Comparison of methods for sampling Thysanoptera on basswood ( Tilia americana L.) trees in mixed northern hardwood deciduous forests

Abstract

Canopy arthropods play integral roles in the functioning, biodiversity, and productivity of forest ecosystems. Yet quantitative sampling of arboreal arthropods poses formidable challenges. We evaluated three methods of sampling the introduced basswood thrips, Thrips calcaratus Uzel (Thysanoptera: Thripidae), from the foliage of basswood canopies with respect to statistical variability and practical considerations (legal, economic and logistical accessibility). All three methods involved removal of foliage, which was performed using a pole-pruner, shotgun, and certified tree-climber. We also tested a fourth method, in which the tree-climber enclosed samples in a plastic bag to estimate losses that occur when branches fall to the ground, even though this is often not practical. The climber plus bag and pole-pruning methods obtained the highest numbers of thrips. Mean number of larval thrips did not vary significantly among the three main sampling methods. Site had a stronger effect on the number of larval thrips obtained than on the number of adults. A significant method by site interaction was observed with adults but not larvae. Significant collection date (which corresponds to thrips life stage) by site interaction was also observed. We regressed sampling methods to determine if the number of thrips obtained using one method can be used to predict the number obtained with another. Tree-climber and pole-pruner data were highly predictive of each other, but shotgun data cannot be used to estimate other methods. Pole-pruning is the most cost-effective and legally permissible technique, but is limited to trees with accessible lower branches. The shotgun method is cost-effective and useful in sampling trees at least up to 27 m, but is prohibited close to human activity. The tree-climber is effective and broadly applicable, but incurs the highest costs. This study shows the need to evaluate a variety of techniques when sampling arboreal insects with respect to predictability, pragmatics and life stages.