Dental microwear texture reflects dietary tendencies in extant Lepidosauria despite their limited use of oral food processing

Abstract

Lepidosauria show a large diversity in dietary adaptations, both among extant and extinct tetrapods. Unlike mammals, Lepidosauria do not engage in sophisticated mastication of their food and most species have continuous tooth replacement, further reducing the wear of individual teeth. However, dietary tendency estimation of extinct lepidosaurs usually rely on tooth shape and body size, which allows only for broad distinction between faunivores and herbivores. Microscopic wear features on teeth have long been successfully applied to reconstruct the diet of mammals and allow for subtle discrimination of feeding strategies and food abrasiveness. Here, we present, to our knowledge, the first detailed analysis of dental microwear texture on extant lepidosaurs using a combination of 46 surface texture parameters to establish a framework for dietary tendency estimation of fossil reptilian taxa. We measured dental surface textures of 77 specimens, belonging to herbivorous, algaevorous, frugivorous, carnivorous, ovivorous, insectivorous, molluscivorous, as well as omnivorous species. Carnivores show low density and shallow depth of furrows, whereas frugivores are characterized by the highest density of furrows. Molluscivores show the deepest wear features and highest roughness, herbivores have lower surface roughness and shallower furrows compared to insectivores and omnivores, which overlap in all parameters. Our study shows that despite short food–tooth interaction, dental surface texture parameters enable discrimination of several feeding strategies in lepidosaurs. This result opens new research avenues to assess diet in a broad variety of extant and extinct non-mammalian taxa including dinosaurs and early synapsids.

1. Introduction

Lepidosauria originated in the Early–Middle Triassic and have since immensely diversified [1]. Today, they are represented by more than 9000 extant species with a global distribution and huge ecological diversity [2,3]. They further show great dietary diversity (figure 1), with specialized herbivores such as frugivores and algaevores, specialized faunivores such as molluscivores, insectivores and carnivores as well as generalistic forms with omnivorous diets composed of different proportions of animal- and/or plant-matter [8,9]. This diversity is represented in dental morphologies (figure 1) within, as well as between, dietary groups. The large disparity in tooth morphologies complicates directly correlating form and function and thus makes it difficult to estimate diet, especially in fossils where behavioural observations or stomach content analyses are impossible. In a morphological approach for diet tendency estimation, tooth complexity (i.e. occlusal tooth crown topography) measured using orientation patch count has been found to increase with higher proportions of plant material in the diet [10], however, overlap in tooth complexity is greater, and thus the relationship between diet and tooth complexity is weaker, compared to mammals [10]. Moreover, unlike mammals, several lepidosaurian species undergo a dietary change during ontogeny (often accompanied by morphological changes in tooth morphology, e.g. Varanus niloticus [11]), which allows us to study changes of tooth wear within the same species feeding on different food types. Their heterogeneity in dietary adaptations reflected in dental morphology makes Lepidosauria an ideal test case to apply microscopic dental wear analyses. Dental wear for dietary tendency estimation is frequently investigated by dental microwear texture analysis (DMTA) in extinct and extant mammals [12–16]. So far, this method has not been used for dietary tendency estimation in extant reptilian taxa. Both, feeding experiments [17–19] and studies of museum material belonging to wild extant species [12,13,20–23] have facilitated our understanding of how microwear textures are formed in mammals and generated reference data for comparison with texture data from fossil mammalian species. DMTA, as well as the stereomicroscopic dental microwear (which uses the ratio of light-microscopically observable ‘pits’ and ‘scratches’, e.g. [24,25]) were long restricted to large herbivorous mammals [13,22,26] and extant and extinct primates as well as for reconstruction of hominin diets [27–30]. Stereomicroscopic dental microwear has also been applied to dinosaur species, mainly sauropods, hadrosaurs and ceratopsians [31–38]. Further work focused on different Triassic amniotes [39] or extinct heterodont crocodiles [40,41]. However, most of these studies used orientation of microwear features to reconstruct chewing movements. Some studies also used microwear to infer dietary niche partitioning between sympatric species and suggested harder, more abrasive diets for species with a higher pit count [32,42]. This work suggests that methods of dental wear analysis such as microwear are applicable to reptilian taxa, and it would be highly promising to evaluate them in extant species with known dietary tendencies.

Figure 1.

Simplified cladogram of lepidosaur species after [3] analysed in this study with examples of dental disparity in extant Lepidosauria (scale bar 1 cm). Dietary groups are colour-coded, tooth symbols with capped base indicate acrodont ankylosis, i.e. no continuous tooth replacement; simple tooth symbols indicate continuous tooth replacement [4–7]. Exemplary three-dimensional images of enamel surface textures from one individual of each dietary group are shown. Scan size 160 × 160 µm. Note that all scans are scaled to the same vertical scale given. Branch length is not informative. (Online version in colour.)

Several researchers have attempted to apply DMTA in other non-mammalian taxa such as durophagous fishes [43] and synapsids, i.e. Tritylodontia [44]. For DMTA, either surface texture parameters from scale-sensitive fractal analysis (SSFA) and/or International Organization for Standardization (ISO)-normed surface roughness and flatness parameters combined with further motif, furrows, direction and isotropy parameters (three-dimensional surface texture analysis, 3DST) are usually used. Here, we propose a combined approach for extant Lepidosauria, because 3DST was successfully employed for non-mammalian species [43] and homodont mammals [45] and the two SSFA parameters Asfc (complexity) and epLsar (anisotropy) have proven highly effective in distinguishing between feeding tendencies in herbivorous and carnivorous (mammal) taxa [46,47].

Still, the relationship between diet and microwear (and DMTA) in non-mammalian tetrapods remains poorly known compared to that of mammals. Mammals, in contrast to reptiles, have evolved sophisticated, three-dimensional oral-food processing through mastication and specialized heterodont dentitions [48–50]. Their teeth repeatedly interact with the ingesta and occluding tooth surfaces, resulting in distinct wear patterns that are informative of diet abrasiveness and material properties [12–14,24,51,52]. We do not know if wear marks on reptile teeth reflect dietary abrasiveness and material properties of the diet or if, without sophisticated mastication, the contact between ingesta and teeth is sufficient to form characteristic and distinctive surface (texture) wear patterns that relate to specific dietary groups.

Within lepidosaurs, several squamatan species and Sphenodon are known to perform intra-oral food processing [53–58], which makes the formation of distinct diet-related tooth wear more likely. However, the majority of lepidosaurian taxa (with several exceptions, e.g. Sphenodon and Uromastyx) only use a simple arcilineal oral food processing strategy [59,60]. Their functional repertoire for food processing consists of puncture crushing, lingual side-to-side transport of the prey item, repositioning of prey within the jaws for another processing stroke, and palatal crushing [56,61].

Moreover, most mammals possess heterodont dentitions, which have specialized tooth positions for different tasks during food gathering and oral processing (incisors, premolars, molars) [50]. Molar occlusal wear is used for dietary tendency estimation, whereas incisal wear is less informative and sometimes even contradictive to molar wear [62,63]. Incisors are often only used for cropping the food or for non-diet related activities (e.g. grooming, burrowing) [64–66]. Mammalian molars show wear facets, which are formed during attritional tooth-tooth and abrasive food-tooth contacts [67]. Dental wear analyses usually focus on the functionally homologous wear facets for inter-species comparisons. However, in predominantly homodont taxa-like Lepidosauria, it is unclear which areas of the tooth crown are best for the analysis of dental wear. As a result, determining which tooth position and facet-like area on the tooth would be most suitable for dietary tendency estimation in non-mammalian taxa remains challenging and will be addressed in this study.

We propose that dietary specializations are reflected in diet group specific dental microwear texture (DMT) patterns, as abrasiveness of ingesta is expected to influence formation of surface textures consistently across taxa [37]. Here, we compile, to our knowledge, the first DMT reference dataset of extant lepidosaurs to address the following key research questions:

• (i) do lepidosaurs show variation in enamel surface texture patterns that reflect dietary tendencies?

• (ii) are surface texture patterns independent of the tooth position within a dentition? and

• (iii) are ontogenetic changes in dietary tendencies from insectivory to molluscivory associated with differences in enamel surface texture patterns (e.g. in Varanus niloticus)?

2. Results

(a). Dataset 1: four scans per specimen

Dietary groups of extant Lepidosauria show a total of 186 significant pairwise differences in 42 out of 46 surface texture parameters (electronic supplementary material, table S4). The best discriminating surface texture parameters between dietary groups were selected as those displaying the highest number of pairwise differences, involving density and depth of furrows (medf, metf), height (Sa, Sq, Sxp), amplitude and complexity (Sdq, Sdr), areal material ratio (Smc) and volume (Vmc, Vv, Vvc, Vvv) (electronic supplementary material, figure S1). A principal components analysis (PCA) with these 12 best-discriminating parameters reveals separation of dietary groups in a dietary space defined by PC1 (78%) and PC2 (15.5%), which together explain 93.5% of the observed variance (figure 2a). The strongest overlap in dietary space exists between herbivores, omnivores and insectivores. Specialized frugivores are partly separated from generalistic herbivores along PC1. Carnivores are also partly separated from herbivores along PC1 but show some overlap. Separation of carnivores from other faunivores (omnivores, insectivores) results from PC1. Molluscivores are well separated from all other dietary groups along PC1. Height and volume parameters, areal material ratio (Sa, Smc, Sq, Sxp, Vmc, Vv, Vvc, Vvv) and mean depth of furrows (metf) load most heavily on the PC1 axis, whereas mean density of furrows (medf), amplitude and complexity (Sdq, Sdr) weigh mostly on PC2. As the two furrow parameters, medf and metf, contribute strongly to only one PC each, a biplot of these two parameters produced thus a similar resolution of dietary groups as the PCA (figure 2a,c) and an even better separation of carnivores and frugivores. The biplot of SSFA parameters Asfc and epLsar shows less separation between dietary groups (figure 2b) as compared to the 3DST parameters metf and medf. The direction-based parameters (Tr1, Tr2, Tr3) are not useful for dietary discrimination in our dataset. Descriptive statistics for all parameters are given in the electronic supplementary material, table S5.

Figure 2.

Plots of DMTA data with convex hulls around the data points for each dietary group. Possible outliers in each plot are represented by filled circles colour-coded according to their dietary group and species abbreviation. (a) Principal component analysis of DMT parameters. (b) Biplot of the SSFA parameters Asfc and epLsar for dataset 1 (up to four scans per specimen). (c) Biplot of the 3DST parameters metf and medf for dataset 1 (up to four scans per specimen). (d) The same as (c) for dataset 2 (all obtained scans). Species abbreviations: choous = Furcifer oustaletti, cycor = Cyclura cornuta, gaaua = Gallotia auaritae, opapo = Pseudopus apodus, sppun = Sphenodon punctatus, tisci = Tiliqua scincoides, tuteg = Tupinambis teguixin, urhar = Uromastyx hardtwickii, uromastyx = Uromastyx sp. (Online version in colour.)

(b). Herbivores

Herbivores (n = 21) like Iguana iguana show medium to low parameter values for all height and volume parameters (Sa, Sq, Sxp, Vmc, Vv, Vvc, Vvv) with a large variability (electronic supplementary material, figure S1). Their enamel surface textures have numerous, shallow furrows as signified by high medf and low metf. Complexity (Asfc) is medium to low.

(c). Algaevores

Algaevores (n = 10), represented by a single species (Amblyrhynchus cristatus), plot in the herbivore dietary space for most roughness parameters (electronic supplementary material, figure S1; figure 2a,c). They are distinguished from herbivores that are more generalistic by Sdr (developed interfacial area ratio, i.e. a roughness measure influenced by all texture features), and larger Asfc values (complexity), though pairwise differences with herbivores are not significant for any parameter.

(d). Frugivores

The only two frugivorous taxa in the dataset, Gallotia auritae (n = 4) and Gallotia galloti (n = 1), plot mostly within the herbivore dietary space (electronic supplementary material, figure S1; figure 2a,c), but are characterized by having the highest density of furrows (medf) and higher complexity (Asfc), as compared to generalistic herbivores.

(e). Omnivores and insectivores

Omnivores (n = 11) like Pogona vitticeps and insectivores (n = 10) like Smaug giganteus are very similar to herbivores (electronic supplementary material, figure S1); but they can be distinguished from herbivores by their higher surface roughness parameters (Sa, Sq, Sxp, Vmc, Vv, Vvc, Vvv) and by larger depth of furrows (metf). Omnivores have significantly higher values in surface roughness (Sa, Smc, Vv) and depth of furrows (metf) than herbivores. There are, however, no significant differences between herbivores and insectivores. Significant differences in parameter values are found between omnivores/insectivores and carnivores as well as molluscivores. Omnivores and insectivores are similar, they overlap in their parameter values and cannot be distinguished from each other.

(f). Carnivores and ovivores (egg-specialists)

Carnivorous taxa (n = 10) like Varanus salvator and ovivorous (egg-feeding) specialists (genus Heloderma, n = 4) are only differentiated by density of furrows (medf), with the ovivores having a significantly lower density of furrows (electronic supplementary material, figure S1). Carnivores fall into the herbivore dietary space in all parameters, except for Str and medf, in which carnivores show significantly lower values than herbivores. Asfc (complexity) is lowest in carnivores and ovivores and significantly lower than in omnivores.

(g). Molluscivores

Molluscivorous species (n = 6) like the adult V. niloticus are well characterized by high and voluminous profiles with large peaks and deep furrows (electronic supplementary material, figure S1; figure 2c). They show significantly higher values compared to the other dietary groups for several roughness parameters (Sa, Sq, Sxp, Sdq, Smc, Vvv, Vmc, Vvv) and mean depth of furrows (metf). Furthermore, their Asfc (complexity) values are higher than all other groups except for algaevores.

(h). Influence of tooth position on dental microwear texture analysis results

For selected specimens, we compared individual results per tooth position along the maxillary tooth row for the two furrow parameters metf and metf to the median values ± s.d., for their assigned dietary group (electronic supplementary material, figure S2). These parameters were chosen as they best differentiate between dietary groups. Variability of parameter values along the tooth row does not exceed the standard deviation within the assigned dietary group in omnivores, carnivores and molluscivores. In the algaevore (ZFMK-84170), tooth positions are highly variable in parameter values. In the frugivorous (ZFMK-Gallotia3), omnivorous (ZFMK-4131) and molluscivorous specimen (SMNS-10707), the more anterior and more posterior tooth positions show distinctly different parameter values compared to the more central tooth positions. In the omnivore (ZFMK-4131, electronic supplementary material, figure S2b), the posterior-most bulbous tooth lies outside of the interquartile range for this individual when all scans are pooled. In the selected herbivore (SMF-33228), a central tooth position (M7) is an extreme outlier compared to the remainder of the dentition.

For individual SMF-33228, random draws of four scans from four different tooth positions were performed 10 times for the parameters metf and metf (electronic supplementary material, figure S3). Median values of the ‘specimens’ created by these random draws all lie within the median values ± s.d. for their assigned dietary group (herbivore).

For the omnivorous ZFMK-4131, 10 random draws resulted in three instances in which medf values for the ‘specimen’ exceeded the median values ± s.d. of their assigned dietary group (electronic supplementary material, figure S3). For the metf parameter values for all ‘specimens’ created by random draws lie outside the dietary space of median values ± s.d. of omnivores, but show very little variation.

(i). Ontogenetic shift in Varanus niloticus

We found distinct differences in parameter values between V. niloticus of different age classes. The juvenile V. niloticus (n = 1) is characterized by low surface roughness values, small wear features and a high density of furrows (medf) (figure 3). Roughness values (Sq) and volume (Vmc) of the surface texture increase in subadult individuals (n = 4) and are highest in the adult specimens (n = 3), whereas density of furrows (medf) is decreasing along the ontogenetic sequence.

Figure 3.

Selected 3DST parameters highlighting the change in diet from insectivory to molluscivory during ontogeny of Varanus niloticus. Boxes with coloured lines indicate the interquartile range of insectivore and molluscivore dietary space determined for all lepidosaur species. Assignment of ontogenetic age is based on relative skull length and dentition. Schematic maxillary shapes: juvenile = smallest skull, all teeth pointed and curved; subadult = larger skull than juvenile, dentition still composed of pointed teeth and some rounded, blunt teeth; adult = largest skull, dentition mainly composed of rounded, blunt teeth. (Online version in colour.)

3. Discussion

(a). Diet-specific dental microwear textures

Based on the enamel surface texture data from 23 extant lepidosaurian taxa we conclude that dietary traits are reflected in tooth wear. The best discrimination between traits is achieved by combining the parameters medf (mean density of furrows) and metf (mean depth of furrows). Dietary specialists, which include harder objects such as molluscs (molluscivores) or seed-containing fruits (frugivores) in their diet, are hereby best characterized and show no or only few overlap in DMTA with other dietary groups (figure 2c,d). There is, however, overlap between omnivores and insectivores. Thus, surface texture parameters alone cannot be used to unambiguously assign the dietary traits omnivory and insectivory to a species. This is, however, not unexpected as omnivores sometimes include larger proportions of insects in their diet [68] resulting in a surface texture signal similar to insectivores and therefore should not be seen as a limitation of the method. Moreover, most lepidosaurs that are classified as insectivores also consume small amounts of plant material (see the electronic supplementary material, table S1). A similar reasoning applies for herbivores. Some taxa, like I. iguana and Uromastyx, are strictly herbivorous [8], whereas others frequently feed on animal matter (e.g. Cyclura cornuta), especially during earlier ontogenetic stages [8,69]. This high plasticity in dietary behaviour could explain the observed partial overlap between insectivores and omnivores. Still, omnivores and insectivores tend to have higher surface roughness parameters compared to herbivores (electronic supplementary material, figure S1), even significantly distinguishable between herbivores and omnivores.

All carnivorous taxa included in this study are either monitor lizards or belong to the genus Heloderma, the latter being more specialized on egg-predation [70]. Monitor lizards have a wide prey spectrum, ranging from insects and annelids to large vertebrates [71,72]. However, they rarely perform prey-reduction behaviour and instead more commonly swallow their prey whole [72,73] (with the exception of Varanus komodoensis, which can deflesh a carcass [74]). Thus, the only interaction between their teeth and the prey is during capture, killing and swallowing. Therefore, unidirectional puncturing of tissue is the only mode of tooth contact expected to cause dental wear, resulting in the observed low roughness, density and depth of surface texture features.

All species in the remaining dietary groups (herbivores, algaevores, frugivores, omnivores, insectivores, molluscivores) seem to display a more pronounced oral food-processing behaviour and particle size reduction than carnivorous or ovivorous species, as reflected in their higher complexity (Asfc) and a higher density of surface texture features (medf).

(b). Influence of tooth position and number of scans

Variability of the surface texture signal along the maxillary tooth row for selected specimens showed that the anterior and larger posterior tooth positions are more prone to display wear-features that represent outliers (electronic supplementary material, figure S2a,b). Variability was lowest in the omnivores, carnivores and molluscivores. Nevertheless, inclusion of all obtained surface texture scans produced a similar number of significant pairwise differences compared to dataset 1 (reduced to up to four scans per specimen, see the electronic supplementary material). This indicates that the number of scans obtained from various tooth positions should be around four, but does not need to be much higher to increase differentiation in dietary space of DMT data in general. If possible, the most anterior and most posterior teeth should be excluded. When focusing only on the best differentiating parameters, mean furrow depth (metf) and density (medf) (figure 2c,d), datasets 1 and 2 performed equally well in distinguishing dietary groups. We further tested how random selection of four scans from four different tooth positions affects parameter values in a herbivorous and an omnivorous specimen (electronic supplementary material, figure S3). All 10 randomly created sub-datasets for each individual fell within the herbivore and omnivore parameter range derived from all individuals in each dietary category, except for the omnivore with respect to mean depth of furrows (metf). However, this specimen is not detected as an outlier in the complete dental microwear texture dataset and does not show large variability in the random draws, thus indicating that four scans from different tooth positions are sufficient to gain a representative dental microwear texture signal. Our results are in accordance with results from Williams et al. [37] for two-dimensional microwear on hadrosaurids, who found no wear differences between sampling sites within a given tooth, and few wear differences along the tooth row. Although even isolated teeth can be used for tooth wear analyses, tooth specimens for dietary analyses in homodont species should be selected carefully. If tooth morphology cannot be employed to assign a tooth to a certain position within the jaw, isolated teeth might potentially stem from the less representative most anterior/posterior positions. Furthermore, in taxa with continuous tooth replacement (including the majority of Lepidosauria, figure 1), newly erupted teeth might exhibit no or biased wear patterns. Such peculiarities, however, will only become apparent when a larger sample is analysed. Owing to this constraint, our extended dataset 2 only comprises 25% more teeth than dataset 1 given that we excluded all ambiguous (i.e. with potentially biased wear-patterns) specimens (about 30% of the scans taken).

(c). Diet-related ontogenetic changes

A distinct dietary shift during ontogeny of V. niloticus is well documented [11,72]. Our surface texture results are consistent with a more insectivorous diet during younger ontogenetic stages and a more durophagous (i.e. molluscivorous) diet in the later ontogenetic stages. We note that a dietary transition can already be observed in subadult individuals, which show intermediate surface texture parameter values between the youngest and the oldest individuals (figure 3). Thus, DMTA is a useful tool for understanding the ontogenetic dietary transition in V. niloticus independent of the obvious change in dental morphology. It is, therefore, likely that DMTA might also be used to assess ontogenetic dietary changes in other reptilian taxa that do not undergo a transformation in tooth morphology (e.g. crocodiliformes).

(d). Limitations of dental microwear texture analysis in Lepidosauria and data outliers

Even though a significant separation of some dietary groups by DMTA is observed (electronic supplementary material, table S4 and figure S1), there are limitations to this study. The number of taxa analysed is limited for some dietary trait groups: algaevores are only represented by one species each, (indeed, algaevory only occurs in one dietary specialist, A. cristatus). Frugivores are only represented by two species belonging to the same genus, but frugivory does also occur among the monitor lizards (Varanus mabitang). However, specimens belonging to this taxon are difficult to access. It would be interesting to test whether the very distinct surface texture patterns which characterize the frugivores in this study (genus: Gallotia) are also exhibited by V. mabitang. Furthermore, for some taxa only one specimen could be obtained for DMTA, thus its surface texture is possibly not fully representative for the diet of the whole species.

In this context, the outliers in figure 2a–d need to be addressed. In the PCA and both metf versus medf biplots, one individual of Sphenodon punctatus plots with the herbivores or frugivores, and not within the insectivores. However, the diet of Sphenodon has been described as highly variable, ranging from insectivorous [75,76] to carnivorous (feeding on eggs and chicks) [77]. Some observations document that Sphenodon includes a substantial proportion of plant material into its diet [8], which may however be incidental [75]. The variability in DMT between different individuals of Sphenodon is thus not surprising. In mammals, DMT is expected to be the result from the last few meals [18,78] an animal consumed and, consequently, a short-term change in food availability will probably impact tooth wear in lepidosaurs, too. The second outlier in both analyses is Tiliqua scincoides. In the surface texture biplot, T. scincoides plots close to molluscivores, but in the literature it is described as an omnivore [79]. T. scincoides has blunt teeth (figure 1) which allow for the incorporation of hard items, such as molluscs, into its diet. In suburban areas, for example, it is known to feed on garden snails [80]. A more molluscivorous diet is in accordance with surface texture results, suggesting that the analysed individual of T. scincoides exhibited a preference for molluscs and/or fed primarily on molluscs prior to death. Finally, the outlier for Tupinambis teguixin (plotting more with herbivores) is also in accordance with its' potential dietary spectrum, as Tupinambis is a flexible omnivore [81,82]. The two herbivores (I. iguana, C. cornuta) plotting higher than other taxa in figure 2c,d showed visibly more pronounced, untypical tooth wear than other specimens. Such outliers strengthen the importance of having a large sample size and treating single individuals with caution.

4. Conclusion

Lepidosauria show ingesta-related DMTs that can be used for characterization of dietary groups, despite their limited oral food processing behaviour, frequent tooth replacement and lack of developed wear facets characterizing mammals. Carnivorous, molluscivorous and frugivorous taxa are significantly distinguishable from other feeding groups, whereas herbivorous, omnivorous and insectivorous taxa partially overlap. An ontogenetic dietary shift from insectivory to molluscivory is detectable for V. niloticus. Surface texture data is indicative of both the material properties of the diet and prey-consumption behaviour (e.g. carnivorous monitor lizards). However, owing to the large dietary flexibility in most omnivorous and insectivorous taxa, inferences of the diets of taxa represented by only a single specimen must be interpreted with caution in DMTA. Variability of the DMT is highest for the most anterior and posterior teeth in the maxillary jaw and thus should preferably be measured on central tooth positions for which no significant inter-tooth variations were detected within an individual. Increasing the overall number of surface texture scans above four per specimen does not improve resolution between dietary groups. We, therefore, suggest that four surface texture scans, preferably from four different tooth positions to provide a representative sample of the dentition, should be performed on the labial side of maxillary teeth. Overall, non-destructive DMTA has great potential for application beyond the traditionally studied mammals to fossil teeth of homodont (reptilian) taxa for palaeodietary tendency estimation of extinct species such as early synapsids and dinosaurs, and for macroecological studies across time and space.

5. Material and methods

(a). Tooth samples

The 77 reptile specimens analysed in this study are curated at the following institutions: Senckenberg Forschungsinstitut und Naturmuseum Frankfurt (Germany) (SMF), Zoologisches Forschungsmuseum Alexander Koenig Bonn (Germany) (ZFMK), Staatliches Museum für Naturkunde Stuttgart (Germany) (SMNS), Muséum National d'Histoire Naturelle Paris (France) (ZA-CA-MNHN), Zentrum Anatomie der Universität zu Köln (Germany) (Köln). Wild-ranging, adult individuals with known provenance were preferably chosen to obtain a diet-related tooth wear signal characteristic of a natural diet. Additionally, some wild caught specimens with unknown provenance collected between 1874 and 1967 were included to increase sample size (electronic supplementary material, table S1). For V. niloticus, three ontogenetic stages (see the electronic supplementary material) were sampled to capture the dietary shift from insectivory to molluscivory, which occurs in this species along with morphological changes in its dentition [11].

(b). Dental microwear texture analysis

We obtained surface texture scans from the labial side of the maxillary teeth of 77 specimens from 23 taxa, representing eight dietary traits (figure 1). Study taxa were assigned to these dietary traits based on observed dietary tendencies and percentage plant-matter consumption as reported in the literature [8] (electronic supplementary material, table S1 and references therein). Only the labial side of upper teeth was used in all lepidosaurian species because the lingual side of upper teeth, as well as the labial side of the lower tooth row, interlock in many faunivorous taxa and show occluding attrition dominated facets, which might obliterate ingesta-related tooth wear [83]. Moreover, the curvature of upper teeth is less pronounced than in lower teeth in most taxa (e.g. A. cristatus in which lower teeth bend strongly inwards), which allows for easier access to the areas of interest and direct DMT measurements on the original tooth surface. High-resolution silicone moulds were made using Provil novo Light C.D. fast set EN ISO 4823, type 3, light (Heraeus Kulzer GmbH, Dormagen, Germany) and used for surface texture scans, for a few specimens, either because the skulls were too large to fit under the lens of the confocal microscope, or because a loan of the material was not possible (compare electronic supplementary material, table S1). Scans were taken from the apex of the tooth progressing upwards. All teeth were carefully cleaned to remove any adherent dust prior to DMTA measurements or moulding, using cotton swabs and ethanol. In case of adherent superglue, teeth were additionally cleaned with acetone.

(c). Surface texture scanning

Surface texture measurements were acquired on the high-resolution confocal disc-scanning surface measuring system μsurf Custom (NanoFocus AG) with a blue LED (470 nm) and high-speed progressive-scan digital camera (984 × 984 pixel), set to a 100× long-distance objective (numerical aperture, 0.8; resolution in x, y = 0.16 µm; step size in z = 0.06 µm), and processed with MountainsMap Premium, v. 7.4.8676, software (DigitalSurf, Besançon, France; www.digitalsurf.com) using the default scanning area as established in mammals (160 × 160 µm). A set of 46 surface texture parameters were quantified using the following analyses: ISO 25178 (roughness), ISO 12871 (flatness, Gaussian filtering, size 0.025 nm), SSFA, motif, furrow, direction and isotropy (electronic supplementary material, table S1). Further detailed information is included in the electronic supplementary material.

(d). Dataset preparation and statistics

We propose that labial teeth from a central position in the maxilla are best suited for DMT measurements, because the central position of such teeth ensures that they are frequently in contact with food items. To assess whether there are any significant inter-tooth differences in wear within dentitions we generated two datasets for statistical analyses:

• (1)selecting four scans (or all available) per specimen (from up to four different tooth positions) from central maxillary teeth (n = 246 teeth, n = 285 scans), and
• (2)including all obtained scans from all maxillary molar and canine-like teeth analysed in each specimen (n = 307 teeth, n = 461 scans).

The same statistical procedure was employed for both datasets. Results for dataset 2 and detailed description of statistics can be found in the electronic supplementary material.

Data accessibility

This article has no additional data.

Competing interests

We declare we have no competing interests.

Funding

The project was funded by the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation programme (ERC CoG grant agreement no. 681450) to T.T.