Science of the Total Environment 439 (2012) 211–219

Contents lists available at SciVerse ScienceDirect

Science of the Total Environment

j ourna l homepage: www.e lsev ie r .com/ locate /sc i totenv

Soil fertility and plant diversity enhance microbial performance inmetal-polluted soils

Anna M. Stefanowicz a,⁎, Paweł Kapusta a, Grażyna Szarek-Łukaszewska a, Krystyna Grodzińska a,Maria Niklińska b, Rolf D. Vogt c

a W. Szafer Institute of Botany, Polish Academy of Sciences, Lubicz 46, 31-512 Kraków, Polandb Institute of Environmental Sciences, Jagiellonian University, Gronostajowa 7, 30-387 Kraków, Polandc Department of Chemistry, University of Oslo, P.O. Box 1033, Blindern, 0315 Oslo, Norway

H I G H L I G H T S

► We examined effects of habitat properties on microbial parameters in polluted soil.► Nutrient content increased microbial activity.► Plant species richness increased microbial activity and functional richness.► Toxic effects of trace metals were ameliorated by nutrient content.► Bacterial and fungal communities were affected by habitat properties differently.

⁎ Corresponding author. Tel.: +48 124241770; fax: +E-mail addresses: a.stefanowicz@botany.pl (A.M. Ste

p.kapusta@botany.pl (P. Kapusta), g.szarek@botany.pl (k.grodzinska@botany.pl (K. Grodzińska), maria.niklinskr.d.vogt@kjemi.uio.no (R.D. Vogt).

0048-9697/$ – see front matter © 2012 Elsevier B.V. Allhttp://dx.doi.org/10.1016/j.scitotenv.2012.09.030

a b s t r a c t

a r t i c l e i n f o

Article history:Received 6 March 2012Received in revised form 13 September 2012Accepted 13 September 2012Available online 13 October 2012

Keywords:Soil microbial communitiesHeavy metalsPlant species richnessPlant functional diversitySoil physicochemical properties

This study examined the effects of soil physicochemical properties (including heavymetal pollution) and vegetationparameters on soil basal respiration, microbial biomass, and the activity and functional richness of culturable soilbacteria and fungi. In a zinc and lead mining area (S Poland), 49 sites were selected to represent all commonplant communities and comprise the area's diverse soil types. Numerous variables describing habitat propertieswere reduced by PCA to 7 independent factors, mainly representing subsoil type (metal-rich mining waste vs.sand), soil fertility (exchangeable Ca,Mg andK, total C andN, organic C), plant species richness, phosphorus content,water-soluble heavy metals (Zn, Cd and Pb), clay content and plant functional diversity (based on graminoids,legumes and non-leguminous forbs). Multiple regression analysis including these factors explained much of thevariation inmostmicrobial parameters; in the case ofmicrobial respiration and biomass, itwas 86% and 71%, respec-tively. The activity of soil microbeswas positively affectedmainly by soil fertility and, apparently, by the presence ofmining waste in the subsoil. The mining waste contained vast amounts of trace metals (total Zn, Cd and Pb), but itpromotedmicrobial performance due to its inherently high content ofmacronutrients (total Ca, Mg, K and C). Plantspecies richness had a relatively strong positive effect on all microbial parameters, except for the fungal component.In contrast, plant functional diversitywas practically negligible in its effect onmicrobes. Other explanatory variableshad only a minor positive effect (clay content) or no significant influence (phosphorus content) on microbialcommunities. The main conclusion from this study is that high nutrient availability and plant species richnesspositively affected the soil microbes and that this apparently counteracted the toxic effects of metal contamination.

© 2012 Elsevier B.V. All rights reserved.

1. Introduction

High concentrations of heavy metals can severely reduce thegrowth and survival of soil microorganisms and thus adversely affectmany ecosystem functions driven by these organisms (Bååth, 1989;

48 124219790.fanowicz),G. Szarek-Łukaszewska),a@uj.edu.pl (M. Niklińska),

rights reserved.

Giller et al., 2009; McGrath et al., 1995; Ramsey et al., 2005). Param-eters describing the condition of microbial communities, such asmicrobial respiration and biomass, or functional and structural diver-sity, are widely used to assess soil health in industrial and urban areas(Avidano et al., 2005; Shukurov et al., 2005; Zhang et al., 2008).

In natural situations, however, biological indicators are influencedby a wider range of environmental factors. Belowground microbiotaare particularly sensitive to changes in soil physicochemical propertiessuch as temperature, moisture, pH, nutrient availability or clay content(Balogh et al., 2011; Lauber et al., 2008; Teklay et al., 2010). Thesefactors directly control the abundance and activity of microbes and

212 A.M. Stefanowicz et al. / Science of the Total Environment 439 (2012) 211–219

indirectly affect them through regulating the bioavailability of the con-taminants (Bååth, 1989; Giller et al., 2009; Vig et al., 2003). For exam-ple, microbial communities benefit from high contents of organicmatter and nutrients in soil. This is due not only to the abundanceand quality of food resources but also to the efficient immobilizationof hard or type B metals by strong binding to the organic material.Such alleviation of the negative effects of heavy metal pollution onsoil biotamay be the cause for the weak or absent correlations betweentrace metal concentrations and microbial activity reported by someauthors (Niklińska et al., 2005; Schipper and Lee, 2004).

Recently, much attention has been given to the role of vegetationin shaping soil microbial communities, as this seems to be similarlyimportant as abiotic factors (see e.g. Garbeva et al., 2008; García-Palacios et al., 2011; Liu et al., 2010; Marschner et al., 2004;Sanaullah et al., 2011). Plants supply soil bacteria and fungi withcarbon substrates through litter production and root exudates(Grayston et al., 1997; Hooper et al., 2000; Wardle et al., 2004).They can also substantially influence the chemistry of soil solutionand thus make local sources of nutrients (or toxicants) available orunavailable for soil biota (Gobran and Clegg, 1996; Hinsinger et al.,2006). Plant species vary in these abilities, so mixtures of them maygenerate a higher biochemical diversity of belowground ecosystemsas compared to monocultures, and in the process may maintainmore diverse, presumably better-functioning, microbial communities.Such a mechanism may explain the positive relationships betweenplant diversity and microbial performance reported by some authors(Eisenhauer et al., 2011; Loranger-Merciris et al., 2006; Stephan et al.,2000; Zak et al., 2003).

Although chronic pollution or other anthropogenic impacts makeindustrial and urban wastelands inhospitable habitats for livingorganisms, they can be colonized by relatively dense vegetation. Arelevant example is the mining environ of Olkusz, the study area ofthis paper, where exploitation of shallow deposits of non-ferrousmetals (Ag, Pb, Zn) has been carried out since medieval times(Cabala et al., 2009). Many of the indigenous plant species havedeveloped some tolerance to the heavy metal toxicity (Olko et al.,2008; Wierzbicka and Panufnik, 1998; Wierzbicka and Pielichowska,2004; Załęcka and Wierzbicka, 2002). They are thereby able to buildspecies-rich communities on the heavily polluted soil. It is possiblethat the greater plant diversity or density in such habitats may resultin an improved microbial performance.

The present work is part of a broader study examining theresponse of soil microbes to stress induced by the high concentrationof trace metals in the soils of the Olkusz ore-bearing region. In theprevious paper (Stefanowicz et al., 2010), we reported relativelyweak correlations between metal concentrations and microbialparameters despite the high level and steep gradients of the former.We hypothesized that the toxic effect of metals on soil microbiotawas ameliorated by the favorable influence of such soil properties asorganic matter content and pH since the values of these variableswere higher in highly polluted sites. In this study we took into ac-count not only the above mentioned (and other) soil physicochemicalproperties, but also parameters of the herbaceous vegetation (i.e. spe-cies richness, species composition, functional diversity and coverage),as factors potentially affecting microbes. By means of multivariatemethods, we identified the main sources of variation in the data stud-ied and estimated their effects on soil basal respiration, microbialbiomass, and the activity and functional richness of culturable soilbacteria and fungi.

2. Materials and methods

2.1. Study area and sampling procedure

The study was carried out in the Olkusz ore-bearing region, anarea heavily contaminated with trace metals due to centuries of

mining and smelting activity (Cabala et al., 2009). The study areawas mostly a mixture of post-mining sites and fallow farmland (aban-doned in the 1970s after a buffer zone was established around themine and a metallurgical plant in Bukowno), afforested or spontane-ously colonized by grassland vegetation. Within this area, 49 studysites were selected to represent the dominant vegetation types(grassland and pine forest) and substrates (mining waste and sand).They were classified in 6 habitat categories described in more detailby Kapusta et al. (2011) and Stefanowicz et al. (2010).

In May 2009, three soil samples of the top mineral soil horizonwere collected from the center of each of the 49 study sites after care-ful removal of the top soil organic layer. The three samples werebulked into one composite sample. The sampling places were situatedaround a circular plot (4 m2) from which the density of particularherbaceous species on the Braun–Blanquet scale (ranging between 1and 6) as well as the total percent plant cover were estimated duringthe growing season. Plant species were identified in the field or weresampled for identification in the laboratory, following the nomencla-ture by Mirek et al. (2002).

2.2. Chemical and biological analyses

The composite fresh soil samples were sieved through 2 mmmesh. Then the samples were split into two parts: one air-dried atroom temperature for analysis of physicochemical properties, andthe other kept moist at 4 °C in a climate chamber for analysis ofmicrobial parameters. Soil dry weight was measured after drying at105 °C for 12 h, and organic matter content was determined asweight loss upon ignition at 550 °C. The particle size distribution(sand, silt and clay fractions) was determined by a combination ofsieving and sedimentation. Soil suspension pH was measuredelectrometrically using a pH electrode following the extraction withH2O at a 1:5 (w:v) ratio. Total and organic C contents were analyzedby dry combustion technique with a LECO SC-144DRPC analyzer(Leco) — organic C was determined after treating the soil sampleswith 2 N H2SO4 and 5% FeSO4 in order to remove carbonates. TotalN was determined by a method based on Kjeldahl digestion using aKjeltec 2300 (FossTecator). Contents of Ca, Mg, K, Zn, Cd and Pb inthe soils were measured by flame or graphite furnace atomic absorp-tion spectrometry (Varian 220 FS) after hot HClO4 digestion (metalstermed in this paper as “total”). Exchangeable Ca, Mg, K, Zn and Cdwere determined by extraction with 0.1 M BaCl2 at pH 7, andwater-soluble Zn, Cd and Pb were measured after the extractionwith deionized water at a 1:10 (w:v) ratio. The pool of total P wasreleased from the soil by mineralization with HClO4 and available Pwas extracted using 0.5 M NaHCO3 (Olsen method). The amount ofphosphorus in the extracts was determined by the molybdenumblue method using a Hach-Lange DR 3800 spectrophotometer.

Soil basal respiration (BR) and microbial biomass (Cmic) assubstrate-induced respiration (SIR) were measured by a chemicalmethod with the absorption of CO2 in 0.2 M NaOH followed by titra-tion of excess hydroxide with 0.1 M HCl. Glucose (10 g kg−1 soil dryweight) was added to soil samples 4 h before SIR measurements. Fordetails of BR and SIR measurements see Stefanowicz et al. (2010).

The activity and functional richness of microbial communitieswere estimated with Biolog GN2 (bacteria) and SFN2 plates (fungi),containing 95 sole carbon substrates for microbes. The procedurestarted from shaking fresh soil samples (3 g dry weight) in 30 ml0.9% NaCl for 1 h. Then the extracts were frozen in liquid nitrogenand kept at −70 °C until analysis (Boivin et al., 2007). Thawed soilsolutions were diluted 10 times in physiological salt (GN2 plates) orin 0.1% water agar+0.04% Tween 80 (polyoxyethylene (20) sorbitanmonooleate) (SFN2 plates). Adding gelling agents and surfactantsallows a uniform dispersion of fungal spores in a solution (Buyeret al., 2001; Kraus et al., 2004). Streptomycin sulfate (1 μg per well)and chlortetracycline (0.5 μg per well) were added to the solution

213A.M. Stefanowicz et al. / Science of the Total Environment 439 (2012) 211–219

for fungal plates to prevent bacterial growth (Dobranic and Zak,1999). The solutions were vortexed vigorously and one plate persoil sample was inoculated with 100 μl inoculum per well. The plateswere incubated at 22 °C and color development was measuredspectrometrically (590 nm) twice a day for ca. 115 h (GN2 plates)or turbidity (650 nm) was measured once a day for ca. 216 h (SFN2plates) with a μQuant reader (BIO-TEK Instruments).

2.3. Data handling

Microbial biomass was calculated according to the equation ofAnderson and Domsch (1978). Soil basal respiration and microbialbiomass were expressed both per soil dry weight (BR, Cmic) and perunit organic matter (BROM, CmicOM). Expressing microbial parametersper unit of organic matter enables a reduction of the influence ofdifferences in soil organic matter content on these parameters; thedifferences sometimes obscure effects of other factors (e.g. heavymetals) on the microbial indicators. Bacterial activity and fungalactivity on each Biolog substrate were expressed as area undercurve (AUC), and calculated as described by Stefanowicz et al.(2010). Average bacterial and fungal activities on each plate wereexpressed as average area under curve (AAUCb and AAUCf, respec-tively). The functional richness of bacterial (S′b) and fungal (S′f)communities was expressed as the number of substrates utilized ona plate. Bacteria were compared with fungi as regards these parame-ters using the t-test for dependent samples.

Plant communities were characterized by plant species richness(the number of species) and percent cover of the herbaceous layer.Then the plants were classified into three functional groups, namely,graminoids (grasses and sedges), legumes and non-leguminousforbs (herbaceous plants not included in the former groups; thisgroup also included tree seedlings). This classification is based ontaxonomy and traits that are potentially relevant to soil microbialcommunities (i.e. N-fixing, root morphology, litter quality). Plantfunctional diversity was calculated using the Shannon–Wiener index:

H′ ¼ −XS

i¼1

pi ln pi;

where S is the total number of functional groups and pi is the numberof species belonging to the ith functional group as a proportion of thenumber of all species. The H′ index was used because it bore someinformation about the contribution of a given functional group inthe mixture and was characterized by continuous variation acrosssites, in contrast to functional richness which received only twovalues: 2 or 3. Also, for each functional group plant density (thesum total of Braun–Blanquet scores of species belonging to a givenfunctional group) was determined. Frequent species (occurring in atleast 10% of the study sites) were submitted to detrended correspon-dence analysis (DCA) to identify the main ecological gradients in theplant dataset. Sample scores of the first two axes (DCA 1 and DCA 2)were used in subsequent analyses as representations of plant speciescomposition (Lepš and Šmilauer, 2003).

The PCA method for factor extraction was performed on practical-ly all variables describing soil physicochemical properties and plantcommunity characteristics; only legume density was censored dueto its considerable deviation from a normal distribution (high propor-tion of 0 values). Factors with eigenvalues >1 were varimax-rotatedand used further on as non-correlated predictor variables in multipleregression and simple non-parametric correlation tests.

Multiple regression analysis was used to assess the influence ofhabitat properties (soil physicochemical properties and vegetationcharacteristics expressed as factors) on soil basal respiration, microbi-al biomass, and the activity and functional richness of soil bacteriaand fungi. Spearman correlation tests were performed to investigate

a functional change in the microbial communities, i.e. a shift in thepattern of utilization of 95 carbon substrates on the Biolog plates bybacteria and fungi, induced by habitat properties. Before this analysis,values of substrate utilization (AUC) were divided by AAUC for agiven site in order to reduce the influence of inoculum density onthe metabolic fingerprint of the microbial communities (Garlandand Mills, 1991).

Statistical analyses were done with STATISTICA 9 (Statsoft Inc.) orCANOCO 4.5 (ter Braak and Šmilauer, 2002). Prior to statistical teststhe data were transformed with logarithmic or exponential functionsto obtain a normal or at least symmetric distribution. After transfor-mation, all variables were normalized — scaled between 0 and 1values.

3. Results

Pollution levels varied widely across the study area. Total contentof trace metals ranged from 0.1 to 72.1 g kg−1 for Zn, from 2 to506 mg kg−1 for Cd, and from 0.1 to 33.2 g kg−1 for Pb. Theirexchangeable and water-soluble forms also varied over a largerange (Table 1). Most other soil properties varied to a similar degree,with many of them having coefficients of variation exceeding 100%(e.g., total Ca and Mg). In contrast, soil pH showed less variation,with values between 6.5 and 7 at most sites. The particle size distribu-tion did not differ much between sites either.

The wide gradients of soil physicochemical properties werereflected in a high variation of biotic parameters (Table 1). Theplant community types ranged from poor psammophilous grasslandto more fertile xerothermic grassland developed on mining waste(calcareous substrate) and fallows. This transition of plant speciescomposition along the main environmental gradients is clear in theDCA diagram (Fig. 1). The minimum number of species (plant speciesrichness) recorded from a sample plot was 2, and the maximum 25.Plant cover ranged from 20% to 100%. All microbial parameters variedsubstantially across sites (Table 2), with coefficients of variationaround 100%. Fungi were slightly more active than bacteria and uti-lized more of the substrates on the Biolog plates (P>0.05).

Factor analysis reduced the original number of variables describ-ing habitat properties to seven independent factors (Table 3). Thewhole model accounted for 80.5% of the variation in the data. Themost important factor (F1) reflected the influence of mining waste(which formed the subsoil of many sites) on the physicochemicalproperties of the top mineral horizon. The next factor (F2) describedsoil fertility and was correlated with exchangeable metals, nitrogenand carbon. Other soil properties were grouped in F4 (total contentand availability of phosphorus), F5 (water-soluble forms of tracemetals) and F7 (clay content). Vegetation parameters were repre-sented by F3 and F6. The former reflected mainly plant species rich-ness and forb density, and the latter reflected plant functionaldiversity and graminoid density.

Microbial parameters were dependent on most of the habitatproperties (Table 4). Subsoil type and, related to it, the total concen-tration of trace metals and macronutrients, determined microbialrespiration and biomass. When the variables were expressed perunit of dry weight (BR and Cmic) the effect was clearly positive.When calculated per unit of organic matter the relationship becameweak (BROM) or negative (CmicOM). A similar inversion applied towater-soluble metals (F5). This factor correlated positively with BRand negatively with BROM. Simple correlations of microbialparameters with soil physicochemical properties (original variables)confirmed that trace metals were responsible for the negativerelationships between F1 and CmicOM and between F5 and BROM

(Table 5). Soil fertility (F2) benefitted almost all microbial parame-ters. Plant species richness (F3) played a similar though less signifi-cant role (see also Fig. 2); F3 was important especially for thebacterial component. Plant functional diversity (F6), on the other

Table 1Soil and vegetation characteristics for 49 study sites.

Variable Minimum Lower quartile Median Mean (SD) Upper quartile Maximum

Sand (%) 78 89 91 91 (4) 94 98Silt (%) 1 3 4 4 (2) 5 9Clay (%) 1 3 5 5 (3) 6 19pH 5.0 6.4 7.0 6.8 (0.7) 7.2 8.2Organic C (%) 0.2 0.7 1.1 1.9 (1.9) 1.8 7.9Total C (%) 0.5 1.1 2.3 3.4 (3.2) 5.0 15.2Total N (%) 0.03 0.08 0.13 0.20 (0.20) 0.19 0.86Total Ca (g kg−1) 0.1 1.4 3.5 10.4 (13.3) 16.6 52.4Total Cd (mg kg−1) 2 10 25 76 (114) 88 506Total K (g kg−1) 0.16 0.36 0.72 0.76 (0.50) 1.02 2.37Total Mg (g kg−1) 0.1 0.5 1.5 4.1 (5.8) 4.6 19.4Total P (g kg−1) 0.08 0.24 0.42 0.43 (0.25) 0.56 1.23Total Pb (g kg−1) 0.1 0.3 0.8 3.7 (7.0) 3.7 33.2Total Zn (g kg−1) 0.1 1.1 2.5 9.8 (15.2) 12.9 72.1Exchangeable Ca (g kg−1) 0.02 0.31 1.20 1.32 (1.20) 2.08 4.52Exchangeable Cd (mg kg−1) 0.1 1.3 3.7 8.5 (13.5) 6.4 60.5Exchangeable K (mg kg−1) 2 8 30 55 (60) 79 270Exchangeable Mg (mg kg−1) 3 33 145 210 (211) 316 861Exchangeable Zn (mg kg−1) 7 24 54 124 (173) 115 772Available P (mg kg−1) 0.6 3.1 4.0 6.2 (5.2) 7.5 23.0Water-soluble Cd (mg kg−1) 0.003 0.009 0.016 0.025 (0.030) 0.029 0.154Water-soluble Pb (mg kg−1) 0.013 0.101 0.185 0.246 (0.208) 0.305 0.938Water-soluble Zn (mg kg−1) 0.6 2.7 3.6 4.2 (2.4) 4.9 11.0Plant species richness 2 6 11 11 (6) 16 25Plant functional diversity 0 0.50 0.63 0.61 (0.17) 0.69 0.99Plant coverage (%) 20 50 70 68 (23) 80 100Forb density 2 7 12 14 (10) 18 43Graminoid density 0 5 6 6 (3) 8 12DCA 1 0 2.0 2.5 2.6 (1.1) 3.4 4.1DCA 2 0 0.9 1.4 1.4 (0.8) 1.9 3.5

SD — standard deviation; plant species richness— the number of herbaceous species; plant functional diversity— the Shannon–Wiener index calculated on the basis of the numberof functional groups and the number of species belonging to each functional group (for details see the text); plant coverage — the percentage cover of herbaceous vegetation; forb(or graminoid) density — the sum total of Braun–Blanquet scores of forb (or graminoid) species; DCA 1 and DCA 2 — sample scores of the first two axes obtained by DCA (Fig. 1);vegetation characteristics were determined for plots of 4 m2.

214 A.M. Stefanowicz et al. / Science of the Total Environment 439 (2012) 211–219

hand, was almost negligible in its effect on microbes, and the amountof phosphorus in soil (F4) had no impact on the response variables atall.

Habitat properties (F1–F7) caused a shift in the functional struc-ture of microbial communities. They affected (largely positively) theutilization of 59 Biolog substrates by bacteria and 49 by fungi. Mostsignificant correlations were found for F2 and F3 within bacterialcommunities. Bacteria from sites of higher soil fertility and plantspecies richness were able to utilize a wider spectrum of organiccompounds such as amino acids, amines and amides, carbohydratesand carboxylic acids (Table 6). Similarly, the fungal substrate utiliza-tion profile was shaped by F2 and F3, but to a much less extent thanthe bacterial one. Moreover, fungal activity responded positively tothe concentration of water-soluble metals (F5; Table 6).

4. Discussion

Industrial areas are considered to provide low quality habitats forsoil biota, mostly because of soil destruction (e.g. top layer removal)and pollution (Batty and Hallberg, 2010). The environs of Olkusz arenot an exception in this respect. The mining waste sites and thesurrounding sandy ground are characterized by unfavorable soilstructure, water and nutrient (particularly N and P) deficiency, andmetal contamination. Under such conditions soil microbial activityis expected to be low. In fact, in the present study soil microbialrespiration and biomass were poor. For instance, the latter constitut-ed on average only 0.3% of soil organic matter, whereas typical valuesrange between 1% and 4% (Brookes, 2001). These results are in linewith the findings of the studies carried out in other mining areas(Chodak and Niklińska, 2010; Rosenvald et al., 2011).

The Olkusz ore-bearing region has a patchy distribution of metalcontamination as a result of the spatial diversity of its natural geologyand of its anthropogenic impacts, particularly ore exploitation and

mining waste disposal. The study sites reflected this; total soilmetal concentrations varied from low values to extremely high, farexceeding environmental standards (Commission of EuropeanCommunities, 1986). Along this large gradient one expects strongnegative correlations between trace metal content and microbialactivity. In a preceding investigation (Stefanowicz et al., 2010), how-ever, these relationships were found to be moderate, and significantonly within grassland communities. Some results suggested that thesoil microbes in the studied area were controlled more by soil fertilitythan by contamination level. This is not surprising since in fieldconditions the trace metals, even at high total concentrations, mayhave low toxicity to organisms because the metals interact withother abiotic factors that influence their bioavailability (Giller et al.,2009). This study focused on determining the effects of particularsoil properties and vegetation parameters on soil microbiota, andaimed at assessing how they might compensate for the negativeeffects of metal contamination.

Habitat fertility is one of the most important factors determiningthe distribution of aboveground and belowground biota. In this anal-ysis this property was represented mainly by F1 (subsoil type) and F2(soil fertility). Both factors explained a large portion of the variationof microbial parameters. The former reflected the mining waste con-tent of the subsoil of a given site. The mining waste is a metal-richamalgam of minerals, mainly dolomite, calcite and clay minerals(Cabala and Teper, 2007). Although it may contain huge amounts ofZn, Cd and Pb, it functions as a fertilizer and/or pH neutralizer ratherthan a source of contaminants due to its high concentrations of Ca, Mgand K. Due to this high content of nutrients it supports microbialperformance (Lalande et al., 2009; Rooney and Clipson, 2009). Soilmicrobes may also benefit from C and P supply, which appears to bebetter at the mining waste sites than elsewhere (both C and P hadrelatively high loadings on F1: 0.55 and 0.41, respectively). The con-tent of these nutrients is a limiting factor in the ecosystems studied,

Fig. 1. DCA diagram of plant species data. The first and the second axes explain 16.8% and 7.9% of the variance in the plant species composition. The axes reflect the response ofspecies to changes in fertility and moisture conditions (soil nutrient content generally increases from left to right, and soil moisture increases from top to bottom). The species(crosses on diagram) used in ordination are Achillea millefolium (Achi mil), Agrostis gigantea (Agrs gig), Anthyllis vulneraria (Anty vul), Armeria maritima (Arme mar), Campanularotundifolia (Camp rot), Cardaminopsis arenosa (Carp are), Carex hirta (Carx hir), Convolvulus arvensis (Cono arv), Crepis biennis (Crep bie), Daucus carota (Dauc car), Deschampsiaflexuosa (Desc fle), Dianthus carthusianorum (Dian car), Euphrasia stricta (Eups str), Festuca ovina (Fest ovi), Galium album (Galm alb), Gypsophila fastigiata (Gyps fas), Knautia arvensis(Knau arv), Leontodon hispidus subsp. hastilis (Leon has), Leontodon hispidus subsp. hispidus (Leon his), Lotus corniculatus (Lotu cor), Melandrium album (Meld alb), Molinia caerulea(Moli cae), Orthilia secunda (Orth sec), Pimpinella saxifraga (Pimp sax), Pinus sylvestris (Pinu syl), Plantago lanceolata (Plan lan), Potentilla arenaria (Pote are), Potentilla erecta (Poteere), Ranunculus acris (Ranu acr), Rumex thyrsiflorus (Rume thy), Scabiosa ochroleuca (Scab och), Silene vulgaris (Sile vul), Thymus pulegioides (Thys pul), Valeriana officinalis (Valeoff) and Viola tricolor (Viol tri).

215A.M. Stefanowicz et al. / Science of the Total Environment 439 (2012) 211–219

so it can be even more important for microbial communities than thecontent of alkali and alkaline earth metals. The trace metals (total Zn,Cd and Pb) exerted a negative effect, revealed when microbial param-eters expressed per unit organic matter were used in the multipleregression or simple correlation analysis, but it was of secondaryimportance under good nutrient conditions. Other authors havemade similar observations (Niklińska et al., 2005; Schipper and Lee,2004): among other things, they found that soil acidity, organicmatter content and nutrient status were stronger determining factorsfor microbes than metal contamination.

Table 2Microbial characteristics of the soil for 49 study sites.

Variable Minimum Lower quartile

BR (mM CO2 kgdw−1 24 h−1) 0.19 0.60BROM (mM CO2 kgom−1 24 h−1) 8.2 14.0Cmic (g kgdw−1) 0.04 0.09CmicOM (g kgom−1) 1.51 2.36AAUCb 0.1 2.8S′b 1 17AAUCf 4.2 20.3S′f 11 51

BR— soil basal respiration; Cmic—microbial biomass; AAUC— average area under curve (microvariables corrected for organic matter content; dw — dry weight; om — organic matter; b — bac

High nutrient availability (exchangeable Ca, Mg and K) and organ-ic matter content (represented by organic C as well as total C and N)benefitted the soil microbiota, as seen in the correlation between F2and microbial parameters. This may be an indirect relationship: inmetal-contaminated soils a high concentration of organic matter oralkaline earth cations can favor microorganisms through effectiveimmobilization of trace metals (Farrell et al., 2010). On the otherhand, F2 together with F7 may also reflect moisture conditions, be-cause organic matter and clay content – having high loadings on F2and F7, respectively – are known to enhance the water-holding

Median Mean (SD) Upper quartile Maximum

1.02 1.40 (1.08) 1.82 5.2719.0 19.9 (7.4) 25.5 34.80.13 0.24 (0.26) 0.29 1.572.71 3.05 (1.26) 3.41 7.79

20.5 23.9 (21.1) 43.0 80.058 49 (28) 74 8730.7 33.2 (18.5) 47.7 74.962 61 (21) 77 91

bial activity); S′ — number of substrates metabolized (functional richness); OM— indicatesterial; f — fungal; SD — standard deviation.

Table 3Results of factor analysis performed on soil and vegetation characteristics (30 variablesfrom Table 1).

Factors Varianceexplained (%)

Habitat properties with the highest loadings (>0.5) on agiven factor

F1 19.6 Total Zn (0.88), total Cd (0.88), total Pb (0.87), total Mg(0.70), total Ca (0.69), pH (0.67), total K (0.59), total C(0.55), exchangeable Cd (0.54)

F2 17.7 Exchangeable Mg (0.87), exchangeable Ca (0.83),exchangeable K (0.78), organic C (0.72), exchangeableCd (0.67), total N (0.66), total C (0.55), exchangeable K(0.51), exchangeable Zn (0.51)

F3 13.4 Plant species richness (0.90), forb density (0.90), DCA1(−0.76), pH (0.59), DCA 2 (−0.51)

F4 9.7 Silt (0.81), available P (0.74), sand (−0.62), total P(0.55)

F5 7.9 water-soluble Cd (0.75), water-soluble Pb (0.72),water-soluble Zn (0.69)

F6 6.8 Graminoid density (0.87), plant functional diversity(0.82)

F7 5.4 Clay (0.71), sand (−0.59)

216 A.M. Stefanowicz et al. / Science of the Total Environment 439 (2012) 211–219

capacity of soil. This may be particularly important for microbial com-munities of warm, dry habitats such as the grasslands and pine forestsstudied here.

Unexpectedly, water-soluble Zn, Cd and Pb (F5) correlated posi-tively with some microbial parameters, especially the fungal compo-nent. A possible explanation is that microbes respond more to theavailability of carbon substrates (organic C, with a loading equal to0.47, made a considerable contribution to F5) than to the concentra-tion of water-soluble metals. The negative relationship between F5and BROM supports this suggestion. Another reason for such a phe-nomenon might be the high sensitivity of bacteria to metals. Fungi,which are known to be more tolerant to this type of contamination(Bååth, 1989; Rajapaksha et al., 2004; Stefanowicz et al., 2008),

Table 4Effects of habitat properties on soil microbial parameters, illustrated by standardized coeffi

Dependent variable Regressionsummary

F1Subsoil type

F2Soil fertility

F3Plant species richn

R2 P

BR 0.86 b0.001 0.61⁎⁎⁎ 0.51⁎⁎⁎ 0.34⁎⁎⁎

BROM 0.39 b0.001 0.32⁎⁎ −0.17 0.43⁎⁎⁎

Cmic 0.71 b0.001 0.31⁎⁎⁎ 0.64⁎⁎⁎ 0.27⁎⁎

CmicOM 0.24 b0.01 −0.45⁎⁎⁎ 0.21 0.31⁎

AAUCb 0.43 b0.001 −0.20 0.37⁎⁎ 0.54⁎⁎⁎

S′b 0.44 b0.001 −0.04 0.49⁎⁎⁎ 0.49⁎⁎⁎

AAUCf 0.44 b0.001 0.05 0.56⁎⁎⁎ 0.14S′f 0.45 b0.001 0.05 0.49⁎⁎⁎ 0.25⁎

Abbreviations of dependent variables are given in Table 2; R2 — adjusted R2; P— significance⁎⁎, Pb0.01; ⁎⁎⁎, Pb0.001).

Table 5Pearson's correlations between metal contamination (different forms of trace metals) and

Zntot Cdtot Pbtot Znws

BR 0.754⁎⁎⁎ 0.754⁎⁎⁎ 0.655⁎⁎⁎ 0.35BROM 0.221 0.191 0.184 −0.19Cmic 0.525⁎⁎⁎ 0.536⁎⁎⁎ 0.427⁎⁎ 0.36CmicOM −0.329⁎ −0.333⁎ −0.389⁎⁎ −0.11AAUCb −0.083 −0.075 −0.175 0.04S′b 0.103 0.111 −0.013 0.18AAUCf 0.246 0.282⁎ 0.174 0.41S′f 0.249 0.279 0.174 0.37

Abbreviations of microbial parameters are given in Table 2; tot — total; ws —water-soluble; ex⁎⁎, Pb0.01; ⁎⁎⁎, Pb0.001).

might have benefitted from weakened competitive pressure andtherefore performed better.

Unlike total and water-soluble metals, exchangeable Zn and Cdwere not extracted as a separate factor (they contribute to F1 andF2), so their direct effect on microbes could not be assessed. It isknown, however, that there may be an indirect negative relationshipbetween these variables. Examining soil mesofauna activity in thesame study plots, Kapusta et al. (2011) found that high concentra-tions of exchangeable Zn and Cd reduced enchytraeid density, ad-versely affecting microbial parameters, which were shown to bepositively dependent on enchytraeids.

Vegetation parameters such as plant diversity and coverage areoften strongly determined by soil physicochemical properties, butnot in this study. Relatively weak correlations were found betweenthese groups of variables (they were almost perfectly separated byfactor analysis). Probably, the reason behind this result is that besidestrophic status and contamination level there are other factors(uncontrolled) greatly shaping plant communities. In the Olkusz min-ing area, history of land-use, age of vegetation cover and adjacentarea type are most likely to have a strong influence on the patternof plant species occurrence.

Vegetation is an important driver of soil microbial communities.Microbes benefitted mainly from plant diversity, represented by F3.In fact, this variable reflected forb species richness (note that forbdensity had almost the same high loading on F3 as plant species rich-ness), because forbs are a highly dominant functional group in thestudy area. The factor also contained much of the variation of plantspecies composition described by DCA1 and DCA2. However, thesecharacteristics are not so important for microbes as plant species rich-ness since they contributed less to F3 and correlated with microbialparameters to a lesser degree (revealed by simple correlations notshown here).

Some authors have suggested that the positive influence of vege-tation on belowground biota might be due to plant biomass since it

cients of multiple regression (β coefficients).

essF4P content

F5Water-soluble metals

F6Plant functional diversity

F7Clay content

−0.07 0.19⁎⁎ 0.19⁎⁎ 0.24⁎⁎⁎

0.08 −0.36⁎⁎ 0.16 0.04−0.05 0.30⁎⁎⁎ 0.08 0.27⁎⁎

0.07 −0.005 −0.03 0.10−0.14 0.09 −0.09 0.06−0.11 0.14 −0.07 0.06−0.18 0.37⁎⁎ −0.11 0.06−0.17 0.40⁎⁎⁎ −0.15 0.08

level; F1–F7 — factors from Table 3; significant β coefficients are asterisked (⁎, Pb0.05;

microbial parameters.

Cdws Pbws Znex Cdex

2⁎ 0.280 0.208 0.432⁎⁎ 0.622⁎⁎⁎

7 −0.202 −0.144 −0.011 0.0737⁎⁎ 0.263 0.255 0.376⁎⁎ 0.486⁎⁎⁎

5 −0.167 0.077 −0.119 −0.1841 0.029 0.131 −0.054 0.0220 0.080 0.143 0.106 0.1911⁎⁎ 0.191 0.346⁎ 0.375⁎⁎ 0.384⁎⁎

0⁎⁎ 0.197 0.427⁎⁎ 0.308⁎ 0.331⁎

— exchangeable; significant Pearson's correlation coefficients are asterisked (⁎, Pb0.05;

Fig. 2. Mean values (±SE) of (a) soil respiration and (b) microbial biomass (variablescalculated per unit organic matter) and of (c) number of substrates utilized by bacteriafor 6 categories of plant species richness.

217A.M. Stefanowicz et al. / Science of the Total Environment 439 (2012) 211–219

usually increases together with plant species richness (Hooper et al.,2005; Marquard et al., 2009; Spehn et al., 2005). In this study such amechanism probably did not operate. Plant cover, which can be con-sidered a rough indicator of plant biomass, neither was correlatedwith the number of plant species (these variables are in separatefactors) nor had a significant impact on microbial parameters.

Table 6Number of significant (pb0.05) Spearman correlations (positive/negative) between factorsby bacteria and fungi.

Chemical guilds N Bacteria

F1 F2 F3 F4 F5 F6

Amino acids 20 1/3 4/0 13/0Amines/amides 6 2/0 5/0Carboxylic acids 24 0/2 6/0 9/0 0/1 3/0 0/1Carbohydrates 28 1/1 12/0 5/0 0/1 1/0Miscellaneous 12 2/0 3/0 0/1Polymers 5 1/0 0/1 1/0Total 95 2/6 27/0 35/0 0/3 4/0 1/2

N — total number of substrates on a Biolog plate; F1–F7 — factors from Table 3.

The assessed effect of plant species richness was clearly positivebut not to the same extent within the whole microbial community.Plants seemed to have a much stronger effect on bacteria than onfungi. Some studies (Rousk et al., 2008, 2010) indicated that bacteriaare effective competitors and are able to control fungal growth.Therefore, an agent potentially favorable to fungi may show no effecton their performance if at the same time it promotes bacteria. Possi-bly the antagonistic relationship between these two groups ofmicrobiota was responsible for the weak response of the fungi toplant species richness.

Different mechanisms are proposed to explain the positive rela-tionship between plant diversity and ecosystem functioning. Manypapers provide evidence for the redundancy hypothesis (Naeem andLi, 1997; Walker, 1992), which assumes that most species within anecosystem duplicate functions of others, particularly those of speciesbelonging to the same functional group. In other words, the numberof functional groups in a plant mixture is of greater importance thanthe number of species. The study findings do not support this. Plantfunctional diversity, extracted as a separate factor (F6), was foundto be almost negligible in its effect on microbial parameters.

Species singularity seems a better explanation of the present re-sults. According to the singular hypothesis, all species contribute toecosystem functioning because each of them is unique in its function-al traits (Eisenhauer et al., 2010; Naeem et al., 2002). For example,plant species differ in the quality and quantity of their litter or theirroot exudates, so a continuous increase in species number shouldcause continuous enhancement of microbial performance due to animprovement of the food supply (this is in line with observationsmade in this study— the ability of microbial community to utilize var-ious Biolog substrates increased with increasing plant species rich-ness). The relationship is not expected to be solely linear since it isknown that species contributions differ considerably. In particular,the addition of a keystone species to a plant assemblage may provokesubstantial changes in the system. Most studies carried out in exper-imental grassland ecosystems show that some legumes can play sucha role (Milcu et al., 2008; Mulder et al., 2002; Spehn et al., 2002;Temperton et al., 2007). In the present survey this group of plantsconsisted mainly of species occurring sporadically (on less than 5%of sites). Only two legumes – Lotus corniculatus and Anthyllisvulneraria – were recorded more frequently (on ca. 20% of the sites)and would be important components of the investigated plant com-munities. The latter one, which commonly grows on metal-rich soilsand has been recognized as an efficient nitrogen fixer (Mahieu etal., 2011), is especially likely to participate in the plant diversityeffect.

The influence of plant diversity on microbial performance hasbeen investigated repeatedly. The reported positive relationshipscome mainly from experiments with manipulated plant communities(De Deyn et al., 2011; Kowalchuk et al., 2002; Lamb et al., 2011;Loranger-Merciris et al., 2006; Milcu et al., 2008; Spehn et al., 2000;Stephan et al., 2000; Zak et al., 2003). The very few studies using nat-ural gradients of plant species richness have yielded similar results

representing habitat properties and Biolog substrates utilized (AUCs divided by AAUC)

Fungi

F7 F1 F2 F3 F4 F5 F6 F7

0/2 4/1 1/00/1 0/12/0 5/0 2/0 1/0 1/0 0/12/2 6/1 5/1 1/3 5/0 0/1 1/10/1 2/0 1/0 2/0 2/0

1/0 1/0 2/01/0 4/6 17/2 10/2 2/3 10/0 0/2 3/1

218 A.M. Stefanowicz et al. / Science of the Total Environment 439 (2012) 211–219

(Eisenhauer et al., 2011; Rodríguez-Loinaz et al., 2008). One probablereason for the paucity of natural studies is the difficulty in separatingthe effects of multiple factors. For example, in research on the influ-ence of plant diversity and soil physicochemical properties on theactivity of soil enzymes, Rodríguez-Loinaz et al. (2008) found bothtypes of predictors to be positively correlated with some microbialparameters, but they were also inter-correlated, making causal rela-tionships hard to identify. The present study also experienced someinter-correlations between explanatory variables; specifically, it wasnot possible to determine the magnitude of the impact of metal con-tamination because it was masked by the opposite influences of otherhabitat properties, especially soil fertility. It was, however, possible todistinguish the effect of vegetation from the effect of soil chemistrysince these variables were in separate (orthogonal) factors. In themanner it was possible to disentangle the effects of important plantcommunity parameters – species richness and functional diversity –

from each other. Thus this study adds valuable findings to the sparsebody of field-based experimental work in this area.

A strong negative impact of contamination on microbial commu-nities was not found, neither on their activity, nor on the functionalstructure. In the study area, trace metals are deposited mainly withmining waste, which contains large amounts of dolomite, calciteand clay minerals. Elevated pH and the high concentration of nutri-ents at highly contaminated sites effectively ameliorate the toxiceffects of metals on the belowground ecosystem and promote micro-bial growth and activity. Plant species richness is another importantfactor determining the performance of soil microbes. It was foundthat plant species richness positively affected almost all microbialparameters but was more beneficial to bacteria than to fungi. Plantfunctional diversity, based on forbs, legumes and graminoids, wasshown to be negligible in its effect on microbes. These findings shouldbe helpful in understanding the interactions between abovegroundand belowground biota developing in contaminated conditions. Theresults imply that the quality of heavy metal contaminated soils ashabitats for organisms may be improved not only by liming or addingorganic material, but also by maintaining high diversity of the plantcover.

Acknowledgments

The studywas performedunder project FMEEAPL0265 supported bya grant from Iceland, Liechtenstein and Norway through the FinancialMechanism of the European Economic Area. The statutory fund of theInstitute of Botany of the Polish Academy of Sciences also providedpartial funding. We thank Michael Jacobs for his final-editing of themanuscript before submission and two anonymous reviewers for theirvaluable comments on the manuscript. We are grateful to the staff ofthe BolesławMine andMetallurgical Plant and the Olkusz Forest Inspec-torate, and to the authorities of Bolesław and Bukowno municipalities,for their kind cooperation.

References

Anderson JPE, Domsch KH. A physiological method for the quantitative measurementof microbial biomass in soils. Soil Biol Biochem 1978;10:215–21.

Avidano L, Gamalero E, Cossa GP, Carraro E. Characterization of soil health in an Italianpolluted site by using microorganisms as bioindicators. Appl Soil Ecol 2005;30:21–33.

Bååth E. Effects of heavymetals in soil onmicrobial processes and populations (a review).Water Air Soil Pollut 1989;47:335–79.

Balogh J, Pintér K, Fóti S, Cserhalmi D, Papp M, Nagy Z. Dependence of soil respiration onsoil moisture, clay content, soil organic matter, and CO2 uptake in dry grasslands. SoilBiol Biochem 2011;43:1006–13.

Batty LC, Hallberg KB. Ecology of industrial pollution. Cambridge: Cambridge UniversityPress; 2010.

Boivin M-EY, Greve GD, García-Meza JV, Massieux B, Sprenger W, Kraak MHS, et al.Algal-bacterial interactions in metal contaminated floodplain sediments. EnvironPollut 2007;145:884–94.

Brookes P. The soil microbial biomass: concept, measurement and applications in soilecosystem research. Microbes Environ 2001;16:131–40.

Buyer JS, Roberts DP, Millner P, Russek-Cohen E. Analysis of fungal communities by solecarbon source utilization profiles. J Microbiol Methods 2001;45:53–60.

Cabala J, Teper L. Metalliferous constituents of rhizosphere soils contaminated byZn–Pb mining in southern Poland. Water Air Soil Pollut 2007;178:351–62.

Cabala J, Krupa P, Misz-KennanM. Heavy metals in mycorrhizal rhizospheres contaminatedby Zn–Pb mining and smelting around Olkusz in southern Poland. Water Air Soil Pollut2009;199:139–49.

Chodak M, Niklińska M. Effect of texture and tree species on microbial properties ofmine soils. Appl Soil Ecol 2010;46:268–75.

Commission of the European Communities Council Directive of 12 June 1986 on theprotection of the environment, and in particular of the soil, when sewage sludgeis used in agriculture. Off J Eur Commun 1986;No L181(86/278/EEC):6-12.

De Deyn GB, Quirk H, Bardgett RD. Plant species richness, identity and productivitydifferentially influence key groups of microbes in grassland soils of contrastingfertility. Biol Lett 2011;7:75–8.

Dobranic JK, Zak JC. A microtiter plate procedure for evaluating fungal functional diversity.Mycologia 1999;91:756–65.

EisenhauerN, BeβlerH, Engels C, GleixnerG,HabekostM,MilcuA, et al. Plant diversity effectson soil microorganisms support the singular hypothesis. Ecology 2010;91:485–96.

Eisenhauer N, Yee K, Johnson EA, Maraun M, Parkinson D, Straube D, et al. Positiverelationship between herbaceous layer diversity and the performance of soilbiota in a temperate forest. Soil Biol Biochem 2011;43:462–5.

Farrell M, Griffith GW, Hobbs PJ, Perkins WT, Jones DL. Microbial diversity and activityare increased by compost amendment of metal-contaminated soil. FEMS MicrobiolEcol 2010;71:94-105.

Garbeva P, van Elsas J, van Veen J. Rhizosphere microbial community and its responseto plant species and soil history. Plant Soil 2008;302:19–32.

García-Palacios P, Bowker MA, Chapman SJ, Maestre FT, Soliveres S, Gallardo A, et al.Early-successional vegetation changes after roadside prairie restoration modifyprocesses related with soil functioning by changing microbial functional diversity.Soil Biol Biochem 2011;43:1245–53.

Garland JL, Mills AL. Classification and characterization of heterotrophic microbial commu-nities on the basis of patterns of community-level sole-carbon-source utilization. ApplEnviron Microbiol 1991;57:2351–9.

Giller KE, Witter E, McGrath SP. Heavy metals and soil microbes. Soil Biol Biochem2009;41:2031–7.

Gobran GR, Clegg S. A conceptual model for nutrient availability in the mineral soil–rootsystem. Can J Soil Sci 1996;76:125–31.

Grayston SJ, Vaughan D, Jones D. Rhizosphere carbon flow in trees, in comparison withannual plants: the importance of root exudation and its impact on microbial activityand nutrient availability. Appl Soil Ecol 1997;5:29–56.

Hinsinger P, Plassard C, Jaillard B. Rhizosphere: a new frontier for soil biogeochemistry.J Geochem Explor 2006;88:210–3.

Hooper DU, Bignell DE, Brown VK, Brussaard L, Dangerfield JM, Wall DH, et al. Interactionsbetween aboveground and belowground biodiversity in terrestrial ecosystems:patterns, mechanisms, and feedbacks. Bioscience 2000;50:1049–61.

Hooper DU, Chapin FS, Ewel JJ, Hector A, Inchausti P, Lavorel S, et al. Effects of biodiversityon ecosystem functioning: a consensus of current knowledge. Ecol Monogr 2005;75:3-35.

Kapusta P, Szarek-Łukaszewska G, Stefanowicz AM. Direct and indirect effects of metalcontamination on soil biota in a Zn–Pb post-mining and smelting area (S Poland).Environ Pollut 2011;159:1516–22.

Kowalchuk GA, Buma DS, de Boer W, Klinkhamer PGL, van Veen JA. Effects ofabove-ground plant species composition and diversity on the diversity ofsoil-borne microorganisms. Anton Leeuw Int J G 2002;81:509–20.

Kraus GF, Druzhinina I, Gams W, Bissett J, Zafari D, Szakacs G, et al. Trichodermabrevicompactum sp. nov. Mycologia 2004;96:1059–73.

Lalande R, Gagnon B, Royer I. Impact of natural or industrial liming materials on soilproperties and microbial activity. Can J Soil Sci 2009;89:209–22.

Lamb E, Kennedy N, Siciliano S. Effects of plant species richness and evenness on soilmicrobial community diversity and function. Plant Soil 2011;338:483–95.

Lauber CL, Strickland MS, Bradford MA, Fierer N. The influence of soil properties on thestructure of bacterial and fungal communities across land-use types. Soil BiolBiochem 2008;40:2407–15.

Lepš J, Šmilauer P. Multivariate analysis of ecological data using CANOCO. Cambridge:Cambridge University Press; 2003.

Liu Z, Fu B, Zheng X, Liu G. Plant biomass, soil water content and soil N:P ratio regulatingsoil microbial functional diversity in a temperate steppe: a regional scale study. SoilBiol Biochem 2010;42:445–50.

Loranger-Merciris G, Barthes L, Gastine A, Leadley P. Rapid effects of plant speciesdiversity and identity on soil microbial communities in experimental grasslandecosystems. Soil Biol Biochem 2006;38:2336–43.

Mahieu S, Frérot H, Vidal C, Galiana A, Heulin K, Maure L, et al. Anthyllisvulneraria/Mesorhizobium metallidurans, an efficient symbiotic nitrogen fixingassociation able to grow in mine tailings highly contaminated by Zn, Pb and Cd.Plant Soil 2011;342:405–17.

Marquard E, Weigelt A, Temperton VM, Roscher C, Schumacher J, Buchmann N, et al.Plant species richness and functional composition drive overyielding in asix-year grassland experiment. Ecology 2009;90:3290–302.

Marschner P, Crowley D, Yang CH. Development of specific rhizosphere bacterialcommunities in relation to plant species, nutrition and soil type. Plant Soil2004;261:199–208.

McGrath SP, Chaudri AM, Giller KE. Long-term effects of metals in sewage sludge onsoils, microorganisms and plants. J Ind Microbiol Biotechnol 1995;14:94-104.

219A.M. Stefanowicz et al. / Science of the Total Environment 439 (2012) 211–219

Milcu A, Partsch S, Scherber C, Weisser WW, Scheu S. Earthworms and legumes controllitter decomposition in a plant diversity gradient. Ecology 2008;89:1872–82.

Mirek Z, Piękoś-Mirkowa H, Zając A, Zając M. Flowering plants and pteridophytes ofPoland — a checklist. Kraków: W. Szafer Institute of Botany, Polish Academy ofSciences; 2002.

Mulder CPH, Jumpponen A, Högberg P, Huss-Danell K. How plant diversity and legumes af-fect nitrogen dynamics in experimental grassland communities. Oecologia 2002;133:412–21.

Naeem S, Li S. Biodiversity enhances ecosystem reliability. Nature 1997;390:507–9.Naeem S, Loreau M, Inchausti P. Biodiversity and ecosystem functioning: the emer-

gence of a synthetic ecological framework. In: Naeem S, Loreau M, Inchausti P,editors. Biodiversity and ecosystem functioning. Oxford: Oxford UniversityPress; 2002. p. 3-11.

Niklińska M, Chodak M, Laskowski R. Characterization of the forest humus microbialcommunity in a heavy metal polluted area. Soil Biol Biochem 2005;37:2185–94.

Olko A, Abratowska A, Żyłkowska J, Wierzbicka M, Tukiendorf A. Armeria maritimafrom a calamine heap — initial studies on physiologic–metabolic adaptations tometal-enriched soil. Ecotoxicol Environ Saf 2008;69:209–18.

Rajapaksha RMCP, Tobor-Kapłon MA, Bååth E. Metal toxicity affects fungal and bacterialactivities in soil differently. Appl Environ Microbiol 2004;70:2966–73.

Ramsey PW, Rillig MC, Feris KP, Gordon NS, Moore JN, Holben WE, et al. Relationshipbetween communities and processes; new insights from a field study of a contami-nated ecosystem. Ecol Lett 2005;8:1201–10.

Rodríguez-Loinaz G, Onaindia M, Amezaga I, Mijangos I, Garbisu C. Relationshipbetween vegetation diversity and soil functional diversity in native mixed-oakforests. Soil Biol Biochem 2008;40:49–60.

Rooney D, Clipson N. Phosphate addition and plant species alters microbial communitystructure in acidic upland grassland soil. Microb Ecol 2009;57:4-13.

Rosenvald K, Kuznetsova T, Ostonen I, Truu M, Truu J, Uri V, et al. Rhizosphere effectand fine-root morphological adaptations in a chronosequence of silver birch standson reclaimed oil shale post-mining areas. Ecol Eng 2011;37:1027–34.

Rousk J, Demoling LA, Bahr A, Bååth E. Examining the fungal and bacterial niche overlapusing selective inhibitors in soil. FEMS Microbiol Ecol 2008;63:350–8.

Rousk J, Brookes PC, Bååth E. Investigating the mechanisms for the opposing pH relation-ships of fungal and bacterial growth in soil. Soil Biol Biochem 2010;42:926–34.

Sanaullah M, Blagodatskaya E, Chabbi A, Rumpel C, Kuzyakov Y. Drought effects on mi-crobial biomass and enzyme activities in the rhizosphere of grasses depend onplant community composition. Appl Soil Ecol 2011;48:38–44.

Schipper LA, Lee WG. Microbial biomass, respiration and diversity in ultramafic soils ofWest Dome, New Zealand. Plant Soil 2004;262:151–8.

Shukurov N, Pen-Mouratov S, Steinberger Y. The impact of the Almalyk Industrial Complexon soil chemical and biological properties. Environ Pollut 2005;136:331–40.

Spehn EM, Joshi J, Schmid B, Alphei J, Körner C. Plant diversity effects on soil heterotrophicactivity in experimental grassland ecosystems. Plant Soil 2000;224:217–30.

Spehn EM, Scherer-Lorenzen M, Schmid B, Hector A, Caldeira MC, Dimitrakopoulos PG,et al. The role of legumes as a component of biodiversity in a cross-European studyof grassland biomass nitrogen. Oikos 2002;98:205–18.

Spehn EM, Hector A, Joshi J, Scherer-Lorenzen M, Schmid B, Bazeley-White E, et al.Ecosystem effects of biodiversity manipulations in European grasslands. EcolMonogr 2005;75:37–63.

Stefanowicz AM, Niklińska M, Laskowski R. Metals affect soil bacterial and fungalfunctional diversity differently. Environ Toxicol Chem 2008;27:591–8.

Stefanowicz AM, Niklińska M, Kapusta P, Szarek-Łukaszewska G. Pine forest andgrassland differently influence the response of soil microbial communities tometal contamination. Sci Total Environ 2010;408:6134–41.

Stephan A, Meyer AH, Schmid B. Plant diversity affects culturable soil bacteria in exper-imental grassland communities. J Ecol 2000;88:988–98.

Teklay T, Shi Z, Attaeian B, Chang SX. Temperature and substrate effects on C & Nmineralization and microbial community function of soils from a hybrid poplarchronosequence. Appl Soil Ecol 2010;46:413–21.

Temperton V, Mwangi P, Scherer-Lorenzen M, Schmid B, Buchmann N. Positive interac-tions between nitrogen-fixing legumes and four different neighbouring species ina biodiversity experiment. Oecologia 2007;151:190–205.

ter Braak CJF, Šmilauer P. CANOCO reference manual and CanoDraw for Windowsuser's guide: software for canonical community ordination (version 4.5). Ithaca:Microcomputer Power; 2002.

Vig K, Megharaj M, Sethunathan N, Naidu R. Bioavailability and toxicity of cadmium tomicroorganisms and their activities in soil: a review. Adv Environ Res 2003;8:121–35.

Walker BH. Biodiversity and ecological redundancy. Conserv Biol 1992;6:18–23.Wardle DA, Bardgett RD, Klironomos JN, Setälä H, van der Putten WH, Wall DH.

Ecological linkages between aboveground and belowground biota. Science2004;304:1629–33.

Wierzbicka M, Panufnik D. The adaptation of Silene vulgaris to growth on a calaminewaste heap (S. Poland). Environ Pollut 1998;101:415–26.

Wierzbicka M, Pielichowska M. Adaptation of Biscutella laevigata L, a metalhyperaccumulator, to growth on a zinc–lead waste heap in southern Poland. I:differences between waste-heap and mountain populations. Chemosphere2004;54:1663–74.

Zak DR, Holmes WE, White DC, Peacock AD, Tilman D. Plant diversity, soil microbialcommunities, and ecosystem function: are there any links? Ecology 2003;84:2042–50.

Załęcka R, Wierzbicka M. The adaptation of Dianthus carthusianorum L. (Caryophyllaceae)to growth on a zinc–lead heap in southern Poland. Plant Soil 2002;246:249–57.

Zhang Y, Zhang H-W, Su Z-C, Zhang C-G. Soil microbial characteristics under long-termheavy metal stress: a case study in Zhangshi wastewater irrigation area, Shenyang.Pedosphere 2008;18:1-10.