SEARCH

SEARCH BY CITATION

Keywords:

  • denitrifiers;
  • community structure;
  • function;
  • plant communities

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

To explore potential links between plant communities, soil denitrifiers and denitrifier function, the impact of presence, diversity (i.e. species richness) and plant combination on nirK-type denitrifier community composition and on denitrifier activity was studied in artificial grassland plant assemblages over two consecutive years. Mesocosms containing zero, four and eight species and different combinations of two species were set up. Differences in denitrifier community composition were analysed by canonical correspondence analyses following terminal restriction fragment length polymorphism analysis of PCR-amplified nirK gene fragments coding for the copper-containing nitrite reductase. As a measure of denitrifier function, denitrifier enzyme activity (DEA) was determined in the soil samples. The presence as well as the combination of plants and sampling time, but not plant diversity, affected the composition of the nirK-type denitrifier community and DEA. Denitrifier activity significantly increased in the presence of plants, especially when they were growing during summer and autumn. Overall, we found a strong and direct linkage of denitrifier community composition and functioning, but also that plants had additional effects on denitrifier function that could not be solely explained by their effects on nirK-type denitrifier community composition.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

There is growing interest in how plant communities and soil microbial community structure and functioning are linked. Of particular interest are the implications that this inter-relationship may have on ecosystem functioning, because microorganisms are the crucial mediators of nutrient (e.g. carbon, nitrogen) and greenhouse gas (e.g. CO2, methane, N2O) cycling. It was shown in several studies that individual plant species can influence microbial communities, in particular in their rhizosphere (Grayston et al., 1998; Wardle et al., 2003; Carney & Matson, 2006; Patra et al., 2006; Bremer et al., 2007). Furthermore, plant species diversity may play a significant role in controlling ecosystem processes and overall ecosystem functioning (Beierkuhnlein & Jentsch, 2005; Hector et al., 2005). However, few studies have investigated the effects of plant community composition and diversity on soil microbial communities and their functioning, and their results have been conflicting (Broughton & Gross, 2000; Brodie et al., 2002; Kowalchuk et al., 2002; Johnson et al., 2003; Wardle et al., 2003; Zak et al., 2003; Carney & Matson, 2006). With regard to nitrogen cycling, no relation between plant diversity and soil microbial community diversity and composition was found in studies of the functional group of β-ammonia oxidizers (Kowalchuk et al., 2000; Carney et al., 2004). However, it has been suggested that other microbial functional groups that interact more directly with plant roots than ammonia-oxidizers might be influenced more strongly by plant diversity and species composition than those with less direct root associations (Kowalchuk et al., 2002).

The objectives of this study were to explore in a manipulative experiment the effects of the presence of plants, of different plant combinations and of plant species diversity on the composition and functioning of denitrifier communities in soil. Although most denitrifiers are not obligately associated with plant roots, plants may influence this functional community in the rhizosphere by affecting the availability of oxygen, nitrate and soil organic carbon. Microbial denitrification is a crucial process in nitrogen cycling and a major source of nitrous oxide (N2O) (Conrad, 1996), which is known to be a potent greenhouse gas that contributes to ozone depletion and global warming (Crutzen, 1970; Dickinson & Cicerone, 1986). Over half of the total N2O input into the atmosphere is due to emissions from soils (Conrad, 1996). However, it is still largely unknown whether denitrifier functioning and community composition are interlinked and how the composition and activity of this soil community is influenced by plant communities. To our knowledge, only one pioneering study has explored the effects of plant species diversity and composition on nitrogen cycling and the trace gas balance of soils (Niklaus et al., 2006). This study found that N2O emissions decreased with increasing plant diversity, but increased in the presence of legumes. Plant community composition and its interaction with soil type strongly affected N2O production rates. However, the study did not investigate the composition of microbial communities involved in N2O metabolism, although denitrifier community composition may be a major determinant of N2O production and thus of ecosystem functioning. The organisms involved in the denitrification process are phylogenetically diverse, but are characterized by the occurrence of two structurally different but functionally equivalent forms of the key enzyme of the denitrification process: a copper- and a heme cd1-containing nitrite reductase encoded by the genes nirK and nirS, respectively. Both genes have frequently been targeted as functional marker genes in culture-independent approaches to explore denitrifier community composition in the environment. We focused on the analysis of nirK-type denitrifier communities, because in several former studies nirK could be more readily amplified from soils than nirS (Prieméet al., 2002; Wolsing & Priemé, 2004; Sharma et al., 2005). Moreover, nirK was used as a functional marker gene in a preceding study to unravel a possible influence of individual plant species on nirK-type denitrifier community composition (Bremer et al., 2007). In the present study, we set up mesocosms that were either devoid of plants or planted with different combinations of nonleguminous plants typical for temperate grasslands, and that varied in diversity (two, four and eight species). Soil denitrifier communities were structurally and functionally characterized by terminal restriction fragment length polymorphism (T-RFLP) fingerprints and potential denitrification activity, respectively. Subsequent statistical analysis aimed at unravelling potential inter-relationships between above-ground plant species composition and diversity and the composition and functioning of this soil microbial community.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Experimental set-up and soil sampling

In autumn 2001, 28 mesocosms (lysimeters, 1.3 × 1.3 × 1.3 m) were arranged randomly in the Ecological Botanical Garden of the University of Bayreuth and kept under ambient environmental conditions for the duration of the experiment. The soil to fill the lysimeters had been taken from a meadow next to the Ecological Botanical Garden and was a stagnic gleysol developed on sandstone. It had the following characteristics: inline image 4.9, NH4+-N(1 M KCl) 19.6 mg kg−1, NO3-N(1 M KCl) 27.0 mg kg−1, and consisted of 7% sand, 78% silt and 16% clay. The soil was thoroughly mixed and steamed (12 h at 100 °C) to kill weed seeds. In the lysimeters, eight plant species (four grasses: Alopecurus pratensis, Anthoxanthum odoratum, Arrhenatherum elatius and Holcus lanatus, and four forbs: Geranium pratense, Plantago lanceolata, Ranunculus acris and Taraxacum officinale) were grown in different combinations and with different levels of plant diversity (i.e. number of plant species). The plant combinations were combinations of two plant species (2a, H. lanatus with A. elatius; 2b, H. lanatus with G. pratense; 2c, H. lanatus with P. lanceolata), a combination of four plant species (4, H. lanatus, A. elatius, G. pratense and P. lanceolata) and a combination of eight plant species (8, all species mentioned). Five replicate mesocosms were set up for each combination and diversity level. Three replicate mesocosms remained unplanted as a control (0). The plants were grown for 3 years, watered if necessary and weeds were removed manually. In summer (June), autumn (September) and winter (December) 2003, and in summer (June) and autumn (September) 2004, four topsoil samples were randomly taken from each mesocosm (three in winter 2003) with a corer to a depth of 7 cm (about 50 g soil), pooled, thoroughly mixed and homogenized with a spoon. All samples were taken for molecular analysis of nirK-type denitrifier communities, but nitrate concentration and denitrifier enzyme activity (DEA) was measured only for samples taken in winter 2003, summer 2004 and autumn 2004, as described below. For DNA extraction and T-RFLP analyses, about 20 g of the soil sample was transferred to a sterile 50-mL Falcon tube and immediately stored at −20 °C. The remainder of the soil sample was transferred to a plastic bag and stored at 4 °C for 5 days before DEA was measured. All soil samples contained plant roots, because the high root density of the plants prevented the separation of rhizosphere and bulk soil. The soil samples that had been stored at 4 °C were sieved (<2 mm) to homogenize the samples and remove plant roots before DEA was measured.

DNA extraction, PCR amplification of nirK and T-RFLP analysis of nirK-type denitrifiers

These procedures were carried out as described elsewhere (Bremer et al., 2007). Briefly, from 0.5 g soil of each mesocosm, DNA was extracted by bead beating and purified by phenol–chloroform extraction, followed by isopropanol precipitation and cleanup with the Wizard® DNA Clean-up-System (Promega, Mannheim, Germany). Terminally labelled nirK fragments were amplified from these DNA extracts using primers nirK1F and nirK5R (5′-end 6-carboxyfluorescein-labelled) and the conditions described previously (Bremer et al., 2007). For each sample, four PCRs were pooled to minimize PCR artefacts, and amplicons of the expected size were gel-purified using the QIAquick gel extraction kit (Qiagen GmbH, Hilden, Germany). For T-RFLP analysis, PCR products hydrolysed with the restriction endonuclease HaeIII were analysed on an automated sequencer. The lengths of the fluorescently labelled terminal restriction fragments (T-RFs) were defined by comparison with an internal length standard (Internal Lane Standard 600; Promega). Because T-RFs can vary slightly in size, T-RFLP patterns were inspected visually and peak size differences of one or two base pairs were confirmed by comparing the respective peaks of all patterns. Peaks with a fluorescence of 50 U over the background fluorescence and larger than 60 bp were analysed by peak height. The relative abundances of T-RFs in a sample, given in percent, were calculated after normalization of peak heights in an iterative standardization procedure (Dunbar et al., 2001).

Nitrate concentrations

Nitrate concentrations in soil were assessed with an ion chromatograph (IC, Sykam, Fürstenfeldbruck) after extraction from soil. An amount of soil equal to 2 g dry weight was suspended in 10 mL distilled and ultrapurified water and shaken at 150 r.p.m. for 1 h at 4 °C. Afterwards the soil suspension was filtered (round filter 2095, Schleicher & Schuell GmbH, Dassel, Germany) and nitrate was measured with an ion chromatograph. NaNO3 (1 mM) was used as a standard. Data were evaluated with the software peaksimple (version 2.66, SRI Instruments, Torrence).

DEA method

For the DEA assay (Tiedje, 1994), an amount of soil equal to 10 g dry weight was placed into serum bottles. The slurries were amended with 25 mL filtered tap water that contained glucose, potassium nitrate (both 10 mM) and 0.1 g L−1 chloramphenicol (Murray & Knowles, 1999). The serum bottles were capped with butyl stoppers, oxygen was removed by flushing the bottles with N2 for 30 min and the pressure was adjusted to normal pressure. In addition, the soil slurries were provided with a volume of acetylene that equals 10% acetylene in the gas phase to inhibit N2O reduction. Soil slurries were incubated at 25 °C and shaken on a horizontal shaker at 200 r.p.m. Gas samples were taken hourly from 0 to 5 h with Pressure-Lock-Syringes® (VICI, Baton Rouge, LA). The syringes were flushed with N2 before gas samples were taken as suggested (Murray & Knowles, 2004). Moreover, slurries were manually shaken to equilibrate between the gas and the liquid phase of the slurries. N2O was measured with a gas chromatograph (Carlo Erba Instruments, GC 8000) connected to a 63Ni-electron capture detector. Data were evaluated with the software peaksimple.

Statistical analyses

Statistical analyses were performed with canoco 4.5 and spss 12.0 for Windows. T-RFs of different lengths were considered to be indicative of different nirK operational taxonomic units (OTUs) present in a sample, and the relative peak heights were used as a measure of the relative abundance of nirK-OTUs. Using ordination techniques, the effects of the experimental factors plant diversity (two, four or eight species), plant treatment (the combinations of plants and absence of plants), plant presence (the contrast between mesocosms with and without plants), plant combination (the various combinations of plants), time and their interactions on T-RFLP profiles were explored. After an initial detrended correspondence analysis (CA) had indicated that a unimodal response model was more appropriate than a linear model, the data were analysed by CA and canonical correspondence analysis (CCA). CA is a method to reduce the dimensionality of a multivariate dataset and ordinate samples in an ordination space that corresponds best to the dissimilarities in species composition, i.e. samples with a similar community composition are placed closer together and samples with a more dissimilar community are positioned further apart. In our case, the relative abundance of the T-RFs obtained from each replicate mesocosm was used as species abundance. CCA is a method of direct gradient analysis that relates variation in community composition to environmental variation (Ter Braak, 1986). Because of the nested design of the experiment, covariables and dummy variables were included in the analyses. Covariables are concomitant variables whose effect is eliminated when analysing the effects of the variables of interest. Dummy variables are nominal variables defined as 1 or 0 that code for the levels of a factor.

First, single CAs were performed with the T-RFLP datasets for the five sampling times to view the structure of the data and the quality of replication. Second, several CCAs were performed to analyse the effect of the factors of interest. These analyses reflected the nested design of the experiment (Zar, 1999) and thus had different numbers of replicates. The effect of plant diversity (i.e. number of plant species) was tested using a dataset consisting of average OTU abundances for the individual plant combinations. The effect of plant treatment and plant presence on nirK-OTUs was tested using a dataset consisting of average OTU abundances for the individual mesocosms. To test the effect of plant combination, the same dataset was used, but with plant presence as a dummy-coded covariate. The effect of time on nirK-OTU abundance was tested using the dataset containing the measurements taken for the individual mesocosms at the five sampling times and with the variation due to mesocosms eliminated. The effect of the interaction between time and plant treatment on nirK-OTU abundance was tested using the dataset containing the measurements taken for the individual mesocosms at the five sampling times and with the variation due to mesocosms and time eliminated. Monte Carlo permutation tests (based on 9999 random permutations) as available in canoco 4.5 (Ter Braak & Šmilauer, 2002) were used to test the hypothesis that the relative abundances of nirK-OTUs were related to the factors of interest.

The effects of the experimental factors plant diversity, plant treatment, sampling time and the interactions of interest on the measured variables, nitrate concentration and DEA, were studied by nested anova (Zar, 1999). The effects of plant diversity were tested against the variation among the five plant combinations, that of plant treatment against the variation among the individual mesocosms, and the effect of time against the residual. The effect of plant treatment was divided into two orthogonal contrasts: a contrast between mesocosms with and without plants present and the effect of the different plant combinations.

A possible direct effect of plant treatment on DEA that was not mediated through its effect on denitrifier community composition was studied by controlling for denitrifier community composition by including sample scores along the first and the second ordination axes (calculated in a CA of the T-RFs from sampling dates for which data on enzyme activity were available) as covariates in a nested general linear model.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Effects of experimental factors on the community composition of nirK-type denitrifiers

Gene fragments (nirK) were successfully amplified from the soil of all plant mesocosms, indicating the presence of nirK-type denitrifier communities. Subsequently, community composition was resolved by T-RFLP analysis of nirK amplicons. When DNA from three different samples was extracted several times, the coefficient of variation of the relative abundance of single T-RFs was <1% (data not shown). Thus, the reproducibility of the procedure was high and therefore differences in the community profiles were considered to represent true variation in nirK-type denitrifier community composition. In total, 59 different T-RFs occurred across all soil samples at all sampling dates with a minimum of four and a maximum of 15 different peaks. The T-RFLP dataset was first analysed by CA to view the similarity of T-RFLP profiles from replicate mesocosms at the individual sampling times (see Supporting Information, Fig. S1a–e). Several profiles from replicate mesocosms of a given sampling time were similar and thus clustered closely together in the ordination diagram, for example those obtained for the lysimeters planted with combinations of H. lanatus and A. elatius (2a1–5; see Fig. S1b) and of H. lanatus and P. lanceolata (2c1–5; see Fig. S1b) in autumn 2003. In contrast, for some treatments, the biological variability among individual mesocosms was rather high (e.g. the unplanted lysimeters in summer 2004; see Fig. S1d).

The T-RFLP dataset was then analysed by CCA to evaluate the influence of the factors of interest on the composition of the denitrifier communities. Plant diversity (i.e. the number of plant species) had no effect on the composition of the nirK-type denitrifier community in the soil [trace (sum of all canonical eigenvalues)=0.096, F=6.0, P=0.13; Monte Carlo permutation test within CCA]. However, the statistical power to test this hypothesis was rather low because only three levels of diversity were used in the experiment (two, four and eight plant species) and relatively few replicates (mesocosms with different plant species composition). In contrast to diversity, the different plant treatments significantly affected the composition of the microbial community (trace=0.144, F=2.5, P=0.0001). The effect of plant treatment on the nirK-type denitrifier composition was due to the effect of the presence of plants [i.e. the dissimilarity of unplanted and planted mesocosm soil samples (Fig. 1a); trace=0.077, F=6.3, P=0.0006] and the differences between the various plant combinations (Fig. 1a; trace=0.067, F=1.5, P=0.0004).

image

Figure 1.  (a) CCA ordination plot for the effect of plant treatment on the composition of the nirK-type denitrifier community in soil, based on the relative abundance of nirK T-RFs from the soil of plant mesocosms. Filled circles indicate plant treatments. The eigenvalues of the first and second axes in the ordination diagram are as follows: λ1=0.078, λ2=0.025. Plant treatments: 0, unplanted mesocosms; 2a, planted with Holcus lanatus and Arrhenatherum elatius; 2b, planted with H. lanatus and Geranium pratense; 2c, planted with H. lanatus and Plantago lanceolata; 4, planted with H. lanatus, A. elatius, G. pratense and P. lanceolata; 8, like 4, but in addition planted with Alopecurus pratensis, Anthoxanthum odoratum, Ranunculus acris and Taraxacum officinale. (b) CCA ordination plot showing the effect of sampling times on the composition of the nirK-type denitrifier community, based on nirK-T-RFLP data. The eigenvalues of the first and second axes in the ordination diagram are as follows: λ1=0.042, λ2=0.018. Arrows indicate development over time.

Download figure to PowerPoint

The composition of the nirK-type denitrifier community differed significantly among sampling times (trace=0.079, F=1.75, P=0.0001; Monte Carlo permutation test; Fig. 1b). The composition of the denitrifier communities in the soil samples from summer, autumn and winter 2003 was as different from each other as from that of the samples taken in autumn 2004, but the community composition of nirK-type denitrifiers in the soil sampled in winter 2003 and summer 2004 was similar. When sampling time was used as a linear contrast to analyse whether there was a continuous development of the community over time, it also affected the composition of the nirK-type denitrifier soil community (trace=0.039, F=3.41, P=0.0001; Monte Carlo permutation test within CCA), indicating a certain continuous trend in mesocosm development.

The effect of the interaction of plant treatment with sampling time on the nirK-type denitrifier communities was significant (trace=0.31, F=1.29, P=0.020; Monte Carlo permutation tests), indicating that the effect of plant treatment varied with the sampling time.

Soil nitrate concentrations and DEA

Soil nitrate concentrations were not influenced by plant diversity (F1, 3=3.73, P=0.15), but varied strongly among plant treatments. The effect of plant treatments was nearly exclusively due to the much higher nitrate concentrations in the upper soil layer of unplanted mesocosms compared with those in mesocosms with plants, because different plant combinations had no influence on soil nitrate (Table 1, Fig. 2). The effect of plant presence varied over time (significant time by plant treatment interaction in Table 1): while the presence of plants had strong effects on soil nitrate in winter 2003 and autumn 2004, the effect was much smaller in summer 2004 (Fig. 2).

Table 1.   Results of a nested general linear model of the effects of the experimental factors on soil nitrate concentration and DEA
Source of variationNitrateDEA
dfSSF-valueSSF-value
  • The effect of the plant combinations was tested against the variation among individual mesocosms. In addition, the effect of plant treatment was partitioned into two orthogonal contrasts ‘no plants vs. plants present’ and ‘plant combination’. Given are the degrees of freedom (df), the sums of squares (SS) and the F-values for the various factors.

  • *

    P<0.05.

  • **

    P<0.01.

  • ***

    P<0.001.

  • NS, not significant.

Plant treatment5381.158.3***6208.523.7***
 Plant presence1380.6291.0***2795.053.4***
 Plant combination40.50.10 NS3413.516.3***
Mesocosm2228.80.25 NS1151.00.9 NS
Sampling time213.81.34 NS438.03.9*
Plant treatment × time1071.91.39 NS4117.47.3***
 Plant presence × time271.86.95**358.63.2 NS
 Plant combination × time80.1<0.01 NS3758.88.3***
Residual34175.7 1919.3 
Total73671.3 13 834.1 
image

Figure 2.  Nitrate concentrations of soil samples from the upper soil layer (0–7 cm) of experimental mesocosms. Data are means (n=5; treatment 0, n=3), bars indicate ±1 SE. Different letters indicate significant differences (Tukey's HSD test, P<0.05). Plant treatments: 0, unplanted; 2a, planted with Holcus lanatus and Arrhenatherum elatius; 2b, planted with H. lanatus and Geranium pratense; 2c, planted with H. lanatus and Plantago lanceolata; 4, planted with H. lanatus, A. elatius, G. pratense, P. lanceolata; 8, like 4, but in addition planted with Alopecurus pratensis, Anthoxanthum odoratum, Ranunculus acris and Taraxacum officinale.

Download figure to PowerPoint

The DEA (Tiedje, 1994) was determined to assess the potential maximum activity of existing denitrifying enzymes in soil. DEA was not influenced by plant diversity (F1, 3=1.83, P=0.27), but depended on plant treatment (Table 1). In unplanted soil, the denitrifying enzyme activity was 80% lower than in planted soil (Table 1, Fig. 3), but there was also a significant variation among the various plant combinations. The DEA activity was generally higher in summer than in autumn, but differences depended on plant treatment (significant time × treatment interaction, Table 1). While enzyme activities in winter and autumn were similar within planted soils, in summer, there was a trend towards higher activity with increasing plant diversity.

image

Figure 3.  Effects of different plant treatments on DEA in mesocosm soil in winter 2003, summer 2004 and autumn 2004. Data are means±1 SE. Bars with different letters are significantly different (Tukey's HSD test, P<0.05). Plant treatments: 0, unplanted mesocosms; 2a, planted with Holcus lanatus and Arrhenatherum elatius; 2b, planted with H. lanatus and Geranium pratense; 2c, planted with H. lanatus and Plantago lanceolata; 4, planted with H. lanatus, A. elatius, G. pratense and P. lanceolata; 8, like 4, but in addition planted with Alopecurus pratensis, Anthoxanthum odoratum, Ranunculus acris and Taraxacum officinale.

Download figure to PowerPoint

Relationships between plants, nirK-type denitrifier community composition and DEA

To analyse whether the effects of the plants on the DEA were exclusively due to their effects on nirK-type denitrifier community composition, the effect of denitrifier community composition was eliminated in a separate analysis using the sample scores of the first two axes of a CA of the nirK-type denitrifier community data from the three sampling dates for which DEA data were also available as covariables. The effects of the sample scores on denitrifier functioning (DEA assay) were highly significant (Table 2), showing that nirK-type denitrifier community composition had a strong influence on denitrifier functioning. However, the effects of the presence of plants and the combination of plants on DEA were still highly significant. This indicates that the plants influenced enzyme activity both indirectly through their effect on the composition of the denitrifier community and directly, i.e. independent of their effect on the microbial community.

Table 2.   Effects of experimental factors (sample scores along axis 1, sample scores along axis 2, plant treatment, etc.) on DEA
Source of variationdfSSF-value
  • To control for possible effects of the composition of the microbial community, sample scores along axes 1 and 2 of a CA were used as covariables. The effect of plant treatment was partitioned into two orthogonal contrasts ‘no plants vs. plants present’ and ‘plant combination’. Given are the degrees of freedom (df), the sums of squares (SS) and the F-values for the various factors.

  • *

    P<0.05.

  • **

    P<0.01.

  • ***

    P<0.001.

  • NS, not significant.

CA scores axis 11738.613.7**
CA scores axis 21741.113.7**
Plant treatment54733.117.6***
 Plant presence11808.634.3***
 Plant combination42924.513.9***
Mesocosm221158.90.98 NS
Sampling time2477.24.4*
Plant treatment × time104529.37.9***
 Plant presence × time2364.53.4*
 Plant combination × time83894.59.0***
Residual321725.9 
Total7413 834.1 

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Plant diversity and combination effects on denitrifier communities

We found a strong effect of the different plant treatments on the composition of nirK-type denitrifiers in the soil and their functioning. The strongest differences in the composition of nirK-type denitrifiers were found between mesocosms with and without plants, but plant combinations also significantly influenced microbial community composition, whereas plant species diversity had no effect. Our finding that the presence of higher plants exerted the most pronounced effect on the composition of soil microbial groups was supported by results of a parallel study that compared the microbial communities (Alpha- and Betaproteobacteria, Bacteriodetes, Chloroflexi, Planctomycetes and Verrucomicrobia) of the mesocosms devoid of plants and of those with plants (Zul et al., 2007). The presence of plants strongly influenced the communities of the studied microorganisms, with one-third of the phylotypes distinguished occurring either only in mesocosms with plants or in those without. Similarly, in several studies, the microbial communities of the rhizosphere were found to be distinct from those of the bulk soil in the vicinity of plants (Kowalchuk et al., 2002; Kuske et al., 2002; Costa et al., 2006). Moreover, the culturable denitrifier community in the rhizosphere of maize plants also differed from that of bulk soil (Cheneby et al., 2004). These differences have been attributed to the effect of root exudates on microbial communities (Bais et al., 2006). This is supported by the fact that in laboratory experiments, the addition of artificial root exudates altered the composition of soil microbial communities (Kozdroj & van Elsas, 2000; Baudoin et al., 2003; Bürgmann et al., 2005).

In a previous study with the same plant species, we found that individual plant species selected for distinct nirK-type denitrifier compositions in soil (Bremer et al., 2007). These results suggested that individual plant species differed in the amount, spectrum and composition of root exudates (Griffiths et al., 1999; Jaeger et al., 1999) or decomposing root material and so resulted in specific denitrifier communities. Organic carbon would be predicted to be the most important factor selecting for distinct soil denitrifier communities, because the majority of denitrifiers are aerobic heterotrophs that may seldom use their denitrification capacity. Accordingly, as hypothesized, different plant combinations also strongly affected denitrifier community composition. Our findings are in line with previous reports that have shown effects of plant combination on the composition of soil microbial communities in general (Kowalchuk et al., 2002; Johnson et al., 2003; Zak et al., 2003; Carney & Matson, 2006).

A distinctive nirK-type denitrifier community was sustained by the combination of H. lanatus/A. elatius (2a in Fig. 1a). This might be due to the particularly high plant productivity (sum of above-ground and below-ground plant biomass in 2003), which was significantly higher for this combination than for any other plant combination (Reuter, 2006). Furthermore, in contrast to all other plant treatments, combination 2a consisted only of grasses, i.e. members of a single functional group. The functional diversity of plants in the mesocosms planted with species belonging to two functional groups, grasses and forbs, may have influenced nirK-type denitrifier community composition by increasing the number of ecological niches or functional redundancy (Hector et al., 1999; Diaz & Cabido, 2001).

Despite significant differences in denitrifier community composition associated with different combinations of plants, our study did not reveal an effect of plant richness on denitrifier community composition. The results of other studies investigating a potential coupling of plant diversity and soil microbial community composition have been conflicting, with some studies also reporting no effects (Broughton & Gross, 2000; Kowalchuk et al., 2000; Carney et al., 2004; Niklaus et al., 2007), while others reported effects of plant diversity on soil microbial communities (Stephan et al., 2000; Kowalchuk et al., 2002; Carney & Matson, 2006; Grüter et al., 2006; Loranger-Merciris et al., 2006; Chung et al., 2007). However, Carney & Matson (2006) suggested that changes in microbial community composition may not be related to plant richness per se, but to the likelihood of a plant community containing a plant that exerts a strong effect on the microbial community composition. For instance, functional traits of individual plant species such as the propensity to form mycorrhiza may exert crucial effects on the microbial community structure (Johnson et al., 2003).

Sampling time effects

In our study, sampling time influenced the nirK-type denitrifier community composition. A major impact, causing differences in community composition between the two years of sampling, could have been exerted by the exceptional heat and drought in 2003, which resulted in an estimated Europe-wide reduction in primary productivity of 30% (Ciais et al., 2005). However, we also observed a seasonal variation in denitrifier community composition. Seasonal effects on soil microbial communities (Grayston et al., 2001; Carney & Matson, 2006) as well as on denitrifiers (Wolsing & Priemé, 2004; Boyle et al., 2006) have been reported previously. The composition of the denitrifier community over time also varied depending on plant combination, indicating that mesocosms with specific plant combinations developed differently over time. This could have been due to differences in the composition of root exudates, depending on the species and the developmental stage of plants (Jaeger et al., 1999), or due to shifts that occurred in the relative contribution of individual plant species to above-ground biomass. For instance, the biomass of H. lanatus decreased strongly during the cold winter 2003, and this species, adapted to oceanic climates, suffered most from the extreme drought in summer 2003. As a consequence, in summer 2003, H. lanatus comprised <50% of the shoot biomass in all two species variants, while P. lanceolata dominated plant communities in the mesocosms planted with H. lanatus/P. lanceolata (2c), as well as the four and eight species variants (Reuter, 2006).

However, the nirK-type denitrifier community composition did not only vary seasonally but also changed directionally with time, suggesting a succession, as indicated by the significant linear trend component of time. This could be due to the development of the plant mesocosms with age either due to ageing of the plants themselves or due to progressing root development in the mesocosms.

Plant and time effects on DEA

The DEA assay (Tiedje, 1994) is commonly used to compare the potential activity of denitrifiers of different sites and experimental treatments in terrestrial environments (Groffman et al., 2006). Our DEA values of <10 to c. 60 ng N2O-N gdw−1 soil h−1 are at the lower end of rates reported for soils in other studies (1–3000 ng N g−1 soil h−1) (Rich et al., 2003). There is no consensus in the literature on the relationship between soil nitrate content and DEA. Most studies from forested ecosystems suggested that DEA is limited by NO3 availability in the soil (Griffiths et al., 1998; Rich et al., 2003) and large positive responses of DEA to NO3 added to soil were observed (Barnard et al., 2006), but other studies found no relationship (Boyle et al., 2006). In our study, a high variability of DEA was observed in the planted mesocosms with low nitrate concentrations. However, the DEA activity was much lower at very high nitrate concentrations that were observed in some of the samples without plants. This suggests that other soil parameters impacted by the plants had a major influence on DEA. Plant presence, plant combination and time affected the DEA, emphasizing the importance of the amount, composition or spectrum of organic carbon supplied by the plants. DEA was significantly increased by the presence of plants, especially when they were actively growing during summer and autumn. In previous studies, increased DEA was found in the rhizosphere (soil adhering to the roots) of potted tree seedlings (Priha et al., 1999), maize (Mahmood et al., 1997) and barley (Højberg et al., 1996). Similarly, the addition of maize mucilage (Mounier et al., 2004) and artificial maize root exudates (Henry et al., 2008) to soil enhanced denitrification activity. However, the results of studies on the impact of different plants species on denitrification activity in the soil have been conflicting. While in one previous study, similar to our results, differences in DEA dependent on the plant species were found (Patra et al., 2006), another study found no effect of plant species (Priha et al., 1999). Effects of time on DEA have been reported previously (Boyle et al., 2006; Wallenstein et al., 2006a), but in the present study, the effect of time also varied with plant combination. This indicates that the development of the microbial communities with plant age depended on plant composition.

Linkage of plant communities, denitrifier communities and functioning

A major, but yet largely unresolved question in microbial ecology is whether microbial community structure and function are interlinked. Studies on denitrifier communities and their functioning suggest that structure–function inter-relationships of these communities may be ecosystem specific (Rich & Myrold, 2004), and that the activity of denitrification enzymes may sometimes be coupled to community composition, while in other cases, it may be determined by environmental factors (Wallenstein et al., 2006b). Some studies reported that potential denitrification rates (Cavigelli & Robertson, 2000; Rich et al., 2003), the oxygen and pH sensitivity of the denitrification enzymes (Cavigelli & Robertson, 2000) and the ratio of gaseous denitrification products (Cheneby et al., 1998) were influenced by the composition of the underlying denitrifier communities, but others found no clear inter-relationship of community composition and enzyme activity (Rich & Myrold, 2004; Boyle et al., 2006) or N2O emission (Ma et al., 2008). Our study showed a strong effect of nirK-type denitrifier community composition on overall denitrifier functioning (Table 2). Because the composition of only part of the denitrifier community was studied, an even more comprehensive picture of the interactions between denitrifier community composition and functioning might have been obtained had nirS-type denitrifier communities also been included. However, the effects of the plants and of sampling time on DEA remained even after statistical elimination of the influence of the composition of the nirK-type denitrifier community. Thus, plants and sampling time had direct effects on DEA, which could not be explained by their effects on the nirK-type denitrifier community composition. This is in line with the results of a previous study in which different artificial maize root exudates had distinctive effects on soil denitrifier activity, although the composition and density of denitrifier communities was not influenced (Henry et al., 2008). However, in contrast to our study, that study was short term (4 weeks), and there might have been effects on the composition of the denitrifier community in the long term.

In conclusion, our results indicate that different plant communities sustain distinctive nirK-type denitrifier communities in the vicinity of plant roots, and that plant composition directly modulates the functioning of below-ground denitrifier communities. Our results further suggest that the largest effects were due to the presence of plants in general and the composition of the plant community, whereas plant species richness hardly had any influence. In our study, we designed plant communities that were similar to nearby semi-natural grasslands. The range of plant diversity studied was rather small (two to eight species) and limited to only two functional groups of plants (grasses and forbs), and it is not yet known how well our mesocosm experiments mimic in situ conditions. However, ecological and biodiversity experiments with model ecosystems have become established tools in ecosystem research during the last decades (Hector et al., 1999; Beierkuhnlein & Jentsch, 2005; Beierkuhnlein & Neßhoever, 2006), because they allow a better control and replication of site conditions. Interestingly, apart from the plant-induced effects, the composition of nirK-type denitrifier communities also exerted a strong effect on denitrifier functioning. These results contribute towards an enhanced understanding of the ecosystem processes involved in nitrogen and greenhouse gas cycling.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

We are grateful to Dr Guido Kossmann for his continuous help with soil sampling. This work was supported by a grant from the Max Planck Society (Munich, Germany) and the German Federal Ministry for Education and Research within the BIOLOG Biodiversity Program (01LC0021).

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • Bais HP, Weir TL, Perry LG, Gilroy S & Vivanco JM (2006) The role of root exudates in rhizosphere interactions with plants and other organisms. Annu Rev Plant Biol 57: 233266.
  • Barnard R, Le Roux X, Hungate BA, Cleland EE, Blankinship JC, Barthes L & Leadley PW (2006) Several components of global change alter nitrifying and denitrifying activities in an annual grassland. Funct Ecol 20: 557564.
  • Baudoin E, Benizri E & Guckert A (2003) Impact of artificial root exudates on the bacterial community structure in bulk soil and maize rhizosphere. Soil Biol Biochem 35: 11831192.
  • Beierkuhnlein C & Jentsch A (2005) Ecological importance of species diversity. A review on the ecological implications of species diversity in plant communities. Plant Diversity and Evolution: Genotypic and Phenotypic Variation in Higher Plants (HenryRJ, ed), pp. 249285. CAB International, Wallingford, UK.
  • Beierkuhnlein C & Neßhoever C (2006) Biodiversity experiments – artificial constructions or heuristic tools? Prog Bot 67: 486535.
  • Boyle SA, Rich JJ, Bottomley PJ, Cromack K & Myrold DD (2006) Reciprocal transfer effects on denitrifying community composition and activity at forest and meadow sites in the Cascade Mountains of Oregon. Soil Biol Biochem 38: 870878.
  • Bremer C, Braker G, Matthies D, Reuter A, Engels C & Conrad R (2007) Impact of plant functional group, plant species, and sampling time on the composition of nirK-type denitrifier communities in soil. Appl Environ Microb 73: 68766884.
  • Brodie E, Edwards S & Clipson N (2002) Bacterial community dynamics across a floristic gradient in a temperate upland grassland ecosystem. Microb Ecol 44: 260270.
  • Broughton LC & Gross KL (2000) Patterns of diversity in plant and soil microbial communities along a productivity gradient in a Michigan old-field. Oecologia 125: 420427.
  • Bürgmann H, Meier S, Bunge M, Widmer F & Zeyer J (2005) Effects of model root exudates on structure and activity of a soil diazotroph community. Environ Microbiol 7: 17111724.
  • Carney KM & Matson PA (2006) The influence of tropical plant diversity and composition on soil microbial communities. Microb Ecol 52: 226238.
  • Carney KM, Matson PA & Bohannan BJM (2004) Diversity and composition of tropical soil nitrifiers across a plant diversity gradient and among land-use types. Ecol Lett 7: 684694.
  • Cavigelli MA & Robertson GP (2000) The functional significance of denitrifier community composition in a terrestrial ecosystem. Ecology 81: 14021414.
  • Cheneby D, Hartmann A, Henault C, Topp E & Germon JC (1998) Diversity of denitrifying microflora and ability to reduce N2O in two soils. Biol Fert Soil 28: 1926.
  • Cheneby D, Perrez S, Devroe C, Hallet S, Couton Y, Bizouard F, Luretig G, Germon JC & Philippot L (2004) Denitrifying bacteria in bulk and maize-rhizospheric soil: diversity and N2O-reducing abilities. Can J Microbiol 50: 469474.
  • Chung HG, Zak DR, Reich PB & Ellsworth DS (2007) Plant species richness, elevated CO2, and atmospheric nitrogen deposition alter soil microbial community composition and function. Glob Change Biol 13: 980989.
  • Ciais P, Reichstein M, Viovy N et al. (2005) Europe-wide reduction in primary productivity caused by the heat and drought in 2003. Nature 437: 529533.
  • Conrad R (1996) Soil microorganisms as controllers of atmospheric trace gases (H2, CO, CH4, OCS, N2O, and NO). Microbiol Rev 60: 609640.
  • Costa R, Gotz M, Mrotzek N, Lottmann J, Berg G & Smalla K (2006) Effects of site and plant species on rhizosphere community structure as revealed by molecular analysis of microbial guilds. FEMS Microbiol Ecol 56: 236249.
  • Crutzen PJ (1970) Influence of nitrogen oxides on atmospheric ozone content. Q J Roy Meteor Soc 96: 320325.
  • Diaz S & Cabido M (2001) Vive la difference: plant functional diversity matters to ecosystem processes. Trends Ecol Evol 16: 646655.
  • Dickinson RE & Cicerone RJ (1986) Future global warming from atmospheric trace gases. Nature 319: 109115.
  • Dunbar J, Ticknor LO & Kuske CR (2001) Phylogenetic specificity and reproducibility and new method for analysis of terminal restriction fragment profiles of 16S rRNA genes from bacterial communities. Appl Environ Microb 67: 190197.
  • Grayston SJ, Wang SQ, Campbell CD & Edwards AC (1998) Selective influence of plant species on microbial diversity in the rhizosphere. Soil Biol Biochem 30: 369378.
  • Grayston SJ, Griffith GS, Mawdsley JL, Campbell CD & Bardgett RD (2001) Accounting for variability in soil microbial communities of temperate upland grassland ecosystems. Soil Biol Biochem 33: 533551.
  • Griffiths BS, Ritz K, Ebblewhite N & Dobson G (1999) Soil microbial community structure: effects of substrate loading rates. Soil Biol Biochem 31: 145153.
  • Griffiths RP, Homann PS & Riley R (1998) Denitrification enzyme activity of Douglas-fir and red alder forest soils of the Pacific Northwest. Soil Biol Biochem 30: 11471157.
  • Groffman PM, Altabet MA, Bohlke JK, Butterbach-Bahl K, David MB, Firestone MK, Giblin AE, Kana TM, Nielsen LP & Voytek MA (2006) Methods for measuring denitrification: diverse approaches to a difficult problem. Ecol Appl 16: 20912122.
  • Grüter D, Schmid B & Brandl H (2006) Influence of plant diversity and elevated atmospheric carbon dioxide levels on below ground bacterial diversity. BMC Microbiology 6: 68. DOI:DOI: DOI: 10.1186/1471-2180-6-68.
  • Hector A, Schmid B, Beierkuhnlein C et al. (1999) Plant diversity and productivity experiments in European grasslands. Science 286: 11231127.
  • Hector A, Loreau M, Schmid B et al. (2005) Biodiversity manipulation experiments: studies replicated at multiple sites. Biodiversity and Ecosystem Functioning. Synthesis and Perspectives (LoreauM, NaeemS & InchaustiP, eds), pp. 3646. Oxford University Press, Oxford, UK.
  • Henry S, Texier S, Hallet S, Bru D, Dambreville C, Cheneby D, Bizouard F, Germon JC & Philippot L (2008) Disentangling the rhizosphere effect on nitrate reducers and denitrifiers: insight into the role of root exudates. Environ Microbiol 10: 30823092.
  • Højberg O, Binnerup SJ & Sørensen J (1996) Potential rates of ammonium oxidation, nitrite oxidation, nitrate reduction and denitrification in the young barley rhizosphere. Soil Biol Biochem 28: 4754.
  • Jaeger CH, Lindow SE, Miller S, Clark E & Firestone MK (1999) Mapping of sugar and amino acid availability in soil around roots with bacterial sensors of sucrose and tryptophane. Appl Environ Microb 65: 26852690.
  • Johnson D, Booth RE, Whiteley AS, Bailey MJ, Read DJ, Grime JP & Leake JR (2003) Plant community composition affects the biomass, activity and diversity of microorganisms in limestone grassland soil. Eur J Soil Sci 54: 671677.
  • Kowalchuk GA, Stienstra AW, Heilig GH, Stephen JR & Woldendorp JW (2000) Molecular analysis of ammonia-oxidizing bacteria in soil of successional grasslands of the Drentsche A (The Netherlands). FEMS Microbiol Ecol 31: 207215.
  • Kowalchuk GA, Buma DS, De Boer W, Klinkhamer PGL & Van Veen JA (2002) Effects of above-ground plant species composition and diversity on the diversity of soil-borne microorganisms. A Van Leeuw J Microb 81: 509520.
  • Kozdroj J & Van Elsas JD (2000) Response of the bacterial community to root exudates in soil polluted with heavy metals assessed by molecular and cultural approaches. Soil Biol Biochem 32: 14051417.
  • Kuske CR, Ticknor LO, Miller ME, Dunbar JM, Davis JA, Barns SM & Belnap J (2002) Comparison of soil bacterial communities in rhizospheres of three plant species and the interspaces in an arid grassland. Appl Environ Microb 68: 18541863.
  • Loranger-Merciris G, Barthes L, Gastine A & Leadley P (2006) Rapid effects of plant species diversity and identity on soil microbial communities in experimental grassland ecosystems. Soil Biol Biochem 38: 23362343.
  • Ma WK, Bedard-Haughn A, Siciliano SD & Farrell RE (2008) Relationship between nitrifier and denitrifier community composition and abundance in predicting nitrous oxide emissions from ephemeral wetland soils. Soil Biol Biochem 40: 11141123.
  • Mahmood T, Ali R, Malik KA & Shamsi SRA (1997) Denitrification with and without maize plants (Zea mays L.) under irrigated field conditions. Biol Fert Soils 24: 323328.
  • Mounier E, Hallet S, Cheneby D et al. (2004) Influence of maize mucilage on the diversity and activity of the denitrifying community. Environ Microbiol 6: 301312.
  • Murray RE & Knowles R (1999) Chloramphenicol inhibition of denitrifying enzyme activity in two agricultural soils. Appl Environ Microb 65: 34873492.
  • Murray RE & Knowles R (2004) Trace amounts of O2 affect NO and N2O production during denitrifying enzyme activity (DEA) assays. Soil Biol Biochem 36: 513517.
  • Niklaus PA, Wardle DA & Tate KR (2006) Effects of plant species diversity and composition on nitrogen cycling and the trace gas balance of soils. Plant Soil 282: 8398.
  • Niklaus PA, Alphei J, Kampichler C, Kandeler E, Korner C, Tscherko D & Wohlfender M (2007) Interactive effects of plant species diversity and elevated CO2 on soil biota and nutrient cycling. Ecology 88: 31533163.
  • Patra AK, Abbadie L, Clays-Josserand A et al. (2006) Effects of management regime and plant species on the enzyme activity and genetic structure of N-fixing, denitrifying and nitrifying bacterial communities in grassland soils. Environ Microbiol 8: 10051016.
  • Priemé A, Braker G & Tiedje JM (2002) Diversity of nitrite reductase (nirK and nirS) gene fragments in forested upland and wetland soils. Appl Environ Microb 68: 18931900.
  • Priha O, Hallantie T & Smolander A (1999) Comparing microbial biomass, denitrification enzyme activity, and numbers of nitrifiers in the rhizospheres of Pinus sylvestris, Picea abies and Betula pendula seedlings by microscale methods. Biol Fert Soil 30: 1419.
  • Reuter A (2006) Biomassebildung und Nährstoffvermögen der Wurzeln in experimentellen Grünlandbeständen mit unterschiedlicher Pflanzenartzusammensetzung. PhD Thesis, University of Bayreuth, Bayreuth, Germany.
  • Rich JJ & Myrold DD (2004) Community composition and activities of denitrifying bacteria from adjacent agricultural soil, riparian soil, and creek sediment in Oregon, USA. Soil Biol Biochem 36: 14311441.
  • Rich JJ, Heichen RS, Bottomley PJ, Cromack K & Myrold DD (2003) Community composition and functioning of denitrifying bacteria from adjacent meadow and forest soils. Appl Environ Microb 69: 59745982.
  • Sharma S, Aneja MK, Mayer J, Munch JC & Schloter M (2005) Diversity of transcripts of nitrite reductase genes (nirK and nirS) in rhizospheres of grain legumes. Appl Environ Microb 71: 20012007.
  • Stephan A, Meyer AH & Schmid B (2000) Plant diversity affects culturable soil bacteria in experimental grassland communities. J Ecol 88: 988998.
  • Ter Braak CJF (1986) Canonical correspondence analysis: a new eigenvector technique for multivariate direct gradient analysis. Ecology 67: 11671179.
  • Ter Braak CJF & Šmilauer P (2002) Canoco Reference Manual and CanoDraw for Windows Users Guide: Software for Canonical Community Ordination, Version 4.5. Microcomputer Power, Ithaca, NY.
  • Tiedje JM (1994) Denitrifiers. Methods of Soil Analysis (WeaverRW & MickelsonSH, eds), pp. 245267. Soil Science Society of America, Madison, WI.
  • Wallenstein MD, Peterjohn WT & Schlesinger WH (2006a) N fertilization effects on denitrification and N cycling in an aggrading forest. Ecol Appl 16: 21682176.
  • Wallenstein MD, Myrold DD, Firestone M & Voytek M (2006b) Environmental controls on denitrifying communities and denitrification rates: insights from molecular methods. Ecol Appl 16: 21432152.
  • Wardle DA, Yeates GW, Williamson W & Bonner KI (2003) The response of a three trophic level soil food web to the identity and diversity of plant species and functional groups. Oikos 102: 4556.
  • Wolsing M & Priemé A (2004) Observation of high seasonal variation in community structure of denitrifying bacteria in arable soil receiving artificial fertilizer and cattle manure by determining T-RFLP of nir gene fragments. FEMS Microbiol Ecol 48: 261271.
  • Zak DR, Holmes WE, White DC, Peacock AD & Tilman D (2003) Plant diversity, soil microbial communities, and ecosystem function: are there any links? Ecology 84: 20422050.
  • Zar JH (1999) Biostatistical Analysis. Prentice Hall, New York, NY.
  • Zul D, Denzel S, Kotz A & Overmann J (2007) Effects of plant biomass, plant diversity, and water content on bacterial communities in soil lysimeters: implications for the determinants of bacterial diversity. Appl Environ Microb 73: 69166929.

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Fig. S1. CA ordination plot for the mesocosm soil samples based on nirK T-RFLP-data obtained from soil collected in (A) summer 2003, (B) autumn 2003, (C) winter 2003, (D) summer 2004, and (E) autumn 2004. The samples, i.e. the three or five replicate mesocosms for each plant combination, are denoted by filled circles and subscripts denote the individual replicates. The eigenvalues of the 1st and 2nd axes in figure 1A are λ1 = 0.228, λ 2 = 0.218, in figure 1B λ 1 = 0.149, λ 2 = 0.119, in figure 1C λ 1 = 0.296, λ 2 = 0.131, in figure 1D λ 1 = 0.106, λ 2 = 0.093, and in figure 1E λ 1 = 0.146, λ 2 = 0.106. 01–3, unplanted mesocosms (^); 2a1–5 (⋆), planted with Holcus lanatus and Arrhenatherum elatius; 2b1–5 (▪), planted with Holcus lanatus and Geranium pratense; 2c1–5 (Δ), planted with Holcus lanatus and Plantago lanceolata; 41–5 (□), planted with Holcus lanatus, Arrhenaterum elatius, Geranium pratense, Plantago lanceolata; 81–5 (), like 4, but in addition planted with Alopecurus pratensis, Anthoxanthum odoratum, Ranunculus acris and Taraxacum officinale.

Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

FilenameFormatSizeDescription
FEM_732_sm_figS1.pdf305KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.