1. Modern literature on plant dispersal by birds focuses mainly on the importance of frugivory and scatter-hoarding, yet recent studies show that endozoochory by migratory waterbirds is an important mechanism of long-distance dispersal for a broad range of plants. Nevertheless, there is a lack of empirical field studies that identify the plants dispersed by waterbirds, and relate them to expectations based on dispersal syndromes. To date, there are no detailed studies of the level of spatial variation in the plant taxa dispersed by a waterbird vector. 2. Five sets of faecal samples (total n = 215) were collected from mallards Anas platyrhynchos on autumn migration in the Tisza and Balaton regions in Hungary, central Europe. Intact diaspores were extracted, identified and their germinability assessed under standard conditions. The plant communities recorded at different sites were compared with PERMANOVA and other multivariate methods. 3. Macrospores of the floating watermoss Salvinia natans were recorded in 32 samples, and a total of 16 macrospores germinated, providing the first field demonstration of endozoochory of ferns by birds. Of 21 angiosperm taxa recorded (of which 8 germinated), 13 were terrestrial species, although the most abundant taxa were aquatic species such as the alkali bulrush Bolboschoenus maritimus and the sago pondweed Potamogeton pectinatus. Two naturalized alien species, the common fig (Ficus carica) and the hackberry (Celtis occidentalis) were also recorded. The plant taxa dispersed varied at two different spatial scales, with minor but significant differences between samples from sites separated by less than 1 km, and major differences between the two regions separated by approximately 220 km. 4. Synthesis. This is a unique study of the spatial variation in plants dispersed by endozoochory by a migratory waterfowl species, with the first demonstration of avian endozoochory of ferns. Most taxa dispersed are generally assumed to disperse by water, wind or self-dispersal, and waterfowl will provide much greater maximum dispersal distances. More such studies are essential before we can understand which plants are dispersed by migratory birds, because morphological dispersal syndromes do not allow us to make adequate a priori predictions.