The way wear goes: phytolith-based wear on the dentine–enamel system in guinea pigs (Cavia porcellus)

Abstract

The effect of phytoliths on tooth wear and function has been contested in studies of animal–plant interactions. For herbivores whose occlusal chewing surface consists of enamel ridges and dentine tissue, the phytoliths might particularly erode the softer dentine, exposing the enamel ridges to different occlusal forces and thus contributing to enamel wear. To test this hypothesis, we fed guinea pigs (Cavia porcellus; n = 36 in six groups) for three weeks exclusively on dry or fresh forage of low (lucerne), moderate (fresh timothy grass) or very high (bamboo leaves) silica content representing corresponding levels of phytoliths. We quantified the effect of these treatments with measurements from micro-computed tomography scans. Tooth height indicated extreme wear due to the bamboo diet that apparently brought maxillary incisors and molars close to the minimum required for functionality. There were negative relationships between a cheek tooth's height and the depth of its dentine basin, corroborating the hypothesis that dentine erosion plays an important role in herbivore tooth wear. In spite of lower body mass, bamboo-fed animals paradoxically had longer cheek tooth rows and larger occlusal surfaces. Because ever-growing teeth can only change in shape from the base upwards, this is a strong indication that failure to compensate for wear by dental height-growth additionally triggered general expansive growth of the tooth bases. The results suggest that enamel wear may intensify after enamel has been exposed due to a faster wear of the surrounding dentine tissue (and not the other way around), and illustrate a surprising plasticity in the reactivity of this rodent's system that adjusts tooth growth to wear.

1. Introduction

By reducing the availability of functional tissue, dental wear threatens the survival of organisms that rely on teeth for food processing, and has a detrimental impact on individual and population fitness [1]. A large number of reports ranging from shrews [2] to primates [3] and elephants [4] suggest that progressed dental wear is associated with reproductive or somatic senescence, and hence limits survival and fitness.

The wear process caused by tooth-on-tooth attrition or tooth–ingesta abrasion ranges somewhere between ‘equal wear on all structures’ and detrimental crack formation resulting in severe loss of individual tissue chips [5]. The abrasive potential of materials like quartz dust, enamel chips and phytoliths is dependent on their mechanical properties, such as hardness and particle size. In particular, the comparative hardness of phytoliths and tooth enamel is a topic of ongoing research and controversy [6–12], with implications also for palaeontology [13] and the reconstruction of diets in hominins and other species [14]. The ability of phytoliths to wear enamel is often questioned in these investigations. Enamel, as the hardest and most mineralized body tissue, features prominently in tooth wear discussions. Yet mammalian teeth, and notably those of most herbivores, are complex composite structures of enamel, dentine and cementum [15] that evolved not only to resist plastic deformation (as well as fracturing), but also to optimize the masticatory comminution of food. This latter function is often achieved by a composite surface of enamel and dentine that facilitates, due to the hardness difference between the two, the establishment of enamel ridges with cutting function. Because phytoliths are harder than dentine [8], they will erode the dentine tissue in between the enamel ridges, changing the chewing forces on these ridges as well as their exposure to abrasion and attrition, and hence indirectly facilitating enamel wear [6]. This hypothesis, to our knowledge, has not been investigated empirically.

A variety of adaptions that ensure functionality in spite of wear validate dental wear as an environmental selective pressure and a factor of phenotype modification. Morphophysiological adaptations to increase the dentition's functional life can roughly be grouped into increased wear resistance (resistance to abrasion and/or crack propagation) or specific strategies to increase the amount of dental tissue [16]. Increased dental durability is achieved by incorporation of harder minerals into the enamel structure [17], alteration of enamel properties [18], adaption of the enamel microstructure in the form of Hunter-Schreger bands [19] or generally increasing enamel thickness [20], or a combination of these strategies. The trade-off with increased wear resistance is higher risk of cracks due to lack of elasticity or the inability to form sharp-edged enamel blades [16].

To increase the amount of available dental tissue, herbivorous mammals either evolved larger teeth, or a form of continuous tooth growth. In teeth with finite growth, the investment in more volume through higher crowns (hypsodonty) or an increase in chewing surface, can, in principle, delay or avoid the negative impact of dental wear [21,22]. Even if growth is terminated, apposition of cementum in the root area may act as partial compensation [23,24]. Hypselodont, aradicular teeth as in lagomorphs and rodents grow continuously [21]. So far, this growth has been considered strictly height-compensatory: experiments with diets of differing abrasiveness [25–28], or with the unilateral clipping of incisors [29,30], indicate that growth is proportional to wear, and accelerates in the absence of occlusion. The signalling pathways responsible for controlling the compensatory growth, which is tooth-specific and can, at the same point in time, differ between incisors and molars [25] or between incisors of the same individual [29], are reliant on pressure signals during occlusion [31]. The response of the hypselodont dental system to complete absence of abrasion, such as during prolonged parenteral or liquid feeding, has not been documented (although Latour et al. [32] did not report any dental problems in rabbits fed a liquid diet for 1 year). To our knowledge, the other extreme—the response to a diet of such excessive abrasion that tooth regrowth cannot compensate—has also never been reported.

Guinea pigs have an exclusively proal chewing motion, where the mandible moves forward, down and laterally to bring the teeth into occlusion [33]. Their molars and molariform premolars are asymmetrically hypselodont, with deep folds dividing the occlusal surface [15]. These enamel folds surrounding the dentine tissue of the antero- and posteroloph show an adapted schmelzmuster in the form of a combination of radial enamel layers and layers of irregular, decussating enamel [19]. The leading edge of the power stroke can be identified due to its missing inner radial layer and shelters the subsequent dentine, resulting in minimum wear of dentine on the leading edge and maximum wear of dentine close to the trailing edge [34]. As in certain other hypsodont herbivores, cementum is not only present at the open root of the molars, but also as cementum pearls on the peripheral surface, as acellular cementum on buccal (lower) or lingual (upper) side of the enamel folds, and as cartilage-like cementum incompletely filling the deep furrows between the enamel folds [35] (see also electronic supplementary material, figure S1). The cartilage-like cementum is mainly thought to supply anchorage points for the periodontal ligaments [36]; its potential role in stabilizing the enamel folds from the outside or any specific involvement in the chewing process has, to our knowledge, not been explored yet.

(a). Hypothesis

In a study designed for measuring 3D enamel surface textures on guinea pig molars when the animals were fed with either dry or fresh forage of different silica contents, differences in 3D texture between the diets were documented [37]. Using micro-computed tomography (mCT) to accurately survey the dentition, our intention was to test the hypothesis that different phytolith concentrations in a natural diet lead to a corresponding macroscopic difference in dental tissue between the feeding groups, similar to previously demonstrated differences in CT-based measurements in guinea pigs on a different set of diets [26]. In particular, we predicted that shorter (i.e. more worn-down) teeth are characterized by deeper (more eroded) dentine basins, indicating that dentine is eroded at a faster rate in the process of wear. Alternatively, if dentine was worn proportionately to enamel, no difference between the two tissues would be expected, and if the rate of enamel wear was higher than that of dentine wear, shallower dentine basins would be expected in shorter teeth. We took different morphometric measures to assess the potential effect of body size on skull and dental measurements, which led to an additional unexpected observation.

2. Material and methods

(a). Feeding experiment

Subadult female Dunkin Hartley (HsdDhl:DH) guinea pigs (n = 36; initial body mass: 263 ± 14 g) were randomly allocated to six groups for a controlled feeding experiment approved by the Cantonal Veterinary Office in Zurich, Switzerland (licence no. ZH135/16). They were floor-housed outdoors as previously described [37]. Vitamin C-supplemented water (200 mg l⁻¹) and food were offered for ad libitum consumption over the course of 21 days. The animals were fed exclusively with a forage diet, consisting either of fresh lucerne (LF, Medicago sativa), timothy grass (GF, Phleum pretense), bamboo leaves (BF, Phyllostachys aureosulcata f. spectabilis) or the corresponding dried hay made of the same batches of forages (LD, GD, BD). Acid detergent insoluble ash (ADIA) content as a proxy for silica was analysed from the diets, and faeces and daily dry matter (DM) intake recorded as previously described [37,38]. Body mass (BM) was recorded every other day and the animals were evaluated daily for clinical abnormalities. At the end of the experiment, the animals were euthanized, and the skulls were macerated. Subsequently, the skulls were submitted to mCT scanning using a Nikon XT H 225 scanner. The resulting images were exported as DICOM files for further analysis.

(b). Dental measurements

The mCT studies were analysed using a multi-planar reconstruction in Horos (v. 3.3.3). Morphometric measurements are explained in electronic supplementary material, figure S2. Measurements of the skull were taken on the axial semi-oblique plane to determine the length of the hard palate, the width between the zygomatic processes and the width of the os nasale. The incisors were measured on a sagittal plane (electronic supplementary material, figure S3) quantifying rostral and caudal height along the curved tracing of the enamel borders as well as the caudal height of the maxillary functional crown. Measurements on the cheek teeth included, on a frontal plane, buccal and lingual height along a curved tracing of the enamel borders (electronic supplementary material, figure S4) and maximal depth of the anterior dentine basin (maximal distance from deepest point in the enamel basin to a plane through the tips of the buccal and lingual enamel ridges; electronic supplementary material, figure S5). Additionally, length of each tooth and the entire tooth row, width and occlusal area (including enamel, dentine and cartilage-like cementum) of each tooth and tooth angle of the upper cheek teeth were measured as shown in electronic supplementary material, figures S6–S8. It should be noted that for each single measurement taken on the 3D reconstructions, the corresponding structure was individually aligned according to standardized procedures (explained in the electronic supplementary material). All mCT scans were evaluated by the same examiner (L.F.M.), who was blind to the diets represented by the respective scans.

(c). Statistics

Morphometric data was analysed using one-way ANOVA with Tukey's post hoc after testing for normality and homoscedasticity of residuals with Shapiro–Wilks's is and Levene's test respectively. For palate length, residuals were not normally distributed, nor when log-transformed, so ranked data was used in the ANOVA, without BM as covariate. The dental measurements were analysed using nested general linear models (GLM) with the diet effect nested within tooth (cheek teeth only, i.e. P4, M1, M2 or M3; no ‘tooth’ term for incisor), and that term within jaw (UR, UL, LR and LL).

The diet effect included two factors: the diet itself (B, L or G) and fresh or dry, as well as the interaction between them. If the interaction was not significant, the model was run without it, and we present only those results. For cases where residuals did not meet parametric assumptions (as above), we log-transformed the data; if it was still not parametric, we used ranks. Body mass was included as a covariate in all models to account for any effects of differences in body size across treatments. All analyses were carried out in R 3.4.2 (R Core Team, 2015).

3. Results

(a). Diets and intake

ADIA content differed in the diets and in the faeces of the six groups (p < 0.001; table 1) corroborating the planned experimental design. Since most plant silica occurs in the form of phytoliths [39], the diets complied with the experimental goal of offering different phytoliths contents in the diet (see also electronic supplementary material, figure S3 in [37]). DM intake for the bamboo groups was more than double that for the control lucerne groups (table 1), indicating that silica intake was highest in animals fed bamboo. Additional analyses revealed no contamination of the feed with external abrasives [37].

Table 1.

Information on diet dry matter (DM) content, diet acid detergent insoluble ash (ADIA), daily (DM) intake and faecal ADIA in the respective guinea pig (Cavia porcellus) dietary groups feeding on lucerne (LF, lucerne fresh; LD, lucerne dry), grass (GF, grass fresh; GD, grass dry) or bamboo (BF, bamboo fresh; BD, bamboo dry) for three weeks.

		lucerne		grass		bamboo
		fresh	dry	fresh	dry	fresh	dry
diet DM content	g kg−1 as fed	166	918	232	930	493	922
diet ADIA	g kg−1 DM	4.7	3.9	5.8	6.7	32.1	32.5
daily DM intake	g	16.2 ± 8.4	14.0 ± 6.3	29.7 ± 7.3	27.4 ± 5.2	32.0 ± 1.8	36.0 ± 6.6
faeces ADIA	% DM	1.01 ± 0.15	0.97 ± 0.18	3.31 ± 0.17	2.87 ± 0.19	5.55 ± 0.27	5.16 ± 0.35

(b). Body mass and morphometry

Body mass at the start of the experiment did not differ between the groups (p = 0.77; electronic supplementary material, table S1). Even though the animals were fed for ad libitum consumption, body mass gain in the two bamboo groups lagged behind the other groups at day 10 as well as at the end of the experiment (p < 0.001; figure 1a; electronic supplementary material, table S1), indicating that in spite of the higher intake, bamboo animals could not maintain normal growth. The body mass did not differ between the L and G diet, nor between fresh and dry forages (p = 0.648 and p = 0.998 resp.; electronic supplementary material, table S1). One-way ANOVA showed no systematic differences between diet groups for distances between the zygomatic arches (p = 0.09; electronic supplementary material, table S1) but the smaller BD animals had shorter hard palates than the LD animals (p = 0.031; figure 1b; electronic supplementary material, table S1), which is expected based on the overall body mass development.

Figure 1.

End body mass of guinea pigs (Cavia porcellus) fed with diets of different phytolith concentrations for three weeks (LF, lucerne fresh; LD, lucerne dry; GF, grass fresh; GD, grass dry; BF, bamboo fresh; BD, bamboo dry) in correlation with body mass at the start of the experiment (a) and length of the hard palate as a proxy for skull size (b). (Online version in colour.)

(c). Dental measurements

On macroscopic inspection of the mCT 3D reconstructions (figure 2) it was possible to assign the animals to their respective forage types (L, G and B), based on a few key characteristics, such as the height and shape of the upper incisors as well as the height of the maxillary cheek teeth. Visually, it appeared that the caudal part of the maxillary incisors, as well as the maxillary cheek teeth, barely protruded from the underlying bony maxillary structures in the B groups.

Figure 2.

Lateral and frontal view of mCT-scan-derived 3D volume skull reconstructions of guinea pigs (Cavia porcellus) fed with plant diets of different phytolith concentrations for three weeks. In the frontal view, the incisors are masked out as well as the zygomatic arch on the lateral view. White scale bar is 1 cm. Note the height of the functional crown in the upper incisors as well as the upper cheek teeth.

Actual measurements confirmed distinctive differences. The rostral height of the incisors differed across diets, and between maxilla and mandible, but not between the dry or fresh forages (p < 0.001, p < 0.01 and p = 0.409 resp.; electronic supplementary material, table S2). On L, the animals had higher functional crowns on the maxillary incisors than on B (p < 0.001; electronic supplementary material, table S2), and also the caudal height from apex to tip of the tooth was greater on G and L than on B (p < 0.001; electronic supplementary material, table S2).

For the cheek teeth, the buccal and lingual height was greater on L and G than on B (p < 0.001 to 0.024; electronic supplementary material, table S7, figure 3a and electronic supplementary material, figure S9a resp.), and the lower cheek teeth were generally longer than the upper ones (p < 0.001, electronic supplementary material, table S7). The depth of the dentine basin was generally deeper on B than on G and L (p < 0.001 to 0.027, except for the upper P4 where G and B did not differ), and there were generally negative relationships between the depth of the dentine basin and tooth height (shown for lingual height in figure 3b; buccal height in electronic supplementary material, figure S9b) on each tooth position, consistent with the idea of differential abrasion that hollows out the dentine basin at a faster rate. The occlusal angle was steeper in some teeth for L and G than on B (P4 and M1, and M2 (grass only), p < 0.001 to 0.042; electronic supplementary material, table S7).

Figure 3.

Lingual height for the cheek teeth row (a) of guinea pigs (Cavia porcellus) fed with diets of different phytolith concentrations for three weeks. Abbreviations as in figure 1. The premolars are generally longer due to the widening of the mouth gape (cf. figure 2). Note the significantly shorter cheek teeth for guinea pigs fed with bamboo (electronic supplementary material, table S7). Dentine basin depth in relation to the lingual height of each individual tooth (b). Note the steady relationship between shorter teeth and lower dentine basin for each tooth position. (Online version in colour.)

The length of the upper tooth rows was longest for the BD animals (UL p < 0.01, UR p < 0.001; figure 4a; electronic supplementary material, table S1) even though they were the smallest group regarding body mass (electronic supplementary material, table S1). On B, the occlusal area was significantly larger for all jaws and teeth than on G or L (p < 0.001; figure 4b), resulting in a (paradoxical) negative relationship between body mass and occlusal area in this dataset.

Figure 4.

Length of the upper right cheek tooth row (a) and the post-canine occlusal area (b) of guinea pigs (Cavia porcellus) fed with diets of different phytolith concentration for 21 days in correlation with body mass at the end of the experiment. Abbreviations as in figure 1. Note that the guinea pigs fed with bamboo were significantly lighter but had longer tooth rows and a larger post-canine occlusal area (PCOA) than the animals on lucerne or grass. (Online version in colour.)

4. Discussion

The present study demonstrates a clear effect of a plant-derived silica-rich diet during dental wear and indicates an intuitive sequence where dentine wears ‘first’ (i.e. at a faster rate than enamel). As previously described, hypselodont tooth growth does not completely compensate for wear, leading to measurable differences in tooth height among diets [25–27]. The feeding of bamboo to the guinea pigs had a negative impact on their body mass and, to some extent, on the corresponding development of skull width (figure 1). Bamboo-induced wear reduced the height of both upper molars and incisors to what could be considered—in live animals—close to the gum line (and hence functionality). Unexpectedly, in spite of the lower body mass, guinea pigs fed on bamboo displayed a cheek tooth row that was not only longer but had a notable larger occlusal area than the animals on the other diets (figure 4). It is difficult to interpret this in any other way than an attempt to increase the dental base tissue in a situation of extreme abrasion, where simple increase in tooth growth rate alone does not achieve adequate compensation. This case of phenotypic plasticity, in animals all deriving from the same batch of genetic background, is remarkable because of the short duration of the experiment and the selective increase in dental area, in spite of the concomitant decrease in body mass and absence of other indications of increased skull growth.

Silica deposits in plants occur mainly as phytoliths [39] and represent one of the mechanical defence mechanisms of plants [40,41], but whether they actually wear down teeth or only mimic or enhance the abrasive effect of dust or grit has been a matter of ongoing discussions [5,7,12,13,42]. These have mainly focused on the physical hardness of phytoliths in comparison with dental structures [6–11]. In practice, without the possibility to grow sufficient amounts of the same plant species with differing silica levels, experimenters working on tooth wear have to make a decision. They can either only vary abrasive (silica) content in artificial diets, eliminating differences due to natural forage structure, but also preventing fully natural chewing behaviour; or they can choose different forage species of different silica content, assuring natural chewing behaviour, but incurring putative uncontrolled effects of structural differences between the forage species beyond sheer phytolith concentrations. In vivo experiments with a phytolith-rich diet consisting of grass and rice hull pellets have shown increased dental wear in guinea pigs on this diet compared with phytolith-poorer diets [26]. Similar results were found on a microscopic scale for 3D enamel surface texture measurement of rabbits on a phytolith-rich pelleted diet, where the surface relief in the enamel showed constant overwriting of wear features on the enamel ridges [43]. Providing a grit-free diet consisting of natural forages, our feeding experiment could establish a distinct wear pattern based on phytolith concentrations of the individual diets (figure 3a), but the findings of the present study must be considered with differences in the structure between the forages in mind.

The bamboo diet induced significantly more wear on the teeth and concurrently changed the occlusal morphology: the occlusal angle was flatter on the rostral teeth than on the caudal ones, and the dentine basin was deepest on the first cheek teeth (electronic supplementary material, table S7). Certain differences in dental morphology along the cheek tooth row can be due to skull anatomy. In guinea pigs, the distance between the maxilla and the mandible increases rostrally, requiring rostral cheek teeth to be higher than caudal ones to achieve occlusion (cf. figure 2). In large herbivores, feeding on grass contaminated with dust and soil, the rostral premolar is at greater risk to encounter external abrasives before they are homogeneously mixed throughout the ingesta bolus [44]. When fed only with endogenous abrasives, as in the present study, the wear gradient probably develops as a result of the force distribution during the chewing cycle. In animals with a lateral chewing stroke, rostral cheek teeth may experience a greater lateral deflection and therefore more wear than the posterior ones [45]. In animals with a proal chewing motion, such as guinea pigs, other reasons for cheek teeth wear gradients need to be invoked; a potential candidate could be systematic intra-oral differences in the ingesta load on the different tooth positions. The increased DM intake of the bamboo group (table 1), requiring more chewing strokes for processing, probably exacerbated the abrasive potential of phytoliths.

Many dental wear studies focus only on the enamel structures in herbivores. Surface texture analysis specifically singles out enamel ridges to perform 3D measurements, mesowear scoring looks at the relief built by the buccal enamel surfaces, and finally macrowear measurements focus on height or volume measurements of enamel structures. Far fewer studies have looked at the properties of dentine and its reaction to the wear process [8,9,11,24]. Protected by the enamel ridges, in particular the leading edge, dentine is mainly exposed to tooth–food mediated abrasion [24], although evidence from stillborn guinea pigs show that dentine basins can at certain times be worn in the absence of food [34]. Kaiser et al. [8] showed with hardness tests that phytoliths are harder than dentine, and therefore are more likely to cause wear of dentine than of enamel. Following the assumption that dentine stabilizes the surrounding enamel ridges, we advocate that the deepening of the dentine basin is not regulated by the height of the enamel ridges, but by the abrasiveness of the ingesta. When the phytoliths wear down the dentine basin close to the trailing edge (electronic supplementary material, figure S6), further chewing could increase attrition on the opposing enamel ridges as they are more exposed to each other, and a decrease of the stabilizing effect of the dentine could make the enamel subject to increased wear during the chewing process. Ever-growing teeth allow us to take a snapshot of a dynamic system where differences in the enamel–dentine equilibrium (the basin depth) emerge between different diets. We suggest that the principle of dentine basin deepening and subsequently exacerbated wear of the enamel ridges should occur throughout all hypsodont herbivores. In animals without ever-growing teeth, however, this sequence will result in a constant equilibrium across diets, so that no systematic variation in dentine basin depth between populations can be measured [24]. Whether the cartilage-like cementum between the trailing edge of the anteroloph and the leading edge of the posteroloph is subjected to the same effect could be subject for further research, as well as its function apart from anchoring the periodontal ligaments.

As an additional finding, the bamboo groups showed significantly longer tooth rows and a larger post-canine occlusal area (PCOA) than the grass and lucerne groups at the end of the feeding period (figure 4b). The PCOA was proposed by Gould [46] as the summation of mesio-distal length times the buccolingual width. This simple measurement is an approximation of the complete chewing area that can be taken with callipers. A standard DICOM viewer can nowadays outline and calculate the exact occlusal area in just as little time. In interspecific models, the size of the PCOA was proposed to scale with species' body masses either metabolically [46] or isometrically [47], but various factors, including chewing rate, tooth wear and gender, may influence this relationship [48]. Relatively little research has been conducted on intra-specific allometry [49], and the potential for a flexible PCOA to emerge in animals with hypselodont dentition has, to our knowledge, not been mentioned so far. Due to the open apex and the ever-growing root in guinea pigs, the cross-section area cannot increase if a tooth is worn down unless more tissue is added at the base of the tooth during adaptive growth, resulting in longitudinal as well as horizontal growth. Assuming that guinea pig cheek teeth grow at a similar rate as their incisors, at about 1.36 mm per week on grass [26], an adaptive broadening of the tooth base should be visible in sagittal (electronic supplementary material, figure S6) or frontal cross-section (electronic supplementary material, figure S4) through a tooth of the animals of our experiment (that lasted three weeks). However, the guinea pig's teeth expand, during their constant growth, from a narrow apex to their final size, and any definition of the finally attained ‘base’ dimension (apart from the occlusal area itself) would appear arbitrary. Therefore, such an assessment is not possible. This growth form—starting from a narrower apex and first expanding up to a ‘final base dimension’—means that one would expect teeth worn down beyond that ‘base’ to be smaller, not larger, in their occlusal area, as seen in the bamboo diet. Further studies would have to determine if the increase in chewing area affects enamel, dentine and cementum equally, whether it is a response to excessive dental wear and the corresponding absence of an occlusal pressure signal, or an attempt to facilitate an increased caloric intake if metabolic needs are not properly met. Similarly, it remains to be shown to what degree the PCOA enlargement is in itself plastic—whether it would reduce again on a different diet, or maintain its size even after a diet change. These hypotheses could have broad implications for stem cell biology: studies of the signalling pathways in the stem cells at the base of hypselodont teeth (mainly mouse incisors) have provided insights into the mechanisms controlling stem cell maintenance, differentiation and tissue proliferation [50]. If our results can be repeated, the stem cells of hypselodont cheek teeth could be additionally studied for the potential of tissue re-patterning.

5. Conclusion

In the present study, a phytolith-rich natural diet abraded teeth of guinea pigs by deepening the dentine basins and causing shortening of the enamel ridges to an extent that could be considered close to compromising functionality. The compensatory response of the hypselodont teeth included not only a presumed longitudinal but also a horizontal growth of dental tissue, showing an additional adaptation mechanism to counter excessive tooth wear in these animals.

Ethics

Experimental procedures were approved by the Cantonal Veterinary Office in Zurich, Switzerland (licence no. ZH135/16).

Data accessibility

The electronic supplementary material includes additional figures and tables offering additional information on methods and summarizing the results of this study.

Authors' contributions

T.T. and M.C. designed the experiments with advice from J.-M.H. and D.C.; D.W., A.D.C. and M.C. performed the feeding experiment; L.F.M. performed measurements and analysed the data with assistance from D.C.; A.D.C. facilitated the nutrient analyses; L.F.M. and M.C. wrote the manuscript with input from the other authors.

Competing interests

The authors declare no competing interests.

Funding

This study was funded by the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation program (ERC CoG grant agreement no 681450) to T.T. The work of L.F.M. was funded by the Candoc Forschungskredit of the University of Zurich.