Taxonomic and functional diversity of insect herbivore assemblages associated with the canopy-dominant trees of the Azorean native forest

Abstract

Oceanic islands have been providing important insights on the structuring of ecological communities and, under the context of the present biodiversity crisis, they are paramount to assess the effects of biological invasions on community assembly. In this study we compare the taxonomic and functional diversity of insect herbivore assemblages associated with the dominant tree species of Azorean native forests and investigate the ecological processes that may have originated current patterns of plant-herbivore associations. Five dominant trees-Erica azorica, Ilex perado subsp. azorica, Juniperus brevifolia, Laurus azorica and Vaccinium cylindraceum-were sampled in the remnants of the native forest of Terceira Island (Azores) using a standardised methodology. The taxonomic and functional diversity of insect herbivore assemblages was assessed using complementary metrics and beta diversity partitioning analysis (species replacement and richness differences) aiming to evaluate the variation in insect herbivore assemblages within and between the study plant species. Sixty two insect species, mostly bugs (Hemiptera) and caterpillars (Lepidoptera), were found in the five study plants with indigenous (endemic and native non-endemic) insects occurring with higher species richness and abundance than introduced ones. Species replacement was the most important component of insect herbivore taxonomic beta diversity while differences in trait richness played a major role on functional beta diversity. The endemic E. azorica stands out from the other study plants by having associated a very distinct insect herbivore assemblage with a particular set of functional attributes, mainly composed by large bodied and long shaped species that feed by chewing. Despite the progressive biotic homogenization witnessed in the Azores during the last few decades, several strong associations between the endemic trees and their indigenous insect herbivores remain.

Fig 1. Taxonomic and functional diversity indices of insect herbivores associated with the different Azorean native forest trees.

(A) Species richness estimated with Jackknife2, (B) species evenness measured with the index Evar, (C) functional richness estimated with Jackknife 2 (D) functional evenness measured with the index R_U, (E) functional originality and (F) functional specialization. Significant differences of the diversity metrics between study plants were evaluated by linear mixed models followed by Tukey HSD tests. Study plants names were abbreviated to genus. Different letters associated with study plant names mean that they are significantly different from each other for that specific diversity metric.

Fig 2

Two-dimensional ordination based on the constrained analyses of principal coordinates for taxonomic (A, B) and functional (C, D) beta diversity. (A) Tβ_total measured with Jaccard index; (B) Tβ_repl, the component of Tβ_total that is due to species replacement between sites; (C) Fβ_total measured with Jaccard index incorporating branch lengths of the functional dendrogram and; (D) Fβ_repl, the component of Fβ_total that is due to trait replacement between sites.

Fig 3. Insect herbivore and functional trait/attribute indicators for the five study plants.

The order (Lep = Lepidoptera, Hem = Hemiptera and Coleo = Coleoptera) and the biogeographical origin (E = endemic, N = native non-endemic, I = introduced and U = Unknown) of each herbivore species/morphospecies is indicated. Functional traits and attributes were computed as CWM before analysis. Associations between plants and herbivore species were assessed using the IndVal index while the associations between plants and insect herbivore functional traits/attributes were evaluated with the point-biserial correlation coefficient r_bp. Significant positive associations (using coloured squares) between host plants and insect herbivore species and traits are shown jointly with the significances of the results from permutation tests (P-value).