Inventorying and estimating subcanopy spider diversity using semiquantitative sampling methods in an Afromontane forest

We investigated the effect of plot-based and unrestricted (plot-less) sampling on an inventory of a megadiverse taxon, spiders, in an Afrotropical forest for the purpose of species richness estimates. We also investigated the efficiency of human-based sampling methods and the effect of allocation of sampling effort to different sampling methods to cover as many microhabitats as possible. In the 10-d sampling period in the montane forest of the Uzungwa Scarp Forest Reserve in Tanzania, eight collectors sampled spiders for 350 h and 800 pitfall trap-days. Two hundred hours of sampling were restricted to a 1-ha plot and 150 It of sampling took place outside the plot. The sampling team included both experienced and inexperienced collectors using five different hand collecting methods during day and night sampling periods. Sampling yielded 9,096 adult spiders representing 170 species in total. Number of species and adult spiders per sample and overall species composition depended mainly on the sampling methods used and time of day. Whether the sampling took place within or at random outside the plot did not affect species composition or number of species per sample. Collector experience did affect the number of species collected per hour and thereby overall species composition of the sample but was less important than sampling methods used and time of day.