Fruit production and patterns of seed dispersal by birds were studied at two elevations in the mediterranean scrublands of southern Spain. Fleshy—fruit—producing species represent a very prominent fraction of woody plants in terms of cover (57—76%) and species number (49—66%). Fruit production occurs year round in the lowland site but is confined to August—February upslope. Ripe fruits are most abundant (>105 ripe fruits/ha) in November—December. Fruit abundance fluctuates widely between years at the highland locality but only slightly in the lowlands. In both communities, the dominant species ripen fruits in autumn—winter, display the highest within—plant fruit densities, and tend to have the most lipid—rich fruits. Fruits differ in pulp nutritive value, seediness, and relative amount of pulp among species but are remarkably uniform in size (mostly 5—10 mm transverse diameter). Two—thirds of the passerine species at each site eat some fruit. Of these species, 69% (highland) and 26% (lowland) are resident "fruit predators," feeding on either pulp or seeds alone, and damaging the seeds when eating pulp and seeds together. The rest are overwintering or migratory seed dispersers that ingest whole fruits without damaging seeds. Seed dispersers are most common in late autumn—winter, coincident with the peak in fruit abundance and the predominance of lipid—rich fruits. A few small (12—18 g body mass) disperser species (Erithacus rubecula, Sylvia atricapilla, Sylvia melanocephala) account for most of the frugivory at each site and disperse the majority of seeds. Fruit predators either are relatively scare or eat fruits infrequently, or fruits represent a negligible fraction of their diets. Fruit removal was very high (89—100% of crops) among species with fruits smaller than the gape width of the abundant small—sized dispersers, and very low among species with fruits larger than gape width. Removal success was negatively correlated with fruit size among species having fruits smaller than dominant dispersers' gape width. No relation has been found between removal success and fruit quality, fruiting time, ripening rate, or within—plant fruit density. The principal dispersers at each site ate mainly the most nutritious fruits, although not to the exclusion of less nutritious fruits. Substantial pairwise plant—bird reciprocity is not common. (The avian species disperses a substantial fraction of a plant's seeds, which in turn provide the bulk of the bird's energy supply.) Current bird—plant seed dispersal interactions are the result of evolutionary, climatic, and geographical factors in the Mediterranean. Mutualistic congruency largely is, in these cases, an epiphenomenon of these factors, not resulting necessarily from mutual adaptations (coevolution). It is suggested that actual coevolution involving a smaller set of bird and plant species may facilitate the persistence of noncoevolving (or very slowly coevolving) plant species, thus favoring the existence of a chronic "anachronism load" (with regard to dispersal) in the plant community.