Is pitfall trapping a valuable sampling method for grassland Orthoptera?

Abstract

Common methods to assess diversity and abundance of Orthoptera are sweep netting, transect counts and box-quadrat sampling. Pitfall trapping, by contrast, is rarely used, and the value of this method is still being questioned. In 2008, we studied Orthoptera species richness and abundance in five vegetation types along a gradient of dune succession on the Baltic Sea island of Hiddensee (NE Germany) by comparing transect-count and pitfall-trapping data. Using transect counts, 12 species were detected in the study area. With pitfall traps, three chorto- and thamnobiont Ensifera species (C. dorsalis, M. roeselii and T. viridissima) were not caught at all, and it was only in low-growing and sparsely-vegetated grey dunes that all present species were detected. With pitfall traps, the proportion of present species recorded strongly declined with increasing height and density of the vegetation type. Assuming that transect counts are a good proxy for relative Orthoptera densities, densities ascertained by pitfall traps are strongly biased by vegetation structure and locomotive behaviour of the species. More than 80% of all individuals were caught in sparsely-vegetated grey dunes. Frequency patterns of the species also differed. Using pitfall traps, especially chortobiont species were significantly underrepresented. Qualitative and quantitative sampling of Orthoptera using pitfall traps seems only reasonable in habitats with low and sparse vegetation and a high proportion of geobiont species.