1. Ungulate herbivores may greatly alter the functional composition of plant communities through their selection of leaves with high nutrient concentrations and low structural or chemical defences. These impacts may have broader implications for ecosystem functions via changing rates of litter decomposition and the relative abundance of foliar traits. However, it is unlikely that ungulate impacts will be consistent across different vegetation types and different environmental conditions.

2. We compared shifts in abundance-weighted trait means and functional diversity indices in 88 pairs of plots fenced to exclude ungulate herbivores (exclosures) and adjacent control plots for eight foliar traits related to ungulate palatability in forest ecosystems. The stem density and basal area per species were both used to weight trait means and functional diversity indices, since these are proportional to recruitment and litter production, respectively. We also regressed exclosure effects on environmental and stand development variables to test whether context influenced the impact of ungulate exclusion.

3. There was a consistent, significant trend across seven of the eight leaf traits examined for increased palatability inside exclosure plots relative to control plots when trait means were weighted by the number of stems per species, but not when weighted by basal area. This suggests that ungulate exclosure consistently increased recruitment of palatable species, but is unlikely to have yet affected mean litter quality. Exclosure effects were linked to stand development variables, with effects being greatest in disturbed stands (where total basal area decreased and the number of stems increased), and least in stands undergoing competitive thinning.

4. Synthesis. The consistency of exclosure effects indicates that ungulate impacts on plant communities in New Zealand’s indigenous forest may be reversible at zero deer density, while their dependence on stand development suggests that the ungulate control impacts may be greatest in seral shrubland or forest communities.

Introduction

Ungulate herbivores alter the functional composition and diversity of plant communities through their selection of plants with high nutrient concentrations in leaves and low resource allocation to structural or chemical defences (e.g. Coley 1987; Augustine & McNaughton 1998; Forsyth, Richardson & Menchenton 2005; Lloyd et al. in press). These ungulate-induced changes in plant communities may have wider consequences for ecosystem functions, such as litter decomposition, since the foliar traits which make plant species palatable to vertebrate herbivores, such as high nitrogen and low structural carbohydrate content, also influence the rate at which leaves decompose (e.g. Grime et al. 1996; Cornelissen et al. 1999; Pérez-Harguindeguy et al. 2000) and photosynthesize (Reich et al. 1998). It is unrealistic to expect these impacts to be consistent across all vegetation types and under varied environmental conditions (Bardgett & Wardle 2003; Bee et al. 2009). The importance of context for functional impacts of ungulate exclusion on plant communities has been little studied (but see Coomes et al. 2003; Bee et al. 2009), and research is required to identify the ecosystems that will be most impacted by ungulates and those where restoration of natural ecosystem processes through ungulate control is most likely.

While plant communities in most regions of the world have co-evolved with large mammalian herbivores, such herbivores were absent from New Zealand until the arrival of Europeans c. 1800 ad (though the New Zealand flora was subject to herbivory by large ratites – moa – until several centuries after human colonization at c. 1280 ad, Holdaway & Jacomb 2000; Bond, Lee & Craine 2004). The widespread introduction of ungulate herbivores (particularly deer and goats) has caused declines of New Zealand’s palatable plant species and impacted on many communities (e.g. Coomes et al. 2003). This study tests whether exclusion of ungulate herbivores alters the functional composition and functional trait diversity of New Zealand’s indigenous forest communities and whether context influences the effects of ungulate exclusion.

It is now well known that functional composition (e.g. aggregate functional trait values), such as the abundance-weighted mean of trait values of plant communities, may control ecosystem processes such as primary productivity (Shipley, Vile & Garnier 2006) and litter decomposition (Cornwell et al. 2008; Fortunel et al. 2009). There is also evidence that high functional trait diversity enhances ecosystem functions including productivity (Petchey, Hector & Gaston 2004) and litter decomposition (Scherer-Lorenzen 2008). As a consequence, changes in the functional composition and diversity of plant communities that result from ungulate herbivore invasions could have significant consequences for primary productivity, nutrient cycling and carbon storage (Pastor et al. 1993; Bardgett & Wardle 2003). For example, the exclusion of ungulate herbivores in New Zealand forests increased the abundance of species with palatable leaves, while litter produced by these species decomposed more rapidly than that of unpalatable species (Wardle, Bonner & Barker 2002). However, it remains unclear whether ungulate herbivores cause detectable shifts in the foliar trait composition of whole forest communities.

In organisms such as plants, where individuals of the same species can vary in size by several orders of magnitude, the relevance of functional composition can be highly dependent on the type of abundance measure used. For example, the impact on ecosystem processes exerted by a few individuals that comprise a large total biomass will generally be greater than that of many small individuals with a small total biomass (Diaz & Cabido 2001). Hence in forest communities functional composition is most relevant for ecosystem processes when trait values are weighted by species biomass, or abundance measures that are proportional to biomass. Similarly, the relevance of functional diversity values for indices that use abundance weightings will depend on the measure of abundance used. There are instances when weighting by the number of individuals, rather than biomass, may be informative. For example, shifts in functional composition weighted by the number of individuals for palatability traits in response to exclusion of ungulates may reveal community-level demographic impacts of ungulates (i.e. reduced recruitment or increased mortality of palatable species), even if there is no biomass-weighted response.

Exclosures are widely used to evaluate the impact of ungulates on vegetation structure and composition (e.g. Bråthen & Oksanen 2001; Kuijper et al. 2010), yet large-scale syntheses of exclosure studies have pointed to highly variable, or ‘idiosyncratic’ effects (e.g. across New Zealand forests, Wardle et al. 2001), without demonstrating the causes of this variability. For ungulate exclusion to reveal impacts on functional composition, palatable species selected by ungulates must be able to disperse back into the patch, and establish under local environmental conditions and in competition with established individuals or recruits of other species. In stressed environments, the recruitment of palatable species may be naturally low since traits that increase palatability are also related to reduced leaf longevity, and plants with long-lived leaves generally have a selective advantage in stressed environments (Grime 2001). Recruitment may also be impeded by intense competition for light from established individuals, since light competition is size-asymmetric (i.e. small individuals are disproportionately disadvantaged, e.g. Coomes & Allen 2007). We may then expect ungulate exclusion effects to be greatest in forest patches of low canopy density (e.g. in forest regenerating after catastrophic disturbance or mortality of large individuals), and lower when light competition is most intense (i.e. during the competitive thinning phase of stand development).

A nationwide system of 85 permanent paired exclosure and control plots in New Zealand’s indigenous forests has been maintained for up to 30 years in some areas. This system of exclosures presents a rare opportunity to test whether ungulate herbivore exclusion effects on the functional composition and functional diversity of foliar traits are consistent at a national scale. Since they cover a wide range of environmental conditions and different stages of stand development, these plots also allow us to test the relative influence of environment and stand development over ungulate removal effects.

Materials and methods

Exclosure plot selection and sampling

A total of 85 pairs of fenced exclosure plots and control plots were chosen for analyses (Fig. 1). The fences were designed to exclude introduced ungulates – mainly red deer (Cervus elaphus), but also other deer species (listed in Fraser, Cone & Whitford 2000), goats (Capra hircus) and wild pigs (Sus scropha domesticus). The plots were spread throughout New Zealand’s three main islands (Fig. 1), and were generally located in primary indigenous forest. Collectively these plots encompass the forest types where ungulate impacts have been of concern to conservation managers. While accurate data on historical ungulate densities are unavailable, ungulates have been present long enough, and have at some point attained high enough densities, to have a marked impact on forest understorey species composition (Fraser, Cone & Whitford 2000). Only plot pairs where the integrity of the fence had been maintained between measurements were included in analyses. All plots were surveyed at least twice following the methods detailed in Hurst & Allen (2007), with the time interval between measurements varying from 5 to 28 years and an average interval of 15 years. Analysing functional shifts between measurements provides additional power in detecting exclosure effects, since they are less likely to be confounded by differences in initial composition between paired exclosure and control plots. There was an average of 0.11 ± 0.08 years (ranging from 0 to 5) between the establishment of the exclosures and the initial diameter measurement. There was no evidence that variation in time between exclosure establishment and initial diameter measurement influenced observed exclosure effects. All exclosure plots and control plots were 20 × 20 m. The diameter and species identity of all stems ≥2.5 cm diameter at 1.35 m (d.b.h.) were recorded at each measurement.

Details are in the caption following the image — **Figure 1**
Open in figure viewer PowerPoint

Map of New Zealand showing location of exclosure plots and trait collection sites.

Foliar traits

We sampled foliar traits for 236 common indigenous vascular species from 14 localities (Fig. 1) which sampled the full latitudinal range of New Zealand’s three main islands. Fully expanded current-year foliage was collected from adult plants. Leaves were collected during the late austral summer (February–April) in 2006 and 2007. Leaf mass per unit area (LMA, g m⁻²) was estimated using fresh leaf area (Winfolia software and Epson flatbed scanner) and leaf dry mass (oven-dried at 60 °C for 48 h). This oven-dried material was analysed for leaf nutrients and secondary defence compounds following Forsyth, Richardson & Menchenton (2005). Total nitrogen (N) and phosphorus (P) were determined using the acid digest and colorimetric methods of Blakemore, Searle & Daly (1987). The % acid detergent fibre (ADF), % acid detergent cellulose (ADC) and % acid detergent lignin (ADL) were determined following Rowland & Roberts (1994), % condensed tannins following the vanillin method of Broadhurst & Jones (1978) and % total phenolics following Price & Butler (1977). On average, species for which trait data were collected contributed over 97% of the total basal area within the plots studied.

Statistical analyses

Ungulate exclusion effects on functional composition

The functional composition of plots was expressed using aggregate (i.e. abundance-weighted mean) foliar trait values (Shipley, Vile & Garnier 2006):

(eqn 1)

where X_i is the trait value for species i, p_i is its proportional abundance and S is species richness of the community.

For each plot, the aggregate trait values for the initial measurement were subtracted from those for the final measurement to give a shift in foliar trait mean. For each pair of plots, the control plot shift in aggregate trait value was subtracted from the corresponding value in the exclosure plot. For each pair of plots, and each foliar trait, this gave a value expressing the difference in foliar trait shifts between the exclosure and the control. A positive value indicates that the foliar trait mean of the exclosure plot has increased more or decreased less between measurements than in the control plot. A negative value indicates that the exclosure trait mean has decreased more or increased less than the control trait mean. As noted above, this approach is powerful in that examining the shift in foliar trait means, rather than differences between exclosure and control plots at a single point in time, minimizes possible confounding effects of initial differences in species composition between exclosure and control when the plots were established.

Two measures of abundance were used to weight aggregate foliar trait values. The number of stems for each species in each plot was used as an indicator of recruitment and mortality effects on trait means. A significant positive exclosure effect on trait means weighted by the number of stems indicates that species with certain traits are more likely to be recruited, or less likely to die, within exclosures relative to control plots. A significant positive exclosure effect on trait means weighted by basal area indicates that the biomass of species with certain traits increased more, or decreased less, in exclosure plots relative to control plots. The amount of litter produced by a species is scaled to basal area (Bray & Gorham 1964; Stohlgren 1988; Liao et al. 2006), so a significant exclosure effect on trait means weighted by basal area may indicate a change in litter quality in exclosures relative to control plots (although it is possible that the traits of live and freshly fallen leaves might differ considerably for some species). To allow for the multiple tests of significance, the number of significant results obtained for either abundance measure was compared to that expected at random using a binomial probability distribution.

Ungulate exclusion effects on functional diversity

Functional diversity values were calculated for all plots where two or more species occurred. Since functional diversity is composed of three primary components – functional richness, functional evenness and functional divergence (Mason et al. 2005) – several indices are required to quantify functional diversity completely. Here we use the range of trait values within a community as an index of functional richness (FR_i, Mason et al. 2005), which is the univariate equivalent of the convex hull volume of Cornwell, Schwilk & Ackerly (2006). The functional regularity index (FRO Mouillot et al. 2005) was used to quantify functional evenness and the ‘FDiv’ index (Villeger, Mason & Mouillot 2008) was used to estimate functional divergence. Each of these indices was applied to each foliar trait individually and mean values were taken across traits for each plot. The FD_var index (Mason et al. 2003) was used as a composite of functional richness and functional divergence. For indices that include abundance weightings (i.e. all except FR_i), values were calculated using both the number of stems or total basal area of each species in each plot. As for functional composition, ungulate exclusion effects on functional diversity were calculated by subtracting the shift in functional diversity observed in the control plot from the corresponding value in its paired exclosure plot. As for shifts in functional composition, the number of significant results obtained for each functional diversity index was compared to random expectation.

Significance of exclosure effects on foliar functional composition and functional diversity

A Monte Carlo permutation test analogous to a paired t-test was used to assess the significance of differences in functional composition and functional diversity values between paired exclosure and control plots. The test statistic is the summed difference between paired exclosure and control plots. The null model randomizes values between exclosure and control plots, but within the same pair of plots as in the observed data. For each test the permutation procedure was repeated 10 000 times to generate a random distribution of values for the test statistic. P-values were calculated as the proportion of simulations giving a value for the test statistic as or more extreme than the observed value and doubled to give a two-tailed test. Tests with 10 000 randomly generated data sets (using a uniform probability distribution) demonstrate that this simulation test provides appropriate Type I error rates. These permutation tests were used to test the significance of shifts in aggregate foliar trait values and univariate functional diversity index values for each functional trait and for multivariate functional diversity index values. A significant result indicates that shifts in functional composition or functional diversity values differ more consistently across pairs of exclosure and control plots than expected at random. Observed paired differences were expressed relative to expected values using the Standardized Effect Size (SES) of Gotelli & McCabe (2002)

(eqn 2)

where I_Obs is the value obtained from the observed data, I_Exp is the mean of the randomizations and σ_Exp is the standard deviation across the randomizations.

The SES index allows comparison of effect sizes while taking into account differences in power between separate analyses.

Modelling context dependence of exclosure effects on functional composition

Exclosure effects on functional composition were regressed on a variety of stand development and environmental variables. Change in mean stem diameter, the number of stems and total basal area between initial and final measurements were used as indicators of stand development, as used by Coomes & Allen (2007). The assumption was that an increase in the number of stems with a decrease in basal area and mean diameter is indicative of canopy opening as a result of the loss of large individuals and increased recruitment of smaller stems in response to reduced competition for light. An increase in basal area with a decrease in the number of stems and increase in mean diameter is likely to indicate competitive thinning, with greater probability of death for smaller individuals due to asymmetric competition for light. A mixture of soil, climatic and topographic variables were also used as predictors of exclosure effects: acid soluble soil phosphorus (ACIDP), soil calcium (CALCIUM), soil drainage (DRAINAGE), soil water deficit (DEFICIT), mean annual solar radiation (MAS), mean annual temperature (MAT), soil particle size (PSIZE), the ratio of rainfall to potential evapotranspiration (R:PET), slope and vapour pressure deficit (VPD). Slope and drainage were measured at the site of each pair of plots while values for all other environmental variables were obtained from the Land Environments New Zealand data base (LENZ, Leathwick et al. 2003). These soil and climatic variables were chosen since they may influence functional traits that occur within a community (Grime 2001) and impact on woody plant growth and distribution (Leathwick et al. 2003).

We applied two modelling approaches to assess the influence of predictor variables on ungulate exclusion impacts. The first involved comparing three different types of linear multiple regression. The stand development (StandDev) regression included change in mean diameter, total basal area and the number of stems in exclosure plots, as well as time between measurements. The environmental regression (Env) included all the environmental variables, and the full model (EnvStandDev) included all predictor variables. Goodness of fit of each model was assessed using the corrected Akaike Information Criterion (AIC_c) value (Burnham & Anderson 2002). An Akaike weight (w_i), which gives an estimate of the probability that a model gives the most parsimonious fit to the data amongst the models considered, was calculated for each model following the method described by Johnson & Omland (2004). A w_i value of >0.9 provides a high degree of confidence (i.e. greater than 90% confidence) that the model in question is the best-performing amongst those considered:

(eqn 3)

where, R is the number of models under consideration and Δ_i is the difference between the AIC_c value of model i and the minimum AIC_c value across all models. The sum of w_i values across all models adds to unity.

The second approach examined the relative explanatory power of individual predictors across models including all possible combinations of predictor variables using linear least-squares regression. To do this, an Akaike weight was calculated for each combination of predictor variables (as in 1, but across models with all possible combinations of predictor variables). A weight for each variable was then calculated by summing the weights across each of the models in which it was included:

(eqn 4)

where, w_k is the Akaike weight of predictor variable k and R_k is the number of models including k (this is the same for all variables when all possible combinations of variables are examined). This weighting provides a measure of the probability that a variable is in the set of the most parsimonious predictors. Since our aim was not to identify a single optimal model (i.e. a model with much lower AIC_c than all other models), but rather the variables most strongly related to exclosure effects, we report the w_k for each of the predictor variables across all possible combinations of these variables. For descriptive purposes we give the R² and significance of the model with the lowest AIC_c score.

Results

None of the foliar traits was very strongly correlated with any of the other traits across species, although there were relatively strong correlations between leaf N and P, and between fibre, cellulose and lignin (Table 1), suggesting that each trait provided at least some information independent of the others and that little would be gained by reducing the dimensionality of trait space through ordination (indeed the first two PCA axes only explained slightly more than 50% of the total variation in trait values). Phosphorus, phenols and tannins were the traits that varied most across species (Table 1), while fibre appeared to be relatively invariant, which conforms with existing literature (Sanger et al. 1996; Sariyildiz & Anderson 2005).

Table 1. (a) Pearson correlation coefficients between functional traits taken across species. (b) Summary statistics taken across species

(a)
	Nitrogen	Phosphorus	LMA	Phenols	Tannins	Fibre	Cellulose
Phosphorus	0.74
LMA	−0.48	−0.45
Phenols	−0.25	−0.25	0.18
Tannins	−0.25	−0.22	0.29	0.52
Fibre	−0.26	−.14	0.19	−0.21	0.16
Cellulose	−0.15	−0.02	0.01	−0.34	−0.12	0.69
Lignin	−0.23	−0.21	0.32	0.08	0.35	0.63	−0.09
(b)
Trait	Unit	Mean	SD	Min	Max	Max/Min	CV
Nitrogen	%DW	1.42	0.65	0.46	5	10	0.5
Phosphorus	%DW	0.13	0.07	0.02	0.58	29	0.5
LMA	g m⁻²	98.1	61.6	22.9	347.7	15	0.6
Phenols	mg g⁻¹	4.8	4.8	0.5	23.5	46	1
Tannins	mg g⁻¹	1.5	2.07	0.2	12	24	1.4
Fibre	%DW	39.8	10.7	15.3	70.2	5	0.3
Cellulose	%DW	22.33	7.49	3.1	44.2	15	0.3
Lignin	%DW	16.5	7.75	3.2	40.16	13	0.5

LMA refers to dry leaf mass per unit of leaf surface area.
%DW: Percentage of dry weight.

Consistency of ungulate exclusion effects on functional composition and functional diversity

Exclosure effects on functional composition (i.e. aggregate trait values) were significantly different from random for seven of the eight traits when weighting by the number of stems (Fig. 2). Observed differences in shifts for N and P content were more positive than expected at random, while those for LMA, phenolics, tannins, fibre and lignin were significantly more negative. These results indicate that the exclusion of ungulates resulted in increased recruitment of species with high nutrient content, low concentrations of herbivore defence compounds and low structural carbohydrate content. This in turn suggests that the exclusion of ungulate herbivores has had a nationally consistent positive demographic impact on populations of palatable plant species.

For functional composition weighted by basal area, no significant departure from random expectation was observed for any on the foliar traits (Fig. 2). This result suggests that there was no consistent effect of ungulate exclusion on the biomass of palatable species. This may, in turn, be seen as evidence that ungulate exclusion is unlikely to have had a nationally consistent effect on the average quality of leaf litter over the periods for which the exclosures have been maintained.

Exclosure effects on functional diversity were less clear (3, 4). Only FD_var weighted by the number of stems yielded more significant results than expected at random using binomial probability (for fibre, cellulose and lignin content, as well as the mean taken across traits, Fig. 3). This result indicates that both functional richness (the volume of functional trait space occupied) and functional divergence (the degree to which abundance is distributed amongst species to maximize abundance-weighted differences) increased more rapidly or decreased less rapidly in exclosures relative to control plots, since FD_var incorporates both functional richness and functional divergence. In more concrete terms, it seems that exclusion of ungulates increased FD_var through a combination of (i) increased colonization of species previously absent from exclosures, with very low values for fibre and cellulose and (ii) an increase in the number of stems of such species that were already present in exclosure plots. However, it seems that the impact of ungulate exclusion has less effect on functional diversity than it does on functional composition.

Context-dependence of ungulate exclusion effects

Stand development was a strong predictor of exclosure effects on shifts in functional composition weighted by the number of stems. The model based solely on stand development, without the inclusion of any environmental variables, received strong (w_i > 0.9) Akaike weight support for all traits (Table 2), and the regression was significant for five (LMA, phenolics, tannins, cellulose and lignin) of the eight traits examined. When functional composition was weighted by basal area, the stand development regression received strong support for all foliar traits, and was significant for all except fibre, cellulose and lignin (Table 2, the influence of individual predictors is outlined below and in Tables 3 and 4). These results indicate that stand development processes are much more important in regulating ungulate exclusion impacts on functional composition than environmental variables.

Table 2. Comparison of models explaining exclosure effects on functional composition (trait means) weighted by the number of stems and basal area

Functional trait	Model type	Weighted by number of stems			Weighted by basal area
Functional trait	Model type	AIC weight	R ²	P	AIC weight	R ²	P
Nitrogen	StandDev	0.999	0.087	0.106	1	0.243	<0.001
	Env	0	0.089	0.67	0	0.088	0.685
	EnvStandDev	0	0.175	0.365	0	0.283	0.025
Phosphorus	StandDev	1	0.107	0.089	1	0.116	0.035
	Env	0	0.041	0.971	0	0.08	0.751
	EnvStandDev	0	0.136	0.643	0	0.168	0.415
LMA	StandDev	1	0.245	>0.001	1	0.192	0.001
	Env	0	0.09	0.666	0	0.108	0.509
	EnvStandDev	0	0.285	0.023	0	0.253	0.061
Phenolics	StandDev	1	0.124	0.025	1	0.195	0.001
	Env	0	0.033	0.987	0	0.042	0.968
	EnvStandDev	0	0.165	0.435	0	0.232	0.108
Tannins	StandDev	1	0.210	0.001	1	0.181	0.002
	Env	0	0.122	0.397	0	0.096	0.613
	EnvStandDev	0.001	0.297	0.015	0	0.246	0.074
Fibre	StandDev	0.999	0.101	0.57	0.998	0.068	0.205
	Env	0.001	0.090	0.094	0.001	0.086	0.698
	EnvStandDev	0	0.145	0.582	0	0.12	0.754
Cellulose	StandDev	0.994	0.155	0.007	0.999	0.027	0.676
	Env	0	0.122	0.399	0.001	0.042	0.967
	EnvStandDev	0.007	0.291	0.019	0	0.066	0.979
Lignin	StandDev	1	0.162	0.005	0.995	0.106	0.052
	Env	0	0.093	0.64	0.004	0.143	0.252
	EnvStandDev	0	0.208	0.19	0.001	0.214	0.167

Akaike Information Criterion (AIC) weight gives an indication of the likelihood that a model is the most parsimonious of the three models being compared. The relationships between individual predictor variables and exclosure effects are outlined in Tables 3 and 4 The Env model contains only environmental variables, that StandDev model only stand development variables and the EnvStandDev model all variables from the previous two models. LMA refers to leaf dry mass per unit of surface area.

Table 3. Multi-model comparisons across models including all possible combinations of predictor variables for functional composition (trait means) weighted by the number of stems

	Nitrogen		Phosphorus		LMA		Phenolics		Tannins		Fibre		Cellulose		Lignin
	W _i	Dir	W _i	Dir	W _i	Dir	W _i	Dir	W _i	Dir	W _i	Dir	W _i	Dir	W _i	Dir
ΔNstems					0.989	−			0.928	−	0.845	-			0.616	−
ΔBA	0.832	−	0.920	−	0.956	+	0.974	+	0.981	+			0.995	+	0.844	+
ΔMeanDiam					0.589	+
TimeDiff															0.601	-
ACIDP
CALCIUM
DRAINAGE
DEFICIT	0.603	+											0.645	+
MAS													0.582	+
MAT									0.557	+
PSIZE
Rainfall:PET									0.412	+
SLOPE
VPD													0.466	-
	R ²	0.108	R ²	0.085	R ²	0.244	R ²	0.109	R ²	0.244	R ²	0.082	R ²	0.212	R ²	0.140
	P	0.008	P	0.006	P	<0.001	P	0.002	P	<0.001	P	0.007	P	<0.001	P	0.002

W _i refers to the summed Akaike Information Criterion weight of each predictor across all models in which it is included. Dir refers to the direction of the model-averaged regression coefficient. Results are presented only for variables with W_i values of over 0.5, and those included on the most parsimonious model (highlighted in bold). R² and P-values are for the most parsimonious model.

Table 4. Multi-model comparisons across models including all possible combinations of predictor variables for functional composition (trait means) weighted by basal area

	Nitrogen		Phosphorus		LMA		Phenolics		Tannins		Fibre		Cellulose		Lignin
	W _i	Dir	W _i	Dir	W _i	Dir	W _i	Dir	W _i	Dir	W _i	Dir	W _i	Dir	W _i	Dir
ΔNstems							0.782	+			0.698	−	0.345	−	0.586	−
ΔBA	0.998	−	0.759	−	0.958	+									0.643	+
ΔMeanDiam	0.567	−	0.512	−	0.604	+	0.778	+	0.791	+
TimeDiff			0.514	+	0.615	−	0.930	−	0.903	−
ACIDP															0.691	−
CALCIUM
DRAINAGE
DEFICIT															0.446	−
MAS
MAT					0.448	+
PSIZE									0.405	−
Rainfall:PET
SLOPE									0.531	+
VPD															0.570	+
	R ²	0.216	R ²	0.098	R ²	0.215	R ²	0.192	R ²	0.209	R ²	0.055	R ²	0.013	R ²	0.168
	P	0.000	P	0.013	P	0.000	P	0.000	P	0.001	P	0.027	P	0.294	P	0.009

W _i refers to the summed AIC weight of each predictor across all models in which it is included. Dir refers to the direction of the model-averaged regression coefficient. Results are presented only for variables with W_i values of over 0.5, and those included on the most parsimonious model (highlighted in bold). R² and P-values are for the most parsimonious model.

The importance of the stand development model was due to ungulate exclusion effects on functional composition being greatest in disturbed stands with high rates of recruitment (Tables 3 and 4). Although no single stand development variable had a dominant effect across all traits, collectively the results indicate that exclosure effects on functional composition were enhanced in disturbed stands, irrespective of the abundance measure used. Changes in the total basal area or the number of stems in exclosures were the strongest predictors of shifts in functional composition weighted by the number of stems (strongest Akaike weight support across all traits; Table 3). In each case an increase in the number of stems enhanced exclosure effects (i.e. was positively related to differences in nutrient content and negatively related to differences in structural carbohydrate content and defence compound concentrations), while an increase in basal area tended to dampen them (Fig. 5). For each trait, the model with the lowest AIC_c value was statistically significant, with the variation explained ranging from 8% to 24%. When weighting by basal area, results were similar, although change in mean stem diameter was also selected for several traits (with a decrease in mean diameter enhancing exclosure effects in each case), and for phenolics and tannins the time between measurements (TimeDiff) received the strongest support, while for lignin and P no single predictor was supported ahead of the others (Table 4). Also, the model with the lowest AIC_c value for cellulose was not significant.

Discussion

The clear effects observed in this study illustrate the potential power of foliar traits to reveal ecological processes. Our approach demonstrates that exclusion of ungulate herbivores from New Zealand forest ecosystems has caused a qualitatively consistent shift in functional composition towards foliar traits indicating increased palatability at a national scale and across a range of environments. Our results also demonstrate that this effect is greatly dependent on the measure of abundance used to weight these values. The clear results obtained for functional composition weighted by the number of stems demonstrates that ungulate exclusion causes an increase in the net recruitment of palatable species. Conversely, the lack of evidence for ungulate exclusion effects on functional composition weighted by basal area suggests that, over a period of one-to-several decades, ungulate exclusion may be unlikely to cause marked shifts in the average leaf litter quality of forest ecosystems. The magnitude of ungulate exclusion effects was influenced consistently by the stage of forest stand development. This implies that (i) even though the direction of effects for functional composition was qualitatively consistent, the magnitude of these effects was context-dependent, and (ii) although there was no consistent effect for functional composition weighted by basal area, in disturbed stands ungulate exclusion could cause an increase in mean leaf litter quality.

These findings provide new insights into the potential of ungulate herbivores to alter ecosystem processes in forest ecosystems and management efforts aimed at moderating these impacts. Firstly, it appears evident that ungulate herbivory is unlikely to cause a major shift in litter quality at the community level over a time span of several decades. However, the demographic impacts on palatable species may give cause for great concern in the long-term. It seems also that conservation managers may be able to identify forests where ungulate herbivore control efforts will have the greatest impact.

Ungulate impacts on litter quality in the medium to long-term

Past studies have noted a weak, but significant effect of exclosure impacts on litter chemical properties (and a large degree of variation across plots, Wardle et al. 2001) despite strong relationships between the foliar traits of browse-layer species and response to herbivory (e.g. Wardle, Bonner & Barker 2002). Although we did not measure litter quality directly, our results suggest that ungulate impacts on community-level litter quality are, on average, likely to be marginal at best. There are several reasons for not having observed a consistent exclosure effect on either functional composition weighted by basal area or mean litter quality. Firstly, most of the dominant canopy tree species in New Zealand’s indigenous forests live for at least several centuries and some for more than 1000 years (Wardle 1991), and ungulate impacts are generally observed in the browse layer (although ungulates can induce mortality of adult trees of some species by ring barking, e.g. Mark & Bayliss 1975). It may therefore be unrealistic to expect changes in basal area-weighted trait means at the time scale over which observations have been made. Secondly, in many habitats the dominant canopy species belong either to the genus Nothofagus, or the family Podocarpaceae, which, based on the trait data used in this study, tend to have relatively unpalatable leaves (i.e. low nutrient content and high structural carbohydrate content, e.g. Wardle, Bonner & Barker 2002). There may be limited potential for ungulates to alter the species composition of late-successional forests, even in the long-term, if dominant late-successional species tend to be avoided rather than selected. Nevertheless, New Zealand’s most numerous canopy species (Weinmannia racemosa, which commonly reaches 300 years in age) showed a significant increase in recruitment in response to ungulate exclusion in the exclosures studied (data not shown and as shown locally in previous studies, e.g. Husheer 2007), suggesting there may be some potential for its abundance in forest canopies to be reduced by ungulates in the long-term (though other studies have shown no net recruitment deficit for this species; Bellingham, Stewart & Allen 1999).

Linking short-term (i.e. decadal) fluctuations in deer populations and the different longevities of major woody species (decades to centuries) is challenging when interpreting the potential of ungulate herbivores to cause long-term shifts in the functional composition of late-successional forests. One approach that might be applied is using forest dynamics models such as SORTIE (Pacala et al. 1996) or LINKAGES (Post & Pastor 1996) and its New Zealand-specific derivative LINK-NZ (Hall & Hollinger 2000). A recent study (Didion, Kupferschmid & Bugmann 2009) provides an example of how a forest gap model may be used to estimate potential long-term ungulate impacts on the species composition of forest canopies. This suggests that gap models should also, in principle, be useful in estimating long-term ungulate effects on the functional composition of canopies in New Zealand forests, especially given that the national system of exclosures provides an excellent opportunity to quantify species-specific growth and mortality rates with and without ungulates.

Contextual dependence of ungulate impacts

Our finding that exclosure effects are greatest in disturbed stands suggests that competition may have a strong influence over the potential of ungulate control to alter the functional composition of forests. Impacts of ungulate exclusion on species composition are likely to be greatest in forests with open canopies, or in regenerating scrubland and early-successional forest (Coomes & Allen 2007), and there is evidence from New Zealand forests that this is the case (Allen, Payton & Knowlton 1984; Smale, Hall & Gardner 1995). However, our study appears to be the first demonstrating a consistent relationship between stand development processes and exclosure effects on a national scale. While the amount of variation explained by stand development variables is small, the consistency of the results across multiple traits suggests that they may be of use in prioritizing management efforts. If the aim of ungulate control is to maximize the change in functional composition in plant communities, then there may be a case for focussing control efforts on areas of regenerating forest rather than in mature forest. The vegetation of New Zealand has been greatly impacted by human disturbance – initially through burning by Polynesian settlers from c. 1280 ad (Wilmshurst et al. 2008; McWethy et al. 2009) and then by logging and forest clearance for farming and mining after European settlement. As a consequence, large areas are covered either by regenerating forest or non-forest vegetation in habitats that could support forest. One estimate suggests that over 2 million hectares of non-forest land under conservation management could potentially support forest (Mason et al., unpublished report), while large areas of contemporary forest are dominated by early- to mid-successional tree species. It seems then that there is great potential for targeting control efforts in seral vegetation.

The minor role of environmental variables may be due in some part to a lack of measurements made on the plots themselves for some variables. Previous work has suggested soil fertility is a key factor moderating ungulate impacts in New Zealand forests (Wardle 1984). Data for soil variables was available at 100 × 100 m resolution. Given evidence that soil fertility may vary markedly over less than 100 m (Richardson, Allen & Doherty 2008), it is likely that the soil data used here glossed over much of the true variation in soil fertility between plots. Collection of soil fertility data from the plots themselves may have revealed a greater influence of environment on exclosure effects.

Thinking outside the square: moving beyond exclosure experiments

Our results add to existing evidence (e.g. Allen, Payton & Knowlton 1984; Smale, Hall & Gardner 1995) that ungulates have caused a shift in New Zealand forest understoreys toward increased relative abundance of unpalatable species. However, paired exclosure and control plots represent an extreme contrast where ungulate densities are maintained at zero for extended periods in very small areas after having been browsed for at least 50 years. In reality, it is highly unlikely that ungulates will be eradicated over large areas of forest or shrubland. Consequently, exclosure effects are likely to overestimate the benefits of ungulate control. Coomes et al. (2003) noted various reasons that ungulate impacts on forest species composition and successional processes in New Zealand might, in some instances, be difficult or impossible to reverse through reduction of ungulate densities. Our results do not directly demonstrate that ungulate impacts on forests can be reversed, only that exclusion of ungulates alters shifts in foliar trait composition. A recent study showed a small increase in the number of seedlings and saplings in response to 44 years of deer control of species identified as palatable using expert opinion and empirical observations (Tanentzap et al. 2009), but no evidence was found for larger size classes. This suggests that, at best, the recovery of forest understorey vegetation at reduced, but non-zero, ungulate density requires more than 40 years. The absence of other studies quantifying the impact (or lack of impact) of ungulate control on the functional composition of forests makes it premature to claim that ungulate impacts really are irreversible at a national scale. Indeed, given that our results emphasize the importance of stand development for effects of ungulate exclusion, some of the evidence cited for the irreversibility of ungulate impacts might largely be explained by the context under which studies were conducted. Since control effects are likely to be most noticeable in regenerating vegetation, it may be best to focus studies on regenerating shrubland and early-successional forest. If ungulate control has no effect on the species and functional composition of such vegetation, then we may have to accept the suggestion of Coomes et al. (2003) that ungulate impacts on New Zealand’s forests are irreversible.

Acknowledgements

We thank Sean Husheer for his efforts in obtaining data from the various agencies that collected it, and the agencies that supplied these data. We also thank Elaine Wright for insights, co-operation and funding; William Lee and three anonymous reviewers for comments on a draft manuscript; Jenny Hurst, Anna Marburg and the National Vegetation Survey Databank (NVS) for assistance in extracting data from NVS; Richard FitzJohn, George Ledgard, Chris Morse, Dean Richards, Katiana Tamiana and Eddie Wheoki for field assistance collecting leaf samples; the Department of Conservation, the Piki te Aroha Marae Trust (Ngāpuhi) and the Tūhoe Tuawhenua Trust for permission to sample leaves on lands they administer; and Brian Daly for analysing leaf samples. Research was funded by the New Zealand Foundation for Research, Science and Technology Ecosystem Resilience Outcome-Based Investment (Contract ID C09X0502) the Department of Conservation’s Inventory and Monitoring Programme and Cross Departmental Research Pool Terrestrial Project 3.

References

Allen, R.B., Payton, I.J. & Knowlton, J.E. (1984) Effects of ungulates in structure and species composition in the Urewera forests as shown by exclosures. New Zealand Journal of Ecology, 7, 119–130.
Web of Science®Google Scholar
Augustine, D.J. & McNaughton, S.J. (1998) Ungulate effects on the functional species composition of plant communities: Herbivore selectivity and plant tolerance. Journal of Wildlife Management, 62, 1165–1183.
10.2307/3801981
Web of Science®Google Scholar
Bardgett, R.D. & Wardle, D.A. (2003) Herbivore-mediated linkages between aboveground and belowground communities. Ecology, 84, 2258–2268.
10.1890/02-0274
Web of Science®Google Scholar
Bee, J.N., Tanentzap, A.J., Lee, W.G., Lavers, R.B., Mark, A.F., Mills, J.A. & Coomes, D.A. (2009) The benefits of being in a bad neighbourhood: plant community composition influences red deer foraging decisions. Oikos, 118, 18–24.
10.1111/j.1600-0706.2008.16756.x
Web of Science®Google Scholar
Bellingham, P.J., Stewart, G.H. & Allen, R.B. (1999) Tree species richness and turnover throughout New Zealand forests. Journal of Vegetation Science, 10, 825–832.
10.2307/3237307
Web of Science®Google Scholar
Blakemore, L.C., Searle, P.L. & Daly, B.K. (1987) Methods for Chemical Analysis of Soils. New Zealand Soil Bureau Scientific Report, 80, 103–pp.
Google Scholar
Bond, W.J., Lee, W.G. & Craine, J.M. (2004) Plant structural defences against browsing birds: a legacy of New Zealand’s extinct moas. Oikos, 104, 500–508.
10.1111/j.0030-1299.2004.12720.x
Web of Science®Google Scholar
Bråthen, K.A. & Oksanen, J. (2001) Reindeer reduce biomass of preferred plant species. Journal of Vegetation Science, 12, 473–480.
10.2307/3236999
Web of Science®Google Scholar
Bray, J.R. & Gorham, E. (1964) Litter production in forests of the world. Advances in Ecological Research, 2, 101–157.
10.1016/S0065-2504(08)60331-1
Google Scholar
Broadhurst, R.B. & Jones, W.T. (1978) Analysis of condensed tannins using acidified vanillin. Journal of the Science of Food and Agriculture, 29, 788–794.
10.1002/jsfa.2740290908
CASWeb of Science®Google Scholar
Burnham, K.P. & Anderson, R.S. (2002) Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach. Springer-Verlag, New York.
CASGoogle Scholar
Coley, P.D. (1987) Interspecific variation in plant anti-herbivore properties - the role of habitat quality and rate of disturbance. New Phytologist, 106, 251–263.
10.1111/j.1469-8137.1987.tb04693.x
Web of Science®Google Scholar
Coomes, D.A. & Allen, R.B. (2007) Effects of size, competition and altitude on tree growth. Journal of Ecology, 95, 1084–1097.
10.1111/j.1365-2745.2007.01280.x
Web of Science®Google Scholar
Coomes, D.A., Allen, R.B., Forsyth, D.M. & Lee, W.G. (2003) Factors preventing the recovery of New Zealand forests following control of invasive deer. Conservation Biology, 17, 450–459.
10.1046/j.1523-1739.2003.15099.x
Web of Science®Google Scholar
Cornelissen, J.H.C., Pérez-Harguindeguy, N., Díaz, S., Grime, J.P., Marzano, B., Cabido, M., Vendramini, F. & Cerabolini, B. (1999) Leaf structure and defence control litter decomposition rate across species and life forms in regional floras on two continents. New Phytologist, 143, 191–200.
10.1046/j.1469-8137.1999.00430.x
Web of Science®Google Scholar
Cornwell, W.K., Cornelissen, J.H.C., Amatangelo, K., Dorrepaal, E., Eviner, V.T., Godoy, O., Hobbie, S.E., Hoorens, B., Kurokawa, H., Pérez-Harguindeguy, N., Quested, H.M., Santiago, L.S., Wardle, D.A., Wright, I.J., Aerts, R., Allison, S.D., Van Bodegom, P., Brovkin, V., Chatain, A., Callaghan, T.V., Díaz, S., Garnier, E., Gurvich, D.E., Kazakou, E., Klein, J.A., Read, J., Reich, P.B., Soudzilovskaia, N.A., Vaieretti, M.V. & Westoby, M. (2008) Plant species traits are the predominant control on litter decomposition rates within biomes worldwide. Ecology Letters, 11, 1065–1071.
10.1111/j.1461-0248.2008.01219.x
PubMedWeb of Science®Google Scholar
Cornwell, W.K., Schwilk, D.W. & Ackerly, D.D. (2006) A trait-based test for habitat filtering: convex hull volume. Ecology, 87, 1465–1471.
10.1890/0012-9658(2006)87[1465:ATTFHF]2.0.CO;2
PubMedWeb of Science®Google Scholar
Díaz, S. & Cabido, M. (2001) Vive la difference: plant functional diversity matters to ecosystem processes. Trends in Ecology & Evolution, 16, 646–655.
10.1016/S0169-5347(01)02283-2
Web of Science®Google Scholar
Didion, M., Kupferschmid, A.D. & Bugmann, H. (2009) Long-term effects of ungulate browsing on forest composition and structure. Forest Ecology and Management, 258, S44–S55.
10.1016/j.foreco.2009.06.006
Web of Science®Google Scholar
Forsyth, D.M., Richardson, S.J. & Menchenton, K. (2005) Foliar fibre predicts diet selection by invasive Red Deer Cervus elaphus scoticus in a temperate New Zealand forest. Functional Ecology, 19, 495–504.
10.1111/j.1365-2435.2005.00991.x
Web of Science®Google Scholar
Fortunel, C., Garnier, E., Joffre, R., Kazakou, E., Quested, H., Grigulis, K., Lavorel, S., Ansquer, P., Castro, H., Cruz, P., Doležal, J., Eriksson, O., Freitas, H., Golodets, C., Jouany, C., Kigel, J., Kleyer, M., Lehsten, V., Leps, J., Meier, T., Pakeman, R., Papadimitriou, M., Papanastasis, V.P., Quetier, F., Robson, M., Sternberg, M., Theau, J.P., Thébault, A. & Zarovali, M. (2009) Leaf traits capture the effects of land use changes and climate on litter decomposability of grasslands across Europe. Ecology, 90, 598–611.
10.1890/08-0418.1
CASPubMedWeb of Science®Google Scholar
Fraser, K.W., Cone, J.M. & Whitford, E.J. (2000) A revision of the established ranges and new populations of 11 introduced ungulate species in New Zealand. Journal of the Royal Society of New Zealand, 30, 419–437.
10.1080/03014223.2000.9517633
CASPubMedGoogle Scholar
Gotelli, N.J. & McCabe, D.J. (2002) Species co-occurrence: a meta-analysis of J. M. Diamond’s assembly rules model. Ecology, 83, 2091–2096.
10.1890/0012-9658(2002)083[2091:SCOAMA]2.0.CO;2
Web of Science®Google Scholar
Grime, J.P. (2001) Plant Strategies, Vegetation Processes, and Ecosystem Properties, 2nd edn. Wiley, Chichester, UK.
Google Scholar
Grime, J.P., Cornelissen, J.H.C., Thompson, K. & Hodgson, J.G. (1996) Evidence of a causal connection between anti-herbivore defence and the decomposition rate of leaves. Oikos, 77, 489–494.
10.2307/3545938
Web of Science®Google Scholar
Hall, G.M.J. & Hollinger, D.Y. (2000) Simulating New Zealand forest dynamics with a generalized temperate forest gap model. Ecological Applications, 10, 115–130.
10.1890/1051-0761(2000)010[0115:SNZFDW]2.0.CO;2
Web of Science®Google Scholar
Holdaway, R.N. & Jacomb, C. (2000) Rapid extinction of the moas (Aves: Dinorinthiformes): model, test, and implications. Science, 287, 2250–2254.
10.1126/science.287.5461.2250
CASPubMedWeb of Science®Google Scholar
Hurst, J.M. & Allen, R.B. (2007) The Recce Method for Describing New Zealand Vegetation: Field Protocols. Manaaki Whenua-Landcare Research, Lincoln, New Zealand.
Google Scholar
Husheer, S.W. (2007) Introduced red deer reduce tree regeneration in Pureora Forest, central North Island, New Zealand. New Zealand Journal of Ecology, 31, 79–87.
Web of Science®Google Scholar
Johnson, J.B. & Omland, K.S. (2004) Model selection in ecology and evolution. Trends in Ecology & Evolution, 19, 101–108.
10.1016/j.tree.2003.10.013
PubMedWeb of Science®Google Scholar
Kuijper, D.P.J., Cromsigt, J.P.G.M., Jędrzejewska, B., Miścicki, S., Churski, M., Jędrzejewski, W. & Kweczlich, I. (2010) Bottom-up versus top-down control of tree regeneration in the Białowieża Primeval Forest, Poland. Journal of Ecology, 98, 888–899.
10.1111/j.1365-2745.2010.01656.x
Web of Science®Google Scholar
Leathwick, J.R., Wilson, G.H., Rutledge, D., Wardle, P., Morgan, F.J., Johnston, K., McLeod, M. & Kirkpatrick, R. (2003) Land Environments of New Zealand. David Bateman, Auckland.
Google Scholar
Liao, J.H., Wang, H.H., Tsai, C.C. & Hseu, Z.Y. (2006) Litter production, decomposition and nutrient return of uplifted coral reef tropical forest. Forest Ecology and Management, 235, 174–185.
10.1016/j.foreco.2006.08.010
Web of Science®Google Scholar
Lloyd, K.M., Pollock, M.L., Mason, N.W.H. & Lee, W.G. (in press) Leaf trait-palatability relationships differ between ungulate species: evidence from cafeteria experiments using naïve tussock grasses. New Zealand Journal of Ecology, 34, 219–229.
Google Scholar
Mark, A.F. & Bayliss, G.T.S. (1975) Impact of deer on Secretary Island, Fiordland, New Zealand. Proceedings of the New Zealand Ecological Society, 22, 19–24.
Google Scholar
Mason, N.W.H., MacGillivray, K., Steel, J.B. & Wilson, J.B. (2003) An index of functional diversity. Journal of Vegetation Science, 14, 571–578.
10.1111/j.1654-1103.2003.tb02184.x
Web of Science®Google Scholar
Mason, N.W.H., Mouillot, D., Lee, W.G. & Wilson, J.B. (2005) Functional richness, functional evenness and functional divergence: the primary components of functional diversity. Oikos, 111, 112–118.
10.1111/j.0030-1299.2005.13886.x
Web of Science®Google Scholar
McWethy, D.B., Whitlock, C., Wilmshurst, J.M., McGlone, M.S. & Li, X. (2009) Rapid deforestation of South Islands, New Zealands, by early Polynesian fires. Holocene, 19, 883–897.
10.1177/0959683609336563
Web of Science®Google Scholar
Mouillot, D., Mason, W.H.N., Dumay, O. & Wilson, J.B. (2005) Functional regularity: a neglected aspect of functional diversity. Oecologia, 142, 353–359.
10.1007/s00442-004-1744-7
PubMedWeb of Science®Google Scholar
Pacala, S.W., Canham, C.D., Saponara, J., Silander, J.A., Kobe, R.K. & Ribbens, E. (1996) Forest models defined by field measurements: II. Estimation, error analysis and dynamics Ecological Monographs, 66, 1–43.
10.2307/2963479
Web of Science®Google Scholar
Pastor, J., Dewey, B., Naiman, R.J., McInnes, P.F. & Cohen, Y. (1993) Moose browsing and soil fertility in the boreal forests of Isle Royale National Park. Ecology, 74, 467–480.
10.2307/1939308
Web of Science®Google Scholar
Pérez-Harguindeguy, N., Díaz, S., Cornelissen, J.H.C., Vendramini, F., Cabido, M. & Castellanos, A. (2000) Chemistry and toughness predict leaf litter decomposition rates over a wide spectrum of functional types and taxa in central Argentina. Plant and Soil, 218, 21–30.
10.1023/A:1014981715532
CASWeb of Science®Google Scholar
Petchey, O.L., Hector, A. & Gaston, K.J. (2004) How to different measures of functional diversity perform? Ecology, 85, 847–857.
10.1890/03-0226
Web of Science®Google Scholar
Post, W.M. & Pastor, J. (1996) Linkages – an individual-based forest ecosystem model. Climatic Change, 34, 253–261.
10.1007/BF00224636
CASWeb of Science®Google Scholar
Price, M.L. & Butler, L.G. (1977) Rapid visual estimation of and spectrophotometric determination of tannin content of sorghum grain. Journal of Agricultural and Food Chemistry, 25, 1268–1273.
10.1021/jf60214a034
CASWeb of Science®Google Scholar
Reich, P.B., Walters, M.B., Ellsworth, D.S., Vose, J.M., Volin, J.C., Gresham, C. & Bowman, W.D. (1998) Relationships of leaf dark respiration to leaf nitrogen, specific leaf area and leaf life-span: a test across biomes and functional groups. Oecologia, 114, 471–482.
10.1007/s004420050471
CASPubMedWeb of Science®Google Scholar
Richardson, S.J., Allen, R.B. & Doherty, J.E. (2008) Shifts in leaf N : P ratio during resorption reflect soil P in temperate rainforest. Functional Ecology, 22, 738–745.
10.1111/j.1365-2435.2008.01426.x
Web of Science®Google Scholar
Rowland, A.P. & Roberts, J.D. (1994) Lignin and cellulose fractionation in decomposition studies using acid-detergent fibre methods. Communications in Soil Science and Plant Analysis, 25, 269–277.
10.1080/00103629409369035
CASWeb of Science®Google Scholar
Sanger, L.J., Cox, P., Splatt, P., Whelan, M.J. & Anderson, J.M. (1996) Variability in the quality of Pinus sylvestris needles and litter from sites with different soil characteristics: Lignin and phenylpropanoid signature. Soil Biology & Biochemistry, 28, 829–835.
10.1016/0038-0717(96)00056-9
CASWeb of Science®Google Scholar
Sariyildiz, T. & Anderson, J.M. (2005) Variation in the chemical composition of green leaves and leaf litters from three deciduous tree species growing on different soil types. Forest Ecology and Management, 210, 303–319.
10.1016/j.foreco.2005.02.043
Web of Science®Google Scholar
Scherer-Lorenzen, M. (2008) Functional diversity affects decomposition processes in experimental grasslands. Functional Ecology, 22, 547–555.
10.1111/j.1365-2435.2008.01389.x
Web of Science®Google Scholar
Shipley, B., Vile, D. & Garnier, E. (2006) From plant traits to plant communities: a statistical mechanistic approach to biodiversity. Science, 314, 812–814.
10.1126/science.1131344
CASPubMedWeb of Science®Google Scholar
Smale, M.C., Hall, G.M.J. & Gardner, R.O. (1995) Dynamics of kanuka (Kunzea ericoides) forest on south Kaipara spit, New Zealand, and the impact of fallow deer (Dama dama). New Zealand Journal of Ecology, 19, 131–141.
Web of Science®Google Scholar
Stohlgren, T.J. (1988) Litter dynamics in two Sierran mixed conifer forests. I. Litterfall and decomposition rates. Canadian Journal of Forest Research, 18, 1127–1135.
10.1139/x88-174
Web of Science®Google Scholar
Tanentzap, A.J., Burrows, L.E., Lee, W.G., Nugent, G., Maxwell, J.M. & Coomes, D.A. (2009) Landscape-level vegetation recovery from herbivory: progress after four decades of invasive red deer control. Journal of Applied Ecology, 46, 1064–1072.
10.1111/j.1365-2664.2009.01683.x
Web of Science®Google Scholar
Villeger, S., Mason, N.W.H. & Mouillot, D. (2008) New multidimensional functional diversity indices for a multifaceted framework in functional ecology. Ecology, 89, 2290–2301.
10.1890/07-1206.1
PubMedWeb of Science®Google Scholar
Wardle, D.A., Barker, G.M., Yeates, G.W., Bonner, K.I. & Ghani, A. (2001) Introduced browsing mammals in New Zealand natural forests: aboveground and belowground consequences. Ecological Monographs, 71, 587–614.
10.1890/0012-9615(2001)071[0587:IBMINZ]2.0.CO;2
Web of Science®Google Scholar
Wardle, D.A., Bonner, K.I. & Barker, G.M. (2002) Linkages between plant litter decomposition, litter quality, and vegetation responses to herbivores. Functional Ecology, 16, 585–595.
10.1046/j.1365-2435.2002.00659.x
Web of Science®Google Scholar
Wardle, J.A. (1984) The New Zealand Beeches Ecology, Utilization and Management. New Zealand Forest Service, Christchurch, New Zealand.
Google Scholar
Wardle, P. (1991) Vegetation of New Zealand. Cambridge University Press, Cambridge.
Google Scholar
Wilmshurst, J.M., Anderson, A.J., Higham, T.F.G. & Worthy, T.H. (2008) Dating the late prehistoric dispersal of Polynesians to New Zealand using the commensal Pacific rat. Proceedings of the National Academy of Sciences of the USA, 105, 7676–7680.
10.1073/pnas.0801507105
CASPubMedWeb of Science®Google Scholar

Citing Literature

Volume98, Issue6

November 2010

Pages 1422-1433

Stand development moderates effects of ungulate exclusion on foliar traits in the forests of New Zealand

Summary

Introduction