Post-mortem alteration of diet-related enamel surface textures through artificial biostratinomy: A tumbling experiment using mammal teeth

Highlights

• First experiment on biostratinomy using dental microwear texture analysis

• Altered surfaces do not change original surface texture signature significantly.

• Feeding categories (grazing/browsing) still identifiable after 336 h of tumbling.

• Results suggest visually abraded fossil teeth do not need to be discarded per se.

Abstract

In the fossil record, teeth are often all that remains of a fossil organism. Dental microwear texture analysis (DMTA) is a common proxy for diet using dental wear features at the μm-scale, enabling comparative and quantitative assessments of various feeding traits in extant and extinct species. In extinct species, original diet-related dental wear features may be overprinted by post-mortem alteration including fluvial transport. Here we experimentally investigate the effects of mechanical alteration on diet-related 3D enamel surface texture (3DST) patterns of different mammal teeth. Post canine teeth of Equus sp., Capreolus capreolus and Otomys sp. are tumbled in sediment-water suspensions using three different grain size fractions of sand. The 3DST of the enamel surfaces are measured prior to and after each tumbling interval and characterised using ISO normed surface texture and SSFA parameters. In all species, we find several parameters to be almost unaffected by tumbling (stable parameters), while other parameters show inconsistent-directional shifts (unstable parameters). For Otomys, all three sediment grain size fractions result in abrasion of peaks and a reduction of overall surface roughness. For Equus, tumbling results in visible abrasive changes in the original wear patterns and the introduction of new wear features. Capreolus capreolus shows high variability in surface texture patterns prior to and after the experiment, hence we see ambiguous trends for changes in parameter values. However, even after 336 h of tumbling the browsing C. capreolus can still be distinguished from the grazing Equus sp.

Thus, biostratinomy may potentially modify diet-related 3DST causing non-systematic bias via mechanical abrasion, which is related to sediment grain size, duration of transport and geometry of teeth. However, the original diet-related 3DST is still preserved and a more prominent characteristic in DMT than the experimentally induced diagenetic alteration.

Introduction

Dental microwear texture analysis (DMTA) is a powerful proxy for oral behaviour, reflecting the mechanical properties of ingesta consumed during the last few days or weeks, and for tooth function in general. The tribosystem of occluding teeth is largely controlled by abrasion and attrition as a cause of material loss (Dahlberg and Kinzey, 1962). Patterns of wear features at μm-scales (roughness texture) can be related to feeding traits, enabling quantitative and comparative assessments of diet/ingesta related traits such as soft- or hard-object feeding among extant and fossil vertebrates (Solounias et al., 1988; Scott et al., 2005; Merceron et al., 2007; Semprebon and Rivals, 2007; Schulz et al., 2010; Winkler et al., 2013). Baker et al. (1959) were among the first to link microscopic wear features on teeth with abrasives ingested in the diet. In their study of sheep teeth, they concluded that internal abrasives, such as phytoliths, where hard enough to scratch enamel and thus indicate diet. Since then, many studies of tooth wear analyses at different spatial scales (mesowear, microwear, dental microwear texture analysis) for a large variety of vertebrates (mostly mammals) have been performed, and characteristic wear patterns related to their assumed natural diets have been described (Fortelius and Solounias, 2000; Schulz et al., 2010; Fraser and Theodor, 2011; DeSantis et al., 2013; Purnell et al., 2013; Calandra and Merceron, 2016).

In the fossil record, teeth are often the only preserved parts of animals and are therefore a crucial source of information about the feeding behaviour of extinct taxa. Dietary reconstructions based on the wear features of fossil occlusal tooth surfaces can easily be biased post-mortem due to mechanical and/or chemical modification (King et al., 1999; Dauphin et al., 2018). Therefore, reconstructions should be restricted to areas of a tooth which are relatively unaltered by post-mortem processes. This makes distinguishing between ante- and post-mortem surface alterations a crucial issue for reliable palaeodietary reconstructions based on 3DST (Calandra and Merceron, 2016).

There are two primary sources of post-mortem mechanical surface alteration on teeth (Teaford, 1988):

• Material loss occurring after death or before/during the burial of an animals remains.

• Material loss caused by excavation and preparation during/after the collection of a fossil.

Both sources of mechanical surface alteration can have an impact on ante-mortem occlusal wear features and possibly alter or obscure those wear features. Most post-mortem wear features are distinctively different in their size, shape or orientation (Teaford, 1988). In addition to mechanical alteration, chemical alteration may modify ante-mortem wear features, e.g. by acid attack in the stomach of a predator (e.g. Dauphin et al., 2003, Dauphin et al., 2018) or by dissolution processes in the soil/sediment environment. Overall, there is a substantial lack of experimental data on the effect of biostratinomy on mammalian teeth. Gordon, 1983, Gordon, 1984 tumbled extracted human molars and premolars with four different types of sediment, ranging from volcanic ash to “pea” gravel (~4–8 mm), in aqueous and non-aqueous environments for different time intervals (unspecified). Gordon found significant changes to the ante-mortem dental wear after 5 h and determined that alteration of the microwear pattern correlated positively with sediment grain size. However, tumbling seemed to abrade the identifiable microwear rather than add or produce new wear features (Gordon, 1984). Similarly, King et al. (1999) tumbled three human lower molar teeth from the late Neolithic with three different grain size fractions of sediment and water for 2–512 h and found evidence for changes in dental microwear only with the smallest grain size fraction. Coarse sand (500–1000 μm) produced no change in dental microwear and quartz pebbles (2000–11,000 μm) also resulted in little surface damage on the tooth. However, a noticeable abrasive effect was observed for the grain size fraction between 250 and 500 μm, which completely removed the original microwear features. Both studies showed a polishing effect on the dental surface rather than formation of new wear features comparable in morphology with those of the ante-mortem wear (Gordon, 1984; King et al., 1999). Nevertheless, taphonomically altered dental wear patterns could still be easily identified and distinguished from diet-related wear patterns (Teaford, 1988; King et al., 1999). For example, Martínez and Pérez-Pérez (2004) tested whether taphonomic and non-taphonomic wear features were clearly distinguishable on the buccal enamel surfaces of fossil hominin teeth from Olduvai and Laetoli (Martínez and Pérez-Pérez, 2004). In this sample, they found evidence for both chemical and mechanical alteration, as well as for well-preserved dental surface areas with no obvious post-mortem material loss, enabling the identification of taphonomically altered teeth. The combination of mechanical and chemical post-mortem wear has often drastically reduced sample sizes of analyses of fossil material (e.g., Ungar et al., 2008), however, this may not always be necessary.

All experimental studies conducted thus far are qualitative and have only been performed on human teeth. Although there are numerous studies connecting dental wear in other mammalian taxa with ingesta properties (i.e. Mainland, 2003a, Mainland, 2003b), there is as of yet no experimental approach to assess the influence of post-mortem abrasion on non-human teeth. Here we present the first experiment to quantitatively test the effect of biostratinomy on DMTA in three different herbivorous mammalian species.

We employ an experimental setup with three different sediment grain size fractions (fine to middle sand, 63–500 μm) and different tumbling time periods (from 0.5 to 336 h). Dental surface texture measurements of the same areas on wear facets were performed on teeth of three herbivorous (grazer and browser) mammals (Otomys sp., Capreolus capreolus and Equus sp.) before and after each tumbling interval, using a high-resolution disc-scanning confocal 3D-surface measuring system μsurf Custom. To quantify dental wear and detect potential alteration by the selected sediments during the experiments, two DMTA methods were employed: scale-sensitive fractal analysis (SSFA) and, what we refer to as 3D surface texture analysis (3DST). Both approaches employ optical profilometry to obtain 3D representations of the enamel surface at submicron resolution and evaluate the overall distribution and three-dimensional geometry of topographic features. In SSFA, surface features are characterised using length-scale and area-scale fractal analyses, describing complexity and anisotropy of the surface (e.g. Ungar et al., 2003, Ungar et al., 2012; Scott et al., 2006; Scott, 2012). In 3DST, standardised roughness (ISO 25178) and flatness (ISO 12781) parameters plus additional motif, furrow, direction and parameters are employed to characterise wear features (Schulz et al., 2013a, Schulz et al., 2013b; Purnell and Darras, 2016; Kubo et al., 2017; Purnell et al., 2017). A major focus of diet reconstruction via dental wear has been on the browser-grazer dichotomy in the artiodactyla and perissodactyla (e.g. Rensberger, 1973; Fortelius and Solounias, 2000; Damuth and Janis, 2011; Fraser and Theodor, 2011; Schulz et al., 2013a). The browser-grazer dichotomy is well reflected at the different scales of dental wear evaluation (i.e. mesowear, microwear, and dental microwear texture analysis). Browsers are characterised by less abraded tooth wear, which is reflected in a mesowear profile with sharp cusps (Fortelius and Solounias, 2000; Kaiser et al., 2000). The surface is dominated by high complexity and low anisotropy values (SSFA, Scott, 2012; Ungar et al., 2012) and an overall smooth surface with low microscopic peaks (3DST, Schulz et al., 2013a, Schulz et al., 2013b) and pits (microwear, Walker et al., 1978). In contrast, grazers show mesowear dominated by rounded or blunt cusps and scratches (microwear, Walker et al., 1978), with an overall greater surface roughness, higher and more frequent peaks and deeper dales but a generally less variable pattern (3DST, Schulz et al., 2013a, Schulz et al., 2013b) and high anisotropy/low complexity values (SSFA, Scott, 2012; Ungar et al., 2012).

The aim of this study is to test whether diet-related 3DST (i.e. browser-grazer differences) are resistant against mechanical alteration during sediment transport under a post-mortem abrasive regime. If biostratinomical changes on the wear facets of teeth occur, are they recognizable as post-mortem wear or do they show patterns comparable to ante-mortem wear? We therefore tested if the 3DST of teeth from a grazer (Equus sp.) and browser (Capreolus capreolus) could still be assigned to the correct dietary category after tumbling intervals of up to 336 h. In other words, could the surface texture parameters normally found to distinguish between these two dietary categories still distinguish between browsers and grazers despite mechanical abrasion meant to mimic sediment transport? We expect that different surface texture parameters will be affected differently by our experimental setup. We anticipate that parameters which describe extremes of the surface (e.g. maximum height) will more prone to alteration by abrasion compared to parameters describe mean roughness of the surface. Thus, we intend to identify parameters that show the potential to bias dietary interpretations. Our goal is further to not only test the hypothesis that sediment grain size has an impact on dental microwear texture, but that tooth geometry does as well. Additionally, we analysed the altered parameters after the experiment to test whether the extreme outliers in parameter values can be used as indicators of taphonomic alteration.