Urban bird communities exhibit high population densities and low species diversity, yet mechanisms behind these patterns remain largely untested. We present results from experimental studies of behavioral mechanisms underlying these patterns and provide a test of foraging theory applied to urban bird communities. We measured foraging decisions at artificial food patches to assess how urban habitats differ from wildlands in predation risk, missed-opportunity cost, competition, and metabolic cost. By manipulating seed trays, we compared leftover seed (giving-up density) in urban and desert habitats in Arizona. Deserts exhibited higher predation risk than urban habitats. Only desert birds quit patches earlier when increasing the missed-opportunity cost. House finches and house sparrows coexist by trading off travel cost against foraging efficiency. In exclusion experiments, urban doves were more efficient foragers than passerines. Providing water decreased digestive costs only in the desert. At the population level, reduced predation and higher resource abundance drive the increased densities in cities. At the community level, the decline in diversity may involve exclusion of native species by highly efficient urban specialists. Competitive interactions play significant roles in structuring urban bird communities. Our results indicate the importance and potential of mechanistic approaches for future urban bird community studies.