This is the accepted version of the article: Pina Miquel, Marta; Kikuchi, Yasuhiro; Nakatsukasa, Masato; [et al.]. «New femoral remains of Nacholapithecus kerioi: Implications for intraspecific variation and Miocene hominoid evolution». Journal of human evolution, Vol. 155 (June 2021), art. 102982. DOI 10.1016/j.jhevol.2021.102982 This version is available at https://ddd.uab.cat/record/238974 under the terms of the license 1 New femoral remains of Nacholapithecus kerioi: implications for intraspecific variation 2 and Miocene hominoid evolution 3 4 Abstract 5 The middle Miocene stem kenyapithecine, Nacholapithecus kerioi (16–15 Ma; Nachola, 6 Kenya), is represented by a large number of isolated fossil remains and one of the most 7 complete skeletons in the hominoid fossil record (KNM-BG 35250). Multiple fieldwork 8 seasons performed by Japanese-Kenyan teams during the last part of the 20th century 9 resulted in the discovery of a large sample of Nacholapithecus fossils. Here, we 10 describe new femoral remains of Nacholapithecus. In well-preserved specimens, we 11 evaluate sex differences and within-species variation using both qualitative and 12 quantitative traits. We use these data to determine whether these specimens are 13 morphologically similar to the species holotype KNM-BG 35250 (which shows some 14 plastic deformation), and to compare Nacholapithecus with other Miocene hominoids 15 and extant anthropoids to evaluate the distinctiveness of its femur. The new fossil 16 evidence reaffirms previously reported descriptions of some distal femoral traits, 17 namely the morphology of the patellar groove. However, results also show that relative 18 femoral head size in Nacholapithecus is smaller, relative neck length is longer, and 19 neck-shaft angle is lower than previously reported for KNM-BG 35250. These traits 20 have a strong functional signal related to the hip joint kinematics, suggesting that the 21 morphology of the proximal femur in Nacholapithecus might be functionally related to 22 quadrupedal-like behaviors instead of more derived antipronograde locomotor modes. 23 Results further demonstrate that other African Miocene apes (with the exception of 24 Turkanapithecus kalakolensis) generally fall within the Nacholapithecus range of 25 variation, whose overall femoral shape resembles that of Ekembo spp. and Equatorius 1 26 africanus. Our results accord with the previously inferred locomotor repertoire of 27 Nacholapithecus, indicating a combination of generalized arboreal quadrupedalism 28 combined with other antipronograde behaviors (e.g., vertical climbing). 29 30 Keywords: Miocene hominoids; Femur; Functional morphology; Positional behavior 31 32 1. Introduction 33 Nacholapithecus kerioi is an extinct hominoid (subfamily Kenyapithecinae, tribe 34 Equatorini) known from the middle Miocene of Kenya (Ishida et al., 1999). Fossil 35 remains belonging to this taxon were found within the Aka Aitheputh Formation 36 (Samburu County, Kenya) in Nachola (Fig. 1), that have been dated at 16–15 Ma 37 (Nakatsukasa et al., 1998; Sawada et al., 1998; Ishida et al., 1999; Nakatsukasa and 38 Kunimatsu, 2009). Originally, the material found in this area was attributed to the genus 39 Kenyapithecus, either as Kenyapithecus sp. or Kenyapithecus cf. africanus, by several 40 authors (Ishida et al., 1984; Rose et al., 1996; Nakatsukasa et al., 1998). Later, Ishida 41 and colleagues (1999) erected the new genus and species Nacholapithecus kerioi with 42 sufficient evidence to differentiate this taxon from other fossil hominoids. 43 Initial studies of the postcranial anatomy of Nacholapithecus were based on two 44 dozen isolated fossil remains collected in the 1980s (most described by Rose et al., 45 1996) and a partial skeleton (holotype, KNM-BG 35250), as well as some other 46 specimens recovered in the 1990s (Rose et al., 1996; Nakatsukasa et al., 1998, 2003a, b, 47 2007a, b, 2012; Ishida et al., 2004; Senut et al., 2004; Nakatsukasa and Kunimatsu, 48 2009; Pina et al., 2018; Takano et al., 2018, 2020). The basic body plan of 49 Nacholapithecus (e.g., narrow thorax and long lumbar spine) is similar to that of 50 Ekembo spp., which are considered arboreal quadrupeds engaging in cautious climbing 2 51 and clambering (Ward, 2015). However, Nacholapithecus also shows more derived 52 features, such as longer pedal digits, an anterior projection of the ulnar coronoid 53 process, a more mobile humeroradial joint, and a higher femoral neck-shaft angle 54 (Nakatsukasa et al., 1998, 2004, 2007a; Ishida et al., 2004; Takano et al., 2018, 2020). 55 These derived features provide insights into the positional behavior of Nacholapithecus, 56 which includes some of the earliest evidence of forelimb-dominated behaviors with the 57 enhancement of vertical climbing capabilities. No specific adaptations for below-branch 58 suspension have been identified, although this positional behavior cannot be completely 59 discarded on the basis of the current evidence (Nakatsukasa and Kunimatsu, 2009; 60 Takano et al., 2018, 2020). 61 Fortunately, the Nachola area is very rich in fossil remains, and excavations 62 since 2000 have unearthed a large number of fossils attributed to Nacholapithecus, 63 resulting in an impressive and unusual collection in comparison with those of other 64 Miocene hominoid taxa. The Nacholapithecus fossil collection comprises several fossils 65 representing the same anatomical element, providing an opportunity to assess 66 intraspecific morphological variation, which is rare within the hominoid fossil record. 67 During the past decade, a number of studies have published on these specimens 68 (Kikuchi et al., 2012, 2015, 2016, 2018; Ogihara et al., 2016; Takano et al., 2020). 69 An initial body mass of 20–23 kg was estimated for Nacholapithecus males 70 (Rose et al., 1996; Ishida et al., 2004; Nakatsukasa and Kunimatsu, 2009). Recently, 71 Kikuchi et al. (2018) estimated body mass from femoral head dimensions obtained from 72 12 fragments (see Kikuchi et al., 2018: Table 2). These authors reported the presence of 73 marked sexual dimorphism in Nacholapithecus’ body mass, with males estimated at an 74 average of two times the body mass of females (see also Ishida et al., 1991 for sexual 75 dimorphism estimated from the canines). They assigned six femora of the sample to 3 76 males (larger) and six to females (smaller) based on their body mass results. Kikuchi et 77 al. (2018) briefly described the femoral specimens, but their focus was on investigating 78 sexual dimorphism. In addition, they were highly selective and did not use specimens 79 with damaged femoral heads. Thus, femoral anatomical features and their functional 80 implications have not been thoroughly addressed. 81 This study focuses on reporting and describing femoral remains attributed to 82 Nacholapithecus, as well as investigating the range of intraspecific variation and sex 83 differences in this taxon using a combination of qualitative and quantitative traits. We 84 use these data to determine the extent to which the well-preserved Nacholapithecus 85 femora are similar to those of the holotype (whose distorted nature has been reported 86 elsewhere; e.g., Nakatsukasa et al., 2012). We further compare Nacholapithecus with 87 other African and Eurasian Miocene hominoids and extant anthropoids. Collectively, 88 these comparisons and analyses allow us to review the Nacholapithecus species 89 diagnosis and morphology in detail, and to evaluate the distinctiveness of its femur. A 90 well-defined femoral morphology diagnosis will contribute to a better understanding of 91 the Nacholapithecus positional behavior and its role within the locomotor evolution of 92 the Hominoidea. 93 94 2. Materials and methods 95 2.1. Samples 96 The femoral material from the Nachola fossil sites housed at the National 97 Museums of Kenya (Nairobi) labeled as ‘Nacholapithecus’, ‘Kenyapithecus sp.’, and 98 ‘Hominoidea’ was reviewed to evaluate its taxonomic diagnosis at the species level. A 99 total of 28 femoral remains was available for Nacholapithecus (Table 1). 4 100 Owing to taphonomic damage (described in detail below), morphometric 101 comparisons are limited and only possible for a reduced number of femoral fragments 102 (see Table 1). When possible (exclusively in nondistorted, nondamaged 103 specimens/regions), selected measurements were taken and then used to quantitatively 104 compare Nacholapithecus with other African and Eurasian Miocene hominoids and a 105 wide range of extant anthropoid primates, including platyrrhines, colobines, 106 cercopithecines and hominoids (Fig. 2; Table 2). The Miocene hominoid sample (Fig. 3; 107 taxonomy after Alba, 2012) includes: the afropithecid Morotopithecus bishopi (UMP 108 MORII 94’80; MacLatchy et al., 2000); the proconsulids Proconsul major (combination 109 of NAP IX 46’99, NAP IX B 64, NAP IX 65 P. 67 fragments; Gommery et al., 1998, 110 2002; Senut et al., 2000), Turkanapithecus kalakolensis (KNM-WK 16950I; Leakey et 111 al., 1988), Ekembo nyanzae (KNM-MW 13142A and KNM-RU 5527; Harrison, 1982; 112 Ward et al., 1993), the kenyapithecine Equatorius africanus (BMNH M.16331; BMNH 113 M.16332-3 is used for qualitative comparisons only; Le Gros Clark and Leakey, 1951; 114 McCrossin, 1994), and the dryopithecines Dryopithecus fontani (IPS 41724; Moyà-Solà 115 et al., 2009; Pina et al., 2019) and Hispanopithecus laietanus (IPS 18800.29; Moyà-Solà 116 and Köhler, 1996; Pina et al., 2012). 117 118 2.2. Measurements 119 Linear measurements of the proximal femur were taken to the nearest 0.1 mm 120 using digital calipers (Fig. 2). Superoinferior heights of the femoral head (SIH) and 121 femoral neck (SIN), and anteroposterior depth (APN) of the femoral neck were used to 122 create an index of the relative size of the head: SIH/(√SIN*APN). Relative length of the 123 femoral neck (relative NL) was estimated by dividing neck length by the total 124 mediolateral width of the proximal end of the femur (TotW; relative NL = NL/TotW x 5 125 100). Neck-shaft angle (NSangle) was measured from photographs of femora in anterior 126 view using Fiji version 2.0 (Schindelin et al., 2012). The SIH/(√SIN*APN) and the 127 NSangle have been traditionally associated with range of hip excursion and joint 128 mobility (the larger the relative size of the head and the NSangle, the greater the motion 129 of the joint), whereas relative NL has been related to the actions of the gluteal muscles 130 during locomotion (the longer the length, the greater the length of the moment arm of 131 the muscles and bending forces supported; Fleagle and Meldrum, 1988; Ruff, 1988; 132 Aiello and Dean, 1990; Lovejoy et al., 2002; Harmon, 2007). 133 134 2.3. Analyses 135 Descriptive statistics (sample sizes, means, standard deviations and ranges) were 136 computed for the SIH/(√SIN*APN), relative NL, and NSangle in Nacholapithecus 137 (Table 3; see also Table 1 for Nacholapithecus raw data and Supplementary Online 138 Material [SOM] Table S1). Boxplots were used to visualize the range of within-species 139 variation and to examine whether Nacholapithecus femoral morphology is distinctive 140 from that of other African and European Miocene hominoids. 141 Quantitative statistical analyses were used to evaluate differences in 142 SIH/(√SIN*APN), relative NL, and NSangle between male and female 143 Nacholapithecus. Sex assignation follows the classification provided in Kikuchi et al. 144 (2018; Table 1). In addition, comparisons among Nacholapithecus (sexes pooled) and 145 extant anthropoids were also carried out to check for potential differences/similarities 146 and defining locomotor affinities. The Shapiro-Wilk test was used to check for 147 normality of the data. The null hypothesis of normally distributed data could not be 148 rejected (p > 0.05) for relative NL and NSangle, but the ratio of SIH/(√SIN*APN) was 149 not normally distributed (p < 0.001). Thus, to evaluate sexual dimorphism, mean 6 150 differences between male and female Nacholapithecus were tested using the two-tailed 151 Student’s t-test for NL and NSangle and the nonparametric Mann-Whitney U test for 152 SIH/(√SIN*APN). Likewise, for comparisons of relative NL and NSangle between 153 Nacholapithecus and extant anthropoids, the parametric analysis of variance (ANOVA) 154 was used along with Student’s t-tests for post-hoc comparisons between species, 155 whereas the nonparametric Kruskal-Wallis and Mann-Whitney U tests were used for 156 SIH/(√SIN*APN). The Bonferroni method was used to adjust for all multiple pairwise 157 comparisons for every variable. All analyses were performed using the statistical 158 package R v. 3.6 (R Core Group, 2017). 159 In addition to quantitative analyses, some morphological traits of the 160 Nacholapithecus femora were also qualitatively compared with those of other Miocene 161 hominoids (including African and Eurasian taxa) to better define the distinctiveness of 162 the Nacholapithecus femoral shape. 163 164 3. Results 165 3.1. Morphological descriptions of femoral fragments attributed to Nacholapithecus 166 KNM-BG 17778 This small head fragment was mentioned previously in Ishida et al. 167 (2004) and Nakatsukasa et al. (2012), but has not been formally described (Fig. 4A–B). 168 It is a half-head fragment with a short portion (ca. 4 mm) of the neck. The position and 169 shape of the fovea capitis suggest that it is an anterior hemi-sphere of a left femoral 170 head. The epiphysis is fused. Its small size (SIH = 17.7 mm) indicates it probably 171 belonged to a female. 172 KNM-BG 40844 A femoral head fragment, probably right (Fig. 4C–D). The fragment 173 bears no epiphyseal line and this, together with its small size (ca. 14 mm 7 174 anteroposteriorly and ca. 13 mm proximodistally) suggests it belongs to an adult female. 175 The fovea capitis is marked and placed in the distal half of the head. 176 KNM-BG 40964 A right proximal femur missing the neck, the head and part of the 177 greater trochanter (Fig. 4E–F). The whole fragment is compressed anteroposteriorly. 178 The lesser trochanter is damaged. Since juvenile specimens have been recovered in the 179 same locality, the possibility that this femur belonged to an immature individual cannot 180 be precluded. However, its size is consistent with that of adult female specimens. 181 KNM-BG 42757 A left proximal femoral fragment (Fig. 4G–J). Although it preserves 182 all the gross anatomical structures, both the epiphysis and the shaft are severely crushed 183 anteroposteriorly (see Fig. 4I). It likely belonged to a male specimen due to its large size 184 (SIH = 24.8 mm). Despite its deformation, the femoral head shows a circular shape in 185 anterior view and is positioned slightly below the most proximal peak of the greater 186 trochanter. The notch between the greater trochanter and the head appears to be wide 187 and deep. 188 KNM-BG 44953 This specimen includes a right and a left femur and some left hip 189 bone fragments. The right femur (44953A) was described by Kikuchi et al. (2018) and 190 the left counterpart (44953B) is described here (Fig. 4K–M). It is a left proximal femur 191 fragment which lacks most proximal structures, i.e., the head, part of the neck, and the 192 greater trochanter. This fragment is anteroposteriorly compressed. The lesser trochanter 193 is slightly eroded and faces completely posteriorly. However, this morphology is 194 probably a result of deformation (Fig. 4L). 195 KNM-BG 42779 A shaft fragment (ca. 91 mm) probably belonging to the most distal 196 part of the diaphysis, without the distal epiphysis and any diagnostic trait available (Fig. 197 4N–P). It is crushed anteroposteriorly. Due to its large size, this femoral fragment likely 198 belonged to a male. 8 199 KNM-BG 42738/42756C This proximal femur was recovered from site BG-I west (Fig. 200 5A–C), together with more than 30 other skeletal elements. Although this collection is 201 still being sorted, most of the elements represent a single young adult male. We include 202 this element in our report since it is associated with the KNM-BG 42732 distal femur 203 fragment described below (Fig. 5F–I). A distal femoral shaft of the right counterpart is 204 also described below (KNM-BG 42722; Fig. 5D, E). 205 KNM-BG 42738/42759C is a ca. 51 mm (proximodistally) long proximal 206 portion of a left femur (Fig. 5A–C). Although Kikuchi et al. (2018) described it briefly, 207 we believe this specimen deserves a more detailed description. It comprises two large 208 pieces (the head/neck portion and the shaft/greater trochanter portion) that join 209 perfectly. The tip of the greater trochanter is missing and the lesser trochanter is broken 210 off from the base. The shaft is lost distally from the lesser trochanter base. The 211 epiphyseal line of the head is completely fused and not visible. The head is almost 212 intact, although the cortex is partially worn out on the anterior and posteroinferior 213 aspects. The articular surface is wide anteriorly and posteriorly. In proximal view, the 214 head is offset slightly anteriorly and weakly rotated posteriorly (Fig. 5A). The head and 215 neck surfaces are more confluent in posterior view. However, the center of the head is 216 almost on the central axis of the neck in proximal view. The fovea is lightly weathered 217 and is located in the posteroinferior quarter of the articular surface. The cross-section of 218 the neck is kidney-shaped. It is weakly concave anteriorly, convex posteriorly, and the 219 inferior part is thicker than the superior part. The SIN = 16.5 mm and APN = 12.9 mm. 220 The neck-shaft angle is 121º. The crista trochanterica is present on the neck (Fig. 5C). 221 The posterior bar of the trochanteric fossa is well developed and runs to the base of the 222 lesser trochanter. The posterior cortex of the shaft is badly damaged and displaced to the 223 medullary cavity. The anterior cortex is also fractured, though to a lesser degree, and 9 224 displaced posteriorly, leaving a wide but shallow depression. The distal part of the great 225 trochanter is protuberant laterally and slightly anteriorly (Fig. 5B, C). Distal and 226 posterior to this protuberance is a small swelling along the distal break. This might be 227 the most proximal part of the lower eminence of the gluteal tuberosity. 228 The shaft is ca. 26.5 mm wide (mediolaterally) and is markedly compressed 229 anteroposteriorly at the distal break point. Despite this compression, the anterior cortex 230 thickness (3.1 mm) is well preserved and not distorted (the posterior thickness is 231 difficult to determine, since the cortex looks affected by the compression). 232 KNM-BG 42722 This specimen is a ca. 61 mm long distal shaft piece of a right femur 233 belonging to the same individual as KNM-BG 42732 and KNM-BG 42738/42756C 234 (Fig. 5). It measures >15 mm mediolaterally, 12.8 mm anteroposteriorly at the proximal 235 break point, and 18.0 mm mediolaterally and 14.1 mm anteroposteriorly at the distal 236 break point. Since the breaks are covered by matrix, cortex is not clearly visible. 237 Surface features are not well developed (Fig. 5D–E). However, two blunt ridges are 238 discernible, which help to identify the anatomical position of this fragment when it is 239 compared with the distal femur of Eq. africanus (BNMH M 16332-3; SOM Fig. S1). 240 One of the two ridges, which is sharper at the distal break point, is a continuation from 241 the lateral supra-epicondylar line. The other more rounded ridge divides the shaft 242 surface into the posterior and medial surfaces. The posterior surface is weakly convex 243 mediolaterally. The medial, anterior and lateral surfaces are not clearly differentiated. 244 The shaft cross-section is not symmetrical mediolaterally. Regarding the anteroposterior 245 axis, the medial half is wider than the lateral one. 246 10 247 KNM-BG 42732 This distal femur is associated with the KNM-BG 42738/42756C 248 proximal femur described by Kikuchi et al. (2018; see above). A distal femoral shaft of 249 the right counterpart is also associated (KNM-BG 42722; see above). 250 This specimen is a ca. 68 mm long distal portion of a left femur (Fig. 5F–I). The 251 epiphyseal line is visible all around where the epiphysis is preserved. It lacks the lateral 252 condyle (and epicondyle). The medial condyle is intact. The width of the medial 253 condyle is 15.6 mm and is visually comparable with that of the left femur (KNM-BG 254 35250J) of the holotype specimen (Ishida et al., 2004). Plastic deformation is minor, as 255 is the distortion (slightly stronger on the lateral side of the shaft). The diaphyseal part 256 suffered notably from erosion. At the level of the proximal break point, which is 45 mm 257 apart from the epiphyseal line on the medial side, the original cortex remains only as a 258 very small (ca. 4 mm wide) portion on the anteromedial surface (Fig. 5I). The break is 259 13.4 mm wide and 12.4 mm thick when the missing outer cortex is not taken into 260 account. Assuming a thickness of the eroded outer layer of cortex of 1 mm (an estimate 261 based on photogrammetry), the original dimension of the break would not exceed 16 262 mm mediolaterally and 15 mm anteroposteriorly. This shaft is rather thin, especially 263 anteroposteriorly when compared with the distal femoral end, probably because it likely 264 belonged to a young adult (note the diaphyseal line in Figure 5F, G). In the femur of Eq. 265 africanus from Maboko (BNMH M 16332-3: Le Gros Clark and Leakey, 1951), the 266 anteroposterior thickness at an equivalent point is ca. 18 mm (measurement taken from 267 a museum-produced cast). The medial supra-epicondylar ridge is discernible although it 268 is damaged by erosion. The lateral supra-epicondylar ridge (line) is less clear. 269 The distal end is approximately 32 mm anteroposteriorly (in anatomical position 270 and measured at the midpoint of the proximodistal height of the medial condyle). The 271 medial condyle is mediolaterally wide (ca. 15.6 mm) relative to the epiphysis (Fig. 5H). 11 272 The pit-like insertion for the collateral ligament is well marked on the medial 273 epicondyle. The rim of the medial condyle is posteriorly and distally intact. However, 274 both the anterior part of this rim and the continuing medial rim of the patellar surface is 275 eroded. This damage probably reduced the articular surface width by ~1-2 mm. The 276 lateral patellar surface rim is more eroded than the medial rim. However, the proximal 277 border of the patellar surface is intact and the proximodistal height of both the medial 278 and lateral surface rims are similar, resulting in a quadrilateral shape for the patellar 279 surface. The medial two-thirds of the intercondylar notch is preserved. A round 280 depression is observed on the lateral side of the medial condyle proximally, which is 281 likely the attachment area of the posterior cruciate ligament. Otherwise, there are no 282 remarkable features (e.g., buttresses: MacLatchy et al., 2000) observed. 283 The two specimens described below belong to previously described femoral 284 remains from Nachola, whose former attributions (either anatomical or taxonomic) are 285 revised here: 286 KNM-BG 15533 This fragment was formerly described as a partial femoral head by 287 Rose et al. (1996). However, the articular surface shows an anteroposterior compression 288 that is typical of the humeral head of this taxon (SOM Fig. S2A). 289 KNM-BG 15536 This is a femoral head fragment described by Rose et al. (1996) as 290 belonging to Nacholapithecus (then Kenyapithecus sp.). However, the morphology 291 around the fovea capitis (shallow and with an irregular articular surface depression 292 adjacent to the fovea, with distinct bone absorptive pits) is not common among the 293 femoral heads of this taxon. This femur fragment may, therefore, belong to another 294 nonprimate mammal (SOM Fig. S2B). 295 296 3.2. Within-species variation in Nacholapithecus and sex differences 12 297 A summary of the main femoral traits described in this study and discussed in 298 the literature for Nacholapithecus can be found in SOM Table S2. 299 In quantitative terms, the range of variation is moderately narrow for 300 SIH/(√SIN*APN) and relative NL in Nacholapithecus (Fig. 6A, B), whereas it is greater 301 for the NSangle (Fig. 6C). Previous authors have noted that the KNM-BG 35250A 302 holotype femur could show some plastic deformation (Fig. 7H). In our results, KNM- 303 BG 35250A falls in an intermediate position within the Nacholapithecus range of 304 variation for the relative femoral head size (Fig. 6A). On the other hand, excepting 305 KNM-BG 40826 (whose NSangle value should be considered with caution given the 306 fragmentary nature of this specimen), all the remaining Nacholapithecus specimens 307 show NSangle values below KNM-BG 35250A (Fig. 6C). Our quantitative results do 308 not show a clear trend for KNM-BG 35250A within the whole Nacholapithecus femora 309 sample that allows us to clearly associate it to plastic deformation issues. 310 Qualitatively, the femoral remains of Nacholapithecus display a hemispherical 311 head in all cases and the articular surface is well differentiated from the neck. When 312 preserved, the fovea capitis is generally shallow and is placed on the distal half of the 313 articular surface. Previous authors highlighted the posterior location of the fovea capitis 314 (Nakatsukasa et al., 2012; Kikuchi et al., 2018), but we were unable to confirm the 315 location in these newly described specimens due to their fragmentary nature and/or the 316 poor preservation. 317 Nakatsukasa and colleagues (Nakatsukasa et al., 2012) suggested that the 318 anteversion of the femoral head in KNM-BG 35250A was due to deformation. The 319 newly described specimens shows slight anteversion; the displacement of the femoral 320 head compared to the neck is not marked (Fig. 5A). The configuration observed in the 321 larger femoral sample of Nacholapithecus might confirm the view of Nakatsukasa et al. 13 322 (2012). When the head and the greater trochanter are preserved, the head proximally 323 projects slightly above the greater trochanter (Fig. 7A–C). 324 Due probably to distortion, the neck length could not be measured in KNM-BG 325 35250A; thus, quantitative comparisons are not possible with the holotype. However, 326 qualitative comparisons suggest that the femoral neck of the Nacholapithecus holotype 327 is shorter than that of other Nacholapithecus femoral remains (e.g., KNM-BG 38391A; 328 Fig. 7A, H). 329 The presence of a lateral flare of the greater trochanter cannot be conclusively 330 added to the morphological suite of features that characterize the Nacholapithecus 331 femur. Among the sample, there are specimens with well-defined flaring (e.g., KNM- 332 BG 44954A and KNM-BG 44953A; Fig. 7B), while others are characterized by 333 minimal lateral expansion of this region (e.g., KNM-BG 38391A; Fig. 7A, see also 7C). 334 As previously noted, it seems clear that when the lateral projection is evident, it occurs 335 mainly at the distal part of the greater trochanter (Fig. 7B, F). 336 Previous authors (Ishida et al., 2004; Nakatsukasa et al., 2012; Kikuchi et al., 337 2018) have noted that the lesser trochanter is placed close to the femoral neck and that it 338 faces posteromedially in Nacholapithecus. The best specimen in which to observe this 339 trait is KNM-BG 17816, which perfectly preserves its original form (Fig. 7D, E; Rose et 340 al., 1996). Although not complete, KNM-BG 40800F and KNM-BG 38391A also 341 support this interpretation (Kikuchi et al., 2018; Fig. 7A). The close location of the 342 lesser trochanter to the femoral neck and its posteromedial orientation can be considered 343 diagnostic for Nacholapithecus (Fig. 7D, E). 344 Despite variation in size, there were no significant differences between males 345 and females in any of the analyses performed: SIH/(√SIN*APN) (U = 9, p = 0.90); 346 relative NL (t = 2.26, df = 3, p = 0.12, ), or NSangle (t = 0.62, df = 3, p = 0.58; see also 14 347 Fig. 6). When quantitative comparisons were not possible, the strong sexual 348 dimorphism in body size of Nacholapithecus permitted us to tentatively differentiate 349 between male (large) and female (small) specimens on the basis of size (see 350 male/female size differences in Fig. 7A, C; Kikuchi et al., 2018). 351 Due to the distorted nature of the distal femoral fragments attributed to 352 Nacholapithecus (KNM-BG 35250B, KNM-BG 35250J, and KNM-BG 42779; KNM- 353 BG 42732 is a young adult individual), only qualitative comparisons could be made for 354 this region (Figs. 4, 5; SOM Table S2). As previously described for the 355 Nacholapithecus holotype (KNM-BG 35250B and J; Fig. 7I), the patellar groove is 356 square-shaped, wide, and shallow in the relatively well-preserved juvenile specimen 357 (KNM-BG 42732; Fig. 5F). The intercondylar fossa seems wide in KNM-BG 42732, 358 although it is not possible to verify this trait in the holotype specimen since the remains 359 are highly compressed mediolaterally (KNM-BG 35250B) and anteroposteriorly 360 (KNM-BG 35250J). 361 362 3.3. Comparisons with other Miocene hominoids 363 Previous studies have identified a series of femoral traits that distinguish 364 Nacholapithecus from other Miocene hominoids (especially African taxa), mainly 365 focusing on differences with Ekembo spp. (Nakatsukasa et al., 1998, 2012; Ishida et al., 366 2004). Our analysis of the new femora raises questions about the distinctiveness of 367 Nacholapithecus, given the overall morphological similarities of the proximal end of the 368 femur to that of Ek. nyanzae (e.g., KNM-MW 13142A and KNM-RU 5527; see below). 369 The femoral morphology also shows some resemblance to that of Eq. africanus (BNMH 370 M.16331) and T. kalakolensis (KNM-WK 16950I), and more clearly differs from M. 15 371 bishopi (UMP MORII 94’80), P. major (NAP IX 46’99) and the European 372 dryopithecines in general terms (Fig. 3). 373 The femoral head relative size (SIH/(√SIN*APN)) range of Nacholapithecus 374 only overlaps with Ekembo and Equatorius (Fig. 6A), with the index value of the 375 holotype close to that of Ekembo (KNM-MW 13142A; Fig. 3D). Morotopithecus (UMP 376 MORII 94’80; Fig. 3A) shows the lowest value (smallest relative femoral head) among 377 fossils; whereas Hispanopithecus (IPS 18800.29; Fig. 3H) shows one of the highest 378 values (largest relative femora head) for this index among extinct taxa, slightly below 379 the KNM-BG 38391A Nacholapithecus specimen. 380 On the basis of the holotype (KNM-BG 35250A), which is probably plastically 381 deformed (Fig. 7H), it has been suggested that femoral neck length of Nacholapithecus 382 is relatively short when compared with Ekembo (Fig. 3D–E; Ishida et al., 2004; 383 Nakatsukasa et al., 2012). Kikuchi et al. (2018) subsequently reinforced this suggestion 384 with a larger sample of femoral remains (most of them also included in this study). 385 When quantified, however, our results do not support the characterization of the 386 Nacholapithecus femoral neck as relatively short compared with Ekembo. The range of 387 variation in neck length is small in Nacholapithecus (Fig. 6B) and the Ekembo specimen 388 KNM-MW 13142A falls within the range of Nacholapithecus; while KNM-RU 5527 389 falls below its lower limit but close to its range of variation. It can be noted that 390 Turkanapithecus (KNM-WK 16950I; Fig. 3C) displays a relatively very long femoral 391 neck compared with the Miocene fossil hominoids (as well as the extant anthropoids in 392 our sample; Fig. 6B). The European dryopithecines display relatively shorter necks than 393 Nacholapithecus, being Hispanopithecus the extinct taxon with the shortest relative 394 neck length (Fig. 6B). 16 395 When all the available Nacholapithecus femora are considered, results show that 396 the NSangle of this taxon is not especially high among the Miocene hominoid taxa 397 (contra Nakatsukasa et al., 2012; Figs. 3 and 6C). All the specimens included in the 398 latter group fall within the range of Nacholapithecus except Turkanapithecus, whose 399 NSangle value is slightly higher than the uppermost limit of the Nacholapithecus range 400 (Fig. 6C). 401 Most of the Nacholapithecus specimens display a lateral projection of the 402 greater trochanter but it is not present in all the available femora (e.g., KNM-BG 403 44954A vs. KNM-BG 38291A; see above; Fig. 7A–C; Kikuchi et al., 2018). When it is 404 present, the flare is evident at the base of the greater trochanter (Nakatsukasa et al., 405 2012; Kikuchi et al., 2018; Fig. 7B). This feature is also present in other Miocene 406 hominoids (Fig. 3; Senut et al., 2000; Bacon, 2001) such as Ekembo (KNM-MW 407 13142A), Turkanapithecus (KNM-WK 16950I), Proconsul (reconstructed specimen 408 from NAP IX), Equatorius (BNMH M.16331), and Hispanopithecus (IPS 18800.29). 409 Only Morotopithecus (UMP MORII 94’80) and Dryopithecus (IPS 41724) show a very 410 light lateral protrusion of the greater trochanter (Fig. 3A, G; MacLatchy et al., 2000; 411 Nakatsukasa et al., 2012; Almécija et al., 2013; Pina et al., 2019). Due to the observed 412 variation in the Nacholapithecus femora, neither the presence nor the absence of a 413 greater trochanter lateral flare can be considered diagnostic for Nacholapithecus. 414 The posteromedial facing and relatively proximal position of the lesser 415 trochanter in Nacholapithecus (Fig. 7D, E; e.g., Ishida et al., 2004) not only resembles 416 that observed in Ekembo (KNM-MW 13142A; Ward et al., 1993, but see Nakatsukasa 417 et al., 2012) and Equatorius (BMNH M16331), but also that described for 418 Morotopithecus (UMP MORII 94’80; MacLatchy et al., 2000; Fig. 3A). However, the 419 configuration of the lesser trochanter in Turkanapithecus (KNM-WK 16950I; Leakey et 17 420 al., 1988; Fig. 3C) clearly differs from that defined for Nacholapithecus by facing more 421 posteriorly and by being positioned more distally M. Pina, pers. obs.). This is also the 422 case for Dryopithecus (IPS 41724) and Hispanopithecus (IPS 18800.29), whose lesser 423 trochanter is placed more distally and facing posteriorly or more medially, respectively 424 (Pina 2016; Pina et al., 2019). In the case of Proconsul (reconstructed specimen from 425 NAP IX; Fig. 3B), interpretations of the direction of the lesser trochanter conflict in the 426 literature; Senut et al. (2000) suggested that the lesser trochanter faces posteriorly, 427 whereas Gommery et al. (2002) advocated for a medial direction. 428 The newly available Nacholapithecus femora also confirm other traits formerly 429 highlighted in the literature, e.g., the close proximity of the gluteal tuberosity to the 430 greater trochanter (Fig. 7F, G). This trait differentiates Nacholapithecus from Ekembo 431 (KNM-MW 13142A) since in Ekembo, the gluteal tuberosity is positioned more distally 432 relative to the greater trochanter (Fig. 3D; Nakatsukasa et al., 2012). Although not well 433 preserved, the position of the gluteal tuberosity in Turkanapithecus (KNM-WK 16950I) 434 resembles that of Ekembo more than that of Nacholapithecus (Fig. 3C; Leakey et al., 435 1988). Only Dryopithecus (IPS 41724) displays a marked gluteal tuberosity among the 436 dryopithecines of the sample (Pina et al., 2019). In this case, the gluteal tuberosity is 437 positioned closer to the greater trochanter than in Ekembo and resembles the condition 438 of Nacholapithecus (Fig. 3G). 439 The patellar groove shape of Nacholapithecus is quadrangular and shallow (Fig. 440 5; see also Fig. 7I; SOM Figure S3; SOM 3D Model S1), as in Turkanapithecus (KNM- 441 WK 16950I; Fig. 3C; Leakey et al., 1988), Morotopithecus (UMP MORII 94’80; Fig. 442 3A; MacLatchy et al., 2000) and probably Equatorius (KNM-MB 24727; see 443 McCrossin, 1994:fig. 38 and p. 162). Rose (1983) also noted that the patellar groove of 444 Ekembo (e.g., KNM-RU 5527; Fig. 3E) is square-shaped and shallow. Nakatsukasa et 18 445 al. (2012:238, footnote Fig. 3) highlighted that “the patellar surface is trapezoidal with a 446 more raised lateral rim” in KNM-RU 5527. This trait, together with the asymmetric 447 width of the condyles in KNM-RU 5527, differentiates the Ekembo distal femur from 448 that of Nacholapithecus. The condyles of KNM-RU 5527 (Nakatsukasa et al., 2012; 449 Fig. 3E), KNM-WK 16950I (Leakey et al., 1988; Fig. 3C) and UMP MORII 94’80 450 (MacLatchy et al., 2000; Fig. 3A) display more asymmetric epicondyles than 451 Nacholapithecus. 452 Finally, the asymmetrical cross-section of the distal shaft displayed by 453 Nacholapithecus (medial half wider than the lateral one) resembles that of Equatorius 454 (BNMH M 16332-3; McCrossin, 1994), although it is much thinner anteroposteriorly in 455 the only available Nacholapithecus specimen (KNM-BG 42722), likely reflecting that it 456 is an immature specimen. 457 458 3.4. Comparisons of Nacholapithecus femora with those of extant anthropoids 459 Nacholapithecus, together with African apes and platyrrhines, shows an 460 intermediate position between cercopithecoids (smaller relative femoral heads) and 461 Asian apes (larger femoral heads) for SIH/(√SIN*APN) (Fig. 6A), with some 462 statistically significant differences (Table 4). 463 Nachalopithecus displays a long neck (NL) relative to the total mediolateral 464 width of the proximal femur, not significantly different from that of platyrrhines (i.e., 465 Cebus apella and atelids), orangutans, chimpanzees, and gorillas, but significantly 466 different from colobines, cercopithecines, Hylobates lar and Pan paniscus (Table 4; 467 Fig. 6B; SOM Table S3). We note that variation in relative NL is high in many of the 468 extant anthropoids, particularly in cercopithecines and colobines (whose ranges overlap 469 with those of the other taxa). 19 470 The neck-shaft angle (NSangle) of Nacholapithecus has been described as high, 471 like that of gibbons or Ateles (Natasukasa et al., 2012). However, statistically, the 472 Nacholapithecus NSangle differs significantly from both H. lar and the atelids, as well 473 as Pongo pygmaeus, cercopithecines, colobines, H. lar, and P. t. schweinfurthii (Table 474 4; Fig. 6C; SOM Table S3). 475 Qualitatively, the gluteal tuberosity in the Nacholapithecus sample is well 476 marked and situated close to the greater trochanter (Fig. 7F, G), as originally described 477 for the holotype specimen (Nakatsukasa et al., 2012). Among extant hominoids, only 478 gibbons are reported to show a marked gluteal tuberosity (Stern, 1972; Almécija et al., 479 2013). As in the case of Nacholapithecus, the gluteal tuberosity of gibbons is positioned 480 close to the greater trochanter (Fig. 7K). 481 The distal end of the femur, KNM-BG 42732, which does not suffer from severe 482 distortion, displays an ape-like distal epiphysis, relatively wider medio-laterally than 483 thick antero-posteriorly (Fig. 5H). The patellar groove of Nacholapithecus is shallow 484 and more similar in shape to lesser apes (Hylobates) than platyrrhyines (Cebus) (i.e., 485 with an approximated quadrangular shape); the intercondylar fossa appears wider than 486 in Cebus (SOM Figure S3; SOM 3D Models S1–S3). 487 488 4. Discussion 489 4.1 Within species variation in Nacholapithecus 490 Kikuchi and colleagues (2018; see also Ishida et al., 1991) reported a strong 491 sexual dimorphism for Nacholapithecus. Nonetheless, differences in size among the 492 Nacholapithecus remains might additionally suggest either 1) the presence of several 493 anthropoid taxa or 2) the presence of two different species of Nacholapithecus in the 494 Nachola area. 20 495 Apart from Nacholapithecus, other anthropoid taxa have been recovered in this 496 region, namely Nyanzapithecus (Kunimatsu, 1992, 1997) and Victoriapithecus 497 (Pickford et al., 1987). It is possible that the smaller femora ascribed to 498 Nacholapithecus belong to Nyanzapithecus, which might be smaller in overall body size 499 (Kunimatsu, 1992, 1997). Female Nacholapithecus and male Nyanzapithecus reportedly 500 overlap in size based on dental dimensions (Kunimatsu, 1997). Thus, the smaller femora 501 in our sample could represent male Nyanzapithecus remains (see also Kikuchi et al, 502 2018). Most of the femora included in the current study were collected in the BG-K 503 locality, from which no Nyanzapithecus specimens have been formally identified thus 504 far. From the extensive collection of primate fossils (~240 dental specimens) recovered 505 from this locality, only a single fragment of maxilla has been preliminarily catalogued 506 as a non-cercopithecoid small catarrhine (Y. Kunimatsu, unpublished data). This 507 fragment could be potentially accommodated as a nyanzapithecine, but due to its poor 508 preservation, its attribution remains provisional (a formal description has not been 509 published yet). All of the other dental material of Nyanzapithecus, as well as material 510 tentatively attributed to this genus as cf. Nyanzapithecus (29 specimens including 511 published and unpublished ones), have been collected from another locality (BG-X) 512 together with ca. 190 Nacholapithecus dental specimens (Y. Kunimatsu, unpublished 513 data). Only two small femora (KNM-BG 17775 and KNM-BG 17778) were recovered 514 from BG-X. Rose et al. (1996) described nine hominoid postcranial specimens collected 515 from this locality. Other than these two femoral specimens, only one proximal phalanx 516 (KNM-BG 15531: of unknown ray) is considerably smaller than pedal proximal 517 phalanges (median rays) of male Nacholapithecus (Nakatsukasa et al., 2003b; see also 518 Nakatsukasa et al., 2012). This proximal phalanx thus could belong to either a female 519 Nacholapithecus or a male Nyanzapithecus, but it could also be a male Nacholapithecus 21 520 phalanx from a paramedian ray. Although further analysis would help to corroborate 521 this hypothesis, given that Nyanzapithecus remains are rare in the area of Nachola, the 522 possibility of currently assigning postcranial remains to non-Nacholapithecus taxa is 523 very low (Ishida et al., 1984; Kunimatsu, 1997). 524 The species Nacholapithecus kerioi was erected on the basis of the KNM-BG 525 35250 skeleton by Ishida et al. (1999). In their article, these authors included other 526 hominoid specimens discovered from Nachola to the species, although they did not 527 specify catalogued numbers. Subsequent works have underpinned the homogeneity of 528 the dental morphology (regardless the differences in size; e.g. Kunimatsu et al., 2004). 529 The evident differences in size among our femoral sample between the smallest and the 530 largest fragments are not reflected in the femoral morphology. Nacholapithecus body 531 mass estimated by Kikuchi et al. (2018), based on the femoral diameter, ranges 8.7–10.8 532 kg for females and 17.3–25.8 kg for males (see their Table 2 BM1 estimates). The 533 largest femur (KNM-BG 40800F) is associated with unpublished dental rows that do 534 not show any diagnostic features to distinguish them from other specimens assigned to 535 Nacholapithecus smaller specimens. The same applies for the second largest femur in 536 the Kikuchi et al., (2018) dataset, the subadult KNM-BG 42738/42756C (Y. Kunimatsu, 537 personal observation). These facts lead us to conclude that the femoral sample analyzed 538 here are all attributable to a single species of Nacholapithecus. 539 In addition, the sample of femoral fragments attributed to Nacholapithecus 540 allows us to shed light on the Nacholapithecus holotype and broaden the range of 541 variation for comparison with other species. With the additional femoral remains, we 542 have found that the holotype femora display some quantitative and qualitative 543 differences from other Nacholapithecus specimens that might support the idea that the 544 KNM-BG 35250 femora display rather severe deformation. KNM-BG 35250A has been 22 545 traditionally used to define the femoral diagnostic traits of Nacholapithecus, but our 546 results suggest that drawing morphofunctional inferences from the holotype of this 547 taxon would be ill advised (see below). 548 549 4.2. Functional interpretations and positional behavior in Nacholapithecus 550 The proximal end of the femur has been the focus of extensive study due to its 551 relation with the functionality of the hip and its potential association with different 552 locomotor modes in primates. Some of the most characteristic traits are found at the 553 head-neck complex. Ruff (1988) suggested that the increase in the femoral head 554 decoupled from that of the neck will result in the enhancement of the hip joint excursion 555 (see also Ward et al., 1993; Harmon, 2007; Hammond, 2014, among others). Likewise, 556 the higher the NSangle, the greater the mobility at the hip (Rose, 1983; Ward et al., 557 1993; Lovejoy et al., 2002; Hammond, 2014). Thus, a large relative head and a high 558 NSangle enhance the capacity for hip abduction and external rotation of the leg. This 559 configuration also increases the angle of the hind limb related to the midline of the 560 body, which facilitates antipronograde behaviors such as vertical climbing (Stern and 561 Susman, 1981; Rose, 1983; Ward et al., 1993; Harrison, 1986; MacLatchy, 1996; 562 Hammond, 2014). By contrast, quadrupedal monkeys display relatively small heads, 563 short necks, and lower NSangles, which help to resist bending forces and movements of 564 the hind limb preferentially in the parasagittal plane (Fleagle, 1977; Rose, 1983; Fleagle 565 and Meldrum, 1988; Cooke and Tallman, 2012). 566 Overall, the enlarged sample of Nacholapithecus femora suggests that the femur 567 of this taxon is characterized by a relatively small head (SIH/[√SIN*APN]), a 568 moderately long neck relative to the total mediolateral width of the proximal end of the 569 femur, and an intermediate (or moderate) NSangle, compared with extant anthropoids 23 570 (Fig. 6). These analyses suggest that previous functional interpretations based on the 571 femoral morphology of KNM-BG 35250, the Nacholapithecus holotype (Fig. 7H, I), 572 should be revised. Although KNM-BG 35250A does not always represent an extreme 573 condition in the Nacholapithecus sample for the variables analyzed here (Fig. 6), 574 functional inferences based exclusively on this specimen must be considered cautiously 575 due to its plastic deformation. Therefore, as the relative femoral head, neck length and 576 neck-shaft angle have been related to hip mobility and the capability of abduction of the 577 hind limb (Grand, 1968; Fleagle and Meldrum, 1988; Ruff, 1988; Aiello and Dean, 578 1990; Hammond; 2014), our results have important implications regarding locomotor 579 inferences in Nacholapithecus, suggesting it might have a less mobile hip joint than 580 previously proposed (e.g., Nakatsukasa et al., 2012). 581 Although quantitative analyses are not possible for the distal femur, the new 582 evidence supports previous descriptions of Nacholapithecus as having an apparently 583 shallow and square-shaped patellar groove (Fig. 5F and Fig. 6I; SOM 3D Model S1). 584 The shape of the patellar groove is controversial since some authors have suggested a 585 high degree of intraspecific variation in this feature in Nacholapithecus, with some 586 individuals displaying a square shape and others a more trapezoidal outline of the 587 groove (Rose et al., 1996). The shallow patellar groove of the Nacholapithecus distal 588 femur accords with its previously described morphology of the patella, which exhibits 589 some living ape-like affinities that probably foretell the specialized patellae of living 590 great apes (Fig. 6L; Ward et al., 1995; Rose et al., 1996; Pina et al., 2014, 2020). 591 Nacholapithecus also displays condyles that are subequal in size (Nakatsukasa et al., 592 2012), a trait typical of quadrupedal anthropoid monkeys in which loadings are equally 593 distributed through the distal end of the femur (Rose, 1983; Georgiou et al., 2018; 594 Sukhdeo et al., 2018). 24 595 The general evidence found for the Nacholapithecus femur suggests that only 596 the patellofemoral articulation might show enhanced range of motion, since this taxon 597 does not display the stabilization traits of this joint characteristic of quadrupedal 598 monkeys (e.g., deep patellar groove at the femur and compartmentalized articular 599 surface of the patella; Harrison, 1986; Ward et al., 1995; DeSilva et al. 2013; Pina et al., 600 2014, 2020). On the other hand, the hip joint might maintain more restricted movements 601 in the parasagittal plane (also supported by distal femoral condyles and epicondyles 602 similar in size). Although Nacholapithecus does not show the whole set of 603 characteristics of the proximal femoral end traditionally related to abduction movements 604 and external rotation of the hip (long neck, low greater trochanter related to the head, 605 relatively large femoral head, among others; Lovejoy et al., 2002; Richmond and 606 Jungers, 2008; Almécija et al., 2013), these movements cannot be completely ruled out 607 from its positional repertoire. This morpho-evolutionary gradation at the femur is also 608 found in Morotopithecus (limited hip abduction and less-restricted movements at the 609 knee; MacLatchy et al., 2000) and departs from the femoral evidence observed in 610 Ekembo, Turkanapithecus and Equatorius. The reviewed morphology of the 611 Nacholapithecus femur presented in this work is completely compatible with the 612 positional repertoire formerly proposed for this Miocene taxon, which probably 613 combined generalized above-branch quadrupedalism with other antipronograde 614 behaviors, such as vertical climbing (no clear evidence for suspension is found in its 615 femur or elsewhere; Nakatsukasa et al., 2003b, 2012; Ishida et al., 2004; Nakatsukasa 616 and Kunimatsu, 2009; Ogihara et al., 2016; Takano et al., 2018, 2020). 617 Results presented in this work suggest that the Nacholapithecus femur resembles 618 those of Ekembo, Turkanapithecus, and Equatorius, showing a general primitive (stem 619 hominoid-like) appearance. In contrast, Nacholapithecus differs from those femora 25 620 showing more derived (living hominoid-like) traits, such as Morotopithecus and 621 Proconsul in Africa (MacLatchy et al., 2000; Senut et al., 2000; Gommery et al., 2002) 622 and Sivapithecus in the late Miocene of Asia (Kelley, 2005; Madar et al., 2002). 623 Nacholapithecus is also unlike the younger European hominids. Although also 624 displaying quadrupedal affinities (Moyà-Solà et al., 2009; Pina et al., 2019), our results 625 show that Nacholapithecus differs from Dryopithecus at the proximal end of the femur 626 and clearly departs from those taxa with well-defined affinities for forelimb-dominated 627 behaviors (i.e., Hispanopithecus, Rudapithecus, Danuvius, and Oreopithecus; Straus, 628 1963; Hürzeler, 1968; Harrison, 1986; Jungers, 1987; Begun 1992, 2013; Rose, 1993; 629 Moyà-Solà and Köhler, 1996; Begun and Kordos, 2011; Begun et al., 2012; Pina et al., 630 2012; Böhme et al., 2019; Ward et al., 2019). 631 When the whole anatomy is taken into account, similarities between 632 Nacholapithecus and Ekembo are less clear. Although the general body plan in these 633 two taxa is similar (narrow and deep trunk; Ward et al., 1993; Nakatsukasa et al., 634 2007a), Nacholapithecus clearly departs from Ekembo regarding forelimb shape. The 635 former displays a series of characteristics more related to the stabilization of the 636 humeroantebrachial complex (e.g., anterior projection of the coronoid process of the 637 ulna and globular humeral capitulum) and enhancement of the pronation-supination 638 movements, as shown in the elbow of living apes (Nakatsukasa and Kunimatsu, 2009; 639 Takano et al., 2018, 2020). As occurs in other Miocene taxa (not only in Africa, but also 640 in Eurasia; see e.g., Pilbeam et al., 1980; Begun, 1992, 2015; Moyà-Solà and Köhler, 641 1996; Madar et al., 2002; Almécija et al., 2013; Ward, 2015; Böhme et al., 2019), the 642 postcranial morphology of Nacholapithecus shows a unique combination of primitive 643 and derived features; in general, a more derived forelimb, foot, and lumbar region, and a 644 primitive hind limb compared with Ekembo (Ishida et al., 2004; Nakatsukasa and 26 645 Kunimatsu, 2009). These results are compatible with general inferences made for this 646 taxon, but also for the rest of middle Miocene hominoids included in this work. Overall, 647 these African primates potentially combined general arboreal quadrupedalism with 648 other antipronograde behaviors (Rose, 1983; Nakatsukasa and Kunimatsu, 2009; Alba, 649 2012; Begun, 2012; Ward, 2015). In the case of Nacholapithecus, its forelimbs and feet 650 were apparently more derived than its hind limbs and trunk. This fact could suggest the 651 presence of selective pressures on the upper half of the body in Miocene African taxa, 652 underpinning the development of forelimb-dominated behaviors such as vertical 653 climbing and, more recently, below-branch suspension (Nakatsukasa and Kunimatsu, 654 2009). Although no specific traits for the latter locomotor mode or slow-cautious 655 quadrupedalism have been identified in Nacholapithecus (Nakatsukasa and Kunimatsu, 656 2009; Takano et al., 2018, 2020), they cannot be entirely ruled out for the positional 657 behavior repertoire of this taxon. Nonetheless, like the other middle Miocene hominoid 658 taxa in this work, Nacholapithecus was likely adapted for an arboreal life, and some 659 antipronograde behavior (e.g., vertical climbing, clambering, and/or cautious and 660 eclectic climbing) could have been a component of its locomotor repertoire, though less 661 than in middle and late Miocene Eurasian taxa (Nakatsukasa et al., 2003b; 2007a; 2012; 662 Senut et al., 2000; Takano et al., 2018, 2020). 663 664 5. Conclusions 665 The study of a larger sample of femora assigned to Nacholapithecus allowed us 666 to review the original description and species diagnostic femoral traits reported for 667 KNM-BG 35250 (holotype). Our results show that the morphology of the better- 668 preserved femora differs in some respects from that of the original description derived 669 from the holotype (mainly at the femoral head and neck). These findings suggest that 27 670 previous interpretation of the femoral morphology of Nacholapithecus may have been 671 influenced by the presence of some distortion and/or deformation in the holotype 672 femora (see also Nakatsukasa et al., 2012). Consequently, in contrast with previous 673 work, we found that the overall femoral morphology in Nacholapithecus is more similar 674 to that of early and middle Miocene taxa (mainly Ekembo spp. and Equatorius) than 675 previously thought. At the same time, the new femoral fragments provide qualitative 676 support for some of the formerly proposed differences with other taxa, such as the 677 relative position of the gluteal tuberosity. 678 In addition, our results highlight the more primitive (stem hominoid-like) 679 appearance of the proximal femur in Nacholapithecus, in contrast to the more derived 680 (extant hominoid-like) traits found in its forelimb, which clearly depart from those 681 shown in Ekembo spp. Taking into account the whole evidence from Nacholapithecus, 682 the mosaic condition of its postcranial skeleton fits well within the positional behavior 683 scenario inferred for the early-middle Miocene of Africa (femoral morphology of 684 Nacholapithecus clearly departs from those of the Eurasian Miocene hominoids). Most 685 of the extinct hominoids found thus far would have a positional behavior repertoire that 686 might include frequent use of general arboreal quadrupedalism combined with other 687 ape-like antipronograde behaviors, such as vertical climbing. However, 688 Nacholapithecus might have displayed certain enhancement of forelimb-dominated 689 behaviors. Although further evolutionary studies are needed to corroborate this 690 hypothesis, such a unique combination in the Nacholapithecus skeleton could be the 691 origin of the more-derived behaviors found in younger Eurasian hominoids. 692 693 References 28 694 Aiello, L.C., Dean, C., 1990. The Hominoid Femur. In: Aiello, L.C., Dean, C. (Eds.), 695 An Introduction to Human Evolutionary Anatomy. Academic Press, London, pp. 696 457-482. 697 Alba, D.M., 2012. Fossil apes from the Vallès-Penedès Basin (NE Iberian Peninsula): 698 Phylogenetic, paleobiogeographic and paleobiological implications. Evol. 699 Anthropol. 21, 254-269. 700 Alba, D.M., Almécija, S., Casanovas-Vilar, I., Méndez, J.M., Moyà-Solà, S., 2012. A 701 partial skeleton of the fossil great ape Hispanopithecus laietanus from Can Feu 702 and the mosaic evolution of crown-hominoid positional behaviors. PLoS One 7, 703 e39617. 704 Almécija, S., Tallman, M., Alba, D.M., Pina, M., Moyà-Solà, S., Jungers, W.L., 2013. 705 The femur of Orrorin tugenensis exhibits morphometric affinities with both 706 Miocene apes and later hominins. Nat. Commun. 4, 2888. 707 Almécija, S., Tallman, M., Sallam, H.M., Fleagle, J.G., Hammond, A.S., Seiffert, E.R., 708 2019. Early anthropoid femora reveal divergent adaptive trajectories in catarrhine 709 hind-limb evolution. Nat. Commun. 10, 4778. 710 Bacon, A.-M., 2001. La locomotion des Primates du Miocène d'Afrique et d'Europe. 711 Cahiers de paléoanthropologie. CNRS Éditions, Paris. 712 Begun, D.R., 1992. Phyletic Diversity and Locomotion in Primitive European 713 Hominids. Am J. Phys. Anthropol. 87, 311-340. 714 Begun, D.R., 2013. The Miocene Hominoid Radiations. In: Begun, D.R. (Ed.), A 715 companion to paleoanthropology. Blackwell Publishing, Chichester, UK, pp. 398- 716 416. 29 717 Begun, D.R., 2015. Fossil Record of Miocene Hominoids. In: Henke, W., Tattersall, I. 718 (Eds.), Handbook of Paleoanthropology. Springer, Berlin Heidelberg, pp. 1261- 719 1332. 720 Begun, D.R., Kivell, T.L., 2011. Knuckle-walking in Sivapithecus? The combined 721 effects of homology and homoplasy with possible implications for pongine 722 dispersals. J. Hum. Evol. 60, 158-170. 723 Begun, D.R., Kordos, L., 2011. New postcrania of Rudapithecus hungaricus from 724 Rudabánya (Hungary). Am J. Phys. Anthropol. S52, 86. 725 Begun, D.R., Nargolwalla, M.C., Kordos, L., 2012. European Miocene hominids and 726 the origin of the African ape and human clade. Evol. Anthropol. 21, 10-23. 727 Böhme, M., Spassov, N., Fuss, J., Tröscher, A., Deane, A.S., Prieto, J., Kirscher, U., 728 Lechner, T., Begun, D.R., 2019. A new Miocene ape and locomotion in the 729 ancestor of great apes and humans. Nature 575, 489-493. 730 Cooke, S.B., Tallman, M., 2012. New endemic platyrrhine femur from Haiti: 731 Description and locomotor analysis. J. Hum. Evol. 63, 560-567. 732 DeSilva, J.M., Morgan, M.E., Barry, J.C., Pilbeam, D., 2010. A hominoid distal tibia 733 from the Miocene of Pakistan. J. Hum. Evol. 58, 147-154. 734 DeSilva, J.M., Holt, K.G., Churchill, S.E., Carlson, K.J., Walker, C.S., Zipfel, B., 735 Berger, L.R., 2013. The Lower Limb and Mechanics of Walking in 736 Australopithecus sediba. Science 340. 737 Fleagle, J.G., 1977. Locomotor behavior and muscular anatomy of sympatric Malaysian 738 leaf-monkeys (Presbytis obscura and Presbytis melalophos). Am J. Phys. 739 Anthropol. 46, 297-307. 30 740 Fleagle, J.G., Meldrum, D.J., 1988. Locomotor behavior and skeletal morphology of 741 two sympatric pitheciine monkeys, Pithecia pithecia and Chiropotes satanas. Am 742 J. Primatol. 16, 227-249. 743 Georgiou, L., Kivell, T.L., Pahr, D.H., Skinner, M.M., 2018. Trabecular bone patterning 744 in the hominoid distal femur. PeerJ 6, e5156. 745 Gommery, D., Senut, B., Pickford, M., 1998. Nouveaux restes postcrâniens 746 d'Hominoidea du Miocène inférieur de Napak, Ouganda. Annales de 747 Paléontologie 84, 287-306. 748 Gommery, D., Senut, B., Pickford, M., Musiime, E., 2002. Les nouveaux restes du 749 squelette d'Ugandapithecus major (Miocène inférieur de Napak, Ouganda). Ann. 750 de Paléontol. 88, 167-186. 751 Grand, T.I., 1968. The functional anatomy of the lower limb of the howler monkey 752 (Alouatta caraya). Am J. Phys. Anthropol. 28, 163-181. 753 Hammond, A.S., 2014. In vivo baseline measurements of hip joint range of motion in 754 suspensory and nonsuspensory anthropoids. Am J. Phys. Anthropol. 153, 417- 755 434. 756 Hammond, A.S., Rook, L., Anaya, A.D., Cioppi, E., Costeur, L., Moyà-Solà, S., 757 Almécija, S., 2020. Insights into the lower torso in late Miocene hominoid 758 Oreopithecus bambolii. Proc. Natl. Acad. Sci. USA 117, 278. 759 Harmon, E.H., 2007. The shape of the hominoid proximal femur: A geometric 760 morphometric analysis. J. Anat. 210, 170-185. 761 Harrison, T., 1982. Small-bodied apes from the Miocene of East Africa. Ph.D. 762 Dissertation, University College London. 763 Harrison, T., 1986. A reassessment of the phylogenetic relationships of Oreopithecus 764 bambolii Gervais. J. Hum. Evol. 15, 541-583. 31 765 Hürzeler, J., 1968. Oreopithecus bambolii Gervais. A preliminary report. 766 Verhandlungen Naturforchenden Gesellschaft Basel 69, 1-48. 767 Ishida, H., Pickford, M., Nakaya, H., Nakano, Y., 1984. Fossil anthropoids from 768 Nachola and Samburu Hills, Samburu District, Kenya. Afr. Study Monogr. Suppl. 769 Issue 2, 73-85. 770 Ishida, H., Mbua, E., Nakano, Y., Yasui, K., 1991. Sexual dimorphism in canine size of 771 Kenyapithecus from Nachola, Northern Kenya. In: Ehara, A., Kimura, T., 772 Takenaka, O., Iwamoto, M. (Eds.), Primatology Today: Proceedings of the XIIIth 773 Congress of the International Primatological Society, Nagoya and Kyoto, 18-24 774 July 1990. Elsevier, Amsterdam, pp. 517-520. 775 Ishidal, H., Kunimatsu, Y., Nakatsukasa, M., Nakano, Y., 1999. New Hominoid Genus 776 from the Middle Miocene of Nachola, Kenya. Anthropol. Sci. 107, 189-191. 777 Ishida, H., Kuminatsu, Y., Takano, T., Nakano, Y., Nakatsukasa, M., 2004. 778 Nacholapithecus skeleton from the Middle Miocene of Kenya. J. Hum. Evol. 46, 779 69-103. 780 Jungers, W.L., 1987. Body size and morphometric affinities of the appendicular 781 skeleton in Oreopithecus bambolii (IGF 11778). J. Hum. Evol. 16, 445-456. 782 Kelley, J., 2005. Twenty-five years contemplating Sivapithecus taxonomy. In: 783 Lieberman, D.E., Smith, R.H., Kelley, J. (Eds.), Interpreting the Past: Essays on 784 Human, Primate, and Mammal Evolution in Honor of David Pilbeam. Brill 785 Academic Publishers, Boston, pp. 123-143. 786 Kikuchi, Y., Nakano, Y., Nakatsukasa, M., Kunimatsu, Y., Shimizu, D., Ogihara, N., 787 Tsujikawa, H., Takano, T., Ishida, H., 2012. Functional morphology and anatomy 788 of cervical vertebrae in Nacholapithecus kerioi, a middle Miocene hominoid from 789 Kenya. J. Hum. Evol. 62, 677-695. 32 790 Kikuchi Y, Nakatsukasa M, Nakano Y, Kunimatsu Y, Shimizu D, Ogihara N, 791 Tsujikawa H, Takano T, Ishida H. 2015. Morphology of the thoracolumbar spine 792 of the middle Miocene hominoid Nacholapithecus kerioi from northern Kenya. J. 793 Hum. Evol., 88, 25-42. 794 Kikuchi, Y., Nakatsukasa, M., Nakano, Y., Kunimatsu, Y., Shimizu, D., Ogihara, N., 795 Tsujikawa, H., Takano, T., Ishida, H., 2016. Sacral vertebral remains of the 796 Middle Miocene hominoid Nacholapithecus kerioi from northern Kenya. J. Hum. 797 Evol. 94, 117-125. 798 Kikuchi, Y., Nakatsukasa, M., Tsujikawa, H., Nakano, Y., Kunimatsu, Y., Ogihara, N., 799 Shimizu, D., Takano, T., Nakaya, H., Sawada, Y., Ishida, H., 2018. Sexual 800 dimorphism of body size in an African fossil ape, Nacholapithecus kerioi. J. Hum. 801 Evol. 123, 129-140. 802 Köhler, M., Alba, D.M., Moyà-Solà, S., MacLatchy, L., 2002. Taxonomic affinities of 803 the Eppelsheim femur. Am J. Phys. Anthropol. 119, 297-304. 804 Kunimatsu, Y., 1992. New finds of a small anthropoid primate from Nachola, northern 805 Kenya. Afr. Study Monogr. 13, 237-249. 806 Kunimatsu, Y., 1997. New Species of Nyanzapithecus from Nachola, Northern Kenya. 807 Anthropol. Sci. 105, 117-141. 808 Kunimatsu, Y., Ishida, H., Nakatsukasa, M., Nakano, Y., Sawada, Y., Nakayama, K., 809 2004. Maxillae and associated gnathodental specimens of Nacholapithecus kerioi, 810 a large-bodied hominoid frm Nachola, northern Kenya. J. Hum. Evol. 46, 365- 811 400. 812 Leakey, R.E., Leakey, M.G., Walker, A.C., 1988. Morphology of Turkanapithecus 813 kalakolensis from Kenya. Am J. Phys. Anthropol. 76, 277-288. 33 814 Le Gros Clark, W.E., Leakey, L.S.B., 1951. The Miocene Hominoidea of East Africa. 815 Foss. Mamm. Afr. 1, 1-117. 816 Lovejoy, C.O., Heiple, K.G., Burnstein, A.H., 1973. The gait of Australopithecus. Am 817 J. Phys. Anthropol. 38, 757-780. 818 Lovejoy, C.O., Meindl, R.S., Ohman, J.C., Heiple, K.G., White, T.D., 2002. The Maka 819 femur and its bearing on the antiquity of human walking: Applying contemporary 820 concepts of morphogenesis to the human fossil record. Am J. Phys. Anthropol. 821 119, 97-133. 822 MacLatchy, L.M., 1996. Another look at the australopithecine hip. J. Hum. Evol. 31, 823 455-476. 824 MacLatchy, L., Gebo, D., Kityo, R., Pilbeam, D., 2000. Postcranial functional 825 morphology of Morotopithecus bishopi, with implications for the evolution of 826 modern ape locomotion. J. Hum. Evol. 39, 159-183. 827 Madar, S.I., Rose, M.D., Kelley, J., MacLatchy, L., Pilbeam, D., 2002. New 828 Sivapithecus postcranial specimens from the Siwaliks of Pakistan. J. Hum. Evol. 829 42, 705-752. 830 McCrossin, M.L., 1994. The phylogenetic relationship, adaptations, and ecology of 831 Kenyapithecus. Ph.D. Dissertation, University of California. 832 McHenry, H.M., Corruccini, R.S., 1976. Fossil hominid femora and the evolution of 833 walking. Nature 259, 657-658. 834 McHenry, H.M., Corruccini, R.S., 1978. The femur in early human evolution. Am J. 835 Phys. Anthropol. 49, 473-488. 836 Morgan, M.l.E., Lewton, K.L., Kelley, J., Otárola-Castillo, E., Barry, J.C., Flynn, L.J., 837 Pilbeam, D., 2015. A partial hominoid innominate from the Miocene of Pakistan: 838 Description and preliminary analyses. Proc. Natl. Acad. Sci. USA 112, 82-87. 34 839 Moyà-Solà, S., Köhler, M., 1996. A Dryopithecus skeleton and the origin of great-ape 840 locomotion. Nature 379, 156-159. 841 Moyà-Solà, S., Köhler, M., Rook, L., 1999. Evidence of hominid-like precision grip 842 capability in the hand of the Miocene ape Oreopithecus. Proc. Natl. Acad. Sci. 843 USA 96, 313-317. 844 Moyà-Solà, S., Köhler, M., Alba, D.M., Casanovas-Vilar, I., Galindo, J., Robles, J.M., 845 Cabrera, L., Garcés, M., Almécija, S., Beamud, E., 2009. First partial face and 846 upper dentition of the Middle Miocene hominoid Dryopithecus fontani from 847 Abocador de Can Mata (Vallès-Penedès Basin, Catalonia, NE Spain): Taxonomic 848 and phylogenetic implications. Am J. Phys. Anthropol. 139, 126-145. 849 Nakano, Y., Ogihara, N., Makishima, H., Shimizu, D., Kagaya, M., Kunimatsu, Y., 850 Ishida, H., 2004. The locomotor adaptation in the pelvic morphology of 851 Nacholapithecus. Anthropol. Sci. 112, 301. 852 Nakatsukasa, M., 2008. Comparative study of Moroto vertebral specimens. J. Hum. 853 Evol. 55, 581-588. 854 Nakatsukasa, M., Kunimatsu, Y., 2009. Nacholapithecus and its importance for 855 understanding hominoid evolution. Evol. Anthropol. 18, 103-119. 856 Nakatsukasa, M., Yamanaka, A., Kunimatsu, Y., Shimizu, D., Ishida, H., 1998. A 857 newly discovered Kenyapithecus skeleton and its implications for the evolution of 858 positional behavior in Miocene East African hominoids. J. Hum. Evol. 34, 657- 859 664. 860 Nakatsukasa, M., Tsujikawa, H., Shimizu, D., Takano, T., Kunimatsu, Y., Nakano, Y., 861 Ishida, H., 2003a. Definitive evidence for tail loss in Nacholapithecus, an East 862 African Miocene hominoid. J. Hum. Evol. 45, 179-186. 35 863 Nakatsukasa, M., Kuminatsu, Y., Nakano, Y., Takano, T., Ishida, H., 2003b. 864 Comparative and functional anatomy of phalanges in Nacholapithecus kerioi, a 865 Middle Miocene hominoid from northern Kenya. Primates 44, 371-412. 866 Nakatsukasa, M., Ward, C.V., Walker, A., Teaford, M.F., Kunimatsu, Y., Ogihara, N., 867 2004. Tail loss in Proconsul heseloni. J. Hum. Evol. 46, 777-784. 868 Nakatsukasa, M., Kunimatsu, Y., Nakano, Y., Ishida, H., 2007a. Vertebral morphology 869 of Nacholapithecus kerioi based on KNM-BG 35250. J. Hum. Evol. 52, 347-369. 870 Nakatsukasa, M., Kunimatsu, Y., Nakano, Y., Egi, N., Ishida, H., 2007b. Postcranial 871 bones of infant Nacholapithecus: Ontogeny and positional behavioral adaptation. 872 Anthropol. Sci. 115, 201-213. 873 Nakatsukasa, M., Kunimatsu, Y., Shimizu, D., Nakano, Y., Kikuchi, Y., Ishida, H., 874 2012. Hind limb of the Nacholapithecus kerioi holotype and implications for its 875 positional behavior. Anthropol. Sci. 120, 235-250. 876 Ogihara, N., Almécija, S., Nakatsukasa, M., Nakano, Y., Kikuchi, Y., Kunimatsu, Y., 877 Makishima, H., Shimizu, D., Takano, T., Tsujikawa, H., Kagaya, M., Ishida, H., 878 2016. Carpal bones of Nacholapithecus kerioi, a Middle Miocene Hominoid from 879 Northern Kenya. Am J. Phys. Anthropol. 160, 469-482. 880 Pickford, M., Ishida, H., Nakaya, Y., Yasui, K., 1987. The Middle Miocene fauna from 881 the Nachola and Aka Aiteputh Formations, Northern Kenya. Afr. Study Monogr. 882 Suppl. Issue 5, 141-154. 883 Pickford, M., Senut, B., Gommery, D., Treil, J., 2002. Bipedalism in Orrorin tugenensis 884 revealed by its femora. Comptes Rendus Palevol. 1, 191-203. 885 Pilbeam, D.R., Young, D.L., 2001. Sivapithecus and hominoid evolution: Some brief 886 comments. In: de Bonis, L., Koufos, G.D., Andrews, P. (Eds.), Hominoid 887 Evolution and Climatic Change in Europe, Vol. 2. Phylogeny of the Neogene 36 888 Hominoid Primates of Eurasia. Cambridge University Press, Cambridge, pp. 349- 889 364. 890 Pilbeam, D.R., Rose, M.D., Badgley, C., Lipschutz, B., 1980. Miocene hominoids from 891 Pakistan. Postilla 181, 1-94. 892 Pina, M., 2016. Unravelling the positional behaviour of fossil hominoids: 893 Morphofunctional and structural analysis of the primate hindlimb. Ph.D. 894 Dissertation, Universitat Autònoma de Barcelona. 895 Pina, M., Alba, D.M., Almécija, S., Fortuny, J., Moyà-Solà, S., 2012. Paleobiological 896 inferences on the locomotor repertoire of extinct hominoids based on femoral 897 neck cortical thickness: The fossil Great ape Hispanopithecus laietanus as a test- 898 case study. Am J. Phys. Anthropol. 149, 142-148. 899 Pina, M., Almécija, S., Alba, D.M., O’Neill, M.C., Moyà-Solà S., 2014. The Middle 900 Miocene ape Pierolapithecus catalaunicus exhibits extant great ape-like 901 morphometric affinities on its patella: Inferences on knee function and evolution. 902 PLoS One 9, e91944. 903 Pina, M., Kikuchi, Y., Nakatsukasa, M., Nakano, Y., Kunimatsu, Y., Ogihara, N., 904 Shimizu, D., Takano, T., Tsujikawa, H., Ishida, H., 2018. Revisiting the femoral 905 morphology of Nacholapithecus kerioi. Anthropol. Sci. 126 (3): 188. 906 Pina, M., Alba, D.M., Moyà-Solà, S., Almécija, S., 2019. Femoral neck cortical bone 907 distribution of dryopithecin apes and the evolution of hominid locomotion. J. 908 Hum. Evol. 136, 102651. 909 Pina, M., DeMiguel, D., Puigvert, F., Marcé-Nogué, J., Moyà-Solà S., 2020. Knee 910 function through finite element analysis and the role of Miocene hominoids to 911 understand the origin of antipronograde behaviours: The Pierolapithecus 912 catalaunicus patella as a test-case study. Palaeontology 63, 459-475. 37 913 R Core Team, 2017. R: A language and environment for statistical computing. R 914 Foundation for Statistical Computing, Vienna. 915 Raza, S.M., Barry, J.C., Pilbeam, D., Rose, M.D., Shah, S.M.I., Ward, S., 1983. New 916 hominoid primates from the middle Miocene Chinji Formation, Potwar Plateau, 917 Pakistan. Nature 306, 52-54. 918 Richmond, B.G., Whalen, M., 2001. Forelimb function, bone curvature and phylogeny 919 of Sivapithecus. In: de Bonis, L., Koufos, G.D., Andrews, P. (Eds.), Hominoid 920 Evolution and Climatic Change in Europe, Vol. 2. Phylogeny of the Neogene 921 Hominoid Primates of Eurasia. Cambridge University Press, Cambridge, pp. 326- 922 348. 923 Richmond, B.G., Jungers, W.L., 2008. Orrorin tugenensis femoral morphology and the 924 evolution of hominin bipedalism. Science 319, 1662-1665. 925 Rook, L., Renne, P., Benvenuti, M., Papini, M., 2000. Geochronology of Oreopithecus- 926 bearing succession at Baccinello (Italy) and the extinction pattern of European 927 Miocene hominoids. J. Hum. Evol. 39, 577-582. 928 Rose, M.D., 1983. Miocene hominoid postcranial morphology. Monkey-like, ape-like, 929 neither, or both?. In: Ciochon, R.L., Corruccini, R.S. (Eds.), New Interpretations 930 of Ape and Human Ancestry. Plenum Press, New York, pp. 503-516. 931 Rose, M.D., 1986. Further hominoid postcranial specimens from late Miocene Nagri 932 Formation of Pakistan. J. Hum. Evol. 15, 333-367. 933 Rose, M.D., 1993. Locomotor anatomy of Miocene hominoids. In: Gebo, D.L. (Ed.), 934 Postcranial Adaptation in Nonhuman Primates. Northern Illinois University Press, 935 Dekalb, pp. 252-272. 936 Rose, M.D., Nakano, Y., Ishida, H., 1996. Kenyapithecus postcranial specimens from 937 Nachola, Kenya. Afr. Study Monogr. Suppl. 24, 3-56. 38 938 Ruff, C., 1988. Hindlimb articular surface allometry in hominoidea and Macaca, with 939 comparisons to diaphyseal scaling. J. Hum. Evol. 17, 687-714. 940 Sawada, Y., Pickford, M., Itaya, T., Makinouchi, T., Tateishi, M., Kabeto, K., Ishida, 941 S., Ishida, H., 1998. K-Ar ages of miocene hominoidea (Kenyapithecus and 942 Samburupithecus) from Samburu Hills, Northern Kenya. C. R. Acad. Sci. IIA 943 Earth Planet. Sci. 326, 445-451. 944 Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch, T., 945 Preibisch, S., Rueden, C., Saafeld, S., Schmid, B., Tinevez, J.-Y., White, D. J., 946 Hartenstein, V., Eliceiri, K., Tomancak, P., Cardona, A., 2012. Fiji: An open- 947 source platform for biological-image analysis. Nat. Methods 9, 676-682. 948 Senut, B., 2016. Morphology and environment in some fossil hominoids and pedetids 949 (Mammalia). J. Anat. 228, 700-715. 950 Senut, B., Pickford, M., Gommery, D., Kunimatsu, Y., 2000. A new genus of Early 951 Miocene hominoid from East Africa: Ugandapithecus major (Le Gros Clark & 952 Leakey, 1950). C. R. Acad. Sci. IIA Earth Planet. Sci. 331, 227-233. 953 Senut, B., Pickford, M., Gommery, D., Mein, P., Cheboi, K., Coppens, Y., 2001. First 954 hominid from the Miocene (Lukeino Formation, Kenya). C. R. Acad. Sci. IIA 955 Earth Planet. Sci. 332, 137-144. 956 Senut, B., Nakatsukasa, M., Kunimatsu, Y., Nakano, Y., Takano, T., Tsujikawa, H., 957 Shimizu, D., Kagaya, M., Ishida, H., 2004. Preliminary analysis of 958 Nacholapithecus scapula and clavicle from Nachola, Kenya. Primates 45, 97-104. 959 Stern, Jr., J.T., 1972. Anatomical and functional specializations of the human gluteus 960 maximus. Am J. Phys. Anthropol. 36, 315-339. 39 961 Stern, Jr., J.T., Susman, R.L., 1981. Electromyography of the gluteal muscles in 962 Hylobates, Pongo, and Pan: Implications for the evolution of hominid bipedality, 963 Am J. Phys. Anthropol. 55, 153-66. 964 Straus, W.L., Jr., 1963. The classification of Oreopithecus. In: Washburn, S.L. (Ed.), 965 Classification and Human Evolution. Aldine, Chicago, pp. 146-177. 966 Sukhdeo, S., Parsons, J., Niu, X.M., Ryan, T.M., 2019. Trabecular bone structure in the 967 distal femur of humans, apes, and baboons. Anat. Rec. doi:10.1002/ar.24050. 968 Susman, R.L., 2004. Oreopithecus bambolii: An unlikely case of hominidlike grip 969 capability in a Miocene ape. J. Hum. Evol. 46, 105-117. 970 Takano, T., Nakatsukasa, M., Kunimatsu, Y., Nakano, Y., Ishida, H., 2003. Functional 971 morphology of the Nacholapithecus forelimb long bones. Am J. Phys. Anthropol. 972 120, 205-206. 973 Takano, T., Nakatsukasa, M., Kunimatsu, Y., Nakano, Y., Ogihara, N., Ishida, H., 2018. 974 Forelimb long bones of Nacholapithecus (KNM-BG 35250) from the middle 975 Miocene in Nachola, northern Kenya. Anthropol. Sci. 126, 135-149. 976 Takano, T., Nakatsukasa, M., Pina, M., Kunimatsu, Y., Nakano, Y., Morimoto, 977 N., Ogihara, N., Ishida, H., In press. New forelimb long bone specimens 978 of Nacholapithecus kerioi from the Middle Miocene of northern 979 Kenya. Anthropol. Sci. 128. 980 Ward, C.V., 2015. Postcranial and locomotor adaptations of hominoids. In: Henke, W., 981 Tattersall, I. (Eds.), Handbook of Paleoanthropology. Springer, Berlin Heidelberg, 982 pp. 1363-1386. 983 Ward, C.V., Walker, A., Teaford, M.F., Odhiambo, I., 1993. Partial skeleton of 984 Proconsul nyanzae from Mfangano Island, Kenya. Am J. Phys. Anthropol. 90, 77- 985 111. 40 986 Ward, C.V., Ruff, C.B., Walker, A., Teaford, F., Rose, M.D., Nengo, I.O., 1995. 987 Functional morphology of Proconsul patellas from Rusinga Island, Kenya, with 988 implications for other Miocene-Pliocene catarrhines. J. Hum. Evol. 29, 1-19. 989 Ward, C.V., Hammond, A.S., Plavcan, J.M., Begun, D.R., 2019. A late Miocene 990 hominid partial pelvis from Hungary. J. Hum. Evol., 102645. 991 992 Figure captions 993 Figure 1. Map of the fossil locality of Nachola in Kenya. 994 995 Figure 2. Linear measurements taken of the proximal femora for quantitative analyses. 996 A) proximal view; B) posterior view; and C) anterior view. Abbreviations: APN = 997 anteroposterior depth of the femoral neck; SIH = maximum superoinferior height of the 998 femoral head; SIN = minimum superoinferior height of the femoral neck; NL = femoral 999 neck length (between the most lateral edge of the femoral head to the medialmost limit 1000 of the trochanteric crest); NSangle = femoral neck-shaft angle; TotW = total 1001 mediolateral width of the proximal femur from the medialmost point of the femoral 1002 head to the lateralmost point of the greater trochanter. Modified from Pina, 2016. 1003 1004 Figure 3. Proximal fossil femoral fragments used for comparisons with 1005 Nacholapithecus kerioi (anterior view). When available, the distal portion is also 1006 displayed (anterior view above, posterior view below). A) Morotopithecus bishopi 1007 (UMP MORII 94’80); B) Proconsul major (combination of NAP IX 46’99, NAP IX B 1008 64, NAP IX 65 P. 67 fragments; reversed); C) Turkanapithecus kalakolensis (KNM- 1009 WK 16950I); D) Ekembo nyanzae (KNM-MW 13142A); E) Ekembo nyanzae (KNM- 1010 RU 5527; reversed); F) Equatorius africanus (BMNH M.16331; reversed; pictures from 41 1011 cast); G) Dryopithecus fontani (IPS 41724); and H) Hispanopithecus laietanus (IPS 1012 18800.29). Femora are displayed with the same maximum superoinferior height of the 1013 femoral head (SIH) to facilitate morphological comparisons. Dashed lines at the right 1014 bottom of A, C, and E represent the outline of the patellar groove. Scale bars = 20 mm. 1015 1016 Figure 4. Fossil femoral fragments belonging to Nacholapithecus kerioi described in 1017 this work (A–P). KNM-BG 17778 (A, anterior view; B, posterior view); KNM-BG 1018 40844 (C, anterior view; D, posterior view); KNM-BG 40964 (E, anterior view; F, 1019 posterior view); KNM-BG 42757 (G, proximal; H, anterior; I, medial; and J, posterior 1020 views); KNM-BG 44953B (K, anterior; L, medial; and M, posterior views); and KNM- 1021 BG; 42779 (N, anterior; O, side; and P, posterior views). Scale bar = 20 mm. 1022 1023 Figure 5. Fossil femoral fragments belonging to Nacholapithecus kerioi described in 1024 this study. These remains are associated with the same individual (probably an 1025 immature). A–C) KNM-BG 42738/42756C (proximal fragment); D–E) KNM-BG 1026 42722 (shaft fragment); F–I) KNM-BG 42732 (distal fragment). A, I) proximal views; 1027 B, D, F) anterior views; C, E, G) posterior views; H) distal view. Black arrows denote 1028 (A) the absence of anteversion of the head, (C) the intertrochanteric line, and (F) the 1029 most proximal point of the lateral and medial rims of the patellar groove. Dashed white 1030 lines denote the supracondylar ridges (E). Dashed black line represents the square- 1031 shaped outline of the patellar groove (F). Scale bar = 20 mm. 1032 1033 Figure 6. Boxplots showing (A) the relative size of the head index: SIH/(√SIN*APN) 1034 (SIH = superoinferior height of the femoral head; SIN = superoinferior height of the 1035 femoral neck; APN = anteroposterior depth); (B) the relative neck length (RelativeNL); 42 1036 and (C) neck-shaft angle (NSangle). Vertical lines represent the median, boxes the 1037 interquartile range, IQR (between the 25th and the 75th percentiles), whiskers the 1038 1.5*IQR, and circles the outliers. Colors represent major taxonomic groups: dark green, 1039 African apes; light green, Asian apes; yellow, cercopithecines; orange, colobines; dark 1040 blue, atelids; light blue, cebines (for colors see online version). 1041 1042 Figure 7. Nacholapithecus kerioi femoral fragments showing some of the anatomical 1043 features discussed in the text. A) KNM-BG 38391A (anterior view); B) KNM-BG 1044 44953A (anterior view); C) KNM-BG 48093 (anterior view); D–E) KNM-BG 17816 1045 (D, anterior view; E, posterior view); F–G) KNM-BG 44954A (F, posterior view; G, 1046 lateral view); H) KNM-BG 35250A (holotype; anterior view); I) KNM-BG 35250J 1047 (holotype; anterior view); J) AMNH 103659, Macaca fascicularis (anterior view) ; K) 1048 AMNH 103344, Hylobates klossi (anterior view); L) AMNH 86857, Pan paniscus 1049 (anterior view). J–L are depicted at the same femoral length (from the head to the distal 1050 end). A–B) differences in greater trochanter lateral flare (fragments scaled to the same 1051 superoinferior height of the femoral head). A, C) sexual dimorphism: C, smallest femur 1052 belonging to a female (Kikuchi et al., 2018); its small size can be compared with A, the 1053 largest male femur (except for H; Kikuchi et al., 2018). A and C femoral fragments are 1054 depicted to the same scale. D–E) lesser trochanter close to the femoral neck; F–G) 1055 marked gluteal tuberosity close to the greater trochanter. I) square-shaped patellar 1056 groove. Black arrows highlight the referenced anatomical traits. Scale bar = 20 mm. 43 Table 1 Femoral remains attributed to Nacholapithecus kerioi. Sex is provided for those specimens included in the sexual dimorphism analysis in Kikuchi et al. (2018).a Accession number Description Side Sex Locality First reference APN SIH SIN NL TotW NSangle KNM-BG 17775 Head with neck ? BG-X Rose et al. (1996) KNM-BG 17778 Head ?L BG-X This studyb KNM-BG 17816 Proximal shaft, including LT R BG-I Rose et al. (1996) KNM-BG 17819 Head with neck R BG-I Rose et al. (1996) KNM-BG 17820 Proximal fragment L BG-I Rose et al. (1996) KNM-BG 17821 Head with neck ? BG-I Rose et al. (1996) KNM-BG 35250A Proximal half R M BG-K Nakatsukasa et al. (1998)c 12.1 22.3 15.5 129.4 KNM-BG 35250B Distal fragment R BG-K Nakatsukasa et al. (1998) KNM-BG 35250D Proximal fragment L BG-K Nakatsukasa et al. (1998) KNM-BG 35250J Shaft and distal end L BG-K Nakatsukasa et al. (1998) KNM-BG 35250U Proximal fragment L BG-K Nakatsukasa et al. (1998) KNM-BG 38391A Proximal fragment L M BG-K Kikuchi et al. (2018)c 9.6 22.0 16.2 23.4 49.7 115.3 KNM-BG 40794A Head with neck R F BG-K Kikuchi et al. (2018)c 7.8 18.5 14.5 KNM-BG 40826 Proximal fragment R F BG-K Kikuchi et al. (2018)c 8.9 17.3 13.4 130.1 KNM-BG 40844 Head ?R BG-K This study KNM-BG 40933 Head with neck L M BG-K Kikuchi et al. (2018)c 12.3 21.9 16.9 KNM-BG 40964 Proximal fragment without head and neck R BG-K This study KNM-BG 42713A Proximal half L F BG-K Kikuchi et al. (2018)c KNM-BG 42722 Shaft fragment R BG-I This study KNM-BG 42732 Distal fragment L BG-I This study KNM-BG 42738/42756C Proximal fragment L M BG-I West Kikuchi et al. (2018)c 12.9 24.2 16.5 25.8 54.2 121.0 KNM-BG 42757 Proximal fragment L BG-K This studyb 114.9 KNM-BG 42779 Shaft fragment (distal) L BG-K This study KNM-BG 44953A Proximal fragment R F BG-K Kikuchi et al. (2018)c 8.1 17.1 13.1 17.3 37.7 108.3 KNM-BG 44953B Proximal fragment without head and neck L BG-K This study KNM-BG 44954A Proximal fragment L M BG-K Kikuchi et al. (2018)c KNM-BG 48092A Proximal fragment R F BG-K Kikuchi et al. (2018)c 9.1 18.0 13.8 18.0 38.5 119.2 KNM-BG 48093 Proximal fragment L F BG-K Kikuchi et al. (2018)c 8.2 17.0 13.3 17.3 38.3 107.1 Abbreviations: APN = anteroposterior depth of the femoral neck (mm); BG = Baragoi; F = female; L = left; LT = lesser trochanter; M = male; NL = neck length (mm); NSangle = neck-shaft angle (degrees); R = right; SIH = superoinferior height of the femoral head (mm); SIN = superoinferior height of the femoral neck (mm); TotW = total mediolateral width of the proximal femur from the medialmost point of the femoral head to the lateral-most point of the greater trochanter (mm); ? = uncertain. a These femoral fragments are fully open access for further analytical studies. b Used for comparative purposes (Ishida et al., 2004; Nakatsukasa et al., 2012), but never formally described. c Fragment included in the quantitative analyses. Table 2 Extant anthropoid taxa included in the analyses. Number of females/males/unknown sex in parentheses.a n Species SIH/(√SIN*APN) Relative NL NSangle Cebus apella 33 (13/20/-) 27 (12/15/-) 27 (12/15/-) Ateles sp. 8 (5/2/1) 8 (2/2/4) 8 (2/2/4) Alouatta sp. 45 (22/19/4) 28 (15/8/5) 30 (14/10/6) Presbytis sp. 34 (20/14/-) 25 (14/11/-) 33 (16/17-/) Colobus sp. 28 (13/15/-) 32 (12/18/2) 27 (12/14/1) Nasalis larvatus 25 (12/13/-) 25 (12/13/-) 25 (12/13/-) Chlorocebus sp. 16 (6/8/2) 10 (4/5/1) — Cercopithecus sp. 49 (19/30/-) 37 (11/20/6) 14 (5/8/1) Macaca sp. 30 (15/15/-) 27 (13/14/-) 26 (13/13/-) Lophocebus sp. 15 (2/12/1) 7 (1/5/1) — Mandrillus sp. 13 (4/8/1) 10 (3/6/1) 10 (3/6/1) Papio sp. 25 (5/11/9) 18 (2/8/8) 20 (4/9/7) Hylobates lar 26 (13/13/-) 25 (12/13/-) 26 (13/13/-) Pongo pygmaeus 12 (4/5/3) 11 (5/4/2) 12 (4/5/3) Pan t. troglodytes 29 (14/15/-) 17 (4/9/4) — Pan t. schweinfurthii 25 (8/17/-) 21 (7/10/4) 26 (8/16/2) Pan paniscus 20 (11/9/-) 20 (11/9/-) 20 (11/9/-) Gorilla g. gorilla 31 (13/18/-) 20 (10/10/-) 26 (13/13/-) Gorilla b. graueri 21 (8/13/-) 22 (8/14/-) 21 (8/13/-) Gorilla b. beringei 10 (5/5/-) 7 (4/3/-) 8 (4/4/-) Abbreviations: APN = anteroposterior depth of the femoral neck; n = sample size; NL = neck length; NSangle = neck-shaft angle; SIH = superoinferior height of the femoral head; SIN = superoinferior height of the femoral neck. a Data for these femora were collected at the American Museum of Natural History, New York (AMNH, USA), the Museum of Comparative Zoology, Harvard University (MCZ, USA), Peabody Museum of Archaeology and Ethnology, Harvard University (PBMA, USA), and the Royal Museum of Central Africa (RMCA, Belgium). Table 3 Descriptive statistics for the SIH/(√SIN*APN) index, relative NL, and NSangle variables in the Nacholapithecus kerioi sample. n Mean SD Min Max SIH/(√SIN*APN) 9 1.644 0.076 1.519 1.764 Relative NL 5 46.50 0.969 45.17 47.60 NSangle 8 118.2 8.587 107.1 130.1 Abbreviations: APN = anteroposterior neck depth; n = sample size; Max = maximum value; Min = minimum value; NL = neck length; NSangle = neck-shaft angle; SIH = superoinferior height of the femoral head; SIN = superoinferior height of the femoral neck; SD = standard deviation. Table 4 Post-hoc pairwise comparisons between Nacholapithecus kerioi and extant anthropoids for the SIH/(√SIN*APN) index, relative NL, and NSangle.a Nacholapithecus kerioi SIH/(√SIN*APN) Relative NL NSangle Cebus apella NS NS NS Atelids NS NS ** Colobines *** * * Cercopithecines ** * ** Hylobates lar *** *** * Pongo pygmaeus * NS *** Pan troglodytes troglodytes NS NS — Pan troglodytes schweinfurthii NS NS * Pan paniscus NS ** NS Gorilla gorilla gorilla NS NS NS Gorilla beringei graueri NS NS NS Gorilla beringei beringei NS NS NS Abbreviations: SIH/(√SIN*APN) = relative size of the femoral head (APN = anteroposterior neck depth; SIH = superoinferior height of the femoral head; SIN = superoinferior height of the femoral neck); Relative NL = neck length divided by total mediolateral width of the proximal femur multiplied by 100; NSangle = neck-shaft angle of the femur. a NS, no significant differences; *, p < 0.05; **, p < 0.005; ***, p < 0.001.