As most stingless bee species depend on preexisting cavities, principally tree hollows, nesting site availability may represent an important restriction in the structuring of their forest communities. The present study examined the spatial dynamics of stingless bee communities in an area of Atlantic Forest by evaluating their swarming to trap-nests. The field work was performed in the Michelin Ecological Reserve (MER) on the southeastern coast of the state of Bahia, Brazil. Seven hundred and twenty trap-nests were distributed within two forest habitats in advanced and initial stages of regeneration. The trap-nests were monitored between September 2009 and March 2011. Twenty-five trap-nests were occupied by five bee species, resulting in a capture ratio of 0.035 swarms/trap (approximately 0.14 swarms/ha), corresponding to 10 swarms/year (0.056 swarms/ha/year). According to previous study at MER, the most abundant species in natural nests were also the most common in trap-nests in the two forest habitats examined, with the exception of Melipona scutellaris Latreille. Swarms of higher numbers of species were captured in initial regeneration stage forests than in advanced regeneration stage areas, and differences in species compositions were significant between both habitats (p = 0.03); these apparent differences were not consistent, however, when considering richness (p = 0.14) and total abundance (p = 0.08). The present study suggests the existence of a minimum cavity size threshold of approximately 1L for most local species of stingless bees and sustains the hypothesis of a mass effect of Tetragonisca angustula Latreille populations from surrounding disturbed habitats on the MER forest community in terms of propagule (swarm) pressure. Examining swarm densities with trap-nests can be a promising technique for comparative analyses of the carrying capacities of forest habitats for stingless bee colonies, as long as size thresholds of cavities for nesting are taken into consideration.