Aims Factors limiting distributions of species are fundamental to ecology and evolution but have rarely been addressed experimentally for multiple species. The conspicuous linear distribution patterns of plant species confined to river corridors in the Central European lowlands constitute an especially long-standing distribution puzzle. We experimentally tested our novel hypothesis that the tolerance of species to river corridor conditions is independent of the degree of confinement to river corridor habitats, but that species not confined to river corridors are better able to take advantage of the more benign non-river corridor conditions. Methods We grew 42 herbaceous species differing in their confinement to river corridors in a common garden experiment on loamy soil typical for river corridor areas and sandy soil typical for non-river corridor areas, andwith andwithout a flooding period. For a subset of species, we grew plants of both river corridor and non-river corridor origin to test for adaptation to river corridor conditions. Important findings Species more confined to river corridor areas benefited less from the more benign non-flooded and non-river corridor soil conditions than species of wider distributional range did. For subsets of 7 and 12 widespread species, the response to flooding and soil origin, respectively, did not differ between plants from river corridor sites and plants from other sites, suggesting that the habitat tolerance of widespread species is due to phenotypic plasticity rather than to local adaptation. Overall, we found clear support for our novel hypothesis that species not confined to river corridors are more able to take advantage of the more benign non-river corridor conditions. Our study provides a general hypothesis on differences between species confined to stressful habitats and widespread species out for test in further multispecies comparative experiments.