A B S T R A C TPrevious studies reported that landscape-level factors are vital to support diversity of spiders in strongly modified arable lands and disturbed habitats such as managed semi-natural grasslands. Cropland management (ploughing, fertilization, and pest management) Q3 and agricultural practices (mowing and grazing) destroy and/or modify regularly the spider assemblages; thus, continuous recolonization from the surrounding landscape is vital to sustain the species pool. On the contrary, we hypothesized that in unmanaged grasslands, the spider assemblages are stable and the importance of recolonization is limited, the local factors become much more important drivers in shaping spider assemblages than landscape-level factors. We tested the importance of local and landscape-level factors on the abundance and species richness of spiders in unmanaged grasslands. At the landscape-level, we found that only the isolation had significant effect on the total abundance, on the abundance of hunting and habitat specialist species, and on the abundance of a frequent species (Gnaphosa mongolica). At the local scale, however, four out of five studied factors influenced significantly the species richness and abundance of spider assemblages and the abundance of two frequent species (Alopecosa psammophila, Berlandia cinerea). Species richness and abundance increased by plant cover, litter cover, and patch size, while decreased by bare ground cover. We found that in unmanaged grasslands, the local factors had vital role in maintaining the spider species richness; this is just the opposite conclusion that was earlier reported for agricultural ecosystems, where landscape-level effects had crucial role providing the species for continuous recolonization.2015 Published by Elsevier B.V. the local scale and the landscape-level scale) in these habitats 33 (Tscharntke et al., , 2012Batáry et al., 2008