Meeting Abstract

70-6  Wednesday, Jan. 6 09:15  Opportunity for selection: the shape of the fitness landscape in an amphidromous waterfall-climbing Hawaiian goby Sicyopterus stimpsoni MOODY, KN*; KAWANO, SM; BRIDGES, WC; BLOB, RW; SCHOENFUSS, HL; PTACEK, MP; Tulane Univ.; NIMBioS; Clemson Univ.; Clemson Univ.; St. Cloud State Univ.; Clemson Univ. kmoody3@tulane.edu

Habitat heterogeneity can drive the evolution of locally specialized phenotypes and lead to adaptive diversification across an environmental gradient. Natural selection is a primary driver of adaptive evolution, and can be weakened by trade-offs between phenotypic traits that confer different performances. These trade-offs have the potential to generate complex fitness surfaces with multiple local optima in multivariate morphospace amongst different selection regimes. We tested for the presence of such fitness surfaces and their correlation with observed morphological differences among juveniles from subpopulations of the amphidromous Hawaiian waterfall-climbing goby, Sicyopterus stimpsoni. To examine the role of natural selection in promoting local adaptation of body shape in this species, we conducted laboratory experiments to compare linear and nonlinear selection and the opportunity for selection in fish from Kaua’i and Hawai’i due to the opposing pressures of predator evasion and waterfall climbing which vary in strength between the islands. We found that directional selection was strong in traits that enhance climbing or predator evasion, but the opportunity for directional selection was greatest for climbing in both subpopulations, but acted on different traits. Furthermore, the strength of directional selection was constrained by nonlinear stabilizing selection and potential trade-offs in functional capacities. These results demonstrate that natural selection can lead to locally adapted phenotypes because of differences in selective pressures. However, similar selection pressures can potentially produce comparable strength, modes, and opportunity for selection through many-to-one-mapping.