SEARCH

SEARCH BY CITATION

Keywords:

  • nutrients;
  • wetlands;
  • functional ecology;
  • plant traits;
  • growth strategies

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  • • 
    An extensive survey of European wetlands was undertaken to compare the importance of growing conditions vs functional characteristics of vegetation in determining N, P and K contents.
  • • 
    Stress-tolerator dominated stands (S) had consistently lower nutrient contents and higher N : P ratios whereas ruderal-dominated (R) stands displayed the opposite pattern. Competitor (C) and competitor-stress tolerator (CS) stands were intermediate to R and S.
  • • 
    These patterns were mostly preserved after removing covariation between vegetation and environment, thus indicating constitutional differences in nutrient signatures between functionally differentiated vegetation. C and R stands were least likely to be nutrient limited. Half of the S stands were probably P-limited but C, CS and R stands rarely or never experienced P limitation. Inferred colimitation by K was twice as frequent in S stands compared with other vegetation.
  • • 
    This study extends the evidence for syndromes of traits closely linked to nutrient use efficiency that increase fitness under particular growing conditions. It also highlights patterns at a community level across a wide range of wetland types and suggests that tissue nutrient signatures will have diagnostic value in predicting community responses to perturbation in nutrient availability.

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Nutrient availability is a key influence on plant growth; species have different nutrient requirements, exploit nutrients with varying efficiency and store or convert nutrients to biomass at different rates (Chapin, 1980; Marschner, 1995; Aerts & Chapin, 2000). Extrapolated to the level of the vegetation, this makes nutrient supply a fundamental factor in the structuring of plant communities (Tilman, 1982). Having recently documented consistent relationships between mineral nutrition and a wide range of other plant traits from which existing plant growth strategies have been defined, Grime et al. (1997) have described mineral nutrients as ‘the fundamental currency of vegetation processes at scales from the individual to ecosystems and landscapes’. The ease with which human activity can modify nutrient availability, at scales ranging from local to international, only amplifies this importance.

Nutrient supply is widely recognized as a major influence on the composition, structure and productivity of wetland vegetation (Auclair, 1976; Verhoeven et al., 1988; Bedford et al., 1999; Wheeler, 1999; Keddy, 2000). It varies spatially and temporally according to the ion chemistry and relative contributions of precipitation, hillslope, groundwater and riverine inputs, and the interaction of these external sources with local biogeochemical processes and anthropogenic factors (Koerselman et al., 1990; Wassen et al., 1990; Grieve et al., 1995). However, quantifying these components and interpreting the effects of variation in nutrient supply on wetland plant species has often proved difficult for various reasons. These include fine scale spatial and temporal environmental variation, the wide ecological amplitude of some species, intraspecific variation in ecological preferences across biogeographical zones, covariation between nutrients and other environmental factors, and the complex regulation of nutrient availability by hydrochemistry (Malmer, 1962; Hayati & Proctor, 1990; Wheeler et al., 1992; Bootsma & Wassen, 1996; Wheeler, 1999).

Searching for community-level patterns using simple trait-based classifications of wetland vegetation might therefore offer a valuable alternative perspective and in practical terms would be a useful tool in predicting generalized vegetation responses to anthropogenic factors that have a direct or indirect effect on nutrient supply. This approach would have added value if functionally different vegetation types defined mainly on the basis of suites of morphological and regenerative traits were shown to possess characteristic tissue nutrient signatures consistent with their preferred growing conditions and the type of nutrient limitation faced. Tissue nutrient contents in wetland plants collected from the field vary widely between species and habitats (Boyd, 1978; Waughman, 1980; Hayati & Proctor, 1990). At a functional group level several recent studies and reviews have emphasized the importance of life form (i.e. forb, graminoid, woody deciduous or woody evergreen) over habitat as a determinant of plant nutrient content, nutrient use efficiency or response to fertilization (Bowman, 1994; Eckstein & Karlsson, 1997; Aerts et al., 1999; Eckstein et al., 1999). However, within herbaceous vegetation, which dominates many temperate wetlands, recurrent patterns in nutrient contents have proved more elusive or still await investigation.

In this study we use an extensive survey of tissue nutrient contents in herbaceous wetland vegetation to address four issues. First, does wetland vegetation described in functional terms differentiate with respect to soil nutrient concentrations or other environmental variables (that may directly or indirectly control nutrient availability)? Second, we assess how tissue nutrient contents of multispecies vegetation are related to the representation of major plant growth strategy elements or to Functional Vegetation Types (FVTs) and if these relationships have a primarily environmental or evolutionary basis. Third, we make inferences about the nature and extent of nutrient limitation affecting wetland vegetation in general, and functionally distinct assemblages in particular, in an attempt to highlight the nutritional challenges facing different types of vegetation under their typical growing conditions. Recent studies in freshwater wetlands have suggested that nutrient ratios within samples of plant tissue are of diagnostic value and can offer a pragmatic alternative to identifying limiting nutrients and detecting their effects at the community level (Wassen et al., 1995; Koerselman & Meuleman, 1996; Boeye et al., 1997). Finally, we examine if compatible relationships are evident within collective vegetation state variables (e.g. biomass or height) which, across a diverse range of sites, are surrogates for overall change in the composition of species or plant growth strategies.

Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Field sampling

Paired samples of vegetation and soil were collected during the months of July and August from seven major wetland complexes distributed in the area between northern Scotland, central Ireland and central France. Sites were chosen to cover a wide gradient of fertility, waterlogging stress and flooding intensity (Table 1) and are representative of the conditions encountered in minerotrophic wetlands across much of north-west Europe.

Table 1.  Location and distribution of wetlands sampled. Bracketed values indicate number of samples collected from each site that were used in this study. For simplicity main vegetation types are described with reference to the CORINE biotopes system (Devilliers et al., 1991). The ordering of sites reflects a broad gradient from base-rich lowland sites descending to base-poor upland ones
Site code: locationHabitatPrincipal biotopesRepresentative species
VF: Little Ouse/Waveney fens, Norfolk-Suffolk, E. England (52°23′ N, 1°0′ E) (Redgrave & Lopham, Market Weston, Hopton, Roydon, Middle Harling, Thelnetham and Banham fens). (88)complex of 7 calcareous, spring-fed, base-rich fens, occasionally waterlogged to surface but widely affected by drainage and abstraction, locally managed by cutting/scrub clearance.rich fens; reed beds; large sedge communities; stream ash-alder woods.Phragmites, Cladium mariscus, Calamagrostis canescens, Filipendula, Carex acutiformis, elata, Juncus subnodulosus, Molinia, Schoenus nigricans.
LM: Leighton Moss, Lancashire, NW England (54°10′ N, 2°47′ W). (15)flooded valley fen with extensive swamp communities around base-rich, eutrophic standing water.reed beds; large sedge communities; tall rush swamp.Carex paniculata, Juncus effusus, Phragmites, Phalaris.
LA: Loire/Allier floodplain: Decize (46°55′ N, 3°28′ E), Apremont (46°55′ N, 3°5′ E), central France. (46)large, active, extensively grazed, floodplain, including permanent and temporary fluvial land forms, and prone to major spring floods, dominated by free-draining nutrient-poor alluvial sands.Amphibious communities; river mud banks; mediterranean xeric grasslands; riparian willow formations; reed beds; large sedge communities.Carex arenaria, elata, otrubae, vesicaria, Cyperus fuscus,Persicaria hydropiper, Phalaris, Urtica dioica.
SC: mid-Shannon callows: Clonmacnoise (53°20′ N, 7°58′ W) and Little Brosna (53°8′ N, 7°55′ W), Co. Offaly, central Ireland. (54)regularly inundated, traditionally managed, expanses of unimproved, lowland wet grassland and fen formed on alluvial silts overlying peat and lake marl and enclosed by esker ridges.reed beds; large sedge communities; acidic fens; meadow sweet stands; eutrophic humid grasslands.Agrostis stolonifera, Carex disticha, elata, panicea, Festuca rubra, Filipendula ulmaria, Phalaris, Ranunculus repens.
EM: Lower Endrick at Endrick Marshes, Loch Lomond, Central Scotland. (57°6′ N, 3°58′ W). (28)annually mown lowland alluvial grasslands, plus temporary floodplain ponds, in matrix of unmanaged extensively inundated nutrient-rich swamp with peripheral incipient raised mire.amphibious communities; reed beds; large sedge beds; tall rush swamp; eutrophic humid grasslands; acidic fen.Agrostis stolonifera, Carex aquatilis, vesicaria, Filipendula, Glyceria fluitans, Juncus effusus, Phalaris, Lysimachia thrysiflora.
TM: Upper Torridge at Kismeldon Meadows (50°56′ N, 4°21′ W) and Bradford Mill meadows (50°51′ N, 4°14′ W), North Devon, SW England. (31)narrow free-draining floodplain corridor with low permeability head deposits on adjacent footslopes supporting abandoned grassland and part- drained fen meadow with groundwater seepage zones and depressions.mesophile pastures, large sedge communities; tall rush swamp; oligotrophic and eutrophic humid grasslands; acidic fen; meadow sweet stands and related tall herb communities.Molinia, Filipendula, Dactylis glomerata, Juncus acutiflorus, Phalaris, Cirsium dissectum, Carex paniculata, panicea.
IM: Upper Spey at Insh Marshes, Badenoch, Speyside, N. Scotland. (57°6′ N, 3°58′ W). (50)extensive, locally grazed, seasonally inundated upland river floodplain with fluvial waterbodies and ditch network on wet acid peats.wet heath; reed beds; large sedge communities; tall rush swamp; acidic fen; transition mires; oligotrophic humid grasslands.Carex aquatilis, nigra, rostrata, lasiocarpa, vesicaria, Deschampsia cespitosa, Eriophorum angustifolium, Molinia, Potentilla palustris.

At each site, vegetation and soils were sampled within 1 × 1 m quadrats located within larger areas of visually homogenous vegetation situated at points along a transect spanning the natural hydrological gradient. Measurements were obtained from a total of 310 sampling units. Within each quadrat, species composition was recorded based on cover. Mean values were obtained for the following state variables: stem density (number of ramets per 0.1 × 0.1 m quadrat, n = 3–5), canopy height (n = 5), stem diameter (measured with calipers at 50 mm above substrate surface, n = 3), nearest neighbour distance (distance between a randomly chosen ramet and the next nearest plant, n = 3) and density of reproductive structures (number of ramets per m2 bearing reproductive structures). Standing crop in height bands of 0–0.1, 0.1–0.2 and > 0.2 m was also determined based on the dried weight of material harvested by clipping from a 0.2 × 0.2 m quadrat placed centrally in each 1 × 1 m plot. In structurally heterogenous vegetation triplicate measurements of biomass were made within each quadrat. The biomass component was retained for laboratory determination of plant nutrient contents. The extent of litter cover was estimated visually.

Data on basic environmental variables relevant to nutrient supply were collected simultaneously in each unit. Soil samples were collected to a depth of 0.2 m and retained for laboratory analysis. pH and Eh7 (i.e. redox potential standardized to pH 7) were measured in situ at 0.05 m depth using precalibrated electrodes. Depth to water table (GWD) in 0.05 m classes was estimated by digging soil pits (to a maximum of 0.7 m) and allowing these to refill. At a third of the sample points these estimates were validated against records from permanent dipwells. The extent of bare ground (bare) was estimated visually and used as a simple integrative index of disturbance, reflecting the effects of flood scouring or sedimentation and livestock poaching. A simple 3-point index was used to describe management inputs (mgmt), ranging from 0 (unmanaged) to 1 (lightly grazed) and 2 (annual summer mowing or moderately grazed). Flood regime (flood) was used to summarize temporal variation in surface inundation caused specifically by river bank overtopping and was based on a combination of monthly recording at specific sites, the annual river flow record and an interpretation of surface topography, aerial photographs and the distribution of trash lines. Values ranged from 0 (never inundated) to 5 (more or less permanently inundated).

Analytical techniques

Air dried soil was analysed for total nitrogen (soilN) and plant available (extractable) phosphorus (soilP) and potassium (soilK). Total N was considered to reflect, better than extractable inorganic nitrogen, the size of the pool of nitrogen available for plant uptake. Total N was measured using Kjeldahl digestion (Bremner & Mulvaney, 1982) followed by determination of the ammonium-N by automated colourimetry. Phosphorus and potassium extractable in 0.5 M ammonium acetate/acetic acid buffer (pH 4.8) were determined by ascorbic acid/molybdenum blue automated colourimetry and flame photometry, respectively. Use of this extractant permitted the determination of both P and K in a single extract and was appropriate to a set of soil samples with a wide range of pH values and organic matter contents. Concentrations of Fe and Mn in air dried soils (soilFe and soilMn) were estimated by extraction of 1.0 g ground subsamples in 50 ml of 0.1 M sodium pyrophosphate for 2 h (Bascomb, 1968). Concentrations of Fe and Mn in the filtered extracts were measured by atomic absorption spectrometry. Analysis of plant tissue was undertaken at the vegetation level using aggregated above-ground material dried for 48 h at 50°C. Data is therefore in the form of a biomass weighted average for the stand. Total nitrogen was measured by Kjeldahl digestion (Bremner & Mulvaney, 1982) followed by determination of the ammonium-N by automated colourimetry. Total phosphorus and potassium were determined following a mixed acid (perchloric acid plus nitric acid) digestion using ascorbic acid/molybdenum blue automated colourimetry and flame photometry, respectively. All determinations were carried out in duplicate and compared with reference material for quality control. Organic matter content (% org) of soils was estimated from loss on ignition of oven dried material after 4 h at 550°C.

Treatment of data

All species were allocated an established phase growth strategy according to Grime et al. (1988). In the case of species that display dual strategies the strategy most relevant to the individual population was chosen wherever possible, based on observation of the critical traits. Unscreened species (mainly annuals from the French sites) were classified, based on field observations and published data, using the key provided in Grime et al. (1988). The contribution of primary strategy elements (C, competitor; S, stress tolerator; R, ruderal) to the vegetation was calculated simply as their contribution to a specific strategy (e.g. for the competitor strategy ‘C’: pure C, 1; CS or CR, 0.5 and CSR, 0.33), with weighting by cover applied to mixed stands (e.g. for a vegetation composed of plants displaying C, CS and CSR strategies in the ratio 10 : 75 : 15 the representation of the C strategy is 10 + (0.5 × 75) + (0.33 × 15) = 52%). To identify functional vegetation types (FVTs) agglomerative hierarchical cluster analysis was applied to the matrix of sites × vegetation, in which vegetation was reduced to the contribution (by percentage cover) of the major strategies (C, S, R, CS, CR, SR, CSR). The analysis was performed with GENSTAT 5 using Euclidean distance as the measure of similarity with clusters defined by average linkage. Four basic FVTs were defined using this procedure. Group ‘C’ (n = 58) was typified by monodominant stands of pure competitor strategists, principally Phragmites australis, Phalaris arundinacea, Glyceria maxima and Urtica dioica. The major contributors to group ‘S’ (n = 70) were stress tolerators (S, S/SC and SC strategies) such as Carex panicea, Molinia caerulea, Carex rostrata, Carex nigra and Potentilla palustris. Group ‘R’ (n = 59) was dominated by annuals and competitive ruderals (R and CR strategies), principally Agrostis stolonifera, Glyceria fluitans, Ranunculus repens, Lolium perenne, Persicaria hydropiper, Chenopodium polyspermum and Cyperus fuscus. The final group, ‘CS’ (n = 123), incorporated stands of competitive-stress tolerators (SC and C/SC strategies) such as Filipendula ulmaria, Deschampsia cespitosa, Juncus subnodulosus and J. effusus, plus the larger sedges (Carex acutiformis, C. aquatilis, C. elata, C. paniculata and C. vesicaria).

Principal components analysis was used to investigate the relationship between soil extractable nutrient concentrations and other environmental variables. A Principal Components Analysis (PCA) correlation biplot illustrating linear relationships between environmental variables was obtained using CANOCO 4 (ter Braak & Šmilauer, 1998) under the option centring and standardization by species. To examine, indirectly, the relationship between FVTs and growing conditions, all sites were subsequently recoded to their FVT.

A linear regression approach was used to describe interrelationships between tissue contents, environmental variables and strategy elements. Plotting as scatter graphs was used to confirm the simple monotonic nature of these relationships. Before analysis all variables were tested for normality (Shapiro-Wilks test) and transformed (logarithmic, square root or arcsine) where necessary. Redundancy analysis (RDA), which is an extension of multiple linear regression, was used to analyse the relationships between dependent variables (tissue contents) and environmental predictors. Tissue nutrient contents based on nonmanipulative field studies have three potential sources of explained variation: (i) variation in growing conditions independent of vegetation (ii) variation in vegetation independent of growing conditions (iii) covariation in vegetation and growing conditions. In fertilization and glasshouse-based studies two of these components would be controlled while varying the growing conditions or species composition independently. By contrast, in uncontrolled field studies it is difficult to determine if, for example, FVTs have different nutrient signatures because they grow preferentially in different places due to adaptive trait–environment relationships (source iii), or because of past evolutionary constraints (source ii). Consequently we used partial tests (Legendre & Legendre, 1998) to decompose the variance in tissue contents into its different components, followed by Monte Carlo permutation tests to test the significance of these components. These analyses are essentially a form of multiple linear regression (in this case undertaken via RDA in CANOCO 4) in which one component is first fitted to a variable followed by fitting of a second (potentially interrelated) component to the residual variation, thereby eliminating the covariation between them.

Tissue nutrient contents are known to vary over the course of the growing season, particularly in relation to the timing of flowering (Marschner, 1995). Since the onset of growth and flowering also vary between individual species the timing of sample collection with respect to phenology is an additional (and inevitable) source of variation in nutrient contents in extensive field-based studies of the type presented here. We sought to control this variation by minimizing the window during which samples were collected and by staggering sampling according to site latitude in an attempt to synchronize vegetation development at different locations. Most importantly, since all locations provided at least a few examples of each FVT, we ensured that samples of the four vegetation types were distributed similarly between the first and last sampling dates.

The type and severity of nutrient limitation was inferred from a joint consideration of absolute nutrient concentrations and N : P ratios in above-ground tissue based on critical values obtained from a review of fertilization studies conducted in a range of wetlands (Wassen et al., 1995; Koerselman & Meuleman, 1996). An N : P ratio > 16 and tissue [P] < 1.0 mg g−1 was regarded as indicative of P limitation. An N : P ratio < 14 and tissue [N] < 20.0 mg g−1 was regarded as indicative of N limitation. In the range of N : P ratios 14–16 (8% of samples) limitation by one or both nutrients (N-P colimitation) was assigned on the basis of absolute concentrations. In the case of K we used the value of 8 mg g−1 suggested by De Wit et al. (1963) based on Dutch agricultural grasslands as a tentative indicator of K deficiency. Fertilization studies in several low productivity wetlands have endorsed this value (e.g. Hayati & Proctor, 1991; Boeye et al., 1997). Where tissue concentrations exceeded 1.0 mg g−1 P, 20 mg g−1 N and 8 mg g−1 K we regarded the vegetation as being unlimited by nutrient supply irrespective of tissue nutrient ratios (Aerts & Chapin, 2000). Critical values for assigning nutrient limitation have been based on a large number of different vegetation types and plant species (Koerselman & Meuleman, 1996).

Although critical nutrient contents and their ratios have received broad endorsement in subsequent fertilization studies they are an imperfect tool that provide only circumstantial evidence of nutrient limitation (Boeye et al., 1997; Wassen et al., 1998; Van Duren & Pegtel, 2000). Thus we accept that on occasions we will have probably assigned nutrient limitation to unlimited stands or failed to recognize limitation where it may have been significant; there is an inevitable tradeoff between detailed knowledge at the level of individual species or sites and comparative surveys covering a wide range of sites and vegetation types (Keddy, 1992). Although more rigorous testing is required we assumed for the purposes of this study that tissue concentrations and their ratios are valid for exploring the nature of nutrient limitation in functionally distinct stands of vegetation (i.e. that interspecific and interstrategy differences in internal nutrient requirements are relatively insignificant compared to differences in nutrient acquisition or conservation strategies). In our approach biomass samples from multispecies stands also effectively reflect nutrient limitation of dominant species (i.e. the main contributors to stand biomass) when co-occurrence of differentially nutrient-limited subordinates may potentially be widespread (Hayati & Proctor, 1991; Roem & Berendse, 2000). Braakhekke & Hooftman (1999) offer a detailed discussion of these dilemmas.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Soil nutrient concentrations

The wide ranges of soil nutrient concentrations (Table 2) reflect the variation in sites from riparian forests and river floodplain grasslands to acid mires, calcareous fens and open water swamp.

Table 2.  Comparison of mineral nutrient contents in soils and above-ground plant tissue
 NPK
  1. Values refer to means ± one standard error, with ranges given beneath in brackets.

Soil (mg kg−1) (n = 315)11.18 ± 0.47 (0.22–29.87)5.06 ± 0.66 (< 0.1–175)177 ± 9.1 (13–1889)
Tissue (mg g−1) (n = 303)15.13 ± 0.28 (7.25–37.03)1.59 ± 0.05 (0.35–4.83)14.43 ± 0.45 (2.85–57.63)
 N : PN : KP : K
 11.10 ± 0.24 (4.54–22.14)1.23 ± 0.03 (0.33–3.82)0.12 ± 0.004 (0.04–0.49)

Relationships between soil nutrient concentrations and hydrochemical variables at the level of the overall dataset are summarized in the form of a correlation biplot (Fig. 1a). The diagram presented explains 40% of the variation in the environmental data. The first (horizontal) axis of this diagram emphasizes the importance of organic matter as a nutrient source (mainly N and P) and hence the close interdependence of N, P and K concentrations, while the second axis is dominated by the strong inverse correlation between pH and exchangeable iron and manganese. N was weakly inversely correlated with redox and water table depth (r2 = −7.0 and −6.0, respectively; P= 0.01) probably as a result of organic matter accumulation in summer waterlogged sites vs loss of N from free draining floodplains through nitrification. At the more base-poor sites (SC, TM and IM) mineral nutrients were generally positively correlated with manganese and iron concentrations, presumably either through immobilization, or their accumulation as organic matter. However, in the pooled data set these correlations were overwhelmed by the strong positive correlation between pH and P which is a reflection of the abundance of insoluble mineral phosphorus in the calcareous fens. P was also inversely correlated with soil Mn and soil Fe (r2 = −4.7 and −6.3, respectively; P ≤ 0.01). Soil N and P concentrations were positively intercorrelated (r2 = 29.0; P ≤ 0.0001).

image

Figure 1. Interrelationships between measured environmental variables summarized using a Principal Components Analysis (PCA) correlation biplot (top left). Eigen values: axis 1 = 0.22, axis 2 = 0.18. Axis scores are from +1 to −1. Codes for environmental variables that are not self explanatory are: EH (redox potential standardized to pH7); % org (organic matter based on loss on ignition); bare (extent of bare ground resulting from flood scouring, sedimentation or livestock poaching); GWD (depth to groundwater table); mgmt (management inputs due to moving or grazing); flood (temporal variation in surface inundation due to riverbank overtopping). Remaining panels illustrate distribution of wetland Functional Vegetation Types (FVTs) in relation to environmental variables. Centroids for the different FVTs are shown in bold and are located at the median of the sample axis scores. Axis scores are from +2.5 to −2.5.

Download figure to PowerPoint

Distribution of vegetation with respect to environmental variables

The distribution of FVTs with respect to soil nutrient contents and other selected hydrochemical variables is summarized in Fig. 1 by overlaying FVTs on the PCA correlation biplot. The major patterns to emerge are that R stands, including mown or grazed inundation grasslands and pioneer communities, occurred on well aerated, organically impoverished and significantly less nutrient-rich soils (P 2.53 mg g−1 and N 4.66 mg g−1) compared to the other FVTs (P 5.84 mg g−1 and N 12.9 mg g−1 for C, CS and S stands combined). This is borne out by a pronounced inverse correlation between %R and soil N (r2 = −21.9; P ≤ 0.0001) and a weaker negative correlation with soil P (r2 = −6.8; P = 0.001). The proportion of other strategies or the distribution of other FVTs could not be clearly linked to potential nutrient supply. The C strategy showed a very weak increase with soil N and P (r2 = 2.1 and 3.5, respectively; P = 0.01), while the representation of stress tolerant species increased slightly with soil N (r2 = 4.2; P = 0.001) but was independent of soil P. The contribution by cover of the three strategy elements to the above-ground vegetation was independent of soil K concentration. S stands had a sharply contrasting pattern of distribution to R stands, comprising mainly mire communities situated on acid soils with high metal contents, which were often also peaty and poorly drained, with modest nutrient contents. An outlying group comprising short sedge dominated vegetation in spring fed calcareous fens was associated with high P and pH. C and CS stands occupy intermediate positions in the biplot, covering a wide environmental range. C stands favoured slightly more base- and nutrient-rich sites, in situations contrasting between permanent inundation (i.e. open water swamps) or zones of strong groundwater recharge where brief winter flooding was followed by a low summer water table (e.g. river embankments and high elevation zones within riparian forest). CS stands were most associated with moderately peaty, often slightly acidic soils with moderate metal ion concentrations and a variable summer water table, but often experienced prolonged winter and spring flooding.

Decomposition of variance in tissue nutrient concentrations

The relative importance of different sources of variation in tissue nutrient contents or their ratios are depicted in Fig. 2. Note the general predominance of the vegetation-environment covariation and that there are marked differences between nutrients in the relative importance of different sources of variation. For example, variation in K has a strong environmental basis whereas in P variation is largely attributable to the strategy composition of the vegetation alone. In all cases growing conditions have a significant effect irrespective of the vegetation present. Equally, with the exception of the N : K ratio, the strategy composition explains a significant component of the variation in nutrient concentrations independently of growing conditions, indicating that there are constitutional differences between growth strategies in terms of nutrient signatures.

image

Figure 2. Partitioning of variance in nutrient contents and ratios based on multiple linear regression using redundancy analysis (RDA). Bracketed values refer to the proportion of the total variance explained by all variables supplied simultaneously (x). Pie charts refer only to the distribution of the explained variance between three sources. Black, plant strategy effects only (y) (i.e. with environment acting as a covariable); white, environment effects only (z) (i.e. with plant strategy composition acting as a covariable); hatching, plant strategy-environment covariation (i.e. x– (y + z)). All analyses and the variance explained by individual components are significant at P = 0.005 (Monte Carlo tests with 199 random permutations) unless indicated otherwise (for N environment effects P = 0.01). Note that the significance of the covariation cannot be tested since this component cannot be derived independently. In the case of P : K the covariation component is slightly negative since (y + z) > x (Baar & Ter Braak, 1996).

Download figure to PowerPoint

Tissue nutrient concentrations with regard to environment

Tissue nutrient contents (Table 2) ranged very widely and were poorly predicted by a range of environmental variables. Many site-specific correlations (not shown) were not reproduced at the level of the combined data set (Table 3). The total variation explained by the 12 environmental variables ranged from 35% (K) to 21% (P : K), although the contribution of five variables (redox potential, ground water depth, management, soil K and soil P) was negligible. In a partial RDA test, where the environment-vegetation covariation was removed, the significance of almost all correlations was weakened, as would be expected, but the underlying relationships were largely unchanged. Tissue nutrient contents were consistently negatively correlated with soil organic matter content and positively correlated with pH. Negative correlations with extractable iron and manganese concentrations suggested that nutrient availability was being regulated through the effect of low pH and waterlogging on N mineralization and leaching of K. The extent of bare ground, and to a lesser degree the flood regime, was associated with increased availability of all nutrients, especially P, suggesting that flooding, or riverborne sediment specifically, was acting as a nutrient source. Nutrient contents were generally independent of their concentrations in the soil (Table 3). Plant P was significantly negatively correlated with soil N illustrating that an adequate supply of N may, by stimulating growth, compound a deficiency in P, and vice versa.

Table 3.  Correlations between tissue nutrient concentrations and environmental variables
Tissue nutrientKNPN : PN : KP : K
  1. All r2 values are significant at P = 0.05 and refer to the percentage variance explained in a full RDA. Underlined values remained significant in a partial RDA with vegetation strategy composition acting as a covariable. Bracketed values were unique to the partial test. Total explained refers to the percentage variance explained by all variables in a multiple regression based on the full RDA1 or the partial RDA2.

Organic matter     −4.8     −7.510.2 5.2
Soil Mn       −7.4−5.2−2.2  7.1
Soil Fe−10.6−5.4       −1.96.0  5.0
Redox potential
pH      11.211.5       4.5−3.4−1.5
GWD−2.3(−2.5)
Management
Flood regime   (−1.7)        3.3−5.23.79.6
Bare ground            7.310.7 15.6−5.91.8
Soil K–1.9
Soil N  −6.6−10.614.8       7.0
Soil P 1.4
Total explained1     35.422.9 30.826.227.8 20.9
Total explained2      13.2 9.3 10.214.817.4 23.2

Tissue concentrations with regard to strategy elements and FVTs

Correlations between the representation of different strategy elements and the absolute or standardized tissue nutrient concentrations in each sample are listed in Table 4. Selected examples are shown in Fig. 3. The standardized concentrations are calculated using the site scores in a partial RDA where the relative abundance of plant growth strategies provides the predictive variable and environment acts as the covariable. By removing the covariation between vegetation and environment this analysis provides an indication of the predicted relative nutrient contents (re-scaled to a range of 0–1, for ease of comparison) that would be recorded in FVTs grown under standard conditions. It is evident from a comparison of Table 3 and Table 4 that the representation of different strategy elements can predict tissue nutrient contents in multispecies vegetation as well as or better than the best environmental characteristic at the level of the whole data set.

Table 4.  Correlations (r2) between the representation of primary plant growth strategies and plant tissue nutrient contents and ratios
Tissue nutrientKNPN : PN : KP : K
  1. Correlations shown are significant at P = 0.05 (*); P = 0.01 (**) or P = 0.001 (***). Upper line gives correlations with observed nutrient contents. Lower line gives correlations after removal of vegetation-environment covariation. n= 303.

C  2.5*  1.8*  1.7* 1.3*
 
S−13.3***−10.0***−30.6*** 25.0*** 4.8**−3.8**
  −2.0*−11.0*** 21.4*** 9.0***
R 28.1*** 18.7*** 44.7***−19.6***−9.2***
  14.8***  7.1*** 19.9***  −3.7*
image

Figure 3. Change in absolute and standardized N and P contents and the N : P ratio with increasing representation of ruderal (upper) and stress tolerant (lower) species. Regression lines are significant at P = 0.001 except where an asterisk appears which is significant at P = 0.05.

Download figure to PowerPoint

Tissue nutrient contents were clearly independent of the C strategy element but there were marked and contrasting changes with the representation of R and S strategies. In the case of R there was a marked increase in absolute and standardized contents of all three nutrients, but, because the rate of increase in P exceeded N there was a decrease in the N : P ratio. In the case of the S strategy there were negative correlations with observed N, P and K and standardized P. Although absolute P was highly variable there was a marked increase in the observed and standardized N : P ratio. The N : K ratio was significantly but more weakly correlated with growth strategies (R and S) and showed the same pattern of change as the N : P ratio. Standardization resulted in the loss of these trends implying that they had a mainly environmental basis. The P : K ratio was weakly negatively correlated with the S strategy. This pattern was reversed by standardization due to the removal of the negative vegetation-environment covariation (Baar & Ter Braak, 1996).

When these trends are translated to FVTs it is evident that functionally different wetland vegetation has highly characteristic signatures in terms of mineral nutrient contents (Fig. 4). Tissue [K] was lowest in S (9.49 ± 0.36 mgg−1) and highest in R (19.7 ± 0.36 mgg−1). In standardized terms these differences were dampened but K levels in R stands remained relatively high. Tissue [N] was similarly lowest in S (12.3 ± 0.25 mgg−1) and highest in R (17.2 ± 0.94 mgg−1). The differences in standardized terms were significant although much reduced, with N contents remaining lowest in S stands. Tissue [P] exhibited the most pronounced pattern with lowest mean contents in S stands (1.00 ± 0.05 mgg−1) compared to 2.30 ± 0.13 mgg−1 in R stands, with C and CS intermediate. This pattern of differences was maintained after standardization. These results are interpreted in terms of nutrient ratios in Fig. 5. N : P ratios were highest in S stands (14 ± 0.52), intermediate in C and CS stands (11 ± 0.40) and relatively low in R stands (7.87 ± 0.30). The contrast between FVTs remained equally acute in the standardized data. FVTs also differed in N : K ratios although differences were much less pronounced compared to N : P and disappeared upon removing the vegetation-environment covariation. The results for the P : K ratio reflect the affect of standardization discussed above and indicate that P : K ratios will typically be highest in S stands compared to the other FVTs.

image

Figure 4. Comparison of absolute (left hand column) and standardized (right hand column) tissue nutrient content in different functional vegetation types (FVTs). Values are means (±1 SE) with units mgg−1 for absolute contents. Standardized values are relative only and are derived from redundancy analysis (RDA) site scores rescaled to a range of 0–1. Letters connect FVTs not significantly different at P = 0.05 (Fishers LSD test).

Download figure to PowerPoint

image

Figure 5. Comparison of absolute (left hand column) and standardized (right hand column) tissue nutrient ratios in different Functional Vegetation Types (FVTs). Values are means (± 1 SE) with units mgg−1 for absolute contents. Standardized values are relative only and are derived from RDA site scores rescaled to a range of 0 to 1. Letters connect FVTs not significantly different at P = 0.05 (Fishers LSD test).

Download figure to PowerPoint

The extent of nutrient limitation in wetland vegetation

In terms of the overall data set 86% of samples showed some form of nutrient limitation as inferred from tissue concentrations. 70% of the sites were subject to some form of N-limitation (N singly in 73% of these cases). P-limitation alone affected 9% of sites with a further 10% affected by P colimitation with N or K or (very rarely) both. K limitation was tentatively inferred at 19% of sites, mostly (68%) as N-K colimitation. There were no examples of pure K limitation. Tissue nutrient concentrations were strongly intercorrelated. This is illustrated in the context of critical ratios and concentrations in Fig. 6.

image

Figure 6. Intercorrelations between tissue nutrient contents with regard to critical concentrations of N, P and K and the N : P ratio. Regression lines on a–c are significant at P = 0.0001. (a) Plant P vs K with upper limits of contents for K and potential P limitation. (b) Plant N vs K with upper limits of contents for K and potential N limitation. (c) Plant N vs P. Stands falling in the upper right sector were regarded as unlimited and in the upper left sector are probably limited only by N. The lower left sector is expanded in (d). Within the range of potential limitation by both N and P the critical ratio for N limitation is indicated by 14 : 1. Sites falling above this line therefore indicate N limitation. The critical ratio for P limitation is indicated by 16 : 1. Sites falling below this line and below the upper threshold for P limitation (1 mg g−1) therefore indicate P limitation. Between these critical ratios and below the critical contents it is assumed that N and P are colimiting.

Download figure to PowerPoint

When the total data set was subdivided into FVTs several major differences emerged (Fig. 7): (i) C and R stands were the most likely to escape nutrient limitation (31% and 26%, respectively) while S stands were invariably nutrient limited; (ii) few C and CS stands showed evidence of P limitation and R stands were apparently unlimited by P but 50% of S stands were likely to be subject to P limitation; and (iii) colimitation by K was almost twice as frequent in S stands as it was in the other FVTs.

image

Figure 7. Pattern of inferred nutrient limitation in different functional vegetation types (FVTs) and all stands combined n, C : 58; CS : 121; S : 71; R : 53. Closed circles, K; open circles, N; squares, P.

Download figure to PowerPoint

Inter relationships between tissue nutrient concentrations and vegetation state variables

Regression coefficients between a range of vegetation state variables and tissue nutrient contents are given in Table 5. Relationships were dominated by negative correlations between biomass-related variables and tissue concentrations. Several state variables (stem diameter, nearest neighbour distance, litter cover and canopy height) were unrelated to nutrient contents. Although there were significant negative correlations with total biomass these relationships were most dependent on the amount of low level biomass (< 10 cm high) whereas tissue contents were more variable in stands characterized by a high (i.e. > 20 cm high) leaf canopy (Fig. 8). The absence of significant correlations between tissue contents and the relative proportions of biomass in different height classes (data not presented) confirmed that these relationships were based on absolute values rather than patterns of biomass allocation.

Table 5.  Significant correlations (r2) between vegetation state variables and tissue nutrient concentrations
 STEM0–10 BIO10–20 BIO20 + BIOTOT BIOREPR
  1. Correlations shown are significant at P = 0.01 (**) or P= 0.001 (***). n= 303.

[N]−31.8***−22.6***−4.9**−20.5***−7.1***
[P]−4.5**−23.2*** −8.6*** −8.4***
[K]−5.6**−22.6***−13.8***−12.6***−7.9***
N : P    5.1**
image

Figure 8. Correlation between total tissue nutrient concentrations and above ground biomass in two height strata. K, closed circles; N, open circles; P, squares. N= 305.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Studies of wetland vegetation have a long tradition in phytosociology. Consequently there are many highly detailed descriptions of the composition of European wetland plant communities and their environmental affinities. More recent studies have focused on gradients, for example from discharge fens through transitional mires to bogs, and have remarked on the general structural characteristics of the dominant species (Verhoeven et al., 1988). Other studies, focusing on nutrient budgets over similar gradients, have independently emphasized changes in nutrient concentrations in plant tissue (e.g. Waughman, 1980; Wassen et al., 1995). Consequently the present study demonstrates a direct link between habitat, functional characteristics of wetland vegetation and nutrient signatures. It also raises the issue of whether these signatures are primarily phenotypic (i.e. a product of growing conditions) or if they relate more strongly to intrinsic (genotypic) properties of the vegetation.

Habitat characteristics of FVTs in relation to nutrient limitation

Most vegetation in this study was subject to some form of inferred nutrient limitation, generally by N, as appears to be the norm in temperate ecosystems (Aerts & Chapin, 2000; Van Duren & Pegtel, 2000). Different forms of nutrient limitation were not distributed randomly between FVTs partly because to some extent these occupied hydrochemical environments that posed different nutritional challenges (Fig. 1). However, in line with previous studies (e.g. Hayati & Proctor, 1990; Wheeler et al., 1992) soil extractable concentrations of nutrients on their own were not an adequate reflection of bioavailability.

In riverine wetlands the narrow growth window between exposure of alluvial sediment and its re-flooding, erosion or burial precludes the development of slower growing, longer-lived stress tolerators. Similarly, regular mowing, poaching or grazing suppresses the development of strongly competitive species. Species with strongly or partially ruderal strategies are therefore typically well represented in habitats ranging from river gravels and temporary floodplain water bodies to river forelands and traditionally managed floodplain meadows (Menges & Waller, 1983; Rodwell, 1992, 2000; Abernethy & Willby, 1999). In such situations the soil nutrient pool is small due to organic impoverishment or export of biomass and vegetation was frequently subject to N-limitation, as predicted for early successional environments by Vitousek & Walker (1987). However, nutrient availability, particularly of P, may be relatively high due to inputs of P adsorbed to riverine sediment and the limited opportunity for immobilization of phosphorus by iron or organic matter complexes. The low N : P supply ratio is liable to accentuate N-limitation. In open, damp, alluvial habitats, with high surface irradiance conditions should favour mineralization of integrated organic matter and nutrient uptake, as well as rapid germination of seeds. Superficial detritus associated with flood strand lines or exported from adjacent riparian forest and beds of emergent monocots must act as an important supplementary source of nutrients in these habitats in the same way that strandline deposits nurture the vegetation of coastal foredunes (Davy & Figueroa, 1993).

The stress tolerator FVT was characterized by low productivity vegetation in peaty, waterlogged sites in acid mires and calcareous spring fed rich fens. S stands had lower concentrations of all nutrients and relatively high N : P and N : K ratios. Critical nutrient contents implied frequent P limitation or colimitation (usually with K) as often previously observed in these habitats (e.g. Boyer & Wheeler, 1989; Hayati & Proctor, 1991; Boeye et al., 1997; Shaver et al., 1998). Phosphorus limitation operates by dual mechanisms in habitats of contrasting pH. In discharge fens at high pH phosphorus is often immobilized by coprecipitation as forms of calcium phosphate (Boyer & Wheeler, 1989). At low pH organic P is reversibly bound to amorphous iron-organic matter complexes, while inorganic P is commonly immobilized as insoluble iron or aluminium phosphates (Shuman, 1991; Kooijman et al., 1998). High concentrations of Fe, Mn and Al often coincide with low soil fertility (assessed phytometrically) although they may be largely independent of extractable concentrations of principal plant nutrients (this study (Fig. 1), Wheeler et al. (1992). Adaptations to mitigate metal ion stress may pose a further constraint on the availability of P relative to other nutrients in S stands (see below). Several S stands were also situated on East Anglian fens located in intensely agricultural catchments, where export of nutrients following reinstatement of hay-making on hydrologically intact sites has restored species-rich spring-fed fen communities associated with P limitation (Koerselman et al., 1990; Fojt & Harding, 1995). Although P limitation is commonly emphasized in small sedge mires it should be stressed that N or NK limitation still accounted for almost half of the inferred limitation of the stress tolerator FVT in the present study. Mineralization of N is generally inhibited in waterlogged, acid soils (De Laune et al., 1981), which also promote leaching of soluble K due to desorption from the cation exchange complex (Bolt & Bruggewert, 1976). Sites subject to forms of N limitation generally had intermediate pH values (in the range 4.6–6.5) where immobilization of P will be relatively weak thus increasing the likelihood of N limitation.

While the growing conditions of S and R stands were strongly divergent the habitat characteristics of C and CS stands overlapped substantially, both with each other and the other FVTs. Habitats ranged from elevated recharge zones subject to predictable winter flooding by nutrient-rich river water and fens subject to a combination of soligenous and topogenous influences (e.g. areas of groundwater discharge within river floodplains), through to the permanently inundated margins of rivers or floodplain water bodies (i.e. open water swamps). Other habitats included groundwater discharge fens with a history of nitrate enrichment of their aquifer, build up of nutrients due to cessation of harvesting or grazing (potentially releasing vegetation from P or K limitation), or a decline in the water table through over-abstraction (Wheeler & Shaw, 1991; Harding, 1993). Although not obvious from our pooled data set, the distribution of C stands was biased towards zones of major nutrient enrichment when examined at an individual site level (Willby et al., 1997). Similar patterns in riverine wetlands have been described by Menges & Waller (1983), Wassen et al. (1995) and Richardson et al. (1999). The greater incidence of nutrient limitation (mainly by N) that we observed in CS stands is consistent with these observations and suggests that in typically productive habitats CS species will persist mostly as subordinates or where the performance of dominant species is locally reduced (e.g. by high divalent metal ion concentrations, Wheeler et al., 1985).

In well-drained sites the effect of increased release of N through peat mineralization should logically be to increase the rate of P limitation (Roem & Berendse, 2000). However, the evidence from our data and from other surveys of high standing crop vegetation is that C and CS type vegetation in fens and marshes is most commonly N limited (Koerselman & Verhoeven, 1995; Bedford et al., 1999; Van Duren & Pegtel, 2000). Presumably, at the level of the environment, this is a consequence of very effective retention and cycling of P (e.g. through high mineralization of P from deep fen peats) under conditions of simultaneous N enrichment (Verhoeven & Arts, 1987). In floodplains and open water swamps widespread N limitation of vegetation, as reported by Van Oorschot et al. (1997) and Wassen et al. (1995), may jointly reflect the scale of alluvial inputs rich in inorganic P (Klopatek, 1978) and the loss of more mobile N from plant litter via leaching, mineralization or denitrification. Conversely, the relatively high proportions of C and R stands that were free of nutrient limitation (none of which had been fertilized for at least 5 yr before this study) are probably testimony to the overall scale of allochthonous nutrient inputs during winter flooding (Wassen et al., 1995; Willby et al., 1997). In the particular case of swamps, the open water nutrient pool, which is continuously replenished by water movement, must be an important supplementary source of nutrients accessible via adventitious roots and submerged stems.

Nutrient signatures of FVTs in relation to plant growth strategies

This study has revealed that wetland FVTs have significantly different nutrient signatures. Since FVTs are distributed nonrandomly within the overall wetland environment these differences will reflect, at least partly, covariation between FVTs and environment. They may relate to indirect correlations with morphological traits, such as the ratio of light harvesting to supportive tissue, that confer a competitive advantage in particular environments (Chapin, 1980; Grime et al., 1988, 1997). Alternatively, they may reflect selection for traits that affect ability to deplete or respond to changes in nutrient supply in resource limited environments (e.g. specific uptake or storage mechanisms) and which determine tissue nutrient contents (e.g. specific root length) or are governed by them (e.g. relative growth rate) (Tilman, 1982; Aerts & Chapin, 2000). Comparison of Fig. 1 and Fig. 6 also reveals that the variation in nutrient content within FVTs is very small relative to the variation in growing conditions. Thus pronounced differences in tissue contents between FVTs are generally preserved even after the covariation between vegetation and environment is removed (Figs 6 and 7). Hence our results also bear the hallmarks of inherent differences in plant growth strategies due to past evolutionary tradeoffs.

Ruderal species are phenotypically plastic and have small but rapidly proliferating root systems with high absorption capacities for limiting nutrients (Grime, 1979; Chapin, 1980; Crick & Grime, 1987; Grime et al., 1997). Consequently they are highly responsive to the marked spatial variation in nutrient supply that is a common feature of disturbed environments, such as riparian zones. These traits reflect a requirement for effective nutrient acquisition strategies in fast-growing species where rapid leaf turnover results in high nutrient losses which must be offset to maintain photosynthetic efficiency (Grime, 1979; Chapin, 1980; Lambers & Poorter, 1992). Consequently, hydrochory, which removes the barrier of seed size on dispersal, is likely to have an important bearing on postgermination nutrient uptake potential in riverine wetlands. Our results, for absolute and standardized nutrient contents clearly testify to the high nutrient affinity of ruderal wetland vegetation. They mirror observations from terrestrial habitats, for example the high nutrient content of plants occurring in the early stages of woodland succession (Woodwell et al., 1975). Auclair (1979) even observed the same trend in nutrient contents from pioneer to seral species within a suite of essentially dominant, stand-forming emergent monocots. McJannet et al. (1995) found that ruderals showed reduced N and P contents relative to other species when grown in excess fertilizer due to rapid reinvestment in new growth as opposed to storage of resource capital. However, under natural conditions where nutrient limitation, particularly by N, appears to be almost routine reinvestment in biomass on a scale that dilutes tissue nutrient concentrations is not feasible. Resources are then likely to be diverted to seed production, and therefore, as we found, ruderal-dominated vegetation is likely to possess comparatively high concentrations of N, P and K (in addition to other mineral nutrients) (Grime et al., 1997; Thompson et al., 1997).

Our observation that N : P ratios are typically lower in ruderal vegetation and higher in vegetation dominated by stress tolerant species is consistent with the results of several recent comparative surveys of mainly terrestrial plants (Grime et al., 1997; Thompson et al., 1997), as well as more general evidence for species and community level adaptations to acquire or conserve the nutrient most limiting to production (Bowman, 1994). Kooijman et al. (1998) also observed that phosphorus uptake was high relative to nitrogen during the early stages of succession in acid dune systems. Meaningful comparisons of N : P ratios in different functional groups growing under field conditions could potentially be distorted by differences in the relative availability of N and P in early vs late successional environments; it is therefore significant that ruderals have been found to maintain markedly lower N : P ratios than other functional groups of wetland plants when grown under standard conditions (McJannet et al., 1995), or, as in our case, when covariation between vegetation and environment is adjusted for. However, the greater importance attributed to phosphorus conflicts with the conventional view of ruderals as nitrophiles, based on the close link between nitrogen nutrition and carbon assimilation (75% of plant N is localized in chloroplasts, mostly in the carbon-fixing enzyme Rubisco; Chapin, 1987). Since P-deficiency restricts flower initiation and seed development (Barry & Miller, 1989) its avoidance has obvious implications in a group of plants characterized by annual/biennial life histories. Consequently a strong positive correlation between P content and relative growth rate (Thompson et al., 1997) may derive partly from the fact that those species with the highest relative growth rates (i.e. ruderals; Grime et al., 1988) also regenerate extensively by seed. Thus we interpret the low N : P ratio in ruderals as evidence of relatively efficient N use due to high nitrogen productivity (sensuBerendse & Aerts, 1987) compared to rapid consumption and diversion of P into reproduction.

Our results demonstrate that S stands contained significantly lower concentrations of all three nutrients than the other FVTs (Figs 3 and 4). These results support the close correlation between low tissue nutrient status and a suite of typical attributes of stress tolerators, including an ability to sustain yield under low nutrient supply, low relative growth rate, high leaf longevity, low phenotypic plasticity, low root absorption capacity and effective nutrient resorption before sensecence (Chapin, 1980; Reich et al., 1992; Grime et al., 1997; Aerts & Chapin, 2000). Hence they support the theory that competitive ability in nutrient limited habitats is governed by effectiveness in lowering the external concentration of a limiting nutrient (Tilman, 1982, 1990; Mamolos et al., 1995) by increasing its rate of acquisition or reducing its loss. These abilities are reflected in minimal tissue concentrations (Aerts & Chapin, 2000).

After removal of the vegetation-environment covariation, our results indicated that nutrient contents of S stands differed markedly from C or CS stands only in terms of lower P and consequently a higher N : P ratio. This supports the existing evidence that stress tolerators focus uptake on more mobile ions such as nitrate while maintaining P reserves due to low internal P demands and efficient conservation (Chapin, 1980; Evans, 1989; Daoust & Childers, 1999). Given the relatively high P : K ratio that we observed this trend apparently does not extend to other mobile ions such as potassium (cf. Veerkamp & Kuiper, 1982). Specific nitrogen uptake mechanisms include preferential use of ammonia where waterlogging favours the reduction of nitrate (De Graaf et al., 1997), although the ionic antagonism between uptake of ammonium (or manganese) and potassium may then lead to K deficiency (Ernst, 1990). Alternatively, radial oxygen release in the rooting zone may be used to promote microbial oxidation of ammonia to nitrate (Engelaar et al., 1995). It is also now known that some nonmycorrhizal sedges utilize organic nitrogen in arctic and alpine environments (Kielland, 1994). Recent studies (e.g. Raab et al., 1999) confirm that this ability extends to plants in subalpine wetlands, comparable to some sites that we studied, where nitrogen mineralization is limited by factors other than temperature. Uptake via a range of mechanisms will be enhanced by increased nitrogen and biomass allocation to below ground organs or a high root surface area or root length per unit root weight, as previously observed in low productivity species (Aerts et al., 1992; Elberse & Berendse, 1993). The emphasis on nitrogen uptake mechanisms may express the need to recover nitrogen losses due to low nitrogen use efficiency in some low productivity species (Bridgham et al., 1995; Aerts & de Caluwe, 1997). These losses arise as a result of the sequestering of N in cell wall materials or secondary compounds, such as polyphenolics (Niemann et al., 1992; Aerts & de Caluwe, 1997), that prolong leaf life span (Poorter & Bergkotte, 1992), deter herbivores (Grime et al., 1996), and ultimately reduce N cycling, thereby delaying the consequences of ecosystem eutrophication (Aerts, 1997). Stress tolerators however, also retain lower relative growth rates than other primary plant strategies under standard growing conditions (Grime et al., 1988). Therefore, from tissue contents alone, we cannot exclude the possibility that luxury consumption of nutrients rather than their investment in growth (Chapin, 1980) contributes to the reduced differential with other strategies in terms of standardized N (and K).

Mechanisms that enhance P uptake in stress tolerators under infertile conditions are often emphasized, including extensive colonization by arbuscular mycorrhiza (Allen, 1991; Grime et al., 1997) and secretion of organic acids by both roots and associated microorganisms (Marschner, 1995; Pérez Corona et al., 1996). However, Pérez Corona et al. (1996) found that the ability to mobilize P from insoluble inorganic sources in a range of Carex species did not correlate with the P availability of their natural habitats. This supports the view that the primary adaptation among stress tolerators to low P availability is an increase in P resorption prior to senescence (Chapin et al., 1982; Berendse & Aerts, 1987; Bowman, 1994). This is borne out by the constitutionally low P contents we found in S stands. Moreover, van Oorschot et al. (1997) have quantified higher phosphorus use efficiency and retranslocation of phosphorus in wet heath vegetation on one site that we studied (Kismeldon Meadows) that provided several examples of S stands. In wetlands the trend for conservation of P might be compounded by the negative effect of waterlogging on root absorption capacities (Gore & Urquhart, 1968). Plant adaptations to reduce stress from high concentrations of divalent metal ions are also likely to impose specific constraints on P nutrition (Laan et al., 1989; Snowden & Wheeler, 1995; De Graaf et al., 1998). Many of the small sedge species predominating in the S stands in this study have extensive root aerenchyma that promotes radial oxygen leakage to the rhizosphere (Green & Etherington, 1977) thereby reducing assimilation of potentially phytotoxic ferrous ions (Snowden & Wheeler, 1993). However, oxidized root plaques are liable to obstruct P uptake (Christensen & Wigand, 1998) or reduce P availability through formation of Fe-P deposits (Snowden & Wheeler, 1995). Separate internal detoxification mechanisms may also result in Fe being immobilized within shoot tissue at the expense of P translocation from the roots (Tannaka et al., 1966; Jones, 1975).

Competitor species achieve high rates of resource capture through active foraging via extensive root systems that provide a large absorptive volume per unit area (Aerts et al., 1992) and therefore tend to have high nutrient contents and a correspondingly high growth rate (Grime et al., 1997). Nutrients are deployed for the formation of canopy foliage (Grime et al., 1988) which requires supportive tissue with a high fraction of cell wall materials such as lignin. Consequently average nutrient contents were generally lower in this group of plants than in ruderals where nutrients reside mainly in the cytoplasm (Grime et al., 1996). In stable, fertile wetlands this is a reflection of the relative importance of effective light interception in the upper canopy layers in determining competitive ability (Elberse & Berendse, 1993; Aerts & de Caluwe, 1994) and therefore the requirement to position the most nitrogen-rich photosynthetic tissue in a way that optimizes carbon gain (Evans, 1989; Poorter et al., 1990). The graminoid growth form meets many of these requirements by having a basal meristem and erect growth while vertically inclined foliage may help reduce self shading. Extensive aerenchyma and the promotion of root aeration via convective gas flow offer significant additional benefits in wetlands (Armstrong et al., 1992) as well as reducing the influx of reduced metal ions and promoting nutrient availability through increased mineralization of organic matter. The average N : P ratio in C and CS stands approximated the supposed optimum for balanced plant growth (c. 10; Aerts & Chapin, 2000). Lower concentrations of P and a higher N : P ratio relative to ruderals are consistent with a strong emphasis on vegetative propagation (mainly via lateral extension of the rhizome). Rhizome storage also facilitates an early onset of growth through translocation of photosynthate (Grime, 1979). In emergent plants, such as Typha latifolia, Phragmites australis and Sparganium erectum, this supports the elevation of new shoots above the water surface and more generally may help synchronize growth to take advantage of temporal fluctuations in resource supply, such as the nutrient flush from decomposing litter (Smith et al., 1988). Competitor species are renowned for producing copious quantities of relatively nutrient-rich litter (Grime, 1979; Grime et al., 1988). The generally more efficient resorption of P and the immobilization of N in litter might be expected to accentuate N limitation (Aerts & Chapin, 2000). However, investigations of reedswamp species indicate that leaching and recycling of N from senesced material may be very rapid (Mason & Bryant, 1975).

CS stands represented the dominant functional type of wetland vegetation. Despite their low phenotypic plasticity, homogenous growth, moderate potential height, low root absorption capacities and intermediate growth rates relative to competitor strategists (Crick & Grime, 1987; Grime et al., 1988; Aerts et al., 1992; Aerts & de Caluwe, 1994) we were unable to detect a significant difference in nutrient contents between CS and C stands, other than lower observed K contents and a higher incidence of inferred N limitation in CS stands. This may highlight a weakness in our approach. Alternatively it might suggest a shift in selection for different components of nutrient use efficiency, namely from high nutrient productivity in fertile habitats to high nutrient retention in less fertile habitats (Berendse & Aerts, 1987), without any overall net affect on tissue contents. CS stands possessed several features including large below-ground biomass, often with massive rhizomes, moderate to large plant size and high relative growth rates compared to S stands that should have provided a competitive advantage in moderately fertile habitats. In wetlands this may include a greater ability to respond to nutrient pulses associated with topogenous inputs and litter decomposition and to translate this into biomass (Crick & Grime, 1987). Thus, in the absence of management, some species (e.g. Cladium mariscus, Filipendula ulmaria) formed extensive monodominant stands. Recent studies have also indicated higher nitrogen use efficiency in high productivity Carex species compared to low productivity congeners. This was due to higher N productivity and higher N resorption caused by lower investment in leaf cell wall materials and phenolics (Aerts & de Caluwe, 1997) which lead to faster rates of litter decomposition and N cycling. A superior nitrogen economy is the expected response in species confronted with N limitation. However, the relatively extravagant rates of leaf turnover in characteristic CS dominants, such as Molinia, should place them at a long-term disadvantage in chronically P-limited sites (Aerts & van der Peijl, 1993). Similarly, although iron tolerance based on exclusion by oxidative precipitation remains high in many CS species (Snowden & Wheeler, 1993), their greater areal capacity for below ground absorption may increase long-term exposure to toxic metal ions with subsequent negative effects on growth.

Nutrient contents with respect to state variables

A negative correlation between tissue N and total biomass has been observed previously (Polisini & Boyd, 1972; McJannet et al., 1995) and may prove to be a vegetation generality based on the increasing dilution of foliage nutrient concentrations by nutrient-poor but high density, supportive tissue under conditions of increasing fertility. Thus there is a predictable succession across the biomass gradient of FVTs with different nutrient signatures (median of 375, 640, 790 and 880 gDWm−2 for R, S, CS and C stands, respectively). Koerselman & Meuleman (1996) failed to detect a biomass-nutrient content relationship in a review of fertilization studies from a variety of wetland ecosystems but this may have been due to the absence of typical ruderal vegetation from their data set. However, the strongest relationships that we observed were within the lowest height strata (0–10 cm), where the highest biomass was associated with leafy tissue of short growing stress tolerators (median biomass 337 gDWm−2) while the lowest values were represented by the broad, well spaced stems, leaf bases and lower leaves of stand forming competitors and leafy, short-statured ruderals (median biomass in the 0–10 cm height range 198 and 206 gDWm−2, respectively). This relationship is supported by the observation that total leaf area below 15 cm was greatest in low productivity grasses (Elberse & Berendse, 1993). It reflects a general underlying trend for the amount of sclerified tissue, and consequently tissue density, leaf life span and nutrient retention time to increase with the representation of slow growing stress tolerators (Aerts & Chapin, 2000). The negative correlation between contents of N and K and the number of reproductive structures is presumably a reflection of nutrient dilution in flowering stems (Wassen et al., 1995; N. J. Willby unpublished). In fertile environments light becomes proportionally more limiting due to an increase in biomass and height will therefore be a good index of competitive ability (Lepš, 1999). The lack of a significant correlation between nutrient contents and height presumably reflects the inclusion of short growing ruderal vegetation with very high nutrient contents on open sites where high rates of disturbance favour regenerative activity over increases in height. At the opposite extreme nutrient-rich foliage of very tall species, such as Urtica dioica and Phragmites australis, must be supported by rigid, nutrient-poor stems.

Conclusions

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

These results extend the evidence for adaptive divergence in plants based on a tradeoff between effective resource acquisition and growth in productive environments and conservation of resources under infertile conditions (Grime, 1979; Berendse & Aerts, 1987). They are significant, first because they confirm that these patterns exist within a single environment (i.e. minerotrophic wetlands) and not merely between sharply contrasting environments. Hence, despite the unique importance of wetness and associated hydrochemical stresses in wetlands nutrient availability remains a major overarching theme. Second these results relate to a data set that is geographically diverse and covers a very wide range of wetland types and associated plant communities. Third, they demonstrate strong patterns at a community level and hence the validity of scaling up from species to vegetation. Finally, they highlight marked patterns, independent of environment, within purely herbaceous vegetation rather than at the level of coarser subdivisions of plant growth forms. Consequently vegetation merits a priority role over abiotic factors in controlling nutrient recycling within wetlands (Aerts et al., 1999). Wetlands are complex environments in terms of spatial and temporal heterogeneity in hydrology and resource supply and – in turn – support complex vegetation mosaics. Despite this there are predictable relationships across wetlands between plant growth strategies and habitat. Tissue nutrient signatures correspond with inherent ecophysiological differences between growth strategies and with the nutritional challenges posed by the habitats to which these growth strategies are best matched. From knowledge of the impacts of perturbation on nutrient supply, tissue nutrient contents therefore offer the potential to forecast changes in vegetation and linked processes caused by eutrophication, acidification, hydrological change or management, or to detect such changes at an early stage before they become difficult to reverse.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References

Collection of some of the samples in this study was undertaken as part of the Functional Assessment of European Wetland Ecosystems (FAEWE) project that was funded under the EC Science and Technology for Environmental Protection (STEP) programme. We are grateful to other project partners, including B. Clement, D. Hogan, A. Hoyer, R. McInnes and K. Murphy, plus the Royal Society for the Protection of Birds, Bec d’Allier Research station, Suffolk Wildlife Trust, English Nature and Scottish Natural Heritage for sharing information and experience on selected sites and for arranging access. Funding of subsequent analysis was provided by a Natural Environment Research Council small grant (GR9/03534). Fieldwork was undertaken with assistance from V. Abernethy, C. Allan, D. Atkinson, J. Millar, A. McMullen, P. Rattray and M. Tebbitt. W. Wilson, N. O’Farrell, R. Terry and C. Morrow provided assistance with laboratory analyses.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  • Abernethy VJ, Willby NJ. 1999. Changes along a disturbance gradient in the density and composition of propagule banks in floodplain aquatic habitats. Plant Ecology 152: 177190.
  • Aerts R. 1997. Nitrogen partitioning between resorption and decomposition pathways: a trade-off between nitrogen use efficiency and litter decomposability. Oikos 80: 603606.
  • Aerts R, Chapin FS. 2000. The Mineral Nutrition of Wild Plants Revisited: a Re-evaluation of Processes and Patterns. Advances in Ecological Research 30: 267.
  • Aerts R, De Caluwe H. 1994. Effects of nitrogen supply on canopy structure and leaf nitrogen distribution in Carex species. Ecology 75: 14821490.
  • Aerts R, De Caluwe H. 1997. Nutritional and plant-mediated controls on leaf litter decomposition of Carex species. Ecology 78: 244260.
  • Aerts R, De Caluwe H, Konings H. 1992. Seasonal allocation of biomass and nitrogen in four Carex species from mesotrophic and eutrophic fens as affected by nitrogen supply. Journal of Ecology 80: 653664.
  • Aerts R, Van Der Peijl MJ. 1993. A simple model to explain the dominance of low-productive perennials in nutrient-poor habitats. Oikos 66: 144147.
  • Aerts R, Verhoeven JTA, Whigham DF. 1999. Plant-mediated controls on nutrient cycling in temperate fens and bogs. Ecology 80: 21702181.
  • Allen MF. 1991. The ecology of soil mycorrhizae. Cambridge, UK: Cambridge University Press.
  • Armstrong J, Armstrong W, Beckett PM. 1992. Phragmites australis: Venturi- and humdity-induced pressure flows enhance rhizome aeration and rhizosphere oxidation. New Phytologist 120: 197207.
  • Auclair AND. 1979. Factors affecting nutrient concentrations in a Scirpus-Equisetum wetland. Ecology 60: 337348.
  • Auclair AND. 1976. Plant standing crop and productivity relations in a Scirpus-Equisetum wetland. Ecology 57: 941952.
  • Baar J, Ter Braak CJF. 1996. Ectomycorrhizal sporocarp occurrence as affected by manipulation of litter and humus layers in Scots pine stands of different age. Applied Soil Ecology 4: 6173.
  • Barry DAJ, Miller MH. 1989. Phosphorus nutritional requirement of maize seedlings for maximum yield. Agronomy Journ 81: 9599.
  • Bascomb CL. 1968. Distribution of pyrophosphate-extractable iron and organic carbon in soils of various groups. Journal of Soil Science 19: 251268.
  • Bedford BL, Walbridge MR, Aldous A. 1999. Patterns in nutrient availability and plant diversity of temperate North American wetlands. Ecology 80: 21512169.
  • Berendse F, Aerts R. 1987. Nitrogen-use efficiency: a biologically meaningful definition? Functional Ecology 1: 293296.
  • Boeye D, Verhagen B, Van Haesebroeck V, Verheyen RF. 1997. Nutrient limitation in species-rich lowland fens. Journal of Vegetation Science 8: 415424.
  • Bolt GH, Bruggewert MGM. 1976. Soil chemistry, A. Basic elements. Amsterdam, The Netherlands: Elsevier.
  • Bootsma MC, Wassen MJ. 1996. Environmental conditions and fen vegetation in three lowland mires. Vegetatio 127: 173189.
  • Bowman WD. 1994. Accumulation and use of nitrogen and phosphorus following fertilization in two alpine tundra communities. Oikos 70: 261270.
  • Boyd CE. 1978. Chemical composition of wetland plants. In: GoodRE, WhighamDF, SimpsonRL, eds. Freshwater wetlands: ecological processes and management potential. New York, USA: Academic Press, 155168.
  • Boyer MLH, Wheeler BD. 1989. Vegetation patterns in spring-fed calcareous fens: calcite precipitation and constraints on fertility. Journal of Ecology 77: 597609.
  • Ter Braak CJF, šmilauer P. 1998. canoco Reference Manual and User’s Guide to Canoco for Windows: Software for Canonical Community Ordination. Version 4.0. Ithaca, NY, USA: Microcomputer Power.
  • Braakhekke WG, Hooftman DAP. 1999. The resource balance hypothesis of plant species diversity in grassland. Journal of Vegetation Science 10: 187200.
  • Bremner JM, Mulvaney CS. 1982. Nitrogen-total. In: PageAL, ed. Methods of soil analysis II chemical and microbiological properties. Madison, WI, USA: American Society of Agronomy, 596624.
  • Bridgham SD, Pastor J, McClaugherty CA, Richadson CJ. 1995. Nutrient-use efficiency – a litterfall index, a model, and a test along a nutrient availability gradient in North Carolina peatlands. American Naturalist 145: 121.
  • Chapin FS. 1980. The mineral nutrition of wild plants. Annual Review of Ecology and Systematics 11: 233260.
  • Chapin FS, Blom AJ, Field CB, Waring RH. 1987. Plant responses to multiple environmental factors. Bioscience 37: 4957.
  • Chapin FS, Follet JM, O’Connor KF. 1982. Growth, phosphate absorption and phosphate chemical fractions in two Chinochloa species. Journal of Ecology 70: 305321.
  • Christensen KK, Wigand C. 1998. Formation of root plaques and their influence on tissue phosphorus content in Lobelia dortmanna. Aquatic Botany 61: 111122.
  • Crick JC, Grime JP. 1987. Morphological plasticity and mineral nutrient capture in 2 herbaceous species of contrasted ecology. New Phytologist 107: 403414.
  • Daoust RJ, Childers DL. 1999. Controls on emergent macrophyte composition, abundance, and productivity in freshwater everglades wetland communities. Wetlands 19: 262275.
  • Davy AJ, Figueroa ME. 1993. The colonisation of strandlines. In: MilesJ, WaltonDWH, eds. Primary succession on land. Oxford, UK: Blackwell, 113132.
  • De Wit CT, Dijkshoorn W, Noggle JG. 1963. Ionic balance and growth of plants. Wageningen, The Netherlands: Verslagen Van Landbouwjundig Onderzoeken, 69.15.
  • DeLaune RD, Reddy CN, Patrick WHJ. 1981. Organic matter decomposition in soil as influenced by pH and redox conditions. Soil Biological Biochemistry 13: 533534.
  • Devilliers P, Devillers-Terchuren J, Ledant JP. 1991. Habitats of the European Community (CORINE Biotopes Manual: Data Specifications, Part 2). Luxembourg: Office for Official Publications of the European Communities.
  • Eckstein RL, Karlsson PS, Weih M. 1999. Leaf life span and nutrient resorption as determinants of plant nutrient conservation in temperate-arctic regions. New Phytologist 143: 177190.
  • Eckstein RL, Karlsson PS. 1997. Above-ground growth and nutrient use by plants in a subarctic environment: effects of habitat, life form and species. Oikos 79: 311324.
  • Elberse WTH, Berendse F. 1993. A comparative study of the growth and morphology of eight grass species from habitats with different nutrient availabilities. Functional Ecology 7: 223229.
  • Engelaar WMHG, Symens JC, Laanbroek HJ, Blom CWPM. 1995. Preservation of nitrifying capacity and nitrate availability in waterlogged soils by radial oxygen loss from roots of wetland plants. Biology and Fertility of Soils 20: 243248.
  • Ernst WHO. 1990. Ecophysiology of plants in waterlogged and flooded environments. Aquatic Botany 38: 7390.
  • Evans JR. 1989. Photosynthesis and nitrogen relationships in leaves of C3 plants. Oecologia 78: 919.
  • Fojt W, Harding M. 1995. Thirty years of change in the vegetation communities of three valley mires in Suffolk, England. Journal of Applied Ecology 32: 561577.
  • Gore AJP, Urquhart C. 1968. The effects of waterlogging on the growth of Molinia caerulea and Eriophorum vaginatum. Journal of Ecology 54: 617633.
  • De Graaf MCC, Bobbink R, Roelofs JGM, Verbeek DJM. 1998. Differential effects of ammonium and nitrate on three heathland species. Plant Ecology 135: 185196.
  • De Graaf MCC, Bobbink R, Verbeek PJM, Roelofs JGM. 1997. Aluminium toxicity and tolerance in three heathland species. Water Air and Soil Pollution 98: 229239.
  • Green MS, Etherington JR. 1977. Oxidation of ferrous iron by rice (Oryza sativa L.) roots: a mechanism for waterlogging tolerance? Journal of Experimental Botany 28: 678690.
  • Grieve IC, Gilvear DJ, Bryant RG. 1995. Hydrochemical and water source variations across a floodplain mire, Insh Marshes, Scotland. Hydrological Processes 9: 99110.
  • Grime JP. 1979. Plant strategies and vegetation processes. Chichester, UK: John Wiley.
  • Grime JP, Cornelissen JHC, Thompson K, Hodgson JG. 1996. Evidence of a causal connection between anti-herbivore defence and the decomposition rate of leaves. Oikos 77: 489494.
  • Grime JP, Hodgson JG, Hunt R. 1988. Comparative plant ecology: a functional approach to common British plants. London, UK: Unwin-Hyman.
  • Grime JP, Thompson K, Hunt R, Hodgson JG, Cornelissen JHC, Rorison IH, Hendry GAF, Ashenden TW, Askew AP, Band SR, Booth RE, Bossard CC, Campbell BD, Cooper JEL, Davison AW, Gupta PL, Hall W, Hand DW, Hannah MA, Hillier SH, Hodkinson DJ, Jalili A, Liu Z, Mackey JML, Matlews N, Mowforth MA, Neal AM, Reader RJ, Reiling K, Rossfraser W, Spencer RE, Sutton F, Tasker DE, Thorpe PC, Whitehouse J. 1997. Integrated screening validates primary axes of specialisation in plants. Oikos 79: 259281.
  • Harding M. 1993. Redgrave and Lopham Fens, East Anglia, England: a case study of change in flora and fauna due to groundwater abstraction. Biological Conservation 66: 3545.
  • Hayati AA, Proctor MCF. 1990. Plant-distribution in relation to mineral nutrient availability and uptake on a wet-heath site in south-west England. Journal of Ecology 78: 134151.
  • Hayati AA, Proctor MCF. 1991. Limiting nutrients in acid-mire vegetation – peat and plant analyses and experiments on plant-responses to added nutrients. Journal of Ecology 79: 7595.
  • Jones R. 1975. Comparative studies of plant growth and distribution in relation to waterlogging. VIII. The uptake of phosphorus by dune and dune slack plants. Journal of Ecology 63: 109116.
  • Keddy PA. 1992. A pragmatic approach to functional ecology. Functional Ecology 6: 621626.
  • Keddy PA. 2000. Wetland ecology: principles and conservation. Cambridge, UK: Cambridge University Press.
  • Kielland K. 1994. Amino acid absorption by arctic plants: implications for plant nutrition and nitrogen cycling. Ecology 75: 23732382.
  • Klopatek JM. 1978. Nutrient Dynamics of Freshwater Riverine Marshes and the Role of Emergent Macrophytes. In: GoodRE, WhighamDF, SimpsonRL, eds. Freshwater wetlands: ecological processes and management potential. New York, USA: Academic Press, 195216.
  • Koerselman W, Bakker SA, Blom M. 1990. Nitrogen, phosphorus and potassium budgets for two small fens surrounded by heavily fertilised pastures. Journal of Ecology 78: 428442.
  • Koerselman W, Meuleman AFM. 1996. The vegetation N: P ratio: a new tool to detect the nature of nutrient limitation. Journal of Applied Ecology 33: 14411450.
  • Koerselman W, Verhoeven JTA. 1995. Eutrophication of Fen Ecosystems: External and Internal Nutrient Sources and Restoration Strategies. In: WheelerBD, ShawSC, FojtWJ, RobertsonRA, eds. Restoration of temperate wetlands. Chichester, UK: John Wiley, 91112.
  • Kooijman AM, Dopheide JCR, Sevink J, Takken I, Verstraten JM. 1998. Nutrient limitations and their implications on the effects of atmospheric deposition in coastal dunes; lime-poor and lime-rich sites in the Netherlands. Journal of Ecology 86: 511526.
  • Laan P, Smolders Blom CWPM, Armstrong W. 1989. The relative roles of internal aeration, radial oxygen loss, iron exclusion and nutrient balances in flood tolerance of Rumex species. Acta Botanica Neerlandica 38: 131145.
  • Lambers H, Poorter H. 1992. Inherent variation in growth rate between higher plants: a search for physiological causes and ecological consequences. Advances in Ecological Research 22: 187261.
  • Legendre P, Legendre L. 1998. Numerical ecology, 2nd edn. Amsterdam, The Netherlands: Elsevier Science.
  • Lepš K. 1999. Nutrient status, disturbance and competition: an experimental test of relationships in a wet meadow. Journal of Vegetation Science 10: 219230.
  • Malmer N. 1962. Studies on mire vegetation in the Archaean area of south western Gotaland (South Sweden). II. Distribution and seasonal variation in elementary constituents on some mire sites. Opera Botanica 7: 167.
  • Mamolos AP, Veresoglou DS, Barbayiannis N. 1995. Plant species abundance and tissue concentrations of limiting nutrients in low-nutrient grassland: a test of competition theory. Journal of Ecology 83: 485495.
  • Marschner H. 1995. Mineral Nutrition of Higher Plants. London, UK: Academic Press.
  • Mason CG, Bryant RJ. 1975. Production, nutrient content and decomposition of Phragmites communis Trin. & Typha angustifolia. Journal of Ecology 63: 7195.
  • McJannet CL, Keddy PA, Pick FR. 1995. Nitrogen and phosphorus tissue concentrations in 41 wetland plants: a comparison across habitats and functional groups. Functional Ecology 9: 231238.
  • Menges ES, Waller DM. 1983. Plant strategies in relation to elevation and light in floodplain herbs. American Naturalist 122: 454473.
  • Niemann GJ, Pureveen JBM, Eijkel GBH, Poorter H, Boon JJ. 1992. Differences in relative growth rate of 11 grasses correlate with differences in chemical composition as determined by pyrolysis mass spectrometry. Oecologia 89: 567573.
  • Perez Corona MEP, Van Der Klundert I, Verhoeven JTA. 1996. Availability of organic and inorganic phosphorus compounds as phosphorus sources for Carex species. New Phytologist 133: 225231.
  • Polisini JM, Boyd CE. 1972. Relationships between cell wall fractions, nitrogen and standing crop in aquatic macrophytes. Ecology 53: 484488.
  • Poorter H, Bergkotte M. 1992. Chemical composition of 24 wild species differing in relative growth rate. Plant, Cell & Environment 15: 221229.
  • Poorter H, Remkes C, Lambers H. 1990. Carbon and nitrogen economy of 24 wild species differing in relative growth rate. Plant Physiology 94: 621627.
  • Raab TK, Lipson DA, Monson RK. 1999. Soil amino acid utilization among species of the Cyperaceae: Plant and soil processes. Ecology 80: 24082419.
  • Reich PB, Walters MB, Ellsworth DS. 1992. Leaf life span in relation to leaf, plant, and stand characteristics among diverse ecosystems. Ecological Monographs 62: 365392.
  • Richardson CJ, Ferrell GM, Vaithiyanathan P. 1999. Nutrient effects on stand structure, resorption efficiency, and secondary compounds in Everglades sawgrass. Ecology 80: 21822192.
  • Rodwell JS, ed. 1992. British plant communities, 3 grasslands and montane communities. Cambridge, UK: Cambridge University Press.
  • Rodwell JS, ed. 2000. British plant communities, 5 maritime communities and vegetation of open habitats. Cambridge, UK: Cambridge University Press.
  • Roem WJ, Berendse F. 2000. Soil acidity and nutrient supply ratio as possible factors determining changes in plant species diversity in grassland and heathland communities. Biological Conservation 92: 151161.
  • Shaver GR, Johnson LC, Cades DH, Murray G, Laundre JA, Rastetter EB, Nadelhoffer KJ, Giblin AE. 1998. Biomass and CO2 flux in wet sedge tundras: Responses to nutrients, temperature, and light. Ecological Monographs 68: 7597.
  • Shuman LM. 1991. Chemical Forms of Micronutrients in Soils. Micronutrients in Agriculture, 2nd edn. Madison, WI, USA: Soil Science Society of America, 113144.
  • Smith CS, Adams MS, Gustafson TD. 1988. The importance of below ground mineral stores in cattails (Typha latifolia L.). Aquatic Botany 30: 343352.
  • Snowdon RED, Wheeler BD. 1993. Iron toxicity to fen plants. Journal of Ecology 81: 3546.
  • Snowdon RED, Wheeler BD. 1995. Chemical changes in selected wetland plant species with increasing Fe supply, with specific reference to root precipitates and Fe tolerance. New Phytologist 131: 503520.
  • Tannaka A, Loe R, Navasero SA. 1966. Some mechanisms involved in the development of iron toxicity symptoms in the rice plant. Soil Science and Plant Nutrition 12: 158164.
  • Thompson K, Parkinson JA, Band SR, Spencer RE. 1997. A comparative study of leaf nutrient concentrations in a regional herbaceous flora. New Phytologist 136: 67796689.
  • Tilman D. 1982. Resource competition and community structure. Princeton, NJ, USA: Princeton University Press.
  • Tilman D. 1990. Mechanisms of plant competition for nutrients: the elements of predictive theory of competition. In: GraceJ, TilmanD, eds. Perspectives on plant competition. New York, USA: Academic Press, 117141.
  • Van Duren IC, Pegtel DM. 2000. Nutrient limitations in wet, drained and rewetted fen meadows: evaluation of methods and results. Plant and Soil 220: 3547.
  • Van Oorschot MMP, Robemont E, Boerstal M, Van Strien I, Van Kerkhoven-Scmitz M. 1997. Effects of enhanced nutrient availability on plant and soil nutrient dynamics in two English riverine ecosystems. Journal of Ecology 85: 167180.
  • Veerkamp MT, Kuiper PJC. 1982. The uptake of potassium by Carex species from swamp habitats varying from oligotrophic to eutrophic. Physiologia Plantarum 55: 237241.
  • Verhoeven JTA, Arts HMM. 1987. Nutrient dynamics for small mesotrophic fens surrounded by cultivated lands. II. N and P accumulation in plant biomass in relation to the release of inorganic N and P in the peat soil. Oecologia (Berlin) 72: 557561.
  • Verhoeven JTA, Koerselman W, Beltman B. 1988. The vegetation of fens in relation to their hydrology and nutrient dynamics; a case study. In: SymoensJJ, ed. Handbook of vegetation science, Vol. 15, vegetation of inland waters. Dordrecht, The Netherlands: Kluwer, 249282.
  • Vitousek PM, Walker LR. 1987. Colonisation, succession and resource availability: ecosystem–level interactions. In: GrayAJ, CrawleyMJ, EdwardsJP, eds. Colonization, succession and stability. Oxford, UK: Blackwell, 201224.
  • Wassen MJ, Barendregt A, Palczynski A, Desmidt JT, DeMars H. 1990. The relationship between fen vegetation gradients, groundwater-flow and flooding in an undrained valley mire at Biebrza, Poland. Journal of Ecology 78: 11061122.
  • Wassen MJ, Olede Venterink HGM, De Swart EOAM. 1995. Nutrient concentrations in mire vegetation as a measure of nutrient limitation in mire ecosystems. Journal of Vegetation Science 6: 516.
  • Wassen MJ, Van Der Vliet RE, Verhoeven JTA. 1998. Nutrient limitation in the Biebrza fens and floodplain (Poland). Acta Botanica Neerlandica 47: 241252.
  • Waughman GJ. 1980. Chemical aspects of the ecology of some South German peatlands. Journal of Ecology 68: 10251046.
  • Wheeler BD. 1999. Water and plants in freshwater wetlands. In: BairdAJ, WilbyR, eds. Eco-hydrology: plants and water in terrestrial and aquatic environments. London, UK: Routledge, 127180.
  • Wheeler BD, Al-Farraj MM, Cook RED. 1985. Iron-toxicity to plants in base-rich wetlands: comparative effects on the distribution and growth of Epilobium hirsutum L. & Juncus subnodulosus Schrank. New Phytologist 100: 653669.
  • Wheeler BD, Shaw SC. 1990. Dereliction and eutrophication in calcareous seepage fens. In: HillierSH, WaltonDWH, WellsDA, eds. Calcareous grasslands – ecology and management. Bluntisham, UK: Bluntisham Books, 154160.
  • Wheeler BD, Shaw SC, Cook RED. 1992. Phytometric assessment of the fertility of undrained rich-fen soils. Journal of Applied Ecology 29: 466475.
  • Willby NJ, Murphy KJ, Gilvear DJ, Grieve ID, Pulford ID. 1997. Hydrochemical–vegetation interactions on a Scottish floodplain mire. In: LargeARG, ed. Floodplain rivers: hydrological processes and ecological significance. Newcastle-upon-Tyne, UK: British Hydrological Society. 4052.
  • Woodwell GM, Whittaker RH, Houghton RA. 1975. Nutrient concentrations in plants in the Brookhaven oak-pine forest. Ecology 56: 318332.