Phytotoxicity dynamics of decaying plant materials

Authors


Author for correspondence: Stefano Mazzoleni Tel: +39 081 7754850 Fax: +39 081 7760104 Email: stefano.mazzoleni@unina.it

Summary

  • • Allelopathic effects of plant litter have been extensively studied, but less attention has been given to the dynamics of phytotoxicity during the decomposition processes.
  • • Decomposition experiments were carried out on above- and below-ground plant materials of 25 species of different functional groups (nitrogen fixer, forbs, woody and grasses–sedges) in aerobic and anaerobic conditions. The phytotoxicity of aqueous extracts of decomposing material was assessed by bioassay in 30 d of laboratory and 90 d of litterbag decomposition experiments.
  • • Phytotoxicity was widespread with c. 90% of the tested species showing significant phytotoxic releases. Phytotoxicity largely varied between different plant functional groups (nitrogen fixer > forbs = woody >> grasses–sedges) and was higher for leaf compared with root materials. In all species tested during decomposition, phytotoxicity rapidly decreased in aerobic conditions but sharply increased and became stable in anaerobic conditions.
  • • The results demonstrate an unexpectedly widespread occurrence of phytotoxicity with clear dynamic patterns during the decomposition processes of plant materials. The ecological consequences of this might have been underestimated.

Introduction

Allelopathy is the negative effect of chemicals released by plant species on growth of other organisms (Rice, 1984). Allelochemicals have been reported to affect neighbouring plant individuals (Schenk et al., 1999; Vivanco et al., 2004) and soil microbial communities, including bacteria, nematodes (Shaukat et al., 2002) and pathogenic and mycorrhizal fungi (Souto et al., 2000). Allelopathic compounds can be either actively released by plants (Bais et al., 2003) or passively produced during the decomposition process of both above- and below-ground plant residues (Blum et al., 1999; Singh et al., 1999).

Decomposition of plant litter is a central process of ecosystems functions and nutrient cycling. Several studies widely investigated the factors affecting the decomposition rates and the related dynamics of nutrients according to environmental conditions (Gholz et al., 2000), litter chemical characteristics (Couteaux et al., 1995) and litter diversity (Gartner & Cardon, 2004). Furthermore, plant litter has been reported to have both positive (Facelli & Pickett, 1991; Xiang & Nilsson, 1999) and negative (Bergelson, 1990; Wedin & Tilman, 1993; Singh et al., 1999) influences on growth and regeneration of plant species. These effects have been related to different mechanisms such as: physical impediment (Wedin & Tilman, 1993), reduced light penetration and changes in the red/far red ratio (Schimpf & Danz, 1999), effects on predation activity (Facelli, 1994) and release of allelochemical compounds during organic matter decomposition (van der Putten et al., 1997; Blum et al., 1999; Armstrong & Armstrong, 2001; Bonanomi et al., 2005). Although the allelopathic effects of plant litter have been extensively studied, little attention has been given to the dynamics of phytotoxicity during the decomposition processes. The organic compounds produced by litter decomposition undergo several physical, chemical, and biological processes in the soil, such as sorption and polymerization by soil organic matter and clay minerals (Makino et al., 1996) and chemical transformation by microorganisms (Blum et al., 1999). These changes over time of both composition and quantity of allelochemicals can either increase or decrease the phytotoxicity of decomposing plant litter (An et al., 2001). However, the dynamics of phytotoxicity have rarely been investigated in natural plant communities (Jäderlund et al., 1996), but have been documented for several crop species in aerobic conditions (Patrick et al., 1963; Chou & Lin, 1976; Putnam, 1994). In general, the most severe inhibition effects have been observed in the early stages of decomposition, followed by decreases in phytotoxicity (Cochrane, 1948; Jäderlund et al., 1996). Anaerobic conditions have been reported to produce stronger and more durable phytotoxic levels (Patrick, 1971; Armstrong et al., 1996).

It is clear that the dynamic release of phytotoxicity during litter decomposition can potentially affect population dynamics (Rice, 1984) and, consequently, the structure of plant communities. Some information is available on phytotoxicity dynamics during decomposition of leaf litter (An et al., 2001) whereas, despite its potential ecological relevance, phytotoxicity related to root litter decomposition has received no attention. The objective of this work was the analysis of phytotoxicity during the decay process of leaf and root material of 25 species from different Mediterranean plant communities. The study aimed to address the phytotoxicity of different plant functional groups (nitrogen fixers, forbs, woody, grasses and sedges) and plant material (leaves and roots) in both aerobic and anaerobic conditions. We investigated the dynamical patterns of litter phytotoxicity under controlled and standardized conditions.

Materials and Methods

Plant material collection

Twenty-five species of four functional groups were selected from different Mediterranean vegetation types (Table 1). Samples were collected from natural plant communities (in the Campania Region, Southern Italy). For each species, leaves and thin roots (diameter < 2 mm) were collected (plants, n > 20), dried to allow storage (+40°C for 5 d), chopped with scissors (size < 1 cm), and then stored at room temperature.

Table 1. Plant material characteristics and experimental conditions of the decomposition processes
Functional groupVegetation typeSpeciesPlant materialDecomposition
Grasses–sedgesMobile sand dune(1) Ammophila littoralis (Beauv.) Rothm.LeafAerobic
Grassland(2) Carex distachya Desf.Leaf–rootAerobic
Grassland(3) Dactylis hispanica Roth.Leaf–rootAerobic
Woodland(4) Festuca drymeia M. et K.LeafAerobic/anaerobic/litterbag
Wetland(5) Juncus effusus L.Leaf–rootAerobic/anaerobic
Grassland(6) Phleum subulatum (Savi) Asch et Gr.LeafAerobic
Wetland(7) Phragmites australis (Cav.) Trin.Leaf–rootAerobic/anaerobic
N-FixerWoodland(8) Coronilla emerus L.LeafAerobic/anaerobic/litterbag
Mobile sand dune(9) Medicago marina L.LeafAerobic
Grassland(10) Medicago minima L.Leaf–rootAerobic
Grassland(11) Melilotus neapolitana Ten.Leaf–rootAerobic
WoodyShrubland (12) Cistus incanus L.LeafAerobic
Shrubland(13) Cistus monspeliensis L.LeafAerobic/anaerobic
Woodland(14) Fraxinus ornus L.LeafAerobic
Woodland(15) Pinus pinea L.LeafAerobic
Woodland(16) Quercus ilex L.LeafAerobic/anaerobic/litterbag
Shrubland(17) Rosmarinus officinalis L.LeafAerobic
ForbsGrassland(18) Bellis perennis L.Leaf–rootAerobic
Woodland(19) Hedera helix L.LeafAerobic/anaerobic/litterbag
Grassland(20) Lobularia maritima L.Leaf–rootAerobic
Grassland(21) Petrorhagia saxifraga L.Leaf–rootAerobic
Grassland(22) Petrorhagia velutina Guss.Leaf–rootAerobic
Grassland(21) Teucrium chamaedrys L.Leaf–rootAerobic
Grassland(21) Teucrium polium L.Leaf–rootAerobic
Wetland(25) Typha latifolia L.Leaf–rootAerobic/anaerobic

Laboratory decomposition

A decomposition experiment was carried out under laboratory conditions. Dry plant material (both leaves and roots) of all species was wetted by distilled water (5% dry weight −50 g l−1). Decomposition processes were carried out in water inside 2-l beaker at 28 ± 4°C for 30 d. Saturated aerobic conditions were obtained by pumping air into the solution, while keeping the beaker closed generated anaerobic conditions. Redox potential (Eh), pH and electrolytic conductivity (EC) of aqueous extracts were measured after filtration to assess the conditions of the decomposition processes. Redox potentials were corrected according to pH levels. Distilled water was added to the beakers to compensate for evaporation losses. A microbial inoculum was made with soil collected under plants of each different species (10 ml of solution obtained from 1 g of the species soil at 10% w : w solution) in order to improve the start-up of the organic matter decomposition processes. This method of decomposition was used because it is rapid, easily reproducible and allows a standardization that avoids the effect of soil type (Zucconi et al., 1981). The experimental conditions of decomposition in water can be considered comparable to those found in the field, because the soil microbial community always operates in thin water films either surrounding solid particles or inside the soil aggregates (Stotzky, 1997; Nannipieri et al., 2003).

Samples of the aqueous suspensions were collected at different stages of the decomposition process. The first sampling was done after 5 h from the start of the experiment and the following after 5, 10, 20 and 30 d. Samples were centrifuged (2395g for 10 min), sterilized (microfiltration with 0.22-µm pore filter), diluted with distilled water to three concentrations (50, 16.6 and 5 g l−1) and stored at −20°C until bioassay.

Litterbag decomposition

Phytotoxicity dynamics were also studied according to the buried litterbag method (Conn & Dighton, 2000) to simulate field decomposition conditions. Four coexisting species (one for each functional group) were chosen from an old-growth holly oak (Quercus ilex) forest stand (Table 1): Hedera helix, Quercus ilex, Coronilla emerus and Festuca drymeia. Plastic litterbags (mesh size 1 mm) were filled with 6 g of dry plant material prepared as described earlier. Litterbags were buried in large trays (30 cm deep, 100 cm for each side) filled with soil from the same stand, kept in a greenhouse in controlled warm (+25°C night and +30°C day) and wet (soil was watered daily to field capacity) conditions to speed up the decomposition process. Litterbags (n = 4) were sampled after 10, 30 and 90 d of decomposition for a total of 48 samples. The bags were dried in the laboratory (+40°C for 5 d) and the material was then mixed with distilled water in beaker at 5% of dry weight (50 g l−1) and shaken for 5 h. The aqueous suspensions were then treated according to the same protocol of the laboratory litter decomposition, i.e. centrifuged (4300 r.p.m. for 10 min), sterilized (microfiltration with 0.22-µm pore filter), diluted by distilled water to three concentrations (50, 16.6 and 5 g l−1) and stored at −20°C until bioassay.

Bioassay

Root elongation tests were carried out with Lepidium sativum L., which is recognized as a sensitive bioassay for phytotoxic compounds (Zucconi et al., 1981; Heil et al., 2002; Gehringer et al., 2003). The use of one test plant has the advantage of standardizing the results of different litter origins. The experiment was done in a growing chamber at constant temperature (+27°C) and in dark conditions. Ten seeds were placed on 9 cm Petri dish over sterile filter paper with 4 ml of test solution. The experiment was performed using three different concentrations of aqueous extracts (50, 16.6 and 5 g l−1). However, only the highest concentration level is reported in statistical analysis and results because, except for obvious concentration effects, they showed the same trend. Every solution plus the control on distilled water was replicated 10 times for a total of 1600 seeds for each plant material and species.

A second experiment was done according to the same design to test the effects of different pH (4.4, 5.5, 6.3, 7 and 8.8) and EC (2, 4.6, 7.6, 9 and 13.1 mS cm−1) levels. The tested pH levels were obtained using a 2-(N-morpholino) ethanesulfonic acid (MES) buffer (Sigma-Aldrich Co. Steinheim, Germany) adjusted with 1 m NaOH. The EC values were obtained using 20, 50, 80, 100 and 150 mmol l−1 of NaCl solution (Macias et al., 2000).

Petri dishes were arranged in growing room according to a totally randomized design and seedlings root length was measured after 36 h from germination. A total of 86 000 seedlings were measured for all the experiments. Data were always expressed as per cent of inhibition of root length compared to control growth.

Statistical analysis

Three-way anova was performed twice to test the main effects and interactions of conditions of decomposition (aerobic vs anaerobic), plant material (leaf vs root or species functional groups) and decomposition time on the inhibition of Lepidium root length. Two-way anova was applied to test the effects of single species (only for the four species used in the litterbag study) and decomposition time on the inhibition of Lepidium root length for both laboratory and litterbag extract. Two-way anova was also applied to test the effects of conditions of decomposition (aerobic vs anaerobic) and decomposition time on extracts pH and EC. One-way anova was applied to test the effects of pH or EC on Lepidium root length.

Results

Inhibition of root length of Lepidium was significantly affected by the conditions of decomposition, the plant materials and the species functional groups (Table 2). There were also significant interactions between conditions and time of decomposition and between decomposition conditions and species functional groups (Table 2).

Table 2. Synthetic results of two three-way anova of percentage inhibition of root elongation of Lepidium sativum
Effectdf F P-value
  1. Condition of decomposition (aerobic and anaerobic), plant material (leaf and root) and decomposition time (0, 5, 10, 20, 30 d) are the main factors of the first analysis (a). Condition of decomposition, plant functional groups (N–fixers, Forbs, Woody, Grasses–sedges) and decomposition time are the main factors of the second analysis (b). P-values < 0.05 in bold type.

(a) Plant materials
Condition of decomposition142.45 < 0.001
Plant material16.08   0.014
Time of decomposition40.93  0.44
Condition × plant material10.24  0.62
Condition × time42.68   0.032
Plant material × time40.2  0.93
Condition × plant material × time40.08  0.98
(b) Plant functional groups
Condition of decomposition139.09 < 0.001
Functional groups35.60   0.001
Time of decomposition41.12  0.34
Condition × functional groups33.43   0.018
Condition × time4J3. 2 1   0.014
Functional groups × time120.35  0.97
Condition × functional groups time120.19  0.99

Aerobic and anaerobic conditions of decomposition produced contrasting dynamics of phytotoxicity levels. At the initial stage of decomposition, phytotoxicity was relatively high in both cases (> 40% inhibition), but either decreased or increased in aerobic and anaerobic conditions, respectively (Fig. 1). The redox potential of aqueous extracts was always higher in aerobic (average E7 values ± 1 SE between species of Fig. 1 after 30 d of decomposition: 170 ± 17.1 mV) than in anaerobic (58 ± 7.8) conditions. The pH of extracts was significantly affected by the decomposition conditions (anova, P < 0.001) but not decomposition time (anova, P = 0.12), while the interactions between terms was significant (anova, P < 0.001). Specifically, in aerobic condition pH values show a clear increase whereas slightly decrease in anaerobic conditions (Fig. 1). The EC of extracts was unaffected by the conditions and decomposition time (anova, P = 0.9 and 0.48, respectively; Fig. 1). Lepidium root growth was not influenced by both pH and EC levels within the tested range (Table 3).

Figure 1.

Inhibition of root length compared with control (= 0%) of Lepidium sativum by aqueous extracts (50 g l−1) (a), pH (b) and electrolytic conductivity (EC) (c) according to aerobic (open circles) and anaerobic (closed circles) conditions during decomposition processes of plant leaf material from different species. Data are averages of eight species (species 4, 5, 7, 8, 13, 16, 19 and 25 – see Table 1 for names). Bars, ± 1 SE.

Table 3. Effects of pH and electrolytic conductivity (EC) on root length of Lepidium sativum compared with control (distilled water)
pHRoot lengthEC (mS cm−1)Root length
  1. Data are averages (± 1 SE) of 10 replicates, different letters show significant differences (P < 0.05).

4.413.1 ± 3.1 a 217.7 ± 1.7 a
5.518.3 ± 2.2 a 4.615.4 ± 23 a
6.3114.9 ± 1.1 a 7.617.1 ± 0.7 a
7.015.1 ± 0.9 a 9.016.5 ± 1.7 a
8.814.1 ± 1.8 a13.114.6 ± 0.9 a
Control18.2 ± 1.7 aControl18.2 ± 23 a

Leaves showed higher phytotoxicity compared with roots (Fig. 2). This was evident at the initial stage of the decomposition process, but in anaerobic conditions the phytotoxicity increased for both plant materials, reaching similar levels after 30 d. By contrast, in aerobic conditions, the phytotoxicity of leaves progressively decreased in 1 month, whereas the phytotoxicity of roots, after an early slight increase, declined to very low levels within 20 d (Fig. 2).

Figure 2.

Inhibition of root length of Lepidium sativum by aqueous extracts (50 g l−1) of leaves (open circles) and roots (closed circles) according to decomposition conditions. Data are averages of 13 and three species in aerobic and anaerobic conditions, respectively (species 2, 3, 5, 7, 10, 11, 18, 20, 21, 22, 23, 24, 25 in aerobic conditions and 5, 7 and 25 in anaerobic conditions – see Table 1 for names). Positive values indicate inhibition and negative values stimulation compared with control (= 0%). Bars, ± 1 SE.

Species functional groups showed very distinct dynamics of phytotoxicity during the decomposition processes. In anaerobic conditions, despite different initial levels, the inhibitory effect strongly and quickly increased in all functional groups (Fig. 3). Conversely, in aerobic conditions, the functional groups showed different dynamics of phytotoxicity. Following an initial high level, a consistent and steady decline of the inhibitory effect was evident for woody species. Forbs showed an initial increase followed by a significant but less marked reduction of phytotoxicity. Nitrogen fixers maintained very high phytotoxic levels, with only slight reductions, in 30 d (from > 80% to > 60% inhibition effect). By contrast, grasses and sedges showed very low levels of phytotoxicity (< 20%), that were slightly elevated only for a short period after 10 d of decomposition. Leaves and roots of this group produced a positive stimulation of root length after 20 d of decomposition (Fig. 3). Both the forbs and grasses–sedges groups showed a marked variability between species.

Figure 3.

Inhibition of root length of Lepidium sativum by aqueous extracts (50 g l−1) according to decomposition of plant material (leaves and roots) of different functional groups (N-fixers, open circles; forbs, open squares; woody, closed squares; grasses and sedges, closed circles). Data are averages of material of all and eight species for aerobic and anaerobic conditions, respectively (in anaerobic conditions species are 4, 5, 7, 8, 13, 16, 19 and 25 – see Table 1 for names). Positive values indicate inhibition and negative values stimulation compared with control (= 0%). Bars, ± 1 SE.

Comparison of phytotoxic release during decomposition between laboratory and litterbag results showed similar trends (Fig. 4), with the ranking of species phytotoxicity remaining the same for both study methods. Phytotoxicity of aqueous extracts of the four coexisting species (Quercus, Hedera, Coronilla and Festuca) was different for all species (anova, P-values always < 0.01) in relation to time of decomposition (anova, P < 0.001), in both laboratory and litterbag conditions. There was also a significant interaction between species and time of decomposition (anova, P < 0.001), both in laboratory and litterbag experiments. The decreasing trend of phytotoxicity was more evident for Hedera, Coronilla and Quercus than in Festuca (Fig. 4).

Figure 4.

Inhibition of root length of Lepidium sativum by aqueous extracts (50 g l−1) during aerobic decomposition processes under laboratory (above) and field (below) conditions (see text for details) of four species of different functional groups (Hedera helix, crosses; Coronilla emerus, circles; Quercus ilex, squares; Festuca drymeia, triangles). Positive values indicate inhibition and negative values stimulation compared with control (= 0%). Data are averages of 10 replicates; bars, ± 1 SE.

Discussion

This study highlighted several unexpected and general patterns of phytotoxic release during decomposition of plant material for a representative group of Mediterranean species. First of all, phytotoxicity of decomposing plant leaf and root appears to be a general phenomenon not restricted to a few allelopathic species. Only three (all from the grasses–sedges group; species 1, 2 and 3 –Table 1) out of 25 species did not produce any phytotoxic effect during the whole decomposition process. This amounts to 88% of the tested species showing a significant phytotoxic release during decomposition. The experiments provided evidence that phytotoxicity is significantly affected by the type of plant material (above- and below-ground plant litter) and by the type of species functional groups. Our results supported previous observations for both aerobic (Cochrane, 1948; An et al., 2001) and anaerobic decomposition processes (Patrick, 1971) and showed consistent changes of pH in these conditions.

Release, production, transformation and destruction of phytotoxic compounds by the microbial activity simultaneously occur during decomposition processes of decaying plant residues (Blum et al., 1999), and their balances determine the net phytotoxic effect on plants. The aim of this study was not to unravel the chemical nature of phytotoxins, which is being considered in other ongoing studies, but to investigate the net inhibitory effect of decomposition processes on root length. The initial phase of decomposition consists of plant tissue comminution and subsequent release of cell content. At this stage, phytotoxicity showed a larger variability in the forbs and grasses–sedges functional groups. Nitrogen fixer species were the most toxic, followed by the woody group, while the grasses and sedges showed very low levels of phytotoxicity. In relation to plant organs, leaf tissues produced greater phytotoxicity than roots, which always showed very low initial phytotoxic levels. So far, we are still unable to explain these large differences among functional groups and between leaves and roots. However, litter of nitrogen (N) fixer species characterized by high N content and low carbon (C) : N ratio (Dilly & Munch, 1998; Rice et al., 2004) have been reported for their phytotoxicity in previous studies (Miller, 1996).

In the early stages of aerobic decomposition of leaf and root, phytotoxicity showed contrasting trends with either a weak decrease or increase according to the type of plant material (Fig. 2). Relatively constant high levels of phytotoxicity were observed for nitrogen fixers, while a rapid decrease characterized the phytotoxic effects of the woody plant material. The forbs group showed an initial significant increase in phytotoxicity followed by a slow decrease (Fig. 3). By contrast, the grasses and sedges, as mentioned earlier, were characterized by very low phytotoxic effects that showed, in any case, a peak after 10 d of decomposition (Fig. 3). After these initial dynamics, as decomposition proceeded in aerobic conditions, phytotoxicity steadily decreased for both roots and leaves (Fig. 2), in all functional groups (Figs 3 and 4).

In anaerobic conditions the phytotoxicity dynamics were completely different, with the phytotoxic level steadily increasing during the whole decomposition process irrespective of plant material, species and functional groups (Figs 1–3). These observed contrasting effects of aerobic and anaerobic conditions indirectly support the hypotheses that microbial activity is the driving factor of phytotoxicity dynamics.

It is known that pH can affect the solubility of allelochemicals compounds with increasing phytotoxicity in acidic conditions (Armstrong & Armstrong, 1999). This is consistent with our results that show the lack of a direct effect of pH on test plants (Table 3) but higher phytotoxicity at low pH levels in anaerobic conditions (Fig. 1).

A phytotoxic release during decomposition processes can have potential effects on plant nutrient uptake (Inderjit & Duke 2003) and possibly affect species competitive ability. Studies of allelopathy under field conditions are rare and their interpretability is mostly limited by the lack of comparative experimental bioassays and field experiments (Inderjit & Callaway 2003). A major criticism of allelopathy is that although most plant species could produce, by metabolism and concentration processes, some phytotoxic compounds, these would be rapidly degraded by the soil microbial activity into nontoxic molecules with reduced impact on plant population dynamics (Harper, 1977). The results of this work clarify the dynamic patterns of phytotoxicity during the decomposition process. It seems clear that anaerobic conditions do affect plant roots by the persistence of phytotoxic compounds derived from the decomposition of any type of plant litter. This durable ‘phytotoxic stress’ can likely interact with other stresses of anoxic environments, such as oxygen deficit (Callaway & King, 1996), and the presence of inorganic toxic compounds, e.g. acid sulphite, ferric and nitrite ions (Marschner, 1995). Then, the relative importance of these phenomena should be influenced by the occurrence of waterlogging and likely be related to both climatic conditions, soil characters and microtopography.

Conversely, aerobic condition generate a more complex and dynamic scenario where the microbial decomposition changes the phytotoxicity levels with their initial increase and subsequent decrease during the processes. Our results demonstrate the occurrence of a temporal phytotoxic ‘window’ during the decomposition processes, which ranged between 5 and 30 d. In all cases, phytotoxicity was reduced after 90 d of decomposition in the litterbag experiments. It should be noted that both laboratory and litterbag studies were carried out at optimal temperature and water conditions, so further studies should investigate the phytotoxic dynamics in different environmental conditions.

Depression of plant growth after addition of organic residues and delayed root growth into decomposing litter have been reported (Seligman et al., 1986; Hodge et al., 1998; Blum et al., 1999) and related to N net immobilization by microbial competition (Seligman et al., 1986; Michelsen et al., 1995). However, this supposed N starvation does not seem a plausible mechanism to explain growth depression in N-rich conditions (Miller, 1996) (i.e. generally for litter with a C : N ratio lower than 30; Hodge et al., 1998). Moreover, it has been demonstrated that root colonization of decomposing plant litter can require several weeks for herbaceous species in controlled conditions (c. 35 d for several grasses; reviewed by Hodge, 2004) and months for temperate forest trees under field conditions (Conn & Dighton, 2000). These delays occur despite the availability of nutrients such as potassium (K), calcium (Ca), magnesium (Mg) (Liu et al., 2000) and phosphorus (P) in rich litter, but also of N in low C : N ratio litters (Taylor, 1998; Rice et al., 2004), which are all released only after decomposition has begun. This sharply contrasts with observations from desert grasses and shrubs, where root growth has been reported to occur within a few days into enriched patches of mineral nutrients (Jackson & Caldwell, 1989). We suggest that the observed delay between the nutrient release from decomposing organic matter and the root colonization and uptake could be related to the occurrence of phytotoxic ‘windows’ during the early stages of the decomposition process. Some more general consequences of the phytotoxic dynamics during decomposition might affect some critical phases of species regeneration, such as seed germination and seedling establishment. In natural plant communities these effects will be influenced by soil types (e.g. clay soils are known to reduce phytotoxic effects by binding allelochemicals to soil colloids; Rice, 1984; Blum et al., 1999).

In conclusion, this paper demonstrates an unexpected widespread occurrence of litter phytotoxicity with clear dynamic patterns during decomposition processes. Indeed, this may change the view of litter only as a nutrient source, because the simultaneous release of nutrients and phytotoxicity during decomposition can produce unavoidable constraints to plant growth. The ecological consequences of such constraints have been neglected to date, but will require more attention to be given their potential effects on plant community organization. In particular, further studies are needed to elucidate the role and ecological consequences of litter phytotoxicity in natural plant communities. Special attention should be given to the possible mediation of phytotoxicity by different soil types and to the different species-specific effects on coexisting plants.

Acknowledgements

We thank Prof. Franco Zucconi and Dr Paolo Sabatini for the useful discussions on the subject of phytotoxicity of plant litter. We are grateful to two anonymous reviewers for their constructive comments on the manuscript.

Ancillary