This is the accepted version of the article: Orlandi Oliveras, Guillem; Nacarino Meneses, Carmen; Köhler, Meike. «Dental histology of late Miocene hipparionins compared with extant Equus, and its im- plications for Equidae life history». Palaeogeography, palaeoclimatology, palaeoe- cology, Vol. 528 (August 2019), p. 133-146. DOI 10.1016/j.palaeo.2019.04.016 This version is avaible at https://ddd.uab.cat/record/215294 under the terms of the license Accepted Manuscript Dental histology of late Miocene hipparionins compared with extant Equus, and its implications for Equidae life history Guillem Orlandi-Oliveras, Carmen Nacarino-Meneses, Meike Köhler PII: S0031-0182(18)30924-6 DOI: https://doi.org/10.1016/j.palaeo.2019.04.016 Reference: PALAEO 9160 To appear in: Palaeogeography, Palaeoclimatology, Palaeoecology Received date: 8 November 2018 Revised date: 8 April 2019 Accepted date: 15 April 2019 Please cite this article as: G. Orlandi-Oliveras, C. Nacarino-Meneses and M. Köhler, Dental histology of late Miocene hipparionins compared with extant Equus, and its implications for Equidae life history, Palaeogeography, Palaeoclimatology, Palaeoecology, https://doi.org/10.1016/j.palaeo.2019.04.016 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. ACCEPTED MANUSCRIPT Dental histology of late Miocene hipparionins compared with extant Equus, and its implications for Equidae life history Guillem Orlandi-Oliveras a*, Carmen Nacarino-Meneses a,b, Meike Köhler a,c. T IP a Institut Català de Paleontologia Miquel Crusafont (ICP), Carrer de les Columnes s/n, CR Campus de la Universitat Autònoma de Barcelona, 08193 Bellaterra, Barcelona, Spain b US University of Cape Town, Department of Biological Sciences, Private Bag X3, Rhodes Gift, 7700 South Africa AN c Institut Català de Recerca i Estudis Avançats (ICREA), Pg. Lluís Companys 23, 08010, Barcelona, Spain M *Corresponding author e-mail: guillem.orlandi@icp.cat ED PT Abstract Hipparionins were a dominant element of the late Miocene faunas of Europe; however, CE their biology and ecology remain incompletely understood. In this paper, we explore the pace of life history of different-sized hipparionin horses, using dental histology, and AC compare it with extant equids. In doing so we consider (i) the size diversity of hipparionins, (ii) their generally smaller size compared to extant equids, and (iii) the allometric coupling between size and life history. In particular, we reconstruct the dental growth in lower first/second molars and in later-formed lower third molars for three hipparionin taxa: two dwarf species (Hipparion periafricanum and H. gromovae), and a larger species (H. concudense). We also analyze dental growth in an extant zebra (Equus quagga) for ACCEPTED MANUSCRIPT comparative purposes. Our results reveal that, within each species, there are differences in enamel growth parameters between the first/second molars and third molars. These results illustrate the differences in the developmental timing of these teeth and the existence of a relationship between dental growth parameters with somatic growth. We also find that hipparionin teeth grow at slower rates and tend to erupt later in time than in extant Equus. Dwarf hipparionins, moreover, exhibit lower enamel extension rates than the T IP larger species, but similar formation and eruption times. Considering the link between CR dental development and life history, these results suggest a slower life history strategy of selected hipparionins compared to extant equids, and a further slower pace of growth in US the two dwarf forms compared to the larger taxon. Keywords: paleohistology; enamel; growth rate; body size; hipparion; tooth growth AN M 1. Introduction ED The first hipparionins arrived in the Old World around 11 Ma, marking the start of the Vallesian European Land Mammal Age, and signaling one of the most important PT biochronological events of the Neogene in Europe (Bernor et al., 1996; Garcés et al., CE 2003, 1997). After their dispersal, hipparionins became dominant in European mammal associations and rapidly diversified into different forms in circum-Mediterranean areas, AC while they retained a more conservative body plan in central Europe (Alberdi, 1989; Bernor et al., 1996, 1990; Eisenmann, 1995; Woodburne, 1989). Besides their morphological diversity, hipparionin horses were highly diverse in size (Eisenmann, 1995; Ortiz- Jaureguizar and Alberdi, 2003), especially during the Turolian European Land Mammal Age (late Miocene) (Bernor et al., 1996, 1990). In the western Mediterranean late Turolian (MN13), for example, there were at least three sympatric hipparionins present (Alberdi, 1974; Alberdi and Alcalá, 1989; Ortiz-Jaureguizar and Alberdi, 2003; Pesquero, 2003): the ACCEPTED MANUSCRIPT scarce and large Hipparion primigenium truyolsi [138 kg (Ortiz-Jaureguizar and Alberdi, 2003)], the small H. gromovae [59 kg (Ortiz-Jaureguizar and Alberdi, 2003)], and the dwarf H. periafricanum [23 kg (Ortiz-Jaureguizar and Alberdi, 2003)]. Different scenarios have been proposed to explain these size differences. On the one hand, based on the evolutionary size response model in Equini (Alberdi et al., 1995), the two small taxa have been hypothesized as representing species adapted to closed and forested areas (Ortiz- T IP Jaureguizar and Alberdi, 2003). On the other hand, some authors have argued that these CR gracile and small-sized taxa were adapted to open habitats (Eisenmann, 1995; Pesquero, 2003), and that the small size represents an adaptation to energy economization due to US the less nutritive xerophyte plants that typify open habitats (Forsten, 1978, 1968). Recently, however, hipparionin size shifts have been considered as a probable outcome of AN the evolution of life history strategies that indirectly affect body size (Orlandi-Oliveras et al., 2018) as has been previously proposed for island ecosystems (Köhler, 2010; Palkovacs, M 2003). ED Hipparionins have long been suggested to follow faster life histories than Equus based on population dynamics studies (Hulbert, 1982; Kurtén, 1953; Van Valen, 1964). PT Through the study of age classes defined by dental wear stages (Hulbert, 1982; Kurtén, CE 1953; Van Valen, 1964; Woodburne and MacFadden, 1982), and considering their smaller size (Van Valen, 1964), some authors proposed shorter life spans and earlier maturity AC ages in hipparionins compared to extant equids. Body mass frequently correlates with life history (Calder, 1984). However, this is not always the case since there are other factors on which life history traits depend and that are associated with an organism’s ecology (Sibly and Brown, 2007; Stearns, 1992). Indeed, the pace of growth of hipparionins has recently been questioned based on the study of their molar eruption sequence interpreted within the context of Schultz’s Rule (Domingo et al., 2018). Grounded on the idea that the ACCEPTED MANUSCRIPT tooth replacement pattern is a valuable proxy of an animal’s life history, Schultz’s Rule states that the later eruption of the permanent molars compared to that of the replacement teeth (e.g. permanent premolars) is typically found in slow-growing taxa (Smith, 2000). Domingo et al. (2018), therefore, suggested that hipparionins might have followed a slow pace of growth due to their late third molar eruption, which occurs prior to the appearance of the fourth premolar in some extant equids (Easley et al., 2005; Hoppe et al., 2004; T IP Joubert, 1972; Lkhagvasuren et al., 2013; Smuts, 1974). Recent studies, however, claim CR that the eruption sequence is more influenced by phylogeny than by life history in primates (Monson and Hlusko, 2018a) and artiodactyls (Monson and Hlusko, 2018b; Veitschegger US and Sánchez-Villagra, 2016), challenging the predictive value of the Schultz’s Rule. Within equids, the timing of third molar and fourth premolar eruption generally differs between AN extant and extinct representatives of the family, with basal equines, Miocene anchitheriines, and hipparionins showing a later m3 eruption (Domingo et al., 2018). M However, in some taxa of the Equus genus the fourth premolar is the last to erupt (Easley ED et al., 2005; Hoppe et al., 2004; Joubert, 1972; Lkhagvasuren et al., 2013; Smuts, 1974) while in others it is the third molar (Grubb, 1981; McDonald, 1996; Penzhorn, 1982). This PT intrageneric variability suggests an adaptive component in the eruption pattern and takes CE issue with the idea of a phylogenetic meaning of this trait. The differences in the eruption sequence (Domingo et al., 2018), in dietary preferences (Semprebon et al., 2016), and in AC dental wear rates (Famoso and Davis, 2016; Hulbert, 1984) between extant equids and hipparionins make us reconsider the previous life history interpretations founded on population dynamics from the study of the dentition ontogeny (Hulbert, 1982; Kurtén, 1953; Van Valen, 1964). The deviations from the allometric coupling between life history and size (Sibly and Brown, 2007) also challenge the idea that faster life histories are necessarily associated with smaller sizes. Moreover, bearing in mind the hipparionin ecological and size diversity (Hayek et al., 1992; MacFadden et al., 1999; Ortiz-Jaureguizar and Alberdi, ACCEPTED MANUSCRIPT 2003; Tütken et al., 2013), we could further expect diverse life history adaptations within the group. Dental histology has proved to be a valuable tool for inferring life histories of extinct species since tooth formation times and growth rates are closely related to the animal’s pace of life history (Dean, 2006; Jordana et al., 2014; Schwartz et al., 2002). Additionally, T the emergence sequence of the permanent dentition has shown to generally correlate with IP the organism’s postnatal growth (Smith, 2000), and the eruption time of certain teeth is CR linked to key life history traits such as weaning and attainment of maturity (Dean, 2006; Dirks and Bowman, 2007; Engström et al., 1983; Hillson, 2005; Smith, 1991, 2000). In this US way, the study of dental growth has allowed the reconstruction of the life strategy of many extinct mammals (Dirks et al., 2012, 2009; Jordana and Köhler, 2011; Schwartz et al., AN 2002; Smith, 2016). Current advances in dental histology of extant equids, moreover, have M characterized the growth of hypsodont teeth of the genus Equus (Nacarino-Meneses et al., 2017), providing a solid framework for studying and inferring the life history of fossil equid ED taxa. PT Here we aim to explore the life history strategies of hipparionin horses and compare them to those of extant equids, focusing on the differences linked to body size variations CE within the clade. The high diversity and abundance of Turolian hipparionins in the western Mediterranean provides an excellent opportunity to study their pace of life history in AC relation to their size. Therefore, we analyze for the first time the dental development in different-sized hipparionins to shed light on their life history strategies. Specifically, we reconstructed the tooth growth patterns of lower first/second molars and lower third molars of three hipparionin taxa (Hipparion periafricanum, Hipparion gromovae, and the larger Hipparion concudense) together with the first and third molar growth of an extant equid. Amongst wild extant equids, we have chosen Equus quagga as a model because unlike ACCEPTED MANUSCRIPT the larger Equus grevyi or the longer-lived Equus hemionus, it does not show any extreme life history trait (Ernest, 2003; Grubb, 1981; Lkhagvasuren et al., 2013; Smuts, 1974). We have analyzed first and third lower molars as these are formed and erupt during different moments of ontogeny that are characterized by two major life history events, weaning and attainment of maturity, respectively. Furthermore, the eruption of the first molar has been shown to correlate with different life history traits, and thus it can be used as a predictor of T IP the organisms’ pace of life history (Dirks and Bowman, 2007; Smith, 2000; Veitschegger and Sánchez-Villagra, 2016). The dental mineralization sequence – intimately related to CR the enamel matrix formation – is the same between hipparionin horses and extant horses US (Domingo et al., 2018; Hoppe et al., 2004): the first lower molar (m1) begins to mineralize before birth, followed by the m2, and later by the p2 and the p3. Although the p4 eruption AN occurs earlier in hipparionins, the p4 starts to mineralize before the m3 as it happens in extant Equus (Domingo et al., 2018; Hoppe et al., 2004). This similar onset of the tooth M formation, and the same degree of overlap between extant equids and hipparionins (e.g. ED the third molar is still being formed when the m1 and the m2 are fully in wear), endorse our interpretations obtained from individual tooth growth reconstructions. Due to their diversity PT and abundance, hipparionins constitute a key mammalian group widely used in CE paleoecological studies, especially in those involving isotopes (Domingo et al., 2009; Matson and Fox, 2010; Nelson, 2005; Rey et al., 2013; Van Dam and Reichart, 2009). AC Besides the importance of dental growth pattern as a life history proxy, the understanding of hipparionin molar growth patterns provides an important basis for future paleoecological inferences from hipparionin dental analyses. 2. Material and Methods The extremely high-crowned molars of Equus are produced by delaying the root formation (Hoppe et al., 2004; Nacarino-Meneses et al., 2017), which leads to extended ACCEPTED MANUSCRIPT dental growth periods. Eruption of the crown occurs gradually, and tooth wear starts before the roots are formed (Kirkland et al., 1996). For this reason, the entire tooth development in Equus must be reconstructed from compositions of molars with different wear degrees (Nacarino-Meneses et al., 2017). Based on differences in the pattern of growth during the crown formation of unworn and worn extant equid molars, Nacarino-Meneses et al. (2017) defined three crown developmental stages (CDS) that differed in growth rate. The first T IP stage is characterized by a fast and linear growth (CDS I), followed by a significant CR decrease in growth rate (CDS II) and the attainment of an almost residual enamel extension rate in the last formed part of the cervix (CDS III). These CDS described in US Equus, further agree with the morphological appearance of the molar. The transition from stage I to stage II matches the tooth eruption time, and that of CDS II to CDS III represents AN the macroscopic root appearance (Nacarino-Meneses et al., 2017). The hypsodont teeth of hipparionins, however, tend to finish the crown formation shortly after the tooth eruption, M allowing the observation of little worn rooted teeth (Supplementary Material Fig. S1). The ED histological analyses of the unworn or slightly worn hipparionin teeth, thus, allows the reconstruction of the complete tooth growth. PT 2.1 Fossil material CE In this study, we studied isolated lower molars from three different hipparionin species of the Spanish Turolian. The teeth of the dwarf hipparionins H. periafricanum and AC H. gromovae come from the El Arquillo fossil site (Teruel Basin, Spain), also known as Rambla de Valdecebro II. This fossil site is dated to the late Turolian (MN13) and establishes the reference fauna for the MN13 with an estimated age of 6.2 Ma (de Bruijn et al., 1992; Van Dam et al., 2001). The H. concudense molars were from Concud, a classical and rich fossil site from the same Teruel Basin. This locality is less than 10 km away from El Arquillo and is dated around 7 Ma on the middle Turolian (MN12) (Pesquero ACCEPTED MANUSCRIPT et al., 2010) (Fig. 1). All fossil samples are accessioned in the Institut Català de Paleontologia (Barcelona, Spain) collections (IPS-91731, IPS-91732, IPS-91733, IPS- 91734, IPS-91762, IPS-91763, IPS-40644, IPS-40665, IPS-40668, IPS-85504, IPS-85506, IPS-85508, IPS-86850, IPS-86851, IPS-87855, IPS-87856, IPS-87857, Table 1). Based on sectioned material, we analyzed two lower first/second molars (m1/m2) T and three lower third molars (m3) for the smallest Old World hipparionin Hipparion IP periafricanum, four lower m3 molars and two lower m1/m2 molars from the small-sized CR Hipparion gromovae, and four lower m3 molars and two lower m1/m2 molars from the large Hipparion concudense (Table 1). To the greatest extent possible, the less worn teeth US were selected in order to retain the maximal growth record and make analogous comparisons. Following Sondaar (1961) and Pesquero (2003), we used tooth width and AN length measurements at 1 cm above the base of the tooth as a biometric criterion for M interspecific differentiation between the two contemporaneous small species H. gromovae and H. periafricanum (Supplementary Material Fig. S2 and S3). Body mass estimations ED based on first phalanx measurements (Alberdi et al., 1995) were obtained from the literature (Ortiz-Jaureguizar and Alberdi, 2003; Pesquero and Alberdi, 2012). Hypsodonty PT indices calculated from measurements of m1/m2 molars were also compiled from literature CE (Cantalapiedra et al., 2017; Pesquero, 2003). Using the same procedure described in Cantalapiedra et al. (2017), we calculated the hypsodonty indices for m3 molars (Table 1). AC 2.2 Extant Equus material Our sample also includes four first molars of extant Equus quagga from the Hagenbeck Zoo (Hamburg, Germany). These teeth represent different eruption and wear stages (Table 1), thus comprising the entire tooth development of Equus molars (Nacarino-Meneses et al., 2017). We also sectioned three third molars from Equus quagga depicting different eruption and wear stages (Table 1). These third molars came from the ACCEPTED MANUSCRIPT Réserve Africaine de Sigean (Sigean, France), the Barcelona Zoo (Barcelona, Spain) and the Hagenbeck Zoo (Hamburg, Germany). The extant samples from the Hagenbeck Zoo are accessioned in the Centrum für Naturkunde of the Universität Hamburg (Hamburg, Germany), the Barcelona Zoo specimen is accessioned in the Museu de Ciències Naturals de Barcelona (Barcelona, Spain), and the specimen from Sigean is housed at the Institut Català de Paleontologia Miquel Crusafont (Barcelona, Spain). T IP 2.3 Thin section preparation and study CR The dental histological sections were prepared following the standard procedures of our laboratory (Jordana et al., 2014; Nacarino-Meneses et al., 2017). Each molar was US embedded in epoxy resin (Araldite 2020) and sectioned buccolingually through the higher mesial cusps (protoconid and metaconid) using an IsoMet low-speed saw (Buehler). In AN some cases, highly curved third molars were previously cut in a transversal plane. At this M stage, the exposed buccolingual surfaces were polished using a grinder polisher (Buehler, MetaServ 250) and glued to a glass slide using ultraviolet curing glue (Loctite 352). ED Thereafter, the glass-mounted samples were cut using a diamond saw (Buehler, Petrothin) up to a thickness of 300 μm and grounded to around 120-70 μm using the grinder polisher. PT Finally, each sample was dehydrated in alcohol gradients and immersed in a histological CE clearing agent (Histo-Clear II) prior to covering them with a DPX medium. The resulting thin sections were studied under polarized light using a Zeiss Scope AC A1 microscope with an attached digital camera (AxioCam ICc5). To estimate the tooth growth parameters, the acquired images of the overall enamel band (Fig. 2A) were merged using Adobe Photoshop® and analyzed with ZEN 2011® (Carl Zeiss) and Image J software. ACCEPTED MANUSCRIPT 2.4 Histological analysis techniques and dental growth parameters We used measurements and counts of enamel incremental markings to assess the tooth growth patterns of the species. As the dental growth parameters depend on the cusp and enamel side studied (Jordana and Köhler, 2011; Kierdorf et al., 2013), we decided to standardize and analyze the buccal enamel band of the protoconid in agreement with previous work on equid teeth (Nacarino-Meneses et al., 2017). This allows future T IP comparisons between extinct and extant equids. We did not analyze the lingual enamel CR band of the metaconid due to the presence of more marked Hunter-Schreger bands that hindered the correct observation of incremental markings. Enamel band thickness of each US tooth was measured perpendicular to the enamel-dentine junction at a homologous zone situated 1 cm above the root. In order to reconstruct the entire tooth growth curve, the AN length of the cuspal enamel lost in worn molars was estimated by considering both the profile of unworn teeth (Tafforeau et al., 2007) and the hypoconulid wear degree (Stirton, M 1941). We used the mean enamel extension rate of the nearest area to extrapolate the ED time that the worn portion required to be formed. In ungulates, the main short-period enamel incremental features are laminations, PT which have a circadian periodicity and run oblique from the enamel-dentine junction (EDJ) CE (Jordana and Köhler, 2011; Kierdorf et al., 2013; Smith, 2006; Tafforeau et al., 2007). The other common incremental features are the Retzius lines or striae of Retzius, which AC correspond to the successive positions of the ameloblast secretory front (Dean, 2000). These long-period marks represent variations in the secretory activity, hypothesized to be ruled by two independent intern cycles (Smith, 2006) or by a general chronobiological rhythm (Bromage et al., 2016). Due to their known cyclical and regular nature, both laminations and Retzius lines permit the calculation of the following tooth growth parameters: the enamel daily secretion rate (DSR), or the amount of enamel secreted by ACCEPTED MANUSCRIPT an ameloblast per day; the enamel extension rate (EER), or the enamel growth rate along the EDJ; and the total crown formation time (CFT). Firstly, and following prior studies on ungulate teeth (Jordana et al., 2014; Jordana and Köhler, 2011), we calculated the DSR quantifying the distance between laminations through the course of the enamel prisms (Fig. 2B). The DSR values have been seen to increase from inner to outer enamel regions (Kierdorf et al., 2014; Metcalfe and Longstaffe, 2012; Smith, 2008), although in some T IP cases no significant differences between enamel areas have been identified (Nacarino- CR Meneses et al., 2017). We made many random counts in different enamel zones within all the tooth length to gather all the variation and obtain a representative mean. Because we US were not able to perform the same amount of DSR measures in all the crown regions and enamel zones due to fossil preservation, direct comparison between the mean obtained AN DSRs should be taken with caution. We calculated EER following the methodology described in Jordana & Köhler (2011): the time that takes to form the distance that M separate two incremental lines through the EDJ (X, Fig. 2C) is determined by dividing the ED prism length that separates those two incremental features (Y, Fig. 2C) by the rate of daily enamel deposition (DSR). This calculation is based on the ideas that the enamel prisms PT mark the enamel growth path and that accentuated lines indicate the position of the CE developing enamel front at a determinate moment (Dean, 2000). Applying this methodology through the entire enamel band, we obtained EER values in different tooth AC portions. The EERs were calculated at the intersection of clearly visible accentuated Retzius or stress lines, rather than at regularly spaced points (Dirks et al., 2012). In order to permit analogous comparisons, we divided the enamel band of each tooth into three thirds (cuspal, middle and cervical). Finally, the CFT was estimated calculating the overall time required to form the enamel band. ACCEPTED MANUSCRIPT Differences between groups within each dental growth parameter were tested with a degree of significance of α = 0.05. Non-parametric Kruskall-Wallis tests were used to test for differences in the DSR values between species since normality was not always met. Differences in EER values between fossil species were compared using an ANCOVA model, as this considers the decrease in EER along the crown height. Prior to the ANCOVA analysis, we tested for the significance of the interaction between the categoric T IP predictive variable (species) and the continuous covariate (crown height) in the two CR different subsets analyzed (first/second molars and third molars). Moreover, we tested for the homogeneity variances (Levene test). In both cases, when the extant Equus data was US excluded, we did not find significant violations of the assumptions of the ANCOVA model (Supplementary Material Table S1-S8). The statistical tests and graphs were performed AN using IBM SPPS Statistics 20 ® and PAST v. 3.18 (Hammer et al., 2001). M 3. Results Laminations represent the main incremental features observed in our sample. These ED are more appreciable on the central and inner enamel zones, although also identifiable PT between some striae of Retzius that outcrop at the outer enamel surface. In the inner enamel zone, laminations are less regularly spaced and more difficult to distinguish as CE some sub-daily incremental lines appear. Retzius lines are clearer in the outer enamel (Fig. 2C). These Retzius lines are also more discernible in the cervical region (Fig. 2C), AC where the angles between those and the EDJ are more obtuse (~ 10º) than in the middle or cuspal region (<5º). These observations agree with previous findings on equid enamel histology (Hoppe et al., 2004; Nacarino-Meneses et al., 2017). 3.1 Enamel daily secretion rates In our sample, the mean daily secretion rates vary from ~14 to ~17 μm /day (Table 2). Contrary to hipparionins, Equus quagga shows different DSRs within the three different ACCEPTED MANUSCRIPT tooth zones (cuspal, middle, cervical) considered from the reconstructed third molar (p < 0.001). These differences are only observed in the m3, since the m1 does not show significant variations on the DSR through the tooth length (p = 0.08). The first formed region of the third molar presents higher DSRs, while it decreases in the middle zone and reaches its minimum at the final formed root zone (Table 2). Considering this, the calculation of the other parameters in third molars of Equus has been done using the T IP mean DSR of each stage, not the general mean of the taxon. When compared within CR hipparionin species, the DSRs of the first/second molars are significantly higher than those reported in third molars (p < 0.001) (Table 2). We do not find these significant differences US between the DSR of m1 and m3 in extant Equus (p = 0.34), when we use the medial region DSR of the m3 as a general mean for the third molar. AN The DSR is significantly different between species, both in the case of the analyzed M m1/m2 (p < 0.001) and the m3 (p < 0.001). Among the first/second molars, the H. concudense m1/m2 exhibit the lowest DSR values with a mean value of 15.76 μm/day, ED which is significantly different from the higher secretion rates of H. gromovae (p = 0.009) and Equus (p < 0.001) (Table 2). The first/second molars of H. periafricanum have a mean PT DSR of 16 μm/day, which is significantly different (p = 0.009) to the higher rates of Equus CE quagga (~17 μm/day). The DSR of the third molars differ significantly between most species (p < 0,001), except for H. concudense and H. periafricanum, which present similar AC low DSRs (Table 2). In general, therefore, H. gromovae exhibit greater mean DSR values than the other two hipparionin species, both in the third and in the first/second molars, and this is only surpassed by the always higher DSRs of Equus quagga. This pattern is followed both in the m3 and the m1/m2, with lower values of DSR in the later formed third molars (Table 2). It should be noted that the same number of DSR measures within the crown regions and the enamel zones was not possible in our sample. We should, hence, ACCEPTED MANUSCRIPT be prudent when interpreting the differences in DSRs found between species and tooth types. However, since the DSRs in the equids’ first molars tend to do not vary significantly between regions (Nacarino-Meneses et al., 2017), we assume that our results are more reliable than if the same measures were performed on taxa that reported high variations on the DSRs, like pigs (Kierdorf et al., 2019, 2014). T 3.2 Enamel extension rates IP The first/second molars’ ameloblasts differentiated along the EDJ at higher rates CR than in the third molars in all the studied taxa, particularly in the cuspal and medium zones (Table 2). Due to these differences, and in order to make interspecific comparisons, we US separated the higher EER values of the m1/m2 from the lower EERs of the m3. We always observed higher EERs in the first formed enamel area, while the rates were lower towards AN the cervical region (Table 2, Fig. 3 and Supplementary Material Fig. S4). For example, the M first third of the crown of the m1/m2 of H. periafricanum grow at a mean rate of 166.62 μm/day, while at the end of the enamel band the mean EER is 37.27 μm/day. This ED diminution is present in all taxa and in both tooth types, thus representing a general decrease in the tooth growth rate through the development of the tooth (Table 2, Fig. 3 PT and Supplementary Material Fig. S4). Taking into account this growth rate decrease, the CE intraspecific variation in EER within the crown height is high, thus hindering the growth rate comparisons between species. To evaluate the EER differences between species, AC therefore, we considered the crown height as a covariate affecting the EER. By doing this, interspecific differences between EER in fossil first/second molars and third molars are significant (p < 0.001). These differences are found between all groups of hipparionin teeth (p ≤ 0.001), except for the m1/m2 of H. gromovae and H. periafricanum, which grow at similar rates (p = 0.075). In all cases, H. concudense teeth grow at higher rates, while the teeth of the dwarf species grow at a slower pace. Although not directly tested by the ACCEPTED MANUSCRIPT ANCOVA model (see section 2.4), Equus quagga teeth develop at higher rates compared to the three hipparionin species (Table 2 and Fig. 3). The higher EER of extant Equus molars compared to hipparionins molars is comparably more pronounced in m1/m2 than in m3 (Table 2 and Fig. 3). 3.3 Growth reconstruction and crown formation times T To reflect the whole tooth growth pattern, we plotted the crown height against the IP crown formation time (CFT) for each species tooth type (Fig. 4). Extant Equus tooth CR development has been plotted considering the entire tooth growth reconstruction using differently worn teeth (Nacarino-Meneses et al., 2017) (see section 2). All the US reconstructed tooth growth curves fit well with von Bertalanffy growth models (Fig. 4 and Supplementary Material Fig. S5). Compared to recently erupted Equus teeth (Nacarino- AN Meneses et al., 2017), we confirm that unworn or slightly worn hipparionin teeth register M most of the tooth development, since asymptotic size is being reached (Fig. 4). Some residual cervical enamel, however, could be further deposited although without increasing ED the tooth height substantially. Steeper slopes than those for the H. gromovae and H. periafricanum teeth indicate the higher growth rates of the Equus and H. concudense PT molars (Fig. 4). We also identify a higher EER in m1/m2 compared to m3 in the steeper CE slopes of the first molars within each species (Supplementary Material Fig. S5). These intraspecific EER differences between m3 and m1/m2 are more pronounced in the Equus AC quagga molars than in hipparionin horses, as first molars show comparably much higher crown extension rates (Fig. 3 and Supplementary Material Fig. S5). The complete growth period or CFT is explored in the teeth that have not lost cuspal growth register and that have started forming roots (Table 3). Despite having the higher crowns, the m1/m2 of H. concudense take 500 days to be formed, a similar time to that of the smaller species’ teeth (Table 3 and Fig. 4). When we compare the third molars ACCEPTED MANUSCRIPT total CFT, these generally take more time to form than first/second molars. Within species, the m3 of H. gromovae and H. periafricanum require similar time to form (around 600 days), whereas the H. concudense m3 spans 100 days more to finish formation (Table 3). The total crown formation time of the composite Equus tooth is much higher, taking 785 days in the case of the first molar and 1750 days in the third. However, these Equus CFTs are the product of differently worn teeth compositions, as they keep growing after eruption T IP and delay their root closure, depositing residual enamel (Nacarino-Meneses et al., 2017). CR Analogous comparisons of the formation time in Equus and Hipparion teeth that present a similar ontogenetic stage (recently erupted and little worn, Table 1: Equus quagga IPS- US 92341 and Hipparion concudense IPS-91763) show that, although the H. concudense tooth (IPS-91763) lost the final part of enamel due to preservation, it took more time (367 AN days) to be formed than the extant Equus tooth (200 days) (Fig. 5). Growth rate differences between these two teeth are also evident (Fig. 5). M Aside from the total growth span, eruption times can be inferred considering that the ED different crown developmental stages are identifiable in the reconstructed growth curve (Nacarino-Meneses et al., 2017). Hence, here we estimated tooth eruption from the PT transition from the first CDS to the second CDS (arrowheads, Fig. 4), following Nacarino- CE Meneses et al. (2017). That is, from the change of an almost linear growth (CDS I) over a reduced growth rate (CDS II) that represents intermediate rates of growth (Nacarino- AC Meneses et al., 2017). Based on the Equus quagga m1 growth curve, we can identify this point at around 230 days, thus indicating the eruption of the first molar. For the Equus quagga m3, on the other hand, we estimate a time of eruption of around 500 days after the start of the formation (Fig. 4). As the first molars of extant zebras erupt around 9 months after birth (270 days) (Smuts, 1974), our estimation for the m1 seems reliable. In zebras, the third molar erupts at the age of 3 years and 3 months (Smuts, 1974) (39 months). ACCEPTED MANUSCRIPT Assuming that zebras’ m3 starts to develop at a similar age as when the third molar of Equus caballus starts to mineralize [21 months of age (Hoppe et al., 2004)], the estimated eruption time of 500 days (~16.5 months) (Fig. 4) also agrees quite well. The decrease in growth rate is more progressive in hipparionin teeth than in Equus (Fig. 4), thus, the transitions between the CDSs in the hipparionin growth curves are less clear. Nevertheless, a change in the slope of the curves is identifiable (arrowheads, Fig. 4) and T IP might match the eruption time as in the extant equid model. Applying this methodology, we infer a later eruption time in hipparionins’ fist/second molars compared to extant Equus CR [300 days to 350 days of formation (Fig. 4)]. The hipparionin third molars, on the other US hand, likely erupted around 500 days after the beginning of their formation, similar to extant E. quagga third molars. Besides using the equid tooth growth model by Nacarino- AN Meneses et al. (2017), we can further endorse the inferences of the eruption times based on radiographic evidence of hipparionin eruptive sequences (Domingo et al., 2018). M Judging from the radiographs provided in Domingo et al. (2018), the enamel band of the ED mesial lobe of an hipparionin m1 close to eruption measures ~53 mm, and that of a fully erupted one ~60 mm. Considering the tooth growth curve of the larger hipparionin species PT (Fig. 4), these teeth may have taken ~300 to ~350 days to be formed, respectively. In the third molars, an unerupted – still in crypt – hipparionin m3 measures ~45 mm and a totally CE erupted m3 measures ~55 mm, which represent ~350 and ~600 days of formation AC respectively. Thereafter, our inferred eruption times in hipparionin lower first/second molars and third molars are in agreement with the estimated CFTs of the teeth that have erupted or are close to eruption. 4. Discussion Our study represents the first comprehensive analysis of hipparionin tooth growth through dental histology. The addition of extant Equus quagga teeth, moreover, allows us ACCEPTED MANUSCRIPT to compare the growth patterns and to identify differences in the tooth formation between extant and extinct representatives of the Equinae subfamily. Dental growth variations between different sets of permanent teeth have hitherto not been explored in ungulate mammals. Here, we analyze and compare, for the first time in ungulates, two different sets of permanent teeth, thus characterizing the growth of dental tissues during two distinct ontogenetic stages. A few other studies have addressed the differences in growth T IP parameters between tooth loci, but these have been focused on hominoids (Guatelli- CR Steinberg et al., 2012; Shellis, 1984; Smith, 2016; Smith et al., 2007). Besides the growth characterization of the hipparionin teeth, widely used in many paleoecological studies US (Domingo et al., 2009; Matson and Fox, 2010; Nelson, 2005; Rey et al., 2013; Van Dam and Reichart, 2009), our approach also allows us to make further inferences of the life AN history of the studied taxa. The assessment of the dental growth patterns, thus, led us to infer differences in the lifestyles of the different-sized hipparionin taxa. Moreover, the M comparison between hipparionin and Equus molars has highlighted significant tooth ED growth differences that can be both involved in life history differences and in the formation of the higher crowned teeth of Equus. PT 4.1 Dental growth parameters: Daily secretion rate CE We identified that the dominant incremental markings in hipparionin enamel are laminations, which is in agreement with previous findings on ungulate teeth (Jordana and AC Köhler, 2011; Kierdorf et al., 2013, 2014; Nacarino-Meneses et al., 2017; Tafforeau et al., 2007). The daily secretion rates obtained through quantification of the lamination spacing (Table 2) are congruent with those determined in the aforementioned studies focused on ungulate species. Nevertheless, our estimated DSR values ranging from 14 to 17 μm/day do not correspond to previous values of 5 μm/day calculated in domestic horses (Hoppe et al., 2004). As formerly pointed out by other authors (Kierdorf et al., 2013, 2014), and ACCEPTED MANUSCRIPT already addressed in previous studies with equid teeth (Nacarino-Meneses et al., 2017), we consider those undervalued preliminary measures a result of the misinterpretation of laminations and sub-daily incremental marks. In our sample, we identified higher DSRs in the hipparionin first/second molars compared to third molars (Table 2). Although the mean DSRs can be skewed to some T extent by the unbalanced counts between the different enamel and crown portions, the IP higher DSR values in m1/m2 compared to those of the m3 are commonly found within all CR zones. The few previous studies that compared daily enamel secretion rate between hominoid tooth types did not detect variations between the molars’ DSRs (Smith, 2016; US Smith et al., 2007). In a similar way, we do not identify DSR differences between the Equus quagga tooth types. In our case, we interpret the lack of differences in Equus tooth AN types as being a result of the lumped measures, the lower number of observations done in M the Equus first molars (n = 45) compared to the third molars (n = 138), and to the significant DSR decrease observed in the consecutive developmental stages of the Equus ED third molars. Similarly, a decrease in the enamel secretion rate at the end of the third molar formation has been observed in human teeth (Guatelli-Steinberg et al., 2012). We PT suggest that these lowered secretion rates on the Equus third molars might be related to a CE general depletion of tissue depositional rates after the eruption of the third molar (transition from CDS I to CDS II sensu Nacarino-Meneses et al. (2017), as this event is broadly AC correlated with the attainment of skeletal maturity (Dean, 2006). Thus, the lower DSRs of hipparionin third molars compared to the first/second molars can also be linked to the differences in the formation timing. In ungulates, however, the DSR parameter has shown to be more correlated with the tooth morphology (i.e. hypsodonty degree) than to the pace of life history or somatic growth (Jordana et al., 2014). This would agree with the differences observed in the DSRs between the hipparionin tooth types, since third molars ACCEPTED MANUSCRIPT are lower crowned and less hypsodont than first/second molars (Table 1 and 3). The DSR, moreover, can also be related to the enamel band thickness, since the amount of enamel secreted per day (i.e. DSR) and the ameloblast secretory lifespan determine the final width of the enamel (Kierdorf et al., 2014). Accordingly, we measured the linear enamel thickness in our sample (Supplementary Material Fig. S6). Despite their lower DSRs, the third molars tend to have somewhat thicker enamel bands than the first/second molars. T IP This result suggests that enamel thickness differences between tooth types are more CR influenced by the secretory lifespan of the ameloblasts (Kierdorf et al., 2014) or the total CFT (Kierdorf et al., 2019) than by the DSR. Regarding intergeneric comparisons, the US higher DSR values of extant Equus’ molars compared to those of hipparionins are expected considering the proposed relationship between hypsodonty and DSR (Jordana et AN al., 2014) but also considering the Equus’ thicker enamel bands (Supplementary Material Fig. S6). However, interspecific differences in the hipparionin DSR do not seem to be M directly linked to differences in hypsodonty (Table 1), nor to differences in enamel ED thickness (Supplementary Material Fig. S6). For example, the smallest and most hypsodont H. periafricanum presents low DSRs comparable to those of the less hypsodont PT H. concudense, which has considerably thicker enamel bands. Therefore, the ameloblast CE secretory lifespan and the CFTs may have more important roles than DSR in the formation of thicker enamel bands (Kierdorf et al., 2019) and more hypsodont teeth (Nacarino- AC Meneses et al., 2017; Witzel et al., 2018). Hence, the small differences in the hypsodonty between these high-crowned hipparionins might be of little importance in terms of DSR changes. We suggest that the relationship between hypsodonty and DSR (Jordana et al., 2014) could be more easily observed in large-scale comparisons between more brachydont to more hypsodont species. Alternatively, the differences identified here could represent interspecific variations in the quotient between enamel and dentine, a characteristic that together with hypsodonty determine tooth durability (Famoso et al., ACCEPTED MANUSCRIPT 2013; Jordana et al., 2012; Winkler and Kaiser, 2015). Thus, we interpret that DSR is probably influenced by many other factors besides hypsodonty, such as tooth morphology, phylogeny (Dirks et al., 2012), enamel thickness (Kierdorf et al., 2014), tooth formation timing or even general somatic growth. The interplay of many of these factors might determine the overall DSR of the tooth. Further studies addressing the relationship between this parameter and the disposal and proportion of enamel could shed light on this T IP topic. CR 4.2 Dental growth parameters: Extension rate Various isotopic studies have already inferred hipparionin tooth growth in height US thanks to the identification of cyclic variations in δ13C and δ18O signal (Matson and Fox, 2010; Nelson, 2005; Van Dam and Reichart, 2009). These estimations represent AN mineralization rates rather than matrix secretion rates. However, since the enamel starts to M mineralize soon after matrix deposition, the extension rates estimated from isotopic studies tend to resemble those calculated using histology (Metcalfe and Longstaffe, 2012). ED Accordingly, the tooth growth rates calculated from the isotopic cyclicity in hipparionin teeth agree well with the enamel extension rates calculated here from incremental growth PT marks. For example, van Dam and Reichart (2009) estimated a growth rate of 40 mm/year CE (= 109.60 μm/day) for the upper third molar of H. concudense from Concud. In our study, we obtained a similar mean EER of 95.54 μm/day in the middle region of the lower third AC molar from the same species population, supporting the link between formation and mineralization rates, which has also been corroborated in extant equids’ molars (Nacarino- Meneses et al., 2017). In general, the proper estimation of a representative EER mean in high-crowned teeth is difficult since there is a general decrease of this parameter through the course of tooth formation (Fig. 3). This declining in the tooth growth pace from the cusp to the cervix ACCEPTED MANUSCRIPT is a reflection of a non-linear dental growth pattern (Bendrey et al., 2015), and has already been described in other mammalian groups (Guatelli-Steinberg et al., 2012; Jordana and Köhler, 2011; Kierdorf et al., 2013; Shellis, 1984) and in extant equids (Nacarino-Meneses et al., 2017). Due to the larger crowns’ heights, the variation in the EER values through the enamel band is especially high in hypsodont ungulate teeth and should be considered when making comparisons. Moreover, apart from the within-tooth variation, the extension T IP rates of the first/second molars and the third molars differed in our sample, being always CR higher in the first/second molars of all taxa. Similarly, Shellis (1984) found that in humans the first formed deciduous teeth grow eight times faster than third molars do, and higher US crown extension rates have been calculated in first molars compared to third molars in other hominoids (Smith, 2016; Smith et al., 2010, 2007). Higher EERs might be influenced AN by the need to form the tooth within a shorter CFT (Kierdorf et al., 2014), as tooth crowns in m1/m2 tend to take less time to be formed than those of m3 (Table 3 and M Supplementary Material Fig. S5). Moreover, we interpret the higher EER values of the ED m1/m2 compared to those of the m3 as a reflection of the higher growth rates of the organism. Mammalian first molars are formed and erupt earlier in the ontogeny, while third PT molars finish their formation when skeletal maturity is being reached (Dean, 2006; Hillson, CE 2005; Hoppe et al., 2004). For this reason, we can expect greater EERs in m1/m2 since there are more activated ameloblasts due to the higher general somatic growth than AC posteriorly in ontogeny, when m3 are being formed. The fact that the deposition of bone lamella is coupled with the Retzius line formation (Bromage et al., 2009), and that there is an endogenous rhythm controlling both functions (Bromage et al., 2016), endorses the relationship between enamel growth rates with bone growth and, then, the general somatic growth. The same idea has recently been suggested due to the discovery of lower EER in mammaliamorphs that experienced an earlier reduction of the overall skeletal growth rates (O’Meara et al., 2018). The influence of growth hormones (GH and IGF-1) and their ACCEPTED MANUSCRIPT receptors in the differentiation and proliferation of the odontogenic cells involved in the formation of teeth (Symons and Seymour, 2000; Young, 1995; Zhang et al., 1992) further supports this link. Besides the variations in intraspecific somatic growth rates during ontogeny, the differences EERs between taxa might also be indicative of differences in the organism’s T pace of growth (Jordana et al., 2014). The estimation of this parameter, therefore, might IP allow further interpretations to be made on the biology of animals, although caution should CR be taken due to the variation of these parameters between and within tooth types. Considering the differences between the third and first/second molars, and the decreasing US of EER through the enamel band, we could identify higher EER in E. quagga compared to all hipparionin horses, and higher EER in the larger hipparionin species compared to the AN dwarfed taxa. Within a sample of 21 ruminant species and taking into account their M phylogenetic relationships, Jordana et al. (2014) found a significant correlation between the residuals of the age at first reproduction – a proxy of the pace of life history corrected ED by body size – and the tooth enamel extension rate. Hence, lower EER values have shown to be indicative of a slower pace of growth than those expected from body size scaling PT (Jordana et al., 2014). Additionally, slower extension rates have also been associated to a CE later attainment of foraging independence in cebids (Hogg and Walker, 2011) and to the prolonged growth periods of modern humans compared to Homo erectus and earlier AC hominins (Dean et al., 2001). The influence of the EER in the formation of a tooth at the required time for its eruption, might be the reason underlying the EER relationship with the timing of life history events, since the tooth development is strongly tied to the mammals’ life cycle (Smith, 2000, 1991). In terms of EER, therefore, we can infer a slower pace of growth in the analyzed hipparionin species in comparison to extant equids. Moreover, the ACCEPTED MANUSCRIPT extension rates of the dwarf late Turolian Spanish forms further suggest a slower pace of growth compared to H. concudense. Some authors have considered the existence of an allometric coupling of the EER with total crown height (Dirks et al., 2012) or body mass (O’Meara et al., 2018). Taking this into account, the lower EERs of the hipparionins compared to extant equids, and the lower T EER of dwarf hipparionins compared to larger representatives, could represent the IP consequence of the allometric scaling of this parameter. The same relationship between CR tooth size and EER was also identified in humans by Guatelli-Steinberg et al.(2012), estimating higher EERs in those teeth with larger enamel dentine junction lengths. On the US other hand, however, the formation of larger canines in male gorillas has been related to longer formation times rather than to higher growth rates (Schwartz and Dean, 2001). AN Smith (2016) also found that taller fossil teeth from larger extinct pongines had lower EERs M compared to those of extant orangutans. Similarly, Jordana and Köhler (2011) identified lower EERs in the higher crowned insular Myotragus balearicus, which was a slow- ED growing mammal with a delayed life history schedule compared to extant caprines (Jordana et al., 2012; Köhler, 2010; Köhler and Moyà-Solà, 2009). In our case, H. PT periafricanum and H. gromovae were relatively higher crowned – more hypsodont – than CE the larger hipparionin species from the late Miocene (Pesquero, 2003). Thus, to form those more hypsodont teeth, we should expect higher EERs compared to less hypsodont AC species (H. concudense), especially if we consider that they would have advanced the general life history schedule due to their reduced size and the theoretical size coupling with life history traits (Calder, 1984; Stearns, 1992). Our results suggest the opposite scenario, namely more hypsodont dwarf hipparionin species exhibiting lower EER, which we further relate to a slower life history than the expected from their size. ACCEPTED MANUSCRIPT 4.3 Formation and eruption times The previous life history interpretations are not only supported by the EER values, but also by the estimated formation and eruption times, two parameters that are strongly correlated with the life history of an organism (Dean, 2006; Macho and Williamson, 2002; Smith, 2000, 1991). T Regarding the formation times, we generally identified higher CFTs in third molars IP compared to those of first molars. Similarly, CFTs in human third molars are much higher CR than in the case of other teeth (Shellis, 1984), as has also been observed in other hominoids where the first molars form faster and in less time than third molars (Smith et US al., 2010, 2007). The longer formation times in the third molars might be related to their lower EER. When comparing between taxa, the overall crown formation times in Equus AN are much higher than in hipparionins, since Equus CFTs are the result of a composite M growth reconstruction as they extend the molar formation. When we compare growth curves and, thus, observe the total crown formation times of individual teeth of similar ED developmental stages, we observe higher CFTs in hipparionin teeth compared to Equus (Fig. 5), and in some cases higher CFTs in the small hipparionins compared to the larger PT species (Fig. 4). Macho and Williamson (2002) found a significant correlation between the CE crown formation times corrected by body mass and the relative gestation length in African bovids. Although they arguably overestimated the CFTs due to the misinterpretation of AC sub-daily striations with daily increments (Kierdorf et al., 2014, 2013), the similar repeat intervals that they reported relating these features across the studied taxa (Macho and Williamson, 2002) might maintain the proportionality of the correction and thus the significance of the correlation. Hence, taking in mind the high relative CFTs obtained in hipparionins, and especially the two dwarf taxa, we might infer higher relative gestation ACCEPTED MANUSCRIPT lengths, a characteristic of a slower pace of life history compared to Equus and to larger hipparionins. However, a more direct life history proxy not related to the continuous growth of extant equid teeth is the timing of dental eruption (Smith, 2000). On the one hand, the first molar eruption is considered as an indicator of the weaning event (Dirks and Bowman, T 2007; Smith, 2000, 1991), and the age of eruption is an informative estimate of postnatal IP growth rate, with faster-growing animals erupting their first molars earlier in life (Smith, CR 2000). The age of first molar eruption, moreover, has been shown to positively correlate with important life history traits such as longevity or female sexual maturity (Veitschegger US and Sánchez-Villagra, 2016). On the other hand, the onset of the third molar eruption has been shown to be correlated with skeletal maturity (Dean, 2006; Engström et al., 1983). AN Data relative to the weaning and maturity attainment of extant equids coincide with those M of the first and third molar eruption times (Table 4). This endorses our interpretations based on the hipparionin eruption times. In extant foals, the first molar starts to be formed ED close to the birth event, and begins mineralization two weeks after (Hoppe et al., 2004). Taking into account the growth curve reconstruction, we identified the emergence time of PT extant Equus molars (Nacarino-Meneses et al., 2017) at ~230 days (Fig. 4), which agrees CE quite well with that reported in the literature (Smuts, 1974). Assuming a similar age of the first molar formation onset in hipparionins, we can infer later eruption times compared to AC extant equids. By this, we roughly estimate the weaning age of the hipparionins analyzed in a span from 300 to 350 days of age (10-12 months). The absolute eruption age of the third molar is difficult to establish because we do not know at which age hipparionins started to form the third molar. However, the hipparionin tooth mineralization pattern has shown to be analogous to that of extant Equus (Domingo et al., 2018). From the growth curves, we estimated a similar third molar ACCEPTED MANUSCRIPT eruption time of ~500 days after the onset of the formation in all analyzed taxa. If we conservatively assume that hipparionin third molars begin to form as in Equus, where third molar mineralization starts at the age of 21 months (Hoppe et al., 2004), the hipparionin third molars should have erupted around the age of ~37.5 months. Considering the allometric coupling between size and life history, we would expect that the smaller species followed faster life history strategies and, thus, they would have matured earlier (Calder, T IP 1984). However, we observe how hipparionins, and especially the dwarf Spanish species, CR had a similar maturity age to extant equids. In this sense, we could infer a slower lifestyle in hipparionin horses bearing in mind the similar timing despite the smaller size (Sibly and US Brown, 2007). Our results on weaning and maturity schedules, together with the hipparionin smaller size compared to extant equids, further support our previous AN interpretation of slower life histories. Conforming to the Schultz’s Rule (Smith, 2000), moreover, the hipparionin eruption sequence have already indicated a slow-growing M strategy (Domingo et al., 2018). Besides the slower pace of life history compared to ED Equus, our results also suggest that the dwarf taxa from the Spanish late Turolian followed slower lifestyles than the larger middle Turolian species; because although they were PT smaller, which predicts faster life history from body mass scaling (Calder, 1984), they CE followed similar life history schedules. 4.4 Hypsodonty and life history interpretations AC The more hypsodont Equus molars erupt earlier compared to those of hipparionins, probably related to a faster ontogenetic development. In a similar way, the South American notoungulates with ever-growing teeth exhibited an earlier eruption of permanent molars compared to premolars, which has also been linked to an accelerated life history (Gomes Rodrigues et al., 2017). The higher wear exposure due to earlier eruption of the permanent molars would have been compensated by the observed extended dental growth span and ACCEPTED MANUSCRIPT a late root formation. After the earlier Equus molar eruption, therefore, the subsequent slower growth and extended formation time helps to counterbalance the height loss due to wear (Nacarino-Meneses et al., 2017). Moreover, the higher EERs related to earlier formation, and the extended crown formation times in Equus, may be involved in the higher hypsodonty indices exhibited within the genus. As a result, extant equids could have increased the endurance of their masticatory apparatus and, thus, incremented their T IP potential life span (Damuth and Janis, 2011; Jordana et al., 2012; Veiberg et al., 2007). CR Hipparionins, on the other hand, would have erupted their permanent molars later than extant equids, thus having a prolonged potential tooth life span as teeth began to wear US later. Moreover, instead of increased crown heights, hipparionin horses relied on occlusal enamel-band complexity as a way to expand their tooth durability (Famoso et al., 2016; AN Famoso and Davis, 2016). M Previous inferences on the life history of the Hipparionini tribe suggested an earlier maturity attainment and shorter life span compared to modern equids (Hulbert, 1982; ED Kurtén, 1953; Van Valen, 1964). These first surveys were based on age-structure and population dynamics studied from attritional equid fossil populations by means of their PT dentition ontogeny (Woodburne and MacFadden, 1982). Therefore, to define and CE characterize discrete age stages for the reconstruction of the demographic profile, these studies depended on models of age determination for extant species and their eruption AC patterns (Kurtén, 1953; Spinage, 1972). For this reason, the inferences of accelerated life history in hipparionins, and especially the interpretations of potential life span, strongly relied on the lower crown height of these taxa compared to present-day equids (Hulbert, 1982) and on the smaller size of some species (Van Valen, 1964). However, as we have seen, hipparionins do not follow the same formation, nor eruption pattern as extant equids (Domingo et al., 2018), and probably they also had different wear rates (Hulbert, 1984) ACCEPTED MANUSCRIPT due to their different dietary niches (Semprebon et al., 2016). Hence, there are arguments in support of the thesis that hipparionins were slow-growing mammals, as has also been recently proposed from their molar eruption pattern based on the Schultz’s Rule (Domingo et al., 2018). More recent approaches using bone paleohistological analyses have provided a new T framework to test for the life history adaptations of this diverse and successful group IP (Martinez-Maza et al., 2014; Orlandi-Oliveras et al., 2018). Thus, the study of the bone CR histology of Hipparion concudense has provided evidence of the attainment of skeletal maturity during the third year of life (Martinez-Maza et al., 2014), coinciding with our US inferred age of third molar eruption. Although Martinez-Maza et al. (2014) pointed out similar life history patterns between H. concudense and extant equids, they suggested that AN H. concudense developed over a shorter time span. Their interpretation was based on an M estimated age of sexual maturity at three years old in H. concudense, while most extant equids reach reproductive maturity during or after the fourth year of life (Martinez-Maza et ED al., 2014). However, the identification of absolute sexual maturity attainment from bone histology analysis is an issue still under debate (Marín-Moratalla et al., 2013; Nacarino- PT Meneses et al., 2016), as Martinez-Maza et al. (2014) already indicated, and the results CE should be considered carefully. Furthermore, wild equid populations can exhibit a wide range of variability in sexual maturity attainment since there are many possible factors AC influencing it (Choquenot, 1991; Penzhorn and van der Merwe, 1988; Smuts, 1976a, 1976b; Westlin-van Aarde et al., 1988). The dental histology parameters and eruption times calculated here, however, point towards a slower life strategy in hipparionins compared to extant equids. Considering their size, the Spanish dwarf taxa would have followed an even slower lifestyle than the larger hipparionins, as was also proposed by Orlandi-Oliveras et al. (2018) from the study of H. gromovae bone histology. This slower ACCEPTED MANUSCRIPT pace of growth has been related to a different ecology of the smaller species compared to the larger hipparionin forms (Orlandi-Oliveras et al., 2018). In African bovids, Macho and Williamson (2002) also found that the small-sized taxa tend to follow slower life history strategies due to their relatively longer gestation lengths. In this case, small ruminants showed different habitat preferences than the larger ones, browsing in more wooded habitats. Future paleodietary reconstructions using micro or mesowear techniques T IP (Calandra and Merceron, 2016; Kaiser and Solounias, 2003) on the teeth of these dwarf CR hipparionins could shed light on their diet and habitat preferences. 5. Conclusions US Our analyses of the dental histology of hipparionin teeth allow the characterization of their growth and comparisons both between different-sized taxa and with an extant equid AN model. In all cases, hipparionin teeth grow and develop differently from those of the genus M Equus. Variations are also found between hipparionin species and between the two tooth types analyzed within each taxon. We have associated the higher enamel secretion and ED extension rates of the first/second molars compared to the third molars with differences in PT the formation timing, although the variation in the secretion rates could also be influenced by the crown height or other characteristics related to tooth morphology. The higher CE enamel extension rates of the first/second molars are a reflection of the fast somatic growth of early ontogenetic stages, while the third molars are formed close to maturity AC attainment. Moreover, third molars generally take more time to be formed and to erupt than first/second molars, which is probably linked to their lower extension rates. Regarding the differences between groups, we found lower daily secretion rates in hipparionin teeth in comparison to E. quagga. We hypothesize that these distinct secretion rates, together with the disparity between hipparionin species, are the consequence of the interplay of ACCEPTED MANUSCRIPT numerous factors, such as hypsodonty, enamel thickness, secretory ameloblast lifespan, and overall tooth morphology. Due to the link between pace of life history and dental growth development, we could also draw inferences on the life history of the studied taxa. The parameters that are more related to the pace of growth are the enamel extension rate and the formation time. The T formation and eruption of first and third molars are further linked to key life history events IP such as weaning and maturity attainment, respectively. As a result, we were able to CR compare the hipparionin life history strategies with an extant equid model, and also to make comparisons between the two dwarf forms and the larger one. Hipparionin teeth US grow more slowly and erupt relatively later than those of extant equids, suggesting a slower pace of growth. In addition, the dwarf hipparionins exhibit lower extension rates but AN similar formation and eruption schedules than the larger species, which considering their M smaller size indicates that they had slower life histories. These results point toward a slow- growing strategy in hipparionin horses, as has recently been suggested from dental ED eruption patterns (Domingo et al., 2018), and debunk previous hypotheses of general shorter life spans and earlier maturity. Dwarfed hipparionin species from the late Turolian PT of Spain, moreover, do not follow the size scaling of the life history schedule but show a CE similar life history strategy as larger hipparionins, which is slower than predicted from the body mass. AC Acknowledgments We would like to thank T. Kaiser (Centrum für Naturkunde, Universität Hamburg) and J. Quesada (Museu de Ciències Naturals de Barcelona) for their permission to analyze extant Equus material from their respective institutions. B. Lamglait (currently at Université de Montréal) together with the Réserve Africaine de Sigean are acknowledged for the donation of extant equid specimens. We also acknowledge M. Soteras March for the ACCEPTED MANUSCRIPT access to the museum collection of the Institut Català de Paleontologia Miquel Crusafont (ICP), M. Fernández (ICP) for the slide preparation, the ICP’s preparation team for the molding of the specimens, R. Molonwy (Centre de Recerca Ecològica i Aplicacions Forestals) for the statistical advice received, and X. Jordana (Universitat Autònoma de Barcelona) for the help during the choice of the material and the discussion of the first analyses. We thank the reviewers and editors for their helpful comments. Funding for this T IP study was provided by the Ministerio de Economía, Industria y Competitividad (MINECO), CR Gobierno de España (PI: M.K., CGL2015-63777), and the CERCA Programme, Generalitat de Catalunya. The Paleobiology research group (ICP) is recognized without US funding by AGAUR (2017 SGR 960). G.O.-O. was supported by a FI-DGR 2016 grant (2016FI_B00202) awarded by AGAUR (Generalitat de Catalunya). The support of the AN DST-NFR Centre of Excellence in Palaeosciences (CoE-Pal) towards this research for C.N.-M. (COE2019-PD03) is hereby acknowledged. Opinions expressed and conclusions M arrived at, are those of the author and are not necessarily to be attributed to the CoE-Pal. ED Data availability PT Relevant data is shown within the paper and in Supplementary Material. Raw data related to this article can be found at http://dx.doi.org/10.17632/h2ywzbcs59.1, an open- source online data repository hosted at Mendeley Data. CE AC References Alberdi, M.T., 1989. A review of Old World hipparionine horses, in: Prothero, D.R., Schoch, R.M. (Eds.), The Evolution of Perissodactyls. Oxford University Press, New York, pp. 234–261. Alberdi, M.T., 1974. Las “faunas de Hipparion” de los yacimientos españoles. Estud. Geológicos 30, 189–212. Alberdi, M.T., Alcalá, L., 1989. El género Hipparion en la fosa de Alfambra-Teruel. Paleontol. i Evol. 105–109. Alberdi, M.T., Prado, J.L., Ortiz-Jaureguizar, E., 1995. Patterns of body size changes in ACCEPTED MANUSCRIPT fossil and living Equini (Perissodactyla). Biol. J. Linn. Soc. 54, 349–370. doi:10.1111/j.1095-8312.1995.tb01042.x Bendrey, R., Vella, D., Zazzo, A., Balasse, M., Lepetz, S., 2015. Exponentially decreasing tooth growth rate in horse teeth: Implications for isotopic analyses. Archaeometry 57, 1104–1124. doi:10.1111/arcm.12151 Bernor, R.L., Koufos, G.D., Woodburne, M.O., Fortelius, M., 1996. The evolutionary history and biochronology of European and Southwest Asian Late Miocene and Pliocene hipparionine horses, in: Bernor, R.L., Fahlbusch, V., Mittmann, H.-W. (Eds.), The Evolution of Western Eurasian Neogene Mammal Faunas. Columbia University T Press, New York, pp. 307–338. IP Bernor, R.L., Tobien, H., Woodburne, M.O., 1990. Patterns of Old World hipparionine evolutionary diversification and biogeographic extension, in: Lindsay, E.H., Fahlbusch, V., Mein, P. (Eds.), European Neogene Mammal Chronology. Plenum CR Press, New York, pp. 263–319. doi:10.1007/978-1-4899-2513-8_18 Bromage, T.G., Idaghdour, Y., Lacruz, R.S., Crenshaw, T.D., Ovsiy, O., Rotter, B., Hoffmeier, K., Schrenk, F., 2016. The swine plasma metabolome chronicles “many US days” biological timing and functions linked to growth. PLoS One 11, 1–19. doi:10.1371/journal.pone.0145919 AN Bromage, T.G., Lacruz, R.S., Hogg, R.T., Goldman, H.M., McFarlin, S.C., Warshaw, J., Dirks, W., Perez-Ochoa, A., Smolyar, I., Enlow, D.H., Boyde, A., 2009. Lamellar bone is an incremental tissue reconciling enamel rhythms, body size, and organismal life history. Calcif. Tissue Int. 84, 388–404. doi:10.1007/s00223-009-9221-2 M Calandra, I., Merceron, G., 2016. Dental microwear texture analysis in mammalian ecology. Mamm. Rev. 46, 215-218. doi:10.1111/mam.12063 ED Calder, W.A., 1984. Size, Function, and Life History. Harvard University Press, Cambridge. doi:/10.2307/4086807 PT Cantalapiedra, J.L., Prado, J.L., Hernández Fernández, M., Alberdi, M.T., 2017. Decoupled ecomorphological evolution and diversification in Neogene-Quaternary horses. Science 355, 627–630. doi:10.1126/science.aag1772 CE Choquenot, D., 1991. Density-dependent growth, body condition, and demography in feral donkeys: testing the food hypothesis. Ecology 72, 805–813. doi:10.2307/1940583 Damuth, J.D., Janis, C.M., 2011. On the relationship between hypsodonty and feeding AC ecology in ungulate mammals, and its utility in palaeoecology. Biol. Rev. 86, 733– 758. doi:10.1111/j.1469-185X.2011.00176.x de Bruijn, H., Daams, R., Daxner-Höck, G., Fahlbusch, V., Ginsburg, L., Mein, P., Morales, J., Heinzmann, E., Mayhew, D.F., van der Meulen, a J., Schmidt Kittler, N., Telles Antunes, M., 1992. Report of the RCMNS working group on fossil mammals, Reisensburg 1990. Newsletters Stratigr. 26, 65–118. doi:10.1127/nos/26/1992/65 Dean, M.C., 2006. Tooth microstructure tracks the pace of human life-history evolution. Proc. Biol. Sci. 273, 2799–2808. doi:10.1098/rspb.2006.3583 Dean, M.C., 2000. Incremental markings in enamel and dentine: what they can tell us about the way teeth grow, in: Teaford, M.F., Meredith Smith, M., Ferguson, M.W.J. (Eds.), Development, Function and Evolution of Teeth. Cambridge University Press, ACCEPTED MANUSCRIPT Cambridge, pp. 119–130. doi:10.1017/CBO9780511542626.009 Dean, M.C., Leakey, M.G., Reid, D.J., Schrenk, F., Schwartz, G.T., Stringer, C., Walker, A., 2001. Growth processes in teeth distinguish modern humans from Homo erectus and earlier hominins. Nature 414, 628–631. doi:10.1038/414628a Dirks, W., Anemone, R.L., Holroyd, P.A., Reid, D.J., Walton, P., 2009. Phylogeny, life history and the timing of molar crown formation in two archaic ungulates, Meniscotherium and Phenacodus (Mammalia, ’Condylarthra’). Front. Oral Biol. 13, 3– 8. doi:10.1159/000242381 Dirks, W., Bowman, J.E., 2007. Life history theory and dental development in four species T of catarrhine primates. J. Hum. Evol. 53, 309–320. doi:10.1016/j.jhevol.2007.04.007 IP Dirks, W., Bromage, T.G., Agenbroad, L.D., 2012. The duration and rate of molar plate formation in Palaeoloxodon cypriotes and Mammuthus columbi from dental histology. CR Quat. Int. 255, 79–85. doi:10.1016/j.quaint.2011.11.002 Domingo, L., Grimes, S.T., Domingo, M.S., Alberdi, M.T., 2009. Paleoenvironmental conditions in the Spanish Miocene-Pliocene boundary: isotopic analyses of Hipparion US dental enamel. Naturwissenschaften 96, 503–511. doi:10.1007/s00114-008-0500-y Domingo, M.S., Cantero, E., García-Real, I., Chamorro Sancho, M.J., Martín Perea, D.M., Alberdi, M.T., Morales, J., 2018. First radiological study of a complete dental AN ontogeny sequence of an extinct equid: implications for equidae life history and taphonomy. Sci. Rep. 8, 8507. doi:10.1038/s41598-018-26817-3 Easley, J., Dixon, P.M., Schumacher, J., 2005. Equine Dentistry, Third. ed. Elsevier. M doi:10.1016/B0-7020-2724-3/X5001-4 Eisenmann, V., 1995. What metapodial morphometry has to say about some Miocene ED hipparions, in: Vrba, E.S., Denton, G.H., Partridge, T., Burckle, L.H. (Eds.), Paleoclimate and Evolution, with Emphasis on Human Origins. New Haven and London, pp. 149–164. PT Engström, C., Engström, H., Sagne, S., 1983. Lower third molar development in relation to skeletal maturity and chronological age. Angle Orthod. 53, 97–106. doi:10.1043/0003- 3219(1983)053<0097:LTMDIR>2.0.CO;2 CE Ernest, S.K.M., 2003. Life history characteristics of placental nonvolant mammals. Ecology 84, 3402–3402. doi:10.1890/02-9002 AC Famoso, N.A., Davis, E.B., 2016. On the relationship between enamel band complexity and occlusal surface area in Equids (Mammalia, Perissodactyla). PeerJ 4, e2181. doi:10.7717/peerj.2181 Famoso, N.A., Davis, E.B., Feranec, R.S., Hopkins, S.S.B., Price, S.A., 2016. Are hypsodonty and occlusal enamel complexity evolutionarily correlated in ungulates? J. Mamm. Evol. 23, 43–47. doi:10.1007/s10914-015-9296-7 Famoso, N.A., Feranec, R.S., Davis, E.B., 2013. Occlusal enamel complexity and its implications for lophodonty, hypsodony, body mass, and diet in extinct and extant ungulates. Palaeogeogr. Palaeoclimatol. Palaeoecol. 387, 211–216. doi:10.1016/j.palaeo.2013.07.006 Forsten, A., 1978. Hipparion and possible Iberian-North African Neogene connections. ACCEPTED MANUSCRIPT Ann. Zool. Fennici 15, 294–297. Forsten, A., 1968. Revision of the palearctic Hipparion, Acta Zoologica Fennica 119, 1- 134. Garcés, M., Cabrera, L., Agustí, J., Maria Parés, J., 1997. Old World first appearance datum of “Hipparion” horses: Late Miocene large-mammal dispersal and global events. Geology 25, 19. doi:10.1130/0091-7613(1997)025<0019:OWFADO>2.3.CO;2 Garcés, M., Krijgsman, W., Peláez-Campomanes, P., Álva, M.A., Daams, R., 2003. Hipparion dispersal in Europe : magnetostratigraphic constraints from the Daroca area (Spain). Coloquios Paleontol. 1, 171–178. T Gomes Rodrigues, H., Herrel, A., Billet, G., 2017. Ontogenetic and life history trait IP changes associated with convergent ecological specializations in extinct ungulate mammals. Proc. Natl. Acad. Sci. 114, 1069–1074. doi:10.1073/pnas.1614029114 CR Grubb, P., 1981. Equus burchelli. Mamm. Species 157, 1. doi:10.2307/3503962 Guatelli-Steinberg, D., Floyd, B.A., Dean, M.C., Reid, D.J., 2012. Enamel extension rate US patterns in modern human teeth: Two approaches designed to establish an integrated comparative context for fossil primates. J. Hum. Evol. 63, 475–486. doi:10.1016/j.jhevol.2012.05.006 AN Hammer, Ø., Harper, D.A.T. a. T., Ryan, P.D., 2001. PAST: paleontological statistics software package for education and data analysis. Palaeontol. Electron. 4(1), 1–9. doi:10.1016/j.bcp.2008.05.025 M Hayek, L.-A.C., Bernor, R.L., Solounias, N., Steigerwald, P., 1992. Preliminary studies of hipparionine horse diet as measured by tooth microwear. Ann. Zool. Fennici 28, 187– 200. ED Hillson, S., 2005. Teeth. Cambridge University Press. Hogg, R.T., Walker, R.S., 2011. Life-history correlates of enamel microstructure in cebidae PT (platyrrhini, primates). Anat. Rec. 294, 2193–2206. doi:10.1002/ar.21503 Hoppe, K.A., Stover, S.M., Pascoe, J.R., Amundson, R., 2004. Tooth enamel biomineralization in extant horses: Implications for isotopic microsampling. CE Palaeogeogr. Palaeoclimatol. Palaeoecol. 206, 355–365. doi:10.1016/j.palaeo.2004.01.012 AC Hulbert, R.C., 1984. Paleoecology and population dynamics of the early Miocene (Hemingfordian) horse Parahippus leonensis from the Thomas Farm site, Florida. J. Vertebr. Paleontol. 4, 547–558. doi:10.1080/02724634.1984.10012030 Hulbert, R.C., 1982. Population dynamics of the three-toed horse Neohipparion from the late Miocene of Florida. Paleobiology 8, 159–167. doi:10.2307/2400452 Jordana, X., Köhler, M., 2011. Enamel microstructure in the fossil bovid Myotragus balearicus (Majorca, Spain): Implications for life-history evolution of dwarf mammals in insular ecosystems. Palaeogeogr. Palaeoclimatol. Palaeoecol. 300, 59–66. doi:10.1016/j.palaeo.2010.12.008 Jordana, X., Marín-Moratalla, N., DeMiguel, D., Kaiser, T.M., Köhler, M., 2012. Evidence of correlated evolution of hypsodonty and exceptional longevity in endemic insular ACCEPTED MANUSCRIPT mammals. Proc. R. Soc. B Biol. Sci. 279, 3339–3346. doi:10.1098/rspb.2012.0689 Jordana, X., Marín-Moratalla, N., Moncunill-Solé, B., Köhler, M., 2014. Ecological and life- history correlates of enamel growth in ruminants (Artiodactyla). Biol. J. Linn. Soc. 112, 657–667. doi:10.1111/bij.12264 Joubert, E., 1972. Tooth development and age determination in the Hartmann zebra Equus zebra hartmannae. Madoqua 5–16. Kaiser, T.M., Solounias, N., 2003. Extending the tooth mesowear method to extinct and extant equids. Geodiversitas 25, 321–345. T Kierdorf, H., Breuer, F., Richards, A., Kierdorf, U., 2014. Characterization of enamel incremental markings and crown growth parameters in minipig molars. Anat. Rec. IP 297, 1935–1949. doi:10.1002/ar.22951 CR Kierdorf, H., Breuer, F., Witzel, C., Kierdorf, U., 2019. Pig enamel revisited – Incremental markings in enamel of wild boars and domestic pigs. J. Struct. Biol. 205, 48-59. doi:10.1016/J.JSB.2018.11.009 US Kierdorf, H., Kierdorf, U., Frölich, K., Witzel, C., 2013. Lines of evidence-incremental markings in molar enamel of soay sheep as revealed by a fluorochrome labeling and backscattered electron imaging study. PLoS One 8, e74597. doi:10.1371/journal.pone.0074597 AN Kirkland, K.D., Baker, G.J., Marretta, S.M., Eurell, J.A.C., Losonsky, J.M., 1996. Effects of aging on the endodontic system, reserve crown, and roots of equine mandibular cheek teeth. Am. J. Vet. Res. 57, 31–38. M Köhler, M., 2010. Fast or slow? The evolution of life history traits associated with insular dwarfing, in: Pérez-Mellado, V., Ramon, Ma.M. (Eds.), Islands and Evolution. Institut ED Menorquí d’Estudis, Recerca, pp. 261–280. Köhler, M., Moyà-Solà, S., 2009. Physiological and life history strategies of a fossil large mammal in a resource-limited environment. Proc. Natl. Acad. Sci. U. S. A. 106, PT 20354–8. doi:10.1073/pnas.0813385106 Kurtén, B., 1953. On the variation and population dynamics of fossil and recent mammal CE populations. Acta Zool. Fenn. 76, 1–122. Lkhagvasuren, D., Ansorge, H., Samiya, R., Schafberg, R., Stubbe, A., Stubbe, M., 2013. Age determination of the Mongolian wild ass (Equus hemionus Pallas, 1775) by the AC dentition patterns and annual lines in the tooth cementum. J. Species Res. 2, 85–90. doi:10.12651/JSR.2013.2.1.085 MacFadden, B.J., Solounias, N., Cerling, T.E., 1999. Ancient Diets, Ecology, and Extinction of 5-Million-Year-Old Horses from Florida. Science 283, 824–827. doi:10.1126/science.283.5403.824 Macho, G.A., Williamson, D.K., 2002. The effects of ecology on life history strategies and metabolic disturbances during development: An example from the African bovids. Biol. J. Linn. Soc. 75, 271–279. doi:10.1046/j.1095-8312.2002.00013.x Marín-Moratalla, N., Jordana, X., Köhler, M., 2013. Bone histology as an approach to providing data on certain key life history traits in mammals: Implications for conservation biology. Mamm. Biol. 78, 422–429. doi:10.1016/j.mambio.2013.07.079 ACCEPTED MANUSCRIPT Martinez-Maza, C., Alberdi, M.T., Nieto-Diaz, M., Prado, J.L., 2014. Life-history traits of the miocene Hipparion concudense (Spain) inferred from bone histological structure. PLoS One 9, e103708. doi:10.1371/journal.pone.0103708 Matson, S.D., Fox, D.L., 2010. Stable isotopic evidence for terrestrial latitudinal climate gradients in the Late Miocene of the Iberian Peninsula. Palaeogeogr. Palaeoclimatol. Palaeoecol. 287, 28–44. doi:10.1016/j.palaeo.2009.12.010 McDonald, H.G., 1996. Population structure of the late Pliocene (Blancan) zebra Equus simplicidens (Perissodactyla: Equidae) from the Hagerman Horse Quarry, Idaho, in: Stewart, K.., Seymour, K.L. (Eds.), Paleoecology and Paleoenvironments of Late T Cenozoic Mammals. University of Toronto Press, Toronto, pp. 134–155. IP Metcalfe, J.Z., Longstaffe, F.J., 2012. Mammoth tooth enamel growth rates inferred from stable isotope analysis and histology. Quat. Res. 77, 424–432. doi:10.1016/j.yqres.2012.02.002 CR Monfort, S.L., Arthur, N.P., Wildt, D.E., 1994. Reproduction in the Przewalski’s horse, in: Boyd, L., Houpt, K.A. (Eds.), Przewalski’s Horse: The History and Biology of an Endangered Species. State University of New York Press, Albany, pp. 173–194. US Monson, T.A., Hlusko, L.J., 2018a. Breaking the rules: Phylogeny, not life history, explains dental eruption sequence in primates. Am. J. Phys. Anthropol. 167, 217–233. AN doi:10.1002/ajpa.23618 Monson, T.A., Hlusko, L.J., 2018b. The evolution of dental eruption sequence in artiodactyls. J. Mamm. Evol. 25, 15–26. doi:10.1007/s10914-016-9362-9 M Nacarino-Meneses, C., Jordana, X., Köhler, M., 2016. Histological variability in the limb bones of the Asiatic wild ass and its significance for life history inferences. PeerJ 4, ED e2580. doi:10.7717/peerj.2580 Nacarino-Meneses, C., Jordana, X., Orlandi-Oliveras, G., Köhler, M., 2017. Reconstructing molar growth from enamel histology in extant and extinct Equus. Sci. Rep. 7, 15965. PT doi:10.1038/s41598-017-16227-2 Nelson, S. V., 2005. Paleoseasonality inferred from equid teeth and intra-tooth isotopic variability. Palaeogeogr. Palaeoclimatol. Palaeoecol. 222, 122–144. CE doi:10.1016/j.palaeo.2005.03.012 O’Meara, R.N., Dirks, W., Martinelli, A.G., 2018. Enamel formation and growth in non- mammalian cynodonts. R. Soc. Open Sci. 5, 172293. doi:10.1098/rsos.172293 AC Orlandi-Oliveras, G., Nacarino-Meneses, C., Koufos, G.D., Köhler, M., 2018. Bone histology provides insights into the life history mechanisms underlying dwarfing in hipparionins. Sci. Rep. 8, 17203. doi:10.1038/s41598-018-35347-x Ortiz-Jaureguizar, E., Alberdi, M.T., 2003. El patrón de cambios en la masa corporal de los Hipparionini (Perissodactyla, Equidae) de la Península Ibérica durante el Mioceno superior-Plioceno superior. Coloquios Paleontol. 1, 499–509. Palkovacs, E.P., 2003. Explaining adaptive shifts in body size on islands: a life history approach. Oikos 103, 37–44. doi:10.1034/j.1600-0706.2003.12502.x Penzhorn, B.L., 1982. Home range sizes of Cape mountain zebras Equus zebra zebra in the Mountain Zebra National Park. Koedoe 25, 103–108. ACCEPTED MANUSCRIPT doi:10.4102/koedoe.v25i1.607 Penzhorn, B.L., van der Merwe, 1988. Testis size and onset of spermatogenesis in Cape mountain zebra (Equus zebra zebra). J. Reprod. Fertil. 83, 371–375. doi:10.1530/jrf.0.0830371 Pesquero, M.D., 2003. Hipparion del Turoliense superior de Las Casiones (Fosa de Teruel). Coloquios Paleontol. Volumen Ex, 511–548. Pesquero, M.D., Alberdi, M.T., 2012. New evidence of conspecificity between Hipparion primigenium melendezi Alberdi, 1974 from Los Valles de Fuentidueña (Segovia) and Hipparion concudense concudense Pirlot, 1956 from Concud (Teruel) Spain. Estud. T Geológicos 68, 247–260. doi:10.3989/egeol.40499.152 IP Pesquero, M.D., Ascaso, C., Alcalá, L., Fernández-Jalvo, Y., 2010. A new taphonomic bioerosion in a Miocene lakeshore environment. Palaeogeogr. Palaeoclimatol. CR Palaeoecol. 295, 192–198. doi:10.1016/j.palaeo.2010.05.037 Pluháček, J., Bartoš, L., Doležalová, M., Bartošová-Víchová, J., 2007. Sex of the foetus determines the time of weaning of the previous offspring of captive plains zebra US (Equus burchelli). Appl. Anim. Behav. Sci. 105, 192–204. doi:10.1016/j.applanim.2006.05.019 Rey, K., Amiot, R., Lécuyer, C., Koufos, G.D., Martineau, F., Fourel, F., Kostopoulos, D.S., AN Merceron, G., 2013. Late Miocene climatic and environmental variations in northern Greece inferred from stable isotope compositions (δ18O, δ13C) of equid teeth apatite. Palaeogeogr. Palaeoclimatol. Palaeoecol. 388, 48–57. M doi:10.1016/j.palaeo.2013.07.021 Schwartz, G.T., Dean, M.C., 2001. Ontogeny of canine dimorphism in extant hominoids. ED Am. J. Phys. Anthropol. 115, 269–283. doi:10.1002/ajpa.1081 Schwartz, G.T., Samonds, K.E., Godfrey, L.R., Jungers, W.L., Simons, E.L., 2002. Dental microstructure and life history in subfossil Malagasy lemurs. Proc. Natl. Acad. Sci. U. PT S. A. 99, 6124–9. doi:10.1073/pnas.092685099 Semprebon, G.M., Rivals, F., Solounias, N., Hulbert, R.C., 2016. Paleodietary reconstruction of fossil horses from the Eocene through Pleistocene of North CE America. Palaeogeogr. Palaeoclimatol. Palaeoecol. 442, 110–127. doi:10.1016/j.palaeo.2015.11.004 Shellis, R.P., 1984. Variations in growth of the enamel crown in human teeth and a AC possible relationship between growth and enamel structure. Arch. Oral Biol. 29, 697– 705. doi:10.1016/0003-9969(84)90175-4 Sibly, R.M., Brown, J.H., 2007. Effects of body size and lifestyle on evolution of mammal life histories. Proc. Natl. Acad. Sci. 104, 17707–17712. doi:10.1073/pnas.0707725104 Silver, I.A., 1969. The ageing of domestic animals, in: Brothweel, D., Higgs, E. (Eds.), Science in Archaeology. New York, pp. 250–268. Smith, B.H., 2000. ‘Schultz’s Rule’ and the evolution of tooth emergence and replacement patterns in primates and ungulates, in: Teaford, M.F., Meredith Smith, M., Ferguson, M.W.J. (Eds.), Development, Function and Evolution of Teeth. Cambridge University Press, Cambridge, pp. 212–228. doi:10.1017/CBO9780511542626.015 ACCEPTED MANUSCRIPT Smith, B.H., 1991. Dental development and the evolution of life history in Hominidae. Am. J. Phys. Anthropol. 86, 157–174. doi:10.1002/ajpa.1330860206 Smith, T.M., 2016. Dental development in living and fossil orangutans. J. Hum. Evol. 94, 92–105. doi:10.1016/j.jhevol.2016.02.008 Smith, T.M., 2008. Incremental dental development: Methods and applications in hominoid evolutionary studies. J. Hum. Evol. 54, 205–224. doi:10.1016/j.jhevol.2007.09.020 Smith, T.M., 2006. Experimental determination of the periodicity of incremental features in enamel. J. Anat. 208, 99–113. doi:10.1111/j.1469-7580.2006.00499.x T Smith, T.M., Reid, D.J., Dean, M.C., Olejniczak, A.J., Martin, L.B., 2007. Molar development in common chimpanzees (Pan troglodytes). J. Hum. Evol. 52, 201–216. IP doi:10.1016/j.jhevol.2006.09.004 CR Smith, T.M., Tafforeau, P., Reid, D.J., Pouech, J., Lazzari, V., Zermeno, J.P., Guatelli- Steinberg, D., Olejniczak, A.J., Hoffman, A., Radovcic, J., Makaremi, M., Toussaint, M., Stringer, C., Hublin, J.-J., 2010. Dental evidence for ontogenetic differences between modern humans and Neanderthals. Proc. Natl. Acad. Sci. 107, 20923– US 20928. doi:10.1073/pnas.1010906107 Smuts, G.L., 1976a. Reproduction in the zebra mare Equus burchelli antiquorum from the Kruger National Park. Koedoe 19, 89–132. doi:10.4102/koedoe.v19i1.1186 AN Smuts, G.L., 1976b. Reproduction in the zebra stallion Equus burchelli antiquorum from the Kruger National Park South Africa. Zool. Africana 11, 207–220. doi:10.1080/00445096.1976.11447525 M Smuts, G.L., 1975. Pre-and postnatal growth phenomena of Burchell’s Zebra Equus burchelli antiquorum. Koedoe 18, 69–102. doi:10.4102/koedoe.v18i1.915 ED Smuts, G.L., 1974. Age determination in Burchell’s zebra (Equus burchelli antiquorum) from the Kruger National Park. South African J. Wildl. Res. 4, 103–115. PT Sondaar, P.Y., 1961. Les Hipparion de l’Aragon meridional. Estud. Geológicos 17, 209– 305. Spinage, C.A., 1972. Age estimation of zebra. Afr. J. Ecol. 10, 273–277. CE doi:10.1111/j.1365-2028.1972.tb00871.x Stearns, S.C., 1992. The Evolution of Life Histories. Oxford University Press, New York. AC Stirton, R.A., 1941. Development of Characters in Horse Teeth and the Dental Nomenclature. J. Mammal. 22, 434–446. doi:10.2307/1374940 Symons, A.L., Seymour, G.J., 2000. A histological study of the effect of growth hormone on odontogenesis in the Lewis dwarf rat. Arch. Oral Biol. 45, 123–131. doi:10.1016/S0003-9969(99)00123-5 Tafforeau, P., Bentaleb, I., Jaeger, J.-J., Martin, C., 2007. Nature of laminations and mineralization in rhinoceros enamel using histology and X-ray synchrotron microtomography: Potential implications for palaeoenvironmental isotopic studies. Palaeogeogr. Palaeoclimatol. Palaeoecol. 246, 206–227. doi:10.1016/j.palaeo.2006.10.001 Tütken, T., Kaiser, T.M., Vennemann, T., Merceron, G., 2013. Opportunistic feeding ACCEPTED MANUSCRIPT strategy for the earliest old world hypsodont equids: evidence from stable isotope and dental wear proxies. PLoS One 8, e74463. doi:10.1371/journal.pone.0074463 Van Dam, J., Alcalá, L., Zarza, A.A., Calvo, J.P., Garcés, M., Krijgsman, W., 2001. The upper Miocene mammal record from the Teruel-Alfambra region (Spain). The MN system and continental stage/age concepts discussed. J. Vertebr. Paleontol. 21, 367– 385. doi:10.1671/0272-410.1671/0272-4634(2001)021[0367:TUMMRF]2.0.CO;2 Van Dam, J., Reichart, G.J., 2009. Oxygen and carbon isotope signatures in late Neogene horse teeth from Spain and application as temperature and seasonality proxies. Palaeogeogr. Palaeoclimatol. Palaeoecol. 274, 64–81. T doi:10.1016/j.palaeo.2008.12.022 IP Van Valen, L., 1964. Age in two fossil horse populations. Acta Zool. 45, 93–106. doi:10.1111/j.1463-6395.1964.tb00712.x CR Veiberg, V., Mysterud, A., Gaillard, J.M., Delorme, D., Van Laere, G., Klein, F., 2007. Bigger teeth for longer life? Longevity and molar height in two roe deer populations. Biol. Lett. 3, 268–270. doi:10.1098/rsbl.2006.0610 US Veitschegger, K., Sánchez-Villagra, M.R., 2016. Tooth eruption sequences in cervids and the effect of morphology, life history, and phylogeny. J. Mamm. Evol. 23, 251–263. doi:10.1007/s10914-015-9315-8 AN Waran, N.K., Clarke, N., Farnworth, M., 2008. The effects of weaning on the domestic horse (Equus caballus). Appl. Anim. Behav. Sci. 110, 42–57. doi:10.1016/j.applanim.2007.03.024 M Westlin-van Aarde, L.M., van Aarde, R.J., Skinner, J.D., 1988. Reproduction in female Hartmann’s zebra, Equus zebra hartmannae . J. Reprod. Fertil. 84, 505–511. ED doi:10.1530/jrf.0.0840505 Winkler, D.E., Kaiser, T.M., 2015. Uneven distribution of enamel in the tooth crown of a Plains Zebra (Equus quagga). PeerJ 3, e1002. doi:10.7717/peerj.1002 PT Witzel, C., Kierdorf, U., Frölich, K., Kierdorf, H., 2018. The pay-off of hypsodonty - timing and dynamics of crown growth and wear in molars of Soay sheep. BMC Evol. Biol. 18, 207. doi:10.1186/s12862-018-1332-9 CE Woodburne, M.O., 1989. Hipparion horses: a pattern of endemic evolution and intercontinental dispersal, in: Prothero, D.R., Schoch, R.M. (Eds.), The Evolution of Perissodactyls. Oxford University Press, New York, pp. 197–233. AC Woodburne, M.O., MacFadden, B.J., 1982. A reappraisal of the systematics, biogeography, and evolution of fossil horses. Paleobiology 8, 315–327. doi:10.1017/S0094837300007077 Young, W.G., 1995. Growth hormone and insulin-like growth factor-I in odontogenesis. Int. J. Dev. Biol. 39, 263–72. doi:10.1387/IJDB.7626416 Zhang, C.Z., Young, W.G., Li, H., Rolinson, S., Waters, M.J., 1992. Growth hormone regulates nucleolar organizer regions during odontogenesis in the rat. J. Oral Pathol. Med. 21, 395–400. doi:10.1111/j.1600-0714.1992.tb01026.x ACCEPTED MANUSCRIPT Table and Figure Captions Table 1. Teeth analyzed in the present study. Mean body mass (BM) and hypsodonty T IP indices (HI) are provided per each species. CR Body mass data from a Pesquero and Alberdi (2012), b Ortiz-Jaureguizar and Alberdi (2003), and c Cantalapiedra et al. (2017). Hypsodonty indices calculated for first/second US molars are taken from d Pesquero (2003) and c Cantalapiedra et al. (2017). e Hypsodonty indices for third molars were calculated from our sample using the less worn specimens. AN Table 2. Dental growth parameters – daily secretion rate (DSR) and enamel extension rate M (EER) – estimated through dental histology analyses. DSR values of Equus m3 are shown separately due to the significant differences in the DSR values within the different areas of ED Equus third molar. The division of Equus quagga molars for the EER considers the PT complete tooth development and corresponds to the different CDS defined by Nacarino- Meneses et al. (2017). CE Table 3. Crown heights through the enamel band and crown formation times of the most AC complete teeth calculated through the dental histology analysis. Equus quagga calculations (*) relative to the total reconstructed molar growth. Table 4. Eruption dates of first and third molars contrasted with life history information in two extant equid species. ________________________________________________________________________ Fig. 1. Map of the area around the city of Teruel (Aragon, Spain). The fossil sites of El ACCEPTED MANUSCRIPT Arquillo (1) and Concud (2) are indicated with a star. (single column). Fig. 2. A: Protoconid buccal enamel band of a Hipparion periafricanum first/second molar (IPS-91733) sectioned longitudinally and observed under polarized light. B: Detail of the enamel band depicted in (A) illustrating the method used to calculate the daily secretion rate (DSR). The distance between laminations through the prism path (dashed line) is T divided by the number of laminations (arrows). Dashed arrow represents the prisms’ IP direction. C: Detail of the cervical zone in (A), showing the methodology used for the CR enamel extension rate (EER) determination. “X” represents the increase in crown height between two incremental features (black lines) and “Y” indicates the distance following the US enamel prism between those two enamel marks. Black arrows point to Retzius lines that outcrop the outer enamel, and “EDJ” depicts the enamel dentine junction. (2 column AN image, color) M Fig. 3. Enamel extension rate values (EER) plotted against the crown height from the first formed zone (cuspal, left size) to the last formed region (cervical, right side). Left chart ED shows EERs for first/second molars and right chart for third molars. Note the different PT scales between the two charts. (2 column image) Fig. 4. Molar growth curves obtained from plotting the crown height through the enamel CE band (mm, y axis) against the crown formation time (days, x axis), from the cusp to the AC root area. Left plots show mean curves for first/second molars while right plots the third molar mean curves. Upper graphs represent the growth curves of all the species analyzed, while lower graphs only show hipparionin teeth. Equus growth curves obtained from the composition of differently worn teeth, which are represented by different grey scale dots. Note that the plot axes are not equally scaled. Dashed lines represent the 95% confidence intervals and the arrows the inferred eruption times. (1.5 column image) ACCEPTED MANUSCRIPT Fig. 5. Growth curves of two molars of comparable developmental stage (Equus quagga IPS-92341 and Hipparion concudense IPS-91763). The curve represents the growth from the cusp to the root. Scale shown in the teeth images correspond to 1 cm. (single column image, color) T IP CR US AN M ED PT CE AC ACCEPTED MANUSCRIPT TABLE 1 BM Species Tooth HI Code Locality Age Wear (kg) middle H. IPS- m1/m2 Concud Turolian Slight concudense 91762 a d (MN12) 160 2.8 middle H. IPS- m1/m2 Concud Turolian Slight - concudense 91763 (MN12) IPS- late Turolian T H. gromovae m1/m2 El Arquillo Slight b d 91731 (MN13) 59 2.9 IP IPS- late Turolian H. gromovae m1/m2 El Arquillo Slight 91732 (MN13) m1/m2 H. IPS- late Turolian CR m1/m2 El Arquillo Slight periafricanum b d 91733 (MN13) 29 3.2 H. IPS- late Turolian m1/m2 El Arquillo Slight + periafricanum 91734 (MN13) Equus IPS- US m1 Hagenbeck Zoo Extant Slight - quagga 92341 Equus IPS- m1 Hagenbeck Zoo Extant Null quagga c c 92342 250 4.4 AN Equus IPS- m1 Hagenbeck Zoo Extant Crypt quagga 92345 Equus IPS- m1 Hagenbeck Zoo Extant Moderate quagga 92346c M middle H. IPS- m3 Concud Turolian Null concudense 40644 (MN12) ED middle H. IPS- m3 Concud Turolian Slight concudense 40665 a e (MN12) 160 2.0 middle PT H. IPS- m3 Concud Turolian Slight concudense 40668 (MN12) middle H. IPS- CE m3 Concud Turolian Null concudense 87855 (MN12) IPS- late Turolian H. gromovae m3 El Arquillo Slight - 85506 (MN13) m3 AC IPS- late Turolian H. gromovae m3 El Arquillo Slight b e 85508 (MN13) 59 1.9 IPS- late Turolian H. gromovae m3 El Arquillo Slight + 86851 (MN13) IPS- late Turolian H. gromovae m3 El Arquillo Slight + 87856 (MN13) H. IPS- late Turolian m3 El Arquillo Slight - periafricanum 85504 (MN13) H. b e IPS- late Turolian m3 29 2.1 El Arquillo Slight + periafricanum 86850 (MN13) H. IPS- late Turolian m3 El Arquillo Moderate periafricanum 87857 (MN13) Equus c e IPS- Réserve m3 250 2.5 Extant Crypt quagga 104358 Africaine ACCEPTED MANUSCRIPT Sigean MZB- Equus m3 94- Barcelona Zoo Extant Slight + quagga 1229 Equus IPS- m3 Hagenbeck Zoo Extant Moderate quagga 92346e a b (footnote) Body mass data from Pesquero and Alberdi (2012), Ortiz-Jaureguizar and Alberdi (2003), and c Cantalapiedra et al. (2017). Hypsodonty indices calculated for first/second molars are taken from d Pesquero (2003) and c Cantalapiedra et al. (2017). e T IP Hypsodonty indices for third molars were calculated from our sample using the less worn specimens. CR TABLE 2 US Overall DSR Cuspal EER Middle EER Cervical EER Species (µm/day) (µm/day) (µm/day) (µm/day) AN H. 15.76 ±0.96 (N = 211.50 ± 27.25 177.08 ± 24.71 82.52 ± 38.96 (N concudense 90) (N = 6) (N = 6) = 9) M 16.27 ±1.15 (N = 144.73 ± 28.05 131.32 ± 29.63 55.75 ± 35.85 (N H. gromovae 104) (N = 7) (N = 6) = 16) m1/m2 ED H. 15.96 ±1.37 (N = 166.62 ± 33.08 121.14 ± 8.11 (N 37.27 ± 19.84 (N periafricanum 85) (N = 7) = 5) = 15) PT Equus 16.98 ±1.63 (N = 441.69 ± 152.37 316.18 ± 146.18 78.37 ± 73.25 (N quagga 45) (N = 10) (N = 13) = 30) CE H. 14.08 ±1.73 (N = 150.39 ± 45.50 95.54 ± 18.87 (N 41.30 ± 22.43 (N concudense 78) (N = 9) = 12) = 15) AC 15.37 ±1.78 (N = 119.09 ± 10.51 95.36 ± 14.25 (N 43.06 ± 20.59 (N H. gromovae 93) (N = 7) = 11) = 15) m3 H. 13.72 ±1.70 (N = 116.35 ± 18.17 75.14 ± 9.74 (N 43.19 ± 20.36 (N periafricanum 36) (N = 6) = 8) = 12) Cuspal: 18.11 ±1.02 Equus (N = 25) 199.42 ± 66.11 112.44 ± 23.99 25.00 ± 18.99 (N quagga Middle: 17.23 ±1.05 (N = 8) (N = 13) = 44) (N = 40) ACCEPTED MANUSCRIPT Cervical: 15.92 ±1.78 (N = 73) TABLE 3 Crown height Species Tooth Code CFT (days) (mm) H. concudense m1/m2 IPS-91762 64.47 523 H. gromovae m1/m2 IPS-91731 53.39 610 T m1/m2 H. gromovae m1/m2 IPS-91732 49.97 511 IP H. periafricanum m1/m2 IPS-91733 44.10 504 H. periafricanum m1/m2 IPS-91734 45.29 498 CR Equus quagga m1 Composition 125.94* 785* H. concudense m3 IPS-87855 57.38 701 H. gromovae m3 IPS-85506 46.71 584 m3 US H. periafricanum m3 IPS-85504 41.51 591 Equus quagga m3 Composition 98.71* 1750* AN M ED PT CE AC TABLE 4 Species Event Age (months) Reference First molar eruption 8-12 (Hoppe et al., 2004) Weaning age (feral) 8-9 (Waran et al., 2008) Equus caballus Third molar eruption 42-48 (Hoppe et al., 2004) Femoral epiphysis fusion 36-42 (Silver, 1969) Breeding (Przewalski’s horse) 36-48 (Monfort et al., 1994) First molar eruption 9-12 (Smuts, 1974) Equus quagga Weaning age 8-12 (Pluháček et al., 2007) ACCEPTED MANUSCRIPT Third molar eruption 39 (Smuts, 1974) Adult body mass attainment 36 (Smuts, 1975) Sexual physiological maturity 42 (Smuts, 1976a) T IP CR US AN M ED PT CE AC ACCEPTED MANUSCRIPT Highlights  The life history of Miocene hipparionins is reconstructed using dental histology  Equid first/second molars grow differently than later-formed third molars  Hipparionin teeth grew at slower rates and erupt relatively later than in Equus  Tooth growth in hipparions suggests slower pace of growth compared to extant T equids IP  The analyzed dwarf hipparionins followed slower life histories than larger species CR US AN M ED PT CE AC Figure 1 Figure 2 Figure 3 Figure 4 Figure 5