Modelling populations of Erigone atra and E. dentipalpis (Araneae: Linyphiidae) across an agricultural gradient in Scotland

Abstract

Linyphiid spiders are important generalist predators of insects in agricultural habitats. Their ability to rapidly disperse and colonise disturbed habitats makes them important as a natural form of pest control, often present before other predatory groups arrive. This paper examines the distribution of Erigone atra and E. dentipalpis, which are amongst the commonest linyphiid species to be found on agricultural land in Scotland. The habitat preferences of these species were assessed using abundance and proportion data from 71 independent sites sampled using pitfall traps over 2 years and a selection of repeat first-year sites sampled during the second year, incorporating a range of land-uses from extensive moorland, through grasslands to intensive arable fields. E. atra dominated in autumn sown crops, and E. dentipalpis dominated at the other end of the agricultural management spectrum in low-intensity grasslands. Both species were considered absent from upland and moorland habitats. Linear regression modelling was used to create a model which best estimated the variance in proportions of E. atra in the catch across the sites using a selection of vegetation, soil, management and landscape variables. This model was based on data from 58 sites where Erigone species were abundant and explained 66% of the variation in E. atra. The variables included were increasing vegetation biomass above 50 mm and increasing levels of mixed grazing (both positive), and a combination of increasing plant species richness and vegetation stem density (negative). The efficiency of the model was examined using data from 13 sites which were sampled in the subsequent year to assess the effects of no changes and changes in land-use on the proportions of the two species. Only two sites were considered inaccurate (more than 20% different from the model), suggesting a relatively high level of accuracy. Interaction between the two species was tested for and not found to explain the differences in proportion observed. Comparisons with other studies were carried out, showing similar trends in proportions and in some of the variables identified as important. The influence of specific variables on each species is described in light of possible allometric-type effects, and the importance of these species for conservation and pest control in agriculture is discussed.