Introduction

The reconstruction of tooth growth is essential to understanding the biology and palaeobiology of mammals1,2, as dental development is closely related with a species’ life history2,3,4,5,6. For instance, the eruption of the first permanent molar correlates well with weaning4,7. Similarly, the emergence of the third molar correlates with skeletal maturity2. Thus, the estimation of rate and duration of molar growth in extant and extinct vertebrates yields key information about their pace of life1,2,3,8,9,10,11. Furthermore, an understanding of tooth growth is crucial for palaeoecological and palaeobiological studies that involve the analysis of stable isotopes12,13. Tooth crowns preserve a temporal record of isotopic variation that can be related to changes in climatic conditions and/or modifications of an animal’s behaviour14,15,16,17,18. Therefore, the rate and duration of tooth growth must be precisely known to accurately develop isotopic sampling methods19 and correctly interpret the isotopic data obtained in this type of investigation12,13.

The pace of growth and development of teeth is recorded, among others, in dental enamel20. From the cusp to the root, enamel is rhythmically deposited by enamel-forming cells called ameloblasts in the amelogenesis process, which involves a first stage of enamel secretion followed by a second phase of enamel maturation21,22. As a result, the histological microstructure of dental tissue registers the pattern of enamel growth in the form of incremental markings20,23. Incremental features have traditionally been classified as short- or long-period marks24. The first ones include cross-striations and laminations and represent a circadian rate of enamel formation23,25,26. Retzius lines, on the other hand, are long-period lines that indicate the successive positions of the developing enamel front20,23. Counts and measurements of incremental markings in enamel provide the basis for quantifying tooth growth, estimating various dental growth parameters such as daily secretion rate or extension rate, and for calculating crown formation time1.

In large herbivorous mammals, studies aimed at reconstructing tooth growth through the analysis of enamel incremental markings have increased considerably in number over the last years3,8,9,10,11,19,27,28,29. However, an analysis of life history parameters from the enamel microstructure in key groups of evolution such as equids30 is still lacking. Only the work of Hoppe et al.13 provided some data about the periodicity and disposition of incremental lines in the enamel of the domestic horse, but recently their results have been questioned by other authors28,31. Indeed, the development of equid teeth has hitherto been determined from radiographic observations13,32,33,34 or by measurements of the crown height12,35, but it has not yet been studied using dental histology.

In the present study, we aim to analyse the enamel microstructure of several wild equid species to provide information about the dental growth pattern and development in this mammalian group. Nowadays, the genus Equus comprises the wild extant species of zebras (E. zebra, E. grevyi, E. quagga), asses (E. africanus, E. kiang, E. hemionus) and horses (E. ferus) that dwell in different areas of Africa and Asia36,37. Here, we quantified incremental markings and enamel growth parameters in the Asiatic wild ass (E. hemionus), plains zebra (E. quagga) and Grevy’s zebra (E. grevyi). These three taxa are the most appropriate ones to infer the dental growth pattern of the clade, as they cover most of the range of habitat, body mass and life history traits observed in extant wild equids38,39,40,41,42. To apply our results to the equid fossil record, we examined the dental enamel of Pleistocene fossil specimens of E. ferus and E. hydruntinus to infer their pattern of molar growth. These two different-sized equids are frequent in European Late Pleistocene mammal assemblages43. Because body size is a fundamental life history trait that tightly correlates with other biological traits such as growth rate44,45,46, we calculated their molar growth rate as a proxy of their overall growth rate to understand whether the differences in body size resulted from changes in life history.

Results

Dental histology of extant Equus

Enamel, dentine and cementum are the three dental tissues that conform the molar crown in all extant Equus (Figs 1a,b; 2b). A thin layer of enamel is also observable in the region that is macroscopically considered as the root13,32,47 (Fig. 2). Therefore, the molar crown, understood as the part of the tooth composed of enamel22, extends further than macroscopically considered47,48 (Fig. 2). To avoid confusion between the morphological and histological distinction of tooth root in Equus first molars, the term root will be used here to designate the most apical area of the tooth which is enamel-free49 (Fig. 2b).

Figure 1
figure 1

Dental histology of Equus and methodologies used to analyse the pattern of enamel growth. (a) Longitudinal section of Equus hemionus’ first lower molar (IPS83151) mounted on two different slides. White dashed rectangle indicates the magnified area in b. (b) Methodology employed to calculate the crown formation time (CFT; IPS83151). The distance between incremental lines (white dashed lines) is measured following the course of enamel prisms (black lines) and divided by the daily secretion rate to determine the time required to form a specific portion of the enamel dentine junction (EDJ; white doubled arrow). (c) Laminations (black lines) and enamel prisms (black arrows) identified in the enamel of E. hemionus (IPS92347). (d) Laminations (black lines) between consecutive Retzius lines (black dashed lines) in the outer enamel of E. grevyi (IPS84963). (e) Angle between the enamel dentine junction (EDJ; white line) and the enamel formation front (EFF; white dashed line) in E. quagga (IPS92346). C = cementum; D = dentine; E = enamel. Black scale bars: 5 mm; white scale bars = 50 μm.

Full size image
Figure 2
figure 2

Macroscopic (a) versus histological (b) distinction of tooth crown and root. Figure shows that dental enamel extends beyond the limit of the macroscopic crown (black dashed line). (a) First lower molar of E. quagga (IPS92346) in buccal view. (b) Histological section of the buccal cusp in IPS92346. C = cementum; D = dentine; E = enamel.

Full size image

Laminations are fine incremental markings running parallel to the enamel formation front20 and are the most common incremental features identified in the enamel of our equid sample (Fig. 1c). Considering that laminations follow the one-day periodicity previously described in other herbivorous mammals3,8,26,27,28, we calculated the crown formation time (CFT) of unworn teeth. As it is shown in Table 1, CFT estimations are a reasonable match for the age of the specimen previously determined by classical methods, confirming the daily periodicity of laminations in equids.

Table 1 Age and crown formation time (CFT) estimates in unworn teeth of extant species.
Full size table

Long-termed Retzius lines, prominent lines formed at an oblique angle to the enamel prisms23, are also observed in the enamel of extant equids (Fig. 1d), mainly in cervical enamel. The periodicity of these features in our equid sample range from 5 to 7 days in E. hemionus and 5 to 6 days in both zebra species (E. grevyi and E. quagga) (RI, Table 2).

Table 2 Body mass, hypsodonty index and values of the different enamel growth parameters inferred in extant and extinct equids. BM = body mass; HI = hypsodonty index; DSR = daily secretion rate; RI = repeat interval; EER = enamel extension rate; EFFa = enamel formation front angle; N = number of observations; SD = standard deviation; Min = minimum; Max = maximum. The star (*) indicates fossil specimens.
Full size table

A mean daily secretion rate of ≈17–18 μm/d was calculated in the enamel of extant equids (DSR, Table 2) regardless of the part of the crown (occlusal, cervical) or the enamel zone (inner, outer) analysed (Kruskal-Wallis, p-value > 0.05, Supplementary Table S1, S2). The Asiatic wild ass shows the slowest daily secretion rate within the extant species, while Grevy’s zebra presents the fastest rate (Table 2, Fig. 3). However, no significant differences have been observed among the species analysed (Kruskal-Wallis, p-value > 0.05, Supplementary Table S3) (Fig. 3).

Figure 3
figure 3

Boxplot of the daily secretion rate (DSR) of the enamel in the extant and extinct species of Equus analysed.

Full size image

The pace of growth of the first lower molars of extant equids (crown height against crown formation time) is plotted in Fig. 4a and b. As equid teeth start to wear before the crown is completely formed32,49, crown development is reconstructed from the cusp tip in unworn teeth (Fig. 4a) and from the root in worn molars (Fig. 4b). Differences in growth between species have only been identified in the most cervical enamel (Fig. 4b), concretely, at the beginning of the morphological root13,32,47 (Fig. 2a). As Fig. 4b shows, E. hemionus deposits enamel in that part of the tooth for a longer period than both zebras. In all studied species, unworn teeth without roots experienced fast linear growth (Fig. 4a) while growth curves of worn molars fit to a polynomial of quadratic order (Fig. 4b). These results indicate that teeth experience different types and rates of growth during formation. Furthermore, the inflection point of the polynomial growth curve (Fig. 4b) is related to macroscopic anatomical changes in the tooth, as it matches the crown divergence that results in the formation of the morphological roots (Fig. 2a). Enamel, then, is still being deposited after the morphological roots are formed, but at much lower rates (Fig. 4b,e).

Figure 4
figure 4

Crown formation time (CFT) related to crown height (a,b) and enamel extension rates (EER) (c,d,e) in the extant Equus species studied. (a) CFT related to crown height in unworn teeth. (b) CFT related to crown height in worn teeth. Dashed line separates measures obtained for the enamel that is deposited before (upward) and after (downward) the crown divergence. (c) EER of unworn teeth studied. (d) EER in macroscopic crowns of worn teeth. (e) EER of the enamel deposited beyond the limit of the macroscopic crown. Legend for scatter plots is shown at the bottom of the figure. 75% grey boxplot = E. quagga; 25% grey boxplot = E. grevyi; white boxplot = E. hemionus.

Full size image

Three different crown developmental stages (CDS) that significantly differ in enamel extension rate (Kruskal-Wallis, p-value < 0.05, Supplementary Table S4) could be established based on the different growth patterns (Fig. 4). During a first stage of fast linear growth (Fig. 4a) enamel extends at a mean rate of 350–400 μm/d (Fig. 4c, Table 2). At the second stage of development, which corresponds to the fastest period of polynomial growth (Fig. 4b), enamel extension progressively decreases up to rates of ≈130 μm/d (Fig. 4d, Table 2). At the third stage, enamel extension rate (EER) decreases quickly to ≈30 μm/d (Fig. 4e, Table 2), representing the slowest period of polynomial growth (Fig. 4b). The decrease in the rate of enamel extension observed throughout tooth formation is mainly due to changes in the enamel formation front angle (EFFa), which presents a range of values that vary from ≈1° on the tooth cusp to ≈11° near the root (Table 2). The EFFa reflects the number of ameloblasts that are secreting matrix at the same time, with smaller angles indicating higher EERs because more ameloblasts are activated along the EEF3,50. Thus, the number of activated ameloblasts in the first lower molars of extant equids is progressively reduced during the development of their crowns.

No significant differences have been found in the rate of enamel extension between the different species of extant Equus analysed within each CDS (Kruskal-Wallis, p-value > 0.05, Supplementary Table S5). This suggests that EER is conservative and characteristic for each developmental stage in extant Equus. On that basis, we reconstructed the complete growth of the crown for the first lower molars of the Asiatic wild ass (Fig. 5a), establishing the overlap area between unworn and worn teeth where similarities in EER were found (Fig. 5b). The timing of several ontogenetic and structural changes of the first molar, such as the time of emergence, eruption and crown divergence obtained from dental histology (Fig. 5a), agrees well with data in the literature on timing of occurrence of these events in equid’s first molars32,51,52. This confirms the validity of the growth reconstruction based on EER for this species and indicates that the complete crown of the first lower molar of E. hemionus takes about three years to be formed. Changes in curve’s slope coinciding with teeth eruption and crown divergence indicate the start of the different CDS (Fig. 5a).

Figure 5
figure 5

Reconstruction of growth of the first molar crown in Equus hemionus based on similarities in EER. (a) Crown formation time against crown height. Three different crown developmental stages (CDS) are identified from the slope of the curve. (b) Teeth of E. hemionus that differ on degree of wear and root development showing the correspondences of EER (black dashed lines) that enabled growth reconstruction (Stage I = IPS83155; Stage II = IPS92347; Stage III = IPS92339). Em = emergence; Er = eruption; Cd = crown divergence. Colour legend is shown at the bottom of the figure.

Full size image

We performed superimposition of teeth considering similarities in EER to infer data about wear rate in the species. As shown in Fig. 5b, the cervical portion of a recently erupted tooth from a one-year old individual (IPS83155) presents similar extension rates as the occlusal part of the crown of a seven-year-old tooth (IPS92347), while EER values at the mid-crown of IPS92347 match those estimated on the occlusal enamel of a thirteen-year-old individual (IPS92339). Superimposition of teeth indicates that approximately 110 mm of tooth crown has been worn down in 12 years: almost 80 mm during the first six years of life and around 30 mm during the next six (Fig. 5). Thus, a wear rate of around 13 mm/y is estimated for the first six-year period, while this is 5 mm/y for the second one. These results indicate that most of the crown formed during the first year of life (IPS83155, Fig. 5b) is worn away by when the individual is seven years old (IPS92347, Fig. 5b).

Dental histology of fossil Equus

As in extant Equus, both laminations and Retzius lines are identified in the enamel of the Pleistocene fossils studied. Long-termed Retzius lines present a periodicity of 5 to 7 days in E. ferus, while they are deposited every 4 to 6 days in E. hydruntinus (Table 2).

The fossil species analysed (E. ferus and E. hydruntinus) secrete enamel at similar rates (Kruskal-Wallis, p-value > 0.05, Supplementary Table S3) (Fig. 3, Table 2). When compared with extant species, however, differences in daily secretion rate have been found between E. hydruntinus, E. quagga and E. hemionus (Fig. 3) (Kruskal-Wallis, p-value < 0.05, Supplementary Table S3). The small extinct species E. hydruntinus secretes enamel a mean rate of ≈19 μm/d, while the Plains zebra and the Asiatic wild ass do at ≈17 μm/d (Fig. 3, Table 2).

We analysed the rate of enamel extension of each fossil tooth based on its macroscopic appearance and the CDS established for extant equids (Fig. 6). The EER of IPS87523 (E. hydruntinus Stage I, Table 2) matches the first stage of crown development (Kruskal-Wallis, p-value > 0.05, Supplementary Table S6) (Fig. 6) and its macroscopic appearance (slight wear and no presence of morphological roots, Table 3) is also that expected for this developmental stage. The rate of enamel extension of IPS87540, IPS87497 and IPS87509 (E. hydruntinus and E. ferus Stage II, Table 2) corresponds to that estimated for the second phase of crown formation (Kruskal-Wallis, p-value > 0.05, Supplementary Table S3). As these teeth show moderate wear and macroscopic roots (Table 3), a stage II of development was expected for their crowns too. These results on EER of fossil Equus confirm that this dental growth parameter is very conservative within the genus.

Figure 6
figure 6

Enamel extension rate of Pleistocene Equus specimens compared with the mean enamel extension rate of the three crown developmental stages established for extant species (grey boxplots).

Full size image
Table 3 Sample studied. E: erupted; NE: no erupted; mo.: months; y: years; ZIHU: Zoological Institute of Hamburg University (Hamburg, Germany); ICP: Catalan Institute of Paleontology (Barcelona, Spain). The star (*) indicates fossil species.
Full size table

Discussion

The aim of the present research was to reconstruct, for the first time, the pace of growth and development of the first lower molars in Equus based on the analysis of enamel microstructure. Until now, most of the research concerning enamel histology has been limited to low-crowned, brachydont teeth of different primate species1 and only a few studies have focused on hypsodont mammals3,8,28,53. Our findings in three species of extant (E. hemionus, E. quagga and E. grevyi) and two Pleistocene (E. ferus and E. hydruntinus) equids contribute to the knowledge of enamel microstructure in high-crowned vertebrates and provide the basis for future research in Perissodactyls. However, the results obtained here are not only relevant for histological research but also for isotopic studies on fossil and archaeological vertebrates. Equus’ teeth, concretely equid enamel, are widely used in palaeobiological and palaeoecological isotopic studies15,17,18 because their extremely high crowns record several years of the individual’s life13. A thorough understanding of the timing, geometry and rate of enamel maturation is key to interpretation of isotopic results54. Although enamel incremental marks do not provide relevant information about the timing26 and/or the pattern54 of enamel mineralization because they only register the secretory stage of amelogenesis20, these features are known to yield accurate estimations on tooth growth rates19. Indeed, our results on enamel extension rates (≈130 μm/d or ≈48 mm/y for the second CDS, Table 2) are similar to the maturation rates reported for equids (40–60 mm/y)55. Thus, rates of enamel extension seem to be a good proxy of rates of enamel maturation in Equus. This finding, along with the detailed description of the growth and development of the equid crown presented here, will help in the understanding of the isotopic microsamples extracted from this mammalian group12,13.

Crown height and formation time

To date, the timing and pace of crown formation in equid teeth has been assessed by measuring crown height12,35 and by identification of dental roots in radiographic images13,32,33,34. In agreement with previous studies47,48, our results indicate that the external division of teeth into crown and root does not match the histological definition of these structures (Fig. 2), as dental enamel extends over the limit of what is usually considered the macroscopic crown of the tooth47 (Fig. 2b). As enamel is characteristic of a tooth’s crown22, we suggest that this portion of the tooth should be considered part of the crown and not part of root; this view is in contrast to a previous study by Kirkland et al.32. Apart from this terminological issue, it is worth noting that the presence of enamel beyond the crown divergence might affect results of previous studies which considered this anatomical point to be the crown’s end. Thus, reconstructions of crown growth that exclude its final portion12,13 may underestimate total crown formation time, as this cervical enamel takes up to one or two years to be formed in zebras and E. hemionus, respectively (Fig. 4b), due to low rates of enamel extension (Fig. 4e, Table 2). On the other hand, the discrepancy between macro- and microanatomical tooth structure (Fig. 2) may lead to a miscalculation of dental indexes that involve measurements of the complete crown. For instance, the hypsodonty index usually used in palaeoecological studies47,56 requires the identification of the crown end to measure the crown height57. Consequently, difficulty in differentiating the exact point where enamel ends from tooth macroanatomy47 (Fig. 2) makes the estimation of the degree of hypsodonty in Equus a challenging issue. Thus, hypsodonty values previously reported for this genus57,58,59 should be viewed with caution because the crown-root transition might be not homologous when measuring crown height. Furthermore, the enamel identified in the morphological roots represents up to 1–1.5 cm of tooth height (Fig. 4b). Failure to consider this area when measuring crown height leads to underestimation of the hypsodonty index.

An alternative approach to estimating hypsodonty in equids can be made from enamel histology. Our results indicate a longer period of crown formation in E. hemionus than in both African zebra species (Fig. 4b), suggesting a higher hypsodonty in the former species. However, previous studies described similar hypsodonty values for these three equids57,58,59. The increase in crown height that characterises hypsodont teeth has usually been explained as an adaptation which extends the durability of the teeth in animals feeding on an abrasive diet56,57. Like all extant equids, E. hemionus, E. quagga and E. grevyi are classified as grazers38,39,40,41,60 with similar abrasive diets60. However, they dwell in different habitats, which also influence tooth wear61. While the Asiatic wild ass is endemic on the steppe and desert plains of central Asia38,40, E. quagga and E. grevyi usually occur in different types of African grasslands38,41. Our findings of an extended deposition of enamel in the Asiatic wild ass (Fig. 4b) is congruent with previous research that correlates hypsodonty and mean annual precipitation60,61, as this species dwells in a more arid habitat. Furthermore, hypsodonty has been explained to occur in response to an increase in lifespan of the species56,62,63,64. A higher degree of hypsodonty in E. hemionus (Fig. 4b) is consistent with this hypothesis, as the maximum longevity reported for this species in the wild is 29 years65, while it is 18 and 21 years for Grevy’s zebra66 and plains zebra51, respectively. Theory suggests that in a resource-limited environment with low extrinsic mortality, natural selection favours a shift in energy allocation from reproduction to growth and maintenance67, and that triggers extended longevity62,64,68,69. The Asiatic wild Ass inhabits a resource-poor environment (steppes and deserts of Asia40) and faces low rates of predation (grey wolves are its only known non-human predators70). Thus, the increase in crown height observed in E. hemionus can alternatively be explained as an increase in durability of its teeth to extend life span. The extended deposition of enamel in the Asiatic wild Ass has been identified after the crown divergence, at the beginning of the macroscopic root (Fig. 4b). Although, as far as we know, there are no descriptions of equid molars which are worn below the crown bifurcation, teeth with such exceptional degree of wear could still be functional. In very old horses, teeth are known to present minimal reserve crowns and very elongated roots that allow stable alveolar attachment49. Thus, a hypothetical equid molar worn below the crown divergence would be viable. This also supports the hypothesis of a prolonged enamel deposition in E. hemionus related to the extended longevity of the species, as the presence of enamel beyond the crown bifurcation increases the durability of the tooth in animals of advanced ontogenetic stages.

Crown developmental stages (CDS)

Our results indicate that the rate of crown formation exponentially decreases throughout tooth development (Figs 4a,b; 5), as already suggested by Bendrey et al.12. The EER decreases from values of ≈350–400 μm/d at the beginning of crown development to rates of ≈30 μm/d at the end of crown formation (Fig. 4c–e, Table 2). A reduction on EER during crown development has also been observed in other mammalian species1,8,10,28,31. Based on such variation of EER along the tooth’s crown, we determined three different developmental stages of the crown (CDS) in the first lower molars of extant Equus (Fig. 4, Table 2). During a first phase of crown formation, dental enamel grows fast and linearly (Fig. 4a,c, Table 2). The next stages of enamel development, however, follow polynomial growth (Fig. 4b). Thus, the second CDS, which presents extension rates values of around ≈130 μm/d (Fig. 4d, Table 2), corresponds to the fastest period of this polynomial growth (Fig. 4b), while CDS III represents the slowest period (Fig. 4b) with very low rates of enamel extension (Fig. 4e, Table 2). The limit between the different CDS seems to correlate with both structural and ontogenetic changes in the tooth in E. hemionus (Fig. 5a). Therefore, the transition from stage I to stage II matches the first molar’s eruption time52, while the beginning of stage III correlates well with the macroscopic appearance of the tooth roots in related Equus species32 (Fig. 5a). Correspondence between the beginning of CDS III and the divergence of the crown has also been detected in E. quagga and E. grevyi (Fig. 4b). Our results also show that the rate of enamel extension is very conservative within the genus Equus and, as previously mentioned, characteristic for each CDS (Figs 4c–e; 6). This result is especially relevant for palaeodemographic studies, as EER estimations of a fragmentary fossil/archaeological equid tooth allow researchers to assigned it to a CDS and, hence, to an age category. Studies aimed to infer the life history strategy of extinct species, however, require combined analysis of the macroscopic appearance of the tooth and the microscopic examination of its EER. Thus, in the case of the estimated CDS of a fossil tooth failing to match the degree of dental wear and/or root development expected for this stage, differences in growth and, thus, in life history might be deduced for the species.

Estimation of wear rates

Superimposition of teeth of E. hemionus based on similarities of EER (Fig. 5) also yield data on wear rates in Equus. A wear rate of around 13 mm/y has been estimated for the first six years of life, a value far above the 3–5 mm/y previously reported in extant49,71 and in extinct equids72. However, the latter published wear rates agree with the estimated wear rate of 5 mm/y obtained for the next six-year period. Therefore, our results suggest that (i) first lower molars of the Asiatic wild ass wear down at an exponentially decreasing rate and that (ii) wear rates are much higher at the beginning of the animal’s life, as already proposed by Levine35. This is explained by the eruption sequence of the species. As the first molar is the first permanent tooth to erupt, and complete permanent dentition is not visible until almost the fifth year of life52, it seems reasonable that higher wear rates in the first molar would be found at early ontogenetic stages. Due to these greater wear rates, the crown formed during the first year of life is almost worn down in a seven-year-old tooth (Fig. 5b). This observation is of special interest for the correct interpretation of isotopic values obtained from enamel microsampling in equids.

Daily Secretion Rate (DSR)

The DSR of enamel in equids has recently been a matter of discussion28,31. The results obtained here suggest that extant equid species secrete enamel at a mean rate of ≈17–18 μm/d (Fig. 3, Table 2). These values agree well with rates of enamel apposition reported for other hypsodont mammals3,8,27,28, but they differ from previously published estimates of only 5 μm/d for the domestic horse13. Kierdorf et al.29,31 considered that these lower values of enamel apposition rates reported in horses might be due to a misidentification of sub-daily incremental marks as daily marks. The daily periodicity of laminations is well-established in several mammalian taxa using experimental labelling23,28,31, but this type of study has never been conducted in equids. Therefore, we compared the estimated age of the youngest individuals with the time of formation of their first molar crown (Table 1), in order to validate and corroborate the periodicity of these features in Equus. As the formation of the first permanent molar in equids starts around the time of birth13,33, the CFT of the still-developing, unworn tooth must be equivalent to the previously calculated age of the individual. Our results show that both methodologies provide similar times for crown formation (Table 1). These results confirm the hypothesis of Kierdorf et al.29,31 that previous studies misidentified incremental lines in equid enamel13.

First insights into enamel histology of fossil equids

Finally, the enamel microstructure of the Late Pleistocene species, E. ferus and E. hydruntinus, was analysed in a first attempt to infer their pattern of molar growth. Both fossil equids present similar values for enamel secretion and extension rate but differ in the periodicity of Retzius lines, namely the repeat interval (RI) (Table 2). As is shown in Table 2, RI in E. hydruntinus comprises 4–6 days, while RI in E. ferus consists of 5–7 days. In agreement with previous research on primates and proboscideans73, these results suggest that equid repeat interval is related to body mass, as E. hydruntinus presents the lowest body mass as well as the lowest RI within the equid species investigated. The EER of all fossil teeth studied was as expected due to their macroscopic appearance (degree of wear and root development) (Fig. 6), which suggests that the time of crown formation and eruption in both extinct species was similar to that reported for extant equids. When comparing extant and extinct Equus, our results show that the daily secretion rate of enamel in E. hydruntinus significantly differs from that of E. quagga and E. hemionus (Fig. 3, Table 2). According to Dirks et al.10, “the daily secretion rate of enamel is likely to be dependent on a complex interaction of tooth size, morphology and life history”. A full assessment of the factors responsible for the higher rates of enamel apposition detected in E. hydruntinus requires a more detailed analysis involving a larger sample size, beyond the scope of the present study. However, some hypotheses can be drawn from our results. On the one hand, our findings do not match expected DSR values for this species based on body size (life history factor) or hypsodonty (tooth size factor). As E. hydruntinus is the smallest and less hypsodont equid analysed here (Table 2), an enamel secretion rate comparable to that of E. quagga would be expected. The unexpected elevation in daily secretion rates in this extinct equid species might be related to tooth morphology. Although genetically related to the extant Asiatic wild ass, E. hydruntinus is a singular extinct species that shares morphological features with extant asses, zebras and the Pliocene equid, E. stenonis 74,75,76. Specifically, the lower molar enamel of E. hydruntinus presents a primitive pattern that appears to be similar to that of E. stenonis 75. Future histological studies of the enamel in this early Pliocene species might shed light on the factors that lead to the daily secretion rates found in E. hydruntinus.

In conclusion, our histological analysis of enamel in the first lower molars of several extant and extinct Equus species allowed the estimation of dental growth parameters and the reconstruction of the enamel growth pattern in the clade. Our results provide further evidence of the already known discrepancy between dental macro- and microanatomy that might hamper growth inferences using external measurements of the crown. This finding calls for a revision of commonly used palaeontological and palaeoecological indexes, requiring a correct identification of the crown-root transition, such as the hypsodonty index. We identified three crown developmental stages (CDS) based on different growth patterns during the formation of the first molar. The beginning of each CDS is related to ontogenetic and morphological changes of the tooth. EER is very conservative within the genus and characteristic for each of the three CDS, thus showing a high potential for use in palaeodemographic and/or life history studies. We estimated a daily secretion rate for enamel of ≈17–18 μm/d in the three extant Equus species, suggesting that previously reported values of DSR in E. caballus are erroneous. Superimposition of different-aged teeth of E. hemionus reveals a high wear rate during the first six years of the animal’s life. This information should be considered in isotopic studies of tooth enamel to correctly interpret isotopic microsamples. Our results show that there are no differences in extension rate or daily secretion rate in the enamel of E. hemionus, E. quagga and E. grevyi. However, the Asiatic wild ass shows a longer period of crown formation in comparison with plains zebra and Grevy’s zebra, which indicates a higher hypsodonty in the ass. Such increase in tooth height is likely to be an adaptation of the Asiatic wild ass to counterbalance tooth wear resulting from both the more arid habitat and the extended longevity of the species. Finally, enamel histology of the Pleistocene species (E. ferus and E. hydruntinus) reveals that the smallest fossil species secreted enamel at higher rates than extant equids of the same size, E. quagga for example. We suggest that this might be related to the tooth morphology of E. hydruntinus rather than its tooth size or life history. Further studies with a larger sample size and related taxa are needed to corroborate these results.

Methods

Material

In the present study, we analysed 14 first lower molars from three extant and two Pleistocene species of Equus (Table 3). A total of 10 teeth representing different eruption stages and degrees of wear were studied in specimens from three extant species (E. hemionus, E. grevyi and E. quagga) from the Hagenbeck Zoo (Hamburg, Germany). Each extant specimen was aged by the eruption and wear patterns described for the different Equus species51,52,77,78. In view of this methodology becoming less reliable when all permanent teeth are erupted, the age of adult individuals was confirmed by counting annual growth marks present in the cementum of the first lower incisors51,52. In addition to the extant samples, four teeth of the Pleistocene species, E. ferus and E. hydruntinus, were examined. Each fossil tooth was assigned to an age category (sub-adult or adult) based on the degree of wear and root development32. The fossil material came from La Carihuela, a Late Pleistocene cave located in Piñar (Granada, Spain) and dated between 82,500 and 11,200 years BP79,80,81. Fossil samples were housed at the Catalan Institute of Palaeontology (Barcelona, Spain), while extant specimens belonged to the collections of the Zoological Institute of Hamburg University (Hamburg, Germany).

Preparation of histological thin-sections

Histological sections of teeth were prepared in our laboratory following standard procedures8. In the case of extant species, teeth were firstly extracted from the mandible and dehydrated using different concentrations of alcohol for a total period of 72 hours (70, 96 and 100%; 24 h in each). With both extant and fossil samples, each tooth was then embedded in epoxy resin (Araldite 2020) and longitudinally sectioned at the level of the protoconid in the bucco-lingual plane using a low-speed diamond saw (IsoMet, Buehler). The cut surface was later polished using a Metaserv®250 (Buehler) and fixed to a frosted glass with ultraviolet-curing adhesive (Loctite 358). Each sample was then cut and ground with a diamond saw (PetroThin, Buehler) up to a thickness of 150 μm and polished again to obtain a final thickness of approximately 120 μm. Finally, the thin sections obtained were dehydrated again in increasing concentrations of alcohol, immersed in a histological clearing agent (Histo-Clear II) and mounted using DPX medium (Scharlau) to improve visualisation of the dental microscopic features. Due to the high height of equids’ crowns, most of the samples had to be mounted on two separate slides10. The identification of incremental markings enabled confirmation of both slides being cut from the same plane (Fig. 1a).

Analysis of thin histological sections

Thin sections were examined using a polarised light microscope (Zeiss Scope.A1 microscope) and images were captured with a digital camera mounted on the microscope (AxioCam ICc5). Enamel tissue was examined in detail and different types of incremental markings were identified. Counts and measures of daily periodic laminations (Fig. 1c) along the buccal cusp were performed using ImageJ software, allowing estimates of several histological parameters that reflect the growth rate of the enamel. Firstly, the daily secretion rate (DSR) was calculated in different areas of the crown by measuring the distance between adjacent daily lines following the course of enamel prisms8 (Fig. 1c). Secondly, the repeat interval (RI), representing the periodicity of long-period lines, was quantified by counting the number of laminations between Retzius lines1 (Fig. 1d). In those areas where laminations were not well-preserved, this parameter was estimated by dividing the distance between consecutive Retzius lines by the DSR1. Thirdly, the course of incremental marks was used to reconstruct the crown formation time (CFT) of each tooth, following the method of Jordana & Köhler8. As indicated in Fig. 1b, the path of the incremental lines was traced from the enamel-dentine junction (EDJ) to the enamel surface. The distance between these lines was then measured following the course of the enamel prism; this value was divided by the DSR to determine the time required to form a specific portion of EDJ. The sum of all times along the EDJ results in the CFT. In unworn teeth of extant species, in which the crown development has not been completed, the CFT was compared with the estimated age of each specimen51,52 to validate and calibrate the daily periodicity of laminations in Equus. Fourthly, the enamel extension rate (EER) was calculated as the quotient between a determinate length of the EDJ and the number of days that it takes to be formed, as this parameter represents the growth of teeth along the EDJ1. The angle between the enamel formation front (EEF) and the EDJ was also quantified (Fig. 1e) because it is directly related to the EER3,21.

Statistics

Statistical analyses were carried out with Java Gui for R© version 1.7-1682. Kruskal-Wallis and Mann-Whitney U tests were performed to analyse differences between groups. A value of p < 0.05 was considered to be statistically significant after applying the Bonferroni correction.

Data availability

All data generated or analysed during this study are included in this published article (and its Supplementary Information files).