Competitive performance and species distribution in shoreline plant communities: a comparative approach

Ecology, Jan, 1995 by Connie L. Gaudet, Paul A. Keddy

INTRODUCTION

The assumption that competitive ability varies along natural environmental gradients of fertility and standing crop is implicit in many general models of plant community structure (Ellenberg and Mueller-Dombois 1974, Grime 1977, 1979, Tilman 1982, 1985, 1988, Austin 1986, Ellenberg 1988, Southwood 1988, Keddy 1990a, b, Keddy and MacLellan 1991) such that the interaction between competitive ability, resource availability, and disturbance determines vegetation composition (Austin 1986, Campbell and Grime 1992, Turkington et al. 1993). Despite its central importance to plant community theory, the basic relationship between competitive ability and distribution along natural environmental gradients has rarely been tested empirically at a scale that enables generalization beyond the particular study conditions or species.

It is becoming increasingly apparent that properties or processes emergent at one level or scale of interaction may not be predictable from lower scales of interaction (e.g., O'Neill et al. 1986, Allen 1987, Moore and Keddy 1989, Reader et al. 1994). Field experiments may be conducted at many different scales, and at many different locations along an environmental gradient, yielding what often appear to be contradictory results in the development of general competition theory (e.g., Wilson and Shay 1990, Wilson and Tilman 1991, Campbell and Grime 1992, Goldberg and Barton 1992, Turkington et al. 1993). The current study therefore emphasizes a test of the generality of the relationship between competitive ability and distribution at a broad community scale, across a large number of species and sites, and a broad range of resource availability. To enable a test of this relationship at such a broad scale, we explore the use of a relatively novel screening or comparative approach (Grime 1977) for evaluating relative competitive performance for a large number of species (Gaudet and Keddy 1988). Specifically, we tested the hypothesis that species distributions along natural environmental gradients of fertility and standing crop are related to their relative competitive performance. We provide a simple, direct test of the importance of this predicted relationship as the basis for further development of predictive (sensu Peters 1982) theory about community pattern and process.

Shoreline plant communities are highly suited as a test of this relationship. Species composition has been shown to vary along natural environmental gradients of stress and disturbance related to wave exposure in shoreline communities (Keddy 1983, 1984). Grace and Wetzel (1981) and Snow and Vince (1984) present evidence that competition produces zonation patterns in marshes, and Wilson and Keddy (1986a) have shown that, for specific sites, competition intensity increases along a gradient of soil organic content and that the mean position of species along this gradient is related to relative competitive performance (Wilson and Keddy 1986b). This last experiment had only seven species and a single environmental measure (percent soil organic content). Whereas mineral nutrient levels can be directly considered as environmental gradients of resource availability, standing crop and percent organic content are emergent community properties (Austin and Smith 1989). Standing crop is a complex response variable that integrates abiotic factors such as resource availability, exposure, and disturbance, and biotic interactions (Grime 1979, Wilson and Keddy 1986b), and this measure is emerging as a key state variable in predictive community ecology along which important community properties such as competition intensity and species richness vary (e.g., Grime 1979, Wilson and Keddy 1986b). It is, however, difficult to interpret the underlying contribution of environmental resource gradients in studies using only standing crop as a predictor variable, or to compare across studies using different measures of productivity or fertility. Here, we provide a direct test of the relationship between competitive performance and field distribution not only along the standing crop and percent soil organic content gradient, but along environmental resource gradients of fertility (soil phosphorus, magnesium, potassium, nitrate), and pH.

The usual approach to measuring competitive performance is to assess the performance (relative yield) of species in all possible pair-wise combinations (e.g., Harper 1977, de Wit 1960). This approach is logistically demanding and limits the number of species that can be tested. To increase the number of species that could be examined, we used a "phytometer" or indicator (Clements 1933) approach based on a modified additive design (Harper 1977) in which the relative competitive performance of a species is evaluated by measuring the relative performance to suppress the growth of a common indicator or background species (phytometer). Although relatively novel in plant community ecology, a bioassay or indicator approach is a useful approach for examining a large number of species with diverse morphologies (e.g., Grime and Hunt 1975, Reader and Southwood 1981, McCanny et al. 1990). The advantage of the phytometer approach is that it provides directly, values of competitive performance on an arbitrary scale for a large number of species. Though this method has received little experimental attention in studies of plant competition (except see Welbank 1963, Mitchley and Grubb 1986), it is highly suited to the current research with emphasis on a comparative screening of a large number of species.