A spatial model to estimate arthropod diversity in agricultural landscapes. Species numbers of epigeal arthropods from a transect with 19 sampling stations in agricultural areas were correlated with landscape parameters, derived from aerial photographs and landscape manging plans, and fed into a geographical information system (GIS). Stepwise multiple regression highlighted a combination of several relevant factors, whihc in turn allowed a reasonably accurate prediction of species numbers for all 19 habitats tested. The model is a promissing approach for mapping the regional distribution of biodiversity hotspots or sinks in agricultural areas - without collecting a single insect.