Present address: Atelier Sternen and Partner, Tobeleggweg 19, CH-8049 Zürich, Switzerland.
Contrasting effects of nitrogen, phosphorus and water regime on first- and second-year growth of 16 wetland plant species
Article first published online: 12 DEC 2003
Volume 17, Issue 6, pages 754–765, December 2003
How to Cite
Güsewell, S., Bollens, U., Ryser, P. and Klötzli, F. (2003), Contrasting effects of nitrogen, phosphorus and water regime on first- and second-year growth of 16 wetland plant species. Functional Ecology, 17: 754–765. doi: 10.1111/j.1365-2435.2003.00784.x
- Issue published online: 12 DEC 2003
- Article first published online: 12 DEC 2003
- Received 10 January 2003; revised 5 July 2003; accepted 7 July 2003
- Biomass allocation;
- growth experiment;
- limiting nutrient;
- tissue nutrient concentration;
- water level
- Top of page
- Materials and methods
- 1The effects of nutrient enrichment on wetland vegetation may depend on the responses of different plant species to nutrient supply over several years, in waterlogged or flooded soils, and under either nitrogen- or phosphorus-limited conditions. However, most growth experiments comparing species from differently productive sites have focused on their short-term responses to variation in N supply. In this study we investigated whether increased N or P supply affects plant growth differently, whether these effects differ between the first and second year of growth, and whether they are modified by the water regime.
- 2Plants of 16 wetland species were grown during two seasons in tubes with sand under full light. Treatments combined three nutrient levels (low N and P, high N, high P) with three water regimes (constantly wet, periodically aerated, periodically flooded).
- 3In the first year, shoot biomass was enhanced by high N supply, particularly in species from nutrient-rich sites; this was associated with reduced shoot P concentration. In the second year, shoot biomass was generally enhanced by high P supply and reduced by high N supply; responses to high P were strongest in species with low shoot biomass and high N concentration but unrelated to the productivity of the species’ sites.
- 4The total biomass produced during both years was smaller at high N supply than at high P supply. A smaller fraction of the N and P supply was recovered in high-N plants, and these plants allocated less biomass to roots than those grown at high P supply or low N and P supply.
- 5Periodic flooding reduced biomass production and nutrient recovery, but hardly influenced the effects of nutrient supply on plant growth. Species from wet meadows were affected more by flooding than species from fens in the first season, but not in the second season.
- 6We propose that high N supply reduced second-year growth because strong P limitation increased below-ground nutrient losses from plants, whereas high P supply enhanced second-year growth by improving N retention in plants. Our results therefore suggest that N and P enrichment may have quite different effects on wetland vegetation.
- Top of page
- Materials and methods
Nutrient enrichment has modified the plant species composition of natural and seminatural vegetation throughout the Northern Hemisphere (Morris 1991; Bobbink, Hornung &Roelofs 1998). These modifications are due largely to changes in competitive interactions after nutrient enrichment (Bobbink et al. 1998; Aerts 1999). To understand the underlying mechanisms, numerous studies have compared the adaptations of species from nutrient-rich and nutrient-poor sites, showing e.g. differences in relative growth rate (Elberse & Berendse 1993; Schippers & Olff 2000), nutrient use (Aerts & Chapin 2000) or tissue structure (Ryser & Wahl 2000). Species from nutrient-rich sites also appeared more plastic and more responsive to variation in nutrient supply than species from nutrient-poor sites (Fichtner & Schulze 1992; Elberse & Berendse 1993; Garnier 1998).
Growth experiments comparing species responses to nutrient supply have mostly varied the supply of N while the supply of the other nutrients remained constant and non-limiting (e.g. Fichtner & Schulze 1992; van der Werf et al. 1993; Aerts & de Caluwe 1994). Alternatively, different amounts of the same nutrient solution with a N : P ratio less than 10 (mass : mass) were used (Olff 1992; Elberse & Berendse 1993). In both cases plant growth was N-limited, and whenever plant species responded differently to the treatments, this primarily reflected their responsiveness to N supply. Fewer experiments compared the growth responses of plants to P supply or to combined variations in N and P supply (e.g. Shaver & Melillo 1984; Ulrich & Burton 1985). The plant traits associated with responses to P appeared similar to those associated with responses to N (Ryser, Verduyn & Lambers 1997; Perez-Corona & Verhoeven 1999), yet growth was generally less influenced by variation in P than in N supply (Güsewell & Koerselman 2002). Keddy et al. (2001) found that the responses of 21 wetland species to a reduction in P supply were unrelated to their maximal growth rates, whereas a reduction in N supply affected the faster-growing species more than the slower-growing ones. This suggests that the responses to P supply, unlike responses to N supply, do not differ between species from nutrient-rich and those from nutrient-poor sites. The same was suggested by the results of Perez-Corona & Verhoeven (1996), but has not yet been tested directly.
Most comparative growth experiments have lasted for only a few weeks or months (Garnier & Freijsen 1994). Longer-term experiments showed that interspecific differences in growth and in responsiveness to nutrient supply may weaken, disappear, or even reverse with time, particularly when nutrient supply is generally limiting (Elberse & Berendse 1993; Ryser 1996). It is not clear whether the responses to N and P supply change differently with time. Field experiments at sites with low N and P availability suggested that the effects of increased P supply appear more slowly and are more persistent than those of increased N supply (Güsewell, Koerselman & Verhoeven 2002).
Plant responses to nutrient supply may also be influenced by the water regime. Drought or flooding may reduce plants’ nutrient uptake and/or their ability to use nutrients for growth (Schat 1984). As a result, the effects of resource supply and of water regime on growth may interact: drought or flooding may strengthen or weaken growth responses to nutrient supply (Neill 1990). As plant species are differently drought- or flooding-tolerant, these interactions may, in turn, differ among plant species (Schat 1984). In particular, drought or flooding may reduce the ability of fast-growing species to benefit from increased nutrient supply and reduce shifts in species composition caused by nutrient enrichment (Bollens, Güsewell & Klötzli 2001).
All these variations in the effects of nutrient supply are likely to be relevant for understanding and predicting the effects of nutrient enrichment on wetland vegetation, which is dominated by perennial plants, where plant growth and species composition depend strongly on the water level (Lenssen, ten Dolle & Blom 1998; Lenssen et al. 1999), and where biomass production can be limited by N or by P (Olde Venterink, van der Vliet & Wassen 2001; Güsewell & Koerselman 2002). Consequently, the responses of wetland species to N supply in typical short-term growth experiments with good aeration of the rooting zone may bear little relevance to the actual effects of nutrient enrichment under field conditions, unless the interspecific differences revealed by such experiments are independent of experimental duration, water regime and type of nutrient limitation.
In this experiment we cultivated plants of 16 wetland species for two seasons at all combinations of three nutrient treatments (low N and P, high N, high P) and three water regimes (constantly wet, periodically moist, periodically flooded) to test whether the effects of nutrient supply on species from differently productive wetland sites depend on the above factors. We tested the following hypotheses: (1) plant growth responds differently to increased N supply and to increased P supply; (2) the effects of increased N or P supply differ between the first and second seasons of growth; (3) the effects of increased N or P supply depend on the water regime; species from nutrient-rich and nutrient-poor sites respond differently, and these differences depend on (4) the experimental duration as well as (5) the type of nutrient limitation.
Materials and methods
- Top of page
- Materials and methods
The 16 plant species are common herbaceous perennials in Swiss wetlands, where they occur at sites with contrasting nutrient availability and water level (Table 1). Seeds of 30 plant species were collected in the field in summer 1996 or 1997 (D. Ramseier, personal communication) and germinated in March 1998 on wet substrate under a fluctuating light and temperature regime. Some species did not germinate, others only poorly. After germination, seedlings were transplanted to multipot trays with nutrient-poor potting soil, and cultivated first in the greenhouse and later outdoors until a sufficient number of plants from 18 species had reached a height of several centimetres. At the beginning of June 1998, seedlings were transplanted to polyethylene tubes of 6 cm diameter and 30 cm height, which were heat-sealed at the bottom and filled with quartz sand to a height of 25 cm. The tubes were placed into 10 l buckets filled with deionized water. Holes punctured in the bottom of the tubes enabled the water level inside tubes to equilibrate with that in the buckets.
|Reeds and tall-sedge fens|
|Nutrient-rich wet meadows|
|Nutrient-poor wet meadows|
experimental design and treatments
The experiment was carried out in the experimental garden of the Geobotanical Institute, ETH Zürich. The mean annual temperature was 9·6 °C in 1998 and 9·5 °C in 1999; total annual precipitation was 1044 mm in 1998 and 1549 mm in 1999 (Meteorological Station Zürich-SMA).
The treatments combined three nutrient levels with three water levels in a factorial (nutrient × water) design. The resulting nine treatments were replicated six times, with replicates arranged in blocks. Each combination of treatment and block consisted of a group of three buckets, and these 54 groups represented the main experimental units. Eighteen tubes (one per species) were randomly distributed among the three buckets of each group (six tubes per bucket). These tubes were not independent of each other, and were therefore treated as split units within the groups of buckets. Most plants of two species died during the first few weeks of the experiment, so that only 16 species remained.
Nutrients were supplied in the form of a modified Hoagland's nutrient solution (Table 2). Each plant received 2·5 ml of the appropriate solution weekly during 15 weeks in 1998 (June–September) and during 14 weeks in 1999 (March–June). With the low-nutrient treatment, one plant received 0·9 mg N and 0·06 mg P per week; the high-N treatment provided a threefold nitrogen supply (2·7 mg N, 0·06 mg P per week), and the high-P treatment provided a threefold phosphorus supply (0·9 mg N, 0·18 mg P per week). These treatments supplied N and P in mass ratios of 15 (low N and P), 45 (high N) and 5 (high P), respectively. All other nutrients were supplied in non-limiting amounts; the supply of micronutrients was doubled in the second year to exclude any possible deficiency.
|Substance||Concentration (µmol l−1) in:|
|Control||High N||High P|
|KNO3||15 000||45 000||15 000|
|Ca(NO3)2·4H2O||5 000||15 000||5 000|
|KCl||30 000||0||28 486|
|CaCl2·2H2O||10 000||0||10 000|
|Year 1||Year 2|
|C4H6O6 (tartaric acid)||20||40|
The three water treatments (wet, flooded, moist) were applied during the same periods as nutrient treatments. In the constantly wet (W) tubes, the water level was constantly 3 cm below the soil surface. In periodically flooded (F) tubes, the water level was raised to 3 cm above the soil surface during three periods of 2–3 weeks, separated by 1 or 2 weeks of W conditions. In the periodically aerated (A) tubes, the water level was lowered to 10 cm below the soil surface in the first year, and to 23 cm below the surface in the second year during the same periods as for the F tubes. The water level was lowered more in the second year because plants had longer roots. Between the two growth periods, all plants were first kept under wet conditions (September–October), after which the water level was lowered to 2 cm in all tubes to prevent frost damage (November–March). Water treatments were re-established when new shoots started to grow (March 1999). The water level within each bucket was regulated by holes of 5 mm diameter drilled into the side wall of buckets at three heights (2, 15 and 22 cm), which were closed with water-resistant adhesive tape or opened, as required. Evaporated water was replaced weekly with deionized water during dry periods.
Plants were harvested after 3–4 months of growth in each year, on 1 September in the first year and on 22 June in the second year. At the first harvest shoots were clipped 1 cm above ground. Senesced leaves were removed because leaves dying during the growth period were mostly blown away. At the second harvest, the sand was washed from roots with water, and plants were divided into shoots and below-ground biomass (roots + rhizomes from first and second year). All samples were dried at 70 °C until constant mass and weighed. The six replicates per species and treatment were then pooled, ground, and analysed for total N and P content: 1 g dry material was digested for 1 h at 420 °C in 15 ml 98% H2SO4 with a copper sulphate–titane oxide catalyst; concentrations of N and P in digests were determined colorimetrically on a flow injection analyser (Tecator, Höganäs, Sweden).
The dry mass of the three plant fractions (shoots year 1, shoots year 2, below-ground), their sum (total) and the ratio of below-ground to total biomass were log-transformed for data analysis and back-transformed for graphical representation. After checking the absence of block × treatment interactions, the effects of treatments and species were analysed with split-plot anova with block, nutrient and water as main-plot factors and species as subplot factor (all fixed). Species × nutrient interactions were generally significant; data were therefore analysed again for each species separately (three-way anova with fixed factors block, nutrient and water). Because the nutrient × water interaction was always minor compared with main effects, means of the three nutrient levels and means of the three water levels (= main effects) were compared pairwise using the Tukey's HSD test.
Nutrient concentrations were log-transformed and analysed with three-way anova with the main-plot factors nutrient and water and the split-plot factor species. There was no block factor because the six replicates had been pooled. The effects of nutrient and water were tested against the nutrient × water interaction (= main-plot residual), and means were compared using Tukey's HSD test. Nutrient pools were calculated by multiplying the N and P concentrations of each plant fraction by its mean dry mass, then adding together the resulting values for each species and treatment. Nutrient pools represented the mean uptake of nutrients by a plant minus their losses through litter or leaching. Percentage nutrient recovery was the N or P pool divided by the amount of N or P supplied over both years. The effects of nutrient, water and species on nutrient recovery were analysed as described for nutrient concentrations; no data transformation was required.
Growth responses to high N or high P supply were quantified as the mean log-transformed shoot biomass of a species at high N or high P supply minus that at low nutrient supply (across water levels and replicates, for each year). Pearson correlation coefficients were used to analyse pairwise correlations between plant traits, growth responses, and Ellenberg's nutrient indicator values (integers from 1 to 9 describing the nutrient status of the species’ habitat; Ellenberg et al. 1991).
Data from some species and biomass fractions were missing: shoots of Primula farinosa were not harvested in the first year because they were almost entirely shorter than 1 cm; Cirsium palustre and Mentha aquatica incurred high mortality during the winter, with only 20% of the plants surviving (no data for year 2); roots of Lychnis flos-cuculi had largely decayed before the second harvest (no data for below-ground and total biomass). As a result, between 12 and 15 species were included in individual tests. Full data (biomass and nutrient concentrations for each treatment and species) are available from the corresponding author on request.
- Top of page
- Materials and methods
treatment effects on biomass and allocation
Plant biomass (mean of all species) was influenced significantly by nutrient supply and water regime (Table 3a). Nutrient supply had a different effect on each of the three plant fractions: Shoot biomass in year 1 was increased by high N supply, whereas that in year 2 was increased by high P supply and slightly reduced by high N supply. Below-ground biomass increased in the order high N < low N and P < high P, and the total amount of biomass produced by the plants during both years was also greatest with high P supply (Table 3b). Due to the changing effects of nutrient supply, shoot biomass hardly differed between the first and second years at low nutrient supply, whereas it decreased at high N and increased at high P supply (Table 3b). Water regime had a similar effect on all plant fractions: biomass was reduced by periodic flooding and either enhanced or unaffected by periodic aeration (Table 3b). The effects of nutrient supply and water regime generally did not interact; the only exception was total biomass, which was reduced slightly by periodic aeration at low nutrient supply, but enhanced at high N or high P supply (Table 3a). The allocation of biomass to below-ground parts was reduced by high N supply but was independent of the water regime (Table 3).
|df||Shoot year 1||Shoot year 2||Below-ground biomass||Total biomass||% Below ground|
|(a) Tests of effects|
|(b) Treatment means||(mg)||(mg)||(mg)||(mg)||(%)|
|Low nutrient supply||234A||229A||824B||1490A||58·3B|
|High N supply||397B||191A||720A||1510A||49·0A|
|High P supply||241A||329B||1027C||1832B||59·4B|
|SE of means‡||6–10||12–21||23–34||31–43||0·6–0·8|
Significant species-by-treatment interactions (Table 3a) indicate that treatments affected the individual species differently. However, interspecific differences mainly concerned the strength of these effects, some species responding more than others, while the type of effects was essentially the same. The responses of individual species to nutrient and water treatments are summarized in Table 4. In the first year, the shoot biomass was enhanced by high N in all species but two (L. flos-cuculi and Succisa pratensis); unaffected by high P; either reduced or unaffected by flooding; and either unaffected or slightly enhanced by aeration. Seven of the eight species that were significantly reduced by flooding in the first year were those from the drier wetland types (wet meadows; Table 4). In the second year, all species but one (Carex flava) responded either positively to high P, negatively to high N, or both together. Negative effects of flooding were found for some species, but these were not species from drier wetland types (Table 4).
|Species code||Nutrient effect||Water effect|
|Year 1||Year 2||Year 1||Year 2|
|Ma||+N||na||A > F||na|
|Cf||+N||N > P||–||–|
|Ao||+N||+P||A > F||–|
|Fu||+N||+P||A > F||A > F|
|Lf||–||+P–N||A > F||+A–F|
|Ca||+N||–||–||A > F|
|Sc||+N||–||A > F||–|
nutrient concentrations and recovery of supplied nutrients
Nutrient concentrations in all plant fractions differed among nutrient treatments (Table 5a). The concentration of N was increased by high N supply, whereas the P concentration was increased by high P and reduced by high N supply (Table 5b). These effects were generally more pronounced in the second year than in the first; only the effect of high N supply on P concentration was stronger in the first year. Periodic flooding slightly increased the P concentration of shoots in the first year and the N concentration of shoots in the second year (Table 5b). Interactions between the effects of nutrient treatments and those of water regime or species on nutrient concentrations were minor compared to the main effects (Table 5a).
|N concentration||P concentration|
|Shoot year 1||Shoot year 2||Below ground||Shoot year 1||Shoot year 2||Below ground|
|(a) Tests of effects|
|Species × nutrient||2·2**||8·1***||4·0***||4·4***||4·5***||9·7***|
|Species × water||1·7*||1·5||2·2*||3·7***||2·5**||1·7°|
|(b) Treatment means (mg g−1)|
|Low N and P||11·3A||11·5A||7·3B||0·95B||0·83B||0·62B|
|SE of means†||0·1–0·1||0·1–0·2||0·2–0·3||0·02–0·04||0·02–0·07||0·02–0·05|
Half of the N and two-thirds of the P supplied with nutrient solutions was recovered in total plant biomass in the low-nutrient treatment (Fig. 1). Nutrient recovery was significantly affected by nutrient supply (Table 6). High supply of either N or P increased the amount of each nutrient taken up, but reduced the percentage recovered (Fig. 1a,b). High supply of N also reduced the recovery of P, whereas high supply of P increased the recovery of N (Fig. 1a,b). The recovery of both nutrients was slightly reduced by periodic flooding. Half the recovered N was found in the root fraction, regardless of nutrient supply or water regime (Table 6; Fig. 1a), whereas a larger proportion of the recovered P was found in roots at high P supply (58%) than at high N supply (47%; Fig. 1b).
|Total||% Below||Total||% Below|
|Species × nutrient||3·0***||5·9***||9·7***||4·8***|
|Species × water||0·9||1·0||1·0||1·7°|
responses of individual species in relation to plant traits
The 16 species differed considerably in biomass and nutrient concentrations (Tables 3a and 5a). The shoot biomass of individual species ranged from 30 to 1000 mg in both years. Interspecific differences were much greater than the effects of nutrient or water treatments, and were hardly influenced by the latter. Indeed, there were strong positive correlations between the mean biomass of the species at low nutrient supply and their mean biomass at high N or high P supply (Fig. 2). Likewise, species means for the three water regimes were highly correlated (r = 0·97–0·99 in year 1; r = 0·93–0·96 in year 2). By contrast, shoot biomass in the first year was only weakly correlated to that in the second year within each nutrient treatment (r = 0·50–0·53), indicating that shoot biomass increased for some species and decreased for others between the first and the second years.
The N concentration of shoots was negatively correlated with shoot biomass in both years (Fig. 3a,b), whereas P concentration was unrelated to shoot biomass (Fig. 3c,d). The concentrations of N and P in roots were negatively correlated with the biomass of roots (r = −0·90 for N; r = −0·81 for P) and with total biomass (r = −0·89 for N; r = −0·77 for P; all correlations for the low-nutrient treatment).
There was no significant correlation between the shoot biomass or the nutrient concentrations therein and the growth responses of the 16 species to high N supply (Table 7). By contrast, the responses to high P supply were negatively correlated with shoot biomass and positively correlated with the N concentrations and N : P ratios of shoots in both years. The responses of the species to high N and high P supply were not correlated with each other in either year, nor were the responses to either nutrient correlated between the two years of the experiment. However, the responses to high N supply in year 1 were negatively correlated with the responses to high P supply in year 2 (Table 7).
|Ellenberg's N value†||Response to high N||Response to high P|
|Year 1‡||Year 2‡||Year 1||Year 2||Year 1||Year 2|
|N conc shoot||−0·39||0·00||0·00||0·22||0·74**||0·68**|
|P conc shoot||0·44°||0·20||0·35||0·26||0·24||−0·01|
|N : P ratio shoot||−0·60*||−0·20||−0·24||0·10||0·41||0·62*|
|Root weight ratio||−0·32||−0·01||0·39||−0·07||0·20|
|Ellenberg N value||0·31||−0·10||−0·36||−0·07|
|Response to high N year 1||0·37||−0·24||−0·68 **|
|Response to high P year 2||−0·06||0·32|
Species from nutrient-rich sites tended to have a higher shoot biomass than those from nutrient-poor sites in the first year (marginally significant correlation with Ellenberg's N value), but not in the second year (Table 7). In the first year, species from nutrient-rich sites also tended to have a lower N concentration and a higher P concentration, so that the shoot N : P ratio was negatively correlated with Ellenberg's N value. These correlations all disappeared in the second year (Table 7). Growth responses to high N or to high P supply were not significantly correlated with Ellenberg's N values in either year, i.e. the species from nutrient-rich wetland types were not clearly more responsive to increased nutrient supply than those from nutrient-poor wetlands in our experiment (Table 7).
- Top of page
- Materials and methods
time-dependent effects of n and p supply
In the first year of this experiment, the growth of all species was enhanced by a high N supply, whereas a high P supply had no significant effect. This confirms our first hypothesis, that growth would respond differently to increased supply of N and P, and concurs with numerous other studies finding growth at low N and P supply to be mainly N-limited (Güsewell & Koerselman 2002). The basis of this response was the ability of all plant species to reduce their P concentration at high N supply. Biomass N concentration is generally less plastic in response to nutrient supply than is P concentration (Aerts & Chapin 2000; Güsewell & Koerselman 2002), which may explain why plants could not similarly increase their growth at high P supply by reducing their N concentration.
In the second year, however, growth was enhanced mainly by an increased supply of P, whereas it was often reduced by a high supply of N. This change in nutrient limitation was reflected by a change in the ratio of N to P concentration in shoot biomass: the mean N : P ratio of plants grown at low nutrient supply was 11·9 in the first year and 14·1 in the second year. The shift in nutrient status indicated by these N : P ratios (cf. Güsewell & Koerselman 2002) probably had two causes: additional nutrient sources, and different nutrient retention by plants. Additional P was supplied by the quartz sand used as substrate. Each tube contained 800 g sand, from which 0·5 mg kg−1 phosphate-P was extracted with ammonium acetate (S.G., unpublished data). The phosphate content of the sand thus represented up to 60% of the P supplied during the first year in the low-nutrient treatment. Atmospheric deposition in Zurich is about 30 kg N ha−1 per year (Hesterberg et al. 1996) and 1 kg P ha−1 per year (Newman 1995). Assuming a homogeneous distribution of these inputs over the year, each tube would have received an additional supply of 1·6 mg N and 0·053 mg P until the first harvest, and of 5·1 mg N and 0·17 mg P between the first and the second harvest. With these additional inputs the N : P supply ratio in the low-nutrient treatment would have been 10·3 until the first harvest and 17·0 until the second harvest. The actual N : P supply ratios probably differed less between the two years, as the mean shoot N : P ratios of the plants (11·9 and 14·1) were more similar (Güsewell & Koerselman 2002). For example, it may be that only a fraction of the P contained in the sand was available in the first year, or that atmospheric N was leached out during the winter.
Different nutrient retention by plants was indicated by the recovery of P after 2 years, which was less in high-N than in low-nutrient plants. It is unlikely that this was due to N toxicity, given the low N concentrations in plant biomass: at 13·6 mg N g−1 in the first year and 17·0 mg N g−1 in the second, the mean N concentration of shoot biomass at high N supply hardly differed from that of wetland plants in the field or in greenhouse experiments where growth was N-limited (Güsewell & Koerselman 2002). Strong P limitation must therefore have been the cause of the reduced second-year growth. Phosphorus limitation was also indicated by extremely low P concentrations: the mean values of 0·66 mg g−1 (shoots) and 0·47 mg g−1 (roots) were lower than in field plants (Güsewell & Koerselman 2002).
Finding a low P recovery in P-limited plants was unexpected, as P-limited plants take up P very efficiently (Aerts & Chapin 2000). It suggests that plants grown at high N supply lost more P between the first and the second harvest than the low-N and P plants. Nutrient losses through leaf litter or seeds were probably negligible, so that the loss of P (and N) must have occurred mainly below ground (Aerts, Bakker & de Caluwe 1992; Vásquez de Aldana, Geerts & Berendse 1996). Although experimental evidence is not consistent, some studies have shown strong P deficiency to enhance root turnover or exudation (Fisher, Eissenstat & Lynch 2002). High N supply may also have enhanced root turnover (Fransen & de Kroon 2001). Because shoots were harvested after the first season, which removed more biomass from the high-N plants than from the low-nutrient plants, the question might arise whether the loss of nutrients caused by this first harvest was not the actual cause of P deficiency in the second year. In fact, P removal differed little between low-nutrient and high-N plants (Fig. 1) because the greater shoot biomass of high-N plants was compensated by a lower P concentration in these shoots. Moreover, the low P recovery of high-N plants could not be a direct effect of harvesting, as any P contained in the harvested shoot biomass was included in recovery calculations.
In conclusion, our second hypothesis (the effects of nutrient supply differ between the first and second season) was not unequivocally supported by the change from N to P limitation between the first and second years, because a change in external nutrient supply may have contributed to this change. However, the negative effects of high N supply, which appeared only in the second year, and were associated with P losses, could not be due to additional nutrient inputs. They clearly show that the effects of N and P supply are differently time-dependent, as already observed in field experiments (Güsewell et al. 2002).
effects of water regime
None of the three water regimes increased mortality among our wetland species, presumably because the first flood occurred after plants were established (Roth et al. 1999). In addition, as the substrate was quartz sand, flooding would not create strongly reducing conditions (Clevering 1998; de Mars & Wassen 1999). On the other hand, periodic aeration (low water level) did not cause any significant water stress, as the sand was always kept moist by capillary rise of water.
Unlike survival, the shoot growth of some species was reduced by flooding. In the first year, species from the drier wetland types were most affected (cf. Lenssen et al. 1998), but in the second year flooding mainly reduced the species responding negatively to high N supply (Table 4). This might be because flooding, like high N supply, reduced the ability of these species to conserve nutrients between the first and second years (Gebauer, Reynolds & Tenhunen 1995). In addition, reduced energy production in flooded roots might have reduced the plants’ ability to take up and transport nutrients (Schat 1984).
The effects of water regime did not interact, or interacted only weakly, with those of nutrient supply. This is at variance with our third hypothesis, but similar to other growth experiments using sand as substrate (Figiel, Collins & Wein 1995). Water regime also had little effect on nutrient concentrations in plant biomass, which indicates that the effects on plant growth were paralleled by effects on nutrient uptake or retention (cf. Schat 1984; DeLaune, Jugsujinda & Reddy 1999). Thus, despite their reduced growth, periodically flooded plants were as nutrient-limited as plants in the two other water regimes. In the field, interactions between nutrient supply and water regime may be mediated by the effects of flooding on symbiotic N fixation (Bordeleau & Prévost 1994); mycorrhizal infection (Miller et al. 1999); or interspecific competition (Grace 1988). None of these mechanisms played a role in the present experiment, so the absence of an interaction is not surprising. The possible role of the aforementioned factors and mechanisms should be assessed in appropriate experiments before concluding definitely that water regime does not modify the effects of nutrient enrichment on species composition in wetlands.
differences between species from nutrient-rich and nutrient-poor sites
In the first year, the shoot biomass of the 16 species was positively correlated with Ellenberg's N indicator value, meaning that species from nutrient-rich sites grew faster despite the generally low supply of nutrients (Elberse & Berendse 1993; Ryser 1996). This relationship disappeared in the second year, confirming our fourth hypothesis, that interspecific differences depend on experimental duration. However, the same plant trait correlated with interspecific differences in shoot biomass in both years: species with larger shoot biomass had a lower N concentration. The latter was probably a consequence, rather than a cause, of the faster growth, at least in the second year when growth was P-limited.
Species from nutrient-rich sites were not significantly more responsive to high N supply than those from nutrient-poor sites: the responses of all species were similar. There was still a weak positive relationship between growth responses to high N and the shoot biomass or Ellenberg's N values of the 16 species. By contrast, the responses to high P supply in the second year, did not even show such a trend, and the responses of the 16 species to high N supply in the first year correlated negatively with responses to high P supply in the second year (r = −0·68). This lends some support to our fifth hypothesis, that differences between species from nutrient-rich and nutrient-poor sites depend on the type of nutrient limitation. More definite results would be needed for confirmation, and would probably be obtained by choosing greater differences between nutrient levels, as differences between fast- and slow-growing species tend to increase with increasing nutrient supply (Bradshaw et al. 1964; Fichtner & Schulze 1992).
Our experiment differed from previous studies (van der Werf et al. 1993; Perez-Corona & Verhoeven 1996) in that high supply of one nutrient (N or P) caused the other nutrient to be strongly limiting for plant growth. Accordingly, increased biomass production at high N supply in the first year entailed a decrease in shoot P concentration; and increased biomass production at high P supply in the second year entailed a decrease in shoot N concentration. Yet the species’ responses to high N supply in the first year did not correlate with their P concentrations, suggesting that a productive use of N and effective foraging for P caused strong growth responses. These traits are found mainly in fast-growing species (van der Werf et al. 1993). By contrast, the responses to high P supply in the second year were positively correlated with the shoot N concentration in the low-nutrient treatment (r = 0·68), suggesting that a previous accumulation of N enabled some species (mainly those with smaller shoot biomass) to benefit from an increased P supply. The negative correlation between growth responses to high N (in year 1) and high P (in year 2) might reflect a trade-off between N productivity in the first year and N conservation from the first to the second year (Eckstein & Karlsson 1997). The fact that shoots were harvested after the first year may have contributed to this trade-off; it is, however, the typical management of the wet meadows where our species occur in Switzerland.
The differences found here correspond to those reported from field fertilization experiments in nutrient-poor wetlands, where N and P fertilization enhanced the biomass of different species, with P fertilization often enhancing the subordinates (Güsewell, Koerselman & Verhoeven 2003). Also, species that increased after N fertilization did not have an above-average P concentration in shoot biomass, whereas species that increased after P fertilization did have an above-average N concentration (Güsewell et al. 2003). The relationships obtained in our growth experiment might therefore be valid more generally.
- Top of page
- Materials and methods
This experiment has shown that increased supply of N or P (at a short supply of the other nutrient) affects the growth and nutrient use of plants differently, and that these effects are time-dependent. We propose that high N supply increased first-year growth because all species reduced their P concentration; fast-growing species tended to benefit most. High N supply reduced second-year growth because strong P limitation increased below-ground nutrient losses from plants. High P supply enhanced total biomass production over 2 years, probably by increasing biomass allocation to roots and thus, improving N acquisition and retention in plants; species with small biomass tended to benefit most. These effects were largely independent of water regime. Because nutrient supply was limiting in all treatments, interspecific differences may have been reduced, but this made the experiment more comparable to field conditions. Many studies have investigated how increased availability of N affects plant functioning and vegetation processes (Bobbink et al. 1998; Garnier 1998; Aerts & Bobbink 1999; Gough et al. 2000). Our results suggest that the effects of increased P availability may be quite different, at both plant and vegetation levels.
- Top of page
- Materials and methods
We thank P.J. Edwards for advice on the experimental design, D. Ramseier for providing plant seeds from his personal collection and for advice on germination procedures, T. Fotsch, A. Risch, R. De Marchi, H.-P. Suter and T. Stüdeli for assistance with the establishment of the experiment, and R. Trachsler for nutrient analyses. Two referees provided constructive comments on an earlier version of the manuscript. Funding was provided by the Swiss National Science foundation.
- Top of page
- Materials and methods
- 1999) Interspecific competition in natural plant communities: mechanisms, trade-offs and plant–soil feedbacks. Journal of Experimental Botany 50, 29–37. (
- 1999) The impact of atmospheric nitrogen deposition on vegetation processes in terrestrial, non-forest ecosystems. The Impact of Nitrogen Deposition on Natural and Semi-natural Ecosystems (ed. S.J.Langan), pp. 85–122. Kluwer, Dordrecht, the Netherlands. & (
- 1994) Nitrogen use efficiency of Carex species in relation to nitrogen supply. Ecology 75, 2362–2372. & (
- 2000) The mineral nutrition of wild plants revisited: a re-evaluation of processes and patterns. Advances in Ecological Research 30, 1–67. & (
- 1992) Root turnover as a determinant of the cycling of C, N, and P in a dry heathland ecosystem. Biogeochemistry 15, 175–190. , & (
- 1998) The effects of air-borne nitrogen pollutants on species diversity in natural and semi-natural European vegetation – a review. Journal of Ecology 86, 717–738. , & (
- 2001) Vegetation changes in two Swiss fens affected by eutrophication and desiccation. Botanica Helvetica 111, 139–155. , & (
- Bordeleau, L.M. & Prévost, D. (1994) Nodulation and nitrogen fixation in extreme environments. Plant and Soil 161, 115–125.
- 1964) Experimental investigations into the mineral nutrition of several grass species. IV. Nitrogen level. Journal of Ecology 52, 665–676. , , & (
- 1998) Effects of litter accumulation and water table on morphology and productivity of Phragmites australis. Wetlands Ecology and Management 5, 275–287. (
- 1999) Effect of root oxygen stress on phosphorus uptake by cattail. Journal of Plant Nutrition 22, 459–466. , & (
- 1997) Above-ground growth and nutrient use by plants in a subarctic environment: effects of habitat, life-form and species. Oikos 79, 311–324. & (
- 1993) A comparative study of the growth and morphology of eight grass species from habitats with different nutrient availabilities. Functional Ecology 7, 223–229. & (
- 1996) Vegetation Mitteleuropas mit den Alpen 5.Aufl. Eugen Ulmer, Stuttgart, Germany. (
- 1991) Zeigerwerte von Pflanzen in Mitteleuropa. Scripta Geobotanica 18, 1–248. , , , , & (
- 1992) The effect of nitrogen nutrition on growth and biomass partitioning of annual plants originating from habitats of different nitrogen availability. Oecologia 92, 236–241. & (
- 1995) Variation in survival and biomass of two wetland grasses at different nutrient and water levels over a six week period. Bulletin of the Torrey Botanical Club 122, 24–29. , & (
- 2002) Lack of evidence for programmed root senescence in common bean (Phaseolus vulgaris) grown at different levels of phosphorus supply. New Phytologist 153, 63–71. , & (
- 2001) Long-term disadvantages of selective root placement: root proliferation and shoot biomass of two perennial grass species in a 2-year experiment. Journal of Ecology 89, 711–722. & (
- 1998) Interspecific variation in plasticity of grasses in response to nitrogen supply. Population Biology of Grasses (ed. G.P.Cheplick), pp. 155–181. Cambridge University Press, Cambridge, UK. (
- 1994) On ecological inference from laboratory experiments conducted under optimum conditions. A Whole Plant Perspective on Carbon–Nitrogen Interactions (eds J.Roy & E.Garnier), pp. 267–294. SPB Academic Publishing, The Hague. & (
- 1995) Growth and allocation of the arctic sedges Eriophorum angustifolium and Eriophorum vaginatum: effects of variable soil oxygen and nutrient availability. Oecologia 104, 330–339. , & (
- 2000) Fertilization effects on species density and primary productivity in herbaceous plant communities. Oikos 89, 428–439. , , & (
- 1988) The effects of nutrient additions on mixtures of Typha latifolia L. and Typha domingensis Pers. along a water-depth gradient. Aquatic Botany 31, 83–92. (
- 2002) Variation in nitrogen and phosphorus concentrations of wetland plants. Perspectives in Ecology, Evolution and Systematics 5, 37–61. & (
- 2002) Time-dependent effects of fertilisation in Dutch floating fens. Journal of Vegetation Science 13, 705–718. , & (
- 2003) N : P ratios as indicators of nutrient limitation for plant populations in wetlands. Ecological Applications 13, 372–384. , & (
- 1996) Deposition of nitrogen-containing compounds to an extensively managed grassland in central Switzerland. Environmental Pollution 91, 21–34. , , et al. (
- 2001) Responses of 21 wetland species to shortages of light, nitrogen and phosphorus. Bulletin of the Geobotanical Institute ETH 67, 13–26. , & (
- 1977) Ökologische Zeigerwerte zur Schweizer Flora. Veröffentlichungen des Geobotanischen Instituts ETH, Stiftung Rübel, Zürich 64, 1–208. (
- 1996) Flora Helvetica. Paul Haupt, Bern, Switzerland. & (
- 1998) The effect of flooding on the recruitment of reed marsh and tall forb plant species. Plant Ecology 139, 13–23. , & (
- 1999) Control of plant species richness and zonation of functional groups along a freshwater flooding gradient. Oikos 86, 523–534. , , & (
- 1999) Redox potentials in relation to water levels in different mire types in the Netherlands and Poland. Plant Ecology 140, 41–51. & (
- 1999) Mycorrhizal status of the genus Carex (Cyperaceae). American Journal of Botany 86, 547–553. , , & (
- 1991) Effect of nitrogen loading on wetland ecosystems with particular reference to atmospheric deposition. Annual Reviews of Ecology and Systematics 22, 257–280. (
- 1990) Effects of nutrients and water level on emergent macrophyte biomass in a prairie marsh. Canadian Journal of Botany 68, 1007–1014. (
- 1995) Phosphorus inputs to terrestrial ecosystems. Journal of Ecology 83, 713–726. (
- 2001) Nutrient limitation along a productivity gradient in wet meadows. Plant and Soil 234, 171–179. , & (
- 1992) Effects of light and nutrient availability on dry matter and N allocation in six successional grassland species: testing for resource ratio effects. Oecologia 89, 412–421. (
- 1996) Effects of soil P status on growth and P and N uptake of Carex species from fens differing in P-availability. Acta Botanica Neerlandica 45, 381–392. & (
- 1999) Biomass allocation and phosphorus productivity of Carex species in relation to soil phosphorus status. Israel Journal of Plant Sciences 47, 97–102. & (
- 1999) Establishment of helophytes in the course of fen restoration. Applied Vegetation Science 2, 131–137. , , , & (
- 1996) The importance of tissue density for growth and life span of leaves and roots: a comparison of five ecologically contrasting grasses. Functional Ecology 10, 717–723. (
- 2000) Interspecific variation in RGR and the underlying traits among 24 grass species grown in full daylight. Plant Biology 3, 426–436. & (
- 1997) Phosphorus allocation and utilization in three grass species with contrasting response to N and P supply. New Phytologist 137, 293–302. , & (
- 1984) A comparative ecophysiological study on the effects of waterlogging and submergence on dune slack plants: growth, survival and mineral nutrition in sand culture experiments. Oecologia 62, 279–286. (
- 2000) Biomass partitioning, architecture and turnover of six herbaceous species from habitats with different nutrient supply. Plant Ecology 149, 219–231. & (
- 1984) Nutrient budgets of marsh plants: efficiency concepts and relation to availability. Ecology 65, 1491–1510. & (
- 1985) The effects of nitrate, phosphate and potassium fertilization on growth and nutrient uptake patterns of Phragmites australis (Cav:) Trin. ex Steudel. Aquatic Botany 21, 53–62. & (
- 1996) Nitrogen losses in perennial grass species. Oecologia 106, 137–143. , & (
- 1993) Contribution of physiological and morphological plant traits to a species’ competitive ability at high and low nitrogen supply. Oecologia 94, 434–440. , , & (