-Through natural population changes and experimental field removals, we tested the hypothesis that Least Flycatchers (Empidonax minimus) restrict habitat use by socially subordinate American Redstarts (Setophaga ruticilla). On a 10-ha site 2-yr-and-older (ASY), but not yearling (SY), male redstarts avoided the sector occupied by flycatchers from 1975 to 1980, but preferred this sector from 1981 to 1985 when flycatchers were absent. Vegetation changed subtly on the site but could not account for the sudden shift in redstart settlement pattern. On 6 4-ha sites ASY male redstarts were most abundant in years of Least Flycatcher absence. On the 5 4-ha sites from which Least Flycatchers either disappeared independently or were removed experimentally between 1981 and 1984, redstart abundance increased on four and remained constant on the fifth; on three control areas redstart numbers declined during the same period. Least Flycatchers recolonized one removal site, and ASY redstart abundance subsequently declined. SY male redstart abundance varied inversely with that of ASY male redstarts. We conclude that flycatchers influenced the distribution of ASY male redstarts directly, and that of SY males indirectly, more than either vegetation structure or other habitat characteristics. At no spatial scale examined, however, did total redstart abundance (ASY + SY) vary inversely with that of Least Flycatchers; in fact, their total abundances correlated positively at a regional scale. These findings, combined with a model for asymmetric competition for mutually preferred habitat (Pimm et al. 1985, Rosenzweig 1985), illustrate how a socially dominant competitor could lead to a broadening rather than a narrowing of the habitat breadth of a subordinate species. We show that competitor species abundances need not vary inversely and that age classes may be affected differentially. This species interaction illustrates subtleties and complexities of how competition can modify avian habitat selection. Received 2 September 1987, accepted 22 January 1988. INTERSPECIFIC competition is widely believed to restrict the range of habitats or resources exploited by many species (e.g. Svardson 1949, Connell 1983). If each species reduces the other's abundance under different circumstances, two species should vary inversely in abundance in one habitat over time, or across an array of habitats at a particular time. Avian biologists in particular have frequently interpreted negative correlations in abundance and replacements along habitat gradients as evidence for interspecific competition (Svardson 1949, MacArthur 1972, Cody 1974, Terborgh and Weske 1975, Diamond 1978, Noon 1981, Mountainspring and Scott 1985, Grant 1986). Recent experimental studies, especially those involving the removal of individuals of one or more putatively competing species, have confirmed that interspecific competition for habitat occurs regularly in a variety of bird species (Mewaldt 1964, Davis 1973, Williams and Batzli 1979, Dhondt and Eyckerman 1980, Hogstedt 1980, Reed 1982, Garcia 1983, Alatalo et al. 1985). Other investigators recorded no negative correlations between densities, and concluded that species often respond independently to structural and floristic features of their environment (Wiens 1977, Rotenberry and Wiens 1980, Wiens and Rotenberry 1981, Collins et al. 1982, James and Boecklen 1984, James et al. 1984). These findings have been interpreted as evidence that species are distributed independently and are not strongly influenced by interspecific competition. The contrasting results between these two types of studies have contributed to the recent controversy concerning the relative importance of competition in structuring communities (Strong et al. 1984, Connor and Sim-
Read full abstract