Grassland Spider Habitats in North-East England

The spider communities of sixty-nine sites covering a range of grassland types in north-east England were sampled with pitfall traps. Species data were ordi- nated using detrended correspondence analysis (DECORANA) and classified using two-way indicator species analysis (TWINSPAN). 2. Seven habitats were identified from the classification. Discriminant analysis indicated that the management regime was the most important distinguishing variable between habitat types. Vegetation height, soil moisture and site altitude were also significant discriminators. 3. Leaving-one-out error estimates indicated that sites could be assigned to correct habitats with a 67% success rate using these four variables. Classification of ten repeat sampled sites indicated that the original classification was robust. 4. Success rate for the classification of ten new sites on the basis of either their spider species lists and their envi- ronmental variables was similar, suggesting that spider communities present in grassland could be predicted rea- sonably well using these four variables alone. 5. The potential for using spider community data in the monitoring of land use changes and in environmental assessments is discussed.