Abstract. Environmental surrogates (land classes) for the distribution of biodiversity are increasingly being used for conservation planning. However, data that demonstrate coincident patterns in land classes and biodiversity are limited. We ask the overall question, ''Are land systems effective surrogates for the spatial configuration of biodiversity for conservation planning?'' and we address three specific questions: (1) Do different land systems represent different biological assemblages? (2) Do biological assemblages on the same land system remain similar with increasing geographic separation? and (3) Do biological assemblages on the same land system remain similar with increasing land system isolation? Vascular plants, invertebrates, and microbiota were surveyed from 24 sites in four land systems in arid northwest New South Wales, Australia. Within each land system, sites were located to give a hierarchy of inter-site distances, and land systems were classified as either ''low isolation'' (large and continuous) or ''high isolation'' (small patches interspersed among other land systems).Each type of land system supported components of biodiversity either not found, or found infrequently, on other land systems, suggesting that land systems function as surrogates for biodiversity, and that conservation-area networks representing land-system diversity will also represent biological diversity. However, the majority of taxa were found on more than one land-system type, suggesting that a large proportion of the plant, arthropod, and microbial biodiversity may be characterized by widespread species with low fidelity to particular land systems. Significant relationships between geographic distance among sites and differences among assemblages were revealed for all taxa except the microbiota. Therefore, as sites on the same land system were located farther apart, the assemblages at those sites became more different. This finding strongly suggests that conservation planning based on land-system diversity should also sample the geographic range occupied by each land system. Land-system isolation was not revealed to be a significant source of variation in assemblage composition. Our research finds support for environmental surrogates for biodiversity in conservation planning, specifically the use of land systems and similarly derived land classifications. However, the need for explicit modeling of geographic distance in conservation planning is clearly indicated.