Experimental alteration of diet-related dental wear and isotope proxies in teeth - implications for fossil food web reconstruction of Permo-Carboniferous vertebrates
Date issued
Authors
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
License
Abstract
The main focus of this thesis was to determine whether dental microwear texture (DMT) and non-traditional stable isotopes are overprinted by diagenetic influences to provide a more thorough understanding of mechanisms behind, and effect of, post-mortem alteration in dental tissues. First, the teeth of several extant species were subjected to different experimental conditions intended to simulate diagenetic alteration. DMT and non-traditional stable isotopes were measured both before and after each experiment, to characterize dental wear features and in-vivo chemical composition as a result of simulated diagenetic alteration. Second, both dental microwear texture analyses (DMTA) and non-traditional stable isotopes were applied to fossil Permo-Carboniferous non-mammalian tetrapods to reconstruct dietary preferences and niche partitioning in the earliest terrestrial food webs.
In general, DMTA and non-traditional isotope systems are increasingly being applied to a great variety of taxonomic groups, such as Reptilia (including dinosaurs) and fish, as well as in modern and fossil ecosystems, to determine feeding categories and positions in trophic webs. In the fossil record, DMTA, as well as non-traditional stable isotopes might be biased by mechanical or chemical post-mortem alteration. Although, fossil tooth enamel is highly mineralized, and thus less prone to mechanical and chemical alteration compared to other hard tissues such as dentin and bone, it can still be diagenetically overprinted. To evaluate the impact of post-mortem fluvial transport, aeolian sediment transport or digestion of teeth on the dental surface, a tumbling, a sandblasting and an acid etching experiment were performed with different vertebrate species using various sediment grain size fractions and acidity conditions. These experimentally produced datasets were compared to a reference dataset of badly preserved dental surfaces resulting from sample excavation and preparation (e.g. adhesive varnish, bubbles in the mould). Using standardized parameters for dental wear (i.e. ISO surface and SSFA), the results of these analyses demonstrate that many post-mortem dental surface alterations are difficult to distinguish from diet-related ante-mortem wear features using DMT parameters alone. Instead, it is necessary for an observer to identify post-mortem features, and to evaluate the degree of alteration before samples are included in a dataset. In addition to diagenesis, limited oral processing behaviour or variable tooth replacement rates can also bias DMTA. For example, while some suborders with reduced mastication and tooth replacement, such as extant lepidosaurs (Reptilia), show diet-related ante-mortem wear features, the results presented in this thesis indicate that extant shark species are not good candidates for diet reconstruction using dental microwear texture. In sharks, it seems that either tooth-to-food contact is too reduced, tooth replacement rate is too fast, or a combination of both factors prevent diet-related dental wear from forming.
To increase our understanding chemical alteration on dental tissue, i.e. dentin and enamel, an in-vitro alteration experiment with cubes cut from an elephant tooth were placed in aqueous isotope-enriched tracer solution at different temperatures and for different durations. Based on changes in the isotope ratios and Raman spectra of experimentally altered dental cubes, dissolution and re-precipitation was identified as the alteration mechanism. Furthermore, it was demonstrated that tooth enamel is highly resistant to chemical alteration of isotopic composition (at least for the isotopes tested in this study), while dentin is more prone to diagenetic alteration.
Finally, DMTA and non-traditional stable isotopes (δ44/42Ca and δ88/86Sr) have been applied to extinct non-mammalian tetrapods, to identify early herbivorous and faunivorous taxa in fossil ecosystems from different Permo-Carboniferous localities in the U.S.A., South Africa, and Germany. For the stable isotopes, extant lepidosaurs with distinct feeding strategies were measured to establish a modern comparative dataset. δ44/42Ca showed a separation of carnivorous and herbivorous taxa in both extant lepidosaurs and extinct non-mammalian tetrapods, with carnivores being more negative than herbivores. Insectivorous and ovivorous extant lepidosaurs yielded the least negative δ44/42Ca values of the measured samples. In general, no systematic offset between lepidosaur bones and teeth were detected, while in fossil non-mammalian tetrapods bones displayed less negative δ44/42Ca values than tooth enamel. So far it is not clear if this offset is caused by diagenesis or by physiological fractionation processes. δ44/42Ca values of upper Permian specimens show a more pronounced trophic spacing compared to lower Permian species, possibly caused by differences in diet or in physiological fractionation processes. δ88/86Sr showed distinct trophic spacing in extant lepidosaurs, similar to that observed for stable Ca. However in fossil specimens, Sr is clearly diagentically overprinted, resulting in positive δ88/86Sr values. In fossil tetrapods, DMTA revealed no differences between carnivores and herbivores, while insectivores differed significantly from both. Interestingly, differences were detected between taxa in both herbivorous and carnivorous feeding groups, suggesting that the mechanical properties of a food (e.g. type and part) and/or different oral processing behaviours, can introduce significant variation in dental wear signals, even within a feeding category.
Principal component analyses of combined DMTA and δ44/42Ca revealed distinct differences between carnivorous, herbivorous and presumed omnivorous taxa. Therefore, it is assumed that these taxa occupied species-specific niches in Late Carboniferous and Permian ecosystems. However, more faunal and floral samples from the respective ecosystems are needed to reconstruct δ44/42Ca values of the primary producers, i.e. plants, and to better understand the differences in δ44/42Ca values in herbivores. Nevertheless, the combination of DMTA and stable Ca data presented in this thesis illustrate the significant potential of this toolkit for more precise reconstruction of niche partitioning in fossil ecosystems when compared to a single proxy approach.