Influence of aerodynamics on the dominance hierarchy of a hummingbird assemblage in Costa Rica Gerardo Avalos 1,2, Gina Lee 3 and Alejandra Soto 1,2 1 Escuela de Biología Universidad de Costa Rica 2060 San Pedro San José Costa Rica 2 The School for Field Studies Center for Sustainable Development Studies 100 Cummings Center, Suite 534- G, Beverly MA USA 3 Boston University Department of Biology 5 Cummington Street Boston, MA USA Hummingbirds are the most specialized flying vertebrates, having the capacity to hover, which is the most energetically expensive flight type. As a group, they show great divergence in body construction, which may affect flight performance and position along the dominance hierarchy. Hummers have very little room for morphological variation given strong selective pressures on flight energetics which influence hovering flight, because without near aerodynamic optimality, individuals will be outcompeted by more powerful and morphologically efficient hummingbirds. Hypothesis: Given strong selection on hummingbird morphological variation and flight performance, we expect that variation in aerodynamic characters will predict competitive interactions among hummingbird species. Study Sites: a. Highlands: La Georgina, Cerro de La Muerte, (3,100 m, 09º34´N, 83º41´W, 25°C to 0°C, dry season Nov-April, peak of wet season Sept-Oct.) b. Mid- elevations: 4 sites in Monteverde: Hummingbird Gallery, Monteverde Cloud Forest Reserve, Selvatura Park, and San Luis Biological Station. The Hummingbird Gallery and Monteverde Cloud Forest Reserve share the same geographic location (10 ⁰ 18’N, 84 ⁰ 48’W), at an elevation of 1480 masl, whereas Selvatura Park (10 ⁰ 20’ 33.33” N, 84 ⁰ 45’ 55.48” W) is located at 1619 masl. All these sites had similar tropical cloud forest conditions, in contrast to San Luis Biological Station (10°15’47.5” N, 84°49’50.4” W, 745 masl) which is a tropical premontane forest. In addition, Hermit hummingbirds were sampled for aerodynamic parameters in Cariari de Guápiles ((10 o 30’35’’N - 83º44’39’’ W, 45 masl). Measurement of aerodynamic characters. Since April 2012 we have captured 348 individuals of 18 species, and calculated the 4 aerodynamic parameters of Stiles (2004): shape ratio (wing length divided by wing width), aspect ratio (wing length divided by wing area), wing loading (weight divided by wing area), and wing taper (shape ratio divided by aspect ratio). We also used Feinsinger and Chaplin’s (1975) allometric model for hovering power, and measured other morphological characters such as weight, bill lenght, wing lenght and width, wing area and wing span. Competitive interactions. These were quantified around feeders by recording feeder usage (dominant species monopolize the use of feeders) or by counting the number of negative interactions among species including gender within a species. Observations were done in 1min periods, separated by 2min. Videos were used to measured record using 1 min periods. METHODS RESULTS AND DISCUSSION The smallest species (Volcano and Greenthorntail) showed the highest wing loading relative to the largest species (Magnificent and Violeta Sabrewing, Fig. 1). These species are at the bottom of the dominance hierarchy at Cerro and Monteverde, respectively, and are characterized by very steady hovering flight, being highly opportunistic and able to escape in the presence of dominant species. Hermit species were intermediate in wing loading vales. The dominant species in Cerro and Monteverde had wing loading values similar to the average value of the assemblage under analysis. Wing loading did not sort species following the dominance hierarchy. A principal component analysis (PCA) explained 85% of the variation in 12 morphological and aerodynamic characters using 3 components. The first component (43.5%) was dominated by wing area, whereas wing loading (21%) and shape ratio (20%) dominated the second and third components. Principal components followed the same pattern of ordination regarding hummingbird size as in Fig. 1, with larger species (i.e., Violet Sabrewing and Magnificent) occupying the opposite end of the ordination relative to smaller species (Volcano and Green Thorntail). Species with high wing-disc loading had a high level of negative interactions (with males dominating females within a species, as expected, Figure 2). PTMG had an overly high number of interactions, representing an outlier to the overall pattern. CONCLUSIONS Our results indicate that variation in aerodynamic characters among hummingbird species is significant, but not related to competitive dominance. Aerodynamics determines access to nectar, which is spatially and temporally heterogeneous, and influences foraging for insects. The pattern of hummingbird construction responds to many physical and biological variables (not only food competition), showing that flexibility in behavior overcomes the limitations initially imposed by body size and shape. 4 Elon University Department of Environmental Science, NC, USA. FIGURE 2 PTMGM FIGURE 1 At Cerro, 50% of feeder visitations were dominated by Fiery-throated (FT). In Monteverde, the male Purple-throated Mountain Gem (PTMG) had 32% of the feeder visits whereas the females were responsible for 9% of visits (versus a 5.5% of visits by males of the alpha species, the VS). PTMG ran away in the presence of VS. The dominance hierarchy at both sites remained consistent with the pattern previously described before the permanent use of feeders.