The spider (ARANEAE) guilds of four different vegetation types with particular reference to plant structure

Abstract

Four subsites were selected at the eastern end of Hedleyhope Fell, Tow Law, County Durham. These areas were chosen on the basis of differing vegetation types. The study of spiders was conducted using pitfall trapping and sweep-netting with each sample collected five times, at fortnightly intervals, between May and July. The value of these sampling methods is considered, with the conclusion reached that they are the best techniques available, taking into account the restrictions imposed, for this type of investigation. A total of 2795 spiders were caught, comprising 10 femilies and 64 species. Chi-square tests, correlation and regression, percentages, Detrended Correspondence Analysis (DCA) and Simpson's diversity index were used when appropriate on the pitfall data. The totals caught via sweep-netting were too small to analyse statistically. Species diversity and richness for both families and species showed considerable differences when the subsites were compared. The distribution between subsites was analysed by DCA (DECORANA) and subsequent Chi-square tests. These critiques suggested that individual species showed preferences for certain subsites. This is demonstrated particularly well by the predominant Lycosid species whose subsite dispersal was markedly dissimilar. The total numbers of spiders caught varied over the study period, rising to a peak in late June after which the numbers dropped again. The ratio of males : females was high at the beginning of the field work and remained so until it declined towards the end of the investigation. The overall male : female ratio was 2.1 :1.The sweep-net analysis illustrated that a different section of the fauna was being sampled by this technique, although catches were poor, and there was some bias towards certain species in some areas. The vegetation was classified by species composition and architecture. Plant cover at each subsite was shown to be disparate according to the National Vegetation Classification (NVC) system. The architecture was analysed between subsites by DECORANA. This, followed by subsequent statistical tests, indicated that the plots were divergent in their physical arrangement in addition to their plant species composition. The suggestion that significant differences between vegetation structure are a determinant of spider guilds is discussed. The reasons behind this community isolation are also considered, with particular reference to interspecific competition, prey availability and microclimatic factors. A longer study would be needed to reach any firm conclusion as it is difficult to exclude any causal explanation of spider community relations with the results obtained over two and a half months.