Why should traceology learn from dental microwear, and vice-versa?

also learn from traceology, especially regarding sample prepara- tion, experimentation and residue analysis. We hope that this focus article will stimulate more awareness, exchange and collaboration between pa- leontologists and archeologists, and especially between dental and artifact microwear analysts. Paleontology, archeology and the ﬁ eld of surface analysis as a whole would all bene ﬁ t from such cooperation.


Introduction
Dental microwear analysis refers to the study of microscopic marks present on tooth surfaces that result from the wear of food particles and external abrasives (Green and Croft, 2018). It belongs to paleontology, but it also has applications in biology, dentistry and archeology (Merceron et al., 2014;Hara et al., 2016;El Zaatari et al., 2016). It is commonly applied to infer diets of fossil vertebrates (Ungar and Evans, 2016) and to reconstruct paleo-environments (Ungar et al., 2012). Traceology is the study of all physical traces on an artifact's surface, which include use-wear, traces of production, non-utilitarian wear (e.g. transport) and post-depositional alterations Thomas et al., 2011). Surface modifications can be analyzed at different scales; we will focus here on the microscopic scale, and refer to it as artifact microwear analysis. Traceology, in this sense, is a sub-discipline of archeology and is included in functional analyses (Marreiros et al., in prep.). It aims at identifying the function of artifacts in terms of action and worked material, i.e. use, to infer past human behavior.
Both dental and artifact microwear analyses try to address similar questions: what have the objects (teeth/artifacts) been used for (food items/worked material) and how (chewing mechanics/tool kinematics)? While the objects analyzed are different, both fields document the wear produced on the sample's surfaces. Both cases therefore represent tribo-systems, i.e. systems of two contacting objects in relative motion to one another (Brown et al., 2018).
The two methods are decades old and have experienced several rounds of developments, mainly driven by the lack of repeatability and reproducibility of early attempts (see below). More recently, methods to quantify surface textures have been developed to counteract these issues. While archeologists are aware that paleontologists apply similar methods (Evans and Macdonald, 2011;Ibáñez et al., 2019;Martisius et al., 2018;Stemp et al., 2015), the reverse is on the whole not true. Generally, the methodological overlap between archeology and paleontology is rarely recognized or exploited and, accordingly, few  Ungar and Evans (2016, p. 3-4) write that "The papers in this volume demonstrate the value of partnerships between paleontologists or archeologists on the one hand, and surface metrologists on the other …" (our italics).
In this focus article, we will first briefly review the methodological developments of dental and artifact microwear analyses that led to quantitative surface texture analyses (QSTA). We will then detail some aspects that can and should be transferred between the fields. We hope that this article will help both fields move forward so that QSTA can be performed with the same approach and quality standards.
From the 1920s to the 1960s, paleontologists and dentists realized that the orientation of scratches mirrors the direction of chewing movements (Butler, 1952;Dahlberg and Kinzey, 1962;Mills, 1955;Simpson, 1926). DMA became more prevalent in the 1970s with the influential works of Walker et al. (1978) and Rensberger (1978), who showed that the proportions of pits and scratches correlate with diet. These two studies were performed with SEM and, as a result, most subsequent research used this equipment.
In response to these issues, Ungar, Scott, Brown and colleagues proposed to use confocal microscopy to acquire 3D representations of the tooth's surface at high magnifications and to automatically quantify surface textures with Scale-Sensitive Fractal Analysis (SSFA; Scott et al., 2005Scott et al., , 2006Ungar et al., 2003). Building upon this methodology, other groups of researchers started to apply the standardized 2D ISO 4287 (1997; Kaiser and Brinkmann, 2006) and 3D ISO 25178 (2012;Calandra et al., 2012;Purnell et al., 2012;Schulz et al., 2010) parameters to measure surface textures. Many studies have applied the quantitative analysis now termed Dental Microwear Texture Analysis (DMTA) to almost all groups of mammals (Calandra and Merceron, 2016), to fishes (Purnell et al., 2012) and to reptiles (Bestwick et al., 2019;Winkler et al., 2019b). Numerous parameters have been applied to quantify surface textures. The four SSFA parameters are the most used, but 30 ISO 25178 parameters have been regularly applied. Flatness (ISO 12781), motif, furrow, direction and isotropy analyses have also been explored (Schulz et al., 2013a;Winkler et al., 2019a). Interferometry (Estebaranz et al., 2007;Merceron et al., 2014;Souron et al., 2015) and focus variation microscopy (Gill et al., 2014;Purnell et al., 2012Purnell et al., , 2017Zhang et al., 2017) have been tested, but confocal microscopy quickly became the equipment of choice for DMTA (Schulz et al., 2013a).
DMTA has become a well-established method. Several research groups now possess the necessary equipment to acquire 3D surface data. Nevertheless, each piece of equipment and each setting will acquire the surface in a different way (Calandra et al., 2019a). A Fig. 1. Main developmental steps of dental (top) and artifact (bottom) microwear analyses leading to quantitative surface texture analyses (DMTA and QAMA, respectively). Developmental steps of DMA similar to steps on AMA are superimposed in red on the AMA chart at the bottom. All dates correspond to the introduction of a new methodology. In most cases, these new steps were implemented in following studies. See text for details and references. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) processing workflow has been proposed to minimize inter-microscope variability (Arman et al., 2016).
Some researchers follow the blind and repeated methodology of Mihlbachler et al. (2012) to limit intra-and inter-observer errors in DMA, but it is not universally adopted (Rivals et al., 2018;Semprebon et al., 2016).
Although the discipline dates back to the 19th century (Olausson, 1980), the seminal work by Semenov (1957Semenov ( , 1964 is usually recognized as having defined the current practice in AMA (Evans et al., 2014a;Stemp et al., 2016). Some researchers emphasized the need for quantification already in the 1970s (Keeley, 1974;Schiffer, 1979). However, besides some attempts with limited results in the 1980s based on digital imaging (Grace, 1989;Grace et al., 1985), interferometry (Dumont, 1982) and tactile profilometry (Beyries et al., 1988), AMA remains largely qualitative. After a long debate over which scale of observation is best suited, a consensus emerged that the high and lowpower approaches should be combined (Gräslund et al., 1990;Marreiros et al., 2015;Odell, 2001).
The lack of repeatability/reproducibility is still prevalent in the widely applied qualitative approach (Evans et al., 2014a and references therein). Yet, QAMA remains nascent: many different pieces of equipment are used, almost no effort is made to standardize acquisition and analysis settings (Calandra et al., 2019a), most analyses include only a few parameters, and no study has yet applied it to infer the function of archeological artifacts (except d'Errico and Backwell, 2009 on bone tools). Most importantly, papers on QAMA must still justify why QSTA makes sense (Evans et al., 2014a;Ibáñez et al., 2019;Stemp et al., 2016).

What use-wear can learn from dental microwear, and viceversa
The parallels between DMA and AMA should now have become evident. Both have followed the same methodological developments. However, DMTA is a much better established and accepted method in paleontology than QAMA is in archeology. Indeed, DMTA has been applied in many studies (see section 2) and few paleontologists would not recommend DMTA over DMA. In this section we emphasize the main aspects that each community should borrow from the other.
It should be noted that quantification here refers only to the calculation of surface texture attributes. Such quantification requires the surface topography to first be digitalized into scaled reconstructions of surface profile (2D) or areal (3D). In other words, QSTA has two parts: (1) acquisition of surface data and (2) quantification of surface attributes. While quantification requires acquisition, 3D models can also be assessed visually (d' Errico and Backwell, 2016;Wei et al., 2016).
Based on the success of DMTA, traceologists should recognize that QSTA has great potential. Indeed, this type of analysis can be applied to any surface data (Brown et al., 2018). Of course, which parameters are the most relevant depend on the wear processes. Therefore, traceologists should not be skeptical about QSTA, but rather work to adapt it to the constraints of archeological artifacts (e.g. different raw materials).
Many parameters are available to quantify surface texture. Volume, isotropy and direction parameters in particular might prove relevant to QAMA. All of them are available in MountainsMap Imaging Topography (Digital Surf, Besançon, France), in the modules "advanced topography", "particle analysis" and "scale-sensitive analysis".
Protocols to minimize inter-microscope variations have been proposed (Arman et al., 2016;Kubo et al., 2017). There is literature on the (biological) meaning of the ISO/SSFA parameters (see reviews cited before). Finally, the approach of Mihlbachler et al. (2012) could easily be adapted to AMA in order to reduce intra-and inter-observer biases.
To our knowledge, Martisius et al. (2018) is the only study to include specialists from both fields. The study applied an established workflow in DMTA. It also selected some potentially interesting parameters from all classes of ISO 25178 parameters, which eased the subsequent statistical analysis. This collaboration most likely allowed meaningful results to be produced much faster.
The accuracy of molding/casting materials has been investigated by researchers from both fields (Goodall et al., 2015;Macdonald et al., 2018;Mihlbachler et al., 2019). The insights of each study should be transferred to the other field.
Residue analysis, is another method to infer the function of a tool (Fullagar, 2014;Haslam et al., 2009). Phytoliths have been found embedded onto primate teeth (Ciochon et al., 1990;Fox et al., 1994) and residues trapped in dental calculus have also been used to infer past human diets (Henry et al., 2011;Weyrich et al., 2017). The extraction and identification of phytoliths could be manageable in extinct vertebrates. Traces of blood, bones, amino acids, proteins, etc. might be preserved and identifiable in fossils too (Bordes et al., 2017;Borgia I. Calandra, et al. Journal of Archaeological Science 110 (2019) 105012 et al., 2017Monnier et al., 2018).

Future directions
DMTA is a well-established method, more mature than QAMA. Nevertheless, there is scope for refinement and improvement of both methods, especially referring to choice of relevant parameters, comparison of microscopes and understanding of microwear formation.
As for most of our colleagues (Evans et al., 2014a;Kimball et al., 2017;Stemp et al., 2016;Stevens et al., 2010), we do not argue that QAMA should replace AMA; we rather recommend to combine both methods. QAMA could, for example, allow inferring the worked material more precisely once AMA has been performed (Ibáñez et al., 2019;Martisius et al., 2018). Ideally, residue analysis should also be applied in combination Stemp et al., 2016), although thoughts should be given to sample preparation protocols so that residues are not washed off prior to analysis (Rots et al., 2016).
Post-depositional processes have the potential to blur any functional signal. This topic has been barely addressed in DMTA (Böhm et al., 2019;Calandra and Merceron, 2016;El Zaatari, 2010); QAMA has until now mainly focused on experimental tools, so these processes have rarely been taken into account (Caux et al., 2018;Galland et al., 2019;Vietti, 2016;Werner, 2018). This topic remains of major importance for future studies in both fields.
Each field has a lot to learn but some things can and should be learned from other fields rather than re-invented. Paleontologists and archeologists are used to burrowing from other disciplines (geology, geography, ethnography, tribology, biology, pathology, forensics …). They have worked closely with surface metrologists and we hope that they will continue doing so. When working together, synergistic effects will allow both fields to grow faster, and a quantitative analysis of surface wear common to both teeth and artifacts will have a broader resonance in paleontology, archeology and beyond.

Funding
This research has been supported within the Römisch-Germanisches Zentralmuseum -Leibniz Research Institute for Archeology by German Federal and Rhineland Palatinate funding (Sondertatbestand "Spurenlabor") and is publication no. 3 of the TraCEr laboratory. The funding source had no involvement in study design; in the collection, analysis and interpretation of data; in the writing of the report; nor in the decision to submit the article for publication.

Conflicts of interest
The authors declare that they have no conflict of interest.