Measuring dental wear equilibriums—the use of industrial surface texture parameters to infer the diets of fossil mammals

Abstract

Inferring the diet of fossil mammals is a major approach to mammalian palaeobiology and palaeoecology. Dental wear provides a unique record of oral behaviour, available for most extant and fossil mammals. Dental wear facets are thus one of the immediate habitat interfaces allowing analysis of food selection, food availability, and dietary segregation in fossil and extant communities based on the same source of information. As any surface, wear facets can be described as a complex landscape with more or less elevated alterations. In order to describe technical surfaces, a system of surface roughness parameters is available, which are highly indicative statistical tools in characterizing surface micro texture in industrial applications. We test the hypothesis, that dental microwear can be described by established surface roughness parameters, and that these parameters can help to identify the dietary traits of fossil herbivores. We investigated three extant African bovid species, the greater kudu, the bushbuck and the hartebeest, which were selected because each of them represents one of the major dietary classes as browsers, mixed feeders and grazers, respectively. As a fossil species we investigated the upper Miocene hipparionine horse Hippotherium primigenium from the upper MN9 Dinotheriensande. A diamond stylus profiling instrument was used to measure surface roughness on four phase I shearing facets of upper molars. Discriminant analyses were performed to test 24 roughness parameters for their ability to predict dietary classes. Several parameters were identified as confidently separating the grazer and the browser in this comparison, while the mixed feeder is intermediate. For H. primigenium, we find a mixed feeding signal confirming earlier dietary reconstructions of the species based on mesowear and microwear. The degree of attrition in browsers is reflected by a high bearing ratio (RTp). In grazers, the lower bearing ratio reflects the striated and elevated topography of an abrasion dominated surface. The bearing ratio may be understood as immediately reflecting the position of the attrition–abrasion equilibrium, and thus bridging the two major methods of tooth based dietary evaluation, the microwear- and the mesowear method. Surface texture parameters are found to encapsulate information about basic dental function that we hypothesize has fundamentally driven dental adaptation in mammalian evolution.

Introduction

Microwear research has shown that microscopic wear patterns on teeth may provide insight into diversity in diet and tooth use of a variety of mammals, including modern and prehistoric man. The logic behind microwear is relatively straightforward. Essentially antagonistic teeth in a dentition and many abrasives in the environment can leave microscopic scars on teeth. In total, those scars result in tooth wear. Analyses of variations in wear patterns provide a record of the oral behaviour that caused them. Nevertheless, interpretation may be difficult because there is no general theory of wear—thus the effects of different abrasives and the ways in which they modify teeth must be determined empirically.

Microwear can only be evaluated on dental surfaces which are subject to abrasion or attrition to a certain degree. Besides studies devoted to dental microwear in mammalian incisors (e.g. Walker, 1976), analysis of wear in molars has been of two types: those concerned with inference of jaw movements from wear patterns (Kay and Hiiemae, 1974, Ryan, 1979, Kay, 1981, Krause, 1982, Teaford and Walker, 1982, Teaford and Walker, 1983, Gordon, 1984a, Gordon, 1984b, Young and Robson, 1987, Wilkins, 1988, Teaford, 1989) and those concerned with the inference of diet (Walker et al., 1978, Covert and Kay, 1981, Peters, 1982, Gordon, 1988, Solounias et al., 1988, Solounias and Moelleken, 1992, Teaford, 1993, Ungar, 1996, Caprini, 1998, Mainland, 1998, Mainland, 2000, Rivals and Deniaux, 2003, Merceron et al., 2004a, Merceron et al., 2004b, Merceron et al., 2005).

Almost all studies in microwear analysis published so far are based on a microscopic inspection of enamel facets involved in food crushing, shearing and grinding. Magnifications used range from 10× to 500× (e.g. Hayek et al., 1992, Rivals and Deniaux, 2003), using SEM or light microscopy (Solounias and Semprebon, 2002, Merceron et al., 2004a, Merceron et al., 2004b, Merceron et al., 2005). Subsequently, the resulting light- or electron microscopic image is analysed in a semi qualitative approach. In most microwear studies concerned with ungulates (Solounias et al., 1988, Hayek et al., 1992, Solounias and Moelleken, 1992) frequencies of pre defined scar morphs are dichotomized into pits and scratches according to dimensional cut-off points. Some studies also employ a polychotomous system of 3 to 5 scar notations (Solounias and Hayek, 1993, Solounias and Semprebon, 2002, Rivals and Deniaux, 2003, Merceron et al., 2004a, Merceron et al., 2004b, Merceron et al., 2005). From the frequencies of these morphs inference on the diet is made by using extant species with known diets as a reference. In a very broad sense, most authors seem to agree in the observation that scratches are the result of abrasives as e.g. phytoliths and grit, which are particularly frequent in grass-dominated abrasive diets (e.g. Merceron et al., 2004a). Pits are considered to have a much less distinct origin, but seem to be more frequent in e.g. leave-eaters and browsers. All transitional stages between the two extreme morphs of scars may occur in either of the two extreme feeding types. Their frequencies are further largely dependent on a set of independent variables as individual age, tooth position, wear facet and the exact locality within a certain wear facet (Teaford, 1988). A recent approach by Solounias and Semprebon (2002) avoids SEM preparation and scar measuring, but relies on visual light microscopic inspection and counting of pre defined scar morphs. The relation of microwear features with certain dental wear facets is best understood in primates, where the process of facet formation is most clearly related to fundamentally different functions in food comminution as puncture-crushing (only abrasion), shearing (attrition and abrasion) and grinding (more abrasion, less attrition) (Butler, 1972). Very recent work by Ungar et al. (2003) employs confocal microscopy and subsequent fractal analyses to characterise dental wear facets.

There has been some experimental work aiming for a better understanding of the interaction of food and tooth enamel surface (Peters, 1982, Maas, 1991, Gügel et al., 2001). In his review paper Teaford (1988) however points out that: “experimental studies of molar microwear have just begun. As a result, they have probably raised more questions than they have answered”. The problem still remaining is: We know too little about how scars are physically and chemically related to certain components of the diet or the habitat, and in which extent they are the result of tooth morphology and function, foraging and chewing behaviour, digestive physiology, thegosis (Every et al., 1998), salivary and food chemistry, or other unknown factors. The state of the art therefore is that microwear analysis is still much of a black box but nevertheless, it is a useful tool in assessing basic physical properties of the diet and dental function of extant and extinct mammals. For this reason, the presented study abandons the classical microwear approach fundamentally in using surface roughness parameters to characterize the physical topography of enamel facets.

A variety of measurement methods is available to collect topographic data of surfaces on a microscopic scale. These methods can be broadly grouped into contact- and non-contact methods. Contact surface sensors sample surfaces with a needle-sensor (diamond or ruby stylus), which moves across the surface mechanically under constant pressure. It converts the vertical and horizontal movements of the needle into an electrical signal. Non contact methods are e.g. the laser stylus scanning method (Kaiser and Katterwe, 2001), light refraction and diffraction methods, interferometric coherence, acoustic emission and confocal scanning microscopy (Evans et al., 2001, Ungar et al., 2003). All non-contact techniques depend on light reflected or refracted from the surface and thus infer on surface topography indirectly. They therefore require specific optical properties of the sample surface in order to allow accurate measurement. Contact methods as the diamond stylus method, however, are robust and widely independent from such environmental conditions. In addition, the measurement equipment is comparably inexpensive, small and not sophisticated. The present study therefore uses a diamond stylus profiling instrument to measure facet surfaces in ungulate teeth. Other possible applications of surface texture analysis in palaeontology would be the quantification of bone weathering, the quantification of chemical and biogenous surface erosion to bone and teeth as a taphonomic application, or the analysis of trace fossils (Kaiser and Katterwe, 2001).

All surface measuring systems reconstruct 3D surface models by assembling line profiles. In finely striated surfaces, much of the topographic information, however, is encapsulated in a single line profile. Because profiles can much easier be measured then 3D-surface models, most technical applications in surface evaluation base on line profile information rather than 3D-surface models. The resulting surface texture can be generally subdivided into three major classes of geometric information—form, waviness and roughness. Form describes the broad scale geometry, like given e.g. by the radius of a cylinder or bowl shaped body, which usually is of limited use in evaluating tooth wear. Waviness includes the more widely spaced (longer wavelength) alterations of a surface from its nominal shape, while roughness includes the finest (shortest wavelength) alterations.

Any surfaces can be described as a complex landscape with more or less elevated alterations. In many respects, worn tooth enamel equals technical surfaces. That the level of “texture” is the right one for evaluating palaeodiet was already suggested by Rensberger (1978). In order to describe technical surfaces based on surface profiles, a system of surface structure parameters is available for each class of wavelength in surface alterations. There are parameters describing waviness and those describing roughness. Roughness parameters describe vertical topographic features (height parameters), horizontal features (spacing parameters) and both integrated (hybrid parameters). Taken together, roughness parameters are a highly indicative statistical tool to characterize surface micro texture (ISO references).

Besides these high frequency irregularities, tooth wear may expose alterations representing areas of differential wear resistance in the tooth enamel, which are characterised by their uniform frequency and amplitude. Technical surfaces further may have unidirectional lay when striations across the surface manufacturing using a unidirectional tooling process (e.g. sawing, planing, milling, grinding). Similar parallel alignments are also found in herbivorous cheek teeth where they indicate the chewing (power) stroke direction. In the chewing systems of ruminants and many equids striations on phase I shearing facets are therefore oriented bucco-lingually (Greaves, 1973, Janis, 1990). The high degree in correspondence between industrially machined surfaces and the wear related micro scars of herbivorous cheek teeth leads us to propose the following hypothesis, which we wish to test in this study.