Abstract
An attempt to predict spider species occurrence is described, using theridiid spiders as an example. Data from the database of Middle European spiders was evaluated to describe the relationships among the occurrence of spider species using the Jaccard coefficient and cluster analysis.
Common "non-specific neighbour" species are eliminated in this way - they are often found with a spider in question but have a minimum indication value. The result of such an analysis is a list of the species occurring regularly with the given spider without any other relation to the features of the habitat.
A possible use of such knowledge could be, for instance, the search for a certain species in the given place on the basis of the neighbour's occurrence discovered by previous collection.