The aim of this study was to examine top-down effects of cursorial spiders in subsidized coastal food webs. Top-down effects were examined by selectively removing cursorial spiders, mainly wolf spiders, from small islands (26-1834m2) during 2004-2007. The removal success varied among islands and years, and spider densities were reduced by 30-65%. To examine treatment effects, arthropods were sampled using a vacuum sampling device at three occasions each summer. The densities of other arthropod predators, especially web spiders and carabids, were higher on islands where cursorial spiders had been removed compared to control islands. This treatment effect probably occurred through a combination of competitive release and reduced intraguild predation from cursorial spiders. No treatment effects were found on herbivore or detritivore densities and plant biomass. This lack of effect may either be because spiders indeed have fairly weak effects on herbivore and detritivore densities on Baltic shorelines or that the removal success of spiders was insufficient for observing such effects. Treatment effects may also be weak because negative effects exerted by spiders on herbivore and detritivore populations were balanced by increased predation by insect predators.