-CO2’ litteragain decomposed more slowly than ‘low-CO2’ litter,with decomposition, however, being almost identicalin both soil groupings (data not shown).

Effects of litter properties on rate of decomposition

Although the patterns of decomposition of the differ-ent litter species were similar in the autumn and springsamples, the amounts of C decomposed differed withsampling time. In the system with ‘low-CO2’ litterand ‘low-CO2’ soil, 42% of H. lanatus litter C and37% of P. clandestinum litter C was metabolised toCO2–C after 56 days in the autumn samples, but only28 and 18%, respectively, in the spring samples. Inthe system with ‘high-CO2’ litter and soil, the corres-ponding values were 47 and 25%, respectively, for theautumn samples and 27% and 14% respectively, forthe spring samples. In each plant species, CO2–C pro-duction from litter decomposition after 56 days was

280

Table 3. P-values from the analysis of variance of the effects of plant species, litterstatus (‘low- CO2’ and ‘high-CO2’), and soil (‘low-CO2’ and ‘high-CO2’), and theirinteractions, on cumulative CO2–C production (mg g−1 litter) at different incubationtimes; at each time, df values were 1 for the main effects and their interactions and16 for the residual. The interactions of litter status and soils and of litter status, plantspecies and soils were all non-significant (P> 0.123 and 0.313, respectively) and arenot reported

Incubation Plant species Litter Soil Litter status × Soil ×period status species species

(days) interaction interaction

1 < 0.001 0.002 0.610 < 0.001 0.283

7 < 0.001 0.009 < 0.001 < 0.001 0.020

14 < 0.001 0.002 < 0.001 0.002 0.126

21 < 0.001 0.012 < 0.001 0.003 0.299

28 < 0.001 0.127 0.002 0.002 0.336

35 < 0.001 0.404 0.007 0.002 0.315

43 < 0.001 0.779 0.028 0.002 0.416

49 < 0.001 0.967 0.067 0.004 0.488

56 0.001 0.810 0.096 0.010 0.533

Table 4. Correlations (r values; n = 4) between litter CO2–C production, after decom-position for 56 days in ‘low-CO2’ and ‘high-CO2’ soil, and litter properties; CO2–Cvalues used were from ‘low-CO2’ and ‘high-CO2’ litter sampled in autumn ’97 andspring ’97 (data not shown) for H. lanatus and autumn ’97 and spring ’98 (data notshown) for P. clandestinum; ∗ = P< 0.05; nd, not determined

Litter Soil Litter property

Total N C:N Lignin Lignin:total N

H. lanatus ‘Low-CO2’ –0.973∗ 0.996∗ nd nd

‘High-CO2’ –0.847 0.915 nd nd

P. clandestinum ‘Low-CO2’ –0.962∗ 0.981∗ –0.639 0.625

‘High-CO2’ –0.980∗ 0.975∗ –0.429 0.777

related negatively to litter N concentrations and pos-itively and significantly to litter C:N ratios (Table 4).Relationships of CO2–C production from P. clandes-tinum litter with lignin concentrations and lignin:totalN ratios were all non-significant.

Effects of litter status (‘low-CO2’ and ‘high-CO2’litter) on net N mineralization

Patterns of net min-N production from litter decom-position tended to be inconsistent in the different soiland litter systems, with net N immobilization occur-ring in several of the ‘high-CO2’ soil systems afterincubation for 120 days (Table 5). Litter net N min-eralization did not differ significantly in any of the

H. lanatus systems, and in the P. clandestinum sys-tems differed significantly (P< 0.05) only betweensoil groupings with added ‘low-CO2’ autumn litter(Table 5).

Net N mineralization in H. lanatus or P. clandes-tinum litter from the different CO2 environments, afterincubation for 120 days in the different soil groupings,was not closely related (P> 0.30) to either the initialtotal N concentration of the litter or to its C:N ratio.Although lignin concentrations did not differ amongtreatments (Table 2), rates of net N mineralization in‘high-CO2’ and ‘low-CO2’ P. clandestinum litter weresignificantly correlated with initial lignin concentra-tions after incubation of the samples in the ‘high-CO2’soil (P< 0.05), but not in the ‘low-CO2’ soil (P=0.38).

281

Table 5. Net min-N production from ‘low-CO2’ and ‘high-CO2’ litter incubated for 120 din soil sampled under ‘low‘ and ‘high‘ concentrations of atmospheric CO2; data are means,with SD in parentheses, of triplicate samples. For each set of samples, the only significantdifference (P< 0.05) between ‘high-CO2’ and ‘low-CO2’ litter or soil was between soilgroupings for the ‘low-CO2 ‘ autumn sample of P. clandestinum litter

Litter Sampling Litter sample Net min-N produced (g kg−1 litter)

time

‘Low-CO2’ soil ‘High-CO2’ soil

H. lanatus Autumn ‘Low-CO2’ 3.0 (0.2) 2.2 (16.0)

’97 ‘High-CO2’ 2.2 (3.7) 6.8 (16.0)

Spring ‘Low-CO2’ 1.6 (1.1) –4.2 ( 9.4)

’97 ‘High-CO2’ 1.0 (3.0) –11.3 ( 7.8)

P. clandestinum Autumn ‘Low-CO2’ 0 (1.6) –8.8 ( 2.0)

’97 ‘High-CO2’ 1.7 (0.8) 4.3 ( 7.7)

Spring ‘Low-CO2’ 3.1 (2.2) –1.9 ( 3.5)

’98 ‘High-CO2’ 1.1 (1.7) –0.8 ( 2.5)

Discussion

Our data confirm the results of Sowerby et al. (2000a)in showing that the decomposition of leaf litter fromthe C3 grass, H. lanatus, was more rapid than thatof litter from the C4 grass, P. clandestinum. Theyalso show that decomposition was generally slowerin mineral soil from a ‘high-CO2’ than from a ‘low-CO2’ environment, and are thus similar to the resultsof Gahrooee (1998a) using Q. cerris litter in theorganic layer of a forest soil. The influence of atmo-spheric CO2 concentrations during plant growth onsubsequent litter decomposability was, however, morecomplex than that found by Sowerby et al. (2000a),during an 18-day incubation, and was dependent onboth species and incubation times. Elevated CO2 ef-fects were most pronounced during the early stagesof decomposition, and were generally non-significantafter 56 days; overall, decomposition of H. lanatus lit-ter tended to be somewhat greater, and decompositionof P. clandestinum litter lower, when harvested from a‘high-CO2’ than from a ‘low-CO2’ environment.

Although the use of soil from this grassland eco-system for the decomposition experiments was real-istic in terms of field conditions, the high soil C andN mineralizing activities in unamended soil did cre-ate problems in quantifying net C mineralization, andespecially net N mineralization, in the added litter.As in other studies without isotopically labelled added

substrates, no account could be taken of any possiblepriming or conserving effect of the added litter onsoil organic matter decomposition (Jenkinson, 1981;Kuzyakov et al., 2000; van Ginkel et al., 2000); theuse of naturally occurring 13C as a tracer for followingthe decomposition of litter from the high atmospheric-CO2 locations was not practicable because of thesimilarity of the 13C signature of CO2 emanating fromthe spring (–9.6 ‰) and of ambient CO2 (about –7‰) (G L Lyon, pers. comm.). However, the differ-ences in CO2–C production between soil with andwithout added litter were large and likely to have beenmuch greater than any priming effect. The similarity ofnet CO2–C production in the preliminary experimentswith 2-fold differences in added P. clandestinum litteralso suggests that any priming effect would have beensmall (Wu et al., 1993). Even if priming effects hadoccurred, our experimental procedure is still relevantfor indicating differences in the return of CO2–C tothe atmosphere from the soil plus litter systems un-der different concentrations of atmospheric CO2 in thefield.

Influence of plant species on litter decomposition

Based on our preliminary experiments and other stud-ies (Pinck and Allison, 1951), the different amountsof H. lanatus and P. clandestinum litter used in theincubations should not have influenced these speciescomparisons. However, if the different amounts did

282

have an effect they would most likely have resulted inslightly lower values for the decomposition of H. lan-atus, based on the ryegrass-decomposition experimentof Jenkinson (1977), and so given conservative valuesfor our species differences.

Because our H. lanatus and P. clandestinumsamples could not be taken from exactly the sameatmospheric CO2 environment at this spatially vari-able site (Ross et al., 2000), we cannot draw too finea distinction between these two species, especiallyfor the ‘high-CO2’ samples. However, the similarityof the decomposition patterns of both the ‘low-CO2’and ‘high-CO2’ autumn and spring samples of H.lanatus, which were collected under somewhat differ-ent atmospheric CO2 concentrations, suggests that theactual CO2 concentration was not of critical import-ance. Other factors, such as plant growth conditions,temporal effects and age of the litter, might alsohave influenced our results and been responsible forthe differences found between the autumn and springsamples of ‘low-CO2’ P. clandestinum litter on de-composition in the ‘low-CO2’ and ‘high-CO2’ soils.Although Koukoura (1998) observed that litter decom-position was not invariably slower in C4 than in C3material in a natural grassland, our results have shownthat decomposition of P. clandestinum litter from ‘low-CO2’ or ‘high-CO2’ environments was, at all samplingtimes, consistently lower than that of H. lanatus fromcomparable environments, in spite of any possibleeffects of other field variables.

Litter quality and rates of decomposition

In Quercus spp. from the vicinity of a CO2 spring,prolonged exposure to elevated CO2 had no appar-ent effect on litter C:N and lignin:N ratios (Gahrooee,1998a). Based on this result and other published data,Gahrooee (1998a) considered that elevated CO2 didnot affect litter quality when plants were grown at con-tinued high levels of atmospheric CO2 under realisticfield conditions. Norby et al. (2001) also recordedthat elevated CO2 often had no significant effect ontotal N concentrations in litter from plants grown inopen systems in the field. Norby et al. (2001) did,however, recognise that several sources of variation inthe resorption of N from leaves during the senescenceprocess could affect litter N data. Our results stronglysuggest that the effects of elevated CO2 on litter qual-ity can be species dependent. As in plant material fromsome other C3 and C4 species (Ball, 1997; Ball andDrake, 1997; Frank et al., 1997; Wand et al., 1999),

total N concentrations in both our autumn and springsamples were lower in the C3 (H. lanatus) litter, butunchanged or higher in the C4 (P. clandestinum) litter,under increased concentrations of atmospheric CO2.

Our H. lanatus results support other data (Franket al., 1997; Norby and Cotrufo, 1998) in showingthat reductions in foliar N concentrations in plantsgrown under elevated CO2 need not result in decreasedrates of decomposition of plant material (Lambers,1993). In both H. lanatus and P. clandestinum, de-composition was greater in litter with the lower Nconcentration and was positively related to litter C:Nratios. Without further experimental data, we are un-able to establish the factor(s) responsible for theseobservations. Possibilities include a higher proportionof readily metabolisable carbohydrates, or decreaseddecomposer efficiency and less rapid formation of re-calcitrant compounds (Agren et al., 2001), in the litterwith the higher C:N ratios. In contrast to our data,decomposition rates were unrelated to C:N ratios insorghum, another C4 plant (Taylor and Ball, 1994),and were also unrelated to the C:N ratios of roots ofother grasses after growth under elevated CO2 (vanGinkel et al., 1996; van Ginkel and Gorissen, 1998).Relationships between decomposition of plant mater-ials and their C:N ratios therefore appear to be com-plex and species dependent. We consequently concurwith other observations (Gorissen and Cotrufo, 2000;Sowerby et al., 2000b; van Ginkel et al., 2000) thatthe C:N ratio is probably not an appropriate generalindicator for changes in decomposition rates.

Although we were unable to replicate most ofthe lignin analyses because of a shortage of material,our limited data suggest that litter lignin concentra-tions were siimilar under different CO2 environments.They are thus in agreement with previous studiesin which growth under elevated CO2 likewise hadno pronounced effect on lignin concentrations in thelitter of other C3 (Lutze et al., 2000; Sowerby etal., 2000b) and C4 (Ball, 1997; Kemp et al., 1994)grasses. Further samples are needed to confirm the re-liability of the suggested relationships, derived fromonly four observations (Table 4), between the de-composability of P. clandestinum litter and its ligninconcentration and lignin:total N ratio. Negative rela-tionships between litter decomposition and lignin con-centrations have, however, been previously reported(Thomas and Akazawa, 1993). The positive relation-ships between litter decomposition and the lignin:totalN ratios are, in contrast, somewhat surprising in viewof the negative relationships found for tropical grasses

283

and legumes (Thomas and Akazawa, 1993) and treeleaf litters (Cotrufo et al., 1994). As in other systems(van Ginkel et al., 1996, 2000), more detailed work isnow needed to interpret the decomposition rates of ourgrasses in terms of litter chemical components.

Influence of elevated CO2 and soils on litterdecomposition

As found by Frank et al. (1997), rates of litter decom-position can vary with the species of grass grown un-der elevated CO2 and, as with other species (Hirschelet al., 1997; Mooney et al., 1999; Norby and Cotrufo,1998), were not invariably slower in our ‘high-CO2’than ‘low-CO2’ material (Figure 1). Initial rates of de-composition could, however, be influenced by the soilused (Table 1).

Although decomposition of added plant materialcan be initially slower in acid than in neutral soils(Jenkinson, 1971), differences in pH values in the‘low-CO2’ and ‘high-CO2’ gley soil were small, andunlikely to have been responsible for the different ini-tial rates of litter decomposition in these soil groupings(Figure 2). Soil differences in initial litter decom-position rates are also unlikely to have been causedby nutrient limitations for microbial activity (Recouset al., 1995; Torbert et al., 1998), as nitrogen min-eralized readily in all the samples and ‘available’ Pappeared to be adequate (Ross et al., 2000). Differ-ences in biological populations are, however, likelyand have been shown for the nematode fauna (Yeateset al., 1999) and mycorrhizal infection rate (Rillig etal. 2000) at the Hakanoa site. Preferential metabolismby the decomposer populations of labile soil organicsubstrates, rather than added litter, may also have con-tributed to the slower litter decomposition rates foundin some of the ‘high-CO2’ than ‘low-CO2’ soil sys-tems, with the proportions of labile components beingrelatively high in the ‘high-CO2’ soil (Ross et al.,2000). It is also possible that temporary stabilizationof some of the added litter by native soil organic matterhad occurred (Jenkinson, 1977), with the effect beinggreatest in the ‘high-CO2’ soil, in which the organicmatter concentration was higher than in the ‘low-CO2’soil.

Under field conditions, grass litter would generallydecompose within the area of plant growth, and inLolium perenne was more rapid in an elevated-CO2than ambient-CO2 system (Sowerby et al., 2000b).Our plant and soil data taken together indicate thatdecomposability of H. lanatus litter would be roughly

similar in ‘high-CO2’ and ‘low-CO2’ environments. Incontrast, the lower decomposability of the P. clandes-tinum litter in the ‘high-CO2’ system suggests thatorganic matter could accumulate from this speciesunder elevated CO2, provided plant growth was notsuppressed. Further work is needed to assess the effectof any increase in litter production or accumulation ondecomposition processes (Ball, 1997) and to corrobor-ate these results with other C3 and C4 species. As inother ecosystems (Dukes and Field, 2000; Frank et al.,1997; Gahrooee, 1998a; Norby and Cotrufo, 1998),changes in species dominance or composition underelevated CO2 could well have a major effect on lit-ter decomposition patterns in this grassland ecosystemand so influence the accumulation of soil organic C.

Influence of elevated CO2 and soils on themineralization of litter N

Torbert et al. (2000) found that net N mineralizationin plant material can be soil-dependent and was lowerfor plants grown under elevated CO2 than ambientCO2. In our incubations, net N mineralization was notclosely related to either the origin of the litter or toits initial total N concentration or C:N ratio. Althoughthe absence of consistent differences in net N miner-alization between our different litter samples and soilgroupings could be authentic, it may also be partlyattributable to the nature of the soil. Rates of net Nmineralization in unamended soil were high, particu-larly in the ‘high-CO2’ soil (Table 1), and sometimesvery variable (Ross et al., 2000). Moreover, replicatevariability in the litter samples also tended to be high(Table 5). This overall variability made the assessmentof litter net N mineralization difficult and imprecise,but would not have invalidated the comparisons of lit-ter from the different atmospheric-CO2 environments.Although there appeared to be a positive relationship,in the ‘high-CO2’ only, between litter net N mineraliz-ation and lignin concentration in P. clandestinum litter,we cannot interpret this result on a biochemical basis,and consider that further work is necessary before itsauthenticity can be accepted.

Although net immobilization of N by decomposinglitter was indicated in some samples of the ‘high-CO2’ soil, a loss of mineral-N through denitrification,as a result of litter C additions to this strongly nitri-fying soil (Ross et al., 2000), is also possible. NetN immobilization also occurred when Quercus litterfrom the Mediterranean spring was incubated in soil,with no significant difference found between mater-

284

ial from plants grown under elevated or ambient CO2(Gahrooee, 1998a). Again, further work is needed toestablish more clearly the effects of plant growth underelevated CO2 on the release of N from litter and therole that any changes in N availability may have onecosystem functioning (Cannell and Thornley, 1998).

Conclusions

Our data confirm that rates of litter decomposition canvary with plant species, and be slower in a C4 thanin a C3 grass. Elevated CO2 effects can also varywith plant species in the early stages of decomposi-tion, with increased decomposition in H. lanatus litterand decreased decomposition in P. clandestinum litterfrom a ‘high-CO2’ environment. Decomposition of allH. lanatus litter and of ‘low-CO2’ P. clandestimum lit-ter was greater in ‘low-CO2’ than in ‘high-CO2’ soil.Overall, results highlight the complexity that can befound within a grassland ecosystem at a natural CO2spring and, in this respect, are in accord with the com-plex decomposition responses to simulated climatechange found by Robinson et al. (1997). Results alsosupport the conclusions of Norby and Cotrufo (1998)and Mooney et al. (1999) that growth in a ‘high-CO2’environment does not mean that the decompositionrate of litter will necessarily decline.

We recognise that potential decomposability alonewas measured in our study and that the experimentswere, for practical reasons, relatively short-term. Aninvestigation of litter decomposition in situ over alonger period is now needed to extend these obser-vations, and to incorporate the influence of otherenvironmental variables, including the activity of soilanimals (Coûteaux et al., 1998). More detailed studiesare also required to assess the decomposability of lit-ter in terms of specific concentrations of atmosphericCO2.

Acknowledgements

We thank Jim Mortimer for access to the site, CharlesFeltham and Natasha Rodda for total C and N andmineral-N measurements, Linda Hill and Keitha Gid-dens for lignin determinations, Graeme Lyon, Instituteof Geological and Nuclear Sciences, for the 13C ana-lyses, Greg Arnold and Ray Webster for statisticaladvice, and Surinder Saggar, the subject editor and an

anonymous reviewer for helpful comments on earlierversions of our manuscript.

References

Agren G I, Bosatta E and Magill A H 2001 Combining theory andexperiment to understand effects of inorganic nitrogen on litterdecomposition. Oecologia 128, 94–98.

Agren G I, McMurtrie R E, Parton W J, Pastor J and ShugartH H 1991 State-of-the-art models of production-decompositionlinkages in conifer and grassland ecosystems. Ecol. Applic. 1,118–138.

Arp W J, Kuikman P J and Gorissen A 1997 Climate change: the po-tential to affect ecosystem functions through changes in amountand quality of litter. In Driven by Nature: Plant Litter Qualityand Decomposition. Eds. G Cadisch and K E Giller. pp 187–200.CAB International, Wallingford.

Ball A S 1997 Microbial decomposition at elevated CO2 levels:effect of litter quality. Global Change Biol. 3, 379–386.

Ball A S and Drake B G 1997 Short-term decomposition of litterproduced by plants grown in ambient and elevated atmosphericCO2 concentrations. Global Change Biol. 3, 29–35.

Bartha R and Pramer D 1965 Features of a flask and method formeasuring the persistence and biological effects of pesticides insoil. Soil Science 100, 68–70.

Canadell J G, Pitelka J F and Ingram J S I 1996 The effects ofelevated [CO2] on plant-soil carbon below-ground: a summaryand synthesis. Plant Soil 187, 391–400.

Cannell M G R and Thornley J H M 1998 N-poor ecosystems mayrespond more to elevated [CO2] than N-rich ones in the longterm. A model analysis of grassland. Global Change Biol. 4,431–442.

Cotrufo M F, Ineson P and Rowland A P 1994 Decomposition oftree leaf litters grown under elevated CO2: Effect of litter quality.Plant Soil 163, 121–130.

Cotrufo M F, Briones M J I and Ineson P 1998 Elevated CO2 af-fects field decomposition rate and palatability of tree leaf litter:importance of changes in substrate quality. Soil Biol. Biochem.30, 1565–1571.

Coûteaux M-M, McTiernan K B, Berg B, Szuberla D, DardenneP and Bottner P 1998 Chemical composition and carbon min-eralisation potential of Scots pine needles at different stages ofdecomposition. Soil Biol. Biochem. 30, 583–595.

Dukes J S and Field C B 2000 Diverse mechanisms for CO2 effectson grassland litter decomposition. Global Change Biol. 6, 145–154.

Frank V M, Hungate B A, Chapin F S III and Field C B 1997 De-composition of litter produced under elevated CO2: dependenceon plant species and nutrient supply. Biogeochem. 36, 223–237.

Gahrooee F R 1998a Impacts of elevated atmospheric CO2 on litterquality, litter decomposability and nitrogen turnover rate of twooak species in a Mediterranean forest ecosystem. Global ChangeBiol. 4, 667–677.

Gahrooee F R 1968b Effects of elevated atmospheric CO2 on soilorganic carbon dynamics in a Mediterranean forest ecosystem.PhD thesis, Wageningen, 165 pp.

Gorissen A and Cotrufo M F 2000 Decomposition of leaf and roottissue of three perennial grass species grown at two levels ofatmospheric CO2 and N supply. Plant Soil 224, 75–84.

Hirschel G, Körner Ch and Arnone J A III 1997 Will rising atmo-spheric CO2 affect leaf litter quality and in situ decompositionrates in native plant communities? Oecologia 110, 387–392.

285

Hungate B A, Holland E A, Jackson R B, Chapin F S III, MooneyH A and Field C B 1997 The fate of carbon in grasslands undercarbon dioxide enrichment. Nature 388, 576–579.

IGBP Terrestrial Carbon Working Group 1998 The terrestrial carboncycle: implications for the Kyoto protocol. Science 280, 1393–1394.

IPCC 1996 Summary for Policy Makers. In Climate Change 1995,The Science of Climate Change. Eds. J T Houghton, L G MeiraFilho, B A Callander, N Harris, A Kattenberg and K Maskel. pp3–7. Cambridge University Press, UK.

Jenkinson D S 1971 Studies on the decomposition of C14 labelledorganic matter in the soil. Soil Science 111, 64–70.

Jenkinson D S 1977 Studies on the decomposition of plant materialin soil. IV. The effect of rate of addition. J. Soil Sci. 28, 417–423.

Jenkinson D S 1981 The fate of plant and animal residues in soil. InThe Chemistry of Soil Processes. Eds D J Greenland and M H BHayes. pp 505–561. John Wiley and Sons, Chichester, U K.

Kemp P R, Waldecker D G, Owensby C E, Reynolds J F and Vir-ginia R A 1994 Effects of elevated CO2 and nitrogen fertilizationpretreatments on decomposition of tallgrass prairie leaf litter.Plant Soil 165, 115–127.

Koukoura Z 1998 Decomposition and nutrient release from C3 andC4 plant litters in a natural grassland. Acta Oecologia 19, 115–123.

Kuzyakov Y, Friedel J K and Stahr K 2000 Review of mechanismsand quantification of priming effects. Soil Biol. Biochem. 32,1485–1498.

Lambers H 1993 Rising CO2, secondary plant metabolism,plant-herbivore interactions and litter decomposition. Vegetatio104/105, 263–271.

Lutze J L, Gifford R M and Adams H N 2000 Litter quality anddecomposition in Danthonia richardsonii swards in responseto CO2 and nitrogen supply over four years of growth. GlobalChange Biol. 6, 13–24.

McGuire A D, Melillo J M, Kicklighter D W, Pan Y, Xiao X,Kelfrich J, Moore B III,Vorosmarty C J and Schloss A L 1997Equilibrium responses of global net primary production andcarbon storage to doubled atmospheric carbon dioxide: sens-itivity to changes in vegetation nitrogen concentration. GlobalBiogeochem. Cycles 11, 173–189.

Mooney H A, Canadell J, Chapin F S III, Ehleringer J R, Körner Ch,McMurtrie R E, Parton W J, Pitelka L F and Schulze E-D 1999Ecosystem physiology responses to global change. In The Ter-restrial Biosphere and Global Change: Implications for Naturaland Managed Ecosystems. Eds. B Walker, W Stefan, J Canadelland J Ingram. pp 141–189. Cambridge University Press, U K.

Newton P C D, Bell C C, Clark H 1996 Carbon dioxide emissionsfrom mineral springs in Northland and the potential of these sitesfor studying the effects of elevated carbon dioxide on pastures.N. Z. J. Agric. Res. 39, 33–40.

Newton P C D, Clark H, Bell C C, Glasgow E M, Tate K R, Ross D J,Yeates G W and Saggar S 1995 Plant growth and soil processes intemperate grassland communities at elevated CO2. J. Biogeogr.22, 235–240.

Newton P C D, Clark H, Edwards G R and Ross D J 2001 Experi-mental confirmation of ecosystem model predictions comparingtransient and equilibrium plant responses to elevated atmosphericCO2. Ecol. Lett. 4, 344–347.

Norby R J and Cotrufo M F 1998 A question of litter quality. Nature396, 17–18.

Norby R J, Cotrufo M F, Ineson P, O,Neill E G and Canadell JG 2001 Elevated CO2, litter chemistry, and decomposition: asynthesis. Oecologia 127, 153–165.

Paterson E, Hall J M, Rattray E A S, Griffiths B S, Ritz K and Kill-ham K 1997 Effects of elevated CO2 on rhizosphere carbon flowand soil microbial processes. Global Change Biol. 3, 363–377.

Pinck L A and Allison F E 1951 Maintenance of soil organic mat-ter: III. Influence of green manures on the release of native soilcarbon. Soil Sci. 71, 67–75.

Recous S, Robin D, Darwis D and Mary B 1995 Soil inorganic Navailability: effect on maize residue decomposition. Soil Biol.Biochem. 27, 1529–1538.

Rillig M C, Hernández G Y and Newton P C D 2000 Arbuscular my-corrhizae respond to elevated atmospheric CO2 after long-termexposure: evidence from a CO2 spring in New Zealand supportsthe resource balance model. Ecol. Lett. 3, 475–478.

Robinson C H, Michelsen A, Lee J A, Whitehead S J, Callaghan TV, Press M C and Jonasson S 1997 Elevated atmospheric CO2affects decomposition of Festuca vivipara (L.) Sm. litter androots in experiments simulating environmental change in twocontrasting arctic ecosystems. Global Change Biol. 3, 37–49.

Rogers H H, Runion G B and Krupa S V 1994 Plant responsesto atmospheric CO2 enrichment with emphasis on roots and therhizosphere. Environ. Pollution 83, 155–189.

Ross D J, Tate K R, Newton P C D, Wilde R H and Clark H 2000Carbon and nitrogen pools and mineralization in a grasslandgley soil under elevated carbon dioxide at a natural CO2 spring.Global Change Biol. 6, 779–790.

Rowland A P and Roberts J D 1994 Lignin and cellulose fractionsin decomposition studies using acid detergent fibre methods.Comm. Soil Sci. Plant Anal. 25, 269–277.

Sadowsky M J and Schortemeyer M 1997 Soil microbial responsesto increased concentrations of atmospheric CO2. Global ChangeBiol. 3, 217–224.

Schimel D S 1995 Terrestrial ecosystems and the carbon cycle.Global Change Biol. 1, 77–91.

Soil Survey Staff 1996 Keys to Soil Taxonomy, Seventh Edi-tion. United States Department of Agriculture Natural ResourcesConservation Service, 644 pp.

Sowerby A, Ball A S, Gray T R G, Newton P C D and Clark H2000a Elevated atmospheric [CO2] from a natural soda springaffects the initial mineralization rates of naturally senesced C3and C4 leaf litter. Soil Biol. Biochem. 32, 1323–1327.

Sowerby A, Blum H, Gray T R G and Ball A S 2000b The decom-position of Lolium perenne in soils exposed to elevated CO2:comparisons of mass loss of litter with soil respiration and soilmicrobial biomass. Soil Biol. Biochem. 32, 1359–1366.

SYSTAT 1996 SYSTAT 6.0 for Windows: Statistics. SPSS Inc.,Chicago, Il.

Tate K R and Ross D J 1997 Elevated CO2 and moisture effects onsoil carbon storage and cycling in temperate grasslands. GlobalChange Biol. 3, 225–235.

Taylor J and Ball A S 1994 The effect of plant material grown underelevated CO2 on soil respiratory activity. Plant Soil 162, 315–318.

Taylor B and Parkinson D 1988 Respiration and mass loss rates ofaspen and pine leaf litter decomposing in laboratory microcosms.Can. J. Bot. 66, 1948–1959.

Thomas R J and Asakawa N M 1993 Decomposition of leaf litterfrom tropical forage grasses and legumes. Soil Biol. Biochem.25, 1351–1361.

Torbert H A, Prior S A, Rogers H H and Runion G B 1998 Cropresidue decomposition as affected by growth under elevatedatmospheric CO2. Soil Science 163, 412–419.

Torbert H A, Prior S A, Rogers H H and Wood C W 2000 Review ofelevated atmospheric CO2 effects on agro-ecosystems: residue

286

decomposition processes and soil C storage. Plant Soil 224, 59–73.

van Ginkel J H and Gorissen A 1998 In situ decomposition of grassroots as affected by elevated atmospheric carbon dioxide. SoilSci. Soc. Am. J. 62, 951–958.

van Ginkel J H, Gorissen A and Polci D 2000 Elevated atmosphericcarbon dioxide concentration: effects of increased carbon inputin a Lolium perenne soil on microorganisms and decomposition.Soil Biol. Biochem. 32, 449–456.

van Ginkel J H, Gorissen A and van Veen J A 1996 Long-termdecomposition of grass roots as affected by elevated atmosphericcarbon dioxide. J. Environ. Qual. 25, 1122–1128.

Wand S J E, Midgley G F, Jones M H and Curtis P S 1999 Re-sponses of wild C4 and C3 grass (Poaceae) species to elevatedatmospheric CO2 concentrations: a meta-analytic test of currenttheories and perceptions. Global Change Biol. 5, 723–741.

Wu J, Brookes P C and Jenkinson D S 1993 Formation and destruc-tion of microbial biomass during the decomposition of glucoseand ryegrass in soil. Soil Biol. Biochem. 25, 1435–1441.

Yeates G W, Newton P C D and Ross D J 1999 Response of soilnematode fauna to naturally elevated CO2 levels influenced bysoil pattern. Nematology 1, 285–293.

Section editor: G.R. Shaver