The Role of Competition in the Distribution of Andean Birds

The object of our study was to compare the distributions of bird species common to two Andean localities in order to improve our understanding of the factors involved as primary causes in imposing distributional limits. A previous paper (Terborgh 1971) had evaluated the importance of three mechanisms in the avifauna of a control locality: (1) factors that vary continuously and in parallel with the elevational gradient, (2) competitive exclusion, and (3) ecotones. As was discussed in that paper, the method, when applied to a single transect, leads to a systematic underestimation of the incidence of competition—mediated limits and to an overestimation of ecotone effects. By studying appropriately selected test localities, we could overcome many of these deficiencies and obtain results that are several steps closer to expressing reality. The control locality, the Cordillera Vilcabamba, is a sector of the Cordillera Oriental, or eastern chain of the main body of the Peruvian Andes. It carries what may be regarded as a complete avifauna, as the transect sliced through a continuous belt of forested terrain which, at any elevation, is exposed to invasion by appropriately adapted species from above and below as well as laterally. It differs in this respect from the test locality (Cerros del Sira), which is an isolated massif that rises out of the Amazonian plain some 100 km east of the main Andes. Because of this isolation, the upper portions of the Sira are exposed to invasion principally from below. An estimated 80%—82% of the species that would have occupied the summit zone of the Sira, had it been a portion of the main body of the Andes, were missing. Such a species deficit provides an ideal test of the efficacy of competition in limiting distributions, because it invites whatever species are available as colonists (in this case, those lower on the mountain) to invade until the community is effectively filled. Of the species that had the opportunity to expand their elevational ranges in the absence of high—elevation congeners that apparently excluded them on the control transect, a minimum of 71% did so, confirming the original assignments of these as competition—limited species in the control locality. In addition, a majority (58%) of the species that had no high—elevation congener on the control transect (assigned to the gradient mechanism), and that had the opportunity to expand into the species—deficit zone of the Sira, were found to do so. This reveals that diffuse competition (as distinguished from the directly observable exclusion of congeners) was really the primary limiting mechanism for most of the species that had been assigned to the gradient mechanism in the control study. Instead of accounting for approximately a third of all distributional limits, as was concluded in the earlier report, it now appears that competition, both direct and diffuse, accounts for something more than two—thirds of the distributional limits of Andean birds whose ranges end somewhere between the lowland plain and tree line.