Estimating and Testing the Shapes of Adaptive Surfaces: The Morphology and Pollination of Dalechampia Blossoms

Using form-function relationships estimated from field data, I generated predicted adaptive surfaces indicating the expected relative performance of pollination functions for different combinations of morphological characters of Dalechampia blossoms. Natural populations plotted on the adaptive surfaces did not occupy the predicted peaks, although they were generally absent from the lowest parts of the adaptive valleys. The absence of populations from the predicted adaptive peaks may be the result of genetic correlations constraining the course of morphological evolution, or they may reflect the inaccuracy of the hypothesized adaptive topography. The true adaptive surface instead may be a central diagonal ridge rather than a single peak, with the second low spot a result of carbon limitation or reproductive interference from sympatric species.