Search

Multivariate techniques were used to compare and contrast the effects of land cover and farming practice on ground beetle and spider assemblages of Scottish farmland. For both ground beetles and spiders, the ordination and fuzzy clustering of sites were related to land cover rather than geographical location or year of sampling. The same four types of land cover group were identified: that is, heather moorland, semi-natural grassland, intensive grassland and arable land. The robustness of these land cover groups was tested using previously unsampled sites and it was found that 79 and 86% of sites, for ground beetle and spider assemblages respectively, were allocated to the land cover group predicted from their actual land cover. Furthermore, procrustes rotational analysis found a strong relationship between ground beetle and spider assemblages in intensively managed sites, suggesting that the assemblage structure of one group could be used to predict that of the other. The observed relationship between spider and ground beetle assemblages does not necessarily indicate that both groups were responding to agricultural practices in the same way. Indeed, the highest number of beetle species occurred in intensively managed grassland and arable sites while the highest number of spider species occurred in semi-natural grassland and heather sites. When conducting ecological assessments, one might wish to collect information on a wide range of ecologically different taxa; however, financial constraints make this unfeasible. From the results it could be concluded that spiders should be chosen in preference to ground beetles when seeking to make predictions on how farming practices influence invertebrates. However, such a conclusion would be premature since not only were spiders more numerous in the traps, but they were also more time consuming to process. In addition, the strong relationship found between the spider and ground beetle assemblages further justifies carabids as a target group when monitoring the influence of farming practices on biodiversity.

Detailed knowledge of the biodiversity of spider communities on agricultural land is important both in terms of enhancing pest control and understanding the driving forces influencing nature conservation value. Pitfall traps were used to assess spider species diversity at 71 Scottish agricultural sites between May and September during 1996 and 1997. Land-use varied from intensive arable fields, grasslands and extensive heather (Calluna vulgaris) moorland. Spider species richness (S) was found to decrease significantly as farm management intensity increased. Several linear regression models based on the 1996 data (50 sites) and a selection of plant, soil and landscape variables explained up to 88% of the variation in species richness. Four of these models were used to estimate 1997 species richness (36 sites: 15 repeat and 21 new) and up to 58% of sites were correctly predicted to within ± four species of the actual number caught. As only 60% of the repeat 1997 sites had values of S within four units of their 1996 score, this suggested a relatively high level of model accuracy. Model accuracy increased to 64% when all four models were used for each site, suggesting the individual models should be targeted at specific land-use types. We discuss the relevance of these models for predicting the consequences of changes in agricultural land-use for spider diversity.

Stationary water traps were used to sample the aquatic invertebrates from both shallow, drought-sensitive pools and deeper, more permanent pools on blanket bog in the Flow Country, northern Scotland during 1994–1995, an area important for nature conservation. The differences in numbers of invertebrates trapped, and the differences in species composition and abundance of adult aquatic Coleoptera and Hemiptera (the two most abundant groups caught) between these two pool types, and spatially within the deeper pools, were investigated.