Seasonal data on phytoplankton composition of seston and stomach contents of the mussel (Mytilus edulis Linnaeus, 1758) from two contrasting sites, an estuarine mud flat and a rocky open shore, were compared to ascertain: (a) the extent to which differential characteristics of both sites affect this composition and (b) the degree of similarity between stomach contents and microalgal composition of seston of these sites as an index reflecting the complex processes of selection taking place within the feeding-digestive system of mussels. Individuals and water samples were collected monthly from November 2001 to December 2002, when salinity, temperature, and total and organic particulate matter concentration were also recorded in the water column. Preserved samples of seston and stomach contents were analyzed by inverted microscopy according to the Utermöhl method. Phytoplankton cells were counted and the different species grouped, taxonomically and, according to the habitat, into pelagic and tychopelagic. These data served to compute abundance (total cell count) and frequency index. Relative abundances of each group were compared for similarity between sampling sites and stomach and water samples in each site. Similarity analyses were performed using the index of Bray-Curtis, significant differences between samples being determined by the non parametric test of ANOSIM. Results of this test for the comparison between water and stomach contents resulted in significant differences: R = 0.68 in the estuary and R = 0.75 in the open shore area. Stomach contents presented a reduced average number of species (n = 6 in mussels from both sites) and a greater proportion of tychopelagic forms for comparison with the water samples (n = 20 and 24 in the estuary and open shore, respectively). Maximum phytoplankton density in water samples occurred in the May to October period, the group responsible for this increment being the diatoms. The stomach contents of marine mussels displayed two peaks of phytoplankton concentration in May (caused by the dinoflagellate Ensiculifera sp.) and in July (caused by the diatoms Pseudo-nitzschia pungens and Licmophora sp.). In the case of stomach contents of estuarine mussels, a single peak of abundance was recorded in the month of May and was mainly produced by Ensiculifera sp. To conclude, the main result coming from these comparisons is the increased abundance of dinoflagellates in the stomach contents relative to the corresponding seawater samples in the estuarine and open shore media. This result is discussed in the light of previous data concerning the differential utilization of species of phytoplankton by bivalve molluscs.