Optimizing survey methods for spiders and harvestmen assemblages in an Amazonian upland forest

Abstract Invertebrates can be sampled using any of several well-established, rapid and cost-effective methods for documenting species richness and composition. Despite their many differences, different orders of arachnids have been often sampled together in various studies. Active nocturnal search has been long considered the most efficient method for sampling spiders and harvestmen in tropical forests. We compared the number of species and composition of spiders and harvestmen simultaneously sampled using three sampling methods beating tray, active nocturnal search and Winkler traps at areas along the Urucu River, Coari, Amazonas. We found that a reasonable inventorying of harvestmen can be accomplished solely by nocturnal search, whereas the beating tray and Winkler approaches are redundant. For spiders, both the nocturnal and beating tray methods were complementary and are needed to provide a more complete picture of spider assemblages. An inventory based solely on nocturnal search saves 75% of the survey costs for harvestmen assemblages and 46% for spider assemblages. Based on our findings we propose that different taxonomic groups (e.g. harvestmen and spiders) should be sampled separately in tropical forests, especially for monitoring purposes, and different sets of methods should be combined for each according to their most efficient and best cost-benefit performance.