Comparing Wing Shape of Bats: The Merits of Principal-Components Analysis and Relative-Warp Analysis

Abstract

Size and shape are fundamental features of organisms. Ideally, the methods used to describe size and shape must be sensitive enough to detect small differences, and at the same time provide interpretation that is visually satisfying. I compared shape of wings among Eptesicus fuscus, Myotis lucifugus, Pipistrellus hesperas , and Tadarida brasiliensis using three methods of morphometric analysis; principal-components analysis on traditional measures of lengths and areas, principal-components analysis of interlandmark distances, and relative-warp analysis. These species were chosen to test the efficacy of each morphometric analysis at discovering and describing differences in both size and shape. Although principal-components analyses provided similar aggregations of species based on variable loadings, only relative-warps analysis provided clear pictures as to how species differ. The molossid bat, T. brasiliensis , had narrower wings than the vespertilionid bats; this shape was not due to compression of the entire wing, but to a complicated rotation around the fifth digit. The vespertilionid bats were difficult to discern based on wing shape alone, but the warp analysis showed that P. Hesperus and M. lucifugus are more similar in shape of wing than either is to E. fuscus . Each method of analysis of shape has strengths, but relative-warp analysis provided both the power to discriminate among minor differences in shape of species within families, and the graphical capability to display those differences in a visually pleasing way.