|Home | About | Journals | Submit | Contact Us | Français|
Is the release of allelochemicals by the dominant tussock grass Festuca paniculata responsible for its dominance by inhibiting growth of neighbour grasses in subalpine grasslands? As such a community is also structured by mowing practices, what could be the impact of mowing on allelopathy?
A design was used that isolated allelopathy from resource competition by separating donor plants (Festuca paniculata) from target plants (F. paniculata, Dactylis glomerata and Bromus erectus). Leachates from donor pots containing bare soil, unmown F. paniculata or mown F. paniculata continuously irrigated target pots containing seedlings. Activated carbon was added in half of the target pots to adsorb potential allelochemicals. C and N analyses of target potting soil were used to test for any effect of treatments on resources. Total phenol concentration was measured in the solutions flowing from donor to target pots.
Festuca paniculata leachates inhibited seedling growth of D. glomerata and B. erectus. Inhibition was correlated with polyphenol concentration, and was not due to resource competition for nitrogen. Mowing the leaves of the donor plants did not significantly increase this inhibition. The activated carbon treatment was not conclusive as it inhibited the seedling growing under control pots with only bare soil.
The results suggest that allelopathy may be at least partly responsible for F. paniculata dominance in subalpine meadows by inhibition of colonization by neighbouring species.
Interactions between plants can be mediated by resources, but also by other mechanisms, including allelopathic interference (Reigosa et al., 1999; Schenk, 2006; San Emeterio et al., 2007). The link between the presence of a released chemical and its effect on the associated plant, however, remains to be demonstrated (Weidenhamer, 1996). One limitation of allelopathy research is the frequent use of inadequate experimental designs that could not simulate field situations (Inderjit and Weston, 2000; Inderjit and Nilsen, 2003). Numerous laboratory experiments have been conducted, but they probably do not reflect processes under field conditions (Inderjit et al., 2001; Inderjit and Callaway, 2003). In field experiments, allelochemical treatments are more realistic, but allelopathy cannot easily be separated from competition for resources (Harper, 1977; Connell, 1990; Williamson, 1990; Thijs et al., 1994; Weidenhamer, 1996; Nilsen, 2002). Several attempts have been made physically to separate the roots of donor and target plants or to manipulate plant density (Weidenhamer et al., 1989; Nilsson, 1994; Ridenour and Callaway, 2001), but the relative contribution of the two processes to explain plant dominance remains unclear. Based on these considerations, the allelopathic potential of Festuca paniculata in subalpine grasslands was tested by means of a semi-controlled experiment submitted to natural environmental conditions. It partly avoided artefacts of laboratory experiments by limiting allelochemical artificial pulses. Moreover, the physical separation of donor and receiver root systems makes it possible, at least partly, to discriminate resource competition from allelopathy.
Changes in management practices (MacDonald et al., 2000) will alter both composition and diversity of plant communities because such changes could modify the different ways by which plants interfere with one another, including chemical interactions. Damage to leaves is known to modify carbon allocation in plants (Wan and Luo, 2003; Zhang et al., 2005; Bahn et al., 2006) which could influence the synthesis and release of carbon-based secondary metabolites that act as allelochemicals (Siemens et al., 2002; Thelen et al., 2005). Only a few studies have analysed the relationship between biomass removal and allelopathy (and mainly for herbivory: Strengbom et al., 2003; Thelen et al., 2005; Karban, 2007). However, the effects of mowing and herbivory may differ (Nilsen, 2002). Thus, relationships between mowing and allelopathy remain unexplored.
In the subalpine meadows of the southern Alps and other mountains of southern Europe, the meadows are often dominated by F. paniculata, a slow growing cespitous grass that becomes overdominant and reduces biodiversity when mowing is abandoned (Jouglet and Dorée, 1991; Vittoz et al., 2005). In subalpine meadows of the northern Alps, its abundance has been shown to increase from 38 to 70% of biomass with abandonment (Quétier et al., 2007). Allelopathy could be one of the mechanisms determining F. paniculata abundance since: (a) large areas around F. paniculata are often covered by a thick litter layer, and have little vegetation density (F. Viard-Crétat, unpubl. res.); and (b) some other Festuca species are well known for releasing allelochemicals (Bertin et al., 2007), especially polyphenols as in F. arundinacea, F. pratensis and F. rubra (Malinowski et al., 1998; Lipinska and Harkot, 2007). Mowing may modify this release of allelochemicals.
In this study, we addressed the following questions: (a) is the growth of subordinate plants affected by leachates from F. paniculata; (b) does F. paniculata emit allelopathic compounds in leachates and can this be related to its effect on seedlings; and (c) is this allelopathic effect modified by mowing? The experimental design used donor pots planted with mown or unmown adults of F. paniculata, under which a tube brought the leachates to a target pot.
Subalpine grasslands dominated by Festuca paniculata were located around the experimental garden of the Station Alpine Joseph Fourier at the Lautaret Pass (2100 m, France, Hautes-Alpes, 44°59′N, 6°23′E). The pot experiment was placed here, and thus realistic climatic conditions for subalpine species were obtained. The meadows are typical subalpine grasslands of the southern French Alps (Jouglet and Dorée, 1991), where average temperatures vary between −1·6°C in January and 14°C in July/August (Besse-en Oisans reference point, altititude 1470 m, reference period, 1953–1990 Météo France). We adapted the experimental design of Newman and Rovira (1975) that involved the allocation of ‘donor’ and ‘target’ pots to separate resource competition from allelopathy (Fig. 1); 2-L pots were used. Potting soil was made with 8/12 sand, 3/12 vermiculite and 1/12 compost. Compost was used to produce a fertile soil medium allowing minimal growth of plants and limiting water evaporation. Organic matter content was not higher in the potting soil medium (5–6% carbon d. wt) than in the field (7%, Robson et al., 2007). Potting soil medium was preferred to natural soils to standardize the substrate and to keep the soil medium inside the donor pots, with at the same time a good leaching ability (which was not the case with natural soils that have high clay content). The plant species had enough time (1 year) to exert a chemical influence on the soil medium and to drive soil microbial communities (Singh et al., 2006). Donor pots were fertilized in June 2005 and 2006 with 3 g of low leaching rate fertilizer (Fertiltop®, 16–8–10 + 4MgO + oligoelements) in order to reduce nutrient limitation and thus have similar nutrient levels in the leachates among the treatments. Donor pots were put on shelves at 70 cm above the ground (Fig. 1). Donor pots had a 10-cm diameter gap on the bottom. In this gap, a 12-cm diameter plastic funnel was glued to the pot to ensure that it was watertight. The funnel was filled with pebbles, above which wire netting of different meshes was placed to retain the soil medium. The funnel was linked to a polyethylene (PE) tube through which leachates could flow. Under all the tubes 180 target pots and 15 PE flasks (500 mL) were placed for chemical analysis of the leachates. On 31 May 2005, 650 ramets of F. paniculata were collected in the field from adult tussocks. Each ramet was separated, the roots and foliage were cut at 5 cm from the base, and they were planted in groups of five into 130 donor pots. Another 65 pots received no ramet but only soil medium. These F. paniculata plants grew during the 2005 season, were kept outside during winter (but protected from frost by snow cover and partial burial) and the experiment was set up again on 21 June 2006. The pots were watered at the beginning of the season with a solution of natural soil collected in surrounding F. paniculata communities, in order to obtain a natural microbial community in the pots, which can influence the emission of allelochemicals (Malinowski et al., 1998, 1999).
The growth of juveniles of three target species was measured. Target pots included F. paniculata itself and two other species that are subordinate species in F. paniculata-dominated meadows: Bromus erectus and Dactylis glomerata. Seedlings of B. erectus, D. glomerata and F. paniculata were germinated in the laboratory in early May from seeds collected the previous year in the meadows surrounding the garden. On 4–6 July 2006, five seedlings of one of the three species were transplanted into each target pot, with 60 target pots for each species. Target potting soil medium was also made with 8/12 sand, 3/12 vermiculite and 1/12 compost, and was fertilized in June 2006 with 3 g of the same fertilizer as used for the donor pots. The experiment was placed in the experimental garden, so that climatic conditions were similar to those of the native communities, as allelopathy may depend on environmental stresses encountered by plants (Anaya, 1999; Xiao et al., 2007). Target pots were watered both by natural rainfall and by leachates coming from donor pots, with automatic watering delivering 50 mL of water three times a day, so that the pots were constantly moist, but not too wet. During the whole season, natural rains brought around 8·1 L in each donor pot whereas artificial watering added 18 L of water. Strong winds dried the pots very quickly at this high altitude and location (a pass), and substantial addition of water in donor pots was necessary, particularly to provide a sufficient watering of the target pot.
In order to adsorb organic compounds that may come from donor pots, an activated carbon treatment was applied to half of the target pots by mixing the soil medium of the upper third of the target pots with 3 g of activated carbon (0·6–2·4 mm, Chimie Plus®). This quantity was calculated from Shevtosa et al. (2005), who used 140 g m−2. As our pots were small in depth and this quantity high, it was adjusted to 100 g m−1, which corresponded to approx. 3 g of activated carbon. This treatment was applied when the target pots were established. Activated carbon is traditionally used for its high surface of adsorption which can retain organic compounds, and should thus remove allelochemicals if present in the experiment.
Finally, to test whether mowing in the field has an effect on allelopathic behaviour, mowing was simulated in half of the F. paniculata donor pots. On 26 July 2006, ramets of the donor pots were cut with scissors at 5 cm above the soil medium surface. Mown biomass was removed from the pots as done in the field.
With ten replicates per treatment, the experiment comprised 10 × 3 target species × 2 activated carbon treatment × 3 donor pot treatments (bare soil, unmown F. paniculata and mown F. paniculata), for a total of 180 target pots and 195 donor pots (Fig. 1). The 15 extra donor pots were used for chemical analysis.
Chemical analyses were conducted on leachates collected during 1 week, once per month, from July to September. After measuring the volume, each solution was filtered (0·45 µm) and analysed for the concentration of total phenols using Folin–Ciocalteu reactant (with gallic acid as a standard) that detects polyphenols, including phenolic amino acids found in F. rubra (Bertin et al., 2007). Above-ground and below-ground biomass of the seedlings was collected at the end of the growing season (20 September 2006). All samples were dried in an oven and weighed. At the same time, approx. 20 g of target soil medium was collected for soil fertility, C and N analyses, and frozen until extraction. Soil medium samples were thawed and sieved at 2 mm to exclude rock and roots. Half of the soil medium was used for extractions of NO3− and NH4+, and half was weighed, dried for estimation of soil water content and then used for C and N percentage analysis. Extractions of NO3− and NH4+ were done with 2 m KCl and 10 g of soil medium. Samples were put in the solution for at least 30 min and the solution was centrifuged and filtered. NO3− and NH4+ concentrations were obtained with a colorimetric chain ‘Lachat Quikchem 8500 FIA’. Dried soil medium was finely ground (500 µm), and for three samples for each combination of treatments 10–20 mg were used to measure C and N quantities with an elementary Analyser Flash EA 1112 (ThermoElectron®).
Carbon content was analysed to make sure that target pots were not enriched in C by C-rich exudates present in leachates of F. paniculata. It was also hypothesized that nitrogen is the keystone for resource competition, as F. paniculata in the field immobilizes available N resources which decreases its availability for other plants (Robson et al., 2007).
The five individuals per target pot were averaged as the experimental unit was the pot. Seedling biomass, leaf/root biomass ratio, soil NO3− and NH4+ concentration and soil %C and %N were analysed using analysis of variance (ANOVA). In this analysis, the target species and the ‘treatment’ were included as factors. Treatment included three modalities: ‘bare soil donor pot’, ‘unmown Festuca donor pot’ and ‘mown Festuca donor pot’. For significant factors, post hoc Tukey HSD tests were used to identify significantly different treatments. Total polyphenol concentrations were analysed using non-parametric tests (Kruskal–Wallis tests) at each date, and a repeated ANOVA procedure was used to test the effect of time. Correlations between mean polyphenol concentrations (n = 5) in September and seedling biomass were calculated. All analyses were conducted with JMP 5.0.1 software (SAS Institute 2002, Cary, NC).
Festuca paniculata leachates reduced seedling biomass of the other species, but not of F. paniculata seedlings (donor treatment × target species, F2,81 = 92·7; P < 0·0001). Festuca paniculata leachates reduced D. glomerata biomass by half compared with the bare soil, and reduced B. erectus biomass by about one-third (Fig. 2). All three target species increased allocation to leaves relative to roots when treated with F. paniculata leachates (F1,54 = 19·89, P < 0·0001; Fig. 3). The interaction was not significant (F2,54 = 0·27, P = 0·762).
Mowing the donor plants tended to increase the inhibitory effect of F. paniculata leachates on seedling growth, but this was not significant as compared with the inhibitory effect (shown by post hoc tests with ‘donor pot type’ used as the factor, including ‘bare soil’, ‘unmown Festuca’ and ‘mown Festuca’; F2,81 = 92·7; P < 0·0001 and see letter in Fig. 2). Mowing did not modify the leaf/root ratio for any of the target species (F1,1 = 0·07, P = 0·80).
The total polyphenol concentration in leachates was greater from F. paniculata donor pots than from bare soil donor pots [0·98 mg L−1 vs. 0·40 mg L−1, donor pot effect in September 2006, χ2 = 9·82, P = 0·007; the effect was also significant earlier (12 July, χ2 = 3·86, P = 0·0495; and 25 July, χ2 = 5·43, P = 0·020); Fig. 4]. Bare soil pots contained small concentrations of polyphenols that probably came from the compost used in the potting soil medium. At the end of the season, F. paniculata donor pots released leachates with twice the concentration of polyphenols compared with that of the bare soil donor pots. Even if time had an effect on concentrations (overall repeated measures ANOVA, donor pot treatment, F1,4 = 11·4, P = 0·028; time, F2,3 = 18·0, P = 0·021), the date did not influence the effects of F. paniculata on polyphenol concentrations (time × treatment, F2,3 = 0·07, P = 0·93). When comparing target total biomasses per pot with average total concentration of polyphenols in the watering solution per treatment in September, a significant negative correlation was found (r = 0·33, P = 0·0014). Polyphenol concentration was the most negatively correlated with root biomass for D. glomerata (r = 0·87, P < 0·0001) and B. erectus (r = 0·59, P = 0·0006), while F. paniculata root biomass was not affected (r = 0·13, P = 0·49). The correlation coefficient with shoot biomass of B. erectus was also significant (r = 0, P = 24), but lower than for the root biomasses. The lower target species biomass with leachates from mown F. paniculata compared with unmown F. paniculata was not, however, matched by an increase in polyphenol concentration (Fig. 4).
Soil carbon in the target pots was influenced by target species and donor pot treatment, only when activated carbon was added (Table 1). Without activated carbon, no significant differences in C content were detected between treatments. Total soil nitrogen concentration did not show any response to treatments, and only marginally responded to target species (Table 1).
Soil medium analyses did not show any difference in nitrate or ammonium concentrations in target pots irrigated with leachates from donor pots with bare soil compared with donor pots with F. paniculata (Table 2), if activated carbon was not taken into account. The only difference in soil inorganic N was between the target species, with higher nitrate contents in pots with F. paniculata or B. erectus than in pots with D. glomerata, which also had lower ammonium concentrations than pots with B. erectus. There was no significant interaction of soil nitrate and ammonium in the analysis if activated carbon was not taken into account. For instance, NO3− concentrations were about three times higher under mown F. paniculata leachates when activated carbon was added to target pots with F. paniculata, compared with target pots without activated carbon.
Activated carbon significantly reduced D. glomerata seedling growth in the target pots under donor pots with bare soil (Fig. 5). There was a significant interaction between species, donor treatment and activated carbon effects (F4,160 = 2·69, P = 0·0326). Activated carbon did not significantly modify biomasses for other treatments and species (post hoc tests, Fig. 5). For the pots treated with leachates of F. paniculata, activated carbon did not alleviate the negative effect of the leachates on the growth as would have been expected. Activated carbon significantly affected soil NO3− concentrations in the target pots depending on the treatment applied (Table 2). It increased NO3− content in the bare soil pots and the mown F. paniculata pots (+64 and +108%, respectively), whereas it decreased it (−42%) in the treatments with unmown F. paniculata. Ammonium content was not modified significantly (Table 2).
Leachates of F. paniculata reduced the growth of seedlings of other species, especially D. glomerata, with a more pronounced effect on roots. Roots may be the primary target as they are in direct contact with the leachates containing allelochemicals. Many studies have shown that roots are particularly sensitive to allelochemicals (Perry et al., 2005; Bertin et al., 2007). This result is particularly important because subalpine grassland species such as B. erectus and D. glomerata have a high allocation to roots at the end of the season and rely on root nutrient and energy storage during the winter. Therefore, reducing root biomass could have consequences for regrowth in the following spring, reducing nutrient reserves of such long-lived perennials. In the target pots under bare soil, the same magnitude of growth rates of the three species were observed as in the field, i.e a very slow growth rate for F. paniculata, intermediate for B. erectus and fast for D. glomerata. One could argue that the difference in biomass between target plants irrigated with leachates from bare soil and F. paniculata donor pots was due to nutrient pre-emption by F. paniculata in donor pots. Neither soil NO3− and NH4+ concentrations nor total nitrogen percentage decreased in the presence of F. paniculata in the donor pots, whatever the target species. It is thus likely that the inhibition of growth in the presence of F. paniculata in donor pots is not due to nitrogen pre-emption. However, the experiment did not permit the exclusion of effects of resource competition for other nutrients, but these may be of minor significance in subalpine grassland for resource competition. Nitrogen is an important feature of subalpine plant communities (Robson et al., 2007).
Festuca donor pots presented greater polyphenol concentrations in the leachates. Thus, polyphenols are good candidates as allelopathic compounds in F. paniculata leachates, like meta-tyrosine found by Bertin et al. (2007) for F. rubra L., although we cannot rule out participation of other metabolites. As a major portion of the toxicity of the polyphenols may be due to a minor fraction of the compounds, further identification experiments are necessary. Allelochemicals could have an effect either directly on plants or indirectly through soil microbial communities (Bowman et al., 2004; Yu et al., 2005; Meier and Bowman, 2008). However, both could be considered to be allelopathic processes. Plants can drive bacterial communities well (Singh et al., 2006), and competitive strategy could include an indirect effect on neighbours through soil biota or mycorrhizal fungi. Seasonal variations of polyphenol concentrations in F. paniculata leachates may have arisen from variation in polyphenol production during the course of development, but also from a greater dilution of polyphenols due to more precipitation in autumn than in July (Météo France data, Monetier-les-Bains, 1926–1987).
Addition of activated carbon in allelopathic experiments is expected to modify neither the biomass nor the soil fertility in control plots (i.e. without donor plants). Recent studies have raised doubts regarding this hypothesis, and Lau et al. (2008) observed an increase of the target biomass when activated carbon is added. Unlike the results of these authors, we observed that activated carbon decreased the growth of D. glomerata in the absence of a donor plant, while the other two species are not affected. A complex interaction between activated carbon, soil micro-organisms, plants and nutrient availability is likely to explain this outcome in the low fertility conditions of this experiment. With the suspected insensitive species F. paniculata, for which growth was not modified, activated carbon under mown F. paniculata increased the total %C of the pot (which may reflect enrichment provided by the charcoal itself), and also available NO3−. This could arise from a higher nitrification rate (Berglund et al., 2004) if microbial communities are stimulated or modified (Pietikainen et al., 2000) by activated carbon or from chemical release of such nutrients from the clay humic complex. Hence, activated carbon has unexpected effects on fertility, which may be driven by competition for nutrients between plant roots and micro-organisms. Moreover, these effects were different depending on the target species and the donor type. This demonstrates a fine interaction between activated carbon, soil processes and biota, with plants that may over-ride adsorption of allelopathic compounds. Lau et al. (2008) also found an interaction between activated carbon and fertility, as well as a different response of growth depending on the species considered. Interpretation of the results produced by activated carbon is thus not recommended by Lau et al. (2008). If scientists wish to continue to use activated carbon in experiments on allelopathy, further investigations on its effect on different types of soil, on microbial communities and on fine exchanges between soil nutrients and roots are needed.
Clipping did not significantly increase inhibition. In contrast to experiments on clipping or herbivory (Reigosa et al., 1999; Thelen et al, 2005; Karban, 2007), leaf damage did not increase the release of allelochemicals nor did it decrease the growth of targets. Allelopathy may not be a mechanism through which F. paniculata dominance is modified by mowing. Mowing, however, interacted with activated carbon in terms of %C and NO3− contents in target pots containing F. paniculata. The absence of a response in our experiment must be interpreted with caution as we did not test repeated mowing over several years as would occur in the field. In subalpine meadows, mowing occurs once a year, but individual plants may experience numerous cuttings as most of the plants are long-lived perennials.
Even though the method we have developed here did not totally answer the debated question of the relative contribution of competition vs. allelopathy to explain F. paniculata dominance in the subalpine field, it allowed us to conclude that this species releases in its leachates enough polyphenols to reduce the growth of its neighbours. Long-term experiments that include several growth cycles, as well as controls of the soil bacterial communities and their activities, are needed to examine the effects of repeated mowing on F. paniculata allelopathy and on D. glomerata and B. erectus recruitment in F. paniculata-dominated grasslands. By manipulating the donor pots, this experimental method could be used as an efficient tool to study the influence of density or of different stress on allelochemical release by F. paniculata. Target pots could also contain a mixture of species in order to measure to what extent their sensibility to F. paniculata allelochemicals could be modulated by biotic (competition) or abiotic (fertility) factors.
This work was supported by the Centre National de la Recherche Scientifique (ATIP project ‘Plant functional traits and dynamics of alpine ecosystems’; GDR 2574 Utiliterres; GDR 2574 Traits and GDR Ecologie chimique), the French Ministry of Higher Education and Research and Université Montpellier II [to F.V.-C.]. We thank Fanny Bouton, Luce Lotito, Laurent Souchaud, Emmanuelle Hélion, Rachel Mold, Laure Doublier, Cécile Bayle, Marie-Pascale Colace and Geneviève Girard for technical help. We are grateful to Florence Baptist and members of the Ecology group of the Botaniska Institutionen, Stockholm, Sweden, for helpful comments and discussions. We also thank two anonymous reviewers for their help in enhancing the quality of the manuscript.