Similarities and Contrasts in the Local Insect Faunas Associated with Ten Forest Tree Species of New Guinea

Abstract

Insect faunas associated with 10 tree species growing in a submontane area in Papua New Guinea are described and compared. In total, 75,000 insects were collected on these trees during the day and night by hand collecting, beating, branch clipping, intercept flight traps, and pyrethrum knockdown over a l-yr period. Association of chewing insects with the hosts was inferred from feeding trials. Characteristics of the fauna associated with each tree species are briefly outlined, with an emphasis on chewing insects. Four subsets of data, of decreasing affinity with the host, were analyzed by canonical correspondence and cluster analyses: (1) specialist leaf-chewers, (2) proven leafchewers, (3) all herbivores (including transient leaf-chewers and sap-suckers), and (4) all insects (including nonherbivore categories). Analyses of similarity between tree species were performed using number of either species or individuals within insect families. Analyses using number of individuals appeared more robust than those using number of species, because transient herbivore species artificially inflated the level of similarity between tree species. Thus, it is recommended that number of individuals be used in analyses of this type, particularly when the association of insects with their putative host has not been ascertained. Not unexpectedly, the faunal similarity of tree species increased along the sequence (1)-(2)-(3)-(4). Convergence or divergence in faunal similarity among tree species certainly results from many factors. Among those identified, successional status (which can be related more generally to the type of habitat in which the host grows) appeared important for specialist leafchewers; gross features of the host, such as leaf palatability and leaf weight (related to leaf toughness), were important for leaf-chewers; features presumably influencing insect flight and alighting (leaf area, probably related to foliage denseness) seemed be important for all herbivores; and features related to host architecture (tree height, type of bark) were important for all insects. Taxonomic isolation and phylogeny of trees were clearly unrelated to faunal similarity, even for specialist leaf-chewers. We discuss briefly from a conservation perspective the loss of tree species in our system and the outcome for associated insect faunas.