Quantitative extraction of macro-invertebrates from temperate and tropical leaf litter and soil: efficiency and time-dependent taxonomic biases of the Winkler extraction

Winkler extractors, a simple device presumed to extract macro-invertebrates efficiently from soil and litter samples, is being used increasingly in ecological surveys and functional studies of soil macro-invertebrate communities. In this study the extraction efficiency and taxonomic bias of the Winkler extraction are evaluated for extraction periods of 3 h up to 7 weeks, calibrated by hand-sorting after 7 weeks. The method extracts most macro-invertebrates completely or to a proportion of over 90% except Isopoda, Diplopoda and Mollusca. However, for an exhaustive result, a long extraction period of several weeks is necessary. For the most speciose group (adult beetles) and for the commonly most abundant group (ants), a short extraction of 3 days was sufficient to get 70% of the individuals and nearly all species. Three days was also sufficient to recover the rank abundance order of beetle families, while for 'higher taxa' and for Chilopoda species, 4 and 3 weeks were necessary, respectively. Optimum extraction times for the abundant macro-invertebrate groups and possible adjustment factors for the soil macro-invertebrates of temperate woodlands are proposed to compensate the taxonomic bias caused by short extraction periods. However, for recording an accurate snapshot of the soil and litter fauna at a particular time, shorter extraction periods are advisable because of the short life cycle of many soil invertebrates causing emergence of later stages or a second generation during longer extraction periods. The problem of contamination of samples is also discussed.