This is the accepted version of the article: Fortuny, Josep; Zanolli, Clément; Bernardini, Federico; [et al.]. «Dryopithecine palaeobiodiversity in the Iberian Miocene revisited on the basis of molar en- dostructural morphology». Palaeontology, 2021. DOI 10.1111/pala.12540 This version is available at https://ddd.uab.cat/record/240871 under the terms of the license 1 Dryopithecine palaeobiodiversity in the Iberian Miocene revisited on the basis of 2 molar endostructural morphology 3 4 by JOSEP FORTUNY1*, CLÉMENT ZANOLLI2, FEDERICO BERNARDINI3,4, 5 CLAUDIO TUNIZ3,4,5 and DAVID M. ALBA1,* 6 7 1 Institut Català de Paleontologia Miquel Crusafont, Universitat Autònoma de 8 Barcelona, Edifici ICTA-ICP, C/ Columnes s/n, Campus de la UAB, 08193 9 Cerdanyola del Vallès, Barcelona, Spain; emails: josep.fortuny@icp.cat (ORCiD: 10 0000-0003-4282-1619), david.alba@icp.cat (ORCiD: 0000-0002-8886-5580) 11 2 Laboratoire PACEA, UMR 5199 CNRS, Université de Bordeaux, Pessac, France; 12 email: clement.zanolli@gmail.com (ORCiD: 0000-0002-5617-1613) 13 3 Department of Humanistic Studies, Università Ca’Foscari, Venezia, 14 Italy 15 4 Multidisciplinary Laboratory, the ‘Abdus Salam’ International Centre for Theoretical 16 Physics, Via Beirut 31, 34151 Trieste, Italy; emails: fbernard@ictp.it (ORCiD: 0000- 17 0002-3282-8799), ctuniz@ictp.it (ORCiD: 0000-0001-6582-7568) 18 5 Center for Archaeological Science, University of Wollongong, Northfields Ave, 19 Wollongong, NSW 2522, Australia 20 21 * Corresponding authors. 22 23 Abstract: Extensive fieldwork at Abocador de Can Mata (NE Iberian Peninsula) has 24 uncovered a previously unsuspected diversity of catarrhine primates in the middle 25 Miocene (12.5–11.6 Ma) of Europe. However, the distinction of the great ape genera 1 26 Pierolapithecus and Anoiapithecus from Dryopithecus—supported by craniodental 27 differences—has been disputed by some authors. Here we revisit the diversity of great 28 apes (dryopithecines) from the Iberian Miocene based on molar 3D endostructural 29 morphology (relative enamel thickness, enamel distribution, and enamel-dentine 30 junction [EDJ]). Using microtomography, we inspected an extensive sample of 49 31 hominoid molars representing at least five species from 12 localities. 2D and 3D 32 relative enamel thickness values indicate that Dryopithecus and “Sivapithecus” 33 occidentalis (species inquirenda) display the thinnest and thickest enamel, 34 respectively, while the remaining taxa (Hispanopithecus, Anoiapithecus, 35 Pierolapithecus) show intermediate values. Upper molar enamel distribution maps 36 exhibit a similar pattern in P. catalaunicus, A. brevirostris, D. fontani, H. laietanus 37 and H. crusafonti whereas for the lower molars they reveal differences between 38 Hispanopithecus laietanus and Hispanopithecus crusafonti. Lower molar enamel 39 distribution and EDJ morphology of “S.” occidentalis support the distinction of this 40 species but do not resolve if it is a junior synonym of Anoiapithecus brevirostris or 41 Pierolapithecus catalaunicus. Overall our results support the distinction of middle 42 Miocene dryopithecins from late Miocene hispanopithecins, the distinction of 43 Pierolapithecus and Anoiapithecus from Dryopithecus among the former, and the 44 distinct species status of H. crusafonti compared to H. laietanus among the latter. Our 45 results highlight the potential of inner tooth morphology for hominoid alpha- 46 taxonomy. 47 48 Key words: Fossil primates; Hominoidea; Dryopithecinae; Dental morphology; 49 Enamel-dentine junction; Relative enamel thickness 50 2 51 The Miocene record of great apes (Primates: Hominidae) in the Iberian Peninsula is 52 restricted to Catalonia in NE Spain (Vallès-Penedès and Seu d’Urgell basins) 53 (Casanovas-Vilar et al. 2011; Alba 2012). Until two decades ago, all great apes from 54 the Iberian Miocene were subsumed into a single genus, Dryopithecus (e.g., Begun et 55 al. 1990; Harrison 1991; Begun 1992, 2002, 2007; Moyà-Solà & Köhler 1993, 1995; 56 Ribot et al. 1996), although most of the available sample consisted of late Miocene 57 remains (Golpe Posse 1993). Since 2002, new discoveries at the middle to late 58 Miocene Abocador de Can Mata (ACM) composite section (ca. 12.6–11.4 Ma) (Alba 59 et al. 2006, 2017; Casanovas-Vilar et al. 2011, 2016) have unveiled a previously 60 unsuspected diversity of catarrhine primates (Alba 2012; Alba et al. 2017), including 61 pliopithecoids (Alba et al. 2010a, 2012a), the small-bodied putative stem hominoid 62 Pliobates (Alba et al. 2015), and as many as three great ape (hominid) genera, each 63 represented by a single species: Pierolapithecus catalaunicus (Moyà-Solà et al. 2004), 64 Anoiapithecus brevirostris (Moyà-Solà et al. 2009a; Alba et al. 2013), and 65 Dryopithecus fontani (Moyà-Solà et al. 2009b; Alba & Moyà-Solà 2012)—of which 66 the two former were originally described based on ACM material (Moyà-Solà et al. 67 2004, 2009a). 68 An isolated upper molar from els Hostalets de Pierola assigned to Dryopithecus by 69 van der Made & Ribot (1999) most likely comes from younger levels than those 70 recorded at ACM (Alba et al. 2013). Although this specimen was left unassigned to 71 genus in some previous publications (e.g., Alba 2012; Alba et al. 2013), it is here 72 attributed to D. fontani following Alba et al. (2020). In turn, a mandibular fragment 73 from ACM is provisionally assigned to “Sivapithecus” occidentalis (Alba et al. 2020). 74 This nominal species is currently recognized as a species inquirenda because it differs 75 from D. fontani but additional material would be required to discount an attribution to 3 76 either P. catalaunicus or A. brevirostris (Alba et al. 2020). The recovery of cranial 77 remains assigned to Dryopithecus at ACM also prompted the resurrection of genus 78 Hispanopithecus (Moyà-Solà et al. 2009b) and the transferral into it of two late 79 Miocene great ape species from the Vallès-Penedès Basin (Hispanopithecus laietanus 80 and Hispanopithecus crusafonti; see Begun 1992, 2002; Moyà-Solà & Köhler, 1993, 81 1995; Golpe Posse 1993; Alba et al. 2012b) formerly assigned to Dryopithecus (see 82 references above). All these genera are generally considered to belong to a single 83 group, here distinguished as the subfamily Dryopithecinae (Alba 2012), whose 84 phylogenetic affinities are still unclear—being generally considered either stem 85 hominids (Moyà-Solà et al. 2004, 2009a; Alba 2012; Alba et al. 2015) or hominines 86 (Begun 2009, 2015; Begun et al. 2012). The distinction of two species of 87 Hispanopithecus (Begun 1992; Cameron 1999; Alba et al. 2012b), formerly criticized 88 by several authors (Harrison 1991; Andrews et al. 1996; Ribot et al. 1996), also relies 89 on dental differences and therefore requires more detailed studies. 90 While the distinction of Hispanopithecus is currently universally accepted 91 (Casanovas-Vilar et al. 2011; Alba 2012; Begun et al. 2012; Alba et al. 2012b, 2015; 92 Begun 2015; Böhme et al. 2019), the recognition of three middle Miocene 93 dryopithecine genera in the Vallès-Penedès Basin (Casanovas-Vilar et al. 2011; Alba 94 2012; Alba & Moyà-Solà 2012; Alba et al. 2013, 2020; Pérez de los Ríos et al. 2013; 95 Marigó et al. 2014) has been accepted by some (Pickford 2012; Fleagle 2013; Böhme 96 et al. 2019; Andrews 2020) but questioned by others (Begun 2009, 2010, 2015). In 97 particular, both Pierolapithecus and Anoiapithecus have been considered likely junior 98 synonyms of Dryopithecus (Begun 2009). Although the distinction of the former is 99 mainly based on cranial anatomy (Moyà-Solà et al. 2004, 2009a,b; Alba 2012; Pérez 100 de los Ríos et al. 2012), subtle differences in dental morphology have also been argued 4 101 to differentiate these two genera among themselves and relative to both Dryopithecus 102 and Hispanopithecus (Alba & Moyà-Solà 2012; Alba et al. 2013, 2020; Pérez de los 103 Ríos et al. 2013). Although the postcranial record is more restricted (unknown for 104 Anoiapithecus), it also supports the genus distinction between Pierolapithecus and 105 Hispanopithecus (Moyà-Solà et al. 2004, 2005; Almécija et al. 2007, 2009; Alba et al. 106 2010c, 2011, 2012c) and hints at some differences between Dryopithecus and the two 107 former genera (Moyà-Solà et al. 2009b; Almécija et al. 2013; Pina et al. 2012, 2019). 108 Enamel thickness variation (Alba et al. 2010b, 2013, 2020) and, most recently, 109 crown endostructural variation (Alba et al. 2020) have also been investigated to clarify 110 the allocation of fragmentary dentognathic remains (Alba et al. 2020). A large 111 proportion of hominoid specimens from the Vallès-Penedès Basin are fragmentary 112 denthognatic remains or isolated teeth. Therefore, dental morphology plays a very 113 important role in the alpha-taxonomy of these taxa. Unfortunately, taxonomic 114 assessments based on dental morphology on Iberian dryopithecines are complicated by 115 small sample sizes and overall similarities in occlusal morphology. Non-invasive 116 techniques based on X-ray microcomputed tomography (µCT) provide a wealth of 117 additional information on the inner structural morphology of teeth (Macchiarelli et al. 118 2013), which complements and augments that provided by the outer enamel surface 119 (OES). The latter is often affected by occlusal wear or taphonomic damages. In 120 contrast, µCT grants non-destructive access to tooth endostructural morphology and 121 enables the assessment of taxonomic and functionally-related parameters such as 122 enamel thickness distribution over the crown, as well as enamel-dentine junction 123 (EDJ) shape. 124 To test the hypotheses that the Iberian Miocene hominid diversity includes at least 125 four genera (Pierolapithecus, Anoiapithecus, Dryopithecus, Hispanopithecus) and 5 126 further evaluate the taxonomic distinctiveness of these Miocene dryopithecines from 127 Catalonia, we investigate the internal dental morphology based on most of the 128 available upper and lower molars. Enamel thickness, previously investigated for the 129 middle Miocene (Alba et al. 2010b, 2013, 2020) and late Miocene (Andrews & Martin 130 1991; Smith et al. 2019) dryopithecines, is here analyzed in 2D and, for the first time, 131 in 3D. We also describe the EDJ morphology of these taxa in relation to previous 132 observations based on OES morphology, and based on these data we reevaluate the 133 distinction between the investigated dryopithecine taxa. Based on the previously 134 published research cited above, differences in terms of dental endostructural 135 organization are to be expected between (1) middle Miocene and late Miocene 136 hominoids; (2) Dryopithecus as compared with Pierolapithecus and Anoiapithecus 137 (and “S.” occidentalis), particularly in terms of tissue proportions; and (3) H. 138 crusafonti and H. laietanus, with the latter being somewhat more derived relative to 139 the middle Miocene genera. 140 141 MATERIAL AND METHODS 142 Dental terminology 143 The dental terminology employed in the descriptions is depicted in Fortuny et al. 144 (2020, fig. 1). It follows that of Harrison and Gu (1999), except that ‘protoconule’ is 145 favored over ‘paraconule’ (Swindler 2002). 146 147 Studied sample 148 It consists of 49 (30 upper and 19 lower) permanent molars from 25 specimens 149 representing a minimum of 15 individuals from 12 localities—see Table 1 for 150 provenance details. The specimens are housed at the Institut Català de Paleontologia 6 151 Miquel Crusafont, Sabadell, Spain (acronym IPS), except for two specimens housed at 152 the Museu de Geologia del Seminari de Barcelona, Spain (MGSB). 153 154 Computational techniques 155 Microcomputed tomography acquisitions Specimens were imaged by microfocus X- 156 ray microcomputed tomography (µCT) at the Multidisciplinary Laboratory of the 157 ‘Abdus Salam’ International Centre for Theoretical Physics of Trieste. The scans were 158 made with a transportable scanner specifically designed for the investigation of 159 cultural heritage items. X-rays are produced by a Hamamatsu microfocus X-ray source 160 (150 kV maximum voltage, 500 mA maximum current, and 5 mm minimum focal spot 161 size) and the detector is a Hamamatsu CMOS flat panel coupled to a fibre optic plate 162 under the GOS scintillator. The system has been designed to allow large sample-to- 163 detector distances to exploit phase-contrast effects (Tuniz et al. 2013). The 164 acquisitions were performed according to the following parameters: 120 to 150 kV 165 voltage; 62 to 201 µA current; a projection each 0.15° to 0.20° (see Fortuny et al. 166 (2020, table 1 for details). MicroCT raw data requests should be addressed to the 167 Fieldwork & Collection Management Area of the ICP. 168 169 Virtual reconstruction and segmentation The final volumes were reconstructed using 170 Cobra v.7.4.16 (Exxim) and DigiCT v.2.3.3 (DIGISENS) in 8-bit format, with an 171 isotropic voxel size ranging from 14.36 to 20.42 µm (see Fortuny et al. 2020, table 1 172 for details). Using Avizo 7.0 (FEI-Visualization Sciences Group Inc.) and ImageJ 173 v.1.47 (NIH; Schneider et al. 2012), a semiautomatic threshold-based segmentation 174 was carried out (Fortuny et al. 2020, fig. 2) following the half-maximum height 175 method (HMH; Spoor et al. 1993) and the region of interest thresholding protocol 7 176 (ROI-Tb; Fajardo et al. 2002), taking repeated measurements on different slices of the 177 virtual stack (Coleman & Colbert 2007). 178 Digital surface models of the OES and EDJ of the investigated molars are available on 179 MorphoSource (www.morphosource.org). OES models are openly shared from 180 Morphosource, whereas the EDJ are available through Morphosource upon request; 181 see Fortuny et al. 2020, table 2 for digital object identifiers and further details. 182 183 Relative enamel thickness 2D relative enamel thickness (RET) was computed 184 following a protocol originally devised for histological sections (Martin 1985). In 185 order not to overestimate RET due to obliquity, it was assessed on virtual coronal 186 buccolingual sections perpendicular to the best-fit plane of the cervical line and 187 passing through the tips of the mesial dentine horns (Benazzi et al. 2014). This method 188 was already used in previous studies to compute 2D RET values for the middle 189 Miocene specimens included in this study (Alba et al. 2013, 2020). The following 190 formula was employed (Martin 1985; Smith et al. 2005; Alba et al. 2010b): 2D RET = 191 2D AET × 100 / b1/2, wherein b is the dentine and pulp area, 2D AET is average 192 enamel thickness, computed as 2D AET = c (enamel cap area) / e (enamel-dentine 193 junction length). While 2D RET is a dimensionless variable originally developed to 194 compare enamel thickness among species of different tooth size (Martin 1985), the 195 assessment of intraindividual intertooth variation should instead be based on 2D AET 196 (Smith et al. 2005). As done by other authors (e.g., Smith et al. 2005), for several of 197 the analyzed teeth it was necessary to correct RET calculations for tooth wear. As it is 198 frequently done to maximize the sample available for the estimation of tissue 199 proportions (e.g., Smith et al. 2006, 2012a; Martin-Francés et al. 2018), 200 reconstructions of the worn enamel and dentine horn tip were made prior to 8 201 measurement for sections showing light to moderate wear, or when a small amount of 202 cervical enamel was missing (based on the curvature and orientation of the outer 203 enamel surface relative to the EDJ). Tooth wear stages were assessed following the 204 adaptation of a previous protocol established on human teeth (Smith 1984). Specimens 205 that were too heavily worn (above stage 4) were excluded. In consequence, a total of 206 41 specimens (25 upper molars and 16 lower molars) were analyzed for 2D RET. The 207 formulas employed to compute 3D AET and 3D RET are just the three-dimensional 208 extension of those employed to compute 2D RET (Olejniczak 2006; Olejniczak et al. 209 2008a,b,c,d; Benazzi et al. 2014): 3D AET = Ve / SEDJ and 3D RET = 3D AET × 210 100 / Vcdp1/3, wherein Ve is enamel cap volume, Vcdp is dentine and pulp volume, 211 and SEDJ is EDJ surface area. A total of 31 specimens (21 upper molars and 10 lower 212 molars) were analyzed for 3D RET. 213 We refrained from using discrete categorizations of enamel thickness stemming 214 from the thin vs. thick dichotomy (Martin 1985) in considering that they do not 215 adequately reflect either the continuum displayed among different teeth of a single 216 individual or intra- and interpecific variability. Therefore, comparisons between the 217 analyzed Miocene apes and other hominoids were based on statistical comparisons of 218 2D RET and 3D RET values among species. For 2D RET results we also included in 219 the statistical comparisons the data derived from the histological sections used by 220 Andrews & Martin (1991) and Kelley et al. (2001) as reported by Smith et al. (2019). 221 The small available sample sizes precluded performing statistial tests to assess 222 differences in RET for different tooth loci separately, as such comparisons would not 223 have enough statistical power to denote significant differences. Statistical comparisons 224 were instead done by lumping the data from all tooth loci. These comparisons should 225 be taken with great care because there is a trend towards increasing relative enamel 9 226 thickness from first to third molars in both humans and apes (Grine & Martin 1988; 227 Macho 1994; Grine 2002, 2005; Smith et al. 2005, 2006, 2019). To account for this 228 problem, we compared individual data for Vallès-Penedès hominoids with the median 229 values and range of variation displayed by extant great ape genera for each tooth locus 230 separately (Smith et al. 2008, 2012b, 2019). Comparisons of 2D RET and 3D RET 231 values for both extant and extinct taxa were made with PAST v. 4.01 (Hammer et al. 232 2001) by means of Kruskal-Wallis tests for equality of medians and pairwise Mann- 233 Whitney post hoc comparisons, which are non-parametric and hence do not assume 234 normal distributions, with and without Bonferroni correction. Adjusted z-score 235 analyses were performed for 2D RET and 3D RET from the Iberian Miocene great 236 apes and on five extant comparative taxa. This method allows the comparison of 237 unbalanced samples, as it is often the case in the fossil record. 238 239 Enamel distribution maps Enamel distribution maps permit comparisons in the local 240 distribution of enamel over the entire crown surface (Macchiarelli et al. 2008, 2009, 241 2013; Zanolli et al. 2019; Thiery et al. 2017). Enamel thickness topographic variation 242 was rendered for 44 specimens (25 upper molars and 19 lower molars) through 3D 243 cartographies using a chromatic scale where thickness increases from dark blue (thin) 244 to red (thick) (Macchiarelli et al. 2008, 2013; Bayle et al. 2011). The software Avizo 245 7.0 (FEI-Visualization Sciences Group Inc.) was used for this purpose. This 246 visualization technique maps the local enamel thickess by computing the site-specific 247 shortest distance between the OES and EDJ surfaces. 248 249 EDJ morphology The morphology of the EDJ was examined in the same 44 250 specimens. The Miocene hominid molars exhibit expression of non-metric features 10 251 that are not covered with the usual scoring systems like the Arizona State University 252 Dental Anthropology method that was developed for human/hominin teeth (Turner et 253 al. 1991; Scott & Irish 2017). Therefore, we elaborated a qualitative approach based 254 on a limited number of stages for each feature (see descriptions of traits in the Tables 3 255 and 4). 256 257 RESULTS 258 Enamel thickness 259 2D RET was computed for 41 specimens (see Fortuny et al. 2020, table 3 and 260 sections used in Fortuny et al. 2020, figs 3–5). 2D RET values for 20 specimens, 261 attributed to P. catalaunicus, A. brevirostris, D. fontani and “S.” occidentalis, had 262 already been published (see Alba et al. 2013, 2020), while 3D RET was newly 263 reported for 31 specimens (see Fortuny et al. (2020, table 4). 264 Based on average 2D RET values (Table 2, Fig. 1 A; see Fortuny et al. 2020, table 265 5 and figs 6, 7A), D. fontani displays the thinnest enamel (12.3), while “S.” 266 occidentalis displays the thickest (19.6) among analyzed sample, and other 267 dryopithecines display similar intermediate average values; in order of increasing 2D 268 RET: H. laietanus (14.3), H. crusafonti (14.4), A. brevirostris (14.6), and P. 269 catalaunicus (15.4). Such apparent differences in average RET values among these 270 taxa cannot be taken at face value and must be interpreted with great care, given the 271 small samples available for most of the taxa and the variation displayed by extant 272 taxa—as further illustrated by the range of H. laietanus (10.3–19.1, N=17), which 273 broadly overlaps the values for the remaining taxa. Our results for average 3D RET 274 (Table 2, Fig. 1B; see Fortuny et al. 2020, table 5 and fig. 7B) also indicate that D. 275 fontani (11.9) and “S.” occidentalis (19.0) display the thinnest and thickest enamel, 11 276 respectively, with the remaining taxa showing intermediate average values; in order of 277 increasing 3D RET: H. crusafonti (12.1), A. brevirostris (12.9), H. laietanus (13.5), 278 and P. catalaunicus (15.5). Although the “S.” occidentalis sample displays the thickest 279 enamel, it should be taken into account that this result is likely biased by small size 280 coupled with the lack of first molars available for analysis. 281 Statistical comparisons based on Kruskal-Wallis tests for equality of medians show 282 significant differences among extant apes in both 2D RET (χ2 = 38.31, p < 0.001) and 283 3D RET (χ2 = 32.05, p < 0.001). Mann-Whitney pairwise comparisons in 2D and 3D 284 RET (See Fortuny et al. 2020, table 6) indicate that African apes (Gorilla and Pan) 285 and siamangs display similarly thin enamel, whereas orangutans and gibbons display 286 significantly thicker enamel (see average and maximum-minimum values in Fortuny et 287 al. 2020, tables 7 and 8). Chimpanzees display slightly thicker enamel than gorillas 288 only for 3D RET, whereas differences in 2D RET become nonsignificant after 289 Bonferroni correction, and the same occurs between siamangs and gibbons. 290 Comparisons between the fossil samples and extant taxa are possible for both 2D 291 and 3D RET (Fortuny et al. 2020, tables 7 and 8, respectively), but comparisons with 292 other extinct taxa are mostly restricted to 2D RET due to the lack of 3D data (with 293 only two exceptions; see Fortuny et al. 2020, table 8). We therefore compared 2D 294 RET among Iberian dryopithecines, extant hominoids, and two fossil hominoid 295 samples—the kenyapithecine Griphopithecus and the dryopithecine Rudapithecus (see 296 descriptive statistics in Fortuny et al. 2020, table 9 and fig. 7A)—whereas 297 comparisons for 3D RET were restricted to Iberian dryopithecines and extant taxa (see 298 Fortuny et al. 2020, table 10 and fig. 7B). 299 When these enlarged samples are considered, Kruskal-Wallis tests show again 300 significant differences in both 2D RET (χ2 = 23.44, p < 0.001) and 3D RET (χ2 = 12 301 13.84, p < 0.001). Mann-Whitney pairwise comparisons are reported for fossil taxa as 302 compared with the extant hominoid samples discussed above (see Fortuny et al. 2020, 303 table 11). Based on the currently available, restricted fossil samples, 2D RET 304 comparisons indicate that all extinct taxa show thicker enamel than gorillas and 305 chimpanzees, and most also display thicker enamel than siamangs, with the exception 306 of D. fontani (although differences also approach the 0.05 significance threshold, for 307 H. laietanus and Rudapithecus). In contrast, most extinct taxa display lower 2D RET 308 than Pongo, except Griphopithecus, P. catalaunicus and “S.” occidentalis, which do 309 not differ from orangutans. However, like most other extinct taxa, P. catalaunicus 310 does not differ from gibbons, whereas Griphopithecus and “S.” occidentalis show 311 instead thicker enamel (although the “S.” occidentalis sample is biased, as explained 312 above) (Fortuny et al. 2020, fig. 7A). 313 Based on 3D RET, less significant differences are found, probably as a result of 314 smaller sample sizes. Only “S.” occidentalis clearly shows thicker enamel than extant 315 apes (although this is probably biased, see above), whereas H. laietanus and A. 316 brevirostris display thicker enamel than siamangs and gorillas (although comparisons 317 with chimpanzees only approach significance), while P. catalaunicus further displays 318 thicker enamel than chimps, more closely resembling orangs and gibbons. 319 Adjusted z-scores for 2D RET (see Fortuny et al. 2020, fig. 8A, table 12) and 3D 320 RET (see Fortuny et al. 2020, fig. 8B and table 13) for each fossil specimen compared 321 with extant hominoid samples may clarify the results provided above, especially for 322 the smaller samples, as it is the case for “S.” occidentalis in both 2D RET and 3D 323 RET, as well as P. catalaunicus, H. crusafonti, and D. fontani in 3D RET. Due to the 324 extensive overlap among extant taxa, particularly when all tooth loci are considered 325 simultaneously, many specimens fit within the variation of all the comparative 13 326 samples, although others do significantly differ from some. For all the taxa examined, 327 some specimens differ from African apes and, for H. laietanus, P. catalaunicus, and 328 “S.” occidentalis, also from siamangs, whereas most specimens do not differ from 329 either gibbons or orangutans. In fact, in 2D only a single specimen of H. laietanus and 330 two of “S.” occidentalis show significantly thicker enamel than orangutans, whereas in 331 3D there is a single specimen of “S.” occidentalis that shows thicker enamel than both 332 siamangs and humans. While the Miocene ape’s sample size available for 3D 333 estimates is smaller than for 2D analyses, the results show high coherence. 334 335 Enamel distribution maps 336 Enamel distribution maps (Figs 2 and 3) in D. fontani upper molars show that the 337 thickest enamel is on the periphery of the cusps and marginal ridges, with much 338 thinner enamel in the trigon than over the talon basin. The upper molars of P. 339 catalaunicus, A. brevirostris and H. laietanus overall show a similar enamel 340 distribution pattern to D. fontani, except that the enamel is relatively thicker on the 341 trigon basin. A specimen of H. laietanus (IPS58340) somewhat differs by having its 342 thickest enamel limited for the most part to the protocone, while both the trigon and 343 talon basins display thinner enamel. In the upper molars of H. crusafonti, thicker 344 enamel is distributed over the talon basin and lateral walls, although, with the 345 exception of the protocone, the trigon enamel is moderately thin to thin. In the lower 346 molars of H. laietanus, the thickest enamel is found on the periphery of the talonid, 347 while the trigonid (even buccally) generally shows moderately thick enamel. The only 348 exception is IPS1822 (the invalid holotype of “D. piveteaui”, currently included in H. 349 laietanus), which displays its thickest enamel only on the outer aspect of the buccal 350 cusps, as in the paratype of H. crusafonti IPS1816. The molars of H. crusafonti 14 351 MGSB25314 have in contrast their thickest enamel located on the outer aspect of the 352 cusps, their inner aspects being also relatively thick (only the center of the occlusal 353 basin and the cervical part of the crown are thin enameled). The holotype of “S.” 354 occidentalis, in turn, displays a similar enamel distribution pattern as IPS1822 and 355 IPS1816, with the thickest areas mostly located on the buccal half of the crown 356 (notably on the outer aspect of the buccal cusps, up to the cusp apex), and thinner 357 enamel being located lingually. “Sivapithecus” occidentalis IPS41734, although 358 somewhat worn, also approximates the enamel distribution shown by IPS1822 and the 359 holotype of “S.” occidentalis. 360 361 EDJ morphology 362 Outer enamel surface morphology (see Fortuny et al. (2020, figs 9 and 10) of all the 363 dryopithecine upper molars considered in this study displays a similar pattern, which 364 is also reflected at the EDJ but with a sharper topography (Fig. 4; Table 3). All the 365 M1s and M2s display four developed dentine horns at the EDJ, corresponding to the 366 four main cusps visible on the OES (see Fortuny et al. 2020, fig. 9), whereas in the M3 367 one of the two distal cusps is frequently absent in accordance with their generally 368 shorter and more distally-tapering crowns. Thus, no hypocone dentine horn is 369 expressed in the M3 of P. catalaunicus (Fig. 4C), whereas it is well expressed A. 370 brevirostris (Fig. 4J) and H. crusafonti (Fig. 4T,U), and discernible (even if smaller) 371 in D. fontani (Fig. 4G). In the M3 of H. laietanus, the hypocone is present, but the 372 metacone horn is very poorly developed (Fig. 4X,Y), which contrasts with the distinct 373 M3 metacone displayed by the other taxa (with the exception of one M3 of H. 374 crusafonti; Fig. 4T). The dentine horns of the upper molar four main cusps are more 375 centrally situated in P. catalaunicus and especially in D. fontani (Fig. 4D–G) than in 15 376 A. brevirostris, H. crusafonti and especially H. laietanus (Fig. 4V–Y), which have 377 more peripheral and vertically set dentine horns, in accordance with their less flaring 378 crowns. In the M1s and M2s of A. brevirostris (Fig. 4H,I,M–P) and H. crusafonti (Fig. 379 4Q–S), the hypocone dentine horn is clearly more lingual than the protocone, as in the 380 D. fontani M1 (Fig. 4E) but unlike in the remaining specimens. 381 The mesial fovea, slit-like at the OES, appears as a larger, ovoid to subrectangular 382 depression at EDJ level, even if much smaller than the trigon basin, particularly in the 383 M3s. This fovea appears shallower and slightly more inclined mesialward in P. 384 catalaunicus (Fig. 4A–C) and D. fontani (Fig. 4D–G) than in the remaining species, 385 and also more mesially projected in the latter. In A. brevirostris (Fig. 4H–P) and 386 especially Hispanopithecus spp. (Fig. 4Q–Y), the mesial fovea is mesiodistally 387 shorter, more buccally positioned, deeper, and enclosed by a stronger mesial marginal 388 ridge. The mesial fovea is generally separated from the trigon basin by a well- 389 developed hypoparacrista that links the paracone with the protoconule (except in a few 390 specimens, e.g., Fig.4U,V). The protoconule is often obliterated by wear on the OES, 391 but a distinct dentine horn subequal in size to those of the four main cusps is 392 frequently evidenced at the EDJ, even if variably developed: it is generally less 393 developed in M3 (except in P. catalaunicus; Fig. 4C), and only poorly developed in 394 the M2 of D. fontani MGSB48486 (Fig. 4D). The hypoparacrista generally terminates 395 at the protoconule, but in some instances it joins the mesial marginal ridge, as in the 396 M3s of P. catalaunicus, D. fontani, A. brevirostris and H. crusafonti (Fig. 4C,G,J,T), 397 and some M2s of A. brevirostris (Fig. 4L) and H. laietanus (Fig. 4W). In the two latter 398 taxa, a secondary small dentine horn is also present at the junction of the 399 hypoparacrista with the marginal ridge. The hypoparacrista generally originates nearby 16 400 the paracone dentine horn apex, although in the M3 of D. fontani and A. brevirostris 401 (Fig. 4G,J) it originates more mesially. 402 At the EDJ, the trigon basin displays a subrhomboid contour (instead of a triangular 403 one, as in the OES) in all the taxa, but is shallower in D. fontani (Fig. 4D–G) and H. 404 laietanus (Fig. 4V–Y). The crista obliqua is high, complete and generally straight in P. 405 catalaunicus (Fig. 4A–C) and A. brevirostris (Fig. 4H–M), although it is somewhat 406 discontinuous (Fig. 4N–O) or even poorly developed (Fig. 4P) is some molars of the 407 latter species. In D. fontani (Fig. 4D–G), the crista obliqua is low and less straight, due 408 to the slightly curved postprotocrista, resulting in a sinuous crista obliqua in one 409 specimen (Fig. 4G). In Hispanopithecus spp. (Fig. 4Q–Y), the crista obliqua is even 410 lower and more diffuse, particularly in H. crusafonti (Fig. 4Q–U), being interrupted 411 (non-merging postprotocrista and hypometacrista) in the two M3s of H. crusafonti 412 (Fig. 4T,U) and one of the M3s of H. laietanus (Fig. 4X). In all the taxa, the talon 413 basin is subtriangular and smaller than the trigon basin, also being shallower than the 414 latter in P. catalaunicus (Fig. 4A–C), A. brevirostris (Fig. 4H–P) and H. crusafonti 415 (Fig. 4Q–U). In most specimens, the talon basin is divided by a transverse and low 416 hypocone-metacone crista, more clearly discerned at the EDJ than at the OES. This 417 crista is generally more distinct in the M1 than in the M2, and absent from the M3. 418 When present, it delimits a distal fovea from the deeper and more expansive talon 419 basin, originating from the hypocone dentine horn and joining the end of the 420 postmetacrista (only with some exceptions; Fig. 4K,Q,V). 421 The upper molars display a subquadrangular (M1) to subrectangular (M2) occlusal 422 outline, except for the M3 of P. catalaunicus, A. brevirostris and H. laietanus (Fig. 423 4C,J,X,Y), which display an ovoid to subtriangular profile due to the truncated talon. 424 The degree of talon development is quite variable in Hispanopithecus, as previously 17 425 noted based on the OES (Alba et al. 2012b), encompassing differences among the 426 other species in this regard. The degree of buccolingual waisting of the upper molars 427 at EDJ level is more marked in A. brevirostris (Fig. 4H,I,K–P), slightly less developed 428 in D. fontani (Fig. 4D–G), more variable in H. crusafonti and H. laietanus (Fig. 4Q– 429 W), and least developed in P. catalaunicus (Fig. 4A,B). The weak to moderate 430 development of the lingual cingulum at the OES is variably expressed at the EDJ, 431 ranging from the lack of this feature to a shelf bordered by a (semi)continuous crest 432 extending along the lingual aspect of the protocone, sometimes connecting the lingual 433 groove separating the protocone and hypocone. Even if decreasing in expression from 434 the M1 to the M3, the lingual cingulum tends to be more developed in A. brevirostris 435 and H. crusafonti (Fig. 4H–U), only moderately expressed in D. fontani (Fig. 4D–G), 436 and even less so in P. catalaunicus (Fig. 4A, B,C) and H. laietanus (Fig. 4V,W). A 437 buccal cingular remnant (in the form of a short crest enclosing a small fovea) 438 frequently appears at the external end of the buccal groove at the EDJ in the M1 and 439 most M2 of A. brevirostris (Fig. 4H,K,L–P) and H. crusafonti (Fig. 4Q–S). Similarly, a 440 shelf-like structure is visible at the same spot in some specimens of D. fontani (Fig. 441 4E,F) and the M1 of P. catalaunicus (Fig. 4A). In contrast, the buccal aspect is rather 442 smooth in H. laietanus (Fig. 4V,W) and the remaining specimens of P. catalaunicus 443 (Fig. 4B). 444 The overall endostructural pattern of the lower molars (Fig. 5; Table 4) is similar to 445 the OES morphology (see Fortuny et al. 2020, fig. 10), with five well-developed 446 dentine horns corresponding to the five main cusps. The dentine horns are generally 447 vertically set, with the exception of the metaconid dentine horn of “S.” occidentalis 448 specimens (Fig. 5A–C), which is tilted toward the center of the tooth. In some 449 specimens of Hispanopithecus spp., the metaconid dentine horn is also somewhat 18 450 centrally tilted (especially in the M3 of MGSB25314; Fig. 5J), although to a lesser 451 extent. The lingual dentine horns are very peripherally situated relative to the crown 452 margin, whereas those corresponding to protoconid and hypoconid are less peripheral. 453 This is more clear-cut in “S.” occidentalis specimens (Fig. 5A–C) compared to 454 Hispanopithecus, although this feature is variable within both H. crusafonti (Fig. 5G– 455 J) and H. laietanus (Fig. 5K–S). At the OES, H. laietanus (see Fortuny et al. 2020, 456 fig.10K, N–S) displays a more reduced buccal cingulid than the remaining taxa (see 457 Fortuny et al. 2020, fig. 10A–J), in which it is nevertheless discontinuous. This 458 difference is more marked at the EDJ (Fig. 5), where the buccal cingulid becomes 459 shelft-like (at least between protoconid-hypoconid and hypoconid-hypoconulid) 460 except in H. laietanus (Fig. 5K, N–S), in which it is only minimally developed if at all. 461 Buccolingual crown waisting is also more strongly expressed at the EDJ, being most 462 marked in H. laietanus (Fig. 5K, N–S) both buccally and lingually, whereas in “S.” 463 occidentalis (Fig. 5A–C) as well as in H. crusafonti (Fig. 5G–J) waisting is moderate 464 to slight. 465 Some specimens express additional cuspulids at the EDJ level. While no 466 tuberculum sextum (C6) is expressed in any specimen at either the OES or the EDJ 467 levels, the M2 of “S.” occidentalis holotype (Fig. 5B) and an M2 of H. crusafonti (Fig. 468 5G) display a well-developed interconulid-type tuberculum intermedium (C7) at the 469 distal end of the postmetacristid, which at the OES is merely expressed as a secondary 470 (cuspulid-like) thickening of the enamel. Furthermore, the M2 of “S.” occidentalis 471 specimens (Fig. 5A–C) display a mesiodistally-elongated metaconid dentine horn with 472 a distinct tuberculum intermedium or metaconulid-type C7 horn just distally from the 473 main metaconid dentine horn. This ‘twinned’ metaconid morphology, also expressed 474 at the OES as a cuspulid-like enamel thickening, is lacking in H. crusafonti (Fig. 5G– 19 475 J), but variably expressed (although to a lesser extent) in most lower molars of H. 476 laietanus (Fig. 5K, N–S), sometimes being discernible at the OES (see Fortuny et al. 477 2020, fig. 10K,L,N,O,R). 478 The mesial fovea is much shorter mesiodistally than buccolingually broad and not 479 completely isolated from the much deeper and more extensive talonid basin, since the 480 hypoprotocristid and hypometacristid junction is interrupted by a mesiodistal fissure. 481 These cristids, less discernible at the EDJ than the OES, are less marked in 482 Hispanopithecus spp. (Fig. 5G–S) than the remaining taxa (Fig. 5A–C). At the EDJ, 483 the cristid obliqua generally appears more distinct than at the OES, although it is also 484 incomplete (the postprotocristid and prehypocristid junction is interrupted, although 485 the buccolingual groove present at the OES is not discernible at the EDJ). Similarly, 486 the profuse development of secondary enamel wrinkling on the talonid basin displayed 487 by some specimens (“S.” occidentalis and H. laietanus; see Fortuny et al. 2020, fig. 488 10B,C,S) has no concomitant expression at the EDJ. Only the obliquely-oriented crest 489 (postcristid + hypoentocristid) separating the talonid basin from the more restricted 490 distal fovea at the OES of most specimens (even if partially interrupted by a 491 mesiodistal groove; see Fortuny et al. 2020, fig. 10) can generally be discerned at the 492 EDJ. However, on the latter it is variably expressed, ranging from a continuous but 493 low crest in most specimens, to a poorly expressed or even indistinguishable structure 494 in some M2s (Fig. 5I,O,Q) and M3s (Fig. 5C,R,S). 495 496 DISCUSSION 497 Our results allow us to refine these previously-reported differences in RET and 498 dental morphology among middle Miocene hominoids from Catalonia (Alba & Moyà- 499 Solà 2012; Alba et al. 2010b, 2013, 2020; Pérez de los Ríos et al. 2013) and further 20 500 provide additional information regarding the distinction between the late Miocene 501 species H. laietanus and H. crusafonti (Begun 1992; Cameron 1999; Alba et al. 502 2012b). 503 504 Enamel thickness 505 This study extends the previously published results of 2D RET for middle Miocene 506 dryopithecines from Catalonia (Alba et al. 2013, 2020) to Hispanopithecus—for 507 which only limited evidence was available in the case of H. laietanus, based on a few 508 histological sections (Andrews & Martin 1991; Kelley et al. 2001; Smith et al. 2019). 509 Most importantly, however, we first report 3D RET results for all these taxa. The 510 enamel of A. brevirostris was originally reported as similarly thick to that of P. 511 catalaunicus (Moyà-Solà et al. 2009b; Alba et al. 2010b), and both were assessed as 512 considerably thicker-enameled than D. fontani (Alba et al. 2010b), thereby 513 contradicting the previous contention that all the middle Miocene dryopithecines from 514 Catalonia were thin-enameled (Begun 2009) as well as the purported synonymy of the 515 former with D. fontani (Begun 2007, 2009). Later on, the significance of such 516 differences was disputed on the basis that enamel thickness is too variable among 517 extant and extinct hominoids to be reliable as a taxonomic criterion (Begun et al. 518 2012). Such differences were subsequently confirmed to some extent using slightly 519 enlarged samples for A. brevirostris and D. fontani and based on higher resolution 520 scans (the same used in this paper; Alba et al. 2013, 2020). However, the latter studies 521 showed that both A. brevirostris and P. catalaunicus display thinner enamel than 522 originally reported (Alba et al. 2010b), being more similar to extant orangutans, albeit 523 still significantly thicker than D. fontani, most similar to extant African apes (Alba et 524 al. 2013, 2020). 21 525 Caution is required when interpreting the RET values, given the small samples 526 analyzed for most extinct taxa, which preclude analyzing sexes or dental loci 527 separately, given the confounding effects of both sexual dimorphism (lower 2D RET 528 values in males) and tooth position (2D RET increase from first to third molars; Smith 529 et al. 2005, 2012b, 2019). This is highlighted by the 2D RET results newly reported 530 here for H. crusafonti and especially H. laietanus, since the comparatively larger 531 sample for the latter species shows a wide range of variation (10.3–19.1; Table 2, see 532 Fortuny et al. 2020, fig. 6) that almost encompasses all the remaining taxa. Similarly, 533 all the extant taxa represented by adequate samples display a wide range of variation 534 in 2D and 3D RET (with maximum values often almost doubling minimum values). 535 Sexual dimorphism might play some role in this regard, given that Smith et al. 536 (2012b) found higher 2D RET values for male than for female orangutans, although 537 differences were not significant for molar loci except in the M3, and differences 538 between tooth loci appear larger. All these factors cannot be adequately addressed in 539 the studied samples, indicating that caution is warranted for extinct taxa represented 540 by small samples. Among hominoids from Catalonia, only H. laietanus (N=17) is well 541 represented, although the samples of P. catalaunicus (N=5), A. brevirostris (N=9) and 542 H. crusafonti (N=7) are still greater than for most other extinct apes except for 543 Griphopithecus (N=8), Rudapithecus hungaricus (N=8) and Gigantopithecus blacki 544 (N=7). 545 While 3D RET is useful for comparing the Iberian dryopithecines with one another 546 and with extant hominoids, comparisons with most extinct hominoids are generally 547 limited to 2D RET data (Smith et al. 2019). This is problematic because, even though 548 2D and 3D RET results are generally in good agreement, this is not always the case 549 due to differences in enamel distribution throughout the crown, as illustrated by the 22 550 cartographies shown in Figures 2 and 3. Interestingly, the range of 3D RET for H. 551 laietanus (10.5–16.4; Table 2) is less extensive than for 2D RET. Furthermore, 3D 552 RET results are arguably more informative because they reflect the global pattern of 553 enamel distribution. 554 Taken together, the data reported here both in 2D and 3D support the view that 555 Iberian dryopithecines, like Rudapithecus from Hungary and Danuvius from Germany 556 (see below), are thicker-enameled than African apes (especially gorillas), and most 557 similar to the condition displayed by extant gibbons and, to a lesser extent, 558 orangutans—with the exception of “S.” occidentalis (which clearly displays thicker 559 enamel, most similar to Griphopithecus) and also P. catalaunicus (which more closely 560 approaches the orangutan condition, particularly in 3D). The 2D RET for 561 Rudapithecus (14.35, range 11.29–17.48, N=8; Smith et al. 2019) fits well with the 562 ranges reported for most Iberian dryopithecines except “S.” occidentalis (and 563 particularly with the variation displayed by H. laietanus), whereas Danuvius (16.03, 564 N=1; Böhme et al. 2019) appears most similar to P. catalaunicus based on the single 565 reported figure for this taxon (a second molar). There is the possibility that D. fontani 566 and H. crusafonti display thinner enamel than the remaining taxa (including A. 567 brevirostris and H. laietanus), apparently being more similar to chimpanzees, but this 568 cannot be demonstrated based on the small available samples available for these taxa. 569 Assuming that “S.” occidentalis likely belongs to one of the taxa recorded at ACM 570 other than D. fontani (Alba et al. 2020), our RET results tentatively support its 571 synonymy with P. catalaunicus. As already explained, the RET figures for “S.” 572 occidentalis are probably exaggerated because there is no first molar available, so that 573 combining with sample with P. catalaunicus would result in a taxon with an enamel 574 thickness most similar to orangutans. In contrast, combining the “S.” occidentalis 23 575 sample with that of A. brevirostris would result in an even wider range of variation 576 than that documented for H. laietanus. 577 Hominoids as a whole are generally thicker-enameled than other anthropoids 578 (Olejniczak et al. 2008a), although displaying considerable interspecific variation 579 (Martin 1985; Schwartz 2000; Smith et al. 2005, 2008; Olejniczak et al. 2008a). 580 Enamel thickness is labile on evolutionary terms due to convergent and relatively 581 rapid dietary adaptations (Andrews & Martin 1991; Alba et al. 2010b). Thick enamel, 582 in particular, has been classically linked to sclerocarpy—i.e., the consumption of hard- 583 foot items (Martin 1985; Andrews & Martin 1991; Vogel et al. 2008). This has been 584 related to selection pressures for low cusp relief and reduced hearing crests (Andrews 585 & Martin 1991) or the biomechanical need to prevent the propagation of radial cracks 586 from the EDJ during the mastication of hard foods (Vogel et al. 2008). However, 587 although there seems to be some general correspondence between the overall 588 properties of the food habitually consumed and tooth structure in primates, linking 589 enamel thickness with specific diets is not necessarily, especially in instances where 590 phylogenetically closely-related species that consume different kinds of food are 591 considered (Grine and Daegling, 2017). Our 2D RET results suggest that most of the 592 Iberian Miocene dryopithecines present thicker enamel than African apes (especially 593 gorillas) and in some cases also than siamangs, which display folivorous tendencies 594 despite a mainly frugivorous diet. In contrast, RET results from Iberian fossil 595 dryopithecines are virtually identical to gibbons, and apparently somewhat lower than 596 in orangutans. This agrees with a soft frugivorous diet for Iberian dryopithecines, with 597 the exception of Pierolapithecus. The latter more closely resembles orangutans, which 598 unlike gibbons displays a scleroparpic component in its diet. This is consistent with 599 microwear data suggesting a frugivorous diet with an orang-like, arboreal hard-object 24 600 component for Pierolapithecus (DeMiguel et al. 2014), but lacking a specialized hard- 601 object diet as that inferred for Griphopithecus. 602 The hypothesis that thick enamel and other dentognathic adaptations to sclerocarpic 603 feeding were the key adaptation that facilitated the dispersal of hominoids out of 604 Africa into Eurasia (Begun 2003) was favored based on previous data (Alba et al. 605 2010b) indicating that both Pierolapithecus and Anoiapithecus were as thick-enameled 606 as earlier African afropithecids (Afropithecus), Eurasian putative stem hominids 607 (Griphopithecus), and early pongines from Asia (Sivapithecus). Our results do not 608 disprove this view, based on the earliest Eurasian forms (Griphopithecus), but indicate 609 that it cannot be supported further based on Iberian dryopithecines, which display an 610 enamel thickness and microwear signal overall more in agreement with soft frugivory 611 (Alba et al. 2010b; DeMiguel et al. 2014). Whether the apparently thicker-enameled 612 Pierolapithecus (particularly if “S.” occidentalis belongs to the same taxon) retains the 613 plesiomorphic condition or represents a secondary reversal among the dryopithecine 614 radition cannot be determined. 615 616 Enamel distribution maps 617 Enamel distribution maps further enable a more refined assessment of enamel 618 thickness, as it has been shown that molar enamel thickness distribution may differ 619 between taxa with close 3D RET values, indicating that the latter might not adequately 620 reflect molar functional and/or taxonomic signals (Kono 2004; Kono & Suwa 2008; 621 Macchiarelli et al. 2008, 2009; Olejniczak et al. 2008b,c; Suwa & Kono 2005; Suwa 622 et al. 2009). Enamel distribution maps of the upper molars of Miocene dryopithecines 623 from Catalonia show that P. catalaunicus, A. brevirostris, D. fontani and H. laietanus 624 exhibit a similar pattern, with the thickest enamel lying over the talon and lateral cusp 25 625 walls, and the trigon basin being considerably thinner (in D. fontani) to moderately 626 thinner (in the other taxa). In this regard, dryopithecines differ from extant apes, with 627 Pan and Hylobates exhibiting their thickest enamel peripherally (on the external 628 aspect of the cusps) and much thinner occlusal enamel, and Gorilla and Pongo 629 approximating this pattern but displaying relatively thicker occlusal enamel (Kono 630 2004; Kono & Suwa 2008; Suwa et al. 2009). 631 While enamel distribution in the upper molars does not differ among the 632 investigated taxa, for the lower molars it shows differences between H. laietanus and 633 H. crusafonti. The distinction between these species and the inclusion of the mandible 634 MGSB25314 in H. crusafonti instead of D. fontani has been supported by some 635 authors (Begun 1992, 2002; Alba 2012; Alba et al. 2012b, 2013) but questioned by 636 others (Harrison 1991; Golpe Posse 1993; Ribot et al. 1996). The only examined 637 lower molar of H. crusafonti from its type locality differs from most specimens of H. 638 laietanus by displaying the thickest enamel on the buccal aspect of the buccal cusps 639 instead of the periphery of the whole talonid, thereby supporting their distinction. The 640 molars from the mandible MGSB25314 also differ from those of H. laietanus. 641 Nevertheless, they differ in a different way than the aforementioned paratype of H. 642 crusafonti, by displaying the thickest enamel on the external aspect of all cusps. This 643 is consistent with assignment of MGSB25314 to a species other than H. laietanus, but 644 it does not particularly supporting its assignment to H. crusafonti. The taxonomic 645 implications of these comparisons must remain tentative given the small sample of 646 lower molars available for H. crusafonti from the type locality, the lack of enamel 647 distribution maps for mandibular specimens of D. fontani from its type locality (Saint- 648 Gaudens, France), and the fact that a particular specimen of H. laietanus (the invalid 26 649 holotype of Dryopithecus piveteaui nomen nudum) more closely resembles the single 650 examined paratype of H. crusafonti. 651 In turn, the enamel distribution of “S.” occidentalis specimens, characterized by the 652 presence of thicker areas mostly on the buccal half of the crown and thinner lingual 653 enamel further supports their attribution to a single taxon. However, the restricted 654 sample sizes available—in particular, the lack of lower molar distribution maps for 655 both other middle Miocene dryopithecines (particularly P. catalaunicus and A. 656 brevirostris, which might be conspecific with the former; Alba et al. 2020)—and 657 further similarities with some specimens of Hispanopithecus (especially IPS1822) 658 preclude a conclusive assessment of the taxonomic implications of enamel 659 distribution. 660 661 EDJ morphology 662 Contrasting with the traditional emphasis on OES for assessing dental morphology, 663 the usefulness of the EDJ has been recently stressed (Olejniczak et al. 2004; Skinner 664 2008; Skinner et al. 2008a, 2009b, Skinner et al. 2008b, 2009a; Zanolli et al. 2012, 665 2014; Zanolli & Mazurier 2013; Davies et al. 2019; Détroit et al. 2019), given that it 666 provides highly-diagnostic additional information for taxonomic identification (Smith 667 et al. 2006; Skinner et al. 2008a, 2008b, 2009, Zanolli et al. 2012, 2014, 2019); and 668 enables tooth morphology comparisons irrespective of occlusal wear (Tables 3 and 4). 669 The thicker and more inflated crests, secondary enamel folds, and cusp bases that 670 distinguish P. catalaunicus from A. brevisrostris and D. fontani at the OES (Alba et 671 al. 2013; Pérez de los Ríos et al. 2013) are not reflected at the EDJ, thus being 672 probably attributable to the overall thicker enamel of the former. However, P. 673 catalaunicus also differs from these genera in other upper molar features observable at 27 674 the EDJ: from D. fontani, in the deeper trigon basin, the higher and straighter crista 675 obliqua, the M1 hypocone dentine horn more aligned with that of the protocone (in 676 agreement with OES morphology; Alba et al. 2013; Pérez de los Ríos et al. 2013), the 677 less buccolingually waisted upper molars, and the less developed lingual cingulum; 678 and from A. brevirostris, in the shallower and less restricted mesial fovea, the M1 and 679 M2 hypocone horn less lingually situated relative to that of the protocone (in 680 agreement with the OES morphology; Alba et al. 2013; Pérez de los Ríos et al. 2013), 681 the markedly less buccolingually waisted upper molars, and the much less developed 682 lingual cingulum. 683 As previously reported (Alba et al. 2020), the EDJ morphology of the ACM/BCV4 684 specimen supports its conspecificity with the holotype of “S.” occidentalis, only 685 differing in the lack of a tuberculum intermedium in the former—which is variable in 686 H. crusafonti and hence likely atributable to intraspecific variation. These similarities 687 (and those in enamel thickness and distribution mentioned above) strengthen the 688 attribution of the ACM/BCV4 specimen to the same taxon as the holotype of “S.” 689 occidentalis (Alba et al. 2020). The latter specimen (originally consisting of a 690 mandibular fragment, but currently preserved as isolated M2 and M3; Golpe Posse 691 1993) was initially assigned to D. fontani (Villalta Comella & Crusafont Pairó, 1941) 692 but soon thereafter used to erect a new species (Villalta Comella & Crusafont Pairó, 693 1944). Over the years, “S.” occidentalis has been mostly synonimized with 694 Hispanopithecus laietanus (or Dryopithecus laietanus) (Crusafont Pairó & Hürzeler 695 1961; Simons & Pilbeam 1965; Begun et al. 1990; Harrison 1991; Golpe Posse 1993; 696 Ribot et al. 1996), later considered a nomen dubium (Moyà-Solà et al. 2004, 2009a; 697 Casanovas-Vilar et al. 2011; Alba 2012; Marigó et al. 2014; Alba et al. 2017) and 698 recently considered as species inquirenda (Alba et al. 2020). In particular, based on 2D 28 699 RET and both EDJ and OES similarities, it was recently concluded that “S.” 700 occidentalis is not synonymous with D. fontani, but given the lack of well-preserved 701 M2 of A. brevirostris is was not possible to favor a synonymy with P. catalaunicus 702 over the latter species (Alba et al. 2020). The 3D RET results reported above 703 tentatively support the view that “S.” occidentalis represent the otherwise unknown 704 lower dentition of P. catalaunicus, in which case the former species epithet would take 705 priority. On the other hand, we consider it unadvisable to formally synonymize these 706 taxa until an alternate assignment to A. brevirostris can be more convincingly 707 excluded based on additional mandibular material unequivocally assignable to P. 708 catalaunicus (i.e., ideally associated with cranial remains). 709 The EDJ morphology is also informative regarding the alpha-taxonomy of 710 Hispanopithecus, which was erected with H. laietanus as its type species in 1944 711 (Villalta Comella & Crusafont Pairó 1944), but it was later synonymized with 712 Dryopithecus (Simons & Pilbeam 1965). For many years, such synonymy was 713 accepted by most authors (Begun et al. 1990; Harrison 1991; Begun 1992, 2002, 2007; 714 Moyà-Solà & Köhler 1993, 1995, Ribot et al. 1996; Andrews et al. 1996), with only a 715 few exceptions (Golpe Posse 1993; Cameron 1997, 1998, 1999). Indeed, H. crusafonti 716 was originally described within Dryopithecus (Begun 1992), but subsequently 717 reallocated to Hispanopithecus only by a few authors (e.g., Cameron 1999) until the 718 discovery of ACM dryopithecines permitted to re-establish the distinct generic status 719 of Hispanopithecus on a more firm basis (Moyà-Solà et al. 2009a; Begun 2009). This 720 has been subsequently accepted by most authors (Begun 2010, 2015; Moyà-Solà et al. 721 2009b; Casanovas-Vilar et al. 2011; Alba 2012; Alba et al. 2012b, 2013; Alba & 722 Moyà-Solà 2012; Begun et al. 2012; Pérez de los Ríos et al. 2013; Fleagle 2013; 723 Böhme et al. 2019), only with few exceptions (Pickford 2012) regarding H. crusafonti. 29 724 Our results show that the late Miocene H. laietanus and H. crusafonti differ from the 725 investigated middle Miocene taxa in the more peripheral dentine horns of the upper 726 molars (especially in H. laietanus) as well as in the lower and often disrupted crista 727 obliqua. They further differ from P. catalaunicus and D. fontani, but not A. 728 brevirostris, in the deeper and more restricted mesial fovea, and the somewhat more 729 marked buccolingual waisting of the upper molars (although this feature is variable). 730 All these features support the distinction of the genus Hispanopithecus from 731 Dryopithecus and other middle Miocene dryopithecine genera from the Vallès- 732 Penedès Basin. 733 It is noteworthy that the distinction of H. crusafonti from H. laietanus (Begun 734 1992, 2002, 2009; Cameron 1999; Moyà-Solà et al. 2009a; Casanovas-Vilar et al. 735 2011; Alba 2012; Alba & Moyà-Solà 2012; Alba et al. 2012b; Pickford 2012) was 736 questioned (Andrews et al. 1996) or even disputed (Harrison 1991; Ribot et al. 1996) 737 by some authors, who considered the former a junior subjective synonym of the latter. 738 The original diagnosis of H. crusafonti (Begun 1992) mentioned a series of differences 739 in tooth size and shape relative to H. laietanus that were later re-evaluated based on an 740 enlarged sample of upper teeth attributed to the latter species (Alba et al. 2012b). Such 741 re-evaluation concluded that incisor morphology and cheek teeth proportions 742 tentatively supported the distinction of two species, while the development of cingula 743 was too variable to serve as a taxonomically valid criterion. In fact, the original claim 744 that H. crusafonti displays more reduced cingula than H. laietanus (Begun 1992; 745 Cameron 1999) was already rejected based on the OES morphology (Ribot et al. 1996; 746 Alba et al. 2013). In turn, the presence of a more median hypocone and a better 747 developed metacone in the M3 of H. crusafonti was considered potentially diagnostic, 748 although with doubts due to the small available samples and the high variability 30 749 displayed by the M3 of H. laietanus (Alba et al. 2012b). Our assesment of the EDJ 750 morphology suggests that the two latter features are too variable to be diagnostic, but 751 leads us to identify alternative additional diagnostic features in the upper molars, 752 namely the more lingual position of the M1 and M2 hypocone, the less peripheral 753 dentine horns, the deeper trigon basin, and the more developed lingual cinglum in H. 754 crusafonti. It is particularly noteworthy that, at the EDJ, H. crusafonti displays better 755 developed cingula than H. laietanus, contrary to the conclusions in the original 756 description of the former species (Begun 1992), which relied on OES morphology. 757 Concerning the lower molars, several distinctive features noted in the original 758 description (deep and narrow grooves between the buccal cuspids or shallower and 759 more restricted talonid basins; Begun 1992) are not reflected in EDJ morphology. In 760 contrast, H. crusafonti (including the specimen MGSB25314) displays a more 761 developed buccal cingulid at the EDJ and less pronounced buccolingual waisting of 762 the lower molars than H. laietanus. Therefore, the EDJ evidence provided here for 763 both upper and lower molars, coupled with cheek tooth proportions and upper incisor 764 morphology (Alba et al. 2012b), support the distinction of the two Hispanopithecus 765 species. It should also be noted that, as previously noted for the OES (Alba et al. 766 2013), for some features (moderately peripheral upper molar dentine horns) H. 767 crusafonti more closely resembles the middle Miocene hominoids—especially A. 768 brevirostris (deeper trigon basin, lingually-positioned hypocone dentine horn, more 769 developed lingual cingulum)—than H. laietanus does. These features appear 770 plesiomorphic and might indicate a more basal position for H. crusafonti, although 771 additional (especially cranial) remains would be required to further test this 772 hypothesis. 31 773 Finally, although a single dryopithecine species is customarily recognized at Can 774 Llobateres (Begun et al. 1990, Harrison 1991; Begun 1992, 2002; Moyà-Solà & 775 Köhler 1993, 1995; Andrews et al. 1996; Ribot et al. 1996; Cameron 1997, 1999; 776 Casanovas-Vilar et al. 2011; Alba & Moyà-Solà 2012; Alba et al. 2012b), this is 777 worth revising in the light of the EDJ data reported here, particularly given previous 778 proposals that two additional species might be represented at this site (Crusafont Pairó 779 & Hürzeler 1961, 1969; Crusafont-Pairó & Golpe-Posse 1973): a smaller species 780 (“Rahonapithecus sabadellensis”) and a larger one (“Dryopithecus piveteaui”). 781 Although these species are nomenclaturally invalid (nomina nuda) (Simons & Pilbeam 782 1965; Begun et al. 1990; Alba & Moyà-Solà 2012; Alba et al. 2012b), some authors 783 have considered that the purported holotype of “Rahonapithecus sabadellensis” 784 (IPS1802, a mandibular fragment with M1–M3) might belong to the same taxon as the 785 holotype of “S.” occidentalis (Pickford 2012). Even if the M3 of this specimen 786 displays the thickest value among the sample of this locality, this is not the case for 3D 787 RET, and overall the wide range of 2D RET values displayed by H. laietanus conform 788 to the levels of variation displayed by extant great ape species (see Fortuny et al. 2020, 789 table 7 and fig. 6). The lower molars from Can Llobateres 1, despite some variation in 790 EDJ morphology (e.g., in the presence of M2 metaconulid), do not show marked 791 differences that might justify the distinction of additional species and are characterized 792 by the same features (e.g., poorly developed buccal cingulid and marked buccolingual 793 waisting) that distinguish the holotype of H. laietanus from the remaining investigated 794 taxa. The same applies to the EDJ morphology of the purported holotype of “D. 795 piveteaui” (IPS1822, an M3 germ). This specimen merely stands out by its marked 796 development of enamel wrinkling at the OES, which is not reflected in any 32 797 concomitant differences from other H. laietanus specimens at the EDJ and is likely 798 attributable to the lack of wear. 799 800 CONCLUSIONS 801 Our reassessment of Iberian dryopithecine palaeobiodiversity in the light of the 802 reported data on tooth endostructural morphology is consistent with the distinction of 803 four different genera and five species of Miocene dryopithecines in Catalonia—a 804 distinction originally based on cranial morphology and features at the OES. Our 805 conclusions therefore reinforce the view that neither Pierolapithecus, Anoiapithecus or 806 Hispanopithecus can be considered junior synonyms of Dryopithecus. This also holds 807 for “S.” occidentalis, which must remain a species inquirenda until additional lower 808 molars of P. catalaunicus or A. brevirostris enable further comparisons. Our results 809 also show that the two late Miocene species of dryopithecines from Catalonia 810 (included in Hispanopithecus) display several derived features as compared to the 811 middle Miocene taxa. However, H. crusafonti generally retains a more primitive 812 morphology than H. laietanus and is thus more similar to the middle Miocene genera. 813 In the future, other aspects of the internal structure of the post-canine teeth of 814 these Miocene hominids and other extinct apes from Europe will be investigated to 815 extract additional paleobiological information. In particular, analyses of root 816 morphology (e.g., Kupczik and Hublin, 2010; Kupczik et al. 2019; Moore et al. 2013, 817 2016; Pan et al. 2019) and the application of morphometric geometric techniques to 818 quantify the EDJ shape (e.g., Skinner et al. 2009b, 2016; Zanolli et al. 2018, 2019) 819 will further highlight the Miocene hominid diversity. 820 However, pending the recovery and analysis of additional craniodental remains 821 in Europe (such as those of Danuvius; Böhme et al. 2019), this study confirms the 33 822 need for such kind of 3D tooth endostructural analyses in studies aimed at assessing 823 the alpha-taxonomy of fossil apes, as noted by some previous studies (Zanolli et al. 824 2019). The application of these techniques to currently available isolated dentognathic 825 fragments from elsewhere in Europe might ultimately unveil further the previously 826 unrecognized palaeobiodiversity of Miocene apes in this continent. 827 828 Acknowledgments This work has been funded by the Spanish Agencia Estatal de 829 Investigación–European Regional Development Fund of the European Union 830 (CGL2016-76431-P and CGL2017-82654-P, AEI/FEDER-UE), the Generalitat de 831 Catalunya (CERCA Program and consolidated research groups 2017 SGR 086 and 832 2017 SGR 116), the French CNRS, and the Regione Friuli-Venezia Giulia 833 (ICTP/Elettra EXACT Project) in the frame of the SAPIENS Project funded by the 834 Centro Fermi. We thank Sergio Llácer for image processing, and Sebastià Calzada for 835 the loan of specimens housed at the MGSB. We thank Jay Kelley for constructive 836 input on an early version of this manuscript, as well as two anonymous reviewers for 837 their helpful comments on the original version submitted to the journal. 838 839 DATA ARCHIVING STATEMENT 840 841 Data for this study are available in MorphoSource and the Dryad Digital Repository: 842 https://morphosource.org/MyProjects/Dashboard/dashboard/select_project_id/1160 843 844 https://datadryad.org/stash/share/EiJrlF8TAY9FmzkrKqbzIcaiXBO8BD- 845 b1BSPjduY8Eg 846 34 847 REFERENCES 848 ALBA, D.M. 2012. Fossil apes from the Vallès-Penedès Basin. Evolutionary 849 Anthropology, 21, 254–269. 850 ― and MOYÀ-SOLÀ, S. 2012. On the identity of a hominoid male upper canine from 851 the Vallès-Penedès Basin figured by Pickford. Estudios Geológicos, 68, 149– 852 153. 853 ― ― CASANOVAS-VILAR, I., GALINDO, J., ROBLES, J.M., ROTGERS, C., 854 FURIÓ, M., ANGELONE, C., KÖHLER, M., GARCÉS, M., CABRERA, L., 855 ALMÉCIJA, S. and OBRADÓ, P. 2006. Los vertebrados fósiles del Abocador 856 de Can Mata (els Hostalets de Pierola, l'Anoia, Cataluña), una sucesión de 857 localidades del Aragoniense superior (MN6 y MN7+8) de la cuenca del Vallès- 858 Penedès. Campañas 2002–2003, 2004 y 2005. Estudios Geológicos, 62, 295– 859 312. 860 ― ― MALGOSA, A., CASANOVAS-VILAR, I., ROBLES, J.M., ALMÉCIJA, S., 861 GALINDO, J., ROTGERS, C. and BERTÓ MENGUAL, J.V. 2010a. A new 862 species of Pliopithecus Gervais, 1849 (Primates: Pliopithecidae) from the 863 Middle Miocene (MN8) of Abocador de Can Mata (els Hostalets de Pierola, 864 Catalonia, Spain). American Journal of Physical Anthropology, 141, 52–75. 865 ― FORTUNY, J. and MOYÀ-SOLÀ, S. 2010b. Enamel thickness in the Middle 866 Miocene great apes Anoiapithecus, Pierolapithecus and Dryopithecus. 867 Proceedings of the Royal Society B, 277, 2237–2245. 868 ― ALMÉCIJA, S., and MOYÀ-SOLÀ, S. 2010c. Locomotor inferences in 869 Pierolapithecus and Hispanopithecus: Reply to Deane and Begun (2008). 870 Journal of Human Evolution, 59, 143–149. 35 871 ― MOYÀ-SOLÀ, S., and ALMÉCIJA, S. 2011. A partial hominoid humerus from the 872 middle Miocene of Castell de Barberà (Vallès-Penedès Basin, Catalonia, 873 Spain). American Journal of Physical Anthropology, 144, 365–381. 874 ― ― ROBLES, J.M., and GALINDO, J. 2012a. The oldest pliopithecid record in the 875 Iberian Peninsula based on new material from the Vallès-Penedès Basin. 876 American Journal of Physical Anthropology, 147, 135–140. 877 ― CASANOVAS-VILAR, I., ALMÉCIJA, S., ROBLES, J.M., ARIAS- 878 MARTORELL, J. and MOYÀ-SOLÀ, S. 2012b. New dental remains of 879 Hispanopithecus laietanus (Primates: Hominidae) from Can Llobateres 1 and 880 the taxonomy of Late Miocene hominoids from the Vallès-Penedès Basin (NE 881 Iberian Peninsula). Journal of Human Evolution, 63, 231–246. 882 ― ALMÉCIJA, S., CASANOVAS-VILAR, I., MÉNDEZ, J.M., MOYÀ-SOLÀ, S. 883 2012c. A partial skeleton of Hispanopithecus laietanus from Can Feu and the 884 mosaic evolution of crown-hominoid positional behaviors. PLoS ONE, 7, 885 e39617. 886 ― FORTUNY, J., PÉREZ DE LOS RÍOS, M., ZANOLLI, C., ALMÉCIJA, S., 887 CASANOVAS-VILAR, I., ROBLES, J.M. and MOYÀ-SOLÀ, S. 2013. New 888 dental remains of Anoiapithecus and the first appearance datum of hominoids 889 in the Iberian Peninsula. Journal of Human Evolution, 65, 573–584. 890 ― ALMÉCIJA, S., DEMIGUEL, D., FORTUNY, J., PÉREZ DE LOS RÍOS, M., 891 PINA, M., ROBLES, J.M. and MOYÀ-SOLÀ, S. 2015. Miocene small-bodied 892 ape from Eurasia sheds light on hominoid evolution. Science, 350, aab2625. 893 ― CASANOVAS-VILAR, I., GARCÉS, M. and ROBLES, J.M. 2017. Ten years in 894 the dump: An updated review of the Miocene primate-bearing localities from 36 895 Abocador de Can Mata (NE Iberian Peninsula). Journal of Human Evolution, 896 102, 12–20. 897 ― FORTUNY, J., ROBLES, J.M., BERNARDINI, F., PÉREZ DE LOS RÍOS, M., 898 TUNIZ, C., MOYÀ-SOLÀ, S. and ZANOLLI, C. 2020. A new dryopithecine 899 mandibular fragment from the middle Miocene of Abocador de Can Mata and 900 the taxonomic status of ‘Sivapithecus’ occidentalis from Can Vila (Vallès- 901 Penedès Basin, NE Iberian Peninsula. Journal of Human Evolution, 145, 902 102790. 903 ALMÉCIJA, S., ALBA, D. M., MOYÀ-SOLÀ, S., and KÖHLER, M. 2007. Orang- 904 like manual adaptations in the fossil hominoid Hispanopithecus laietanus: first 905 steps towards great ape suspensory behaviours. Proceedings of the Royal 906 Society B, 274, 2375–2384. 907 ― ― ― 2009. Pierolapithecus and the functional morphology of Miocene ape hand 908 phalanges: paleobiological and evolutionary implications. Journal of Human 909 Evolution, 57, 284–297. 910 ― TALLMAN, M., ALBA, D.M., PINA, M., MOYÀ-SOLÀ, S., and JUNGERS, 911 W.L. 2013. The femur of Orrorin tugenensis exhibits morphometric affinities 912 with both Miocene apes and later hominins. Nature Communications, 4, 2888. 913 ANDREWS, P., 2020. Last common ancestor of apes and humans: Morphology and 914 environment. Folia Primatologica, 91, 122–148. 915 ― and MARTIN, L. 1991. Hominoid dietary evolution. Philosophical Transactions of 916 the Royal Society B, 334, 199–209. 917 ― HARRISON, T., DELSON, E., BERNOR, R. L. and MARTIN, L. 1996. 918 Distribution and biochronology of European and Southwest Asian Miocene 919 catarrhines. 168–207. In BERNOR, R.L., FAHLBUSCH, V. and 37 920 MITTMANN, H.-W (eds). The Evolution of Western Eurasian Neogene 921 Mammal Faunas. Columbia University Press, Columbia, 528 pp. 922 BAYLE, P., BONDIOLI, L., MACCHIARELLI, R., MAZURIER, A., PUYMERAIL, 923 L., VOLPATO, V. and ZANOLLI, C. 2011. Three-dimensional imaging and 924 quantitative characterization of human fossil remains. Examples from the 925 Nespos database. 29–46. In MACCHIARELLI, R. and WENIGER G.-C. (eds). 926 Pleistocene Databases. Acquisition, Storing, Sharing. Wissenschaftliche 927 Schriften des Neanderthal Museums, Mettmann 4, 121 pp. 928 BEGUN, D.R. 1992. Dryopithecus crusafonti sp. nov., a new Miocene hominoid 929 species from Can Ponsic (Northeastern Spain). American Journal of Physical 930 Anthropology, 87, 291–309. 931 ― 2002. European hominoids. 339–368. In HARTWIG, W.C. (ed). The Primate 932 Fossil Record. Cambridge University Press, Cambridge, 503 pp. 933 ― 2003. Planet of the apes. Scientific American, 289, 74–83. 934 ― 2007. Fossil record of Miocene hominoids. 921–977. In HENKE, W. and 935 TATTERSALL, I. (eds). Handbook of Paleoanthropology. Springer Verlag, 936 Berlin, 2069 pp. 937 ― 2009. Dryopithecins, Darwin, de Bonis, and the European origin of the African 938 apes and human clade. Geodiversitas, 31, 789–816. 939 ― 2010. Miocene hominids and the origins of the African apes and humans. Annual 940 Review of Anthropology, 39, 67–84. 941 ― 2015. Fossil record of Miocene hominoids. 1261–1332. In HENKE, W. and 942 TATTERSALL, I. (eds). Handbook of Paleoanthropology. Springer Verlag, 943 Berlin, 2624 pp. 38 944 ― MOYA-SOLA, S. and KOHLER, M. 1990. New Miocene hominoid specimens 945 from Can Llobateres (Vallès Penedès, Spain) and their geological and 946 paleoecological context. Journal of Human Evolution, 19, 255–268. 947 ― NARGOLWALLA, M. C. and KORDOS, L. 2012. European Miocene hominids 948 and the origin of the African ape and human clade. Evolutionary Anthropology, 949 21, 10–23. 950 BENAZZI, S., PANETTA, D., FORNAI, C., TOUSSAINT, M., GRUPPIONI, G. and 951 HUBLIN, J.J. 2014. Guidelines for the digital computation of 2D and 3D 952 enamel thickness in hominoid teeth. American Journal of Physical 953 Anthropology, 153, 305–313. 954 BÖHME, M., SPASSOV, N., FUSS, J., TRÖSCHER, A., DEANE, A.S., PRIETO, J., 955 KIRSCHER, U., LECHNER, T. and BEGUN, D.R., 2019. A new Miocene ape 956 and locomotion in the ancestor of great apes and humans. Nature, 575, 489– 957 493. 958 CAMERON, D.W. 1997. A revised systematic scheme for the Eurasian Miocene fossil 959 Hominidae. Journal of Human Evolution, 33, 449–477. 960 ― 1998. Patterns of faciodental sexual dimorphism in Hispanopithecus. Zeitschrift für 961 Morphologie und Anthropologie, 82, 47–58. 962 ― 1999. The single species hypothesis and Hispanopithecus fossils from the Vallés 963 Penedés Basin, Spain. Zeitschrift für Morphologie und Anthropologie, 82, 964 159–186. 965 CASANOVAS-VILAR, I., ALBA, D.M., GARCÉS, M., ROBLES, J.M. and MOYÀ- 966 SOLÀ, S. 2011. Updated chronology for the Miocene hominoid radiation in 967 Western Eurasia. Proceedings of the National Academy of Sciences of the 968 United States of America, 108, 5554–5559. 39 969 ― GARCÉS, M., VAN DAM, J.A., GARCÍA PAREDES, I., ROBLES, J.M. and 970 ALBA, D.M. 2016. An updated biostratigraphy for the late Aragonian and 971 Vallesian of the Vallès-Penedès Basin (Catalonia). Geologica Acta, 14, 195– 972 217. 973 COLEMAN, M.N. and COLBERT, M.W. 2007. CT thresholding protocols for taking 974 measurements on three-dimensional models. American Journal of Physical 975 Anthropology, 133, 723–725. 976 CRUSAFONT-PAIRO, M. and HÜRZELER, J., 1961. Les Pongidés fossiles 977 d'Espagne. Comptes Rendus de l’Académie des Science, Paris 252, 582–584. 978 ― ― 1969. Catálogo comentado de los póngidos fósiles de España. Acta Geologica 979 Hispánica, 4, 44–48. 980 ― and GOLPE-POSSE, J.M. 1973. New pongids from the Miocene of Vallès Penedès 981 Basin (Catalonia, Spain). Journal of Human Evolution, 2, 17–24. 982 DAVIES, T.W., DELEZENE, L. K., GUNZ, P., HUBLIN, J.-J. and SKINNER, M. M. 983 2019. Endostructural morphology in hominoid mandibular third premolars: 984 Discrete traits at the enamel-dentine junction. Journal of Human Evolution, 985 136, 102670. 986 DEMIGUEL, D., ALBA, D.M. and MOYÀ-SOLÀ, S. 2014. Dietary specialization 987 during the evolution of Western Eurasian hominoids and the extinction of 988 European great apes. PLoS One, 9, e97442. 989 DETROIT, F., MIJARES, A.S., CORNY, J., DAVER, G., ZANOLLI, C., DIZON, E., 990 ROBLES, E., GRÜN, R. and PIPER, P.J. 2019. A new species of Homo from 991 the Late Pleistocene of the Philippines. Nature, 568, 181–186. 992 FAJARDO, R.J., RYAN, T.M. and KAPPELMAN J. 2002. Assessing the accuracy of 993 high resolution X-ray computed tomography of primate trabecular bone by 40 994 comparisons with histological sections. American Journal of Physical 995 Anthropology, 118, 1–10. 996 FLEAGLE, J.G. 2013. Primate Adaptation and Evolution, 3rd ed. Academic Press, 997 London, 464 pp. 998 FORTUNY, J., ZANOLLI, C., BERNARDINI, F., TUNIZ, C. and ALBA, D.M. 2020. 999 Data from: Dryopithecine palaeobiodiversity in the Iberian Miocene revisited 1000 on the basis of molar endostructural morphology. Dryad Digital Repository. 1001 https://datadryad.org/stash/share/EiJrlF8TAY9FmzkrKqbzIcaiXBO8BD- 1002 b1BSPjduY8Eg 1003 GOLPE-POSSE, J.M. 1993. Los Hispanopitecos (Primates, Pongidae) de los 1004 yacimientos del Vallès-Penedès (Cataluña, España). II: Descripción del 1005 material existente en el Instituto de Paleontología de Sabadell. Paleontologia i 1006 Evolució, 26–27, 151–224. 1007 GRINE, F.E. 2002. Scaling of tooth enamel thickness, and molar crown size reduction 1008 in modern humans. South African Journal of Science, 98, 503–509. 1009 ― 2005. Enamel thickness of deciduous and permanent molars in modern Homo 1010 sapiens. American Journal of Physical Anthropology, 126, 14–31. 1011 GRINE, F.E. and DAEGLING, D.J. 2017. Functional morphology, biomechanics and 1012 the retrodiction of early hominin diets. Comptes Rendus Palevol, 16, 613–631. 1013 ― and MARTIN, L.B. 1988. Enamel thickness and development in Australopithecus 1014 and Paranthropus. 3–42. In GRINE, F.E (ed). Evolutionary History of the 1015 “Robust” Australopithecines. Aldine de Gruyter, Berlin, 550 pp. 1016 HAMMER, Ø., HARPER, D.A.T. and RYAN, P.D. 2001. PAST: Paleontological 1017 statistics software package for education and data analysis. Palaeontologia 1018 Electronica, 4, 4. 41 1019 HARRISON, T. 1991. Some observations on the Miocene hominoids from Spain. 1020 Journal of Human Evolution, 19, 515–520. 1021 ― and GU, Y. 1999. Taxonomy and phylogenetic relationships of early Miocene 1022 catarrhines from Sihong, China. Journal of Human Evolution, 37, 225–277. 1023 KELLEY, J. DEAN, M.C. and REID, D.J. 2001. Molar growth in the late Miocene 1024 hominoid, Dryopithecus laietanus. 123–134. In BROOK, A. (ed). Dental 1025 Morphology 2001: 12th International Symposium on Dental Morphology. 1026 Sheffield Academic Press, Sheffield, 350 pp. 1027 KONO, R. 2004. Molar enamel thickness and distribution patterns in extant great apes 1028 and humans: New insights based on a 3-dimensional whole crown perspective. 1029 Anthropological Science, 112, 121–146. 1030 ― and SUWA, G. 2008. Enamel distribution patterns of extant human and hominoid 1031 molars: Occlusal versus lateral enamel thickness. Bulletin of the National 1032 Museum of Nature and Science, Series D, 34, 1–9. 1033 KUPCZIK, K. and HUBLIN, J.J. 2010. Mandibular molar root morphology in 1034 Neanderthals and Late Pleistocene and recent Homo sapiens. Journal of 1035 Human Evolution, 59, 525–541. 1036 ― DELEZENE, L.K. and SKINNER, M.M. 2019. Mandibular molar root and pulp 1037 cavity morphology in Homo naledi and other Plio-Pleistocene hominins. 1038 Journal of Human Evolution, 130, 83–95. 1039 MACCHIARELLI, R., BONDIOLI, L. and MAZURIER, A. 2008. Virtual dentitions: 1040 touching the hidden evidence. 426–448. In IRISH, J.D., NELSON, G.C. (eds), 1041 Technique and Application in Dental Anthropology. Cambridge University 1042 Press, Cambridge, 470 pp. 42 1043 ― MAZURIER, M., ILLERHAUS, B. and ZANOLLI, C. 2009. Ouranopithecus 1044 macedoniensis: virtual reconstruction and 3D analysis of a juvenile mandibular 1045 dentition (RPl-82 and RPl-83). Geodiversitas, 31, 851–863. 1046 ― BAYLE, P., BONDIOLI, L., MAZURIER, A., ZANOLLI, C. 2013. From outer to 1047 inner structural morphology in dental anthropology: integration of the third 1048 dimension in the visualization and quantitative analysis of fossil remains. 250– 1049 277. In SCOTT, G.R. and IRISH, J.D. (eds), Anthropological Perspectives on 1050 Tooth Morphology. Genetics, Evolution, Variation. Cambridge University 1051 Press, Cambridge, 582 pp. 1052 MACHO, G.A. 1994. Variation in enamel thickness and cusp area within human 1053 maxillary molars and its bearing on scaling techniques used for studies of 1054 enamel thickness between species. Archives of Oral Biology, 39, 783–792. 1055 MARIGÓ, J., SUSANNA, I., MINWER-BARAKAT, R., MADURELL- 1056 MALAPEIRA, J., MOYÀ-SOLÀ, S., CASANOVAS-VILAR, I., ROBLES, 1057 J.M. and ALBA, D.M. 2014. The primate fossil record in the Iberian 1058 Peninsula. Journal of Iberian Geology, 40, 179–211. 1059 MARTIN, L.B. 1985. Significance of enamel thickness in hominoid evolution. Nature 1060 314, 260–263. 1061 MARTIN-FRANCÉS, L., MARTINÓN-TORRES, M., MARTÍNEZ DE PINILLOS, 1062 M., GARCÍA-CAMPOS, C., MODESTO-MATA, M., ZANOLLI, C., 1063 RODRÍGUEZ, L. and BERMUDEZ DE CASTRO, J.M. 2018. Tooth crown 1064 tissue proportions and enamel thickness in Early Pleistocene Homo antecessor 1065 molars (Atapuerca, Spain). PLoS ONE, 13, e0203334. 43 1066 MOORE, N.C., SKINNER, M.M. and HUBLIN, J.J. 2013. Premolar root morphology 1067 and metric variation in Pan troglodytes verus. American Journal of Physical 1068 Anthropology, 150, 632–646. 1069 ― THACKERAY, J.F., HUBLIN, J.J. and SKINNER, M.M. 2016. Premolar root and 1070 canal variation in South African Plio-Pleistocene specimens attributed to 1071 Australopithecus africanus and Paranthropus robustus. Journal of Human 1072 Evolution, 93, 46–62. 1073 MOYÀ-SOLÀ, S. and KÖHLER, M. 1993. Recent discoveries of Dryopithecus shed 1074 new light on evolution of great apes. Nature, 365, 543–545. 1075 ― ― 1995. New partial cranium of Dryopithecus Lartet, 1863 (Hominoidea, 1076 Primates) from the upper Miocene of Can Llobateres, Barcelona, Spain. 1077 Journal of Human Evolution, 29, 101–139. 1078 ― ― ALBA, D.M., CASANOVAS-VILAR, I. and GALINDO, J. 2004. 1079 Pierolapithecus catalaunicus, a new Middle Miocene great ape from Spain. 1080 Science, 306, 1339–1344. 1081 ― ― ALBA, D.M., CASANOVAS-VILAR, I. and GALINDO, J. 2005. Response to 1082 comment on "Pierolapithecus catalaunicus, a new Middle Miocene great ape 1083 from Spain". Science, 308, 203d. 1084 ― ALBA, D.M., ALMÉCIJA, S., CASANOVAS-VILAR, I., KÖHLER, M., DE 1085 ESTEBAN-TRIVIGNO, S., ROBLES, J.M., GALINDO, J. and FORTUNY, J. 1086 2009a. A unique Middle Miocene European hominoid and the origins of the 1087 great ape and human clade. Proceedings of the National Academy of Sciences 1088 of the United States of America, 106, 9601–9606. 1089 ― KÖHLER, M., ALBA, D.M., CASANOVAS-VILAR, I., GALINDO, J., ROBLES, 1090 J.M., CABRERA, L., GARCÉS, M., ALMÉCIJA, S. and BEAMUD, E. 44 1091 2009b. First partial face and upper dentition of the Middle Miocene hominoid 1092 Dryopithecus fontani from Abocador de Can Mata (Vallès-Penedès Basin, 1093 Catalonia, NE Spain): taxonomic and phylogenetic implications. American 1094 Journal of Physical Anthropology, 139, 126–145. 1095 OLEJNICZAK, A.J. 2006. Micro-computed tomography of primate molars. 1096 Unpublished PhD thesis, Stony Brook University, New York, 242 pp. 1097 ― MARTIN, L.B. and ULHAAS, L. 2004. Quantification of dentine shape in 1098 anthropoid primates. Annals of Anatomy, 186, 479–486. 1099 ― TAFFOREAU, P., FEENEY, R.N.M. and MARTIN, L.B. 2008a. Three- 1100 dimensional primate molar enamel thickness. Journal of Human Evolution, 54, 1101 187–195. 1102 ― ― ― ― 2008b. Three-dimensional molar enamel distribution and thickness in 1103 Australopithecus and Paranthropus. Biology Letters, 4, 406–410. 1104 ― SMITH, T.M., WANG, W., POTTS, R., CIOCHON, R., KULLMER, O., 1105 SCHRENK, F. AND HUBLIN, J.-J. 2008c. Molar enamel thickness and 1106 dentine horn height in Gigantopithecus blacki. American Journal of Physical 1107 Anthropology, 135, 85–91. 1108 ― ― T.M., FEENEY, R.N., MACCHIARELLI, R., MAZURIER, A., BONDIOLI, 1109 L., ROSAS, A., FORTEA, J., DE LA RASILLA, M., GARCIA- 1110 TABERNERO, A., RADOVCIC, J., SKINNER, M.M., TOUSSAINT M., and 1111 HUBLIN, J.-J. 2008d. Dental tissue proportions and enamel thickness in 1112 Neandertal and modern human molars. Journal of Human Evolution, 55, 12– 1113 23. 45 1114 PAN, L., DUMONCEL, J., MAZURIER, A. and ZANOLLI, C. 2019. Structural 1115 analysis of premolar roots in Middle Pleistocene hominins from China. Journal 1116 of Human Evolution, 136, 102669 1117 PÉREZ DE LOS RÍOS, M., MOYÀ-SOLÀ, S. and ALBA, D.M. 2012. The nasal and 1118 paranasal architecture of the Middle Miocene ape Pierolapithecus catalaunicus 1119 (Primates: Hominidae): Phylogenetic implications. Journal of Human 1120 Evolution, 63, 497–506. 1121 ― ALBA, D.M. and MOYÀ-SOLÀ, S., 2013. Taxonomic attribution of the La Grive 1122 hominoid teeth. American Journal of Physical Anthropology, 151, 558–565. 1123 PICKFORD, M. 2012. Hominoids from Neuhausen and other Bohnerz localities, 1124 Swabian Alb, Germany: evidence for a high diversity of apes in the Late 1125 Miocene of Germany. Estudios Geológicos, 68, 113–147. 1126 PINA, M., ALBA, D.M., ALMÉCIJA, S., FORTUNY, J., and MOYÀ-SOLA, S. 1127 2012. Brief Communication: Paleobiological inferences on the locomotor 1128 repertoire of extinct hominoids based on femoral neck cortical thickness: the 1129 fossil great ape Hispanopithecus laietanus as a test-case study. American 1130 Journal of Physical Anthropology, 149, 142–148. 1131 ― ― MOYÀ-SOLÀ, S., and ALMÉCIJA, S. 2019. Femoral neck cortical bone 1132 distribution of dryopithecin apes and the evolution of hominid locomotion. 1133 Journal of Human Evolution, 136, 102651. 1134 RIBOT, F., GIBERT, J. and HARRISON, T. 1996. A reinterpretation of the taxonomy 1135 of Dryopithecus from Vallès-Penedès, Catalonia (Spain). Journal of Human 1136 Evolution, 31, 129–141. 1137 SCHNEIDER, C.A., RASBAND, W.S. and ELICEIRI, K.W. 2012. NIH Image to 1138 ImageJ: 25 years of image analysis. Nature Methods, 9, 671–675. 46 1139 SCHWARTZ, G.T. 2000. Taxonomic and functional aspects of the patterning of 1140 enamel thickness distribution in extant large-bodied hominoids. American 1141 Journal of Physical Anthropology, 111, 221–244. 1142 SCOTT, G.R and IRISH J.D. 2017. Human Tooth Crown and Root Morphology. The 1143 Arizona State University Dental Anthropology System. Cambridge University 1144 Press, Cambridge, 342 pp. 1145 SIMONS, E. L. and PILBEAM, D. R. 1965. Preliminary revision of the 1146 Dryopithecinae (Pongidae, Anthropoidea). Folia Primatologica, 3, 81–152. 1147 SKINNER, M. M. 2008. Enamel-dentine junction morphology of extant hominoid and 1148 fossil hominin lower molars. Unpublished Ph.D thesis, The George 1149 Washington University, Washington D.C., 191 pp. 1150 ― WOOD, B.A., BOESCH, C., OLEJNICZAK, A., ROSAS, A., SMITH, T.M., and 1151 HUBLIN, J.-J. 2008a. Dental trait expression at the enamel-dentine junction of 1152 lower molars in extant and fossil hominoids. Journal of Human Evolution, 54, 1153 173–186. 1154 ― GUNZ, P., WOOD, B. A. and HUBLIN, J.-J. 2008b. Enamel-dentine junction 1155 (EDJ) morphology distinguishes the lower molars of Australopithecus 1156 africanus and Paranthropus robustus. Journal of Human Evolution, 55, 979– 1157 988. 1158 ― WOOD, B.A. and HUBLIN, J.-J. 2009a. Protostylid expression at the enamel- 1159 dentine junction and enamel surface of mandibular molars of Paranthropus 1160 robustus and Australopithecus africanus. Journal of Human Evolution, 56, 76– 1161 85. 1162 ― GUNZ, P., WOOD, B. A., BOESCH, C. and HUBLIN, J.-J. 2009b. Discrimination 1163 of extant Pan species and subspecies using the enamel-dentine junction 47 1164 morphology of lower molars. American Journal of Physical Anthropology, 1165 140, 234–243. 1166 ― DE VRIES, D., GUNZ, P., KUPCZIK, K., KLASSEN, R.P., HUBLIN, J.J. and 1167 ROKSANDIC, M. 2016. A dental perspective on the taxonomic affinity of the 1168 Balanica mandible (BH-1). Journal of Human Evolution, 93, 63–81. 1169 SMITH, H.B. 1984. Patterns of molar wear in hunter-gatherers and agriculturalists. 1170 American Journal of Physical Anthropology, 63, 39–56. 1171 SMITH, T.M., OLEJNICZAK, A.J., MARTIN, L.B. and REID, D.J. 2005. Variation 1172 in hominoid molar enamel thickness. Journal of Human Evolution, 48, 575– 1173 592. 1174 ― ― REID, D.J., FERRELL, R.J. and HUBLIN J.-J. 2006. Modern human molar 1175 enamel thickness and enamel-dentine junction shape. Archives of Oral Biology, 1176 51, 974–995. 1177 ― ― REH, S., REID, D.J. and HUBLIN, J.-J. 2008. Enamel thickness trends in the 1178 dental arcade of humans and chimpanzees. American Journal of Physical 1179 Anthropology, 136, 237–241. 1180 ― ― ZERMENO, J.P., TAFFOREAU, P., SKINNER, M.M., HOFFMANN, A., 1181 RADOVČIĆ, J., TOUSSAINT, M., KRUSZYNSKI, R., MENTER, C., 1182 MOGGI-CECCHI, J., GLASMACHER, U.A., KULLMER, O., SCHRENK, 1183 F., STRINGER, C. and HUBLIN, J.J. 2012a. Variation in enamel thickness 1184 within the genus Homo Journal of Human Evolution, 62, 395–411 1185 ― KUPCZIK, K., MACHANDA, Z., SKINNER, M.M. and ZERMENO, J.P. 2012b. 1186 Enamel thickness in Bornean and Sumatran orangutan dentitions. American 1187 Journal of Physical Anthropology, 147, 417–426. 48 1188 ― TAFFOREAU, P., POUECH, J. and BEGUN, D.R. 2019. Enamel thickness and 1189 dental development in Rudapithecus hungaricus. Journal of Human Evolution, 1190 136, 102649. 1191 SPOOR, F., ZONNEVELD, F. and MACHO, G.A. 1993. Linear measurements of 1192 cortical bone and dental enamel by computed tomography: applications and 1193 problems. American Journal of Physical Anthropology, 91, 469–484. 1194 SUWA, G. and KONO, R. T. 2005. A micro-CT based study of linear enamel 1195 thickness in the mesial cusp section of human molars: Reevaluation of 1196 methodology and assessment of within-tooth, serial, and individual variation. 1197 Anthropological Sciencies, 113, 273–289. 1198 SUWA, G., KONO, R.T., SIMPSON, S.W., ASFAW, B., LOVEJOY, C.O. and 1199 WHITE, T. 2009. Paleobiological implications of the Ardipithecus ramidus 1200 dentition. Science, 326, 94–99. 1201 SWINDLER, D.R. 2002. Primate Dentition. An Introduction to the Teeth of Non- 1202 Human Primates. Cambridge University Press, Cambridge, 296 pp. 1203 THIERY, G., LAZZARI, V., RAMDARSHAN, A. and GUY, F. 2017. Beyond the 1204 map: Enamel distribution characterized from 3D dental topography. Frontiers 1205 in Physiology, 8, 524. 1206 TURNER, C.G. NICHOL, C.R. and SCOTT, G.R. 1991. Scoring procedures for key 1207 morphological traits of the permanent dentition: The Arizona State University 1208 Dental Anthropology System. 13–31. In KELLEY, L., KELLEY, M. and 1209 LARSEN, C.S. (eds), Advances in Dental Anthropology. Wiley-Liss, New 1210 York, 389 pp. 1211 TUNIZ, C., BERNARDINI, F., CICUTTIN, A., CRESPO, M.-L., DREOSSI, D., 1212 GIANONCELLI, A., MANCINI, L., MENDOZA CUEVAS, A., SODINI, N., 49 1213 TROMBA, G., ZANINI, F. and ZANOLLI, C. 2013.The ICTP-Elettra X-ray 1214 laboratory for cultural heritage and archaeology. A facility for training and 1215 education in the developing world. Nuclear Instruments and Methods in 1216 Physics Research Section A, 711, 106–110. 1217 VAN DER MADE, J. and RIBOT, F. 1999. Additional hominoid material from the 1218 Miocene of Spain and remarks on hominoid dispersals into Europe. 1219 Contributions to Tertiary and Quaternary Geology, 36, 25–39. 1220 VILLALTA COMELLA, J.F. DE and CRUSAFONT PAIRO, M. 1941. Hallazgo del 1221 "Dryopithecus fontani", Lartet, en el Vindoboniense de la cuenca Vallés- 1222 Penedés. Boletín Instituto Geológico y Minero de España, 55, 131–142. 1223 ― ― 1944. Dos nuevos antropomorfos del Mioceno español y su situación dentro de 1224 la moderna sistemática de los símidos. Notas y Comunicaciones del Instituto 1225 Geológico y Minero de España, 13, 1–51. 1226 VOGEL, E.R., VAN WOERDEN, J.T., LUCAS, P.W., UTAMI ATMOKO, S.S., 1227 VAN SCHAIK, C.P. and DOMINY, N.J. 2008. Functional ecology and 1228 evolution of hominoid molar enamel thickness: Pan troglodytes schweinfurthii 1229 and Pongo pygmaeus wurmbii. Journal of Human Evolution, 55, 60–74. 1230 ZANOLLI, C., BONDIOLI, L., MANCINI, L., MAZURIER, A., WIDIANTO, H. and 1231 MACCHIARELLI, R. 2012. Two human fossil deciduous molars from the 1232 Sangiran Dome (Java, Indonesia): Outer and inner morphology. American 1233 Journal of Physical Anthropology, 147, 472–481. 1234 ― and MAZURIER, A. 2013. Endostructural characterization of the H. 1235 heidelbergensis dental remains from the early Middle Pleistocene site of 1236 Tighenif, Algeria. Comptes Rendus Palevol, 12, 293–304. 50 1237 ― BONDIOLI, L., COPPA, A., DEAN, C.M., BAYLE, P., CANDILIO, F., 1238 CAPUANI, S., DREOSSI, D., FIORE, I., FRAYER, D.W., LIBSEKAL, Y., 1239 MANCINI, L., ROOK, L., MEDIN TEKLE, T., TUNIZ, C. and 1240 MACCHIARELLI, R. 2014. The late Early Pleistocene human dental remains 1241 from Uadi Aalad and Mulhuli-Amo (Buia), Eritrean Danakil: 1242 macromorphology and microstructure. Journal of Human Evolution, 74, 96– 1243 113. 1244 ― DEAN, C.M., ROOK, L., BONDIOLI, L., MAZURIER, A. and 1245 MACCHIARELLI, R. 2016. Enamel thickness and enamel growth in 1246 Oreopithecus: Combining microtomographic and histological evidence. 1247 Comptes Rendus Palevol, 15, 209–226. 1248 ― PAN, L., DUMONCEL, J., KULLMER, O., KUNDRAT, M., LIU, W., 1249 MACCHIARELLI, R., MANCINI, L., SCHRENK, F. and TUNIZ, C., 2018. 1250 Inner tooth morphology of Homo erectus from Zhoukoudian. New evidence 1251 from an old collection housed at Uppsala. Journal of Human Evolution, 116, 1252 1–13. 1253 ― KULLMER, O., KELLEY, J., BACON, A.-M., DEMETER, F., DUMONCEL, J., 1254 FIORENZA, L., GRINE, F.E., HUBLIN, J.-J., TUAN NGUYEN, A., HUONG 1255 NGUYEN, T.M., PAN, L., SCHILLINGER, B., SCHRENK, F., SKINNER, 1256 M., JI, X. and MACCHIARELLI, R. 2019. Evidence for increased hominid 1257 diversity in the Early to Middle Pleistocene of Indonesia. Nature Ecology and 1258 Evolution, 3, 755–764. 1259 1260 FIGURE CAPTIONS 51 1261 Fig. 1. Boxplots comparing relative enamel thickness among dryopithecine species 1262 recorded in NE Iberian Peninsula: (A) 2D RET; (B) 3D RET. Only taxa represented 1263 by at least three specimens are depicted. Boxes represent the interquartile range (IQR; 1264 25th and 75th percentiles), centerline is median, whiskers denote the maximum and 1265 minimum values within 1.5 times the IQR, dots are outliers, and stars represent 1266 extreme outliers. Abbreviations: PC, Pierolapithecus catalaunicus; DF, Dryopithecus 1267 fontani; AB, Anoiapithecus brevirostris; HC, Hispanopithecus crusafonti; HL, 1268 Hispanopithecus laietanus; SO, “Sivapithecus” occidentalis species inquirenda. 1269 1270 Fig. 2. Upper molar enamel distribution maps of Iberian dryopithecines. (A–C) 1271 Pierolapithecus catalaunicus (IPS21350, holotype): L M1 (A), L M2 (B), and L M3 1272 (C); (D) Dryopithecus fontani (MGSB48486): R M2; (E–G) D. fontani (IPS35026): L 1273 M1 (E), L M2 (F) and L M3 (G); (H–L) Anoiapithecus brevirostris (IPS43000, 1274 holotype): R M1 (H), R M2 (I), R M3 (J), L M1 (K) and L M2 (L); (M–P) A. 1275 brevirostris (IPS35027): L M1 (M), L M2 (N), R M1 (O) and R M2 (P); (Q–U) 1276 Hispanopithecus crusafonti (paratypes): L M1 IPS1818 (Q), R M1 IPS1815 (R), L M2 1277 IPS1820 (S), R M3 IPS1812 (T) and R M3 IPS1814 (U); (V–Y) Hispanopithecus 1278 laietanus: R M2 IPS1844 (V), L M2 IPS58339 (W), L M3 IPS58340 (X) and L M3 1279 IPS1772 (Y). Each tooth has its own color scale of enamel thickness, ranging from 0 1280 to maximum thickness (indicated in mm above the scale). Scale bar represents 5 mm. 1281 1282 Fig. 3. Lower molar enamel distribution maps of Iberian dryopithecines. (A) 1283 “Sivapithecus” occidentalis species inquirenda (IPS41734): R M2; (B–C) “S.” 1284 occidentalis (holotype): L M2 IPS1826+1827 (B) and L M3 IPS1826+1827 (C); (D–F) 1285 Anoiapithecus brevirostris (IPS43000, holotype): L M1 (D), L M2 (E), R M1 (F); (G) 52 1286 Hispanopithecus crusafonti (IPS1816, paratype): R M2; (H–J) H. crusafonti 1287 (MGSB25314): L M1 (H), L M2 (I) and L M3 (J); (L–M) Hispanopithecus laietanus 1288 (IPS1804, holotype): L M2 (L) and L M3 (M); (K, N–S) H. laietanus: R M2 IPS1780 1289 (K), R M1 IPS1797 (N), R M2 IPS1797 (O), R M1 IPS1802 (P), R M2 IPS1802 (Q), R 1290 M3 IPS1802 (R) and L M3 IPS1822 (S). Each tooth has its own color scale of enamel 1291 thickness, ranging from 0 to maximum thickness (indicated in mm above the scale). 1292 Scale bar represents 5 mm. 1293 1294 Fig. 4. Upper molar enamel-dentine junction (EDJ) Iberian dryopithecines. (A–C) 1295 Pierolapithecus catalaunicus (IPS21350, holotype): L M1 (A), L M2 (B), and L M3 1296 (C); (D) Dryopithecus fontani (MGSB48486): R M2; (E–G) D. fontani (IPS35026): L 1297 M1 (E), L M2 (F) and L M3 (G); (H–L) Anoiapithecus brevirostris (IPS43000, 1298 holotype): R M1 (H), R M2 (I), R M3 (J), L M1 (K) and L M2 (L); (M–P) A. 1299 brevirostris (IPS35027): L M1 (M), L M2 (N), R M1 (O) and R M2 (P); (Q–U) 1300 Hispanopithecus crusafonti (paratypes): L M1 IPS1818 (Q), R M1 IPS1815 (R), L M2 1301 IPS1820 (S), R M3 IPS1812 (T) and R M3 IPS1814 (U); (V–Y) Hispanopithecus 1302 laietanus: R M2 IPS1844 (V), L M2 IPS58339 (W), L M3 IPS58340 (X) and L M3 1303 IPS1772 (Y). Scale bar represents 5 mm. 1304 1305 Fig. 5. Lower molar enamel-dentine junction (EDJ) of Iberian drypithecines. (A) 1306 “Sivapithecus” occidentalis species inquirenda (IPS41734): R M2; (B–C) “S.” 1307 occidentalis (holotype): L M2 IPS1826+1827 (B) and L M3 IPS1826+1827 (C); (D–F) 1308 Anoiapithecus brevirostris (IPS43000, holotype): L M1 (D), L M2 (E), RM1 (F); (G) 1309 Hispanopithecus crusafonti (IPS1816, paratype): R M2; (H–J) H. crusafonti 1310 (MGSB25314): L M1 (H), L M2 (I) and L M3 (J); (L–M) Hispanopithecus laietanus 53 1311 (IPS1804, holotype): L M2 (L) and L M3 (M); (K, N–S) H. laietanus: R M2 IPS1780 1312 (K), R M1 IPS1797 (N), R M2 IPS1797 (O), R M1 IPS1802 (P), R M2 IPS1802 (Q), R 1313 M3 IPS1802 (R) and L M3 IPS1822 (S). Scale bar represents 5 mm. 1314 1315 Table 1. Studied samples of Miocene hominoids from Catalonia. 1316 1317 Table 2. Descriptive statistics for relative enamel thickness. 1318 1319 Table 3. Qualitative dental features of the EDJ of the upper molars. 1320 1321 Table 4. Qualitative dental features of the EDJ of the lower molars. 54 Table 1 Studied samples of Miocene hominoids from Catalonia. Catalog No. Tooth Weara Locality Age Taxon Remarks IPS1772 L M3 1–2 CLL1 9.7 Ma Hispanopithecus laietanus — IPS1780 R M2 1–2 CLL1 9.7 Ma Hispanopithecus laietanus — IPS1797 R M1 2–3 CLL1 9.7 Ma Hispanopithecus laietanus — IPS1797 R M2 1–2 CLL1 9.7 Ma Hispanopithecus laietanus — Invalid holotype of IPS1802 R M1 3 CLL1 9.7 Ma Hispanopithecus laietanus “Rahonapithecus sabadellensis” (nomen nudum) Invalid holotype of IPS1802 R M2 1–2 CLL1 9.7 Ma Hispanopithecus laietanus “Rahonapithecus sabadellensis” (nomen nudum) Invalid holotype of IPS1802 R M3 1–2 CLL1 9.7 Ma Hispanopithecus laietanus “Rahonapithecus sabadellensis” (nomen nudum) IPS1804 L M2 2 LT 9.5 Ma Hispanopithecus laietanus Holotype IPS1804 L M3 1–2 LT 9.5 Ma Hispanopithecus laietanus Holotype IPS1812 R M3 1 CP1 10.4–10.0 Ma Hispanopithecus crusafonti Paratype 3 IPS1814 RM 1–2 CP1 10.4–10.0 Ma Hispanopithecus crusafonti Paratype 1 IPS1815 LM 1–2 CP1 10.4–10.0 Ma Hispanopithecus crusafonti Paratype IPS1816 R M2 2 CP1 10.4–10.0 Ma Hispanopithecus crusafonti Paratype IPS1818 L M1 1–2 CP1 10.4–10.0 Ma Hispanopithecus crusafonti Paratype IPS1820 L M2 1–2 CP1 10.4–10.0 Ma Hispanopithecus crusafonti Paratype 1 Invalid holotype of IPS1822 L M3 CLL1 9.7 Ma Hispanopithecus laietanus “Dryopithecus piveteaui” (nomen nudum) “Sivapithecus” occidentalis species Holotype of “Sivapithecus” IPS1826 L M2 1 CV 12.5-11.9 Ma inquirenda occidentalis “Sivapithecus” occidentalis species Holotype of “Sivapithecus” IPS1827 L M3 1 CV 12.5-11.9 Ma inquirenda occidentalis IPS1844 R M1 1 CLL1 9.7 Ma Hispanopithecus laietanus — IPS18000.5 R M1 4 CLL2 9.6 Ma Hispanopithecus laietanus — 2 IPS18000.5 RM 2 CLL2 9.6 Ma Hispanopithecus laietanus — 3 IPS18000.5 RM 1–2 CLL2 9.6 Ma Hispanopithecus laietanus — IPS21350 L M1 1–2 ACM/BCV1 11.9 Ma Pieralopithecus catalaunicus Holotype IPS21350 L M2 1–2 ACM/BCV1 11.9 Ma Pieralopithecus catalaunicus Holotype 3 IPS21350 LM 2 ACM/BCV1 11.9 Ma Pieralopithecus catalaunicus Holotype IPS21350 R M1 2 ACM/BCV1 11.9 Ma Pieralopithecus catalaunicus Holotype IPS21350 R M2 1–2 ACM/BCV1 11.9 Ma Pieralopithecus catalaunicus Holotype IPS35026 L M1 3 ACM/C3-Ae 11.9 Ma Dryopithecus fontani — 2 IPS35026 LM 2 ACM/C3-Ae 11.9 Ma Dryopithecus fontani — IPS35026 L M3 1–2 ACM/C3-Ae 11.9 Ma Dryopithecus fontani — IPS35027 R M1 3 ACM/C1-E* 12.3–12.2 Ma Anoiapithecus brevirostris — IPS35027 R M2 1–2 ACM/C1-E* 12.3–12.2 Ma Anoiapithecus brevirostris — IPS35027 L M1 3 ACM/C1-E* 12.3–12.2 Ma Anoiapithecus brevirostris — IPS35027 L M2 1–2 ACM/C1-E* 12.3–12.2 Ma Anoiapithecus brevirostris — “Sivapithecus” occidentalis species IPS41734 R M2 1 ACM/BCV4 11.9 Ma — inquirenda IPS43000 R M1 1 ACM/C3-Aj 11.9 Ma Anoiapithecus brevirostris Holotype IPS43000 R M2 1–2 ACM/C3-Aj 11.9 Ma Anoiapithecus brevirostris Holotype 3 IPS43000 RM 1 ACM/C3-Aj 11.9 Ma Anoiapithecus brevirostris Holotype 1 IPS43000 LM 2 ACM/C3-Aj 11.9 Ma Anoiapithecus brevirostris Holotype IPS43000 L M2 1–2 ACM/C3-Aj 11.9 Ma Anoiapithecus brevirostris Holotype IPS43000 L M1 3 ACM/C3-Aj 11.9 Ma Anoiapithecus brevirostris Holotype IPS43000 L M2 4 ACM/C3-Aj 11.9 Ma Anoiapithecus brevirostris Holotype IPS43000 R M1 3 ACM/C3-Aj 11.9 Ma Anoiapithecus brevirostris Holotype 1 Likely same individual as IPS58339 L M2 CLL1 9.7 Ma Hispanopithecus laietanus IPS58340 1 Likely same individual as IPS58340 L M3 CLL1 9.7 Ma Hispanopithecus laietanus IPS58339 MGSB48486 R M2 1 HP 12.5–9.7 Ma Dryopithecus fontani — MGSB25314 L M1 4 TF 10.4–10.0 Ma Hispanopithecus crusafonti — MGSB25314 L M2 3–4 TF 10.4–10.0 Ma Hispanopithecus crusafonti — MGSB25314 L M3 3 TF 10.4–10.0 Ma Hispanopithecus crusafonti — Abbreviations: L, left; R, right; IPS = ‘Institut de Paleontologia de Sabadell’, former name of Institut Català de Paleontologia Miquel Crusafont; MGSB = Museu de Geologia del Seminari de Barcelona; ACM, local stratigraphic series of Abocador de Can Mata (els Hostalets de Pierola); C3, Cell 3 (ACM sector); BCV, Barranc de Can Vila (ACM sector); CLL, Can Llobateres (Sabadell); CP, Can Poncic (Sant Quirze); CV, Can Vila (els Hostalets de Pierola); HP, Hostalets de Pierola indet. (els Hostalets de Pierola); LT, La Tarumba; TF, Teuleria del Firal (Seu d’Urgell). Numbers after acronyms refer to stratigraphic levels within a single site, whereas other alphanumeric combinations after ACM sectors (separated by a dash) refer to localities within subsectors. a Wear stages are adapted from Smith (1984). Table 2. Descriptive statistics for relative enamel thickness. The minimum number of individuals (MNI) for each taxon is provided. 2D RET Taxon (localities)a N MNI Mean SD SE 95% CI 95% CI Minimum Maximum Pierolapithecus catalaunicus (ACM/BCV1) 5 1 15.36 1.15 0.51 14.35 16.37 13.99 16.88 Dryopithecus fontani (ACM/C3-Ae, HP) 3 2 12.33 0.71 0.41 10.56 14.10 11.55 12.95 Anoiapithecus brevirostris (ACM/C3-Aj, ACM/C1-E*) 9 2 14.58 1.35 0.45 13.70 15.46 12.97 17.32 Hispanopithecus crusafonti (CP1, TF) 7 3 14.39 1.26 0.48 13.46 15.33 13.03 16.76 Hispanopithecus laietanus (CLL1, CLL2, LT)b 17 5 14.34 2.77 0.67 13.02 15.66 10.35 19.11 "Sivapithecus" occidentalis (ACM/CV, ACM/BCV4) 3 2 19.66 2.35 1.36 17.01 22.32 17.43 22.11 3D RET Taxon (localities)a N MNI Mean SD SE 95% CI 95% CI Minimum Maximum Pierolapithecus catalaunicus (ACM/BCV1) 4 1 15.46 2.03 1.02 13.47 17.45 13.24 18.14 Dryopithecus fontani (ACM/C3-Ae, HP) 3 2 11.87 0.93 0.54 10.83 12.92 10.96 12.81 Anoiapithecus brevirostris (ACM/C3-Aj, ACM/C1-E*) 7 2 12.93 1.11 0.42 12.11 13.75 11.40 14.26 Hispanopithecus crusafonti (CP1, TF) 4 2 12.06 1.51 0.76 10.58 13.53 10.97 14.27 Hispanopithecus laietanus (CLL1, CLL2, LT) 10 3 13.46 2.18 0.69 12.11 14.82 10.49 16.38 "Sivapithecus" occidentalis (ACM/CV) 3 2 18.96 1.74 1.00 16.99 20.92 17.16 20.63 Abbreviations: ACM, local stratigraphic series of Abocador de Can Mata (els Hostalets de Pierola); C3, Cell 3 (ACM sector); BCV, Barranc de Can Vila (ACM sector); CLL, Can Llobateres (Sabadell); CP, Can Poncic (Sant Quirze); CV, Can Vila (els Hostalets de Pierola); HP, Hostalets de Pierola indet. (els Hostalets de Pierola); LT, La Tarumba; TF, Teuleria del Firal (Seu d’Urgell). a Numbers after acronyms refer to stratigraphic levels within a single site, whereas other alphanumeric combinations after ACM sectors (separated by a dash) refer to localities within subsectors. 1 b Hispanopithecus laietanus specimens include histological sections from Andrews & Martin (1991) and Kelley et al. (2001) as reported by Smith et al. (2019). 2 Table 3. Qualitative dental features of the EDJ of the upper molars. P. D. fontani Features A. brevirostris H. crusafonti H. laietanus catalaunicus little to Dentine little little moderately very moderately horns peripheral peripheral peripheral peripheral peripheral shallow and shallow and Mesial deep and deep and shallow and moderately mesially fovea restricted restricted restricted developed located Trigon deeper shallower deeper deeper shallower basin moderately moderately high, Crista high and high, complete low and often high and complete and obliqua centrally and straight disrupted often straight twisted disrupted aligned with markedly aligned with markedly or slightly more markedly more or slightly M1–M2 more lingual more lingual lingual than lingual than more lingual hypocone than than protocone protocone than protocone protocone only in M1 protocone variable variable Buccolingua slight moderate marked (slight to (slight to l waisting moderate) moderate) Lingual slightly moderately well slightly well developed cingulum developed developed developed developed 1 Table 4. Qualitative dental features of the EDJ of the lower molars. Features H. crusafonti H. laietanus “S.” occidentalis Metaconid horn vertical vertical tip centrally tilted Protoconid and entoconid variable but variable but less peripheral horns generally peripheral generally peripheral absent to poorly Buccal cingulid well developed well developed developed Lower molar crown slight marked moderate waisting Tuberculum intermedium sometimes present absent variable (interconulid) M2 metaconulid (twinned absent rarely expressed marked metaconid) Hypoprotocristid and nearly indistinct nearly indistinct marked, although hypometacristid and interrupted and interrupted not merged 1