The topography of diet: Orientation patch count predicts diet in turtles

Use of quantitative morphological methods in biology has increased with the availability of 3D digital data. Rotated orientation patch count (OPCr) leverages such data to quantify the complexity of an animal's feeding surface, and has previously been used to analyze how tooth complexity signals diet in squamates, crocodilians, and mammals. These studies show a strong correlation between dental complexity and diet. However, dietary prediction using this technique has not been tested on the feeding structures of edentulous (toothless) taxa. This study is the first to test the applicability of OPCr to the triturating surface morphology of a beaked clade. Fifty-five turtle specimens, 42 of which preserved both the skull and rhamphotheca, were categorized into dietary categories based on the food sources comprising 90% or 60% of their diets. Photogrammetric models of each specimen were read into molaR, producing OPCr results. Comparison of bone and rhamphotheca OPCr values shows no significant difference in complexity, implying that bone can suffice for predicting diet from morphology when keratin is absent. Carnivorous taxa have significantly lower OPCr values than herbivorous or omnivorous taxa, showing that feeding surface complexity in edentulous animals varies with diet similarly to tooth complexity in toothed taxa. Comparison of bone OPCr values by family shows that Testudinidae (tortoises) are more complex than Cheloniidae (sea turtles) and Chelydridae (snapping turtles), but that Cheloniidae and Chelydridae are not significantly different from each other. We therefore find that OPCr can be used to differentiate between carnivores and other dietary categories in edentulous taxa.