Diet niche relationships among North American grassland and shrubsteppe birds.

We consider the dietary relationships of the numerically dominant breeding bird species in four North American grassland/shrubsteppe habitats, sampled over 2-3 consecutive years. Overall, the diets of these species contained primarily insects: orthopterans comprised 29% of the diet biomass, coleopterans 24%, and lepidopteran larvae 23%, while seeds contributed 15% of the average diet. These diets varied substantially, however, and we evaluated several aspects of this variation. Intersexual differences in diets within a species were few, despite the occurrence of significant sexual size dimorphism in several species. For many species, however, there were substantial shifts in dietary composition between years at a given location; overall, the average between-year similarity of species' dietary composition was 70%. Different species exhibited rather different diet patterns. Horned Larks were relatively omnivorous, had broad diet composition niches, and varied considerably in diets between different locations. Meadowlarks were also broad-niched and geographically variable in their diets, but were the most highly carnivorous of the species we considered. Dietary niche breadths of Grasshopper Sparrows were intermediate, but diet composition was rather stable, both between years and between locations. Chestnut-collared Longspurs exhibited narrow diet niches, but substantial annual variation: each year this species apparently exploited a different but limited set of prey types rather heavily. Larger avian predators generally consumed a broader array of functional groups of prey, but did not differ in the taxonomic variety of their diets from small birds. Variation in diet composition between individuals within local populations was considerable; in most species, an individual contained on the average 30-40% of the prey taxa represented in entire population smaples.Patterns of dietary overlap among species were quite inconsistent from year to year at most locations, although at the shrubsteppe site overlap among all species present was consistently quite high. Relatively few cooccurring species pairs exhibited low diet overlap. The degree of diet niche overlap was unrelated to body size differences of the birds, despite as much as six-fold differences in weight among some coexisting species. Relationships of the bird species on another dimension of the trophic niche, prey size, also differed substantially between sites and years. The ranking of co-occurring species by the mean sizes of the prey they consumed generally did not parallel their rankings by body sizes, and in some cases the smallest and the largest species present ate prey of similar sizes. At the shrubsteppe site, all the breeding species exhibited quite similar frequency distributions of prey sizes in their diets.As species number and diversity increased in the breeding avifaunas, diet niche breadths generally decreased, species packing by prey size decreased, and diet composition niche overlap remained relatively unchanged. These trends are in at least partial agreement with predictions of diffuse competition theory, but the patterns were derived from broad inter-site comparisons of overall site averages, and the relationships generally did not hold within local assemblages of species. In general, our attempts to match values of dietary niche features with site characteristics failed to demonstrate close agreement with the predictions of prevailing ecological theory based upon assumptions of resource limitation and competition. Instead, our findings seem generally most consistent with the suggestion that food is not normally limiting to bird populations in these systems, and individuals and populations are exploiting the food resources in an opportunistic fashion, which leads to considerable individual, between-year, and between-location variation in diet compositions and interspecific overlaps.Our attempts to discern clear relationships that accord with theoretical expectations in these avian assemblages are thwarted by our lack of detailed information on the resource base and by the lack of clear tests that will separate alternative hypotheses of community organization and structuring. We suggest that these complications may compromise the findings of many community studies.