Community structure is often thought to be determined by interspecific competition. Critics of this idea state that we must first demonstrate structure with null models to test whether structure indeed exists. Here, we use 179 predatory and saprophagic species of flies in the family Muscidae (Diptera) that were captured in Malaise traps in six locations in Paraná State during one year of study. To test community structure, we generated five presence-absence (1-0) matrices: two by trophic guilds, two by habitat, and one general matrix (taxonomic). Two indices of co-occurrence (C) and covariance (V) of species were calculated in the matrices developed through 5,000 Monte Carlo randomizations. Randomizations followed two different assumptions: 1) fixed number of species per location, and 2) constant proportions of species at all locations. Comparisons with null-model communities showed that "taxonomic" species assemblage had a false structure, while "ecological" species assemblages had true structure. Although the ecological assemblages are consistent with the theory of interspecific competition as a cause of community structure, it remains possible that other causes of structure may also be important.
Randomization; species co-ocurrence; coexistence; competition; community ecology
