10
Oecologia (2009) 161:719–728 DOI 10.1007/s00442-009-1413-y 123 ECOSYSTEM ECOLOGY - ORIGINAL PAPER Solar ultraviolet radiation alters alder and birch litter chemistry that in turn aVects decomposers and soil respiration Titta Kotilainen · Jari Haimi · Riitta Tegelberg · Riitta Julkunen-Tiitto · Elina Vapaavuori · Pedro Jose Aphalo Received: 17 September 2008 / Accepted: 22 June 2009 / Published online: 14 July 2009 © Springer-Verlag 2009 Abstract Solar ultraviolet (UV)-A and UV-B radiation were excluded from branches of grey alder (Alnus incana) and white birch (Betula pubescens) trees in a Weld experi- ment. Leaf litter collected from these trees was used in microcosm experiments under laboratory conditions. The aim was to evaluate the eVects of the diVerent UV treat- ments on litter chemical quality (phenolic compounds, C, N and lignin) and the subsequent eVects of these changes on soil fauna and decomposition processes. We measured the decomposition rate of litter, growth of woodlice (Porcellio scaber), soil microbial respiration and abundance of nema- todes and enchytraeid worms. In addition, the chemical quality of woodlice feces was analyzed. The exclusion of both UV-A and UV-B had several eVects on litter chemis- try. Exclusion of UV-B radiation decreased the C content in litter in both tree species. In alder litter, UV exclusion aVected concentration of phenolic groups variably, whereas in birch litter there were no signiWcant diVerences in phenolic compounds. Moreover, further eVects on microbial res- piration and chemical quality of woodlice feces were apparent. In both tree species, microbial CO 2 evolution was lower in soil with litter produced under exclusion of both UV-A and UV-B radiation when compared to soil with control litter. The N content was higher in the feces of woodlice eating alder litter produced under exclusion of both UV-A and UV-B compared to the control. In addition, there were small changes in the concentration of individual phenolic compounds analyzed from woodlice feces. Our results demonstrate that both UV-A and UV-B alter litter chemistry which in turn aVects decomposition processes. Keywords Alnus incana · Betula pubescens · Leaf litter decomposition · Leaf phenolics · Ultraviolet radiation Introduction Numerous studies aiming to improve our understanding of the eVects of ultraviolet (UV)-B radiation, especially on plants, have been justiWed by the increasing levels of UV-B radiation reaching ground level due to stratospheric ozone depletion. The eVects of UV-B radiation, through plants, on herbivores, soil organisms and decomposition processes have also been well acknowledged (Caldwell et al. 2007). In many ecosystems, including boreal forests, nutrient min- eralization from dead organic matter by soil microbes determines the availability of nutrients to plants, being therefore essential for ecosystem functioning (van der Heijden et al. 2008). Furthermore, litter consumers, micro- arthropods and some macrofauna, aVect nutrient release via physical and to a lesser extent chemical modiWcation of the Communicated by Stefan Scheu. T. Kotilainen · J. Haimi Department of Biological and Environmental Sciences, University of Jyväskylä, P.O. Box 35, 40014 Jyväskylä, Finland Present Address: T. Kotilainen (&) · R. Tegelberg · P. J. Aphalo Department of Biological and Environmental Sciences, University of Helsinki, P.O. Box 65, 00014 Helsinki, Finland e-mail: titta.kotilainen@helsinki.W R. Julkunen-Tiitto Faculty of Biosciences, University of Joensuu, P.O. Box 111, 80101 Joensuu, Finland E. Vapaavuori Finnish Forest Research Institute, Suonenjoki Research Unit, Juntintie 154, 77600 Suonenjoki, Finland

Solar ultraviolet radiation alters alder and birch litter chemistry that in turn affects decomposers and soil respiration

Embed Size (px)

Citation preview

Oecologia (2009) 161:719–728

DOI 10.1007/s00442-009-1413-y

ECOSYSTEM ECOLOGY - ORIGINAL PAPER

Solar ultraviolet radiation alters alder and birch litter chemistry that in turn aVects decomposers and soil respiration

Titta Kotilainen · Jari Haimi · Riitta Tegelberg · Riitta Julkunen-Tiitto · Elina Vapaavuori · Pedro Jose Aphalo

Received: 17 September 2008 / Accepted: 22 June 2009 / Published online: 14 July 2009© Springer-Verlag 2009

Abstract Solar ultraviolet (UV)-A and UV-B radiationwere excluded from branches of grey alder (Alnus incana)and white birch (Betula pubescens) trees in a Weld experi-ment. Leaf litter collected from these trees was used inmicrocosm experiments under laboratory conditions. Theaim was to evaluate the eVects of the diVerent UV treat-ments on litter chemical quality (phenolic compounds, C, Nand lignin) and the subsequent eVects of these changes onsoil fauna and decomposition processes. We measured thedecomposition rate of litter, growth of woodlice (Porcellioscaber), soil microbial respiration and abundance of nema-todes and enchytraeid worms. In addition, the chemicalquality of woodlice feces was analyzed. The exclusion ofboth UV-A and UV-B had several eVects on litter chemis-try. Exclusion of UV-B radiation decreased the C content inlitter in both tree species. In alder litter, UV exclusionaVected concentration of phenolic groups variably, whereas

in birch litter there were no signiWcant diVerences in phenoliccompounds. Moreover, further eVects on microbial res-piration and chemical quality of woodlice feces wereapparent. In both tree species, microbial CO2 evolution waslower in soil with litter produced under exclusion of bothUV-A and UV-B radiation when compared to soil withcontrol litter. The N content was higher in the feces ofwoodlice eating alder litter produced under exclusion ofboth UV-A and UV-B compared to the control. In addition,there were small changes in the concentration of individualphenolic compounds analyzed from woodlice feces. Ourresults demonstrate that both UV-A and UV-B alter litterchemistry which in turn aVects decomposition processes.

Keywords Alnus incana · Betula pubescens · Leaf litter decomposition · Leaf phenolics · Ultraviolet radiation

Introduction

Numerous studies aiming to improve our understanding ofthe eVects of ultraviolet (UV)-B radiation, especially onplants, have been justiWed by the increasing levels of UV-Bradiation reaching ground level due to stratospheric ozonedepletion. The eVects of UV-B radiation, through plants, onherbivores, soil organisms and decomposition processeshave also been well acknowledged (Caldwell et al. 2007).In many ecosystems, including boreal forests, nutrient min-eralization from dead organic matter by soil microbesdetermines the availability of nutrients to plants, beingtherefore essential for ecosystem functioning (van derHeijden et al. 2008). Furthermore, litter consumers, micro-arthropods and some macrofauna, aVect nutrient release viaphysical and to a lesser extent chemical modiWcation of the

Communicated by Stefan Scheu.

T. Kotilainen · J. HaimiDepartment of Biological and Environmental Sciences, University of Jyväskylä, P.O. Box 35, 40014 Jyväskylä, Finland

Present Address:T. Kotilainen (&) · R. Tegelberg · P. J. AphaloDepartment of Biological and Environmental Sciences, University of Helsinki, P.O. Box 65, 00014 Helsinki, Finlande-mail: [email protected]

R. Julkunen-TiittoFaculty of Biosciences, University of Joensuu, P.O. Box 111, 80101 Joensuu, Finland

E. VapaavuoriFinnish Forest Research Institute, Suonenjoki Research Unit, Juntintie 154, 77600 Suonenjoki, Finland

123

720 Oecologia (2009) 161:719–728

decomposing material and the soil environment (Bardgett2005). Limited knowledge about the eVects of UV-B radia-tion on higher trophic levels limits our understanding of theoverall impacts of UV-B radiation on terrestrial ecosys-tems.

Indirect eVects on decomposition may stem fromchanges in plant chemistry induced by UV radiation dur-ing plant growth. Indeed, a few studies have demonstratedthat even though impacts of UV-B radiation on plants mayoften be benign, indirect eVects on decomposers via plantsmagnify the eVects of stratospheric O3 depletion (Johnsonet al. 2002; Pancotto et al. 2003). Impacts of UV-A radia-tion on soil microbes and microinvertebrates in exclusionstudies have also been observed (Convey et al. 2002).Changes induced by UV radiation aVect the concentrationsof lignin, N, carbohydrates, cellulose, tannins and otherphenolic metabolites and the ratios of lignin:cellulose, C:Nand lignin:N; factors which are considered to have animportant role in the decomposition of litter (Paul et al.1999; Bardgett 2005). However, eVects of UV-B radiationon litter chemistry have been demonstrated to be variableand dependent on the system under focus (Rozema et al.1997; Paul et al. 1999; Moody et al. 2001). There is evi-dence suggesting that species-speciWc production of phe-nolic compounds could have a bearing on soil microbialactivity and nutrient cycling at the ecosystem scale, furtherhaving an eVect on competitive interactions of plants(Bardgett 2005). Although secondary metabolites are notthe whole story in terms of ecosystem responses, they areideal candidates for mediating multiple trophic responsesto UV-B radiation. Since the consequences of their induc-tion can be monitored from the molecular/photochemistrylevel through to higher trophic levels, they can be usedto demonstrate cause and eVect mechanisms (Bassman2004).

Both direct eVects of UV-B on chemical composition ofplant litter and direct and indirect eVects on litter decompo-sition have been studied mostly through supply of UV radi-ation using lamp systems. There are some problems withthis approach due to diVerences between the spectra of UVlamps and in solar spectra following the changes in theozone column, since only the shortest wavelengths of thesolar spectrum at ground level are aVected (Mazza et al.2000; Flint and Caldwell 2003a, b). This in part explainsthe varying results of diVerent experiments (Pancotto et al.2003 and references therein). To overcome the problem ofdiVering spectral output, biological spectral weighting func-tions (BSWF) are used to calculate UV doses supplied bylamps. BSWF are derived mainly from action spectra, whichdescribe the relative eVectiveness of radiation of diVerentwavelengths in producing a biological response (Flint andCaldwell 2003a, b). Furthermore, because most research hasbeen centered on the stratospheric ozone depletion issue

the role of solar UV-A radiation on litter chemistry anddecomposition processes remains poorly understood.

The objective of this Weld study was to assess the respec-tive roles of solar UV-A and UV-B bands in the accumula-tion of phenolics and concentrations of C, N, lignin andsugars in grey alder (Alnus incana) and white birch (Betulapubescens) leaf litter. Additionally, we determined whetherthere are diVerences in responses of a diverse micro- andmeso faunal community and soil respiration and macrofa-una, speciWcally woodlice, feeding directly on litter grownunder diVerent UV treatments. Three separate decomposi-tion experiments were performed in the laboratory: twomicrocosm experiments diVering in duration with litter, soiland natural micro- and meso faunal communities and oneexperiment with woodlice [Porcellio scaber (Latreille);Crustacea: Isopoda: Porcellionidae] feeding on litter.

Materials and methods

Field treatments

The Weld experimental set-up is described in detail inKotilainen et al. (2008). BrieXy, we chose eight grey alder(Alnus incana L.) trees and 12 white birch (Betula pubes-cens Ehrh.) trees (3–4 m high) for our experiment in anabandoned (about 25 years ago) agricultural Weld nearJyväskylä (62°16�N, 25°29�E), central Finland. The experi-ment had three UV treatments created by means of plasticWlms: near ambient control (UV-A+, UV-B+; polythene, 04PE-LD; Etola, Jyväskylä, Finland), exclusion of UV-Bradiation (UV-A+, UV-B¡; polyester, 0.125 mm thick,Autostat CT5; Thermoplast, Helsinki, Finland) and exclu-sion of both UV-A and UV-B radiation (UV-A¡, UV-B¡;Rosco E+# 226, surface-dyed polyester, 0.076 mm thick;West Lighting, Tampere, Finland). The three treatmentswere located on diVerent branches of each tree. The Wlterswere installed before bud break, during the Wrst week ofMay 2004 and the experiment continued until the end of thegrowing season. A spectrophotometer equipped with anintegrating sphere (Shimadzu UV-2501 PC UV–VIS;Kyoto, Japan) was used to measure the spectral transmit-tance properties of the Wlters. For the transmittanceproperties and for the estimated mean biologically eVectiveUV doses throughout the experiment, see Kotilainen et al.(2008).

Plant material

For the decomposition experiments and litter analyses weused litter from Wve grey alder trees and from six whitebirch trees. We chose trees in which there was enough litteravailable for the decomposition experiments. Leaf samples

123

Oecologia (2009) 161:719–728 721

for phenolic analyses were collected at the time the leaveswere ready for abscission and dropping (when pulledsoftly), between 6 September and 10 October 2004 depend-ing on the tree. In most cases there were two branches foreach treatment per tree, but in three alders there was onlyone branch per treatment. One leaf was taken from eachbranch near the center of the Wlter; in the case of the threealders with fewer treated branches, two leaves per branchwere taken. After this initial sampling for phenolics analy-ses all the remaining leaves on the branch were collectedfor use in the decomposition experiments. Litter was driedin a ventilated oven (+30°C) for 5 days. Samples for phenolicanalyses were stored at ¡20°C. Samples for N, C, carbo-hydrate and lignin analyses, and litter used in thedecomposition experiments were stored at room temperature.

Litter sample preparation and chemical analyses

For more detailed description of the phenolics sample prep-aration, see Kotilainen et al. (2008). In brief, for the pheno-lics analyses we took two leaves per tree per treatment.These two leaves were pooled for the extraction. A groundsubsample (20 mg) was suspended into 1 ml of methanol[high performance liquid chromatography (HPLC) grade;Rathburn Chemicals, Walkerburn, Scotland, UK] andhomogenized. This extraction step (adding methanol, cen-trifugation, collecting the supernatant) was repeated 5times. Three aliquots were produced from a 2-ml sample(2 £ 500 �l and 800 �l) and dried under nitrogen. Tanninswere measured from the 800-�l aliquot using a butanol–HCl test (Hagerman 1995) with a spectrophotometer at550 nm and quantiWed using standardization with puriWedtannin from leaves of A. incana and B. pubescens collectedfrom the Weld site. Individual phenolics were analyzed byHPLC from the 500-�l aliquots. The HPLC-diode arraydetection analyses were performed as described in Keski-Saari et al. (2005) with an injection volume of 10 �l, in anAgilent 1100 Series (Agilent, Waldbroon, Germany).Retention times and UV spectra were used to identifythe compounds monitored at 220, 270, 320 and 360 nm.Compound identiWcation is described in Keinänen andJulkunen-Tiitto (1998) and Keski-Saari et al. (2005).

Part of the litter collected after initial sampling for phen-olics analyses was ground to a Wne powder using a mill(Fritsch Pulverisette 14; Fritsch, Idar-Oberstein, Germany)Wtted with a 0.5-mm sieve. C and N (two replicates persample, 3 mg each) were analyzed from the powder with anelement analyzer (EA 1110 CHNS-O; CE Instruments,Milan, Italy).

The same milled litter used for nutrient analyses wasanalyzed to determine litter cell wall composition for lig-nin, total hydrolyzable sugars, �-cellulose and uronic acidconcentrations. Milled leaf powder (0.5 g) was extracted in

acetone (150 ml) by the Soxhlet method and according tothe SCAN-CM (1994) standard for gravimetric measure-ment of acetone-soluble extractives and to yield extractive-free samples. These extractive-free samples were used inthe further measurements of gravimetric and acid-solublelignin, uronic acids, total sugars and �-cellulose asdescribed in Anttonen et al. (2002).

Decomposition experiments

Microcosms with soil and litter

Plastic jars were used as microcosms (volume 1,600 ml,gas exchange allowed through a mesh covering a 15-mm-diameter hole in the plastic lid). Thirty-three microcosmsreceived 700 g of soil (collected from the Weld experimentalsite) and 2 g (d.m.) of litter each, together with 20 ml of tapwater to moisten the soil and the litter. Litter from eachtreatment and tree was placed on the surface of the soil ofseparate microcosms. After 3 days, the litter was weighedand half of the moist litter was placed back into the plasticmicrocosms and the other half was placed into other micro-cosms with woodlice (see description of the woodliceexperiment below).

CO2 production (soil respiration) was measured 13 timesduring the experiment with an infrared carbon analyzer(EQ 92 Universal Carbon Analyzer; PPM Systems, Espoo,Finland). For measuring respiration, 1-ml air samples weretaken through the mesh covering the hole in the lid with aclean syringe. The microcosms were sealed with tape for1 h, after which another set of 1 ml air samples were taken.The evolution rate of CO2 was calculated for each jar fromthe increase in concentration, and the exact time betweensamplings.

The litter was weighed every other week, 9 times in all.The whole microcosms were also weighed every otherweek and water was added to restore the initial weight. Themicrocosms were incubated in a climate chamber (+17°C)for 18 weeks. At the end of the experiment, the leaf litterwas dried (+30°C, 5 days) and weighed. Soil pH was mea-sured [5 g of soil (fresh mass (f.m.), in water at 1:5 ratio].Nematodes and enchytraeids were extracted using wet fun-nel methods from 5 to 65 g (f.m.) of soil per microcosm,respectively (O’Connor 1962; Sohlenius 1979). Both nema-todes and enchytraeids were counted alive. In addition,length of enchytraeids was measured and their fresh bio-mass estimated according to Abrahamsen (1973). Duringthe experiment some microcosms were contaminated.Therefore, at the end of the experiment, for microcosmswith alder litter, there were four replicates of both controland exclusion of UV-B radiation treatments and three repli-cates of exclusion of both UV-A and UV-B treatment. Forall the treatments with birch litter, n = 6.

123

722 Oecologia (2009) 161:719–728

To study the long-term eVects of diVerent UV treatmentson litter decomposition, we had another experiment thatcontinued for 13 months. Eighteen plastic microcosmsreceived 700 g of soil (collected from the Weld experimentsite) and 2 g of litter (dry mass; d.m.) each. During theexperiment some microcosms were contaminated andtherefore at the end of the experiment, for all the treatmentswith alder litter n = 3, for microcosms with birch litter n = 3in control and exclusion of UV-B radiation, n = 2 in exclu-sion of both UV-A and UV-B. At the end of this decompo-sition experiment, the same procedures were followed asdescribed above.

Microcosms with woodlice and litter

The woodlouse population used originated from TheNetherlands and was collected from the Weld in October2004. Before the start of the experiment, the animals werereared in glass microcosms and fed with both grey alder andwhite birch leaves collected from the Weld experimental site.

Glass jars with casting made of plaster of Paris andcharcoal at the bottom were used as microcosms (volume500 ml, gas exchange allowed through a mesh covering ahole in the metal lid). Each of the 33 microcosms receivedthree woodlice and litter that had been incubated for 3 daysin plastic microcosms (see above), since woodlice preferlitter infested by microbes (Zimmer and Topp 2000).Again, alder and birch litter were put in separate micro-cosms. Woodlice were weighed before their introductionand all three individuals for one microcosm were treated asa group; the average fresh mass was 97.8 mg § SE17.9 mg with group minimum 51 mg and maximum137 mg. For all the treatments with alder litter n = 5, withbirch litter n = 6.

CO2 production was measured 12 times during theexperiment (see description of the method above). The litterand the woodlice were weighed twice a month, 6 times inall. All the microcosms were incubated in a climate cham-ber (+17°C) for 13 weeks. In comparison, the duration ofthe short-term decomposition experiment with soil andlitter was 18 weeks and CO2 production was measured13 times during the experiment. The microcosms wereweighed every 2 weeks and water was added to replace theamount evaporated. At the end of the experiment litter andwoodlice were weighed. In addition, feces of woodlicewere weighed, dried (+30°C, 5 days) and stored at roomtemperature.

Phenolics were analyzed from woodlice feces collectedat the end of the experiment. The procedure used was thesame as described above for litter, except that sample sizewas 100 mg. Two aliquots were taken from a 2-ml sample(1,000 and 800 �l). The 800-�l aliquot was used for themeasurement of tannins and the 1,000-�l aliquot was used

to analyze for individual phenolics. Similarly as for litter, Cand N also were measured from the feces.

Statistical design and analyses

Treatments were assigned at random to branches withineach tree, each tree considered a complete block for statisti-cal analysis. The data were analyzed by Wtting a linearmixed eVects model, with random eVects for the tree factor.Analysis was performed with R software (R DevelopmentCore Team 2006), using package NLME (Pinheiro andBates 2000). Comparisons between individual pairs oftreatments were done only when the test of overall signiW-cance of the treatments yielded P < 0.05. Package gregmiscwas used to Wt these contrasts (to the model not includingthe interaction). Transformations of the data were not used,but a power of variance function covariate was included inthe model in cases when variances were not homogeneous,as revealed by a likelihood ratio test between models. Nor-mality was checked with quantile–quantile plots of residu-als. For two compounds in litter and for eight compounds inwoodlice feces that were absent from several samples butpresent in others, a generalized linear mixed model wasWtted with function glmmPQL from package MASS(Venables and Ripley 2002) assuming a binomial distribu-tion and treating the presence or absence of the metaboliteas a binary response. For the discussion of results � = 0.05was used as the limit for signiWcance of main eVects andinteractions, whereas for pairwise comparisons � = 0.10was used; P-values from multiple contrasts were adjustedusing Holm’s procedure. Figures were drawn using R withpackage lattice (Sarkar 2008).

Results

Litter chemistry

Phenolics

Alder litter phenolics are grouped into seven phenolicgroups (Table 1). The concentration of cinnamic acids washigher under exclusion of both UV-A and UV-B than in thecontrol (P = 0.063), with a diVerence between UV-exclu-sion treatments (P = 0.015) (Table 1). The concentration ofchlorogenic acids was lower under exclusion of UV-B thanunder the control (P < 0.0001) and higher under exclusionof both UV-A and UV-B than under the control(P < 0.0001), with a diVerence between UV-exclusiontreatments (P < 0.0001) (Table 1). The exclusion of bothUV-A and UV-B decreased the concentration of Xavonoids(P = 0.061), with a diVerence also between UV-exclusiontreatments (P = 0.094) (Table 1). Concentrations of phenolic

123

Oecologia (2009) 161:719–728 723

acids, phenolic glycosides, presumed stilbenes (tentativelyidentiWed) and tannins were not signiWcantly aVected by theUV treatments (Table 1). The concentration of summedphenolics was not signiWcantly aVected by the exclusion ofUV; the mean over all for treatments was 124 mg g¡1 with-out tannins and 184 mg g¡1 with tannins.

Three phenolic groups were identiWed in birch litter;chlorogenic acid, Xavonoids and tannins, and these werenot signiWcantly aVected by the UV treatments (Table 1).The concentration of summed phenolics was not signiW-cantly aVected by the exclusion of UV; means over alltreatments were 15 mg g¡1 without tannins and 103 mg g¡1

with tannins. The results of UV treatment eVects on indi-vidual phenolic compounds in alder and birch collectedduring the growing season are presented in detail in Kotilainenet al. (2008). Results presented above are only from the treesand litter that were used in the decomposition experiments.

C, N and cell wall chemistry

The percentage of N (% of d.m.) was not aVected by theUV treatments, but there was a clear diVerence between thetree species: the percentage was higher in alder comparedwith birch (Table 2). In both tree species, the percentage ofC (% of d.m.) was lower under exclusion of UV-B than inthe control (P = 0.02), with a diVerence between UV-exclu-sion treatments (P = 0.02) (Table 2). The C:N ratio washigher in birch compared to alder, but there were no diVer-ences between the UV treatments (Table 2).

The percentages of total lignin and gravimetric lignin (%of d.m.) were not aVected by the UV treatments, but therewas more gravimetric lignin in birch (Table 2). For the per-centage of acid-soluble lignin, there was an interactionbetween UV treatment and tree species. The percentage of

acid-soluble lignin in alder litter was lower under exclusionof both UV-A and UV-B than in the control (P = 0.002),with a diVerence between UV-exclusion treatments(P = 0.002) (Table 2). In birch litter there were no diVer-ences between treatments (Table 2). The percentages ofuronic acids, �-cellulose and total sugars in alder and birchlitter were not aVected by the UV treatments, percentagesof the Wrst two were higher in birch compared to alder(Table 2).

For the total lignin:�-cellulose ratio there was an interac-tion between treatment and tree species. This ratio in alderlitter was lower under exclusion of UV-B (P = 0.088) andunder exclusion of both UV-A and UV-B than in the con-trol (P = 0.0582) (Table 2). In birch, this ratio was notaVected by the UV treatments (Table 2). The ratio of totallignin:total hydrolyzable sugars was lower under exclusionof both UV-A and UV-B than in the control in both treespecies (P = 0.036, P = 0.023) (Table 2). Total lignin:Nratio was not aVected by the UV treatments (Table 2).

Decomposition experiments

Microcosms with soil and litter

The mean litter mass loss over all treatments was 32.8% ofd.m. in both species. The mean number of nematodes was10 g¡1 soil (f.m.) and the mean biomass of enchytraeidswas 34 �g g¡1 soil (f.m.). Soil pH was higher in micro-cosms with birch litter; the mean over all treatments was5.43, compared to a mean of 5.31 in microcosms with alderlitter. These parameters were not aVected by the UV treat-ments. The cumulative CO2 evolution was lower underexclusion of both UV-A and UV-B (P = 0.046) comparedto the control in both tree species (Fig. 1).

Table 1 Responses of phenolic metabolites by compound group to ultraviolet (UV) radiation in grey alder (A) and white birch (B) litter

For A, n = 5 trees; for B, n = 6 trees. Values sharing the same letter within rows are not signiWcantly diVerent (adjusted P > 0.10)

For all the compound groups in alder UV treatment df = 2 and 8. For all the compound groups in birch UV treatment df = 2 and 10, except fortannins df = 2 and 9

Compound group Species P, F Mean § SE (mg g¡1)

Treatment UV-B+, UV-A+ UV-B¡, UV-A+ UV-B¡, UV-A¡

Phenolic acids A 0.50, 0.752 0.229 § 0.009a 0.252 § 0.078a 0.172 § 0.03a

Cinnamic acids A 0.014, 7.67 1.14 § 0.187a 0.828 § 0.193a 1.79 § 0.246b

Phenolic glycosides A 0.37, 1.14 11.46 § 2.12a 19.64 § 8.53a 8.95 § 1.46a

Presumed stilbenes A 0.91, 0.093 99.6 § 15.66 a 93.0 § 23.89a 89.8 § 18.83a

Chlorogenic acids A <0.0001, 194 6.92 § 1.54 a 4.67 § 1.13b 7.57 § 1.87c

B 0.42, 0.954 2.77 § 0.732a 2.72 § 0.60a 3.83 § 0.932a

Flavonoids A 0.022, 6.44 13.83 § 2.59a 8.72 § 2.07a 4.52 § 0.585b

B 0.57, 0.598 13.01 § 1.27a 11.40 § 1.57a 11.08 § 1.43a

Tannins A 0.58, 0.581 64.6 § 6.11a 54.0 § 7.36a 60.7 § 8.80a

B 0.83, 0.197 89.0 § 5.32a 86.5 § 13.92a 87.8 § 13.54a

123

724 Oecologia (2009) 161:719–728

In the long-term decomposition experiment with soil andlitter (13 months) litter decomposition and the numbers ofnematodes were not signiWcantly aVected by the UV treat-ments (data not shown). The average litter mass loss overall treatments was 45.6% of d.m. The number of nematodeswas on average 9 per 1 g of soil (f.m.). No enchytraeidworms were found in the soil samples. At the end of theexperiment soil pH was slightly altered by the exclusion ofboth UV-A and UV-B (P = 0.04) for both tree species. ThepH ranged from 5.04 in the control to 5.12 under exclusionof both UV-B and UV-A radiation, with intermediatevalues under exclusion of only UV-B.

Microcosms with woodlice and litter

The cumulative CO2 evolution (Fig. 1) and litter loss (datanot shown) were not aVected by the UV treatments. Theaverage litter mass loss over all treatments was 56.2%.Woodlice grew better when feeding on alder litter, but no

signiWcant diVerences were observed between the UV treat-ments. The average growth over all treatments was 38.9%in the microcosms with alder litter and 24.0% in birch.

Feces chemistry

Phenolics

In feces of woodlice fed with alder litter, ten diVerent phe-nolic metabolites were observed, belonging to Wve diVerentgroups according to their chemical properties. Phenolicacids, presumed stilbenes, Xavonoids and tannins were notsigniWcantly aVected by the diVerent UV treatments(Table 3). The concentration of summed phenolics was notsigniWcantly aVected by the exclusion of UV; averages overall treatments were 0.307 mg g¡1 without tannins and2.34 mg g¡1 with tannins. In cinnamic acids there was onlyone metabolite, a P-OH cinnamic acid derivative, with con-centration higher under the exclusion of both UV-A and

Table 2 Responses of diVerent chemical parameters and ratios to UV radiation in grey alder (A) and white birch (B) litter

For A, n = 5 trees; for B, n = 6 trees. Values sharing the same letter within rows are not signiWcantly diVerent (adjusted P > 0.10). d.m.: Dry mass

For all the parameters, UV treatment df = 2 and 18, species df = 1 and 9 and interaction df = 2 and 18

% of d.m. P, F Species Mean § SE

Treatment Species Interaction UV-B+ UV-A+ UV-B¡ UV-A+ UV-B¡ UV-A¡

N 0.24, 1.54 <0.0001, 258 0.065, 3.20 A 2.82 § 0.092a 2.75 § 0.104a 2.67 § 0.094a

B 1.05 § 0.072a 1.10 § 0.067a 1.07 § 0.046a

C 0.017, 5.13 0.0002, 34.71 0.96, 0.05 A 48.75 § 0.261a 48.26 § 0.149b 48.71 § 0.199a

B 46.71 § 0.346a 46.23 § 0.417b 46.72 § 0.288a

Gravimetric lignin 0.17, 1.96 0.026, 7.06 0.98, 0.018 A 29.43 § 0.916a 28.08 § 1.24a 28.77 § 1.77a

B 33.69 § 0.837a 32.36 § 1.17a 32.82 § 1.38a

Acid-soluble lignin <0.0001, 12.74 <0.0001, 29.66 <0.0001, 8.91 A 20.70 § 1.28a 20.60 § 1.30a 18.73 § 1.26b

B 12.75 § 0.624a 13.48 § 0.636a 12.84 § 0.416a

Total lignin 0.11, 2.47 0.059, 4.68 0.47, 0.786 A 50.13 § 0.452a 48.68 § 0.992a 47.50 § 1.14a

B 46.44 § 0.934a 45.84 § 1.18a 45.66 § 1.40a

Uronic acids 0.45, 0.844 0.0004, 29.28 0.58, 0.565 A 3.13 § 0.20a 3.16 § 0.235a 3.19 § 0.374a

B 6.37 § 0.360a 6.13 § 0.513a 6.68 § 0.721a

�-Cellulose 0.93, 0.069 0.0001, 39.58 0.11, 2.49 A 7.68 § 0.176a 8.07 § 0.235a 8.0 § 0.237a

B 10.67 § 0.278a 10.29 § 0.437a 10.46 § 0.377a

Total hydrolyzable sugars 0.68, 0.398 0.10, 3.31 0.79, 0.241 A 20.89 § 0.412a 20.64 § 0.596a 20.46 § 0.397a

B 19.25 § 0.359a 19.44 § 0.507a 19.67 § 0.554a

Ratios

C:N 0.066, 3.17 <0.0001, 98.0 0.24, 1.56 A 17.38 § 0.694a 17.64 § 0.666a 18.34 § 0.772a

B 45.61 § 3.23a 42.85 § 2.69a 43.90 § 1.76a

Total lignin:N 0.87, 0.143 <0.0001, 179 0.49, 0.749 A 17.85 § 0.584a 17.79 § 0.771a 17.90 § 0.938a

B 45.14 § 2.71a 42.37 § 2.43a 42.79 § 1.57a

Total lignin:�-cellulose 0.11, 2.51 0.0001, 47.28 0.023, 4.69 A 6.54 § 0.182a 6.05 § 0.181b 5.94 § 0.10b

B 4.38 § 0.183a 4.51 § 0.274a 4.40 § 0.212a

Total lignin:total hydrolyzable sugars

0.048, 3.62 0.94, 0.006 0.78, 0.251 A 2.46 § 0.051a 2.36 § 0.059ab 2.32 § 0.054b

B 2.42 § 0.080a 2.37 § 0.112ab 2.33 § 0.104b

123

Oecologia (2009) 161:719–728 725

UV-B (P = 0.001), and with a diVerence between the UV-exclusion treatments (P = 0.001) (Table 3).

In feces of woodlice fed with birch litter, nine diVerentphenolic metabolites were observed, belonging to fourdiVerent groups. Concentrations of diVerent compoundgroups (phenolic acids, cinnamic acid, Xavonoids andpresumed stilbenes) were not signiWcantly aVected by the

UV treatments (Table 3). The concentration of summedphenolics was not signiWcantly aVected by the exclusion ofUV; the average over all treatments was 0.664 mg g¡1.Concerning individual metabolites, a presumed stilbenederivative was aVected by UV exclusion; this compoundwas totally absent under the exclusion of both UV-A andUV-B radiation (Fisher test = 0.005) (control, 0.042 §0.016 mg g¡1; exclusion of UV-B, 0.026 § 0.009 mg g¡1).

C and N

Percentage of C (% of d.m.) and C:N ratio in the woodlicefeces were not aVected by the UV treatments (Table 4). Inthe N percentage (% of d.m.) of the woodlice feces therewas an interaction between UV treatment and tree species.In the feces of woodlice fed with alder litter, N was lowerin the treatment that excluded both UV-B and UV-A radia-tion (P = 0.038) compared with the control, with a diVer-ence also between UV-exclusion treatments (P = 0.01)(Table 4). There were no diVerences in the N percentage inthe feces of woodlice fed with birch litter (Table 4).

Discussion

Our results suggest that the striking diVerence betweenthese species in leaf senescence (alder leaves drop whilestill green while birch leaves Wrst become yellow) is reX-ected in their litter phenolic proWles. The species also var-ied in their chemical responses to solar UV radiation. Inalder litter an eVect of mainly UV-A was apparent in theconcentration of cinnamic acids and Xavonoids. Interest-ingly, the cinnamic acids increased and Xavonoids decreaseddue to the exclusion of UV-A. These results indicate that adecrease in Xavonoid accumulation increases the accumulation

Fig. 1 C in cumulative CO2 evolution under diVerent UV treatments.Data for alder (1A) and for birch (1B) measured 13 times during18 weeks in the short-term decomposition experiment with soil and lit-ter. Data for alder (2A) and for birch (2B) measured 12 times during13 weeks in the decomposition experiment with woodlice and litter(means § SE). In the decomposition experiment with soil and litter themicrocosms had 700 g of soil and 1 g (d.m.) of litter each, whereas inthe decomposition experiment with woodlice and litter the microcosmshad three woodlice and 1 g (d.m.) of litter each. White bar Near-ambi-ent control, dark-grey bar exclusion of UV-B radiation, light-grey barexclusion of both UV-A and UV-B radiation. In the decompositionexperiment with soil and litter, for alder n = 4 trees for the near-ambi-ent control and exclusion of UV-B radiation, n = 3 for exclusion ofboth UV-A and UV-B radiation. For birch n = 6 trees in all treatments.In the decomposition experiment with woodlice and litter, n = 5 in alltreatments and in both species

Table 3 Phenolic-group responses to UV radiation in woodlice feces fed with grey alder (A) or white birch (B) litter

Values sharing the same letter within rows are not signiWcantly diVerent (adjusted P > 0.10). For A, n = 5 microcosms; for B, n = 6 microcosms

For all the compound groups in feces of woodlice fed with A litter, UV treatment df = 2 and 8; for those fed with B litter, UV treatment df = 2 and 10

Compound group Species P, F Mean § SE (mg g¡1)

Treatment UV-B+, UV-A+ UV-B¡, UV-A+ UV-B¡, UV-A¡

Phenolic acids A 0.93, 0.077 0.020 § 0.005a 0.022 § 0.001a 0.020 § 0.004a

B 0.78, 0.25 0.027 § 0.004a 0.030 § 0.004a 0.026 § 0.006a

Cinnamic acids A <0.0001, 25.10 0.030 § 0.005a 0.035 § 0.002a 0.080 § 0.008b

B 0.178, 2.07 0.037 § 0.005a 0.047 § 0.006a 0.062 § 0.015a

Presumed stilbenes A 0.66, 0.851 0.190 § 0.061a 0.246 § 0.076a 0.168 § 0.039a

B 0.14,2.36 0.076 § 0.023a 0.051 § 0.014a 0.025 § 0.008a

Flavonoids A 0.73, 0.325 0.037 § 0.018a 0.029 § 0.002a 0.044 § 0.019a

B 0.82, 0.20 0.517 § 0.143a 0.427 § 0.063a 0.512 § 0.125a

Condensed tannins A 0.46, 0.851 2.22 § 0.322a 2.11 § 0.207a 1.78 § 0.201a

123

726 Oecologia (2009) 161:719–728

of cinnamic acids, from which all phenylpropanoids arederived (Dixon and Paiva 1995). For the concentration ofchlorogenic acids, our results suggest opposite eVects ofUV-A and UV-B. These results are, in general, similar tothose from green leaves of this species (Kotilainen et al.2008), indicating negligible eVects of senescence onresponses of alder leaf phenolics. In contrast, three phenolicgroups identiWed in birch litter were not signiWcantlyaVected by the UV treatments. In earlier harvests, the con-centrations of chlorogenic acids and Xavonoids in birchwere aVected by the UV treatments (Kotilainen et al. 2008).

Other chemical properties measured from leaf littershowed other diVerences between the tree species, e.g., per-centage of N was higher in alder and percentages of �-cel-lulose and gravimetric lignin lower in alder than in birch.Moreover, in alder litter, the percentage of acid-solublelignin was negatively aVected by UV-A exclusion and thetotal lignin:�-cellulose ratio by UV-B exclusion. In bothtree species, exclusion of UV-B negatively aVected the totallignin:total hydrolyzable sugars ratio. Data of C concentra-tion suggested opposite eVects of UV-B and UV-A, sinceexclusion of UV-B decreased it slightly, but there were nodiVerences between simultaneous exclusion of both UV-Band UV-A and the control. In contrast, in a Weld study inArgentina, UV-B attenuation during growth of Gunneramagellanica did not have any signiWcant eVects on litterquality (N, P, C:N, lignin, cellulose) (Pancotto et al. 2003).On the other hand, in barley (Hordeum vulgare), the con-centrations of P, cellulose and lignin:N ratio were higherunder near-ambient UV-B than under attenuated UV-B,while the concentrations of N and soluble carbohydrateswere lower. Lignin content and C:N ratio were not aVectedby UV-B treatments (Pancotto et al. 2005). These inconsis-tent observations suggest species-speciWc responses andmay also stem from diVerences in experimental methods.

Nematodes and enchytraeids are highly representa-tive of the soil animal community, since they comprise allsoil animal trophic groups, i.e., herbivores, bacterivores,

fungivores, omnivores, detritivores and predators (Mikolaet al. 2005). In the current study, nematodes and enchytrae-ids were not aVected by the UV treatments, nor were therediVerences in their numbers between the tree species. Incontrast, in a UV-B-exclusion experiment in long-term Weldplots in a Sphagnum peatland in Argentina near-ambientUV-B had negative eVects on the numbers of rotifers, nem-atodes and mites but positive eVects on testate amoebae(Robson et al. 2005). Similarly, in a UV-exclusion experi-ment in Greenland near-ambient levels of both UV-A andUV-B decreased the numbers of microarthropods, particu-larly Collembola (Convey et al. 2002). UV-inducedchanges in vegetation and in the microenvironment werepresumed to have induced these eVects. The variable resultsare not that surprising, since there were diVerences in bothexperimental set-ups and duration.

Woodlice growth and the amount of litter consumed bywoodlice were not aVected by the UV treatments. However,woodlice grew better when fed with alder litter, likely dueto the fact that P. scaber grow better when feeding on litterwith a lower C:N ratio (Zimmer and Topp 2000). ForP. scaber, food quality is largely determined by microbialcolonization, since microbes provide an easily accessible Nsource (Zimmer and Topp 2000). Another possible expla-nation for faster growth of woodlice fed with alder litter inthe current study is that the concentration of tannins waslower in alder litter than in birch litter. An important impactof tannins on terrestrial isopods is their suspected inhibitoryeVects on gut microbiota that degrade phenolic compoundsand cellulose (Zimmer and Topp 1998; Zimmer 1999).There can be major diVerences in the structure of con-densed tannins, even between closely related plant species(Ayres et al. 1997) and subsequently this may have aVectedtheir assimilation by woodlice. Further, there were UV-induced changes in individual phenolics in both species andthe concentration of N in feces of woodlice eating alder lit-ter. These changes in feces chemistry could have an eVecton the decomposition rate, since fecal pellets have a higher

Table 4 Responses of C and N concentration and C:N ratio to UV radiation in woodlice feces fed with grey alder (A) or white birch (B) litter

Values sharing the same letter within rows are not signiWcantly diVerent (adjusted P > 0.10). For A, n = 5 jars; for B, n = 6 jars

For all the parameters, UV treatment df = 2 and 18, sample time df = 1 and 9 and interaction df = 2 and 18

P, F Species Mean § SE

Treatment Species Interaction UV-B+ UV-A+ UV-B¡ UV-A+ UV-B¡ UV-A¡

N (% of d.m.) 0.077, 2.97 <0.0001, 214 0.014, 5.46 A 3.41 § 0.085a 3.39 § 0.153a 2.99 § 0.147b

B 1.55 § 0.074a 1.61 § 0.097a 1.62 § 0.054a

C (% of d.m.) 0.59, 0.538 0.65, 0.227 0.13, 2.26 A 45.64 § 1.19a 44.43 § 1.41a 41.76 § 2.65a

B 42.94 § 1.38a 42.71 § 1.08a 43.90 § 0.989a

C:N ratio 0.243, 1.53 <0.0001, 446 0.755, 0.286 A 13.39 § 0.351a 13.16 § 0.444a 13.95 § 0.442a

B 27.89 § 0.914a 26.88 § 1.32a 27.24 § 0.856a

123

Oecologia (2009) 161:719–728 727

surface-to-volume ratio compared with the original litter,thus enhancing decomposition (Bardgett 2005).

In the current study in the microcosms with soil and lit-ter, the amount of substrate was substantially larger than inmicrocosms with litter and woodlice. UV-induced changesin plant litter did not aVect CO2 production of the micro-cosms with litter and woodlice, additionally there were nodiVerences between the two tree species. In contrast, CO2

evolution measured in the short-term decomposition exper-iment with soil and litter, assumed to be mainly the result ofsoil microbial respiration, was negatively aVected in bothtree species mainly by UV-B exclusion. This result presum-ably relates to changes in litter leachates that had an aVecton microbes. In a Weld experiment in Greenland, solar UV-Battenuation above ground altered the microbial communitycomposition (Rinnan et al. 2005). Similarly, in a 5-yearstudy in subarctic heath in northern Sweden, it was shownthat despite subtle eVects on the aboveground vegetation,UV-B supplementation resulted in a decrease in themicrobial biomass C:N ratio (Johnson et al. 2002). Theimportance of soil respiration stems from the magnitude ofsoil-to-atmosphere CO2 Xux; it is the dominant terrestrialsource of CO2, and it is approximated that »10% of theatmosphere’s CO2 cycles through soils each year. This soilrespiration is produced mainly by heterotrophic organismsand by respiration of living roots, inXuenced mainly by fac-tors such as temperature, moisture and substrate properties(Raich and Tufekcioglu 2000, Raich et al. 2002). Takentogether, results of the current study suggest that UV-induced changes in litter altered microbial communitiesand/or their activity in the soil. These litter-quality-depen-dent eVects that showed up as signiWcant changes in soilCO2 emissions could be important for the planetary Ccycle. This response under exclusion of UV radiation alsosuggests consequences for the diverse soil decomposercommunity (food web structure) in the long run.

There were surprisingly few diVerences between treespecies in the decomposition experiments, despite manydiVerences between species in their litter chemistry (e.g., C,N, �-cellulose, see above); even the soil pH in microcosmswith birch litter was higher than in microcosms with alderlitter. UV-driven changes in soil pH in the current studywere small, and assumed to have little eVect on soil faunaand decomposition. Moreover, UV-induced changes in lit-ter chemistry had no inXuence on the litter mass loss in anyof the decomposition experiments. After 13 months in thelong-term decomposition experiment with soil and litter,approximately half of the original litter mass still remainedin the microcosms. The quality of litter aVects its decompo-sition rate and degradation of the most recalcitrant fraction,mainly lignin, takes place late during the process of decom-position (Bardgett 2005). Therefore, it is plausible thatthe observed changes in the total lignin:total hydrolyzable

sugars ratio in both species and changes in diVerent phenolicsin alder litter might have an eVect on the later phases oflitter decomposition.

Supplemental UV-B radiation has been shown toincrease or decrease or to have no eVect on chemicalmetabolite and element composition of litter (e.g., �-cellu-lose, lignin:N ratio) and decomposition rates, e.g., inCalamagrostis epigeios, Quercus robur and Vacciniumuliginosum and Vaccinium myrtillus (Gehrke et al. 1995;Rozema et al. 1997; Newsham et al. 2001a, b; see alsoreview by Bassman 2004). The variable results suggest spe-cies-speciWc responses but can also partly stem from diVer-ences in spectra of UV lamps and diVerent BSWF used inthe experiments. Our earlier study with exclusion of UVradiation showed that the accumulation of phenolic metab-olites in the green leaves of the same trees used in this cur-rent litter study may follow diVerent action spectra, whichalso suggests that appropriate BSWF extend in most casesinto the UV-A region (Kotilainen et al. 2008). Our Wndingsin the current study—changes in litter chemistry and subse-quently on soil respiration—suggest that ozone depletionand the resulting increased UV-B radiation may have far-reaching consequences for the decomposition process interrestrial ecosystems. Furthermore, these Wndings indicatethat independently of the problem of ozone-depletion, bothUV-B and UV-A radiation play an important role in terres-trial ecosystems as indirect regulators of decomposition.

Acknowledgments We thank Sinikka Sorsa for assisting with theHPLC analyses and Leena Siitonen, Nipa Manosuk, Pasi Kemppainenand Kati Sivander for assisting with the Weld experiment and preparingthe leaf samples, and Mustafa Boucelham and Anna Repo for assistingwith the decomposition experiments. We thank Sari Vilhunen andHanna Ruhanen for assisting with the lignin and sugar concentrationanalyses and Matty Berg for collecting the woodlice population. Wethank Emily Knott for checking the English language of the manu-script. This work was supported by the Maj and Tor Nessling Founda-tion (grant no. 2007046 to T. Kotilainen) and more recently also by theAcademy of Finland (grant no. 116775 to P. J. Aphalo). The experi-ments reported here comply with the current laws of Finland.

References

Abrahamsen G (1973) Studies on body-volume, body-surface area,density and live weight of Enchytraeidae (Oligochaeta). Pedobio-logia 13:6–15

Anttonen S, Manninen A-M, Saranpää P, Kainulainen P, Linder S,Vapaavuori E (2002) EVects of long-term nutrient optimisationon stem wood chemistry in Picea abies. Trees Struct Funct16:386–394

Ayres MP, Clausen TP, MacLean SF, Redman AM, Reichardt PB(1997) Diversity of structure and antiherbivore activity incondensed tannins. Ecology 78:1696–1712

Bardgett RD (2005) The biology of soil. Oxford University Press,Oxford

Bassman JH (2004) Ecosystem consequences of enhanced solarultraviolet radiation: secondary plant metabolites as mediators of

123

728 Oecologia (2009) 161:719–728

multiple trophic interactions in terrestrial plant communities.Photochem Photobiol 79:382–398

Caldwell MM, Bornman JF, Ballaré CL, Flint SD, Kulandaivelu G(2007) Terrestrial ecosystems, increased solar ultraviolet radia-tion, and interactions with other climate change factors. Photo-chem Photobiol Sci 6:252–266

Convey P, Pugh PJA, Jackson C, Murray AW, Ruhland CT, Xiong FS,Day TA (2002) Response of Antarctic terrestrial microarthropodsto long-term climate manipulations. Ecology 83:3130–3140

Dixon RA, Paiva NL (1995) Stress-induced phenylpropanoid metabo-lism. Plant Cell 7:1085–1097

Flint SD, Caldwell MM (2003a) A biological spectral weightingfunction for ozone depletion research with higher plants. PhysiolPlant 117:137–144

Flint SD, Caldwell MM (2003b) Field testing of UV biological spectralweighting functions for higher plants. Physiol Plant 117:145–153

Gehrke C, Johanson U, Callaghan TV, Chadwick D, Robinson CH(1995) The impact of enhanced ultraviolet-B radiation on litterquality and decomposition processes in Vaccinium leaves fromthe subarctic. Oikos 72:213–222

Hagerman AE (1995) Tannin chemistry. Department of Chemistry,Miami University, Miami, pp 24–25

Johnson D, Campbell CD, Lee JA, Callaghan TV, Gwynn-Jones D(2002) Arctic microorganisms respond more to elevated UV-Bradiation than CO2. Nature 416:82–83

Keinänen M, Julkunen-Tiitto R (1998) High-performance liquid chro-matographic determination of Xavonoids in Betula pendula andBetula pubescens leaves. J Chromatogr A 793:370–377

Keski-Saari S, Pusenius J, Julkunen-Tiitto R (2005) Phenolic com-pounds in seedlings of Betula pubescens and B. pendula areaVected by enhanced UV-B radiation and diVerent nitrogen regi-mens during early ontogeny. Glob Change Biol 11:1180–1194

Kotilainen T, Tegelberg R, Julkunen-Tiitto R, Lindfors A, Aphalo PJ(2008) Metabolite speciWc eVects of solar UV-A and UV-B onalder and birch leaf phenolics. Glob Change Biol 14:1294–1304

Mazza CA, Boccalandro HE, Giordano CV, Battista D, Scopel AL,Ballaré CL (2000) Functional signiWcance and induction by solarradiation of ultraviolet-absorbing sunscreens in Weld-grownsoybean crops. Plant Physiol 122:117–126

Mikola J, Nieminen M, Ilmarinen K, Vestberg M (2005) Belowgroundresponses by AM fungi and animal trophic groups to repeateddefoliation in an experimental grassland community. Soil BiolBiochem 37:1630–1639

Moody S, Paul N, Björn LO, et al. (2001) The direct eVects of UV-Bradiation on Betula pubescens litter decomposing at four Euro-pean Weld sites. Plant Ecol 154:27–36

Newsham KK, Anderson JM, Sparks TH, Splatt P, Woods C, McLeodAR (2001a) UV-B eVect on Quercus robur leaf litter decomposi-tion persists over four years. Glob Change Biol 7:479–483

Newsham KK, Splatt P, Coward PA, Greenslade PD, McLeod AR,Anderson JM (2001b) Negligible inXuence of elevated UV-Bradiation on leaf litter quality of Quercus robur. Soil BiolBiochem 33:659–665

O'Connor FB (1962) The extraction of Enchytraeidae from soil. In:Murphy PW (ed) Progress in soil zoology. Butterworths, London,pp 279-285

Pancotto VA, Sala OE, Cabello M, Lopez NI, Robson TM, Ballaré CL,Caldwell MM, Scopel AL (2003) Solar UV-B decreases decom-position in herbaceous plant litter in Tierra del Fuego, Argentina:potential role of an altered decomposer community. Glob ChangeBiol 9:1465–1474

Pancotto VA, Sala OE, Robson TM, Caldwell MM, Scopel AL (2005)Direct and indirect eVects of solar ultraviolet-B radiation on long-term decomposition. Glob Change Biol 11:1982–1989

Paul ND, Callaghan TV, Moody S, Gwynn-Jones D, Gehrke C (1999)UV-B impacts on decomposition biochemical cycling. In: Roze-ma J (ed) Stratospheric ozone depletion: the eVects of enhancedUV-B radiation on terrestrial ecosystems. Backhuys, Leiden, pp117–133

Pinheiro JC, Bates DM (2000) Mixed-eVects models in S and S-plus.Springer, New York

R Development Core Team (2006) R: a language and environment forstatistical computing. R Foundation for Statistical Computing,Vienna

Raich JW, Tufekcioglu A (2000) Vegetation and soil respiration:correlations and controls. Biogeochemistry 48:71–90

Raich JW, Potter CS, Bhagawati D (2002) Interannual variation inglobal soil respiration, 1980–94. Glob Change Biol 8:800–812

Rinnan R, Keinanen MM, Kasurinen A, et al. (2005) Ambient ultravi-olet radiation in the Arctic reduces root biomass and alters micro-bial community composition but has no eVects on microbialbiomass. Glob Change Biol 11:564–574

Robson TM, Pancotto VA, Scopel AL, Flint SD, Caldwell MM (2005)Solar UV-B inXuences microfaunal community composition in aTierra del Fuego peatland. Soil Biol Biochem 37:2205–2215

Rozema J, Tosserams M, Nelissen HJM, van Heerwaarden L, Broek-man RA, Flierman N (1997) Stratospheric ozone reduction andecosystem processes: enhanced UV-B radiation aVects chemicalquality and decomposition of leaves of the dune grassland speciesCalamagrostis epigeios. Plant Ecol 128:285–294

Sarkar D (2008) Lattice: multivariate data visualization with R.Springer, New York

SCAN-CM (1994) Asetoniin liukenevan aineen määrittäminen (inFinnish). Scandinavian Pulp, Paper and Board Testing Committee50. Central Laboratory of Finland, Espoo, Finland

Sohlenius B (1979) A carbon budget for nematodes, rotifers and tardi-grades in a Swedish coniferous forest soil. Holarctic Ecol 2:30–40

van der Heijden MGA, Bardgett RD, van Straalen NM (2008) The un-seen majority: soil microbes as drivers of plant diversity and pro-ductivity in terrestrial ecosystems. Ecol Lett 11:296–310

Venables WN, Ripley BD (2002) Modern applied statistics with S.Springer, New York

Zimmer M (1999) The fate and eVects of ingested hydrolyzable tanninsin Porcellio scaber. J Chem Ecol 25:611–628

Zimmer M, Topp W (1998) Microorganisms and cellulose diges-tion in the gut of the woodlouse Porcellio scaber. J Chem Ecol24:1397–1408

Zimmer M, Topp W (2000) Species-speciWc utilization of food sourcesby sympatric woodlice (Isopoda: Oniscidea). J Anim Ecol69:1071–1082

123