Spider communities in agricultural habitats: the effect of habitat interfaces on the distribution of species.

Abstract

Two Sites representing a change in habitat type at a community interface in an agricultural situation between managed and unmanaged areas were sampled using transects of pitfall traps crossing the interface at right angles from one habitat type to the next, during May/June, 1992.Site 1 consisted of a lightly grazed pasture bordered on one side by a mainly coniferous plantation, with an abrupt interface between the two contrasting community types. Site 2 consisted of a field of spring barley bordered on one side by a semi natural area of mixed deciduous woodland, with the interface consisting of a boundary approximately 2m wide between the two habitat types. Community analysis revealed patterns of total spider abundance, species richness, and diversity across the interlace. The Site 1 interface had a negative effect on the numbers present in the pasture, but no significant effect on the numbers in the plantation. Numbers in the pasture were significantly higher than that at the interface and in the plantation. Species richness decreased progressively from the pasture to the plantation, and species diversity increased, with the interface showing intermediate values for both these factors. The transect regions displayed differences in the dominance of groups, with Telragnathidae and Lycosidae in the pasture replaced by Linyphiidae on approaching the interface and throughout the plantation. Species were influenced in different ways by the habitat change, but most preferred one side of the interface where they occurred in highest abundance. The Site 2 interface had a positive effect on the numbers of spiders present in the crop and in the woodland. Numbers in the interface samples were significantly higher than in the crop samples. Species richness was higher in the woodland than in the crop, and highest in the boundary samples Species diversity was higher in the woodland than in the crop traps. The transect regions showed differences in the dominance of groups, with Linyphiidae dominant in both the crop and woodland, and Lycosidae dominant at the interface. Species showed individualistic responses to the change in habitat, with most preferring one side to the other, some were ubiquitous, and a few increased in abundance at the interface. Similarity indices (Sqirensen and Chi-squared) were calculated to visualise the effects of local fauna 1 exchange between habitat types at each Site, by providing a measure of similarity according to species composition. Both revealed greatest similarity between within-habitat traps than between traps from different sides of the interface. The pasture at Site 1 showed greater similarity to the interface than did the plantation. At Site 2, the interface showed a fairly high degree of similarity to both the crop and woodland. A classification, according to spider species, was performed in order to determine whether the transect regions on either side of the interface at each Site could be characterised by the species present and their abundances. Both Sites revealed a division between the two habitat types on either side of the interface according to the community composition. This indicated that vegetation structure played an important role in determining the species composition, mediated through the effects of disturbance, as this was the most dramatic change in environmental features between the two regions either side of the interface. In order to identify further the factors which determined the spider distribution patterns, ordination of the sites and species were performed. Canonical correspondence analysis revealed that vegetation structure was the most important factor influencing spider distribution across the interface at Site 1, and vegetation structure and % soil moisture content were important in determining the spider species distributions at Site 2. This was considered to be related to the degree of management and disturbance of the habitats.