SEARCH

SEARCH BY CITATION

Keywords:

  • decomposition;
  • functional traits;
  • herbivory;
  • native species;
  • New Zealand;
  • nitrogen fixing plant;
  • non-native species;
  • shrub

Summary

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

1. Non-native invasive and nitrogen (N)-fixing plant species can cause large ecosystem-level impacts, particularly when they differ in functionally important plant traits from native and non N-fixing species. However, it remains unclear as to whether and how plant invasion status and N fixation ability consistently influence key plant leaf and litter traits, and trait-driven processes like herbivory and decomposition.

2. We compared leaf and litter traits, leaf palatability and litter decomposability for 41 co-occurring woody species, including native N-fixers, native non N-fixers, invasive N-fixers and invasive non N-fixers, from a New Zealand floodplain. We tested the hypotheses that: (i) invasive and N-fixing species have higher foliar N and specific leaf area, and lower concentrations of defensive phenolics and structural compounds than do native and non N-fixing species, and (ii) invasive and N-fixing species generally produce more decomposable litter and palatable foliage than do native and non N-fixing species.

3. Consistent with our hypotheses, invaders had higher foliar N and N : P ratio, and lower C : N ratio, than did native species. However, in contrast to our hypotheses, foliar phenolics were higher for the invaders while other leaf and litter traits were unaffected by invasion status. Further, N-fixers had higher N and N : P ratios, and lower C : N ratios than did non N-fixers, but other leaf and litter traits were unaffected by N fixation ability.

4. Leaf palatability was unaffected by invasion status but was higher for N-fixers than for non N-fixers. Litter decomposability was unaffected both by invasion status and N fixation ability. We found a significant positive relationship between leaf palatability and litter decomposability across all species, because similar traits, particularly the C : P ratio and total phenolic concentrations of plant tissues, were correlated with both processes.

5. Our results demonstrate that a small number of key traits, such as C : P ratio and total phenolic concentrations, drive both herbivory and decomposition irrespective of plant invasion status or N fixation ability. As such, they highlight that interspecific differences in particular plant traits, rather than plant functional group memberships based on invasion status and N fixation ability, are more effective in predicting palatability and decomposability.


Introduction

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

Invasion of ecosystems by non-native plant species is one of the most important components of global environmental change. Several studies have shown that invasive plant species can have large effects at the whole ecosystem level, through influencing input of nitrogen by biological fixation (Vitousek & Walker 1989; Hughes & Denslow 2005), rates of biogeochemical processes (Ehrenfeld 2003; Liao et al. 2008), decomposition and mineralization rates of plant litter (Ehrenfeld, Kourtev & Huang 2001; Standish et al. 2004), stocks of nutrients in the soil (Hughes & Denslow 2005) and the functioning of the soil biota (Wolfe & Klironomos 2005; Van der Putten, Klironomos & Wardle 2007). These effects of non-native invasive species should be most pronounced when the invasive species differ from co-occurring native species in key functional traits (Vitousek & Walker 1989; Ehrenfeld 2003; Emery 2007). Some studies have shown significant overall functional trait differences between co-occurring sets of non-native invasive and native species (Baruch & Goldstein 1999; Funk & Vitousek 2007; Leishman et al. 2007), although the magnitude of these differences is influenced by environmental conditions, plant functional group, and the timescale considered (Thompson, Hodgson & Rich 1995; Baruch & Goldstein 1999; Daehler 2003; Burns 2006; Funk & Vitousek 2007; Leishman et al. 2007). As such, invasive species often have faster growth rates, higher specific leaf areas (SLA) and higher foliar nutrient concentrations than do co-occurring native species (e.g. Baruch & Goldstein 1999; Burns 2006; Leishman et al. 2007; Peltzer et al. 2009). These trait differences in turn have functional consequences; for example, in comparison with native species, invasive species may produce litter that decomposes faster (Ehrenfeld et al. 2001; Allison & Vitousek 2004; Standish et al. 2004; Kueffer et al. 2008), yield foliage that is more palatable to herbivores (Agrawal & Kotanen 2003 but see Keane & Crawley 2002; Agrawal et al. 2005), and enhance the rates of biogeochemical processes (Ehrenfeld 2003).

One key plant trait that has large ecosystem-level consequences, especially in nutrient-poor ecosystems, is the capacity of the plant to form symbioses with bacteria that convert atmospheric nitrogen (N) to plant available forms. Colonization of new surfaces by N-fixing plant species results in greatly improved soil fertility, ecosystem process rates and nutrition of co-occurring non N-fixing plant species (Vitousek et al. 1987; Stock, Wienand & Barker 1995; Hooper & Vitousek 1998; Bellingham, Walker & Wardle 2001). Further, N-fixing species often have higher foliar nutrient concentrations (Cornelissen et al. 1997; Peltzer et al. 2009) and SLA (Craine et al. 2002; Wright et al. 2004; but see Tjoelker et al. 2005) when compared with non N-fixing species, and again, these traits can be associated with higher rates of litter decomposition and foliar herbivory (Knops, Bradley & Wedin 2002; Pérez-Harguideguy et al. 2003; Tateno et al. 2007; Cornwell et al. 2008). Furthermore, invasive floras often contain a greater proportion of N-fixing species than do the corresponding native floras (Vitousek & Walker 1989; Stock et al. 1995; Vitousek et al. 2002; Ehrenfeld 2003; Levine et al. 2003). If key traits of invasive and N-fixing species do indeed differ from those of native and non N-fixing species, then the impacts of plant species on ecological processes such as herbivory and decomposition could potentially be predicted from their invasion status and N fixation ability.

To understand how invasion status and N fixation ability influence key functional traits, we compared leaf and litter traits, leaf palatability and litter decomposability for co-occuring woody plant species in an early successional floodplain system in New Zealand (Bellingham et al. 2005; Peltzer et al. 2009). This system is ideal for comparative work of invasive and native species because it contains several species of each type both with and without the ability to fix atmospheric N. From this system, we studied 41 common woody species (shrubs and small trees), i.e. between 9 and 11 species in each of the four categories: native N-fixer, native non N-fixer, invasive N-fixer and invasive non N-fixer (see Tables S1 and S2). For each species, we measured key functional traits in both foliage and litter, as well as decomposability of leaf litter and palatability of foliage to a generalist herbivore. We used these data to test the following two hypotheses:

  • 1
     Leaf and litter traits associated with carbon capture strategies are consistently affected by invasion status (Baruch & Goldstein 1999; Leishman et al. 2007) and by N fixation ability (Cornelissen et al. 1997; Peltzer et al. 2009). Thus, foliage and litter from invasive N-fixing species should have traits most closely associated with resource acquisition (i.e. high SLA and foliar nutrient concentrations, low levels of defensive phenolics and structural carbohydrates), while native non-N fixing species should have traits associated with resource conservation (i.e. low SLA and foliar nutrient concentrations, high levels of defensive compounds and structural carbohydrates) (see Coley, Bryant & Chapin III 1985; Díaz et al. 2004); traits for native N-fixers and invasive non N-fixers should be intermediate.
  • 2
     Litter decomposability and leaf palatability are consistently affected by invasion status and N fixation ability. Thus, litter from invasive N-fixers should decompose faster than that from invasive non N-fixers and native N-fixers while that from native non N-fixers should decompose most slowly. Similarly, invasive N-fixers should produce the most palatable foliage while native non N-fixers should produce the least. This is because the leaf and litter traits associated with carbon capture strategies are also related to litter decomposability and foliage palatability, and because both process are often driven by a similar suite of litter and foliar functional traits (Grime et al. 1996; Cornelissen et al. 1999; Wardle, Bonner & Barker 2002; Cornelissen et al. 2004; but see Kurokawa & Nakashizuka 2008).

The issue of whether trait differences between native and invasive N-fixers are similar to those between native and invasive non N-fixers remains unexplored. In testing our hypotheses, the ultimate goal was therefore to determine whether generalizable trait differences exist between co-occurring native and invasive N-fixing and non N-fixing plant species, and the consequences of these differences for ecological processes driven by herbivores and decomposers.

Materials and methods

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

Sample collection and trait analysis

The foliage and litter materials used for our study were collected on and near a river floodplain on the eastern South Island of New Zealand, located at 42°20′S, 173°33′E, at between 220 and 280 m above sea level. The substrate is highly fractured greywacke sandstone with some argillite and tuffaceous sandstone, and the soil has on average 0·2% C, 0·01% N, 0·0017% P and a pH of 8·0 (Bellingham et al. 2005). The study system is floristically diverse, with 62 native and 48 non-native plant species known to co-occur. Further descriptions of the site are given in Bellingham et al. (2005).

We selected 41 woody species (shrubs and small trees) for our study, all of which are common in the study system (see Tables S1 and S2). These include nine native N-fixers, 11 native non N-fixers, 10 invasive N-fixers and 11 invasive non N-fixers. We collected leaf and litter samples during February to March 2007. For each species, we collected a total of at least 20 g fresh fully expanded leaves from c. 5 individuals. Further, we hand-collected at least 30 g freshly produced litter with no obvious signs of degradation from under c. 5 individuals for each species; it is assumed that any minor degradation that might have occurred prior to litter collection would be consistent across the four groups of plants. For some N-fixing species, such as Carmichaelia odorata and Cytisus scoparius, older dried leaves were used as a substitute for true leaf litter (Wardle et al. 2002) because their very small leaf size made collection of adequate amounts of fresh leaf litter impossible (see Tables S1 and S2). The litter samples were air dried until the decomposition trial started in May 2007.

Specific leaf area (cm2 g−1) was determined for 20–50 fresh leaves for each species. Using these samples, dry matter content (DMC) was also calculated after drying at 60 °C to constant mass. Subsamples of each leaf and litter sample were dried at 40 °C to constant mass, milled and analysed for the concentration of several chemical components. Concentrations of total phenolics and condensed tannins were determined colorimetrically following extraction with 50% acetone for 16 h; the procedure for phenolic measurements follows Price & Bulter (1977); and that for condensed tannin measurements follows Broadhurst & Jones (1978); see also Mole & Waterman (1994). Total phenolic concentrations were determined as tannic acid equivalents and condensed tannins as catechin equivalents. Lignin, cellulose and fibre fractions were determined using the acid-detergent fibre–sulphuric acid procedure (Rowland & Roberts 1994). Briefly, acid-detergent fibre (hereafter called ‘fibre’) was determined as the fraction remaining after treatment of the powdered samples with boiling acid detergent to hydrolyse protein. This fibre residue was then treated with 72% sulphuric acid to destroy the cellulose; cellulose is defined as the difference between the initial and remaining fractions. Finally, the residue was ignited at 550 °C to destroy all remaining organic matter, leaving only the inorganic (ash) fraction; acid-detergent lignin (hereafter called ‘lignin’) is defined as the difference in this residue between pre- and post-ignition. Total carbon (C) and Total N were determined by a Leco autoanalyzer (Laboratory Equipment Corporation, St Joseph, MI, USA). Total phosphorus (P) was determined using automated colorimetric methods (Technicon Instruments 1977) after digestion. Values of ratios of C : N, C : P and N : P were derived from the above measurements.

Litter decomposability

The leaf litter decomposability of each species was assessed using a standardized laboratory bioassay (Wardle et al. 1998; Wardle et al. 2002). For each species, six 9 cm diameter Petri dishes were each filled with a 12 g (wet weight) of standardized humus substrate (26·9% C, 1·38% N, 0·14% P, 59·0 mg kg−1 Olsen P, 6·80 pH; collected from Kaituna Valley at 43°74′S, 172°69′E) and dried to 139% moisture (dry weight basis); a disc of nylon mesh with 1 mm holes was placed on the humus surface. Litter (1 g, air-dried, except for 0·5 g for Coriaria kingiana), cut into approximately 10 × 10 mm fragments when dead leaves were bigger than this size, was placed on the surface of the mesh of each Petri dish; the dish was then sealed with tape to minimize water loss and incubated for 103 days at room temperature (15–25 °C). Upon harvest, all remaining (undecomposed) litter was removed from each Petri dish, rinsed, oven dried (60 °C to constant mass) and dry mass determined. Litter decomposition rate was determined as the percentage mass lost during incubation.

Leaf palatability

To test the relative palatability of each plant species, we used a generalist herbivore, the grey field slug, Derocerum reticulatum (Muller) (Mollusca: Stylommatophora: Agriolimacidae); this species has previously been shown to be an effective test organism for comparative studies of foliage palatability (Wardle et al. 1998). Slugs were collected from vegetable gardens and pasture in North Canterbury, New Zealand (43°39′S, 172°29′E) during February and March 2007, and kept in an incubator (18 °C) until the feeding trial. For each plant species, foliage was cut into approximately 10 × 10 mm fragments when the leaf was larger than this size, and 100 mg dry weight equivalent of fresh leaves (determined using measurements of dry matter content) was placed in each of eighteen 9 cm diameter Petri dishes each lined with a moistened paper towel. Two mature slugs (body weight 300–500 mg) were placed in each of nine of these Petri dishes, while the other nine were kept slug-free. The Petri dishes were incubated in a growth chamber (18 °C; day : night light ratio = 16 h : 8 h) for 4 days. After incubation, all leaf material was oven dried (60 °C to constant mass) and the dry weight remaining was determined. For each plant species the difference in dry weight remaining between the herbivore treated and untreated leaf material was used as a measure of palatability for that species (Wardle et al. 1998).

Data analysis

For all data analysis, individual species served as the unit of replication. The leaf traits used for analyses were DMC, SLA, the concentrations of N and P, the ratios of C : N, C : P, and N : P, and the concentrations of fibre, cellulose, lignin, condensed tannins and total phenolics. The litter traits used for analyses were the same as for the leaf traits except that DMC and SLA were not included. For each leaf and litter trait variable, and for leaf palatability and litter decomposability, data were analysed using two-way analysis of variance, with the main factors being invasion status (non-native or native) and N fixation ability (N-fixer or non N-fixer). Further, the full leaf trait data set and the full litter trait data set (except for N and P concentrations) were each summarized into fewer variables by principal components analysis, an ordination approach that has previously been useful in summarizing trait-based data sets of this type (e.g. Díaz et al. 2004; Wright et al. 2004). Data for N and P concentrations were not included to avoid colinearity with other variables that were included such as the ratios of C : N, C : P and N : P. Scores for the two primary ordination axes for both the leaf and litter traits were analysed by anova to test for invasion status and N fixation ability as described above. Correlation and stepwise multiple regression analyses were performed to determine which combinations of leaf traits best predicted leaf palatability, and which combinations of leaf traits and litter traits best predicted litter decomposability. Linear regression analysis was used to determine the relationship between leaf palatability and litter decomposability. Data variables were log(x + 1)-transformed when required to satisfy the assumptions of parametric data analysis. Statistical data analyses were performed by R (v2·6·1) (R Development Core Team, Vienna, Austria) and JMP 7 (SAS Institute, Cary, NC, USA).

Results

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

Leaf and litter traits among plant groups

Several leaf and litter traits were significantly influenced by invasion status and N fixation ability (Tables 1 and 2), and both leaf and litter C : N ratios were significantly higher for native than invasive species and for non N-fixing than N-fixing species (Fig. 1a,b). Further, both leaf and litter N concentrations were significantly higher for invasive than native species and for N-fixing than non N-fixing species (data not presented here; see Tables S1 and S2). Both leaf and litter N : P ratios were significantly higher for invasive than native species and for N-fixing than non N-fixing species (Fig. 1c,d). The concentrations of condensed tannins and total phenolics in both leaves and litter were significantly higher for invasive than native species, but did not differ between N-fixing and non N-fixing species (Tables 1 and 2, Fig. 2). There was no effect of plant invasion status or N fixation ability on leaf P concentration (overall mean ± SD = 0·197 ± 0·090%), litter P concentration (0·133 ± 0·102%), leaf C : P ratio (288 ± 105), or litter C : P ratio (587 ± 455) (Tables 1 and 2). Neither invasion status nor N fixation ability influenced leaf DMC (33·8 ± 8·31%), SLA (126·9 ± 54·4 cm2 g−1) or concentrations of cellulose (12·6 ± 4·45%) or lignin (10·8 ± 5·47%) (Table 1). A significant interaction of invasion status and N fixation ability for foliar fibre concentration was driven by fibre being lower for N-fixing native plants than for non N-fixing natives (t18 = 2·31, = 0·033) but not differing significantly between N-fixing and non N-fixing invasive plants (t19 = 0·903, = 0·378). By contrast, foliar fibre concentration did not differ significantly between native and invasive N-fixing plants (t17 = 1·21, = 0·242), but tended to be lower for invasive non N-fixing plants compared with native non N-fixers (t20 = 1·91, = 0·071). Further, neither invasion status nor N fixation ability influenced concentrations of litter cellulose (12·7 ± 5·18%) or lignin (13·9 ± 7·25%), although litter fibre concentration was lower for N-fixing species (24·6 ± 8·07%) than for non N-fixing species (29·8 ± 6·61%).

Table 1.   Results from two-way anova (F and P values) testing for the effects of plant invasion status and N fixation ability on leaf traits for 41 co-occurring woody plant species
Leaf traitanova
Invasion status (I) (F, P)N fixation ability (N) (F, P)I × N (F, P)
  1. DMC, dry matter content; SLA, specific leaf area.

  2. *Variables that are transformed by log(x + 1) before analyses. Invasion status refers to native vs. invasive species, and N fixation ability refers to N-fixing vs. non N-fixing species. Significant values at P = 0·05 are in bold. Degrees of freedom are: I = 1, N = 1, I × N = 1, residual = 38. PCI and PC II are the first and second ordination axes from PCA and explain 32·0% and 24·5% of the total variability respectively.

DMC (%)1·15, 0·2900·15, 0·7041·89, 0·177
SLA (cm2 g−1)*1·58, 0·2160·65, 0·4272·66, 0·112
N (%)*7·82, 0·00830·08, <0·0010·04, 0·839
P (%)*0·14, 0·7110·39, 0·5390·08, 0·775
C : N*5·45, 0·02526·45, <0·0010·85, 0·364
C : P0·39, 0·5380·74, 0·3941·44, 0·239
N : P*6·90, 0·01278·64, <0·0010·16, 0·697
Fibre (%)0·04, 0·8510·40, 0·5314·45, 0·042
Cellulose (%)*3·77, 0·0603·43, 0·0722·57, 0·118
Lignin (%)*1·71, 0·1990·24, 0·6312·40, 0·130
Condensed tannins (%)*7·92, 0·0081·69, 0·2020·49, 0·490
Total phenolics (%)*7·58, 0·0090·36, 0·5513·35, 0·075
PC I0·03, 0·8691·04, 0·3151·41, 0·242
PC II11·43, 0·0029·62, 0·0045·03, 0·031
Table 2.   Results from two-way anova (F and P values) testing for the effects of plant invasion status and N fixation ability on litter traits for 41 co-occurring woody plant species
Litter traitanova
Invasion status (I) (F, P)N fixation ability (N) (F, P)I × N (F, P)
  1. *Variables that are transformed by log(x + 1) before analyses. Invasion status refers to native vs. invasive species, and N fixation ability refers to N-fixing vs. non N-fixing species. Significant values at = 0·05 are in bold. Degrees of freedom are: I = 1, N = 1, I × N = 1, residual = 38. PCI and PC II are the first and second ordination axes from PCA and explain 33·7% and 25·1% of the total variability respectively.

N (%)*4·47, 0·04168·23, <0·0010·61, 0·439
P (%)*0·34, 0·5630·66, 0·4210·05, 0·830
C : N*5·01, 0·03161·99, <0·0012·96, 0·094
C : P*0·17, 0·6820·41, 0·5270·73, 0·397
N : P*14·32, <0·00190·25, <0·0010·17, 0·682
Fibre (%)0·54, 0·4665·06, 0·0310·77, 0·386
Cellulose (%)*0·29, 0·5960·47, 0·4960·10, 0·754
Lignin (%)*1·41, 0·2432·88, 0·0982·31, 0·137
Condensed tannins (%)*5·86, 0·0200·80, 0·3780·13, 0·725
Total phenolics (%)*12·27, 0·0010·18, 0·6760·75, 0·392
PC I3·54, 0·0683·77, 0·0600·72, 0·400
PC II16·38, <0·00148·10, <0·0013·04, 0·090
image

Figure 1.  Variation in ratios of (a) leaf C : N, (b) litter C : N, (c) leaf N : P and (d) litter N : P between four functional groups of co-occurring plant species (between 9 and 11 species per group), as shown by box and whisker plots. Boxes indicate 50% of values and bars indicate 95% of values. Black dots indicate outliers. Bars topped with the same letter are not significantly different at = 0·05 (Least Significant Difference test).

Download figure to PowerPoint

image

Figure 2.  Variation in concentrations of (a) leaf condensed tannins, (b) litter condensed tannins, (c) leaf total phenolics and (d) litter total phenolics between four functional groups of co-occurring plant species (between 9 and 11 species per group), as shown by box and whisker plots. Boxes indicate 50% of values and bars indicate 95% of values. Black dots indicate outliers. Bars topped with the same letter are not significantly different at = 0·05 (Least Significant Difference test).

Download figure to PowerPoint

Principal components analysis of the leaf traits data showed that the first (PCI) and second (PCII) ordination axes explained 32·0% and 24·5% of the total variability respectively. Further, PCI was strongly related to C : P ratio, DMC and C : N ratio, while PCII was strongly related to concentrations of total phenolics and cellulose (see Table S3). Analysis of variance on the leaf PC axes revealed significant overall effects of invasion status, N fixation ability and their interaction for PCII but not PCI (Table 1). As to the interaction effect, leaf PCII did not differ significantly between N-fixing and non N fixing invasive plants (t19 = 0·658, = 0·516), but was higher in N-fixing natives compared with non N fixing natives (t18 = 3·78, = 0·001). On the other hand, leaf PCII did not differ significantly between native and invasive N-fixing plants (t17 = 0·536, = 0·599), but was significantly higher for invasive non N-fixing plants compared with native non N-fixers (t20 = 4·47, < 0·001). Principal components analysis on the full litter traits data set showed that the PCI and PCII explained 33·7% and 25·1% of the total variability respectively, and that PCI was strongly related to the concentrations of lignin, fibre and condensed tannins, and C : P ratio. Further, PCII was strongly related to the N : P and C : N ratios, and total phenolic concentration (see Table S4). Analysis of variance on those PC axes showed significant overall effects on both invasion status and N fixation ability on PCII but not PCI (Table 2). There were no interactive effects between invasion status and N fixation ability on litter trait variables (Table 2).

Palatability and decomposability

Analysis of variance showed that leaf consumption rate by slugs was significantly higher for N-fixing than for non N-fixing species, although there was no difference between invasive and native species (Fig. 3a). Litter decomposition rate was unaffected by either invasion status or N fixation ability (Fig. 3b). By testing the effect of N fixation ability within both invasive and native plants, it was revealed that both consumption and decomposition did not significantly differ with N fixation ability for invasive plants (t19 = 1·75, = 0·097 and t18 = 0·922, = 0·369 respectively) or for native plants (t18 = 1·75, = 0·097 and t18 = 1·74, = 0·099 respectively). By testing the effect of invasion status within both N-fixers and non N-fixers, it was revealed that both consumption and decomposition did not significantly differ with invasion status for non N-fixing plants (t20 = 1·38, = 0·182 and t19 = 0·351, = 0·729 respectively). For N-fixing plants, the consumption did not differ with invasion status (t17 = 0·625, = 0·540), but the decomposition did marginally differ (t17 = 1·88, = 0·078). There was a significant positive relationship between leaf consumption rate and litter decomposition rate for the entire data set (Fig. 4a) and for both invasive and native N-fixing plants (Fig. 4b), but not for invasive or native non N-fixing plants (Fig. 4c). The slope for regression using ANCOVA differed significantly between N-fixers and non N-fixers (F7,32 = 7·71, = 0·009), with the N-fixing plants having a steeper slope; slopes did not significantly differ between native and invasive species (F7,32 = 0·114, = 0·738) (Fig. 4b,c). Further, there was a significant interaction between invasion status and N fixation ability (F7,32 = 4·40, = 0·04); this interaction was driven by the consumption of native plant species not differing markedly between N- and non N-fixing plants, whereas consumption of invasive species declined strongly for non N-fixers. Within N-fixing plants, there was no significant difference in slopes between native and invasive species (Fig. 4).

image

Figure 3.  Variation in (a) leaf consumption rate by slugs and (b) litter decomposition rate between four functional groups of co-occurring plant species (between 9 and 11 species per group), as shown by box and whisker plots. Black dots indicate outliers. For consumption, F and P values from two-way anova are; native vs. invasive: F1,38 = 1·28, = 0·265, N-fixer vs. non N-fixer: F1,38 = 6·12, = 0·018, interaction term: F1,38 = 0·00, = 0·996. For decomposition, F and P values from anova are; native vs. invasive: F1,38 = 2·13, = 0·153, N-fixer vs. non N-fixer: F1,38 = 0·39, = 0·534, interaction term: F1,38 = 3·61, = 0·065. Bars topped with the same letter are not significantly different at = 0·05 (Least Significant Difference test).

Download figure to PowerPoint

image

Figure 4.  Relationship between leaf consumption rate and litter decomposition rate for (a) all species (R2 = 0·370, < 0·001, solid line), (b) N-fixing species only (native N-fixers: R2 = 0·811, < 0·001, dotted line; invasive N-fixers: R2 = 0·602, = 0·008, dashed line), (c) non N-fixing species only (native non N-fixers: R2 = 0·002, = 0·910; invasive non N-fixers: R2 = 0·035, = 0·606). Leaf consumption rate was represented in log-transformed values for each panel.

Download figure to PowerPoint

Traits related to palatability and decomposability

Across all species, leaf consumption and litter decomposition were related to similar sets of traits. Consumption and decomposition rates were both significantly negatively related to leaf DMC, C : N ratio, C : P ratio, total phenolic concentration, and PCI. In addition, consumption rate was also significantly negatively related to condensed tannin concentrations of leaves while decomposition rate was significantly positively related to SLA (Table 3). Decomposition rate was significantly negatively related to most litter traits except for the N : P ratio, the concentrations of cellulose and condensed tannins, and PCII (Table 3). Between 43% and 48% of the total variability in both consumption and decomposition could be explained by combinations of plant traits using multiple regression analysis (Table 4). Leaf or litter C : P ratio and total phenolics emerged as having significant explanatory power for all regression relationships and therefore serve as important predictors of both consumption and decomposition rates (Table 4, Fig. 5). Other traits that had significant explanatory power in the multiple regression analysis were SLA and foliar N : P ratio for consumption rate, and litter lignin concentration for decomposition rate (Table 4).

Table 3.   Pearson’s correlation coefficients between leaf consumption rate and litter decomposition rate with leaf and litter traits for 41 co-occurring woody plant species
TraitConsumption‡ vs. leaf traitDecomposition vs. leaf traitDecomposition vs. litter trait
  1. DMC, dry matter content; SLA, specific leaf area. PCI and PCII are the first and second ordination axes from PCA.

  2. ‡Variables transformed by log(x + 1) before analyses.

  3. †Litter C : P ratio was transformed by log(x + 1), but leaf C : P ratio was not transformed.

  4. *, **, ***Correlation coefficient significantly different to 0 at = 0·050, 0·010, and 0·001 respectively.

DMC (%)−0·520***−0·398*
SLA (cm2 g−1)‡0·1190·329*
C:N‡−0·492**−0·355*−0·381*
C:P†−0·504***−0·559***−0·680***
N:P‡0·039−0·191−0·263
Fibre (%)−0·281−0·179−0·349*
Cellulose (%)‡−0·136−0·0310·066
Lignin (%)‡−0·269−0·188−0·412**
Condensed tannins (%)‡−0·390*−0·303−0·294
Total phenolics (%)‡−0·418**−0·512***−0·512***
PC I−0·566***−0·508***−0·672***
PC II−0·211−0·3070·076
Table 4.   Multiple regression models relating leaf consumption and litter decomposition to plant traits
Dependent variableIndependent variables in relationshipR2
  1. The data set used and dependent and independent variables are the same as for Table 3. Only those terms that explain a statistically significant proportion of the variation of the response variable at = 0·05 are included. < 0·001 for each regression. (+) and (−) after each variable indicate a positive and negative relationship between the independent and dependent variables. Consumption rate and all independent variables in relationships were transformed by log(x + 1) before analyses.

  2. SLA, specific leaf area; TP, total phenolics.

Consumption rate vs. leaf traitsSLA(−); C:P(−); N:P(+); TP(−)0·48
Decomposition rate vs. leaf traitsC:P(−); TP(−)0·43
Decomposition rate vs. litter traitsC:P(−); lignin(−); TP(−)0·46
image

Figure 5.  Relationships of leaf consumption rate with (a) leaf C : P ratio (R2 = 0·254, < 0·001) and (b) leaf total phenolic concentrations (R2 = 0·175, = 0·007), and of decomposition rate with (c) litter C : P ratio (R2 = 0·463, < 0·001) and (d) litter total phenolic concentrations (R2 = 0·262, < 0·001), for 41 co-occurring plant species. Leaf consumption rate, litter C : P ratio, and leaf and litter total phenolic concentrations were represented in log-transformed values for each panel.

Download figure to PowerPoint

Given that N-fixing and non N-fixing plants differed significantly in their relationships between decomposability and palatability, we also performed multiple regression analyses using leaf and litter traits as predictors separately for N-fixers and non N-fixers. When all N-fixing plant species were considered irrespective of plant invasion status, multiple regression showed that consumption was predicted from leaf C : P ratio and total phenolic concentrations (model R2 = 0·806, < 0·001, d.f. = 18), and all of these variables were negatively associated with consumption. For non N-fixing plant species only, consumption was negatively associated with leaf DMC and SLA but positively associated with leaf total phenolic concentrations (model R2 = 0·484, = 0·002, d.f. = 21). Furthermore, when all N-fixing plant species were considered irrespective of invasion status, decomposition was best predicted from litter C : P ratio and total phenolic concentrations (model R2 = 0·780, P = <0·001, d.f. = 18), and all of these variables were negatively associated with decomposition. For non N-fixing plant species only, decomposition rate was not significantly related to any measured variable at = 0·05.

Discussion

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

Do traits differ among functional groups?

We found only partial support for our first hypothesis that invasive and N-fixing species have traits that are more closely associated with rapid acquisition of resources than do native and non N-fixing species (Baruch & Goldstein 1999; Funk & Vitousek 2007; Leishman et al. 2007). In our study, leaves and litter of invasive species had higher concentrations of N than did native species. Because C and P concentrations did not differ among the two groups, this difference in N concentration also had important implications for nutrient stoichiometry, with invasive species having elevated N : P ratio and depressed C : N ratio relative to native species (Fig. 1). This is consistent with earlier work pointing to invasive and native species differing in their elemental stoichiometry (Davis, Grime & Thompson 2000; Blumenthal 2005; Dassonville et al. 2008). We also found, unsurprisingly, that leaves and litter of N-fixing species had higher N concentrations than did non N-fixing species (Tables 1 and 2), a trend that was consistent across native and invasive species, and one that is well established in the literature (Cornelissen et al. 1997; Peltzer et al. 2009). Moreover, this serves to elevate N : P ratio and depress C : N ratio for N-fixing species compared with non N-fixing species.

On the contrary, leaf DMC, SLA, leaf and litter cellulose and lignin did not significantly differ between invasive and native species, and between N-fixing and non N-fixing species (Tables 1 and 2). Interestingly, in our study system, SLA and foliar N did not vary in the same way across functional groups, despite having previously been shown to vary in tandem in global data sets (Reich et al. 1999; Wright et al. 2004). In the same locations as that used for the present study, Peltzer et al. (2009) showed that invasive species have significantly higher SLA and foliar N overall than did co-occuring native species. In their study, the invasive species were mostly herbs (70·2%) or grasses (22·8%), while the native species were largely woody species (67·0%). However, in our study, all plants were of a similar growth form (i.e. woody shrubs and small trees). Kueffer et al. (2008) previously showed that differences in SLA between natives and invasive species depended on growth form and habitat; invasive and native woody species differed in SLA for pioneer and canopy species, but not for understorey species. This highlights the importance of considering plant life form or habitat when key traits between native and invasive species are being compared. Moreover, we found that both leaf fibre and leaf PCII were influenced by interactive effects of invasion status and N fixation ability; both variables were affected by N fixation ability only for native species, and by invasion status only for non N-fixers (although only at = 0·071 in the case of fibre). This highlights that differences in particular traits (in this case leaf fibre concentration and those traits that drive variation in PCII) between native and invasive species, or between N-fixing and non N-fixing species, may exist for only some subsets of the flora and not others.

Further, invasive species in our study had higher concentrations of condensed tannins and total phenolics in their leaves and litter than did the native species, which conflicts with our first hypothesis. Interestingly, the native plants in our system have almost no condensed tannins in their leaves and litter. This result suggests that the invasive species present in our study system may have evolved under conditions that select for the production of secondary metabolites to conserve nutrients (Hättenschwiler & Vitousek 2000) and prevent nutrient loss (e.g. through reduced herbivory; Coley et al. 1985; Bowen & Van Vuren 1997) while the native species have not. Thus, the invasive species would appear to be superior to native species both in terms of nutrient acquisition (leading to higher foliar N) and nutrient conservation (through having higher total phenolics and condensed tannins). This could potentially contribute to the relative success of invasive species in our study system, although the precise role of leaf traits in determining the success of woody invasive vs. native species in systems such as ours remains largely unexplored. Overall, therefore, we found only mixed support for the prediction of our first hypothesis that invasive species have traits consistent with acquisition rather than conservation of resources. There was little evidence of interactive effects between plant invasion status and N fixation ability on leaf and litter traits, meaning that trait differences between native and invasive species are consistent when comparing N-fixing and non N-fixing components of the flora.

What traits relate to palatability and decomposability?

In our study, we found that the traits most consistently related to both consumption and decomposition included leaf or litter C : P ratio and total phenolic concentration, and this would appear to contribute to the positive relationship, we found between palatability and decomposability (Table 4, Figs 4 and 5). However, a few additional traits also significantly influenced only foliage palatability (i.e. SLA and N : P ratio) or only litter decomposition (i.e. litter lignin concentration, Table 4). These results suggest that tissue stoichiometry, chemical defensive compounds, and, to some extent, structure, may act in concert to control leaf palatability by generalist herbivores and litter decomposition across contrasting species. Our results are therefore consistent with other studies that have shown several leaf or litter attributes, including nutrient concentrations, secondary metabolites and structural compounds, can be involved in determining either litter decomposability or foliage palatability (e.g. Grime et al. 1996; Cornelissen et al. 1999; Hättenschwiler & Vitousek 2000).

The positive relationship between leaf consumption and litter decomposition was still significant when only N-fixing species were considered but not when only non N-fixing species were considered, and the regression slopes between decomposability and palatability differed significantly for N-fixing and non N-fixing species (Fig. 4). This is because, although consumption and decomposition for the N-fixers were both strongly influenced by some of the same traits, i.e. the C : P ratios and total phenolic concentrations of plant tissues, these processes for non N-fixers were only weakly predicted by traits (and not by the same traits). Further, the spread of values of both consumption and decomposition data (Fig. 4) was greater for the N-fixers than the non N-fixers, which should result in a stronger relationship between decomposition and consumption for the N-fixers. Finally, the weak but significant interactive effect between N fixation ability and plant invasion status in determining the slope of the regression between consumption and decomposition was presumably driven by invasion status contributing to a steeper slope for N-fixing than non N-fixing species. In summary, these results demonstrate that the mechanisms determining the relationship between palatability and decomposability differs greatly between N-fixing and non N-fixing species, but little between native and invasive species.

Palatability and decomposability across functional groups

Nitrogen-fixing plants produced more palatable leaves than did non N-fixing plants, which is at least partially in support of our second hypothesis. However, litter decomposability was unaffected by N fixation ability. Further, there was no evidence that invasion status per se affects either palatability or decomposability, which is not consistent with this hypothesis. Thus, our results suggest that a priori assignment of plant species to functional groups based on their invasion status and N fixation ability are not necessarily useful for predicting key functional attributes of these species such as foliar palatability and litter decomposability (Wright et al. 2006). This is despite invasive and native species in our study differing in several leaf and litter traits that have been previously reported to drive both processes, including concentrations of N, total phenolics and condensed tannins (Swift, Heal & Anderson 1979; Tuomi 1992; Wardle et al. 2002). Of the two leaf or litter traits that most consistently affected both decomposability and palatability according to multiple regression analysis (Table 4), total phenolics differed between invasive and native species while the C : P ratio did not (Tables 1 and 2). Further, some other traits identified as being important for driving palatability and decomposability in this and the previous studies [i.e. N : P ratio and SLA for palatability (see also Cornelissen et al. 1999); litter lignin concentration for decomposability (see also Berg & McClaugherty 2003)] did not differ between invasive and native species. This pattern was also shown by ordination analyses; plant invasive status and N fixation ability only differed along PC II (Tables 1 and 2), but PC I had significant correlations with palatability and decomposability whereas PC II did not (Table 3). In addition, P may be a more critical limiting factor than N for ecological processes such as herbivory and decomposition (Forsyth, Richardson & Menchenton 2005; Wardle et al. 2009) in this ecosystem. As such, although N in plant tissues distinguishes invasive from native species or N-fixing from non N-fixing species, the availability of P (and therefore tissue C : P ratio) was more important in determining leaf palatability and litter decomposability.

Conclusions

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

Our results have several implications. First, they show that native and invasive species can be distinguished by particular leaf or litter traits, but that both foliage palatability and litter decomposability were not different between the two groups of species. This is because the leaf functional traits that are associated with plant invasion status (notably N) are not necessarily consistent with those that drive differences among plant species in foliage palatability and litter decomposability (notably C : P ratio and total phenolics). Secondly, N fixation ability has much stronger effects on ecological processes than does invasion status, presumably because of fundamental differences in the foliar nutrition, chemistry and structure between N-fixing and non N-fixing plants. Thirdly, they suggest that traits used for assignment of species to plant functional groups need not necessarily be the same as those that are important in driving ecological processes like herbivory and decomposition. Thus, our results highlight that interspecific differences in particular foliar and litter traits, rather than plant functional group memberships based on invasion status and N fixation ability, are more important in predicting ecological processes like herbivory and decomposition, even when those attributes are different to those that are routinely used to assign plant species to functional groups.

Acknowledgements

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

We thank Karen Boot, Gaye Rattray and Renske Terhürne for laboratory work, Pablo Manzano, Peter Bellingham, Rowan Buxton, Larry Burrows for advising and helping to collect leaf samples, Ian Dickie, Robinne Weiss, Stewart Oliver, Romina Rader, Gary Barker, Ellen Cieraad for helping to collect slugs used for the feeding trial, Hugh Gourlay and Ellen Cieraad for the incubator for the feeding trial, Peter Heenan for helping to identify the species, Sarah Richardson for very helpful comments on the manuscript, and the Palmerston North Chemistry Laboratory, Landcare Research for chemical analyses of the samples. This work was supported by the New Zealand Foundation for Research, Science and Technology (Ecosystem Resilience Outcome Based Investment), the Japan Society for Promotion of Science Research Fellowship for Young Scientists to HK and the Global Centre of Excellence Program ‘Eco-Risk Asia’ at Yokohama National University.

References

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

Supporting Information

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

Table S1. Leaf and litter traits for each native species.

Table S2. Leaf and litter traits for each exotic species.

Table S3. Correlation coefficients between PCI and PCII and each leaf trait.

Table S4. Correlation coefficients between PCI and PCII and each litter trait.

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

FilenameFormatSizeDescription
FEC_1676_sm_FEKurokawaST.doc168KSupporting info item

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