Functional Ecology 2010, 24, 196–207

doi: 10.1111/j.1365-2435.2009.01612.x

Are competitive effect and response two sides of the same coin, or fundamentally different? Ping Wang*,1,2, Tara Stieglitz1, Dao Wei Zhou3 and James F. Cahill Jr1 1

Department of Biology, University of Alberta, Edmonton, Alberta T6G 2E9, Canada; 2State Environmental Protection Key Laboratory of Wetland Ecology and Vegetation Restoration, College of Urban and Environmental Sciences, Northeast Normal University, Changchun, Jilin 130024, China; and 3Northeast Institute of Geography and Agroecology, Chinese Academy of Science, Changchun, Jilin 130012, China

Summary 1. The ability to suppress neighbour growth and the ability to withstand growth suppression are widely viewed as two forms of competition, competitive effect and competitive response. 2. We conducted a greenhouse experiment to determine whether these two forms of competition were functionally linked, and to determine which plant traits are associated with effect and response competitive abilities among seedlings of 22 perennial North American prairie species. We further explored the trait-function relationship by growing plants under different soil fertilities and with different neighbour species. 3. To determine competitive abilities, we used a phytometer approach with two phytometer species: Poa pratensis and Achillea millefolium grown in competition with each of the 22 focal species under low and high fertility. Root and shoot morphological traits were measured and principal component analyses were used to reduce the dimensionality of the data. Three axes were extracted, which roughly corresponded to size, root and shoot architecture. 4. The hierarchy of competitive effect ability of the target species did not vary with either soil fertility or neighbour identity, while the hierarchies of competitive response abilities were highly variable among the treatments. Competitive effect ability was closely associated with size-related traits under high nutrient conditions, and with root-related traits under low nutrient conditions. In contrast, few plant traits axes were related to competitive response. 5. These findings indicate significant differences between competitive effect and response ability. We suggest competitive effect ability is a consistent trait of a species, linked to specific plant traits. In contrast, we found little evidence to support the idea that competitive response ability is itself a species trait, and instead it appears this may be simply a collection of different ways of avoiding or tolerating competition and ⁄ or low nutrient conditions. Supporting this argument was a lack of any consistency in which traits were associated with competitive response ability. 6. We recognize the limitations of a single study of seedlings under greenhouse conditions. However, we suggest these findings indicate a need to critically examine current assumptions about plant competition, how it is defined, and the traits which control a species’ competitive ability. Key-words: comparative study, competitive effect, competitive response, functional ecology, plant competition, plant traits, soil fertility

Introduction Species differ substantially in their ability to compete for soil and light resources (Goldberg 1996; Keddy et al. 2002; Fraser & Miletti 2008). When resources are limited, a plant’s competitive ability may influence a plant’s ability to survive, grow, and reproduce (Fargione & Tilman 2006; *Correspondence author. E-mail: [email protected]

Wilson & Keddy 1986; but see Neytcheva & Aarssen 2008). Competitive ability has been suggested to include two components: competitive effect, the ability of one plant to suppress the growth of another, and competitive response, the ability of a species to resist suppression from its neighbours (Goldberg 1990, 1996). As functional traits are supposed to be surrogates of plant competitive ability (Goldberg & Fleetwood 1987; Keddy et al. 2002), we might predict that a certain plant trait or traits should be consistently associated

 2009 The Authors. Journal compilation  2009 British Ecological Society

Competitive effect and response ability 197 with each component of competitive ability. Identifying these traits could provide insight into a number of key issues in ecology, such as the control on the phylogenetic and species composition of communities (Fridley, Grime & Bilton 2007; Lamb 2008), species invasion (Suding, LeJeune & Seastedt 2004), phenotypic integration in response to multiple ecological challenges (Kembel & Cahill 2005), and prediction of functional changes in response to changing environments. However, traits that confer competitive effect ability might not also confer competitive response ability. Though the relationship between competitive effect and response is positive in some studies (Wilson & Keddy 1986; Goldberg & Fleetwood 1987; Miller & Werner 1987; Gurevitch et al. 1990; Novoplansky & Goldberg 2001; Thomsen, Corbin & D’Antonio 2006), it is uncorrelated in others (Peart 1989; Goldberg & Landa 1991; Cahill, Kembel & Gustafson 2005; Fraser & Miletti 2008). Among genotypes of Arabidopsis thaliana, size was the dominant factor determining competitive effect, but it had less influence in determining a genotype’s competitive response ability (Cahill, Kembel & Gustafson 2005). Comparative studies linking plant traits to both competitive effect and response in a number of plant species appear generally to be lacking, and are critical to understanding of what exactly ‘plant competition’ is, how plant traits influence a species’ competitive ability, and what factors influence trait dispersion in natural communities (Goldberg & Fleetwood 1987; Goldberg & Landa 1991; Johansson & Keddy 1991; Keddy, Fraser & Wisheu 1998; Hager 2004; Fraser & Miletti 2008). The most thorough studies linking plant traits to competitive ability have identified plant height and size as dominant factors (Goldberg & Fleetwood 1987; Gurevitch et al. 1990; Keddy et al. 2002; Cahill & Lamb 2007) with larger species

Table 1. Species used in this experiment, including 22 target species (T) and 2 phytometers (P). Species 1–15 competed with Poa pratensis and Achillea millefolium respectively; species 16–22 were only in competition with P. pratensis

generally able to suppress the growth of smaller species. This finding of a size-advantage for competitive ability is widely assumed to be generally true, though has recently been called into question (Schamp, Chau & Aarssen 2008). One possible explanation for the consistency of these prior findings is that most comparative studies of plant competition have been conducted under high nutrient conditions, where root competition is likely of limited importance (Gaudet & Keddy 1988; Keddy, Fraser & Wisheu 1998; Keddy et al. 2002). In natural systems, resources are variable, and root competition is often as strong, or stronger, than competition for light (Belcher, Keddy & Twolanstrutt 1995; Lamb, Shore & Cahill 2007; Lamb 2008). General body-size may be less important when competition is for soil resources, as root competition is sizesymmetric (Cahill & Casper 2000). We grew 22 perennial North American prairie species with and without competition (with different competitor) at two levels of soil fertility, measuring a number of morphological traits, to determine whether these two forms of competition were functionally linked, and to determine which plant traits are associated with effect and response competitive abilities. We further explored the trait-function relationship by growing plants under different soil fertilities and with different neighbour species to test the generality of any identified traitfunction relationships.

Materials and methods SPECIES USED

A phytometer method (Gaudet & Keddy 1988; Keddy, Fraser & Wisheu 1998) was used to assess the competitive performance of 22 plant species. These species, referred to as target species (Table 1), consisted

ID

Species

Family

Duration

Status

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Agropyron dasystachyum Artemisia frigida Artemisia ludoviciana Aster ericoides Bouteloua gracilis Bromus inermis Deschampsia cespitosa Festuca hallii Gaillardia aristata Heterotheca villosa Koeleria macrantha Linum lewisii Rumex triangulivalvis Solidago missouriensis Stipa viridula Aster laevis Campanula rotundifolia Coreopsis tinctoria Descurainia sophia Galium boreale Rumex crispus Poa pratensis Achillea millefolium

Poaceae Asteraceae Asteraceae Asteraceae Poaceae Poaceae Poaceae Poaceae Asteraceae Asteraceae Poaceae Linaceae Polygonaceae Asteraceae Poaceae Asteraceae Campanulaceae Asteraceae Brassicaceae Rubiaceae Polygonaceae Poaceae Asteraceae

Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Perennial Biennial Biennial Perennial Perennial Perennial Perennial

T T T T T T T T T T T T T T T T T T T T T P⁄T P

 2009 The Authors. Journal compilation  2009 British Ecological Society, Functional Ecology, 24, 196–207

198 P. Wang et al. primarily of perennial herbaceous species from a local native rough fescue prairie (see Lamb 2007 for more information). These species were chosen based upon the availability of viable seed, diversity of growth forms, and the ability to grow within a greenhouse. Seed were obtained through field collection and purchase through a local seed supplier (Bedrock Seed Bank, Alberta, Canada). Two species, Rumex crispis and Coreopsis tinctoria are not commonly found in this system, however were included because seed were available and their rapid growth rates would enhance the variation in plant traits among the target species. As we were testing the basic functional ecology of plant trait – competitive ability relationships, as opposed to the evolution of niche differentiation or coexistence, the source and identity of our seed and species should not influence the qualitative nature of the results. Poa pratensis and Achillea millefolium, both common in the local grassland, were chosen as two phytometer species (species with which all target species compete). We chose these species as they have dramatically different growth forms (grass vs. forb), and are widely used in other experiments by a diversity of biologists (e.g. Reader et al. 1994; Lamb & Cahill 2006). Individuals of these two phytometer species were grown in competition with individuals of 22 other species for P. pratensis and 15 species for A. millefolium. Additionally, P. pratensis, but not A. millefolium, was included as a target species. The reduced usage of A. millefolium was due to unexpectedly poor seed germination.

EXPERIMENTAL TREATMENTS

The experiment was conducted in a greenhouse in the Department of Biological Sciences Biotron facilities at the University of Alberta. Seeds of all plant species were germinated under 15 ⁄ 25 C (12 ⁄ 12 h) conditions in greenhouse flats filled with sterilized seedling starter mix (Altwin Distributors Inc., Medicine Hat, Alberta, Canada). Seedlings were transplanted into tall and narrow pots (5 · 6 · 20 cm height) 10 days following germination, with the number of individuals per pot dependent upon the competition treatment (see below). Soil inside the pots was a very low nutrient 3 : 1 sand : top soil mixture. The soil was used without amendment for the low nutrient treatment, while slow-release fertilizer (N : P : K = 15 : 15 : 15%, 8 mg cm)3) was added to the high nutrient pots. Soon after transplanting, height, length of longest leaf and leaf number of each plant were used to estimate initial plant total biomass. To estimate the initial biomass of each plant (used in calculation of relative growth rate), we measured plant height, length of longest leaf, leaf number, and dry total biomass of 15 additional plants of each species. These data (ln transformed) were then placed in a stepwise multiple regression (dry biomass served as the response variable), with the resulting regression equations (see Appendix S1 in Supporting Information) applied to the data from the plants in the main experiment to estimate initial biomass. In the no competition pots, one individual of a single target species was planted in the centre of a pot. In the competition treatments, a single individual of a single target species and a single individual of one of the two phytometer species were planted in the centre of a pot with approximate 2 cm space between them. Plant density can have strong effects on both competitive effect and response (Goldberg & Landa 1991). However, the size of this experiment necessitated us to use only two plant densities (1 or 2 plants per pot). As is seen in the results section, this planting density was sufficient to cause competitive encounters for the majority of target species. Each treatment combination (competition · resource level) was replicated six times in a fully factorial randomized block design (five replicates when A. millefolium as a phytometer due to poor germination).

Plants were watered as needed, and the ambient temperature was kept at 25 C with 16 ⁄ 8 h lighting throughout the experiment. When plant growth was such that there was the potential for shading among plants in neighbouring pots, cylinders of transparency film were placed around the perimeter of each pot to ensure the plants were unable to escape laterally from aboveground competition. Twelve circular holes of 1Æ5 cm diameter were punched in the cylinders to keep effective ventilation and avoid high temperature inside. We recognize that these growth conditions may have prevented one possible competitive response mechanism (lateral escape), though was necessary to ensure the plants in adjacent pots would not interfere with each other.

PLANT HARVEST

Plant harvest was initiated 60 days following seedling transplant. Blocks were harvested sequentially, with the total harvest lasting 10 days. We recognize that some species may possess mechanisms that influence competitive encounters which may be expressed only under specific circumstances (e.g. during regeneration phase). As our study only measured plants during the initial stages of growth, 60 days will be insufficient for capturing the full diversity of competitive traits that occur in nature. However, after 60 days there were visibly obvious fertilization and competition effects, and one species C. tinctoria, was at peak anthesis in all treatments. Some individuals of other species (A. ludoviciana, A. laevis, B. gracilis and R. triangulivalvis) were also flowering, though not enough to warrant statistical analysis. As a result, we are confident that plants were competing during this study, and if morphological traits are related to a species’ competitive abilities, we should be able to detect any such correlations. At harvest, shoots were clipped at the soil surface and immediately placed into a 4 C refrigerator. Pots were then placed in a freezer (including the roots and soil), until roots could be extracted. Roots were recovered by thawing the pots, and washing the soil from the roots. Due to the low plant density, sandy soil, and general texture ⁄ colour differences between our phytometers and target species, we have high confidence in our root extraction efficiency and the ability to assign roots to individuals (Cahill 2002).

TRAIT MEASUREMENT

In addition to the plant traits measured at the time of transplant (see above), we measured other morphological characteristics (Table 2). Immediately prior to harvest, ‘natural height’ (the distance between the soil and the upper most surface of the plant) was measured. At harvest, three mature leaves of each individual plant were randomly selected for calculating specific leaf area and were scanned using WinFOLIA v5.1a software (Regent Instruments Inc., Sainte-Foy, Quebec, Canada). Biomass of these leaves was recorded after the tissues were oven dried. Root surface area and root length of each individual plant were measured using WinRHIZO v4.1b (Regent Instruments Inc.). All remaining plant tissues (root, stem, leaf, flower) were dried in a 65 C oven for at least 48 h, and weighed separately. Over the course of the experiment, only 5 of 1102 plant died. Thus, although mortality can be an important aspect of competitive encounters, we focused our attention to biomass responses. From these data, we then calculated shoot and total biomass, total leaf area (TLA), specific leaf area (SLA = TLA ⁄ leaf weight), root ⁄ shoot weight ratio, stem ⁄ leaf weight ratio, and root ⁄ leaf area ratio for each plant. In addition, the relative growth rate (RGR) of each plant was calculated as RGR = (lnW1)lnW0) ⁄ (t1)t0) following Hunt (Hunt 1978).

 2009 The Authors. Journal compilation  2009 British Ecological Society, Functional Ecology, 24, 196–207

Competitive effect and response ability 199 Table 2. The rotated component matrix of extracted axes from 12 traits of target and phytometer species growing alone under low and high nutrient condition. The numbers in bold indicate eigenvector scores of traits are equal or more than 0.5 Low nutrient

Relative growth rate Shoot biomass Stem biomass Leaf biomass Leaf area Root biomass Root area Root length Root ⁄ shoot ratio Root ⁄ leaf area ratio Stem ⁄ leaf ratio Natural height

High nutrient

F1L (46%)

F2L (20%)

F3L (15%)

F1H (49%)

F2H (21%)

F3H (11%)

)0Æ044 0Æ305 0Æ071 0Æ446 0Æ104 0Æ602 0Æ874 0Æ894 0Æ624 0Æ900 )0Æ204 0Æ406 Root

0Æ698 0Æ650 0Æ325 0Æ769 0Æ914 0Æ676 0Æ343 0Æ272 0Æ416 )0Æ251 )0Æ107 0Æ010 Size

0Æ067 0Æ656 0Æ913 0Æ201 0Æ037 )0Æ009 0Æ138 0Æ184 )0Æ272 0Æ082 0Æ888 0Æ696 Shoot architecture

0Æ593 0Æ465 0Æ157 0Æ864 0Æ880 0Æ899 0Æ754 0Æ653 0Æ696 )0Æ070 )0Æ237 0Æ441 Size

0Æ408 0Æ864 0Æ957 0Æ319 0Æ252 0Æ062 0Æ212 0Æ303 )0Æ281 )0Æ106 0Æ897 0Æ769 Shoot architecture

)0Æ125 )0Æ025 )0Æ038 0Æ038 )0Æ209 0Æ040 0Æ499 0Æ521 0Æ121 0Æ940 )0Æ049 0Æ186 Root foraging

2 12

3 4

P. pratensis A. millefolium

5 20

3 5

6 14

2 14

Numbers in parentheses were the percentage of variance explained by each factor. Factors were named as F1L-Root, F2L- Size, F3L-Shoot architecture, F1H-Size, F2H-Shoot architecture, F3H-Root foraging. Numbers below each factor of P. pratensis and A. millefolium were the order of that factor score among 23 species (low value represents high score). Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. COMPETITIVE EFFECT AND RESPONSE

Competitive effect and response were calculated for each target species as (Keddy, Twolanstrutt & Wisheu 1994): CET ¼ PT =P CRT ¼ TP =T Where CET represents the competitive effect of a given target species, PT is phytometer biomass when grown with the target species, and P is phytometer biomass when the phytometer grew alone. Low CE values indicate strong ability of the target species to suppress the growth of the phytometer. CRT is the competitive response of a given target species, TP is target plant biomass when grown with a phytometer, and T is target plant biomass when grown alone. High CR values indicate strong ability of the target species to mitigate the cost of competition with the phytometer. CR and CE values were calculated for each target species when grown with each of the two phytometer species. Measures of CE and CR were made within each block, such that all target species have up to six replicate values when competing with P. pratensis and five replicate values when competing with A. millefolium.

STATISTICAL ANALYSIS

Competitive effect and response abilities along soil fertility and neighbour identity To determine whether CE and CR are consistent along soil fertility and neighbour identity, we conducted a series of non-parametric Spearman correlations. Specifically, we used the CE and CR ratios calculated above to determine the relative CE and CR ranking of each species in each treatment · phytometer combination. In all cases, values of ‘1’ represented the strongest competitor and larger values corresponded to decreasing competitive abilities. We then tested for

correlations in rankings between the two fertilizer treatments and the two phytometer treatments.

Correlation between competitive effect and response To detect whether a species’ competitive effect ability was correlated with its competitive response ability, we tested for a correlation between competitive effect and response of target species. This was done for each phytometer and nutrient level separately, and then with the data combined across these treatments. These correlations were done both of the CE and CR ratio data (using Pearson correlations), and on the CE and CR rankings calculated above (Spearman correlations).

Linking plant traits to competitive abilities Due to the large number of plant traits measured, and the strong covariance among the data (such as height and shoot biomass, root area and root length, etc.), we used principal components analysis (PCA) for data reduction. Separate PCAs were conducted for each fertilization treatment using SPSS 15.0 (SPSS Inc., Chicago, Illinois, USA). Plant trait data were collected for each target species when grown without competition, and thus represents the morphology of the species without competition. All measured traits (Table 2) were included untransformed in the analyses. Axes with eigenvalues greater than 1 were extracted and rotated (Varimax method), with a Kaiser-MeyerOlkin (KMO) test (0Æ64 ⁄ 0Æ72, low ⁄ high nutrient) (Kaiser 1970, 1974). To determine which traits axes (suites of plant traits) were most strongly associated with competitive effect and response abilities of the target species, we conducted a stepwise multiple-regression for each of the four phytometer · nutrient combinations using SPSS 15.0. In these analyses, we used the CE and CR ratios (see above) as our measures of competitive ability. In each regression, the three axes served as independent variables, and competitive effect or response served as the dependent variable.

 2009 The Authors. Journal compilation  2009 British Ecological Society, Functional Ecology, 24, 196–207

200 P. Wang et al.

Over the course of the experiment, 5 of 1102 plants died, and large amounts of soil ran out of the bottoms of additional 10 pots. The data from both of these sets of pots were excluded from all data analysis. Soil fertility had clear effects on plant growth, increasing average plant size nearly 10-fold, and plants were generally smaller when grown with a neighbour than when grown individually (Figs 1 and 2).

were less suppressed by their neighbours. This result was consistent in the analysis of both ranks and competitive effect and response ratios. When the data were analyzed separately for the two phytometer species, competitive effect and competitive response were correlated at both nutrient levels for A. millefolium, but only under low nutrient conditions for P. pratensis (Fig. 5). These results were generally consistent regardless of whether raw or rank CE and CR measures were used (Fig. 5).

COMPETITIVE EFFECT AND RESPONSE ABILITIES

PLANT TRAITS

Results

ALONG SOIL FERTILITY AND NEIGHBOUR IDENTITY

A species ranking of competitive effect ability were generally consistent among fertilization treatments and across phytometer (Figs 3 and 4). Specifically, the rank of competitive effect ability under low nutrient conditions was correlated with a species rank under high nutrient conditions (Fig. 3). Similarly competitive effect rankings for a species were consistent regardless of which phytometer was used under high nutrients, but not under low nutrient conditions (Fig. 4). Visual examination of the data suggests the lack of correlation in ranks among phytometers at low nutrients was driven by a single outlying data point. In contrast to competitive effect rankings, competitive response rankings were highly variable, with no correlations among nutrient levels or phytometers (Figs 3 and 4).

RELATIONSHIP BETWEEN COMPETITIVE EFFECT AND RESPONSE ABILITIES

We found a generally negative relationship between competitive effect and response at both nutrient levels (Fig. 5), indicating that species which suppressed more to their neighbours

When species were grown alone in the low nutrient treatment, their traits were extracted into three axes which explain 81% of the variation in plant traits. The interpretation of PCA axes is always a subjective process, and most variables contribute some weighting to each axis. However, we have tried to focus on the suites of factors with the largest weightings on each axis, and thus interpret the axes from the unfertilized trait data as representing a ‘root axis’ that may represent the ability to capture soil resources. This axis is correlated with high root investment (biomass, area, length), larger root : shoot ratios, and larger root: leaf area ratio (F1L; 46% of variation); a ‘size axis’ representing general body size and growth rate (F2L; 20%); and a ‘shoot architecture axis’ consisting primarily of plant height and stem biomass. (F3L; 15%) (Table 2). In the high nutrient treatment, plant traits were again extracted onto three axes, explaining 82% of the variation in plant traits. However, there were some differences in which traits were most strongly correlated with these axes (Table 2). We interpret the axes as a ‘size axis’ (F1H 49% of variation) influenced by a combination of traits associated with the first two axes in the unfertilized plants (both

(a)

(b) Fig. 1. The total biomass of target species growing alone, growing with P. pratensis, growing with A. millefolium at low (a) and high (b) nutrient levels. Species name were showed in Table 1. Species 1–15 competed with both P. pratensis and A. millefolium, while species 16–22 only competed with P. pratensis. Error bars represent Mean + 1Æ0 SE. Note different axes scales in the two figures. The difference between the ‘alone’ and ‘competition’ columns is indicative of the competitive response of a species.  2009 The Authors. Journal compilation  2009 British Ecological Society, Functional Ecology, 24, 196–207

Competitive effect and response ability 201 (a)

(b)

(c)

Fig. 2. The total biomass of P. pratensis (a, b) and A. millefolium (c, d) growing alone (line), growing with target species (bars) at low and high nutrient levels. Species name were showed in Table 1. Species 1–15 competed both with P. pratensis and A. millefolium, while species 16–22 only competed with P. pratensis. Error bars represent Mean + 1Æ0 SE. Note different axes scales in the figures. Values below the horizontal line indicate competition, and those above the line indicate facilitation.

(d)

root allocation and overall plant size); a ‘shoot architecture axis’ [F2H (21%)] similar to the third axis in the unfertilized treatment; and a ‘‘root foraging axis’[F3H (11%)] positively correlated with both root length and area, and root: leaf area ratio. In a visual examination of the weighting of each species on the extracted axes (as opposed to each trait as described above), it was apparent that the two phytometers differed greatly. Specifically, P. pratensis grew larger with a bigger root system and greater height than most of species under both nutrient levels, while A. millefolium was relatively large high for size, but in the middle or lower of species on the root and shoot architecture axes in both nutrient treatments (Table 2).

RELATIONSHIPS BETWEEN COMPETITIVE ABILITY AND PLANT TRAITS

Competitive effect ability was well described by the plant traits measured here (Fig. 6), however the exact relationships between the PCA axes and competition varied as a function of which phytometer served as the competitor. When competing with P. pratensis, the competitive effect of the target species was related with their size axis under low nutrients (Fig. 6c), and both plant size and shoot architecture axes

under high nutrient level (Fig. 6d,f). When competing with A. millefolium, the competitive effects of target species were related with their root trait axis under low nutrient level (Fig. 6a), and related with size trait axis under high nutrient level (Fig. 6d). In contrast, competitive response ability of the target species was poorly explained by these plant traits (Fig. 7). Of the four multiple-regressions run, only one was significant, where the competitive response of A. millefolium was significantly related to both the root and size axes in the low nutrient treatment (Fig. 7a,c).

Discussion We found mixed support for the idea that competitive effect and response are each discrete aspects of a plant’s competitive ability. The hierarchy of competitive effect ability was consistent between soil fertility and neighbour identity treatment, while competitive response ability was not (Figs 3 and 4). Further, the morphological traits measured here were able to predict competitive effect abilities (Fig. 6), but few traits were consistently related to competitive response abilities (Fig. 7). As a whole, these data suggest that competitive effect ability exists as a discrete plant ‘trait’, controlled by a predictable set of other plant traits. In contrast, the concept of ‘competitive

 2009 The Authors. Journal compilation  2009 British Ecological Society, Functional Ecology, 24, 196–207

202 P. Wang et al. (a)

(a)

(b)

(b)

Fig. 3. The ranks of competitive effect (a) and response (b) to P. pratensis (CEP, CRP, solid circles), and to A. millefolium (CEA, CRA, hollow circles) in low and high nutrient conditions. Numbers outside the brackets are Spearman coefficients, and numbers inside brackets are P values.

response ability’ does not appear to represent any single and consistent aspect of a plant, and there do not appear to be a discrete set of traits which confer enhanced response ability. As discussed below, we suggest that validity of current assumptions about these dual forms of plant competition require further consideration.

DIFFERENCES OF COMPETITIVE EFFECT AND RESPONSE ABILITY

Competitive effect and response varied in their responses to changes in fertilization and the identity of the phytometer. In general, competitive effect appeared fairly stable, such that a species’ competitive effect ability when grown with one phytometer was correlated to its effect ability against the second phytometer (Fig. 4; Fraser & Miletti 2008; Goldberg & Landa 1991). Similarly, competitive effect ability at low nutrients was correlated with that under high nutrient conditions (Fig. 3) (Keddy, Twolanstrutt & Wisheu 1994; Keddy et al. 2002; Fraser & Miletti 2008). In contrast, but similar to con-

Fig. 4. The ranks of competitive effect (a) to P. pratensis (CEP) with to A. millefolium (CEA), and competitive response (b) to P. pratensis (CRP) with to A. millefolium (CRA) under low (solid circles) and high (hollow circles) nutrient condition. Numbers outside the brackets are Spearman coefficients, and numbers inside brackets are P values.

clusions of other researchers (Keddy, Twolanstrutt & Wisheu 1994; Novoplansky & Goldberg 2001; Suding & Goldberg 2001; Fraser & Miletti 2008), competitive response abilities were not correlated as a function of phytometer nor nutrient level (Figs 3 and 4). These findings suggest that competitive effect ability is a general characteristic of a species, while competitive response ability is highly labile and contingent upon specific conditions. As a result, we suggest that strong effect competitors are likely to do well in most communities in which competition occurs, and this could contribute to species invasion (Suding, LeJeune & Seastedt 2004; Thomsen, Corbin & D’Antonio 2006). In contrast, competitive response ability is unlikely to serve as a general plant strategy, as the conditions under which a particular species has strong competitive response abilities appear to be narrow. Such variation between competitive effect and response may have important evolutionary implications as selection for competitive effect should impact growth throughout a species’ range, while selection for competitive response should have impact in only

 2009 The Authors. Journal compilation  2009 British Ecological Society, Functional Ecology, 24, 196–207

Competitive effect and response ability 203 (a)

(c)

(b)

(d)

Fig. 5. The relationship between competitive effect (CE) and response (CR) of target species competing with P. pratensis (solid) and A. millefolium (hollow) at low (left) and high (right) nutrient level. The a and b are the Pearson correlation analysis on raw CE and CR data, while c and d are the Spearman correlation analysis on rank of CE and CR. Numbers outside the brackets are Spearman coefficients, and numbers inside brackets are P values.

a narrow range of conditions. The sensitivity of competitive response ability to soil nutrient and neighbour identity may contribute to the variable results of studies (i.e. Gaudet & Keddy 1995; Fraser & Keddy 2005; Lamb & Cahill 2006; Engel & Weltzin 2008) testing whether natural communities are structured by competition. Despite these differences between competitive effect and response abilities, these two competitive forms were correlated (Fig. 5), such that better effect competitors were also generally better response competitors (Wilson & Keddy 1986; Goldberg & Fleetwood 1987; Miller & Werner 1987; Gurevitch et al. 1990; Novoplansky & Goldberg 2001; Thomsen, Corbin & D’Antonio 2006). This finding contrasts with other studies, where no such correlations were found among species (Peart 1989; Goldberg & Landa 1991; Fraser & Miletti 2008), or genotypes (Cahill, Kembel & Gustafson 2005). We do not have a clear explanation for this discrepancy, though we found that those correlations occurred more frequently in infertile soil (Wilson & Keddy 1986; Goldberg & Fleetwood 1987; Miller & Werner 1987) and ⁄ or limited space (Gurevitch et al. 1990; Thomsen, Corbin & D’Antonio 2006) than in fertile soil (Goldberg & Landa

1991; Fraser & Miletti 2008) or unlimited space (Peart 1989). In this experiment, higher correlation coefficients and significance demonstrated in lower soil nutrient condition also supported this finding (Fig. 5). Once strong competitors acquired most of limited resources, their neighbours could do nothing to improve growth without minimal resources even though kinds of response strategies they may have, which result in worse effect competitors being worse response competitors, and vice versa.

PLANT TRAITS AND COMPETITIVE ABILITY

We found no single set of traits that could totally explain either competitive effect or competitive response abilities under all environments (Figs 6 and 7). At high nutrient conditions, competitive effect ability was strongly related to plant size under high nutrient condition (Fig. 6). This finding is consistent with other studies (Goldberg & Fleetwood 1987; Gurevitch et al. 1990; Keddy et al. 2002), and is intuitive. Bigger plants usually have an advantage in capturing light (Grime 1977; Wildova et al. 2007) which is generally limiting under high nutrient conditions (Belcher, Keddy & Twolan-

 2009 The Authors. Journal compilation  2009 British Ecological Society, Functional Ecology, 24, 196–207

204 P. Wang et al. (a)

(b)

(c)

(d)

(e)

(f)

Fig. 6. The multiple-regression (stepwise method) of competitive effect to P. pratensis (CEP, solid) and to A. millefolium (CEA, hollow) on extracted trait factors (by principal component analysis method): root (a and b), size (c and d), shoot architecture (e and f) under low (left) and high (right) nutrient condition.* and *** indicates significant regression at P = 0Æ05 and 0Æ001 level, ns indicate the independent variables were excluded from the regression.

strutt 1995; Cahill 1999). We suggest the accumulation of evidence leads to a general conclusion that bigger plants are best able to suppress the growth of other plants when resources are in high supply. In contrast, although competitive effect ability under low nutrient conditions was correlated with that under high nutrient conditions (Figs 2 and 5), plant size only predicted

competitive effect ability when P. pratensis was the phytometer (Fig. 6). When A. millefolium was used, a general measure of roots was associated with competitive effect ability. At a most basic level, this finding indicates that the trait-function relationship for competitive effect is contingent upon the specific set of species undergoing competition and the fertility of the soil. This idea has been presented before (Gurevitch et al.

 2009 The Authors. Journal compilation  2009 British Ecological Society, Functional Ecology, 24, 196–207

Competitive effect and response ability 205 (a)

(b)

(c)

(d)

(e)

(f)

Fig. 7. The multiple-regression (stepwise method) of competitive response to P. pratensis (CRP, solid) and to A. millefolium (CRA, hollow) on extracted trait factors (by principal component analysis method): root (a and b), size (c and d), shoot architecture (e and f) under low (left) and high (right) nutrient condition. * and *** indicates significant regression at P = 0Æ05 and 0Æ001 level, ns indicate the independent variables were excluded from the regression.

1990; Dekker, Verkerk & den Ouden 2008), and the empirical support provided here indicates that the species composition of a community may influence the strength of competition found within the community. We suggest the differences among these two phytometers under low nutrient conditions were due to differences in root growth. P. pratensis had an extensive fibrous root system (the 2nd largest root system

among all 23 species), while A. millefolium has a much smaller root system (the 12th largest system among 23). Cahill & Casper (2000) observed that increased root biomass increased the intensity of root competition over a very narrow range, and a point further increases in root biomass resulted in no increase in the intensity of root competition. Following this result, we suggest that increased root growth would be more successful

 2009 The Authors. Journal compilation  2009 British Ecological Society, Functional Ecology, 24, 196–207

206 P. Wang et al. in suppressing the growth of species which produce relatively few roots, such as A. millefolium, but not those which produce substantial root growth (P. pratensis). It is also possible that the differences observed are due to physiological, rather than morphological traits, and this area requires further study. Competitive response was not correlated with any trait-axis in all but one of the four nutrient · phytometer combinations. Our results are consistent with the ideas put forth by Keddy, Fraser & Wisheu (1998), in which enhanced competitive response abilities may result from a number of different traits. This argument is further strengthened by our finding that competitive response was highly variable among nutrient treatments, suggesting that there are multiple ‘solutions’ to enhanced competitive response ability for each unique set of environmental conditions. As a result, we suggest that competitive response ability is the product of numerous plant traits, which can occur in a large number of possible combinations. Although it is possible that we simply did not measure the ‘right’ trait, we suggest that the accumulation of data on this issue (Table 2 and Fig. 7), also leads to a general conclusion: there is no single mechanism, nor set of traits which lead to enhanced competitive response ability.

WHAT IS COMPETITIVE ABILITY?

The idea that plants possess a ‘competitive ability’, and that this can vary among species is a central concept in plant ecology (Grime 1979; Tilman 1988; Goldberg 1996; Keddy et al. 2002). This idea is at the core of fundamental theories of community organization and the concept of plant strategies (MacArthur & Levins 1967; Grime 1979). Competitive ability is often regarded as a combination of two components: competitive effect and response ability (Goldberg & Landa 1991). Of these, competitive effect is the form most widely discussed, and studied. We suggest that competitive effect ability might be a ‘real’ ability for competition. It is relatively stable with resource level and neighbour’s identity. However, the response ability of species changed with resource level and neighbour’s identity. In other words, the plants’ competitive response abilities are not only related to the traits of target plant species, but also those of neighbours (Pennings et al. 2005). We suggest our understandings of the details of how plants actually compete are not as well understood as widely thought. A related difficulty arises in trying to understand competitive response. Competitive response is generally poorly explained in plant trait studies yet is highly variable within and among species. One possible explanation for the lack of ability to link competitive response to specific plant traits is if the term ‘competitive response’ is simply a synonym for stress-tolerance. We suggest low resource levels, derived by biotic or abiotic causes, will impose stress on the plants. As a result, the strategies plants possess for competitive response ability may be the large diversity of options plants have for tolerating low levels of different resources, and it would not be surprising to find a lack of consistency between specific morphological plants traits and competitive response ability.

In conclusion, we suggest that the nature of plant competition results in few strategies for suppressing others, while there may be a large number of strategies available to avoid or reduce the impacts of competition. The lack of consistent correlations between plant traits and competitive response ability raises doubts about whether competitive response is in fact a single concept. Further research is needed to determine whether this concept is redundant with the idea of stress-tolerance, and if so, it may be time the term was abandoned. These ideas have important implications for a number of questions in conservation biology and basic understanding of the structure of natural communities.

Acknowledgements We thank K. Thompson, G. G. McNickle, J. Bennett, S. White for their valuable comments. Thanks to the biotron staff in Department of Biology in University of Alberta. This work was supported by a Project of the State Administration of Foreign Experts Affairs to P. Wang, a Natural Science and Engineering Research Council Discovery grant to J. F. Cahill.

References Belcher, J.W., Keddy, P.A. & Twolanstrutt, L. (1995) Root and shoot competition intensity along a soil depth gradient. Journal of Ecology, 83, 673–682. Cahill, J.F. (1999) Fertilization effects on interactions between above- and belowground competition in an old field. Ecology, 80, 466–480. Cahill, J.F. (2002) What evidence is necessary in studies which separate root and shoot competition along productivity gradients? Journal of Ecology, 90, 201–205. Cahill, J.F. & Casper, B.B. (2000) Investigating the relationship between neighbor root biomass and belowground competition: field evidence for symmetric competition belowground. Oikos, 90, 311–320. Cahill, J.F., Kembel, S.W. & Gustafson, D.J. (2005) Differential genetic influences on competitive effect and response in Arabidopsis thaliana. Journal of Ecology, 93, 958–967. Cahill, J.F. & Lamb, E.G. (2007) Interactions between root and shoot competition and plant traits. Hortscience, 42, 1110–1112. Dekker, M., Verkerk, P.J. & den Ouden, J. (2008) Target species identity is more important than neighbor species identity. Forest Ecology and Management, 255, 203–213. Engel, E.C. & Weltzin, J.F. (2008) Can community composition be predicted from pairwise species interactions? Plant Ecology, 195, 77–85. Fargione, J. & Tilman, D. (2006) Plant species traits and capacity for resource reduction predict yield and abundance under competition in nitrogen-limited grassland. Functional Ecology, 20, 533–540. Fraser, L.H. & Keddy, P.A. (2005) Can competitive ability predict structure in experimental plant communities? Journal of Vegetation Science, 16, 571– 578. Fraser, L.H. & Miletti, T.E. (2008) Effect of minor water depth treatments on competitive effect and response of eight wetland plants. Plant Ecology, 195, 33–43. Fridley, J.D., Grime, J.P. & Bilton, M. (2007) Genetic identity of interspecific neighbours mediates plant responses to competition and environmental variation in a species-rich grassland. Journal of Ecology, 95, 908–915. Gaudet, C.L. & Keddy, P.A. (1988) A comparative approach to predicting competitive ability from plant traits. Nature, 334, 242–243. Gaudet, C.L. & Keddy, P.A. (1995) Competitive performance and species distribution in shoreline plant-communities - a comparative approach. Ecology, 76, 280–291. Goldberg, D.E. (1990) Components of resource competition in plant communities. Perspectives in Plant Competition (eds J. Grace & D. Tilman), pp. 27– 49. Academic Press, New York. Goldberg, D.E. (1996) Competitive ability: definitions, contingency and correlated traits. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences, 351, 1377–1385. Goldberg, D.E. & Fleetwood, L. (1987) Competitive effect and response in four annual plants. Journal of Ecology, 75, 1131–1143.

 2009 The Authors. Journal compilation  2009 British Ecological Society, Functional Ecology, 24, 196–207

Competitive effect and response ability 207 Goldberg, D.E. & Landa, K. (1991) Competitive effect and response hierarchies and correlated traits in the early stages of competition. Journal of Ecology, 79, 1013–1030. Grime, J.P. (1977) Evidence for the existence of three primary strategies in plants and its relevance to ecological theory. American Naturalist, 111, 1169–1194. Grime, J.P. (1979) Plant Strategies and Vegetation Processes. Wiley, London. Gurevitch, J., Wilson, P., Stone, J.L., Teese, P. & Stoutenburgh, R.J. (1990) Competition among old-field perennials at different levels of soil fertility and available space. Journal of Ecology, 78, 727–744. Hager, H.A. (2004) Competitive effect versus competitive response of invasive and native wetland plant species. Oecologia, 139, 140–149. Hunt, R. (1978) Plant Growth Analysis. Arnold, London. Johansson, M.E. & Keddy, P.A. (1991) Intensity and asymmetry of competition between plant pairs of different degrees of similarity - an experimental study on 2 guilds of wetland plants. Oikos, 60, 27–34. Kaiser, H.F. (1970) A second generation little jiffy. Psychometrika, 35, 401– 415. Kaiser, H.F. (1974) An index of factorial simplicity. Psychometrika, 39, 31–36. Keddy, P., Fraser, L.H. & Wisheu, I.C. (1998) A comparative approach to examine competitive response of 48 wetland plant species. Journal of Vegetation Science, 9, 777–786. Keddy, P.A., Twolanstrutt, L. & Wisheu, I.C. (1994) Competitive effect and response rankings in 20 wetland plants - Are they consistent across 3 environments. Journal of Ecology, 82, 635–643. Keddy, P., Nielsen, K., Weiher, E. & Lawson, R. (2002) Relative competitive performance of 63 species of terrestrial herbaceous plants. Journal of Vegetation Science, 13, 5–16. Kembel, S.W. & Cahill, J.F. (2005) Plant phenotypic plasticity belowground: a phylogenetic perspective on root foraging trade-offs. American Naturalist, 166, 216–230. Lamb, E.G. (2007) Competition and Community Structure in a Rough Fescue Grassland. PhD Thesis, University of Alberta, Canada. Lamb, E.G. (2008) Direct and indirect control of grassland community structure by litter, resources, and biomass. Ecology, 89, 216–225. Lamb, E.G. & Cahill, J.F. (2006) Consequences of differing competitive abilities between juvenile and adult plants. Oikos, 112, 502–512. Lamb, E.G., Shore, B.H. & Cahill, J.F. (2007) Water and nitrogen addition differentially impact plant competition in a native rough fescue grassland. Plant Ecology, 192, 21–33. MacArthur, R. & Levins, R. (1967) The limiting similarity, convergence, and divergence of coexisting species. American Naturalist, 101, 377–385. Miller, T.E. & Werner, P.A. (1987) Competitive effects and responses between plant-species in a 1st-year old-field community. Ecology, 68, 1201–1210. Neytcheva, M.S. & Aarssen, L.W. (2008) More plant biomass results in more offspring production in annuals, or does it? Oikos, 117, 1298–1307. Novoplansky, A. & Goldberg, D.E. (2001) Effects of water pulsing on individual performance and competitive hierarchies in plants. Journal of Vegetation Science, 12, 199–208.

Peart, D.R. (1989) Species interactions in a successional grassland .2. colonization of vegetated sites. Journal of Ecology, 77, 252–266. Pennings, S.C., Clark, C.M., Cleland, E.E., Collins, S.L., Gough, L., Gross, K.L., Milchunas, D.G. & Suding, K.N. (2005) Do individual plant species show predictable responses to nitrogen addition across multiple experiments? Oikos, 110, 547–555. Reader, R.J., Wilson, S.D., Belcher, J.W., Wisheu, I., Keddy, P.A., Tilman, D., Morris, E.C., Grace, J.B., McGraw, J.B., Olff, H., Turkington, R., Klein, E., Leung, Y., Shipley, B., Vanhulst, R., Johansson, M.E., Nilsson, C., Gurevitch, J., Grigulis, K. & Beisner, B.E. (1994) Plant competition in relation to neighbor biomass - an intercontinental study with Poa Pratensis. Ecology, 75, 1753–1760. Schamp, B.S., Chau, J. & Aarssen, L.W. (2008) Dispersion of traits related to competitive ability in an old-field plant community. Journal of Ecology, 96, 204–212. Suding, K.N. & Goldberg, D.E. (2001) Do disturbances alter competitive hierarchies? mechanisms of change following gap creation. Ecology, 82, 2133– 2149. Suding, K.N., LeJeune, K.D. & Seastedt, T.R. (2004) Competitive impacts and responses of an invasive weed: dependencies on nitrogen and phosphorus availability. Oecologia, 141, 526–535. Thomsen, M.A., Corbin, J.D. & D’Antonio, C.M. (2006) The effect of soil nitrogen on competition between native and exotic perennial grasses from northern coastal California. Plant Ecology, 186, 23–35. Tilman, D. (1988) Plant Strategies and the Dynamics and Structure of Plant Communities. Princeton University Press, Princeton. Wildova, R., Gough, L., Herben, T., Hershock, C. & Goldberg, D.E. (2007) Architectural and growth traits differ in effects on performance of clonal plants: an analysis using a field-parameterized simulation model. Oikos, 116, 836–852. Wilson, S.D. & Keddy, P.A. (1986) Species competitive ability and position along a natural stress disturbance gradient. Ecology, 67, 1236–1242. Received 23 April 2009; accepted 15 June 2009 Handling Editor: Ken Thompson

Supporting information Additional Supporting Information may be found in the online version of this article: Appendix S1. Regression equations of initial plant biomass. 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 material) should be directed to the corresponding author for the article.

 2009 The Authors. Journal compilation  2009 British Ecological Society, Functional Ecology, 24, 196–207