Thomas (1982) compared the distance between maxillar lamellae and the frequencies of occurrence of certain foods of critical sizes (six seed species ranging from 0.5 to 5.5 mm and four animal taxa ranging from 0.3 to 1 mm) in the gut contents of four dabbling ducks (mallard, pintail, teal and shoveler A. clypeata), and did not observe food partitioning by size, although small sample size and the limited size range considered may explain this. Nummi & Väänänen (2001) studied diet overlap among six sympatric dabbling ducks (mallard, pintail, teal, shoveler, wigeon A. penelope and

garganey A. querquedula) and failed to demonstrate any difference in diet size, proposing that the high level selleck inhibitor of diet overlap was promoted by abundant food resources in their study area (hence no competition). The latter studies are, however, typical snapshot studies. For the present meta-analysis, we used a very large compilation of data, from all over the flyway, and we were able to show that there are consistent differences JAK inhibitor in mean size of ingested seeds between species over large geographic areas and over seasons. The differences in seed diet therefore appear to have an important role in community structure, as lamellar density largely dictates which particle sizes are going to dominate the diet of individual ducks of a given

species (see Gurd 2006 for details about the complexity of food filtering in dabbling ducks). Moreover, the ANOSIM analyses revealed that the seeds consumed by mallard and teal differ by family (and a fortiori species). The size segregation hence also reflects differences in seed species composition

in the diet, which may also partly explain the coexistence of these two species under a paradigm of resource-limited competition-structured communities. Pintail, however presented similarities with mallard and teal diet. As stated earlier, the analyses are based on seed families. Segregation see more might also occur in a more subtle manner at the seed species level. Specializing in different food sizes (and species) may be an adaptation reducing niche overlap in times of high interspecific competition. Apart from lamellar density, there are other physiological and ecological differences between species that may influence diet. Species with fewer lamellae (but larger, longer bodies) indeed tend to feed in deeper, open microhabitats, while species with denser lamellae (but smaller, shorter bodies) tend to feed in shallower and more vegetated microhabitats (Pöysä et al., 1996; Pöysä & Sorjonen, 2000), which could also have an effect on food particle size through different plant composition. A combination of differences in bill lamellar density, body length and feeding habits may therefore be required for genuine food resource partitioning among dabbling ducks (cf. Nudds et al., 2000; Guillemain et al., 2002).