Skip to main content
NCBI home page
UKPMC Funders Author Manuscripts logoLink to UKPMC Funders Author Manuscripts
. Author manuscript; available in PMC: 2015 Apr 23.
Published in final edited form as: Perspect Plant Ecol Evol Syst. 2012 Dec 20;14(6):402–410. doi: 10.1016/j.ppees.2012.09.003

Nurse plant effects on plant species richness in drylands: the role of grazing, rainfall and species specificity

Santiago Soliveres 1,2,*, David J Eldridge 1, Frank Hemmings 1, Fernando T Maestre 2
PMCID: PMC4407966  EMSID: EMS63070  PMID: 25914602

Abstract

The outcome of plant-plant interactions depends on environmental (e.g. grazing, climatic conditions) and species-specific attributes (e.g. life strategy and dispersal mode of the species involved). However, the joint effects of such factors on pairwise plant-plant interactions, and how they modulate the role of these interactions at the community level, have not been addressed before. We assessed how these species-specific (life strategy and dispersal) and environmental (grazing and rainfall) factors affected the co-occurrence of 681 plant species pairs on open woodlands in south-eastern Australia. Species-specific attributes affected the co-occurrence of most species pairs, with higher co-occurrence levels dominating for drought-intolerant species. The dispersal mechanism only affected drought-tolerant beneficiaries, with more positive co-occurrences for vertebrate-dispersed species. Conversely, the percentage of facilitated species at the community scale declined under higher rainfall availabilities. A significant grazing × rainfall interaction on the percentage of facilitated species suggests that grazing-mediated protection was important under low to moderate, but not high, rainfall availabilities. This study improves our ability to predict changes in plant-plant interactions along environmental gradients, and their effect on community species richness, by highlighting that: 1) species-specific factors were more important than environmental conditions as drivers of a large amount (~30%) of the pairwise co-occurrences evaluated; 2) grazing and rainfall interaction drive the co-occurrence among different species in the studied communities, and 3) the effect of nurse plants on plant species richness will depend on the relative dominance of particular dispersal mechanisms or life strategies prone to be facilitated.

Keywords: competition, diversity, facilitation, herbivory, open woodland, plant-plant interactions, semi-arid, stress-gradient hypothesis

Introduction

Positive interactions among plants are crucial for maintaining species richness in plant communities, particularly in harsh environments (Hacker and Bertness, 1999; Callaway, 2007; Soliveres et al., 2011a). These positive (facilitatory) plant-plant interactions co-occur with negative (competitive) ones, and their relative importance and net effect on diversity depend strongly on the environmental conditions (Bertness and Callaway, 1994; Hacker and Bertness, 1999; Holmgren and Scheffer, 2010). An increasing body of research suggests that the effects of positive interactions on species richness at the community level are prevalent at moderate levels of stress (Hacker and Gaines, 1999; Michalet et al., 2006). However, such effects also depend upon the nature of, and interrelationships among, the different stressors present within the environmental gradients studied (Maestre et al., 2009; Smit et al., 2009).

The joint effects of herbivory and abiotic stress on plant-plant interactions have seldom been addressed, even though these two stressors commonly co-occur in nature and strongly shape plant communities (Ibañez and Schupp, 2001; Gómez-Aparicio et al., 2008; Sasaki et al., 2008; Suzuki et al., 2008). The effects of grazing on plant communities and plant-plant interactions are known to vary with water and nutrient status (e.g. Illius and O’Connor, 1999; Veblen, 2008). Some research suggests that the impact of grazing on plant performance can be extremely important in water-limited environments due to low vegetation productivity, which may cause a higher grazing pressure in the remaining palatable tissues (Illius and O’Connor, 1999). Protection against herbivory by nurse plants can, therefore, be crucial for the maintenance of diversity in water-limited environments (e.g. Rebollo et al., 2002; Graff et al., 2007; Smit et al., 2007; Soliveres et al., 2011b). However, other lines of inquiry suggest that the role of herbivory-mediated facilitation at the community level is important at moderate to high, but not at low, levels of water availability (Smit et al., 2009; Holmgren and Scheffer, 2010; but see Gómez-Aparicio et al., 2008). This is explained by the relatively low abundance of herbivores under such conditions that, together with the lower palatability and productivity of drought-adapted plants, would explain the relatively weak impact of herbivory on arid and semiarid plant communities regarding the prevalent harsh abiotic conditions (reviewed in Smit et al., 2009; see also Howard et al., 2012). The contrasting results and theories found in the literature, together with the elevated number of factors affecting plant-plant interactions, require that more empirical studies be undertaken to advance our understanding about the role of plant-plant interactions on community diversity along environmental gradients involving both biotic and abiotic stressors (Bertness and Shumway, 1993; Smit et al., 2009; Bulleri et al., 2011).

Irrespective of the environmental conditions, it is well known that species-specific features of particular species, such as their ecological traits, dispersal mechanisms or evolutionary relationships, determine, to a large degree, the magnitude and direction of the interactions among them (e.g. Liancourt et al., 2005; Gómez-Aparicio, 2008; Pueyo et al., 2010; Soliveres et al., in press). The response of plant-plant interactions to herbivory is also known to be highly dependent on the plant species involved (Callaway et al., 2000; Skelinen, 2008). The outcome of these interactions will depend largely on the tolerance to herbivory of the facilitated species, and on the ability of the nurse to moderate the effects of herbivores (Baraza et al., 2006; Vanderberghe et al., 2009). Alternatively, nurse plants not only protect neighbours against herbivory, but also enhance their recovery after grazing by providing higher levels of resources (Rand, 2004; Acuña-Rodriguez et al., 2006). This may also lead to species-specific responses depending on the identity of the limiting resource for each beneficiary species, and on how its availability is affected by both herbivory and nurse plants (Wise and Abrahamson, 2005; Soliveres et al., 2011b). To understand the role of plant-plant interactions on the maintenance of plant diversity, therefore, we need to know how these multiple environmental and species-specific attributes affect pairwise interactions and how this influences overall trends at the community level. Undoubtedly, this will “help in the development of general theory and in explaining apparently contradictory results found in the literature” (Brooker et al., 2008).

The main aim of our study was to determine how different levels of both rainfall availability and grazing pressure, and their interaction, affected plant-plant interactions. Specifically, we endeavoured to answer the following questions: 1) what is the relative importance of environmental factors (grazing, rainfall) and species-specific attributes as drivers of pairwise plant-plant interactions?, and 2) do grazing and rainfall modulate the effects of these interactions on plant species richness at the community level?

Methods

Study area and establishment of the rainfall and grazing gradients

The study was conducted within 27 open eucalypt woodland plots (50 m × 50 m) located in south-eastern Australia. The plots were positioned along a 1200 km west–east transect with average annual rainfall decreasing from western Sydney (33°33′43”S, 150°43′51”E), in the east, to Wentworth (34°06′40”S, 142°11′27”E), in the west. Rainfall is the most important abiotic stressor for these plant communities (e.g. Eldridge et al., 1991). Consequently, it was the predominant abiotic factor forming the environmental gradient studied, as it varied from 220 mm in Wentworth (mean minimum to mean maximum annual temperature: 10.3 to 23.7 °C) to 900 mm in western Sydney (mean minimum to mean maximum annual temperature: 11.3 to 24.4 °C). The plots had similar soil types (clay-loam soils with gradational profiles), and were located on landscapes of similar aspect and slope to minimize the influence of factors other than climate in our conclusions. Vegetation was characterised by the presence of large eucalypt trees (Eucalyptus populnea F. Muell., E. intertexta R. T. Baker, E. microtheca F., E. tereticornis S., E. moluccana Roxb. and E. crebra F. Muell.) and shrubs (Eremophila mitchellii Benth, Dodonaea viscosa (L.) Jacq., and Bursaria spinosa Cav.).

We established a fully factorial design with three levels of rainfall (western Sydney: 800-900 mm, Cobar: 350-450 mm, and Wentworth: 220-280 mm) and grazing (high, moderate and low), each replicated three times (n = 27 plots). We selected plots with a similar cover of both trees and shrubs to minimize any potential effects of such cover on plant species richness at the community level. The average cover of woody plants, determined from a 200 m transect in all the plots, was 30 ± 3% for trees and 11 ± 2% for shrubs (mean ± SE, n = 27). The three contrasted levels of grazing pressure were selected based on four attributes; 1) the standardized stocking rates, 2) land use, 3) distance to water points, and 4) vegetation composition. We converted the different animal densities to a common unit: DSE ha−1 (dry sheep equivalents), which rates was calculated using unpublished, long-term average sheep stocking rate figures from western NSW (Office of Environment and Heritage – OEH, Rangeland Assessment Program) and density data for different herbivores obtained from OEH and Department of Lands data (see Moles et al., 2011; Howard et al., 2012) and unpublished reports (Table 1). These variables have been repeatedly shown as good surrogates of herbivore activity in Australia and elsewhere (Andrew, 1988; Landsberg et al., 1999; Barnes, 2001). Plots in the low grazing regime were located in either National Parks or Nature Reserves at distances of more than 5 km from water, i.e. outside the grazing range of sheep and goats (Pringle and Landsberg, 2004). These sites were characterised by a plant community of high cover, abundance and biomass of native plant species. Among these species, palatable grasses were abundant, suggesting low grazing levels (Illius and ÓConnor, 1999). Sites in the moderate grazing category were either on individual grazing properties or on areas adjacent to major roads along which livestock are moved periodically (travelling stock reserves). We combined these different land management types because both are characterized by intermittent grazing. The travelling stock reserves are grazed intermittently because strict regulations require that livestock must move a minimum distance each day, thereby preventing overgrazing. On individual grazing properties, grazing by feral herbivores is restricted by the use of herbivore-proof fencing, and domestic herbivores are moved regularly between paddocks to avoid overgrazing and soil disturbance (time-controlled grazing; Johnston et al., 1996), which also promotes intermittent grazing. Plots under the high grazing regime were all pastoral properties that practice a set-stocking strategy, with livestock being grazed continuously in the same paddock. Sites were generally close to livestock watering points (< 2 km), a reasonable piosphere (i.e., area of high grazing impact around water points) for this region (Andrew, 1988; Pringle and Landsberg, 2004). Therefore, these sites are heavily grazed by sheep, cattle, and sometimes horses, and also have variable, but often large, populations of feral grazers (goats, rabbits). The stocking rates and management of the sampled plots has been consistent over the past 50 years. Our approach, therefore, does not underestimate the well-known mid- and long-term effects of herbivory on soil properties, seed banks or the existing plant composition prior to sampling (Milchunas and Lauenroth, 1993; Tongway et al., 2003).

Table 1.

Characteristics of the 27 study sites in relation to species richness, rainfall and grazing levels, the composition and density of grazers. Grazers: H = horses, C = cattle, S = sheep, G = goats, K = kangaroos, R = rabbits. Different animal densities have been converted to a common unit (dry sheep equivalents - DSE ha−1) using published and unpublished literature (see maintext).

Site Species richness Rainfall level Grazing level Grazers composition DSE ha−1
Caravans 37 High High H, C, R, K 20.6
Hoxton Canal 29 High High H, C, R, K 17.3
Swan’s nest 39 High High H, C, R, K 19.0
Cowra reserve 23 High Medium K, R 2.5
Quandialla 19 High Medium K, C 6.9
Rickaby’s 36 High Medium K, R 3.8
Cowra Hwy 32 High Low K, R 0.2
Hoxton woodland 34 High Low K 0.1
Millenium 36 High Low K 0.1
Etiwanda 22 Medium High S, K, R, G 3.8
Florida 29 Medium High S, K, R, G 4.3
Truganini 18 Medium High S, K, R, G 4.2
Cobar Common 25 Medium Medium K, G, R 2.9
Minesite 28 Medium Medium K, G, R 3.2
The Wire 31 Medium Medium K, G, R 2.9
Gilgunnia 40 Medium Low K 0.9
Wagga Tank 38 Medium Low K, G, R 2.4
Yathong 34 Medium Low K, G, R 2.4
LG1 12 Low High S, C, G, K, R 5.2
WG1 12 Low High S, C, G, K, R 4.2
WR1 11 Low High G, K, R 4.0
PG1 14 Low Medium S, K, G, R 3.5
WG3 16 Low Medium K, C, S, R 3.4
WR2 19 Low Medium K, R, G 1.2
KP1 16 Low Low K 1.0
KP2 19 Low Low K 1.0
LP1 13 Low Low K 1.0
Open in a new tab

Any study reporting differences in herbivore intensities across environmental gradients will invariably encounter differences in the composition of herbivores across the gradient. Similar changes on the composition of plants and the identity of nurse plant species are expected across the gradient. However, in most natural systems and ecological models, a different herbivory intensity is often (implicitly or not) related to a different composition of herbivores and plant species across environmental gradients (Milchunas et al., 1988; Cingolani et al., 2005). Changes in plant-herbivore composition, some including livestock and some others not, could surely affect our results. However, grazing by livestock and wild herbivores is a major driver of plant community dynamics in drylands, and their relative densities/compositions -together with their effect on plants- changes depending on the environmental conditions and land management (Holechek, 1988; Osem et al., 2002; Chamaillé-Jammes et al., 2007). Thus, we believe that the differential composition of both plants and herbivores across the studied gradients is a more realistic test of our hypotheses, rather than a limitation of our study.

Vegetation survey

We sampled three markedly different microsites (Open, Shrub, Tree) at each site. Open microsites were located more than 1 m away from the canopy edge of any shrub or tree. Shrub species (hereafter Shrub microsite) differed across the gradient, but all were inverse cone-shaped (e.g., Eremophila mitchellii, Dodonaea viscosa, Senna artemisioides [Gaudic. ex DC.] Randell, or Bursaria spinosa). The tree species sampled (hereafter Tree microsite) also changed across the gradient, but were always large trees from the genus Eucalyptus with the same general form and morphology, and were characterised by a deep litter layer beneath their canopies (see Appendix A). The two types of nurses studied here (trees and shrubs) ameliorate their abiotic environment by altering shade, temperature and soil fertility compared with Open microsites (Weltzin and Coughenour, 1990; Warnock et al., 2007). However, we studied the effects of trees and shrubs separately to broaden the generality of our conclusions, as these species differ in a number of ways (Appendix A), and their interactions with neighbours would therefore be expected to change along the grazing and rainfall gradients studied. The shrubs at our study sites were generally multi-stemmed, and their stem arrangement created a structure that restricts access to large grazers. Tree microsites were single-stemmed Eucalyptus spp. trees; their growth form allowed access of grazing animals to the area beneath their canopy. Indeed the canopies of these trees are often used as resting or camping sites for kangaroos and domestic livestock (Eldridge and Rath, 2002; Wilson et al., 2007). Large eucalypt trees deposit substantial quantities of litter (up to 2.2 t ha−1), and have high levels of organic matter (4.4%) beneath their canopies (Eldridge and Rath, 2002). These canopies are substantially larger (area up to 300 m2) and denser than those of the shrubs (area up to 10 m2); thus more shade is expected under tree than under shrub canopies. In spite of their positive effect on soil fertility and of their previously reported facilitative effects (Eldridge and Rath, 2002; Soliveres et al. 2011a), Eucalypt litter is known to have allelopathic effects on nutrient uptake and germination of some species (e.g. May and Ash, 1990). These allelopathic effects could substantially reduce their positive effects for some species and/or increase the species-specific nature of plant-plant interactions (Callaway, 2007).

We sampled thirty 0.5 m × 0.5 m quadrats within each microsite and site. The quadrats were distributed among at least three trees and five shrubs at each plot, which were of similar size and separated by distances of more than 2 m. In each quadrat we measured the cover and number of individuals of every perennial plant species. Annual plant composition in arid and semi-arid areas may substantially change through the year and among different years (Whitford 2002). Thus, we did not include annual species in our observational design to avoid confounding effects in the differences among study sites derived from sampling “incomplete” communities depending on the timing or the year of sampling. From these data we assessed interactions at both community and pairwise levels (see below). We acknowledge that plant density and cover (our surrogates of plant performance) might not be the best possible measurements for plant performance, and that it would have been preferable to measure directly either survival or growth. However, the recruitment of semi-arid plants, including those studied here, is mainly limited by seedling survival (Eldridge et al., 1991). Thus, the sole presence of an adult plant, and therefore the number of individuals, is a good surrogate of its survival, regardless of their size. Conversely, plant cover is related to productivity in semi-arid environments (Flombaum and Sala, 2009) and is a reasonable surrogate of plant performance, particularly if we consider the large amount of pairwise interactions assessed here and the logistic constrains of measuring directly plant performance in all of them.

Assessing plant-plant interactions and measuring species-specific factors

We divided all of the pairwise interactions included within our 27 plots into four “interaction-status” groups: Shrub-selective, Tree-selective, Non-selective (those facilitated by either shrubs or trees), and non-facilitated species. We performed χ2 tests by microsite pairs (Shrub vs Open and Tree vs Open; to assign each species to a given group). We did not consider the difference in cover between the different microsites in the plot since we sampled exactly the same number of quadrats and therefore the sample effort was identical. Those species with significantly more individuals than expected under Shrub or Tree microsites than in the open (i.e. a significant χ2 for Shrub vs Open but not for Tree vs Open or vice versa) were considered as Shrub or Tree-selective, respectively. Those species with significantly more individuals than expected under Shrub and Tree microsites than in the open (i.e. significant χ2 for Shrub vs Open and for Tree vs Open) were considered as Non-selective. The remaining species were considered as non-facilitated. When a given species had fewer than 15 individuals in a given plot, no statistical test was performed, and we considered as Shrub or Tree-selective those species found only within a Shrub or Tree microsite with no individuals in the open. When species with fewer than 15 individuals were found under the canopy of both nurses (Shrub and Tree), but not in the open, were referred to as a Non-selective. Among the species with fewer than 15 individuals, those with individuals in the open were referred to as non-facilitated, regardless of the number of individuals located under Shrub or Tree microsites. Since we had the same sampling effort for each microsite, it is very unlikely that these species with low abundances affected our results either by over- or underestimating the effect of nurse plants on the species richness of the community because they should have exactly the same probability of occurring under a given microsite by chance.

Although the degree of co-occurrence has successfully been used as a proxy of facilitation in numerous studies worldwide (e.g. Tewsbury and Lloyd, 2001; Valiente-Banuet et al., 2006), assuming that a higher level of co-occurrence is solely related to positive interaction between the species involved is risky. Plant spatial patterns are not only driven by these interactions but by a variety of different processes such as resource heterogeneity or dispersal (Alados et al., 2006). Regrettably, our observational study does not allow us to tease apart these different processes; however the results from empirical studies and mathematical models are in agreement in that both dispersal and facilitation/competition shifts are the major drivers of the spatial co-occurrence among species under semiarid conditions (e.g. Tirado and Pugnaire, 2005; Gómez-Aparicio, 2008; Pueyo et al., 2008; Soliveres et al. in press). We pooled different dispersal mechanisms into two classes: vertebrate-dispersed and dispersed by other vectors. This improved our interpretation of the co-occurrences by discriminating between 1) species found beneath a given nurse plant as a result of dispersal and facilitative mechanisms from 2) species occurring solely through facilitative processes. Plants dispersed by vertebrates (mostly birds and small mammals) were those with fleshy fruits or adhesive appendages (adapted to attach to fur, hair or feathers of vertebrates). These seeds are more likely to be dropped close to the canopies of woody plants, leading to a nucleation process (Dean et al., 1999). Conversely, seeds dispersed by wind, ants or gravity (classified in our study as “dispersed by other vectors”), do not contribute to this nucleation process. Wind-dispersed seeds are also likely to accumulate beneath nurses such as our Shrub microsites, which capture wind-blown sediments. Other patch-types (e.g. dead branches or grass patches) also retain these wind-dispersed seeds and. Although these other patch-types are smaller than Eucalyptus or shrub canopies, they are fixed permanently to the soil surface, making them equally effective as traps for wind dispersed seeds compared with Eucalyptus trunks or shrub canopies. We believe, therefore, that wind-dispersed seeds are less likely to co-occur more than expected by chance with woody plants by dispersal alone. The primary dispersal mechanisms of each of our species were determined by consulting literature sources, and were assigned to one of our two groups (Appendix B).

We used three categories to define the life strategy of all the species sampled: drought tolerator, competitor under wetter conditions, or generalist. We based our classification of “competitor” or “stress-tolerator” specifically on the adaption to water scarcity (drought tolerators) or the ability to compete with neighbours under wetter conditions (competitor) of each species. Since ecological traits indicating these adaptations for all species found in the 27 plots were unavailable in the literature, we developed a simple indicator based on our own observations. We defined drought tolerators as those species that: 1) were able to produce more biomass (indicated by higher cover in the quadrats) at the driest end of the gradient (see below), or 2) occurred only in the driest part of the gradient, suggesting that their recruitment is limited by factors other than low water availability. In the latter case we considered only those species with at least five occurrences in this extreme of the gradient. Competitor species were defined in a similar way: those with higher cover in the wettest part of the gradient or that only occurred in the wettest plots. We measured the limitation on biomass production (cover) under each rainfall level with χ2 tests, with average cover of each species by rainfall level as the response variable (considering cover as count data for biomass sections) and rainfall level as the predictor. Unfortunately, we could not exclude the possibility that some species were overlooked because they were grazed; however, we pooled the data of the three different levels of grazing studied to perform these χ2 tests. Thus, the possibility of overlooking grazing-sensitive species should wane as these species should appear at least in the ungrazed plots. Thus, a strong effect of herbivory on this analysis is unlikely. To avoid confusion with the facilitative effect of nurses per se, only individuals growing in the Open microsite were considered for both classifications. Any species that failed to meet these requirements (i.e. non-significant χ2 test and fewer than five individuals sampled in any of the rainfall extremes) was characterised as “generalist strategy” (see classification summary in Table 2). To increase the confidence in our classification, we compared it with data of specific leaf area (SLA) obtained for over 80 species from the published literature (Westoby and Wright, 2003; Wright et al., 2004; Taylor and Eamus, 2008) by using ANOVA (drought-tolerator/competitor/generalist classification as predictor and SLA as response variable). Specific leaf area is a widely used functional trait, known to be strongly related to drought and shade tolerances, growth rate and other important ecological functions (Westoby and Wright, 2003; Wright et al., 2004). High SLA values are found in shade tolerant, highly competitive plant species and low SLA values found in less competitive but drought-tolerant species (Wright et al., 2004). We found that SLA increased significantly from drought-tolerators (60.6 ± 10.7 cm2/g, n = 10) to generalists (112.1 ± 14.7 cm2/g, n = 53) to competitors (161.2 ± 17.4 cm2/g, n = 19; F2,81 = 7.42; P = 0.001). Our metric of plant-plant interactions at the community level was the percentage of facilitated species in each plot. Facilitated species were considered as those species with more occurrences under any nurse than in the open (i.e., Shrub-selective, Tree-selective and Non-selective species).

Table 2.

Summary of the different classifications used to define the pairwise interactions included in the study.

Classification Category Description
Interaction status Shrub-selective Species with more co-occurrences under shrubs than in open areas or under trees
Tree-selective Species with more co-occurrences under trees than in open areas or under shrubs
Non-selective Species with more co-occurrences either under shrubs or trees than in open areas
Non-facilitated Those species with significantly less occurrences under shrubs or trees than in the open, or with non-significant χ2 tests
Species-specific factors Dispersal mechanism Vertebrate Species with fleshy fruits or structures to assist dispersal by mammals
Other Species dispersed by vectors other than birds or mammals (ants, wind, gravity)
Life strategy Drought-tolerators Species only occurring in the driest part of the gradient, or with significantly more cover under the driest conditions
Competitors Species only occurring in the wettest part of the gradient, or with significantly more cover under the wettest conditions
Generalists Species not fulfilling the requirements to be either drought-tolerators or competitors
Open in a new tab

Statistical analyses

We used log-linear models to assess the role of environmental (grazing and rainfall) and species-specific (dispersal and life strategies) factors in determining the outcome of specialist and generalist pairwise plant-plant interactions. We had balanced replication within the different grazing and rainfall levels, but our data were based on natural occurrences of each life strategy and dispersal class, which are known to change along abiotic gradients (Grime, 1977; e.g. more drought-tolerant species were found under lower rainfall levels). Thus, we had to account for this in our analyses. For doing so, we analyzed changes in the relative frequency of facilitated/unfacilitated cases, rather than the absolute number. Separate analyses were performed for the relative frequency of Shrub-specialist, Tree-specialist, and Generalist interactions. As neither the life strategy nor SLA data could be considered completely independent of rainfall level (given that rainfall was used to determine life strategy, see above, and SLA significantly increased with water availability), we used separated log-linear models for species-specific and environmental variables to avoid multicollinearity. Since we were interested in assessing the different role played by species-specific and environmental factors as drivers of each of the interaction-types considered (Shrub- and Tree-selective and Non-selective species), we ran separated analyses for each of them. Thus, six different hierarchical log-linear models were used. Three of them included Grazing and Rainfall as categorical predictors, with separate models each for frequency of non-selective, or Shrub or Tree-selective as response variables. The remaining models included Strategy and Dispersal as categorical predictors, with the same three different interaction-status attributes (non-selective, or Shrub or Tree-selective) as response variables.

The total number of species (richness), and the percentage of facilitated species were analysed using a split-plot ANOVA, with Rainfall and Grazing as fixed factors. The first stratum considered Rainfall, and the second stratum the Grazing × Rainfall interaction. The ranges of grazing intensity changed substantially across the rainfall gradient (Table 1). Hence, to control for the different grazing ranges along the rainfall levels, we introduced the exact DSE ha−1 (dry sheep equivalents) as a covariate in the ANOVA. Diagnostic tests indicated that the data met assumptions of homoscedasticity and normality, and therefore were not transformed. ANOVA analyses were performed using Minitab, while Log-linear models were performed using SPSS 13.0 for Windows (SPSS Inc., Chicago, Illinois, USA).

Results

Species-specific and environmental factors as drivers of plant-plant interactions

The frequency of non-selective positive interactions (18% of the 681 pairwise interactions studied) changed significantly in relation to both species-specific (life strategy and dispersal) and environmental (rainfall and grazing) factors (Table 3). The frequency of positive generalist interactions increased linearly with decreasing rainfall or increasing grazing pressure (LOG-LINEAR: Rainfall χ2 = 8.97; DF = 2; P < 0.05, Grazing χ2 = 6.91; DF = 2; P < 0.05; Table 3A). These positive interactions were also less frequent for species following a “drought tolerator” strategy than for those following “competitor” or “generalist” strategies (χ2 = 14.52; DF = 2; P < 0.001), and for vertebrate-dispersed species than for those dispersed by other mechanisms (χ2 = 5.10; DF = 1; P < 0.05; Table 3B).

Table 3.

Percentage of positive interactions by category in relation to environmental (A) or species-specific (B) variables (N = 681 species). Life strategies are species more abundant in 1) high rainfall plots [Competitors], 2) low rainfall plots [Drought-tolerators] and 3) neither Competitors nor Drought-tolerators [Generalists]. Dispersal mechanism: species dispersed by birds or mammals [Vertebrate] or by other vectors, such as gravity, wind or ants [Other]). Shrub- and Tree-selective beneficiaries = species only found beneath Shrub or Tree canopies, respectively, and Non-selective beneficiaries = species found beneath Shrub and Tree canopies but not in the open.

Rainfall Grazing Shrub-selective Tree-selective Non-selective
A) Environmental variables High High 21.9 21.0 16.2
Moderate 15.4 17.9 15.4
Low 10.8 26.5 9.8
Moderate High 29.0 18.8 30.4
Moderate 23.0 35.6 14.9
Low 14.3 23.2 17.0
Low High 22.9 20.0 28.6
Moderate 24.5 30.6 26.5
Low 27.1 18.8 18.8
B) Species-specific variables Life strategy Dispersal mechanism
Competitor Other 17.9 18.3 19.1
Vertebrate 17.1 17.1 35.4
Drought-tolerator Other 19.1 17.6 8.8
Vertebrate 34.8 13.0 4.3
Generalist Other 23.5 30.4 15.2
Vertebrate 12.0 52.0 16.0
Open in a new tab

Shrub and Tree-selective interactions represented 25% and 19% of all 681 pairwise interactions studied, respectively. Shrub-selective positive interactions increased with grazing pressure (χ2 = 7.19; DF = 2; P < 0.05; Table 3A), and marginally increased with aridity (χ2 = 5.71; DF = 2; P = 0.058). However, they were only marginally affected by the Life strategy × Dispersal mechanism interaction (χ2 = 5.47; DF = 2; P = 0.065), with a slightly higher frequency of positive interactions for drought-tolerator beneficiaries when those were vertebrate dispersed. Conversely, tree-selective interactions were only significantly affected by the life strategy of the beneficiary species, with a lower frequency of positive interactions for drought-tolerator species (χ2 = 25.60; DF = 2; P < 0.001). Neither the dispersal mechanism of the beneficiary species, nor the levels of rainfall or grazing, affected the frequency of tree-selective positive interactions (P > 0.1 in all cases; Table 3A).

Community level responses to grazing and rainfall

Species richness differed significantly in relation to rainfall but was not significantly affected by grazing, the dry sheep equivalents (DSE), or the rainfall × grazing interaction (Fig. 1A). Richness was greatest under high or medium rainfall (~30 species by plot) and lowest under low rainfall (~15 species). Interestingly, the lack of effect of both grazing and the rainfall × grazing interaction was caused by the introduction of DSE as covariate, which was non-significant (P = 0.358). Even after controlling by the different ranges in grazing intensity across the rainfall/grazing categories, the percentage of facilitated species was significantly affected by rainfall, grazing and their interaction. A reduction in the percentage of facilitated species was found under high rainfall levels (from ~70% at moderate-low rainfall levels to ~50% under high rainfall; Fig. 1B). The increase in grazing pressure (either to moderate or high levels) increased the percentage of facilitated species, from ~55% at low grazing levels to ~69% at moderate and high grazing levels. However, this grazing effect disappeared under the highest rainfall availability. It must be noted that the widest range in grazing density (DSE) was precisely found under this high rainfall availability (Table 1).

Figure 1.

Figure 1

Open in a new tab

Species richness (A) and percentage (B) of facilitated species at the community level at different rainfall and grazing levels. Statistical significance of the ANOVAs performed are shown in each panel. Degrees of freedom [factor, error] were 2,6 for rainfall; 2,12 for grazing; and 4,12 for rainfall × grazing interaction. Abbreviations are: R = Rainfall, G = Grazing and R×G= Rainfall × Grazing interaction. Data are means ± SE (n = 3).

Discussion

Species specificity in plant-plant interactions

The low frequency of positive co-occurrences when beneficiary species were drought-tolerators is consistent with the literature (e.g. Hoffman, 1996; Gómez-Aparicio et al., 2005; Gross et al., 2010). These species are more likely to be outcompeted beneath nurse canopies by either other facilitated species or the nurse itself, and are less likely to benefit from greater water availability beneath the nurse canopy (Liancourt et al., 2005; Holmgren and Scheffer, 2010). Furthermore, most adaptations to drought stress, such as hairy and thick leaves, lignified tissue or short stature are also related to grazing-resistance, reducing the probability of these species to be facilitated even by grazing protection (Baraza et al., 2006; Smit et al., 2009).

Most of criticism surrounding the use of co-occurrence measures as surrogates of biotic interactions in plants is based on habitat-sharing or seed trapping. Our study is not free of this criticism, since we are not able to differentiate between positive co-occurrences caused by dispersal, dispersal + positive interactions, positive interactions, or other factors (i.e. soil heterogeneity) from our observational dataset. However, we partially reduced the uncertainty of our results by considering the dispersal mechanism of the species studied. Interestingly, positive co-occurrence among species pairs was not related to the dispersal strategy of the target species, with the exception of drought-tolerator species. This suggests that the positive co-occurrences found for the remainder life strategies (competitors or generalists) are caused by facilitative interactions, and not by the role of the nurse species as bird perches or refugia for mammals. As stated above, drought-tolerator species are much less likely to benefit from the ameliorated microclimatic conditions and, therefore, the higher frequency of positive co-occurrence for vertebrate-dispersed species in this group found seems to reflect the important role of nucleation processes, rather than true facilitative interactions, to determine co-occurrence patterns among drought-tolerant species and their nurses.

Tree-selective species (involved in over 19% of all pairwise interactions, 30% for positive interactions only) were only significantly affected by species-specific attributes, which also played a major role in the outcome of interactions involving non-selective beneficiary species. Therefore, our results suggest that species-specific factors can be more important than environmental variables in defining a large number of the pairwise interactions in the studied ecosystem. Interestingly, an elevated number of the species facilitated by either shrubs or trees did not correspond to a specific life strategy (Table 3B), suggesting that species-specific attributes other than the life strategy play an important role in such interactions. These could include indirect facilitatory mechanisms (Levine, 1999; Soliveres et al., 2011a), or differential responses to the allelopathic effects of eucalypts, litter type and amount, or the nature of mycorrhizal associations among the beneficiary species (Callaway, 2007). Unfortunately, our experimental design does not allow us to disentangle the multiplicity of possible mechanisms driving the outcome of these species-specific interactions. Irrespective of the particular mechanism(s) involved, our results question the notion of the overwhelming importance of environmental factors as drivers of plant-plant interactions, which has pervaded current literature. In line with previous work (Callaway, 2007; Brooker et al., 2008), our study suggests that species-specific factors, such as life strategy and dispersal mechanisms, should also be considered when studying the relationship between plant–plant interactions and environmental gradients. The reduction in the relative abundance of competitor species (sensu Grime, 1977) with declining rainfall has been suggested as the mechanism responsible for the waning of facilitation at the community level under extreme conditions (Michalet et al., 2006), and could explain the reduced number of facilitated species detected at the low rainfall end of our gradient.

Community level responses to grazing and rainfall

Grazing affected plant-plant interactions at the community level at low to moderate, but not high, water availabilities. The greater influence of grazing on facilitative interactions under low and moderate, but not high, rainfall conditions (Fig. 1B) is at odds with current predictions (Smit et al., 2009; Holmgren and Scheffer, 2010) and empirical research (Howard et al., 2012). In contrast to our results, previous research has demonstrated the crucial role of plant-plant interactions in the maintenance of plant species richness under grazing across a range of ecosystems and productivity levels (Callaway et al., 2000; Smit et al., 2007). However, other research clearly supports the notion of a higher grazing impact on plant performance, and therefore a higher importance of grazing-mediated facilitation, in water-limited environments (Illius and O’Connor, 1999; Rebollo et al., 2002; Gómez-Aparicio et al., 2008). It is difficult to explain these contrasting results in the literature, but we speculate that the interaction between grazing and abiotic stress could be idiosyncratic and depend on 1) the relative abundances of grazing-intolerant vs drought-intolerant species, and 2) the degree of protection against grazing provided by nurse canopies (Baraza et al., 2006; Gómez-Aparicio et al., 2008, Vanderberghe et al., 2009). Put simply, if grazing-intolerant species are abundant and nurses provide a high degree of protection against herbivory, then a high importance of grazing-mediated facilitation should be expected under any abiotic condition. Conversely, if plants are more sensitive to drought than to grazing, or the degree of grazing protection provided by the nurses is insufficient to shield them from higher herbivory pressure caused by the lack of alternative food sources under harsh conditions, then grazing-mediated protection should wane under extremely high levels of grazing pressure (e.g. Smit et al., 2007; Soliveres et al., 2011b).

It must be noted, however, that our results and conclusions regarding the grazing effect on plant-plant interactions could be strongly affected by the unequal range of grazing densities (dry sheep equivalents [DSE]) over the rainfall gradient and within/between our categorical grazing levels (Table 1). Arguably, we believe that separating grazing levels by these categorical classes instead of using the DSE per se is a reasonable procedure by three main reasons: 1) our classification was also strongly supported by the other indicators of grazing intensity used (land use, plant composition and distance to water points). All of them repeatedly demonstrated to be good indicators of ecologically important differences in grazing intensity affecting plant communities, 2) plant productivity is much higher under higher rainfall levels and, therefore, a much higher density of grazers can be sustained regarding the low productivity of water-limited environments (Holechek, 1988). Therefore, the effect of a relatively low grazing density under water-limited conditions is expected to have the same effect on plants than a much higher grazing density under wetter conditions (~21 vs ~13 DSE in the high vs low rainfall level, Table 1; Osem et al., 2002). Hence, we believe that our categorical classification reflects better than the raw DSE numbers the ecologically significant effect of the different grazing intensities on plants, and how this effect changes across different water availabilities, 3) to properly analyze the grazing × rainfall interaction using the DSE we would need to have the same ranges of DSE across rainfall levels, which did not happen. However, when we controlled for these different grazing densities introducing DSE as a covariate in the split-plot ANOVA, the results did not change substantially and therefore we can be confident about our results and conclusions.

Concluding remarks

To our knowledge, this is the first study to assess how both grazing and abiotic stress influence plant-plant interactions at the community-level in a terrestrial system. Though based solely on observational data, our work provides a first assessment of how the species-specific nature of some plant-plant interactions can affect the role they play at the community level along contrasting environmental conditions. We found strong species-specific components in all pairwise interactions, with positive interactions prevailing for less drought-adapted species or for vertebrate-dispersed species when these were tolerant to drought. These species-specific attributes were particularly important for Tree-selective and non-selective interactions, which accounted for ~37% of all the interactions studied, with some interactions independent of environmental conditions. However, and irrespective of the important role of species-specific attributes in defining the outcome of pairwise interactions, our results show shifts between plant-plant interactions at the community level depending on the environmental conditions. This could result from either the influence of the environmental conditions on the outcome of some pairwise interactions (non-selective and Shrub-selective beneficiaries) or, more likely, from changes in the relative abundances of the different life strategies along the gradient. These two factors could define the number of facilitated species at each plot, and therefore determine the response of the whole community.

Our study highlights two important issues: 1) the greater importance of species-specific attributes, regarding abiotic conditions, in defining the outcome of a large fraction of the pairwise co-occurrences evaluated, and 2) a greater dependence of grazing-mediated facilitative interactions under low to moderate, than under high, rainfall levels. Our results also suggest that the effects of plant-plant interactions on plant species richness are dependent on the relative predominance of their particular dispersal mechanisms or life strategies.

Supplementary Material

Supplementary Data

Acknowledgments

We thank James Val, Erin Roger and Madeleine Rankin for assistance with field work. Paul Theakston assisted with plant identification and many collaborating pastoralists allowed us access to their properties. SS and FTM were supported by the European Research Council under the European Community’s Seventh Framework Programme (FP7/2007-2013)/ERC Grant agreement n° 242658 (BIOCOM).

Footnotes

Supporting Information

Appendix A. Pictures of the different grazing-rainfall levels and the two nurse-types included in this study

Appendix B. Life strategy and dispersal mode of the species surveyed.

Reference list

  1. Acuña-Rodríguez IS, Cavieres LA, Gianoli E. Nurse effect in seedling establishment: facilitation and tolerance to damage in the Andes of central Chile. Revista Chilena Hist. Nat. 2006;79:329–336. [Google Scholar]
  2. Andrew MH. Grazing Impact in Relation to livestock watering points. Trends Ecol. Evol. 1988;3:336–339. doi: 10.1016/0169-5347(88)90090-0. [DOI] [PubMed] [Google Scholar]
  3. Baraza E, Zamora R, Hódar JA. Conditional outcomes in plant-herbivore interactions: neighbours matter. Oikos. 2006;113:148–156. [Google Scholar]
  4. Barnes RFW. How reliable are dung counts for estimating elephant numbers? Afr. J. Ecol. 2001;39:1–9. [Google Scholar]
  5. Bertness MD, Callaway RM. Positive interactions in communities. Trends Ecol. Evol. 1994;9:191–193. doi: 10.1016/0169-5347(94)90088-4. [DOI] [PubMed] [Google Scholar]
  6. Brooker RW, Kikvidze Z, Pugnaire FI, Callaway RM, Choler P, Lortie CJ, Michalet R. The importance of importance. Oikos. 2005;109:63–70. [Google Scholar]
  7. Brooker RW, Maestre FT, Callaway RM, Lortie CJ, Cavieres LA, Kunstler G, Liancourt P, Tielbörger K, Travis JMJ, Anthelme F, Armas C, Coll L, Corcket E, Delzon S, Forey E, Kikvidze Z, Olofsson J, Pugnaire FI, Saccone P, Schiffer K, Seifan M, Touzard B, Michalet R. Facilitation in plant communities: the past, the present, and the future. J. Ecol. 2008;96:18–34. [Google Scholar]
  8. Bulleri F, Cristaudo C, Alestra T, Benedetti-Cecchi L. Crossing gradients of consumer pressure and physical stress on shallow rocky reefs: a test of the stress-gradient hypothesis. J. Ecol. 2011;99:335–344. [Google Scholar]
  9. Callaway RM. Positive Interactions and Interdependence in Plant Communities. Springer; New York: 2007. [Google Scholar]
  10. Callaway RM, Kikodze D, Kikvidze Z. Facilitation by unpalatable weeds may conserve plant diversity in overgrazed meadows in the Caucasus Mountains. Oikos. 2000;89:275–282. [Google Scholar]
  11. Chamaillé-Jammes S, Valeix M, Fritz H. Managing heterogeneity in elephant distribution: interactions between elephant population density and surface water availability. J. Appl. Ecol. 2007;44:625–633. [Google Scholar]
  12. Cingolani AM, Posse G, Collantes MB. Plant functional traits, herbivore selectivity and response to sheep grazing in Patagonian steppe grasslands. J. App. Ecol. 2005;42:50–59. [Google Scholar]
  13. Dean WRJ, Milton SJ, Jeltsch F. Large trees, fertile islands, and birds in arid savanna. J. Arid Environ. 1999;41:61–78. [Google Scholar]
  14. Eldridge DJ, Rath D. Hip holes: kangaroo resting sites enhance the physical and chemical environment of woodland soils. Aust. Ecol. 2002;27:527–536. [Google Scholar]
  15. Eldridge DJ, Westoby M, Holbrook KG. Soil-surface characteristics, microtopography and proximity to mature shrubs: effects on survival of several cohorts of Atriplex vesicaria seedlings. J. Ecol. 1991;79:357–364. [Google Scholar]
  16. Flombaum P, Sala OE. Cover is a good predictor of biomass in arid systems. J. Arid Environ. 2009;73:597–598. [Google Scholar]
  17. Gómez-Aparicio L. Spatial patterns of recruitment in a Mediterranean tree (Acer opalus subsp. granatense): linking the fate of seeds, seedlings, and saplings in heterogeneous landscapes at different scales. J. Ecol. 2008;96:1128–1140. [Google Scholar]
  18. Gómez-Aparicio L, Valladares F, Zamora R, Quero JL. Response of tree seedlings to the abiotic heterogeneity generated by nurse shrubs: an experimental approach at different scales. Ecography. 2005;28:757–768. [Google Scholar]
  19. Gómez-Aparicio L, Zamora R, Castro J, Hódar JA. Facilitation of tree saplings by nurse plants: microhabitat amelioration or protection against herbivores? J. Veg. Sci. 2008;19:161–172. [Google Scholar]
  20. Graff P, Aguiar MR,, Chaneton EJ. Shifts in positive and negative plant interactions along a grazing intensity gradient. Ecology. 2007;88:188–199. doi: 10.1890/0012-9658(2007)88[188:sipanp]2.0.co;2. [DOI] [PubMed] [Google Scholar]
  21. Grime JP. Plant strategies, vegetation processes, and ecosystem properties. John Wiley and Sons; Chichester, UK: 1977. [Google Scholar]
  22. Gross N, Liancourt P, Choler P, Suding KN, Lavorel S. Strain and vegetation effects on limiting resources explain the outcomes of biotic interactions. Persp. Plant Ecol., Evol. Syst. 2010;12:9–19. [Google Scholar]
  23. Hacker SD, Bertness MD. The role of positive interactions in the plant species diversity of salt marsh plant communities. Ecology. 1999;80:2064–2073. [Google Scholar]
  24. Hoffmann WA. The effects of cover and fire on seedling establishment in a neotropical savanna. J. Ecol. 1996;84:383–393. [Google Scholar]
  25. Holechek JL. An approach for setting the stocking rate. Rangelands. 1988;10:10–14. [Google Scholar]
  26. Holmgren M, Scheffer M. Strong facilitation in mild environments: the stress gradient hypothesis revisited. J. Ecol. 2010;98:1269–1275. [Google Scholar]
  27. Howard KSC, Eldridge DJ, Soliveres S. Positive effects of shrubs on plant species diversity do not change along a gradient in grazing pressure in an arid shrubland. Basic App. Ecol. 2012;13:159–168. [Google Scholar]
  28. Ibáñez I, Schupp EW. Positive and negative interactions between environmental conditions affecting Cercocarpus ledifolius seedling survival Oecol. 2001;129:543–550. doi: 10.1007/s004420100757. [DOI] [PubMed] [Google Scholar]
  29. Illius AW, O’Connor TG. On the relevance of nonequilibrium concepts to semi-arid grazing systems. Ecol. Appl. 1999;9:798–813. [Google Scholar]
  30. Johnston PW, McKeon GM, Day KA. Objective ‘safe’ grazing capacities for southwest Queensland Australia: development of a model for individual properties. Rang. J. 1996;18:244–258. [Google Scholar]
  31. Kitzberger T, Steinaker DF, Veblen TT. Effects of climatic variability on facilitation of tree establishment in northern patagonia. Ecology. 2000;81:1914–1924. [Google Scholar]
  32. Körner C. Limitation and stress - always or never? J. Veg. Sci. 2003;14:141–143. [Google Scholar]
  33. Landsberg J, Lavorel S, Stol J. Grazing response groups among understorey plants in arid rangelands. J. Veg. Sci. 1999;10:683–696. [Google Scholar]
  34. Levine JM. Indirect facilitation: evidence and predictions from a riparian community. Ecology. 1999;80:1762–1769. [Google Scholar]
  35. Liancourt P, Callaway RM, Michalet R. Stress tolerance and competitive-response ability determine the outcome of biotic interactions. Ecology. 2005;86:1611–1618. [Google Scholar]
  36. Maestre FT, Bautista S, Cortina J. Positive, negative, and net effects in grass-shrub interactions in Mediterranean semi-arid grasslands. Ecology. 2003;84:3186–3197. [Google Scholar]
  37. Maestre FT, Callaway RM, Valladares F, Lortie CJ. Refining the stress-gradient hypothesis for competition and facilitation in plant communities. J. Ecol. 2009;97:199–205. [Google Scholar]
  38. Maestre FT, Cortina J. Do positive interactions increase with abiotic stress? A test from a semi-arid steppe. Proc. R. Soc. London, B. 2004;271:S331–S333. doi: 10.1098/rsbl.2004.0181. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Maestre FT, Valladares F, Reynolds JF. Is the change of plant-plant interactions with abiotic stress predictable? A meta-analysis of field results in arid environments. J. Ecol. 2005;93:748–757. [Google Scholar]
  40. May FE, Ash JE. An assessment of the allelopathic potential of Eucalyptus. Aust J Bot. 1990;38:245–254. [Google Scholar]
  41. Michalet R, Brooker RW, Cavieres LA,, Kikvidze Z, Lortie CJ, Pugnaire FI, Valiente-Banuet A, Callaway RM. Do biotic interactions shape both sides of the humped-back model of species richness in plant communities. Ecol. Lett. 2006;9:767–773. doi: 10.1111/j.1461-0248.2006.00935.x. [DOI] [PubMed] [Google Scholar]
  42. Milchunas DG, Lauenroth WK. Quantitative effects of grazing on vegetation and soils over a global range of environments. Ecol. Monogr. 1993;63:327–366. [Google Scholar]
  43. Milchunas DG, Noy-Meir I. Grazing refuges, external avoidance of herbivory and plant diversity. Oikos. 2002;99:113–130. [Google Scholar]
  44. Milchunas DG, Sala O, Lauenroth WK. A generalized model of the effects of grazing by large herbivores on grassland community structure. Am. Nat. 1988;132:87–106. [Google Scholar]
  45. Osem Y, Perevolotsky A, Kigel J. Grazing effect on diversity of annual plant communities in a semi-arid rangeland: interactions with small-scale spatial and temporal variation in primary productivity. J. Ecol. 2002;90:936–946. [Google Scholar]
  46. Pringle HJR, Landsberg J. Predicting the distribution of livestock grazing pressure in rangelands. Aust. Ecol. 2004;29:31–39. [Google Scholar]
  47. Rand TA. Competition, facilitation, and compensation for insect herbivory in an annual salt marsh forb. Ecology. 2004;85:2046–2052. [Google Scholar]
  48. Rebollo S, Milchunas DG, Noy-Meir I, Chapman PL. The role of a spiny plant refuge in structuring grazed shortgrass steppe plant communities. Oikos. 2002;98:53–64. [Google Scholar]
  49. Sasaki T, Okayasu T, Shirato Y, Jamsran U, Okubo S, Takeuchi K. Can edaphic factors demonstrate landscape-scale differences in vegetation responses to grazing? Plant Ecol. 2008;194:51–66. [Google Scholar]
  50. Skelinen A. Herbivore and neighbour effects on tundra plants depend on species identity, nutrient availability and local environmental conditions. J. Ecol. 2008;96:155–165. [Google Scholar]
  51. Smit C, Vanderberghe C, den Ouden J, Müller-Schärer H. Nurse plants, tree saplings and grazing pressure: changes in facilitation along a biotic environmental gradient. Oecol. 2007;152:265–273. doi: 10.1007/s00442-006-0650-6. [DOI] [PubMed] [Google Scholar]
  52. Smit C, Rietkerk M, Wassen MJ. Inclusion of biotic stress (consumer pressure) alters predictions from the stress gradient hypothesis. J. Ecol. 2009;97:1215–1219. [Google Scholar]
  53. Soliveres S, Eldridge DJ, Maestre FT, Bowker MA, Tighe M, Escudero A. Microhabitat amelioration and reduced competition among understorey plants as drivers of facilitation across environmental gradients: Towards a unifying framework. Persp. Plant Ecol., Evol. Syst. 2011a;13:247–258. doi: 10.1016/j.ppees.2011.06.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Soliveres S, García-Palacios P, Castillo-Monroy AP, Maestre FT, Escudero A, Valladares F. Temporal dynamics of herbivory and water availability interactively modulate the outcome of a grass-shrub interaction in a semi-arid ecosystem. Oikos. 2011b;120:710–719. [Google Scholar]
  55. Soliveres S, Torices R, Maestre FT. Evolutionary relationships can be more important than abiotic conditions in predicting the outcome of plant-plant interactions. Oikos. doi: 10.1111/j.1600-0706.2011.20309.x. in press. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Suzuki M, Miyashita T, Kabaya H, Ochiai K, Asada M, Tange T. Deer density affects ground-layer vegetation differently in conifer plantations and hardwood forests on the Boso Peninsula, Japan. Ecological Research. 2008;23:151–158. [Google Scholar]
  57. Taylor D, Eamus D. Coordinating leaf functional traits with branch hydraulic conductivity: resource substitution and implications for carbon gain. Tree physiology. 2008;28:1169–1177. doi: 10.1093/treephys/28.8.1169. [DOI] [PubMed] [Google Scholar]
  58. Tielbörger K, Kadmon R. Temporal environmental variation tips the balance between facilitation and interference in desert plants. Ecology. 2000;81:1544–1553. [Google Scholar]
  59. Tongway DJ, Sparrow AD, Friedel MH. Degradation and recovery processes in arid grazing lands of central Australia. Part 1. Soil and land resources. J. Arid Env. 2003;55:301–326. [Google Scholar]
  60. Vanderberghe C, Smit C, Pohl M, Buttler A, Freléchoux F. Does strength of facilitation by nurse shrubs depend on grazing resistance of tree saplings? Basic App. Ecol. 2009;10:427–436. [Google Scholar]
  61. Veblen KE. Season- and herbivore-dependent competition and facilitation in a semi-arid savanna. Ecology. 2008;89:1532–1540. doi: 10.1890/07-0973.1. [DOI] [PubMed] [Google Scholar]
  62. Warnock AD, Westbrooke ME, Florentine SK, Hurst CP. Does Geijera parviflora Lindl. (Rutaceae) facilitate understorey species in semi-arid Australia? Rang. J. 2007;29:207–216. [Google Scholar]
  63. Weltzin J, Coughenour A. Savanna tree influence on understorey vegetation and soil nutrients in northwestern Kenya. J. Veg. Sci. 1990;1:325–334. [Google Scholar]
  64. Westoby M, Wright IJ. The leaf size-twig size spectrum and its relationship to other important spectra of variation among species. Oecol. 2003;135:621–628. doi: 10.1007/s00442-003-1231-6. [DOI] [PubMed] [Google Scholar]
  65. Wilson BR, Growns I, Lemon J. Scattered native trees and soil patterns in grazing lands of the Northern Tablelands of NSW, Australia. Aust. J. Soil Res. 2007;45:199–205. [Google Scholar]
  66. Wise MJ, Abrahamson WG. Beyond the compensatory continuum: environmental resource levels and plant tolerance of herbivory. Oikos. 2005;109:417–428. [Google Scholar]
  67. Wood DH. Estimating rabbit density by counting dung pellets. Aust. Wildlife Res. 1988;15:665–671. [Google Scholar]
  68. Wright IJ, Reich PB, Westoby M, Ackerly DD, Baruch Z, Bongers F, Cavender-Bares J, Chapin FS, Cornelissen JHC, Diemer M, Flexas J, Garnier E, Groom PK, Gulias J, Hikosaka K, Lamont BB, Lee T, Lee W, Lusk C, Midgley JJ, Navas ML, Niinemets Ü, Oleksyn J, Osada N, Poorter H, Poot P, Prior L, Pyankov VI, Roumet C, Thomas SC, Tjoelker MG, Veneklaas EJ, Villar R. The world-wide leaf economics spectrum. Nature. 2004;428:821–827. doi: 10.1038/nature02403. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Data
NIHMS63070-supplement-Supplementary_Data.docx (1.6MB, docx)

RESOURCES