On Assessing the Significance of Interspecific Competition to Guild Structure

Two varieties of empirical information have been collected for guilds of animals: first, information related to functional organization as revealed by field manipulations; second, descriptive data that are used to characterize guilds by measuring such quantities as niche breadth, niche segregation, niche overlap, etc. Only functional data can be used to test proximate theoretical models because these data can be organized such that specific hypotheses can be falsified. Descriptive data can only be consistent in the predictions of hypotheses, although inconsistencies clearly imply falsification of such hypotheses. A three—phase procedure that can be used to specify the influence of interspecific competition on guild structure is presented. These phases are: (1) measurement of overlap of resource usage (° descriptive data); (2) establishment of interaction, which entails the experimental demonstration that the presence of other species leads to altered patterns of resource usage; and (3) demonstration of depressive competition, in which the interaction is shown to be responsible for altered characteristics of populations (e.g., population density, age—structure, etc.) or individuals within the populations (e.g., fecundity, mortality, relative fitness). The potential drawbacks associated with the experimental demonstration of interaction or depressive competition are listed. These include the relative strengths of interspecific and intraspecific density effects, choices of times and locations for perturbations, possible effects of competitive mutualism, frequency—dependent predation, and choice of variables with which to assess competition. Depressive competition has been adequately shown to occur in a number of studies, particularly those that concentrate on small guilds (less than six species). Functional organization will be more difficult to show or delineate in guilds consisting of larger numbers of species because such guilds are more likely to be influenced by alternative processes.