Dental adaptive traits to herbivory in Plio-Pleistocene African wild pigs: new analyses of enamel complexity and thickness

Abstract

Wild pigs (Suidae) have a complex evolutionary history that is well documented in the rich fossil record from African Plio-Pleistocene sites. These ungulates have adapted to major environmental changes (expansion of arid and open grasslands) during the last million years in Africa, with the iterative convergent evolution of specialized dental morphologies (third molars with long and high crowns). Our work aims at better characterizing the dental adaptive traits to herbivory in African suids. For this purpose, we built a reference data set of 17 third molars of extant Eurasian and African suids (Babyrousa, Sus, Porcula, Potamochoerus, Phacochoerus, Hylochoerus), spanning the full spectrum from omnivory to specialized herbivory, and we used that data set as an interpretative framework for the paleoecological inferences of extinct African suids (Potamochoerus, Kolpochoerus, Metridiochoerus, Notochoerus; 16 fossil teeth). Using X-ray microtomography, we characterized the inner structures of the lower and upper third molars of suids. Using the Avizo® software, we segmented the dental tissues and generated slices perpendicular to the crown height to simulate flat occlusal morphologies at various wear stages. On each slice, we then automatically compiled several dental parameters using the Fiji® software: 1) the enamel complexity (EDBI = the ratio of the length of enamel blades to the square root of the occlusal surface; 2) the average enamel thickness (ET); and 3) the dentin to enamel ratio (DER = the ratio of dentin area to occlusal area) as a proxy for the degree of wear. For each tooth, bivariate plots of the enamel complexity and thickness versus the degree of wear (= “wear spaces”) allow to quantify the changes that occur during the life of the individual, and allow standardized comparisons among taxa with various dental morphologies. Our new approach represents a major improvement compared to the previous methods as it considers the variations of