A Composite Interception Trap for Sampling Arthropods in Tree Canopies

A new design of interception trap is presented based on a combination of features ofwindow and Malaise traps. The trap has been developed for study of the canopy faunas of rainforest trees, specifically Argyrodendron actinophyllum. Five such traps have been operated in SE Queensland for a year with catches being removed weekly. Coleoptera, Nematocera, Homoptera and Hymenoptera were the most abundant groups in these samples with the families Sciaridae, Cicadellidae, Scolytidae and Chrysomelidae dominating the collections. The trap appears to sample the arboreal fauna selectively and was used to identify the peak of arthropod abundance during the year. Introduction Arboreal arthropods are usually sampled in rainforests using light traps (Ricklefs 1975; Sutton and Hudson 1980) or canopy fogging (Erwin 1983; Stork 1987). However the limitations of light traps are well known (Bowden 1982): information is gained on night-flyers only; the attractiveness of the trap varies from one taxon to another; correction factors must be computed for the effect of temperature, moonlight and cloud cover; the catches are directly related to the trap illumination and to the background illumination. This last point is particularly important when the sampling program is concerned with the host specificity of the specimens collected. Light traps operated in the rainforest canopy under different background illuminations are indeed likely to attract insects from a wide range of host trees and vines. The comparison and the interpretation of the qualitative and quantitative data would be difficult in such conditions. This paper describes a simple and inexpensive interception trap which overcomes most of the problems cited above. The temporal distribution of canopy arthropods associated with the rainforest tree, Argyrodendron actinophyllum, was appraised during a first year of continuous trapping with 5 such traps. Materials and methods The study site was situated in a region of subtropical rainforest (complex notophyll vine forest) near Mt Glorious (27"19'S, 152'45'E, 700 m), some 30 kms NW of Brisbane, Qld. Argyrodendron actinophyllum Edlin (Sterculiaceae) is a canopy tree which is common on the site and often reaches 40 m in height in subtropical rainforests. Young ( I 985: group 9) provides a detailed description ofthe floristic data in the study area. Five traps were placed in 5 different A. actinophyllum trees free of vines and several hundreds of metres distant from each other. The traps were positioned at mid-level in the heart of the crowns, at about 25 m above the forest floor, in February 1986. Each trap was surveyed weekly, on the same day, using the single rope technique described by Perry (1978). The fixed position of the trap minimised problems related to changes in trap orientation (and the consequent differences in catches) during this long-term study. The interception trap developed in this project is a composite trap which consists of 2 sub-units (Fig. 1). The top sub-unit is a small Malaise trap similar to the model of Townes (1 972), previously tested in a survey of the arboreal fauna associated with Pinus mugo in Europe (Basset 1985 a and b). The profile area is 0.35 m2. A recent review of the use of Malaise traps in entomology is given by Haenni and Matthey (1 984). The bottom sub-unit is a window trap derived from the design of Wilkening et al. (1 98 I), with a collecting surface of 0.14 m?. Arthropods either fly or crawl up into the collector of the Malaise trap or impact against the plexiglass panels, falling into the collector of the window trap. Both collectors were filled with a collecting fluid (ethylene glycol 20%), which is neither an attractant nor a repellant and has a low evaporation coefficient. A commercial killing agent ("Shelltox Ministrip") in the collectors was replaced at 3-week intervals. The collected material was stored in 70% alcohol and sorted into families or Recognisable Taxonomic Units (RTU's). Sorting is still in progress.