Scale dependence of effective specialization: its analysis and implications for estimates of global insect species richness

Estimates of global insect species richness are sometimes based on effective specialization, a calculation used to estimate the number of insect species that is restricted to a particular tree species. Yet it is not clear how effective specialization is influenced by spatial scale or characteristics of the insect community itself (e.g. species richness). We investigated scale dependence and community predictors of effective specialization using 15,907 beetles (583 species) collected by insecticide fogging from the crowns of 96 trees (including 32 Quercus trees) located in Ohio and Indiana. Trees were distributed across 24 forest stands (∼1 ha) nested within six sites (∼10–100 km²) and two ecoregions (> 1000 km²). Using paired-sample randomization tests, we found that effective specialization ( f _k ) exhibited negative scale-dependence in early (May–June 2000) and late (August–September 2000) sampling periods. Our average effective specialization ( F ) values — those that are comparable to Erwin's (1982) estimates — ranged from 19% to 97%, and increased as spatial scale decreased. We also found that beetle species richness and the number of shared beetle species across host trees were significant and consistent negative predictors of F . This shows that increases in spatial scale, species richness, and the number of trees (and/or tree species) all coincide with decreases in effective specialization. Collectively, our results indicate that estimates of global insect species richness based on effective specialization at a single spatial scale are overestimating the magnitude of global insect species richness. We propose that scale dependence should be promoted to a central concept in the research program on global estimates of species richness.