We conducted reciprocal transplant experiments with four dominant species from salt and tidal freshwater marshes at two tidal elevations. Ten individuals of each species were transplanted as sods to the opposing marsh type and within their native marsh (two sites each). Transplants were placed at the centre of 9-m2 plots along a line parallel to the river bank. In order to disentangle abiotic and biotic influences, we set up plots with and without neighbouring vegetation, resulting in five replicates per site. Freshwater species (Bolboschoenus maritimus and Phragmites australis) transplanted to salt marshes performed poorly regardless of whether neighbouring vegetation was present or not, although 50-70\% of the transplants did survive. Growth of Phragmites transplants was impaired also by competition in freshwater marshes. Salt marsh species (Spartina anglica and Puccinellia maritima) had extremely low biomass when transplanted to freshwater marshes and 80-100\% died in the presence of neighbours. Without neighbours, biomass of salt marsh species in freshwater marshes was similar to or higher than that in salt marshes. Our results indicate that salt marsh species are precluded from freshwater marshes by competition, whereas freshwater species are excluded from salt marshes by physical stress. Thus, our study provides the first experimental evidence from a European estuary for the general theory that species boundaries along environmental gradients are determined by physical factors towards the harsh end and by competitive ability towards the benign end of the gradient. We generally found no significant impact of competition in salt marshes, indicating a shift in the importance of competition along the estuarine gradient.