SEARCH

SEARCH BY CITATION

Keywords:

  • clonal plants;
  • competitive neighbourhood characteristics;
  • determinants of plant community diversity and structure;
  • O-ring statistics;
  • phalanx vs. guerilla strategy;
  • spatial structure;
  • species interaction

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  1. Non-random spatial patterns are a common feature of plant communities. However, the mechanisms leading to their formation remain unknown. The clonal dispersal ability of a species, that is, the average length of spacers between ramets, is commonly acknowledged to influence spatial patterns in clonal plants, although this relationship remains to be demonstrated. Moreover, the clonal dispersal ability of neighbouring species may influence environmental conditions and trigger modifications in clonal characteristics of a focal species. Thus, not only the clonal dispersal ability of a species, but also that of its competitors may influence the fine-scale spatial pattern of a species.
  2. In this article, we compared spatial patterns (in terms of colonization and occupation of space) of species with low (L), intermediate (I) or high (H) clonal dispersal abilities. Twelve species were classified within three groups of clonal dispersal (L, I or H) based on their average spacer lengths, and seven types of experimental assemblages consisting of species from one, two or three dispersal groups were studied. Two questions were addressed: (i) does the species clonal dispersal ability influence their spatial patterns and (ii) are species fine-scale spatial patterns affected by the clonal dispersal of neighbours? Species spatial patterns were recorded for each assemblage and were then analyzed using point pattern analysis.
  3. Despite strong species-specific effects, L-species displayed the highest level of local aggregation, which is indicative of limited space colonization, and the lowest level of local co-occurrence with other species, which is indicative of a high level of space occupation. The opposite pattern was observed in H-species, while that of I-species was intermediate. The species spatial patterns were modified by the clonal dispersal ability of competitors.
  4. Synthesis. This study emphasizes the importance not only of clonal dispersal but also of biotic interactions and, more precisely, of plant neighbour characteristics, in the spatial patterning of grassland plant communities.

Introduction

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

Spatial patterns, that is, non-random distribution of individual plants in space, are a common feature of plant communities (Watt 1947; Kershaw 1963; Herben & Hara 2003). However, the underlying processes that generate such spatial structure have yet to be identified. Both theoretical (Silander & Pacala 1985) and experimental studies (Stoll & Prati 2001; Kikvidze et al. 2005) have suggested that, due to their sessile lifestyle, plant lives are mostly determined at a local scale via abiotic (e.g. resource availability) and biotic factors, such as the specific composition of their neighbourhood (the so-called ‘plant's eye view’ of the community, Turkington & Harper 1979), rather than the spatial average of the overall community (mean field assumption). This viewpoint assumes that spatial structure affects the fate of plant populations or communities, even at a scale of a few centimetres (Mack & Harper 1977; Silander & Pacala 1985; Stoll & Prati 2001), which emphasizes the importance of identifying fine-scale mechanisms of pattern formation. Spatial patterns may be determined by several mechanisms, either solely or in combination (Tilman & Kareiva 1997; Condit et al. 2000; Dieckmann, Law & Metz 2000). Hence, species spatial patterns are expected to result from (i) environmental heterogeneity, which may be generated by the vegetation itself (Condit et al. 2000; Bolker, Pacala & Neuhauser 2003; Montoya et al. 2009), (ii) demographic processes such as dispersal through either seeds or clonal reproduction (Hubbell 1979; Bolker, Pacala & Neuhauser 2003; Seidler & Plotkin 2006) or (iii) local biotic interactions (e.g. competition, facilitation) (Bolker & Pacala 1999; Seabloom et al. 2005). Therefore, experimental studies of simplified multispecies plant communities may prove a valuable tool for disentangling the relative effect of these parameters on plant spatial patterns.

Local dispersal (through seeds or vegetative reproduction) has been proposed as one of the factors that shape spatial correlations within and among species (Law & Dieckmann 2000). Most experimental studies on pattern formation in plants have focused on annual species (Bergelson 1990; Rees, Grubb & Kelly 1996; Stoll & Prati 2001). However, in grassland communities, clonal growth is the major driver of dispersal (Otsus & Zobel 2002; Benson & Hartnett 2006) that generates fine-scale aggregation, because it places daughter ramets at a short distance from their parents (Kershaw 1963; Harada & Iwasa 1996; Herben & Hara 2003) through connective stems (the connections). In clonal plants, this distance between parent and daughter ramets (spacer length sensu Bell 1984) is mainly determined by the clonal dispersal ability, which results from a trade-off between space colonization and space occupation (Wildová, Wild & Herben 2007), and has traditionally been classified along a gradient from low (phalanx strategy) to high clonal dispersal (guerilla strategy) (Lovett-Doust 1981). At one extreme of the gradient, clonal plants with low (L) clonal dispersal abilities are characterized by almost inexistent connection spacers. They are expected to form an uninterrupted front of aggregated ramets (Lovett-Doust 1981; Cheplick 1997; Humphrey & Pyke 1998) and to maximize local occupation of space. At the other extreme of the gradient, plants displaying high clonal dispersal (H) produce poorly ramified connections with long spacers that are expected to generate networks of widely spaced ramets and infiltrate the surrounding vegetation (Lovett-Doust 1981). Hence, high clonal dispersal is expected to maximize the colonization of space. However, although the length of connection spacers is often used as a surrogate of spatial pattern in clonal plants, this relationship has rarely been experimentally demonstrated.

Plant neighbourhood characteristics, such as competitor density, size or identity (i.e. the species they belong to), are known to influence the outcome of competition (Goldberg & Landa 1991). Nevertheless, most studies investigating the role of competitive interactions in plant community structuring have focused on the presence of competitors per se (Bertness & Ellison 1987; Lepš & Kindlmann 1987), neglecting the importance of their identity. Still, competition for space relies on the ability of plants to reach and colonize available patches, which is strongly species-specific (Herben et al. 1993; Herben & Hara 1997, 2003; Wildová, Wild & Herben 2007) and, for perennial species, is mostly determined by clonal dispersal (Schmid & Harper 1985; Cheplick 1997; Humphrey & Pyke 1998). Competitors may also generate fine-scale environmental heterogeneity (Jackson & Caldwell 1993; Herben, During & Law 2000), which may strongly depend on their clonal dispersal abilities (Herben, During & Krahulec 1995). Clonal plants may be able to adjust their foraging behaviour in response to this environmental heterogeneity by altering their clonal traits, especially branching intensity (i.e. the number of clonal connections) and spacer length (Slade & Hutchings 1987; Cain, Dudle & Evans 1996). Therefore, beyond the mere presence of competitors, their clonal dispersal abilities could strongly influence spatial pattern formation.

The aim of this study was to investigate the effect of clonal dispersal on fine-scale pattern formation (i.e. within distance ranges of a few centimetres). In this instance, we set up an experiment using 12 clonal species, primarily classified within three groups of clonal dispersal based on spacer length. We assembled species to constitute seven types of simplified grassland communities differing in the number and nature of co-occurring clonal dispersal groups. We specifically addressed two main questions. First, does the clonal dispersal ability of species influence their fine-scale spatial patterns? We expected the local colonization of space to increase and local occupation of space to decrease along a gradient from low-to-high clonal dispersal. Second, are the fine-scale spatial patterns of species modified by the clonal dispersal ability of neighbours? We used recent techniques of point pattern analysis to characterize the spatial patterns of plants, as a combined description of local colonization and local occupation of space.

Materials and methods

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

Plant material

Clonal plants are constituted as a network of erected shoots (ramets, Harper 1977) connected through modified horizontal stems (connections). The connective part between two consecutive shoots will be described as a spacer (Bell 1984). Such architecture promotes a certain dispersal ability for the plant, which can be evaluated through the spacer length. We delineated three groups of four species each, one dicotyledon and three monocotyledons (Poaceae species), based on their clonal dispersal (Table 1). Among a set of common species of temperate grasslands, the selection of the 12 studied species and their classification within the three dispersal groups were based on their spacer lengths. For each species, we randomly selected and measured a spacer on 20 clonal fragments found on field margins around Rennes (Brittany, west France) to calculate a mean value. These species were all developing through a network of shoots with either above-ground or below-ground connections, spacer length being almost null in species with low (L-species), short in species with intermediate (I-species) and long in species with high dispersal ability (H-species) (Table 1).

Table 1. Species representing each dispersal group. Mean values (± SE) correspond to mean spacer length calculated from measurements on 20 clonal fragments of each species randomly sampled from grasslands. Values are in centimetres
 High dispersal (H)Intermediate dispersal (I)Low dispersal (L)
    
 Elytrigia repens L.2.78 (± 0.11)Brachypodium pinnatum L.1.03 (± 0.05)Lolium perenne L.0.28 (± 0.03)
MonocotyledonAgrostis stolonifera L.7.95 (± 0.53)Festuca rubra L.0.87 (± 0.09)Dactylis glomerata L.0.51 (± 0.04)
 Holcus mollis L.4.03 (± 0.23)Agrostis tenuis Sibth.2.62 (± 0.12)Holcus lanatus L.0.28 (± 0.02)
DicotyledonRanunculus repens L.14.98 (± 0.70)Anthemis nobilis L.1.22 (± 0.06)Centaurea nigra L.0.44 (± 0.04)

Independent clonal fragments of these species were then collected in April 2009. We selected units composed of one mature shoot with one connected spacer for species with an intermediate and high dispersal ability and of three joined shoots forming a tuft for species with low dispersal ability (i.e. with almost inexistent spacers). This standardization of initial individuals allowed them to have stored resources to support survival after transplantation, as well as living axillary buds for clonal development. Before planting, we carefully selected units that were of similar size (in terms of shoot height) to lower differences in initial competitive ability of the units.

Experimental design

To study species spatial patterns in different competitive environments, experimental communities were planted in the experimental garden of the University of Rennes in May 2009.

Seven types of species assemblages were tested, differing in the nature and the number of co-occurring clonal dispersal groups (one, two, or three); all combinations of clonal dispersal were tested: three assemblages of one clonal dispersal group [species with low (L), intermediate (I) or high (H) dispersal – these assemblages were composed of the four species representing the group of dispersal], three assemblages of two clonal dispersal groups [species with low and intermediate dispersal (L-I), species with low and high dispersal (L-H) or species with intermediate and high dispersal (I-H) – each assemblage was composed of eight species, four from each selected dispersal group] and one assemblage with the three clonal dispersal groups [species with low, intermediate and high dispersal (L-I-H) – this assemblage was constituted of 12 species]. Each type of community was replicated eight times. The replicates were randomly positioned within the experimental area, comprising 56 culture plots of 1.30 × 1.30 × 0.25 m3. In each culture plot, 48 clonal units were planted at a distance of 16 cm from each other following a hexagonal pattern (Birch, Oom & Beecham 2007). The substrate was homogeneous within and amongst all experimental plots and was composed of sand (20%) and ground soil (80%). The clonal units were equally distributed over all co-occurring species, and their position on the plantation pattern was randomized.

To focus only on dispersal linked with clonal growth, and avoid seed production, all mature flowers were cut off. Weeds were regularly removed, and the culture plots were watered every two days during the dry season. Above-ground vegetation was mown once a year at the end of the summer.

Data collection

In October 2010, we recorded the spatial patterns of all species present in the experimental communities using a square lattice of 80 × 80 cm² centred on the culture plot. The lattice was positioned in this way as it eliminated individuals at the edge of the plantation pattern, which did not have six neighbouring competitors, and thus minimized the edge effect. In the studied assemblages of clonal plants, the isolation of individual ramets would have been almost impossible or, at best, highly time-consuming and beyond the scope of this study. Thus, we recorded presence/absence data in 5 × 5 cm² cells of the lattice (256 cells in total). The presence of a species was denoted when a ramet rooted in the target cell. We selected this cell size as it is larger than single ramets, but smaller than clonal fragments, and corresponds well to the scale at which interactions are likely to occur for grassland plant species (Purves & Law 2002).

Fine-scale spatial pattern analysis

The experiment was set up so that initial abiotic conditions (light, nutrients and edaphic conditions), as well as initial plant density within and between experimental plots, were homogeneous. Consequently, any emerging spatial pattern could be assumed to be stationary and isotropic (i.e. the density of individuals λ(x) was independent from the location x and the direction) and, thus, to only relate to the relative position of pairs of points (second-order spatial structure). Our aim was to describe fine-scale spatial patterns for each species in all plots. For that purpose, we used distance-dependent pair density probability functions (Wiegand & Moloney 2004).

The O-ring statistics O(r) describes the average density of points at a distance r (Wiegand & Moloney 2004; Law et al. 2009). The pair correlation function g(r) results from the normalization by the intensity of the pattern λ so that g(r) = O(r) / λ (Wiegand & Moloney 2004; Law et al. 2009). A random distribution of points is indicated by O(r) = λ and g(r) = 1. In the present study, we used O(r) rather than g(r), as we were interested in direct measures of the average density of neighbours (see Appendix S1 in Supporting Information for more details on spatial analyses). The average density of a given pattern i was defined as follows:

  • display math(eqn 1)

where N is the number of cells within a grid, and pi(x) is the density of pattern i in cell x, with pi(x) taking values 0 or 1 (Law et al. 2002). Consequently, λi was comprised between 0 (no point of pattern i in the grid) and 1 (all the cells of the grid occupied by pattern i).

Local aggregation

Local aggregation depends on the ability of a plant to expand in the surrounding space. We estimated local aggregation using the univariate pair correlation function O11(r), which directly measures the average density of points of pattern 1 (i.e. the pattern of the target species sp1), at a distance r from all points of pattern 1.

Given the diffuse nature of spatial colonization by clonal plants (i.e. from the cell it initially occupies, an individual plant colonizes neighbouring cells in a continuous way), the density of adjacent cells colonized by a plant is expected to more strongly decrease with distance when spatial colonization is low (on very local distances) than for plants with high spatial colonization (on longer distances). Consequently, we calculated the index of local aggregation, Aloc, as an indicator of local aggregation of conspecific points. This index, which was adapted from Réjou-Méchain et al. (2011) (note that these authors used g(r) rather than O(r)), was defined as minus the slope of O11(r) on log(r) for a radius r ranging between 1 and 3 neighbouring cells (i.e. within the distance range [5–15 cm]). This short distance range corresponded to the fine scale that we were interested in for this study. High values of Aloc indicated a high level of local aggregation of the target species. Aloc should covary negatively with the ability of a plant to expand in the surrounding space (i.e. colonization ability).

Local co-occurrence

We expected the level of spatial co-occurrence of a plant with heterospecific plants to depend on its ability to form a physical barrier that prevents neighbours from colonizing the area it already occupies. We described local co-occurrence using the bivariate O-ring statistics O12(r), which measures the average density of points of pattern 2 at a distance r away from all points of pattern 1, thus describing the spatial association of patterns 1 and 2 at the scale r (Wiegand & Moloney 2004; Law et al. 2009). In these analyses, each point of pattern 1 corresponded to the presence of the target species sp1 within a cell, while each point of pattern 2 corresponded to the presence of at least one species of the assemblage other than the target species (spi≠1) within a cell.

More precisely, we calculated the local co-occurrence index, Cloc, as the value of O12(r) within cells occupied by target points of pattern 1 (i.e. at distance r = 0). Consequently, Cloc corresponded to the average density of cells occupied by pattern 1 that were also co-occupied by pattern 2. This index should covary negatively with the level of local occupation of space.

We calculated univariate and bivariate O-ring statistics using the grid-based approach of the Programita software (Wiegand & Moloney 2004). The smallest spatial unit for analysis corresponded to one 5 × 5 cm² cell of the sampling lattice. All analyses were carried out up to a maximal distance that was equal to half the length of our square study plots, rmax = 8 cells (40 cm), and with ring width dr = 1 cell (5 cm).

Change in relative cover

As an indicator of species spatial performance, we described the change in relative cover (Δ cover) for a species compared with the other species of the assemblage, within the entire experimental plot. We calculated Δ cover as the relative difference between the observed relative percentage cover and the relative percentage cover expected under the assumption of equal colonization of space among species. Δ cover = −1 indicated the disappearance of a species from the plot, negative values indicated that the species was scarcer than expected, and increasing positive values indicated an increasing dominance of the species in the vegetation cover (Table 2).

Table 2. Description of the synthetic indices of local aggregation (Aloc), local co-occurrence with heterospecific plants (Cloc) and change in relative cover (Δ cover)
Index nameIndex calculationInterpretation
Local aggregation (Aloc)Minus the slope of O11(r) against log(r) for r ranging between 1 and 3 cells, that is, [5–15 cm].image with image local aggregation
Local co-occurrence (Cloc)Value of O12(r) at r = 0image with imageco-occurrence with heterospecific plants
Change in relative cover (Δ cover)inline image with inline image and Cover exp- sp1 = 25, 12.5 or 8.33 % in mono-, bi-, and tri-dispersal assemblages, respectivelyimage with image dominance in the cover −1: mortality of all individuals of species < 0: cover weaker than expected = 0: cover equal to expected > 0: cover greater than expected

Statistical analyses

We empirically fixed a threshold of 10 cells colonized by the target species within a plot, under which spatial pattern analysis was not carried out. Below this threshold, the number of points was too low to ensure sufficient statistical power and led to jagged plots that could not be analyzed. Consequently, the number of replicates could dramatically be reduced for some species, mainly due to the disappearance of some clonal fragments during the experiment. In particular, spatial pattern analysis was not carried out for the L-species Lolium perenne, due to high mortality of clonal fragments, regardless of the assemblage. We also did not carry out spatial pattern analysis for the I-species Anthemis nobilis in the L-I-H assemblage or Agrostis tenuis in both the I-H and L-I-H assemblages. Finally, the H-species Holcus mollis was excluded from spatial pattern analysis for the L-H assemblage.

To detect any general relationship between Aloc, Cloc and Δ cover, we tested pairwise correlation between these three indices across all species. For that purpose, we calculated the average value of each index for each species (except for L. perenne, for which spatial analysis was not carried out). As conditions of normality were not satisfied for Δ cover, we used the Spearman's rank correlation coefficient. Then, we tested the effect of (i) clonal dispersal and (ii) the type of assemblage on species spatial patterns. As one species belonged to a single clonal dispersal group (L, I or H), some species were absent from some assemblages. For instance, there were no L-species in the I-H assemblage. Thus, we did not carry out any global analysis on the whole data set. Rather, we partitioned the whole data set into several subsets. To analyse the link between clonal dispersal and species spatial pattern, we partitioned the data set according to the type of assemblage, which led to the formation of five subsets: one subset grouping the assemblages of one clonal dispersal group (L, I and H assemblages) and four subsets representing assemblages of two or three clonal dispersal groups (L-I, L-H, I-H and L-I-H assemblages). For each assemblage, we carried out linear-model anovas on each synthetic index of spatial pattern (Aloc, Cloc) and the change in relative cover (Δ cover), with clonal dispersal group (L, I and H) as the fixed effect and species as fixed effect nested into clonal dispersal group (Crawley 2007). For a target species spi, Δ cover depended on the relative cover of all species within the assemblage. Thus, we did not carry out the analysis for this index in assemblages of one clonal dispersal group, where all species belonged to the same clonal dispersal group (no possible dispersal group effect).

To determine how the type of neighbourhood affected species spatial pattern, we separately carried out three linear-model anovas on Aloc, Cloc and Δ cover for species with low, intermediate and high dispersal ability, using the assemblage (one dispersal group, L-I, L-H, I-H and L-I-H) and the species as fixed factors. Data for a given species were included in the analyses providing that they were available for at least two replicates. Consequently, in some assemblages, the spatial patterns for a given clonal dispersal group did not include all of the four species representing a particular dispersal ability. Data were log-transformed, when necessary, to meet the assumptions of the analyses. Tukey's Honestly Significant Different (HSD) tests were used for post-hoc comparisons. Linear-model anovas were performed using the R software version 2.8.1. (R Development Core Team 2008).

Results

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

Correlations between local aggregation, local co-occurrence and change in relative cover

We found a negative correlation between the index of local aggregation Aloc and the index of local co-occurrence Cloc (ρ = −0.84, P = 0.0026, n = 11). In contrast, no significant correlation was detected between either Aloc and Δ cover (ρ = 0.14, P = 0.69, n = 11) or Cloc and Δ cover (ρ = −0.19, = 0.58, n = 11).

Impact of clonal dispersal on species spatial patterns

Aloc inversely increased with the dispersal ability of the species, being the lowest for the H-species, intermediate for I-species and the highest for the L-species (Fig. 1 and Table 3; see also Fig. S1). This hierarchy remained consistent regardless of the assemblage (Table 3).

Table 3. Results of linear-model anovas testing the clonal dispersal group and species effect on the three synthetic indices of plant spatial patterns. Species effect was nested into strategy
 Clonal dispersal groupClonal dispersal group / SpeciesComparison of means
  1. Comparison of means resulted from Tukey's HSD post-hoc tests. ns P-value > 0.05, *P-value < 0.05, **P-value < 0.01, ***P-value < 0.001.

L, I and H assemblages
Alocn =78 F = 199.63***F = 6.53***H < I < L
Clocn = 78F = 187.38***F = 6.00***L < I < H
L-I assemblage
Δ covern = 64F = 32.07***F = 14.02***I < L
Alocn = 45F = 20.28***F = 1.43nsI < L
Clocn = 45F = 8.04**F = 1.24nsL < I
L-H assemblage
Δ covern = 64F = 6.18*F = 53.73***L < H
Alocn = 41F = 71.08***F = 3.23*H < L
Clocn = 41F = 0.71nsF = 6.91*** 
I-H assemblage
Δ covern = 64F = 264.48***F = 34.74***I < H
Alocn = 44F = 76.21***F = 6.18***H < I
Clocn = 44F = 0.055nsF = 1.91ns 
L-I-H assemblage
Δ covern = 96F = 97.51***F = 23.06***I < L < H
Alocn = 59F = 118.04***F = 8.43***H < I < L
Clocn = 59F = 9.79***F = 2.48*L < I = H
image

Figure 1. Local aggregation index Aloc. Values are averages (± SE) of replicates plots (a, c, e) across all species belonging to a clonal dispersal group and (b, d, f) for each species individually. L low dispersal, I intermediate dispersal and H high dispersal, in assemblages of one, two or three clonal dispersal groups. Different lower-case letters indicate significant differences (post-hoc Tukey's HSD tests).

Download figure to PowerPoint

In the assemblages of one clonal dispersal group, Cloc increased with the dispersal ability of the species, being lowest for L-species, intermediate for I-species and the highest for H-species (Fig. 2 and Table 3; see also Fig. S2). The hierarchy in Cloc between L-species and I-species was maintained regardless of the assemblage. By contrast, Cloc was not significantly different between I-species and H-species when they were grown in the same assemblages (either I-H or L-I-H assemblages) (Table 3). Despite no significant difference in the L-H assemblage, L-species maintained a significantly lower Cloc than H-species when grown in the L-I-H assemblage (Table 3).

image

Figure 2. Local co-occurrence index Cloc. Values are averages (± SE) of replicates plots (a, c, e) across all species belonging to a clonal dispersal group and (b, d, f) for each species individually. L low dispersal, I intermediate dispersal and H high dispersal, in assemblages of one, two or three clonal dispersal groups. Different lower-case letters indicate significant differences (post-hoc Tukey's HSD tests).

Download figure to PowerPoint

Response of species spatial patterns to neighbourhood characteristics

For L and H-species, Aloc was affected by the clonal dispersal ability of neighbouring species (significant assemblage effects on Aloc for both dispersal groups) (Table 4). In H-species, Aloc increased in the presence of L-species (i.e. in L-H and L-I-H assemblages) (Fig. 1e,f), whereas the presence of I-species alone had no effect. Although the presence of H-species alone had no significant effect, Aloc of L-species tended to increase when grown with I-species (i.e. in L-I assemblage) and reached its maximal value in the assemblage of three dispersal groups (L-I-H assemblage) (Fig. 1a,b). In I-species, Aloc was, on average, not significantly affected by the type of assemblage, despite species-specific differences (Fig. 1c,d and Table 4).

Table 4. Results of linear-model anovas testing the assemblage and species effect on the three synthetic indices of spatial pattern
 AssemblageSpeciesAssemblage × Species
  1. ns P-value > 0.05, *P-value < 0.05, **P-value < 0.01, ***P-value < 0.001.

Low clonal dispersal group
Δ covern = 128F = 20.80***F = 103.40***F = 2.55*
Alocn = 91F = 8. 43***F = 8. 47***F = 1.43ns
Clocn = 91F = 59. 56***F = 7. 42***F = 1.07ns
Intermediate clonal dispersal group
Δ covern = 127F = 88.25***F = 41.53***F = 10.76***
Alocn = 72F = 1.36nsF = 5.12**F = 2.44*
Clocn = 72F = 49.80***F = 0.84nsF = 0.66ns
High clonal dispersal group
Δ covern = 128F = 20.20***F = 150.20***F = 4.95***
Alocn = 104F = 13.51***F = 31.69***F = 1.02ns
Clocn = 104F = 33.06***F = 23.21***F = 4.46***

Cloc also varied in response to the clonal dispersal ability of species constituting the neighbourhood. For L-species, Cloc remained weak regardless of the assemblage but significantly increased in the presence of H-species, either alone (L-H assemblage) or in assemblage with I-species (L-I-H assemblage), whereas it was not affected by the presence of I-species alone (L-I assemblage) (Fig. 2a,b and Table 4). Similarly, Cloc of the latter species was not affected by L-species (L-I assemblage), but significantly increased in the presence of H-species, as indicated by values close to 1 in either I-H or L-I-H assemblage (Fig. 2c,d and Table 4). For H-species, Cloc was the highest when they only developed with I-species (I-H assemblage), was intermediate in the H assemblage and the L-I-H assemblage and was the lowest in the presence of L-species alone (L-H assemblage) (Fig. 2e,f and Table 4).

Changes in relative cover

Species relative cover (i.e. the relative number of cells in which a species was present compared with other species constituting the assemblage) was unbalanced at both the species and clonal dispersal group level. Regardless of the assemblage, the observed relative cover was, on average, the lowest for I-species (negative values of Δ cover) and the largest for H-species (positive values of Δ cover) (Fig. 3c-e and Table 3).

image

Figure 3. Relative variation between species relative cover expected under the assumption of equal space colonization among species (Δ cover). Values are averages (± SE) of replicates plots (a, c, e) across all species belonging to a clonal dispersal group and (b, d, f) for each species individually. L low dispersal, I intermediate dispersal and H high dispersal, in assemblages of one, two or three clonal dispersal groups. Different lower-case letters indicate significant differences (post-hoc Tukey's HSD tests). See Table 2 for details about the method of calculation of Δ cover.

Download figure to PowerPoint

The change in relative cover also depended on the assemblage (Table 4). For I- and H-species, relative cover tended to be more balanced when they developed in assemblages with L-species only (with smaller divergence of observed cover from expected cover, Fig. 3c-e and Table 4). For the latter species, the change in relative cover also depended on the clonal dispersal ability of species constituting the assemblage; L-species dominated the cover when grown only with I-species, but expanded less than expected when H-species were present (Fig. 3a).

Interspecific variations in spatial patterns and changes in relative cover

In this study, each clonal dispersal group comprised four species (one dicotyledon and three monocotyledons). Regardless of the clonal dispersal group, our results showed significant species-specific effects on species spatial patterns and their response to neighbouring species (type of assemblage) (significant dispersal group/species effects, except for Aloc and Cloc in L-I and Cloc in I-H assemblages; Table 3; see also Figs S3–S8). These results were confirmed by significant species effects, when data for all assemblages were considered within each clonal dispersal group, except for Cloc with I-species (Table 4). In addition, species that belonged to the same clonal dispersal group tended to respond differently to the neighbourhood, as revealed by significant assemblage × species effects, except for Cloc with I-species (Table 4). Nevertheless, these interspecific differences rarely masked the effects of the main factors (i.e. the clonal dispersal group and assemblage) (Tables 3 and 4).

The change in relative cover was rather homogeneous among I-species and, to a lesser extent, L-species (except for Lolium perenne, for which mortality was high in all assemblages) (Fig. 3b-d). Conversely, H-species proved to be highly variable regarding their relative cover within a plot (Fig. 3f). Elytrigia repens dominated plant cover in all assemblage types, contrasting with Ranunculus repens and Holcus mollis, which were dominated by other species. For Agrostis stolonifera, the change in relative cover depended on the assemblage, only being positive when mixed with I-species (Fig. 3f).

Discussion

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

The present study addressed the gap between observed spatial patterns in plant communities and the processes responsible for these patterns. Despite species-specific effects, our results demonstrated that the spatial patterns of clonal plants grown in assemblages depend on both the species clonal dispersal ability and the clonal dispersal ability of neighbouring species.

Clonal dispersal as a predictor of species spatial patterns

Using a large panel of clonal plant species, our study revealed contrasting plant spatial patterns at a very fine scale (of up to 15 cm), which corresponds to distances at which biotic interactions are likely to occur (Purves & Law 2002). These results provided evidence for the importance of clonal growth as a mechanism of spatial dynamics in grassland plant communities (Herben, During & Krahulec 1995).

Regardless of the assemblage, our study showed that patterns of local colonization (which was negatively linked to the index of local aggregation Aloc) and occupation of space (which was negatively linked to the index of local co-occurrence Cloc) depended on the clonal dispersal ability of the species (i.e. spacer length). In accordance with our first hypothesis, species with very short spacers (L-species in our study; phalanx species sensu Lovett-Doust 1981) maximized spatial occupation, whereas species with long spacers (H-species in our study; guerilla species sensu Lovett-Doust 1981) maximized spatial colonization. Although this relationship is often assumed, this is, as far as we know, the first demonstration that differences in spacer lengths result in different spatial patterns. These observations and, especially, the negative correlation between the indices of local aggregation and local co-occurrence confirmed the existence of a trade-off in clonal plants between colonization and the consolidation of space (Cheplick 1997; Humphrey & Pyke 1998; Gough et al. 2002).

Influence of neighbourhood characteristics on spatial patterns

Although general hierarchies in the spatial pattern of species characterized by different clonal dispersal abilities were maintained regardless of the assemblage, our results demonstrated that the identity and clonal dispersal ability of neighbouring species can influence spatial pattern formation.

For L (low dispersal) and H (high dispersal) species, the pattern of local colonization of space responded differently to the clonal dispersal ability of neighbours, whereas it remained fixed for I (intermediate dispersal) species. First, the colonization of space by L-species was reduced in the presence of I-species alone (tendency in the L-I assemblage) or mixed with H-species (in the L-I-H assemblage). Such modification in local colonization might have enhanced the ability of L-species to form a physical barrier to protect their clonal territory (Humphrey & Pyke 1998; Gough et al. 2002), as indicated by the preservation of a high level of occupation of space in the presence of I-species alone. However, L-species failed to efficiently protect their territory in the presence of H-species alone, which managed, at least to a certain extent, to enter the tuft (increase in the index of local co-occurrence). In these latter species, the presence of L-species generated a significant decrease in the colonization of space, paralleled with either an increase or no change in the occupation of space (in the L-H and the L-I-H assemblages respectively).

Two processes may explain these overall variations in species spatial patterns in response to the neighbourhood, although we could not disentangle their relative effects based on our current experimental design. First, plastic modifications in the clonal architecture of individuals and, especially, of spacer length have already been demonstrated in response to competition (Cheplick & Gutierrez 2000; Humphrey & Pyke 2001; Weijschedé et al. 2008). Such variation in clonal architecture may specifically depend on the identity of their neighbours (Callaway, Pennings & Richards 2003). Second, local events of ramet mortality may free space within individual clonal territory. Hence, such demographic variations could have been primarily responsible for the decrease in occupation of space observed for L-plants in the presence of H-competitors.

As each clonal dispersal group was composed of four species, the number of species differed between assemblages of one, two or three groups. Such differences in species richness might have contributed to the observed differences in species spatial patterns, but this cannot be tested with our experimental design. Nevertheless, differences in species spatial patterns were observed in the assemblages of two groups presenting the same species richness (eight species). We may, thus, assume that this species richness effect was of less importance compared with the effects of the clonal dispersal and of the neighbourhood.

Consequences for species relative cover

Despite a clear trade-off between colonization and consolidation of space, a hierarchy in relative cover among clonal dispersal groups was observed, with H-species being the strongest and I-species being the lowest competitors. However, changes in relative cover in the three dispersal group assemblage (L-I-H) could not be predicted as the average of changes in relative cover recorded in two dispersal group assemblages. This result illustrated the non-additivity of competitive interactions within plant community (Weigelt et al. 2007).

No correlation was detected between either local colonization of space (Aloc) or local occupation of space (Cloc), and species change in relative cover (Δ cover), across studied species. These results indicated that species change in relative cover cannot be directly linked to their average spatial patterns, suggesting that other factors influenced species relative cover in the assemblages. H-species dominated the cover of all types of multidispersal group assemblages, despite the colonization of space requiring a major investment in connections. Consequently, the presence of H-species in the assemblage generated a decrease in the relative cover of the other species, regardless of the complexity of the assemblage (assemblages with either two or three dispersal groups). Nevertheless, their spatial spread was hampered by L-species (second strongest competitors), possibly due to their tightly packed structure (Lovett-Doust 1981; Humphrey & Pyke 1998) and their ability to retain available nutrients (Derner & Briske 1999). These latter species are also usually considered to maximize intraclonal contacts (Lovett-Doust 1981; Schmid 1986; Humphrey & Pyke 1998), reducing the probability of meeting individuals with greater competitive ability (Murrell, Purves & Law 2001; Stoll & Prati 2001). Differences in the success of these two groups of species are consistent with previous studies. Species with high dispersal ability (H) can rapidly colonize and exploit open spaces, being better competitors than species with low dispersal ability (L) in the short term (Schmid & Harper 1985; Cheplick 1997; Humphrey & Pyke 1998). In addition, the infiltration of H-species within the clonal territory of L-species increased the number of interspecific contacts, which might have enhanced the competitive pressure and limited L-species relative cover.

Regardless of the assemblage, I-species were always dominated by other species. This suggested that intermediate clonal dispersal, which lacks the ability to spread extensively or to resist competitor invasions, is not efficient in the competition for space. Nevertheless, initially planting three shoots of L-species and only one shoot of I-species may have given a competitive advantage to the former, which could also explain the weaker performance of I-species.

The competitive hierarchies among clonal dispersal groups were established after only two years of cultivation and confirmed the short-term advantage provided by high dispersal ability. However, species with low or intermediate dispersal abilities, often referred to as phalanx species, are known to be more efficient in the long term (Gough et al. 2002). Therefore, it remains to be tested whether the structure of these hierarchies changes over time.

Interspecific variations in spatial patterns and changes in relative cover

Within the same clonal dispersal group, we found significant species effects on the spatial pattern and spatial performance within a plot (change in relative cover). Other clonal features can explain these interspecific variations. For instance, traits related to physiological integration (Herben 2004) or demographic parameters (e.g. ramet birth and death rate or ramet life span) (Wildová, Wild & Herben 2007) may affect the effective spatial pattern.

Non-clonal traits might also explain observed inter-specific differences, particularly regarding relative cover and survival within the plots. For instance, high mortality rates recorded for the grass L. perenne in all assemblage types cannot be linked to its own clonal dispersal ability or to the clonal ability of neighbours. Instead, these mortality rates may be due to a weak competitive ability in the undisturbed environment encountered within the experimental plots. Furthermore, only two of four H-species (A. stolonifera and E. repens) managed to dominate the cover in the L-I-H assemblage. The success of these species could have been enhanced by ramet traits supporting their competitive ability (e.g. shoot height). Studies have then highlighted that integrating clonal as well as non-clonal traits may improve the description of plant strategies along environmental gradients (Rusch et al. 2011; Gough et al. 2012). As an example, the combination of a high clonal dispersal with a high stature, which respectively maximizes space capture in the horizontal and vertical plans, is expected to enhance plants competitive ability (Gough et al. 2012). As advocated by de Kroon & Schieving (1990), the consideration of larger sets of traits associated with clonal growth, but also of non-clonal traits, may further contribute to our understanding of the role of clonality in plant spatial pattern formation and relative cover in clonal plant communities.

Concluding remarks

Clonal dispersal emerges as a key element of spatial pattern formation in plant communities. Our study provides evidence that clonal dispersal may rapidly (i.e. after only two experimental years) shape the fine-scale spatial structure of the plant community, by generating diverse spatial patterns from an initial random distribution of individuals. In addition, spatial patterns were mediated by the clonal dispersal abilities of the surrounding plants. Therefore, the neighbours' characteristics, rather than the mere presence of competitors, must be considered when investigating spatial patterning in plant communities, especially those dominated by clonal plants.

Acknowledgements

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

The authors thank N. Renaud for assistance with the data collection and J.S. Pierre for helpful comments on statistical analyses. We are also grateful to T. Wiegand, who kindly provided access to Programita software. The English was corrected by G. Schofield. This project benefited from the grant ANR-08-SYSC-012 provided by the Agence Nationale de la Recherche (France).

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • Bell, A.D. (1984) Dynamic morphology: a contribution to plant population ecology. Perspectives in Plant Population Ecology (eds R. Dirzo & J. Sarukhán), pp. 4865. Sinauer, Sunderland, Massachusetts.
  • Benson, E.J. & Hartnett, D.C. (2006) The role of seed and vegetative reproduction in plant recruitment and demography in tallgrass prairie. Plant Ecology, 187, 163178.
  • Bergelson, J. (1990) Spatial patterning in plants: opposing effects of herbivory and competition. Journal of Ecology, 78, 937948.
  • Bertness, M.D. & Ellison, A.M. (1987) Determinants of pattern in a New England salt marsh plant community. Ecological Monographs, 57, 129147.
  • Birch, C.P.D., Oom, S.P. & Beecham, J.A. (2007) Rectangular and hexagonal grids used for observation, experiment and simulation in ecology. Ecological Modelling, 206, 347359.
  • Bolker, B.M. & Pacala, S.W. (1999) Spatial moment equations for plant competition: understanding spatial strategies and the advantages of short dispersal. The American Naturalist, 153, 575602.
  • Bolker, B.M., Pacala, S.W. & Neuhauser, C. (2003) Spatial dynamics in model plant communities: what do we really know? The American Naturalist, 162, 135148.
  • Cain, M.L., Dudle, D.A. & Evans, J.P. (1996) Spatial models of foraging in clonal plant species. American Journal of Botany, 83, 7685.
  • Callaway, R., Pennings, S.C. & Richards, C. (2003) Phenotypic plasticity and interactions among plants: trait plasticity and community dynamics. Ecology, 84, 11151128.
  • Cheplick, G.P. (1997) Responses to severe competitive stress in a clonal plant: differences between genotypes. Oikos, 79, 581591.
  • Cheplick, G.P. & Gutierrez, C.M. (2000) Clonal growth and storage in relation to competition in genets of the rhizomatous perennial Amphibromus scabrivalvis. Canadian Journal of Botany, 78, 537546.
  • Condit, R., Ashton, P.S., Baker, P., Bunyavejchewin, S., Gunatilleke, S., Gunatilleke, N. et al. (2000) Spatial patterns in the distribution of tropical tree species. Science, 288, 14141418.
  • Crawley, M.J. (2007) The R book. John Wiley & Sons, Chichester.
  • Derner, J.D. & Briske, D.D. (1999) Intraclonal regulation in a perennial caespitose grass: a field evaluation of above- and below-ground resource availability. Journal of Ecology, 87, 737747.
  • Dieckmann, U., Law, R. & Metz, J.A.J. (2000) The Geometry of Ecological Interactions: Simplifying Spatial Complexity. Cambridge University Press, Cambridge.
  • Goldberg, D.E. & Landa, K. (1991) Competitive effect and response: hierarchies and correlated traits in the early stages of competition. Journal of Ecology, 79, 10131030.
  • Gough, L., Goldberg, D.E., Hershock, C., Pauliukonis, N. & Petru, M. (2002) Investigating the community consequences of competition among clonal plants. Evolutionary Ecology, 15, 547563.
  • Gough, L., Gross, K.L., Cleland, E.E., Clark, C.M., Collins, S.L., Fargione, J.E., Pennings, S.C. & Suding, K.N. (2012) Incorporating clonal growth form clarifies the role of plant height in response to nitrogen addition. Oecologia, 169, 10531062.
  • Harada, Y. & Iwasa, Y. (1996) Analyses of spatial patterns and population processes of clonal plants. Researches on Population Ecology, 38, 153164.
  • Harper, J.L. (1977) Population Biology of Plants. Academic Press, London.
  • Herben, T. (2004) Physiological integration affects growth form and competitive ability in clonal plants. Evolutionary Ecology, 18, 493520.
  • Herben, T., During, H.J. & Krahulec, F. (1995) Spatiotemporal dynamics in mountain grasslands: species autocorrelations in space and time. Folia Geobotanica and Phytotaxonomica, 30, 185196.
  • Herben, T., During, H.J. & Law, R. (2000) Spatio-temporal patterns in grassland communities. The Geometry of Ecological Interactions: Simplifying Spatial Complexity (eds U. Dieckmann, R. Law & J.A. J Metz), pp. 4864. Cambridge University Press, Cambridge.
  • Herben, T. & Hara, T. (1997) Competition and spatial dynamics of clonal plants. The Ecology and Evolution of Clonal Plants (eds H. de Kroon & J. van Groenendael), pp. 331358. Backhuys Publishers, Leiden.
  • Herben, T. & Hara, T. (2003) Spatial pattern formation in plant communities. Morphogenesis and Pattern Formation in Biological Systems–Experiments and Models (eds T. Sekimura, S. Noji, N. Ueno & P.K. Maini), pp. 223235. Springer Verlag, Berlin.
  • Herben, T., Krahulec, F., Hadincová, V. & Kovářová, M. (1993) Small-scale spatial dynamics of plant species in a grassland community over six years. Journal of Vegetation Science, 4, 171178.
  • Hubbell, S.P. (1979) Tree dispersion, abundance, and diversity in a tropical dry forest. Science, 203, 12991309.
  • Humphrey, L.D. & Pyke, D.A. (1998) Demographic and growth responses of a guerrilla and a phalanx perennial grass in competitive mixtures. Journal of Ecology, 86, 854865.
  • Humphrey, L. & Pyke, D. (2001) Ramet spacing of Elymus lanceolatus (thickspike wheatgrass) in response to neighbour density. Canadian Journal of Botany, 79, 11221126.
  • Jackson, R.B. & Caldwell, M.M. (1993) The scale of nutrient heterogeneity around individual plants and its quantification with geostatistics. Ecology, 74, 612614.
  • Kershaw, K.A. (1963) Pattern in vegetation and its causality. Ecology, 44, 377388.
  • Kikvidze, Z., Pugnaire, F.I., Brooker, R.W., Choler, P., Lortie, C.J., Michalet, R. & Callaway, R. (2005) Linking patterns and processes in alpine plant communities: a global study. Ecology, 86, 13951400.
  • de Kroon, H. & Schieving, F. (1990) Resource partitioning in relation to clonal growth strategy. Clonal Growth in Plants: Regulation and Function (eds J. van Groenendael & H. de Kroon), pp. 113130. SPB Academic Publishing, The Hague.
  • Law, R. & Dieckmann, U. (2000) A dynamical system for neighborhoods in plant communities. Ecology, 81, 21372148.
  • Law, R., Purves, D.W., Murrell, D.J. & Dieckmann, U. (2002) Causes and effects of small-scale spatial structure in plant populations. Interim Report of the International Institute for Applied Systems Analysis. Laxenburg, Austria.
  • Law, R., Illian, J., Burslem, D., Gratzer, G., Gunatilleke, C.V.S. & Gunatilleke, I. (2009) Ecological information from spatial patterns of plants: insights from point process theory. Journal of Ecology, 97, 616628.
  • Lepš, J. & Kindlmann, P. (1987) Models of the development of spatial pattern of an even-aged plant population over time. Ecological Modelling, 39, 4557.
  • Lovett-Doust, L. (1981) Population dynamics and local specialization in a clonal perennial (Ranunculus repens). I. The dynamics of ramets in contrasting habitats. Journal of Ecology, 69, 743755.
  • Mack, R.N. & Harper, J.L. (1977) Interference in dune annuals: spatial pattern and neighbourhood effects. Journal of Ecology, 65, 345363.
  • Montoya, D., Purves, D.W., Urbieta, I.R. & Zavala, M.A. (2009) Do species distribution models explain spatial structure within tree species ranges? Global Ecology and Biogeography, 18, 662673.
  • Murrell, D.J., Purves, D.W. & Law, R. (2001) Uniting pattern and process in plant ecology. Trends in Ecology & Evolution, 16, 529530.
  • Otsus, M. & Zobel, M. (2002) Small-scale turnover in a calcareous grassland, its pattern and components. Journal of Vegetation Science, 13, 199206.
  • Purves, D.W. & Law, R. (2002) Fine-scale spatial structure in a grassland community: quantifying the plant's-eye view. Journal of Ecology, 90, 121129.
  • R Development Core Team (2008) R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna.
  • Rees, M., Grubb, P.J. & Kelly, D. (1996) Quantifying the impact of competition and spatial heterogeneity on the structure and dynamics of a four-species guild of winter annuals. The American Naturalist, 147, 132.
  • Réjou-Méchain, M., Flores, O., Bourland, N., Doucet, J.L., Fétéké, R.F., Pasquier, A. & Hardy, O.J. (2011) Spatial aggregation of tropical trees at multiple spatial scales. Journal of Ecology, 99, 13731381.
  • Rusch, G.M., Wilmann, B., Klimešovà, J. & Evju, M. (2011) Do clonal and bud bank traits vary in correspondence with soil properties and resource acquisition strategies? Patterns in alpine communities in the Scandian mountains. Folia Geobotanica, 46, 237254.
  • Schmid, B. (1986) Spatial dynamics and integration within clones of grassland perennials with different growth form. Proceedings of the Royal Society of London B, 228, 173186.
  • Schmid, B. & Harper, J.L. (1985) Clonal growth in grassland perennials: I. Density and pattern-dependent competition between plants with different growth forms. Journal of Ecology, 73, 793808.
  • Seabloom, E.W., Bjørnstad, O.N., Bolker, B.M. & Reichman, O.J. (2005) Spatial signature of environmental heterogeneity, dispersal, and competition in successional grasslands. Ecological Monographs, 75, 199214.
  • Seidler, T.G. & Plotkin, J.B. (2006) Seed dispersal and spatial pattern in tropical trees. PLoS Biology, 4, e344.
  • Silander, J.A. & Pacala, S.W. (1985) Neighborhood predictors of plant performance. Oecologia, 66, 256263.
  • Slade, A.J. & Hutchings, M.J. (1987) The effects of light intensity on foraging in the clonal herb Glechoma hederacea. Journal of Ecology, 75, 639650.
  • Stoll, P. & Prati, D. (2001) Intraspecific aggregation alters competitive interactions in experimental plant communities. Ecology, 82, 319327.
  • Tilman, D. & Kareiva, P. (1997) Spatial Ecology: The Role of Space in Population Dynamics and Interspecific Interactions. Princeton University Press, Princeton.
  • Turkington, R. & Harper, J.L. (1979) The growth, distribution and neighbour relationships of Trifolium repens in a permanent pasture: II. Inter- and intra-specific contact. Journal of Ecology, 67, 219230.
  • Watt, A.S. (1947) Pattern and process in the plant community. Journal of Ecology, 35, 122.
  • Weigelt, A., Schumacher, J., Walther, T., Bartelheimer, M., Steinlein, T. & Beyschlag, W. (2007) Identifying mechanisms of competition in multi-species communities. Journal of Ecology, 95, 5364.
  • Weijschedé, J., Berentsen, R., de Kroon, H. & Huber, H. (2008) Variation in petiole and internode length affects plant performance in Trifolium repens under opposing selection regimes. Evolutionary Ecology, 22, 383397.
  • Wiegand, T. & Moloney, K.A. (2004) Rings, circles, and null-models for point pattern analysis in ecology. Oikos, 104, 209229.
  • Wildová, R., Wild, J. & Herben, T. (2007) Fine-scale dynamics of rhizomes in a grassland community. Ecography, 30, 264276.

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
jec12066-sup-0001-AppendixS1-FigS1-S8.docWord document6503K

Appendix S1. Details on the calculation of null models for interpreting the shape of univariate and bivariate O-ring statistics.

Figure S1. Univariate O-ring statistics O11(r) for the three clonal dispersal groups.

Figure S2. Bivariate O-ring statistics O12(r) for the three clonal dispersal groups.

Figure S3. Univariate O-ring statistics O11(r) for species with low dispersal abilities.

Figure S4. Bivariate O-ring statistics O12(r) for species with low dispersal abilities.

Figure S5. Univariate O-ring statistics O11(r) for species with intermediate dispersal abilities.

Figure S6. Bivariate O-ring statistics O12(r) for species with intermediate dispersal abilities.

Figure S7. Univariate O-ring statistics O11(r) for species with high dispersal abilities.

Figure S8. Bivariate O-ring statistics O12(r) for species with high dispersal abilities.

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