This is the accepted version of the journal article: Arias-Martorell, Julia; Urciuoli, Alessandro; Almécija, Sergio; [et al.]. «The radial head of the Middle Miocene ape Nacholapithecus kerioi : Morphometric affinities, locomotor inferences, and implications for the evolution of the hominoid humeroradial joint». Journal of Human Evolution, Vol. 178 (May 2023), art. 103345. DOI 10.1016/j.jhevol.2023.103345 This version is available at https://ddd.uab.cat/record/273107 under the terms of the license Short Communication The radial head of the Middle Miocene ape Nacholapithecus kerioi: Morphometric affinities, locomotor inferences, and implications for the evolution of the hominoid humeroradial joint Julia Arias-Martorella,b,*, Alessandro Urciuolic,d,a, Sergio Almécijae,f,a, David M. Albaa, Masato Nakatsukasag a Insitut Català de Paleontologia Miquel Crusafont, Universitat Auntònoma de Barcelona, Edifici ICTA-ICP, c/ Columnes s/n, Campus de la UAB, 08193 Cerdanyola del Vallès, Barcelona, Spain b School of Anthropology and Conservation, Marlowe Building, University of Kent, Canterbury, CT2 7NR, UK c Universitat Autònoma de Barcelona, Campus de la UAB, 08193 Cerdanyola del Vallès, Barcelona, Spain d Division of Palaeoanthropology, Senckenberg Research Institute and Natural History Museum Frankfurt, Frankfurt am Main, Germany e Division of Anthropology, American Museum of Natural History, Central Park West at 79th Street, New York, NY 10024, USA f New York Consortium in Evolutionary Primatology, New York, NY 10024, USA g Laboratory of Physical Anthropology, Graduate School of Science, Kyoto University, 606- 8502 Kyoto, Japan *Corresponding author. E-mail address: julia.arias@icp.cat (J. Arias-Martorell). Acknowledgments We would like to thank NACOSTI for permission to carry out this research in Kenya; we also thank the Director, Head of Earth Sciences and Palaeontology Section Staff at the National Museums of Kenya for their support and collaboration in this project. We would also like to thank the Associate Editor and three reviewers for their help improving earlier versions of this manuscript. This work is part of R+D+I projects PID2020-116908GB-I00 and PID2020- 117289GBI00, funded by the Agencia Estatal de Investigación of the Spanish Ministry of Science and Innovation (MCIN/AEI/10.13039/501100011033/). Research has also been supported by the Generalitat de Catalunya/CERCA Programme, the Agència de Gestió d’Ajust Universitaris i de Recerca of the Generalitat de Catalunya (2022 SGR 01188 and 2022 SGR 00620, and Beatriu de Pinós Programme—H2020 MSCA-Cofund Grant No 801370 to J.A.M), a Margarita Salas postdoctoral fellowship funded by the European Union-NextGenerationEU to A.U, and the grants JPJSBP120216301 and KAKENHI 22H02708 to M.N. 1 Short Communication 2 The radial head of the Middle Miocene ape Nacholapithecus kerioi: Morphometric affinities, 3 locomotor inferences, and implications for the evolution of the hominoid humeroradial joint 4 5 Keywords: Miocene apes; Forelimb; Elbow joint; Pronosupination; Locomotion; 3D 6 geometric morphometrics 7 8 1 9 1. Introduction 10 The systematic position of the Middle Miocene large-bodied ape Nacholapithecus 11 kerioi (see review in Nakatsukasa and Kunimatsu, 2009) is somewhat uncertain (e.g., 12 Almécija et al., 2021), being considered a stem hominid by some researchers (Alba, 2012; 13 Kunimatsu et al., 2019) but being favored as a stem hominoid by recent cladistic analyses 14 (Pugh, 2022). The remains of this species, dated to the Middle Miocene (16 –15 Ma; Sawada 15 et al., 1998, 2006; Nakatsukasa and Kunimatsu, 2009), were originally discovered in the Aka 16 Aiteputh Formation of the Nachola region (northern Kenya) and assigned to Kenyapithecus 17 (Ishida et al., 1984; Rose et al., 1996; Nakatsukasa et al., 1998). However, the discovery of 18 the articulated partial skeleton KNM-BG 35250, of which numerous postcranial elements and 19 the cranium were recovered (Nakatsukasa et al., 1998; Ishida et al., 2004), prompted the 20 description of a new genus and species for the large-bodied hominoid from Nachola (Ishida et 21 al., 1999). 22 Many postcranial elements of N. kerioi have been described in detail from the 23 holotype KNM-BG 25350 skeleton (Nakatsukasa et al., 1998, 2003a, 2007b, 2012; Ishida et 24 al., 2004; Senut et al., 2004; Kikuchi et al., 2012, 2015, 2016; Ogihara et al., 2016; Takano et 25 al., 2018), as well as from numerous (mostly isolated) finds (Rose et al., 1996; Nakatsukasa et 26 al., 2003b, 2007a; Pina et al., 2021; Takano et al., 2020). From these studies, it has been 27 inferred that N. kerioi possessed a pronograde body plan (narrow thorax, long and flexible 28 vertebral column, and limbs used mostly in the parasagittal plane; Nakatsukasa and 29 Kunimatsu, 2009), like other Early and Middle Miocene apes such as Ekembo and Equatorius 30 (see review in Ward, 2015). However, Nacholapithecus exhibits derived postcranial features 31 compared with those of earlier (e.g., Ekembo) and contemporaneous (e.g., Equatorius) stem 32 hominoids, especially in the elbow joint, which exhibits a deep zona conoidea and a large, 33 globular, and medially tilted capitulum in the distal humerus (Nakatsukasa and Kunimatsu, 34 2009; Takano et al. 2020). These derived features indicate that Nacholapithecus displays the 2 35 earliest known evidence of increased forelimb-dominated behaviors with enhanced vertical 36 climbing capabilities among fossil apes (Nakatuskasa and Kunimatsu, 2009; Takano et al., 37 2018, 2020). 38 The distal humeral and proximal ulnar morphology of Nacholapithecus have been 39 interpreted from a locomotor viewpoint (Takano et al., 2018, 2020) but its proximal radial 40 morphology remains to be analyzed from this perspective. The distinctively derived proximal 41 radial morphology of extant hominoids is functionally related to wide ranges of 42 pronosupination coupled with universal stability at the humeroradial joint (Sarmiento, 1988; 43 Rose, 1988, 1993; Sarmiento et al., 2002: Fig. 4) and has been used to make locomotor 44 inferences in fossil catarrhines (Arias-Martorell et al., 2021). Therefore, here we provide a 45 quantitative morphological analysis of the radial head of N. kerioi with the objective of 46 refining previous locomotor inferences for this species. Our analysis is based on the isolated 47 proximal radius fragment KNM-BG 40021 from the fossiliferous site BG-K (see Takano et 48 al., 2020 for a full description of the specimen). We compare it with stem catarrhines and 49 other Miocene hominoids by means of three-dimensional geometric morphometrics (3DGM) 50 to establish its closest morphometric affinities. 51 52 2. Materials and methods 53 2.1. Studied and comparative sample 54 The right proximal radial fragment of N. kerioi KNM-BG 40021 (Fig. 1; Takano et al., 55 2020) is housed at the National Museums of Kenya (KNM, Nairobi, Kenya). KNM-BG 40021 56 is a 63 mm-long fragment preserving the radial head, radial tuberosity, and a small section of 57 the shaft below the radial tuberosity. The specimen has compression damage affecting mainly 58 the shaft and radial tuberosity, which are both flattened anteroposteriorly. The posterior aspect 59 of the radial head is also flattened, affecting the depth of the fovea, which shows an 60 artifactually increased depth. However, the anterior and medial/lateral aspects of the radial 3 61 head are well-preserved and preserve their original morphology, including the outline of the 62 fovea (Takano et al., 2020). Our analyses focused on this undistorted and well-preserved 63 aspect of the radial head, the shape of which we characterized using 3DGM to capture more 64 subtle aspects of variation such as the curvature and outline of the radial head, which are very 65 important aspects of radial head variation and function. 66 The comparative fossil sample includes 3D virtual models of the radii of the 67 dendropithecids Simiolus enjisessi and Dendropithecus macinessi, the pliopithecoid 68 Epipliopithecus vindobonensis, and the stem hominoids Ekembo heseloni and Equatorius 69 africanus (see Supplementary Online Material [SOM] Table S1 for further details). The 70 extant comparative sample is the same as in Arias-Martorell et al. (2021), including 116 radii 71 from 26 anthropoid species including all extant hominoid genera (SOM Table S2). Three- 72 dimensional landmarks were collected from 3D models of the radii listed in SOM Tables S1 73 and S2. Details about the scanning procedures of both the fossil and extant comparative 74 sample are presented in SOM S1 and SOM Table S2. All 3D models from right radii 75 (including fossils) were mirrored to the left side during the process of mesh reconstruction as 76 most extant anthropoid radii in our sample are from the left side. Landmarks were placed 77 using IDAV Landmark Editor v. 3.6 (Wiley et al., 2005) and all statistical analysis were done 78 with the statistical environment R v. 4.1.1 (R Core Team, 2021). 79 80 2.2. Geometric morphometric analyses 81 The shape affinities of KNM-BG 40021 were explored using a landmark protocol 82 specifically designed to capture the most informative aspects of shape preserved in this 83 specimen (Fig. 2; for further details see SOM S1 and Table S3). We performed a generalized 84 Procrustes analysis (GPA) with the ‘Morpho’ v. 2.8 package (Schlager, 2017) in R (R Core 85 Team, 2021). We applied semilandmark sliding (identified with SMvector; Schlager, 2017) 86 on curves defined by adjacent landmarks and identified with the function ‘outline’ (Schlager, 4 87 2017). To identify major patterns of shape variation across the sample, we performed a 88 between-group principal component analysis (bgPCA; Mitteroecker and Bookstein, 2011) on 89 the GPA-transformed coordinates of the extant sample, with major anthropoid clades 90 (platyrrhines, cercopithecines, colobines, hylobatids, and hominids) as the grouping factor. To 91 rule out the presence of spurious groupings in the sample, we computed a cross-validated 92 bgPCA and compared the results to those of the bgPCA without cross-validation (Bookstein, 93 2019; Cardini et al., 2019; Cardini and Polly, 2020). We computed the Z scores and r2 for 94 group differences in the raw shape data (Adams and Collyer, 2016), and the scores of both the 95 non-cross-validated and the cross-validated bgPCAs using the ‘RRPP’ v. 2.5 R package 96 (Collyer and Adams, 2018; SOM S1). The fossils were left ungrouped and plotted a posteriori 97 onto the morphospace identified by the bgPCA based on extant taxa. To assess the affinities 98 of each fossil specimen with the a priori defined groups we computed the squared 99 Mahalanobis distances between each individual and the group means using the D2.dist 100 function of the ‘biotools’ v. 4.2 (Da Silva, 2020) R package, as well as their typicality 101 probabilities using the function typprobClass in ‘Morpho’ (Table 1; SOM S1). To visualize 102 shape changes occurring along the bgPC axes, we identified the extreme landmark 103 conformations for each bgPC and then warped the 3D model of the individual closest to the 104 mean configuration of the sample—identified with the function ‘FindMeanSpec’ within 105 ‘geomorph’ v. 3.1.1 R package (Adams et al., 2020)—toward the obtained configurations. We 106 finally computed an unweighted pair group methods with arithmetic mean (UPGMA) cluster 107 analysis (SOM S1). 108 To assess correlations between size and shape, ordinary least squares (OLS) and 109 phylogenetic generalized least squares (PGLS; Adams, 2014) regressions of bgPC scores vs. 110 log-transformed centroid size (CS; with natural logarithms, ln CS) were computed using the 111 ‘geomorph’ v. 3.1.1 R package (Adams et al., 2020). To compute the PGLS regressions, we 112 used a time-calibrated phylogenetic tree based on molecular data downloaded from 10kTrees 5 113 website v. 3 (Arnold et al., 2010; SOM S1). To evaluate the influence of phylogeny vs. 114 function on the proximal radial shape among extant anthropoids, phylogenetic signal was also 115 quantified by means of both Pagel’s λ and Blomberg’s K statistics (Pagel, 1999; Blomberg et 116 al., 2003; see SOM S1 for further details) using the ‘phytools’ v. 0.6-60 R package (Revell, 117 2012). 118 119 3. Results 120 The bgPCA (Fig. 3) discriminates among extant hominids, hylobatids, and monkeys and 121 correctly classifies 80.2% of cases in the five groups defined a priori (platyrrhines, 122 cercopithecines, colobines, hylobatids, and hominids), while the bgPCA with cross-validation 123 (SOM Fig. S1) correctly classifies 72.4% of cases (SOM S1; SOM Tables S4 and S5). 124 Misclassification cases occur mainly among monkeys, whereas hylobatids are correctly 125 classified in 76% of the cases and hominids in 86% (cross-validated bgPCA; SOM Table S5). 126 There is no perceptible change between the bgPCA (Fig. 3) and the cross-validated 127 bgPCA (SOM Fig. S1) plots, and the Z-scores are similar for the raw data (6.9), bgPCA (8.7), 128 and cross-validated bgPCA (8.7), implying a similar strength of morphological integration in 129 the datasets. Similarly, r2 increases from the raw data comparisons (0.26) to both standard (0.52) 130 and cross-validated (0.52) bgPCAs. This indicates that grouping structure is not spurious 131 because there is a comparable increase in r2 from the raw data to both the standard and cross- 132 validated bgPCAs. Only bgPC1 and bgPC2 (which account for 90% of the variance) are 133 discussed below because bgPC3 (7% variance) and bgPC4 (3% variance) yielded no 134 meaningful patterns. 135 The bgPC1 (78% of variance) embeds significant but low phylogenetic signal (K = 136 0.35, p = 0.027; λ = 0.54, p = 0.016), suggesting a considerable amount of homoplasy, best 137 illustrated by the overlap between Ateles and hominids—see SOM S1 for the different 138 implications of K and λ. Although bgPC1 is significantly correlated with ln CS (OLS: p < 6 139 0.001; PGLS: p = 0.008), size only accounts for a small amount of shape variation (OLS: r2 = 140 0.10 and adjusted r2 = 0.09; PGLS: r2 = 0.26 and adjusted r2 = 0.23). This axis discriminates 141 between hominoids (hominids and hylobatids, overlapping toward negative scores) and 142 cercopithecoids and platyrrhines (which mostly display positive scores, particularly 143 colobines)—except for Ateles, which mostly overlaps with hominoids (Fig. 3). 144 Nacholapithecus displays slightly positive scores within the hominid variation range, whereas 145 all the other fossils, including Ekembo and Equatorius, fall within the cercopithecoid– 146 platyrrhine distribution (albeit occupying different positions along bgPC1) apart from 147 hominoids. Simiolus displays the most positive scores and Ekembo is closer to hominoids (but 148 well distinct from Nacholapithecus) among the remaining fossil sample. Shape differences 149 along bgPC1 are driven by the shape of the radial head (Fig. 3) and the mediolateral tilting of 150 the head. Stem catarrhines and monkeys other than Ateles (more positive scores) display a 151 more elliptic radial head in proximal view and a medial elevation of the head in anterior view, 152 which results into a mediolateral tilting of the head. In contrast, extant hominoids and Ateles 153 (more negative scores) have rounder heads that are not tilted (Fig. 3). 154 bgPC2 (12% of variance) embeds no significant phylogenetic signal (K = 0.22, p = 155 0.312; λ < 0.001, p = 1.000) and is not correlated with ln CS (OLS: p = 0.083; PGLS: p = 156 0.274). bgPC2 distinguishes between hominids (positive and slightly negative scores) and 157 hylobatids (more negative scores) with some overlap, but both groups largely overlap with 158 cercopithecoids and platyrrhines, which are not distinguished along this axis (Fig. 3). All the 159 fossils analyzed cluster close to one another, with Nacholapithecus, Ekembo, and 160 Dendropithecus displaying scores closer to 0, and Simiolus and Equatorius displaying slightly 161 negative scores. Shape differences along bgPC2 are also driven by tilting of the head toward 162 more positive scores. More subtle differences include a more uniformly expanded distal 163 surface area of the radial head toward positive scores, whereas in specimens with more 7 164 negative scores the distally expanded surface area is more restricted to the anteromedial 165 aspect of the head. 166 Based on the typicality probabilities for the fossils (Table 1), Nacholapithecus is 167 classified as a platyrrhine as first option and as a hominid as second option. Ekembo, 168 Equatorius, and Dendropithecus are also classified as platyrrhines, whereas Epipliopithecus 169 and the two Simiolus specimens are classified as cercopithecines. The UPGMA analysis 170 clusters Nacholapithecus with a subcluster including all extant hominoids (SOM S2; SOM 171 Fig. S2). 172 173 4. Discussion and conclusions 174 We used a landmark protocol that characterizes the anterior aspect of the radial head to 175 analyze the single proximal radial fragment available for Nacholapithecus. The overall results 176 are similar to those previously found using a protocol with a more complete proximal radius 177 coverage (Arias-Martorell et al., 2021) but have more limited explanatory power at the 178 morphofunctional level. The shape of the anterior aspect of the radial head is only partially 179 explained by phylogeny, with bgPC2 showing no meaningful phylogenetic signal (as the 180 monkey variation encompasses that of extant hominoids) and bgPC1 displaying significant 181 but low values. The latter denote homoplasy, which might be explained by Ateles partially 182 overlapping with hominids due to its convergently evolved hominoid-like humeroradial joint 183 shape (Larson, 1998; Arias-Martorell et al., 2021). Similarities include a round radial head 184 and a more uniform surface area in the distal expansion of the radial head (not circumscribed 185 to the anteromedial side) than in other monkeys (Arias-Martorell et al., 2021). 186 Functional inferences for the elbow complex of Nacholapithecus have been based 187 mostly on the forelimb evidence from the holotype (KNM-BG 35250; Ishida et al., 2004; 188 Nakatsukasa and Kunimatsu, 2009; Takano et al., 2018). The modern hominoid-like globular 189 capitulum of the humerus (KNM-BG 35250M) indicates enhanced mobility at the 8 190 humeroradial joint and suggests enhanced forearm pronosupination capabilities because of the 191 inferred rounder radial head compared with cercopithecids and stem Miocene hominoids 192 (Takano et al., 2018). As indicated by our shape analysis, Nacholapithecus possesses a fairly 193 circular radial head outline with limited tilting and a distal articular surface uniformly 194 expanded to some extent beyond the anteromedial aspect of the radial head. On the 195 morphospace, KNM-BG 40021 displays an intermediate position between extant hominoids 196 and monkeys, close to hominids and overlapping with Ateles. This is further supported by the 197 fact that the analysis classifies Nacholapithecus as a platyrrhine (first option, owing to its 198 similarities to Ateles) or as a hominid (second option). 199 The anterior aspect of the radial head articulates with the zona conoidea of the 200 humerus in extant apes, whose humeroradial joint is stable throughout all ranges of 201 pronosupination and flexion–extension (Rose, 1988). In contrast, cercopithecoids, have a 202 stable elbow in fully pronated position, where the humeroantebrachial joint achieves a close- 203 packing position (Harrison 1987; Rose, 1988, 1993; Alba et al., 2011). The more extant 204 hominoid-like and atelid-like anterior aspect of the radial head of Nacholapithecus is 205 consistent with a radial head that is able to articulate with both the humerus and the ulna in 206 pronated and semipronated forearm positions due to incipient beveling of the radial head 207 beyond the lateral lip (Rose et al., 1992; Takano et al., 2018, 2020)—achieved, in 208 Nacholapithecus, by a rounder head and a more uniform distal expansion of the articular area 209 of the radial head than in cercopithecoids, nonsuspensory platyrrhines, and both earlier and 210 coeval Miocene apes. Relative to extant hominoids, Nacholapithecus displays a more 211 primitive humeroulnar joint that is not capable of full extension due to a long olecranon 212 process (Takano et al., 2020) and further retains primitive traits at the wrist joint (e.g., 213 ulnocarpal articulation; Ogihara et al., 2016). However, these plesiomorphic features are 214 combined with a humeroradial joint somewhat derived toward the extant hominoid condition 215 than that of early hominoids such as Ekembo and Equatorius (Takano et al., 2018, 2020). The 9 216 enhanced stability in wider ranges of pronosupination (especially in semipronated arm 217 positions) of the radiohumeral joint indicated by our results is concordant with previous 218 locomotor inferences for Nacholapithecus. Its locomotor repertoire has been described as 219 including forelimb-dominated arboreal behaviors with the forelimbs playing an important role 220 in both body support and overhead positions (Takano et al., 2018, 2020)—i.e., vertical 221 climbing, orthograde clambering, transferring, and bridging in higher frequencies than in 222 Early Miocene apes (e.g., Ekembo)—combined with powerful grasping abilities (Ishida et al., 223 2004; Nakatuskasa et al., 2002, 2007a, 2007b, 2012, 2016; Nakatsukasa and Kunimatsu, 224 2009; Alba et al., 2011; Ogihara et al., 2016; Ward, 2015; Takano et al., 2018, 2020). 225 Nevertheless, our results also suggest that the anterior aspect of the radius is not 226 sufficient to distinguish well among groups of quadrupedal taxa with radial heads most suited 227 to maintaining stability in a flexed-elbow and fully pronated hand posture (Rose, 1988)—as 228 illustrated by the considerable overlap between cercopithecines, colobines, and platyrrhines. 229 The fact that small-bodied stem catarrhines and the stem hominoids Ekembo and Equatorius 230 occupy the same region of the morphospace indicates similarities in the anterior aspect of the 231 radial head, especially in the distal expansion of its articular surface area. This is not 232 surprising given that a more distally expanded anteromedial articular surface area is part of 233 the ancestral anthropoid morphotype (Rose, 1988, 1993, 1994, 1997; Senut, 1989). In 234 previous analyses based on the proximal radius (Arias-Martorell et al., 2021), Ekembo and 235 Equatorius displayed a clearly intermediate morphology between extant hominoids and 236 monkeys, rather than closer affinities with monkeys. In our bgPCA plot, Ekembo is somewhat 237 closer to extant hominoids than Equatorius along bgPC1—consistent with differences in their 238 positional behavior (Ward, 1993, 2015; Ward et al., 1993, 1999; McCrossin et al., 1998; Patel 239 et al., 2009)—albeit less so than Nacholapithecus, whose humeroradial joint appears more 240 derived toward the crown hominoid condition (Takano et al., 2018, 2020). Small-bodied stem 241 catarrhines display the radial head morphology characteristic of non-hominoid anthropoids 10 242 and our results broadly agree with those previously obtained (Arias-Martorell et al., 2021) 243 albeit with lower resolution, especially for taxa with a high quadrupedal component. 244 The mosaic configuration of the elbow joint of Nacholapithecus, combining a 245 primitive humeroulnar joint with a quite derived humeroradial joint, supports a stepwise 246 evolution of the anthropoid elbow (Alba et al., 2011, 2015), with extant cercopithecoids and 247 hominoids displaying features derived in opposite directions and stem hominoids displaying 248 mosaic morphologies unlike those of living apes (Alba et al., 2011; Arias-Martorell et al., 249 2021). The uncertain phylogenetic relationships of Nacholapithecus (e.g., Almécija et al., 250 2021; Urciuoli et al., 2021) hinder to some extent the evolutionary implications of its 251 proximal radial morphology. Nevertheless, the proximal radial morphology of 252 Nacholapithecus, more derived than that of Ekembo but more primitive than that of crown 253 hominoids, likely reflects an enhancement of pronosupination movements associated with an 254 emphasis on orthograde positional behaviors—compatible with both the stem hominoid 255 (Pugh, 2022) and stem hominid (Alba, 2012; Kunimatsu et al., 2019; Morimoto et al., 2020) 256 status proposed for this taxon. Regardless of its systematic position, the elbow morphology of 257 Nacholapithecus suggests that the last common ancestor of crown hominoids displayed a 258 humeroradial joint more primitive than extant hominoids—indicating that the 259 humeroantebrachial complex would have evolved to some extent independently between 260 hylobatids and hominids. The latter notion would be consistent with the independent 261 evolution of orthogrady-related features in atelids and various lineages of crown hominoids 262 (Larson, 1998; Alba, 2012; Almécija et al., 2021) and is further reinforced by the retention of 263 primitive features (likely related to above-branch quadrupedalism) in the humeroulnar joint of 264 the Late Miocene great ape Hispanopithecus—recovered as a stem hominid by cladistic 265 analyses (Alba et al., 2015; Pugh, 2022)—despite indicating an elbow complex suitable for 266 preserving stability along the full range of flexion/extension and enabling a broad range of 267 pronosupination as in extant hominoids (Alba et al., 2012). 11 268 269 References 270 Adams, D.C., 2014. A method for assessing phylogenetic least squares models for shape and 271 other high-dimensional multivariate data. Evolution 68, 2675–2688. 272 Adams, D.C., Collyer, M.L. 2016. On the comparison of the strength of morphological 273 integration across morphometric datasets. Evolution 70, 2623–2631. 274 Adams, D.C., Collyer, M.L., Kaliontzopoulou, A., 2020. Geomorph: Software for geometric 275 morphometric analyses. R package version 3.2.1. https://cran.r- 276 project.org/package=geomorph. 277 Alba, D.M., 2012. Fossil apes from the Vallès-Penedès Basin. Evol. Anthropol. 21, 254–269. 278 Alba, D.M., Almécija, S., DeMiguel, D., Fortuny, J., Pérez de los Ríos, M., Pina, M., Robles, 279 J.M., Moyà-Solà, S., 2015. Miocene small-bodied ape from Eurasia sheds light on 280 hominoid evolution. Science 350, aab2625. 281 Almécija, S., Hammond, A.S., Thompson, N.E., Pugh, K.D., Moyà-Solà, S. Alba, D.M. 2021. 282 Fossil apes and human evolution. Science 372, eabb4363. 283 Almécija, S., Smaers, J.B. Jungers, W.L. 2015. The evolution of human and ape hand 284 proportions. Nat. Comm. 6, 1–11. 285 Arias-Martorell, J., Almécija, S., Urciuoli, A., Nakatsukasa, M., Moyà-Solà, S., Alba, D.M. 286 2021. A proximal radius of Barberapithecus huerzeleri from Castell de Barberà: 287 Implications for locomotor diversity among pliopithecoids. J. Hum. Evol. 157, 103032. 288 Arnold, C., Matthews, L.J., Nunn, C.L., 2010. The 10kTrees website: A new online resource 289 for primate phylogeny. Evol. Anthropol. 19, 114–118. 290 Blomberg, S.P., Garland, T., Ives, A.R., 2003. Testing for phylogenetic signal in comparative 291 data: behavioral traits are more labile. Evolution 57, 717–745. 292 Bookstein, F.L., 2019. Pathologies of between-groups principal components analysis in 293 geometric morphometrics. Evol. Biol. 46, 271–302. 12 294 Cardini, A., Polly, P.D., 2020. Cross-validated between group PCA scatterplots: A solution to 295 spurious group separation? Evol. Biol. 47, 85–95. 296 Cardini, A., O'Higgins, P., Rohlf, F.J., 2019. Seeing distinct groups where there are none: 297 Spurious patterns from between-group PCA. Evol. Biol. 46, 303–316. 298 Collyer, M.L., Adams, D.C. 2018. RRPP: An R package for fitting linear models to high- 299 dimensional data using residual randomization. Methods Ecol. Evol. 9, 1772–1779. 300 Da Silva, A.R., 2020. On testing for seed sample heterogeneity with the exact probability 301 distribution of the germination count range. Seed Sci. Res. 30, 59-63. 302 Harrison, T., 1987. The phylogenetic relationships of the early catarrhine primates: A review 303 of the current evidence. J. Hum. Evol. 16, 41–80. 304 Harrison, T., 2010. Dendropithecoidea, Proconsuloidea, and Hominoidea (Catarrhini, 305 Primates). In: Werdelin, L., Sanders, W.J. (Eds.), Cenozoic Mammals of Africa. University 306 of California Press, Berkeley, pp. 429–469. 307 Ishida, H., Kunimatsu, Y., Nakatsukasa, M., Nakano, Y. 1999. New hominoid genus from the 308 Middle Miocene of Nachola, Kenya. Anthropol. Sci. 107. 189–191 309 Ishida, H., Kunimatsu, Y., Takano, T., Nakano, Y., Nakatsukasa, M., 2004. Nacholapithecus 310 skeleton from the Middle Miocene of Kenya. J. Hum. Evol. 46, 69–103. 311 Kikuchi, Y., Nakano, Y., Nakatsukasa, M., Kunimatsu, Y., Shimizu, D., Ogihara, N., 312 Tsujikawa, H., Takano, T., Ishida, H., 2012. Functional morphology and anatomy of 313 cervical vertebrae in Nacholapithecus kerioi, a middle Miocene hominoid from Kenya. J. 314 Hum. Evol. 62, 677-695. 315 Kikuchi, Y., Nakatsukasa, M., Nakano, Y., Kunimatsu, Y., Shimizu, D., Ogihara, N., 316 Tsujikawa, H., Takano, T., Ishida, H., 2015. Morphology of the thoracolumbar spine of the 317 middle Miocene hominoid Nacholapithecus kerioi from northern Kenya. J. Hum. Evol. 88, 318 25-42. 319 Kikuchi, Y., Nakatsukasa, M., Nakano, Y., Kunimatsu, Y., Shimizu, D., Ogihara, N., 13 320 Tsujikawa, H., Takano, T., Ishida, H., 2016. Sacral vertebral remains of the Middle 321 Miocene hominoid Nacholapithecus kerioi from northern Kenya. J. Hum. Evol. 94, 117- 322 125. 323 Kunimatsu, Y., Nakatsukasa, M., Shimizu, D., Nakano, Y. and Ishida, H. 2019. Loss of the 324 subarcuate fossa and the phylogeny of Nacholapithecus. J. Hum. Evol. 131, 22–7. 325 Larson, S.G., 1998. Parallel evolution in the hominoid trunk and forelimb. Evol. Anthropol. 6, 326 87–99. 327 Le Gros Clark, W.E., Thomas, D.P., 1951. Associated jaws and limb bones of Limnopithecus 328 macinnesi. Fossil Mammals Afr. 3, 1–27. 329 McCrossin, L.M., Benefit, B.R., Giteu, S.N., Palmer, A.K., Blue, K.T., 1998. Fossil evidence 330 for the origins of terrestriality among Old World higher primates. In: Strasser, E., Fleagle, 331 J., Rosenberger, A., McHenry, H. (Eds.), Primate Locomotion: Recent Advances. Plenum 332 Press, New York, pp. 353–396. 333 McNulty, K.P., Begun, D.R., Kelley, J., Manthi, F.K., Mbua, E.N., 2015. A systematic 334 revision of Proconsul with the description of a new genus of early Miocene hominoid. J. 335 Hum. Evol. 84, 42–61. 336 Mitteroecker, P., Bookstein, F., 2011. Linear discrimination, ordination, and the visualization 337 of selection gradients in modern morphometrics. Evol. Biol. 38, 100–114. 338 Morimoto, N., Kunimatsu, Y., Nakatsukasa, M., Ponce de Leon, M.S., Zollikofer, C.P., 339 Ishida, H., Sasaki, T. Suwa, G. 2020. Variation of bony labyrinthine morphology in Mio− 340 Plio−Pleistocene and modern anthropoids. Am. J. Phys. Anthropol. 173, 276–292. 341 Nakatsukasa, M., Kunimatsu, Y. 2009. Nacholapithecus and its importance for understanding 342 hominoid evolution. Evol. Anthropol. 18, 103–119. 343 Nakatsukasa, M., Yamanaka, A., Kunimatsu, Y., Shimizu, D., Ishida, H. 1998. A newly 344 discovered Kenyapithecus skeleton and its implications for the evolution of positional 345 behavior in Miocene East African hominoids. J. Hum. Evol. 34, 657–664. 14 346 Nakatsukasa, M., Kunimatsu, Y., Nakano, Y., Ishida, H. 2002. Morphology of hallucial 347 phalanges in extant anthropoids and fossil hominoids. Z. Morphol. Anthropol. 83, 361– 348 372. 349 Nakatsukasa, M., Kunimatsu, Y., Nakano, Y., Takano, T., Ishida, H. 2003a. Comparative and 350 functional anatomy of phalanges in Nacholapithecus kerioi, a Middle Miocene hominoid 351 from northern Kenya. Primates 44, 371–412. 352 Nakatsukasa, M., Tsujikawa, H., Shimizu, D., Takano, T., Kunimatsu, Y., Nakano, Y., Ishida 353 H. 2003b. Definitive evidence for tail loss in Nacholapithecus, and East African Miocene 354 hominoid. J. Hum. Evol. 45, 179–186. 355 Nakatsukasa, M., Kunimatsu, Y., Nakano, Y., Ishida, H. 2007a. Vertebral morphology of 356 Nacholapithecus kerioi based on KNM-BG 35250. J. Hum. Evol. 52, 347–369. 357 Nakatsukasa, M., Kunimatsu, Y., Nakano, Y., Egi, N., Ishida, H. 2007b. Postcranial bones of 358 infant Nacholapithecus: ontogeny and positional behavioral adaptation. Anthropol. Sci. 359 115, 201-213. 360 Nakatsukasa, M., Kunimatsu, Y., Shimizu, D., Nakano, Y., Kikuchi, Y., Ishida, H. 2012. 361 Hind limb of the Nacholapithecus kerioi holotype and implications for its positional 362 behavior. Anthropol. Sci. 120, 12073. 363 Nakatsukasa, M., Almécija, S., Begun, D.R. 2016. The hands of Miocene hominoids. In: 364 Kivell, T.L., Lemelin, P., Richmond, B.G., Schmitt, D. (Eds.), The evolution of the primate 365 hand. Springer, New York, pp. 485–514. 366 Kunimatsu, Y., Nakatsukasa, M., Shimizu, D., Nakano, Y., Ishida, H., 2019. Loss of the 367 subarcuate fossa and the phylogeny of Nacholapithecus. J. Hum. Evol. 131, 22–27. 368 Ogihara, N., Almécija, S., Nakatsukasa, M., Nakano, Y., Kikuchi, Y., Kunimatsu, Y., 369 Makishima, H., Shimizu, D., Takano, T., Tsujikawa, H., Kagaya, M., Ishida, H. 2016. 370 Carpal bones of Nacholapithecus kerioi, a Middle Miocene hominoid from Northern 371 Kenya. Am. J. Phys. Anthropol. 160, 469–482. 15 372 Pagel, M., 1999. Inferring the historical patterns of biological evolution. Nature 401, 877– 373 884. 374 Patel, B.A., 2005. The hominoid proximal radius: Re-interpreting locomotor behaviors in 375 early hominins. J. Hum. Evol. 48, 415–432. 376 Pina, M., Kikuchi, Y., Nakatsukasa, M., Nakano, Y., Kunimatsu, Y., Ogihara, N., Shimizu, 377 D., Takano, T., Tsujikawa, H., Ishida, H., 2021. New femoral remains of Nacholapithecus 378 kerioi: Implications for intraspecific variation and Miocene hominoid evolution. J. Hum. 379 Evol. 155, 102982. 380 Pugh, K.D. 2022. Phylogenetic analysis of Middle-Late Miocene apes. J. Hum. Evol. 165, 381 103140. 382 R Core Team, 2021. R: A language and environment for statistical computing. R Foundation 383 for Statistical Computing, Vienna. 384 Revell, L.J., 2012. Phytools: An R package for phylogenetic comparative biology (and other 385 things). Methods Ecol. Evol. 3, 217–223. 386 Rose, M.D., 1988. Another look at the anthropoid elbow. J. Hum. Evol. 17, 193–224. 387 Rose, M.D., 1993. Locomotor anatomy of Miocene hominoids. In: Gebo, D.L. (Ed.), 388 Postcranial Adaptation in Nonhuman Primates. Northern Illinois University Press, 389 DeKalb, pp. 252–272. 390 Rose, M.D., 1994. Quadrupedalism in some Miocene catarrhines. J. Hum. Evol. 26, 387–411. 391 Rose, M.D., 1997. Functional and phylogenetic features of the forelimb in Miocene 392 hominoids. In: Begun, D.R., Ward, C.V., Rose, M.D. (Eds.), Function, Phylogeny and 393 Fossils: Miocene Hominoid Evolution and Adaptation. Plenum Press, New York, pp. 79– 394 100. 395 Rose, M.D., Leakey, M.G., Leakey, R.E.F., Walker, A.C., 1992. Postcranial specimens of 396 Simiolus enjiessi and other primitive catarrhines from the early Miocene of Lake Turkana, 397 Kenya. J. Hum. Evol. 22, 171–237. 16 398 Rose, M.D., Nakano, Y., Ishida, H. 1996. Kenyapithecus postcranial specimens from 399 Nachola, Kenya. Afr. Study Monogr. 24, 3–56. 400 Rossie, J.B., Gutierrez, M., Goble, E., 2012. Fossil forelimbs of Simiolus from Moruorot, 401 Kenya. Am. J. Phys. Anthropol. 147 (S54), 252. 402 Ruff, C.B., 2002. Long bone articular and diaphyseal structure in Old World monkeys and 403 apes. I: Locomotor effects. Am. J. Phys. Anthropol. 119, 305–342. 404 Ruff, C.B., 2003. Long bone articular and diaphyseal structure in Old World monkeys and 405 apes. II: Estimation of body mass. Am. J. Phys. Anthropol. 120, 16–37. 406 Sawada, Y., Pickford, M., Itaya, T., Makinouchi, T., Tateishi, M., Kabeto, K., Ishida, S., 407 Ishida, H., 1998. K-Ar ages of Miocene Hominoidea (Kenyapithecus and 408 Samburupithecus) from Samburu Hills, Northern Kenya. C. R. Acad. Sci. Paris 326, 445– 409 451. 410 Sawada, Y., Saneyoshi, M., Nakayama, K., Sakai, T., Itaya, T., Hyodo, M., Mukokya, Y., 411 Pickford, M., Senut, B., Tanaka, S., Chujo, T., Ishida, H. 2006. The ages and geological 412 background of Miocene hominoids Nacholapithecus, Samburupithecus, and Orrorin from 413 Kenya. In: H. Ishida, R. Tuttle, M. Pickford, N. Ogihara, M. Nakatsukasa (Eds.), Human 414 origins and environmental backgrounds. Developments in Primatology: Progress and 415 Prospects. Springer, New York, pp. 71-96 416 Schlager, S., 2017. Morpho and Rvcg – shape analysis in R: R-packages for geometric 417 morphometrics, shape analysis and surface manipulations. In: Zheng, G., Li, S., Székely, 418 G. (Eds.), Statistical Shape and Deformation Analysis. Methods, Implementation and 419 Applications. Academic Press, London, pp. 217–256. 420 Senut, B., 1989. Le Coude des Primates Hominoïdes. Anatomie, Fonction, Taxonomie, 421 Évolution. Éditions du Centre National de la Recerche Scientifique, Paris. 17 422 Senut, B., Nakatsukasa, M., Kunimatsu, Y., Nakano, Y., Takano, T., Tsujikawa, H., Shimizu, 423 D., Kagaya, M., Ishida, H. 2004. Preliminary analysis of Nacholapithecus scapula and 424 clavicle from Nachola, Kenya. Primates 45, 97-104. 425 Sherwood, R.J., Ward, R.J., Hill, A., Duren, D.L., Brown, B., Downs, W., 2002. Preliminary 426 description of the Equatorius africanus partial skeleton KNM-TH 28860 from 427 Kipsaramon, Tugen Hills, Baringo District, Kenya. J. Hum. Evol. 42, 63–73. 428 Takano, T., Nakatsukasa, M., Kunimatsu, Y., Nakano, N., Ogihara N., Ishida, H. 2018. 429 Forelimb long bones of Nacholapithecus (KNM-BG 35250) from the middle Miocene in 430 Nachola, northern Kenya. Anthropol. Sci. 126, 135–149. 431 Takano, T., Nakatsukasa, M., Pina, M., Kunimatsu, Y., Nakano, Y., Morimoto, N., Ogihara, 432 N., Ishida, H. 2020. New forelimb long bone specimens of Nacholapithecus kerioi from 433 the Middle Miocene of northern Kenya. Anthropol. Sci. 128, 200116. 434 Urciuoli, A., Zanolli, C., Almecija, S., Alba, D.M. 2020. Reassessment of the phylogenetic 435 relationships of the late Miocene apes Hispanopithecus and Rudapithecus based on 436 vertibular morphology. Proc. Natl. Acad. Sci. USA 118, e2015215118. 437 Walker, A.C., Pickford, M., 1983. New postcranial fossils of Proconsul africanus and 438 Proconsul nyanzae. In: Ciochon, R.L., Corruccini, R.S. (Eds.), New Interpretations of Ape 439 and Human Ancestry. Plenum Press, New York, pp. 325–351. 440 Ward, C.V., 1993. Torso morphology and locomotion in Proconsul nyanzae. Am. J. Phys. 441 Anthropol. 92, 291–328. 442 Ward, C.V., 2015. Postcranial and locomotor adaptations of hominoids. In: Henke, W., 443 Tattersall, I. (Eds.), Handbook of Paleoanthropology, 2nd ed. Springer, Heidelberg, pp. 444 1363–1386. 445 Ward, C.V., Walker, A., Teaford, M.F., Odhiambo, L., 1993. Partial skeleton of Proconsul 446 nyanzae from Mfangano Island, Kenya. Am. J. Phys. Anthropol. 90, 77–111. 447 Ward, S., Brown, B., Hill, A., Kelley, J., Downs, W., 1999. Equatorius: A new hominoid 18 448 genus from the middle Miocene of Kenya. Science 285, 1382–1386. 449 Wiley, D.F., Amenta, N., Alcantara, D.A., Ghosh, D., Kil, Y.J., Delson, E., Harcourt-Smith, 450 W., Rohlf, F.J., St John, K., Hamann, B., 2005. Evolutionary morphing. In: Silva, T.C., 451 Gorller, E., Rushmeier, H. (Eds.), VIS 05 IEEE Visualization. IEEE, Minneapolis, pp. 452 431–438. 453 Zapfe, H., 1958. The skeleton of Pliopithecus (Epipliopithecus) vindobonensis Zapfe and 454 Hürzeler. Am. J. Phys. Anthropol. 16, 441–457. 455 Zapfe, H., 1961. Die Primatenfunde aus der miozänen Spaltenfüllung von Neudorf an der 456 March (Děvínská Nová Ves), Tschechoslowakei. Schweizer. Palaeontol. Abh. 78, 1–293. 457 458 Figure captions 459 460 Figure 1. Right proximal radial fragment (KNM BG 40021) of Nacholapithecus kerioi from 461 the Aka Aiteputh Formation in Nachola, Kenya, in proximal (a), anterior (b), and medial (c) 462 views. 463 464 Figure 2. Landmark protocol illustrated on renderings of a 3D model of the right proximal 465 radial fragment (KNM-BG 40021, mirrored) of Nacholapithecus kerioi, in proximal (a), 466 anterior (b), lateral (c), and medial (d) views. Landmarks (L, bolded) and semilandmarks (SL) 467 are denoted by black and red dots, respectively, and described in SOM Table S3. 468 469 Figure 3. Results of the between-group principal component analysis as depicted by a 470 bivariate plot of bgPC2 vs. bgPC1. Groups distinguished a priori are denoted by color coded 471 convex hulls and symbols: violet = hominids; green = hylobatids; orange = cercopithecines; 472 emerald = colobines; pink = platyrrhines. Extant genera are denoted by different symbols (see 473 legend). The scatter of Ateles is highlighted with a convex hull in darker pink within the 19 474 platyrrhine distribution. Fossil specimens (scores projected a posteriori) are denoted by 475 colored stars (see legend). The percentage of variance explained by each bgPC is reported 476 within parentheses. Renderings along axes represent maximum and minimum shape changes 477 for that axis (corresponding to their position at the positive and negative ends of each axis). 478 Abbreviation: bgPC = between-group principal component. 20 Ekembo heseloni (KNM-RU 2036AI) Simiolus enjiessi (KNM-MO 63) Simiolus enjiessi (KNM-MO 17022B) Epipliopithecus vindobonensis (OE 304PCe) Dendropithecus macinnesi (KNM-RU 2098) Gorilla Equatorius africanus (KNM-TH 28860J) Pongo Nacholapithecus kerioi (KNM-BG 40021) Pan Homo Nomascus Hylobates Symphalangus 0.05 Hoolock Mandrillus bgPC2 (12% variance) Macaca Papio Colobus Nasalis 0.00 Alouatta Cebus Ateles Lagothrix -0.05 -0.1 0.0 0.1 bgPC1 (78% variance) Table 1 Squared Mahalanobis distances (D2) between fossils and group means, and classification results based on typicality probabilities (p) of the fossils computed from the between-group principal component analysis of radial head shape.a Species Catalogue No. D2/p Cercopithecines Colobines Hominids Hylobatids Platyrrhines Nacholapithecus kerioi KNM-BG 40021 D2 9.892 8.325 7.317 6.062 1.782 p 0.014 0.010 0.217* 0.015 0.486** Ekembo heseloni KNM-RU 2036CE D2 3.878 5.265 4.770 11.891 0.603 p 0.529* 0.107 0.062 0.002 0.983** Equatorius africanus KNM-TH 28860-J D2 4.002 4.385 11.520 7.729 2.919 p 0.689* 0.155 0.007 0.001 0.827** Epipliopithecus vindobonensis O.E. 304 PCe D2 7.576 17.238 10.630 16.416 5.385 p 0.127** <0.001 <0.001 <0.001 0.056* Dendropithecus macinnesi KNM-RU 2098 D2 15.063 8.5177 12.022 13.748 1.964 p 0.031 0.118* 0.004 <0.001 0.874** Simiolus enjiessi KNM-MO 63 D2 12.904 12.547 13.622 14.547 3.052 p 0.528** 0.011 <0.001 <0.001 0.233* Simiolus enjiessi KNM-MO 17022B D2 5.533 11.160 16.496 11.045 6.580 p 0.304** 0.003 <0.001 <0.001 0.088* a Group membership was rejected at p < 0.05. Two asterisks (**) denote primary group classification; one asterisk (*) denotes secondary group classification. Shortest D2 and highest p-values are bolded. Supplementary Online Material (SOM): The radial head of the Middle Miocene ape Nacholapithecus kerioi: Morphometric affinities, locomotor inferences, and implications for the evolution of the hominoid humeroradial joint SOM S1 Supplementary materials and methods Scanning The radii of Simiolus enjiessi KNM-RU 17022B and of Epipliopithecus vindobonensis O.E. 304 PCe were scanned from high-quality casts housed in the Institut Català de Paleontologia Miquel Crusafont (ICP, Cerdanyola del Vallès, Spain). All the other fossils (see SOM Table S1) were scanned from original specimens at the Kenya National Museums with a NextEngine surface laser scanner (Next Engine, Santa Monica) using the HD3 macro mode. The 3D models of the extant specimens were obtained using a NextEngine surface scanner and two different high-resolution µCT scanners (SOM Table S2): a BIR ACTIS 225/300 industrial µCT scanner (Department of Human Evolution, Max Planck Institute for Evolutionary Anthropology, Germany) and a Nikon XT 225 ST µCT scanner (Cambridge Biotomography Centre, Department of Zoology, University of Cambridge, UK). Specimens scanned with the NextEngine scanner were obtained using a resolution of >10,000 points per square inch; 6–12 scans were taken at two or more positions and then merged using ScanStudio HD PRO software v. 1.3.2 (Next Engine, Santa Monica). The isotropic voxel size range for the µCT scans sample is 21.9–51.5 µm. Laser scan-derived 3D models were cleaned (fill holes, irregularities in mesh, etc.) using Geomagic Wrap 2017 (3D Systems, Inc. Morrisville), and µCT scans were processed in AVIZO v. 6.3 (Visualization Sciences Group, Berlin). Landmark protocol Our landmark protocol was specifically devised to capture the shape affinities of KNM-BG 40021. It consists of 15 3D landmarks including 9 type II and 6 semilandmarks (Fig. 2; SOM Table S3). Type II landmarks reflect points in anatomical structures that can be recognized by their geometry, such as the maximum point of a curve. The homology of type III landmarks is given relative to the other landmarks around them, which should be type I or type II (therefore, a type III landmark would be the middle point between two 'true' landmarks, for example; Bookstein, 1997; O’Higgins, 2000). The protocol thus captures the anterior aspect of the radial head, including the anterior outline of the fovea (L1–3, SL1–2), the anterior outline of the radial head (L4–6, SL3–4), and the distal expansion of articular surface (L7–9, SL5–6). Radii were oriented in anterior view, either using anatomical orientation (complete radii) or based on the anteromedial position of the radial tuberosity, which was preserved in all individuals, allowing a swift identification of the medial, lateral, and posterior aspects of the radial head (as type II landmarks depend on correct orientation for placement; Zelditch et al., 2012). All type II 2 landmarks (L1–9) were used in a previous study of the proximal radius of fossil catarrhines (Arias-Martorell et al., 2021). We included six additional semilandmarks to ensure a more detailed and accurate representation of the anterior aspect of the shape of the radial head (Bookstein, 1997). Shape changes occurring along the axes of the bgPCA depicted in Figure 3 were visualized by warping the specimen closest to the mean configuration—identified with the ‘findMeanSpec’ function in Geomorph (Adams et al., 2020)—toward the extreme landmark configurations for each bgPC. Statistics We computed the standard deviates of observed statistics as effect sizes from distributions of random outcomes. We used these to compare the strength of morphological integration across morphometric datasets using the statistic test (Z-score) under the null hypothesis of a random association of variables. The method displays a constant expected value and confidence intervals and thus provides a consistent measure of integration suitable for comparisons across datasets (Adams and Collyer, 2016). To assess the similarity of a fossil specimen’s score to each a priori defined group based on their distribution of scores (variability) we used typicality probabilities. These are computed based on the Mahalanobis square distance (D2) between the specimen and the group centroids and represent the p-value to test the null hypothesis of group membership. Hence, a specimen is considered an outlier for a given group when p < 0.05, while higher typicality probabilities denote closer affinities between the individual and the distribution of the group. Note that the sum of the typicality probabilities for a given specimen does not equal 1, as they do not assume that the specimen must belong to one of the groups defined a priori, and rather represent the likelihood of belonging to each a priori defined group separately. In the UPGMA analysis, we used the mean Procrustes coordinates of each group, of the two Simiolus specimens, and the other fossils. To compute the amount of phylogenetic signal embedded in the shape data we computed both Pagel’s λ (Pagel, 1999; Freckleton et al., 2002) and Bloomberg’s K (Blomberg et al., 2003) statistics. Both compare the observed data distribution to that expected under a Brownian motion model of evolution but they are not entirely comparable: λ compares the actual covariance among species with that expected under Brownian motion, whereas K more specifically reflects how variance is partitioned. Pagel’s λ ranges from 0 to 1: λ = 1 implies that trait covariance is exclusively influenced by phylogeny, λ < 1 suggests that other factors besides phylogeny influence trait evolution, and λ = 0 is obtained when no phylogenetic 3 correlation is found in the data. In contrast, Blomberg’s K may vary beyond unity: K ≈ 1 similarly implies a model of evolution that closely resembles that expected under Brownian motion, K < 1 implies that closely related taxa resemble each other less than expected (variance accumulates within clades), possibly because of independent evolution (i.e., homoplasy), and K > 1 implies that closely related taxa are more similar than expected, so that variance accumulates among clades (as the result of stabilizing selection or architectural constraints). SOM S2 Supplementary results Shape changes were warped onto the specimen of Pan troglodytes USNM-220062, which was identified to be the closest to the mean configuration with a Mahalanobis distance of 0.068. In the UPGMA analysis, all the monkeys clustered together, including a subcluster for Lagothrix, Simiolus, and Epipliopithecus, apart from the hominoid + Nacholapithecus cluster. Ekembo did not cluster with any other taxon but is at the base of the monkey + small fossil catarrhines cluster, whereas Equatorius and Dendropitheus clustered apart from all the other taxa. 4 SOM Figure S1. Results of the cross-validated between-group principal component (bgPC) analysis of proximal radial shape among extant anthropoid primates as depicted by a bivariate plot of bgPC2 vs. bgPC1. Groups distinguished a priori are denoted by color-coded convex hulls and symbols: violet = hominids; green = hylobatids; orange = cercopithecines; emerald = colobines; pink = platyrrhines. The scatter of Ateles is highlighted with a convex hull in darker pink within the platyrrhine distribution. Extant genera are denoted by different symbols (see legend). Fossil specimens (scores projected a posteriori) are denoted by colored stars (see legend). 5 Simiolus enjiessi Epipliopithecus vindobonensis Lagothrix Mandrillus Ateles Papio Macaca Alouatta Colobus Nasalis Cebus Ekembo heseloni Hylobatids Gorilla Pan Pongo Homo Nacholapithecus kerioi Dendropithecus macinnesi Equatorius africanus SOM Figure S2. Dendrogram derived from the unweighted pair group methods with arithmetic mean (UPGMA) cluster analysis. Cophenetic correlation was r2 = 0.84. 6 Cebus apella Lagothrix lagotricha Ateles fusciceps Alouatta seniculus Alouatta palliata Pongo pygmaeus Pongo abelii Pan troglodytes Pan paniscus Homo sapiens Gorilla gorilla Gorilla beringei Nomascus gabriellae Symphalangus syndactylus Hylobates lar Hylobates klossii Hylobates agilis Hoolock hoolock Nasalis larvatus Colobus guereza Macaca mulatta Macaca fascicularis Macaca arctoides Papio hamadryas Papio anubis Mandrillus sphinx 50 40 30 20 10 0 SOM Figure S3. Consensus molecular tree downloaded from 10kTrees website v. 3 (www.10Ktrees.org) used to compute phylogenetic signal and conduct the phylogenetic generalized least squares analyses. 7 SOM Table S1 Details of the fossil comparative sample. Species Catalogue no. Site References Simiolus enjiessi KNM-MO 63 Moruorot Rose et al. (1992: Fig. 8); Senut (1989: Fig. 62); Rossie et al. (2012) Simiolus enjiessi KNM-MO 17022B Kalodirr Rose et al. (1992: Fig. 9) Le Gros Clark and Thomas (1951: Pl. 4 Fig. 9, Pl. 5 Fig. 11); Senut (1989: Dendropithecus macinnesi KNM-RU 2098 Rusinga Island Fig. 76 and Pl. X) Zapfe (1958: Pl. 1A and Fig. B5); Zapfe (1961: Fig. 54); Senut (1989: Fig. Epipliopithecus vindobonensis O.E. 304 PCe Devínska Nová Ves 95 and Pl. XV) Ekembo heseloni KNM-RU 2036CE Rusinga Island Walker and Pickford (1983: Fig. 4); Senut, 1989 (Fig. 74 and Pl. VII) Equatorius africanus KNM-TH 28860-J Kipsaraman Ward et al. (1999: Fig 2k); Sherwood et al. (2002: Fig. 1f) SOM Table S2 Details of the extant primate sample used in the study. Media and identifier (when available) are listed for specimens downloaded from Morphosource.org. Species Catalog no. Sex Side Source Media Identifier Scanner Alouatta palliata aequatorialis USNM 338107 M L USNM — — NextEngine Alouatta palliata palliata USNM 282798 F L USNM — — NextEngine Alouatta seniculus AMNH 42316 F L AMNH — — NextEngine Alouatta seniculus AMNH 23333 M L AMNH — — NextEngine Alouatta sp. ZMB 35764 M L ZMB — — µCT (BIR ACTIS) Alouatta sp. ZMS 1973-0330 ? R ZMS — — µCT (BIR ACTIS) Ateles fusciceps robustus USNM 338111 F L USNM — — NextEngine Ateles fusciceps robustus USNM 338112 M L USNM — — NextEngine Ateles sp. ZMB 45255 F R ZMB — — µCT (BIR ACTIS) Ateles sp. ZMB 44814 M R ZMB — — µCT (BIR ACTIS) Ateles sp. ZMB 38734 ? R ZMB — — µCT (BIR ACTIS) Ateles sp. ZMB 44079 ? L ZMB — — µCT (BIR ACTIS) Cebus apella apella USNM 361020 M L USNM — — NextEngine Cebus apella USNM 397940 F R USNM — — NextEngine Cebus apella AMNH 133606 M L AMNH — — NextEngine urn:catalog:AMNH:Mammals:M- Cebus apella paraguanayanus AMNH 133631 M R MS M12099-19605 — 133631 urn:catalog:AMNH:Mammals:M- Cebus apella paraguayanus AMNH 133623 M R MS M12095-19594 — 133623 urn:catalog:AMNH:Mammals:M- Cebus apella paraguayanus AMNH 133628 M R MS M12093-19588 — 133628 9 Colobus guereza AMNH 52223 F L AMNH — — NextEngine Colobus guereza AMNH 52241 F L AMNH — — NextEngine Colobus guereza kikuyuensis USNM 452621 M L USNM — — NextEngine Colobus guereza AMNH 52248 M L AMNH — — NextEngine Colobus guereza USNM 452632 F L USNM — — NextEngine Gorilla beringei beringei AMNH 54091 F L AMNH — — NextEngine Gorilla beringei beringei RMCA 2263 F L RMCA — — NextEngine Gorilla beringei beringei USNM 395636 M L USNM — — NextEngine http://n2t.net/ark:/65665/313444cf4- Gorilla beringei beringei USNM 396934 M L Morphosource M56720-102006 — f1e7-4bbc-ba69-039e4d4557e4 http://n2t.net/ark:/65665/32f41b8f5- Gorilla beringei beringei USNM 396937 F L Morphosource M57009-102295 — 9a15-4f88-af7e-8218ebf0b616 http://n2t.net/ark:/65665/3db306794- Gorilla beringei beringei USNM 397351 M L Morphosource M56268-101554 — 3c8e-4930-bb20-e514ac62bac6 Gorilla beringei graueri AMNH 202932 M R AMNH — — NextEngine Gorilla beringei graueri RMCA 8187 M L RMCA — — NextEngine Gorilla beringei USNM 239883 M L USNM — — NextEngine Gorilla gorilla USNM 586541 F R USNM — — NextEngine Gorilla gorilla gorilla AMNH 1673390 F L AMNH — — NextEngine Gorilla gorilla CMNH 2767 M L CMNH — — NextEngine Gorilla gorilla USNM 174722 M R USNM — — NextEngine Gorilla gorilla USNM 176225 M L USNM — — NextEngine Gorilla gorilla MER 300 F R PCM — — µCT (Nikon) Homo sapiens AMNH 99-8376 F L AMNH — — NextEngine Homo sapiens USNM 1512 F L USNM — — NextEngine Homo sapiens AMNH 20-3501 M L AMNH — — NextEngine 10 Homo sapiens USNM 942 M L USNM — — NextEngine Homo sapiens PSU 105-1793 ? L MS M45359-82651 — — Hoolock hoolock AMNH 83425 F R AMNH — — NextEngine Hoolock hoolock AMNH 83420 M R AMNH — — NextEngine Hylobates agilis AMNH 106575 F L AMNH — — NextEngine Hylobates klossii AMNH 103344 M R AMNH — — NextEngine Hylobates klossii AMNH 103347 M L AMNH — — NextEngine Hylobates lar vestitus NMNH 271047 F L USNM — — NextEngine Lagothrix lagotricha DU-BAA 90 ? R MS M12471-20497 ark:/87602/m4/M20497 — Macaca arctoides AMNH 112727 F L AMNH — — NextEngine Macaca fascicularis USNM 271168 M R USNM — — NextEngine Macaca fascicularis ZMB 48496 ? L ZMB — — µCT (BIR ACTIS) Macaca fascicularis ZMB 49090 ? L ZMB — — µCT (BIR ACTIS) Macaca fascicularis ZMB 49092 ? L ZMB — — µCT (BIR ACTIS) Macaca mulatta DU-BAA 142 ? R MS M12472-20500 — — Macaca mulatta USNM 537241 F L USNM — — NextEngine Macaca mulatta USNM 537253 M L USNM — — NextEngine urn:catalog:AMNH:Mammals:M- Mandrillus sphinx AMNH 89361 M R MS M10169-14599 — 89361 urn:catalog:AMNH:Mammals:M- Mandrillus sphinx AMNH 89365 M R MS M10176-14633 — 89365 Nasalis larvatus USNM 536050 F L USNM — — NextEngine Nasalis larvatus AMNH 106275 M L AMNH — — NextEngine Nasalis larvatus AMNH 198276 M L USNM — — NextEngine Nomascus gabriellae AMNH 87253 F L AMNH — — NextEngine 11 Pan paniscus AMNH 86857 F L AMNH — — NextEngine Pan paniscus RMCA 29045 F L RMCA — — NextEngine Pan paniscus RMCA 27696 M R RMCA — — NextEngine Pan paniscus SBU 87-1 M L SBU — — NextEngine Pan troglodytes USNM 176226 F L USNM — — NextEngine Pan troglodytes USNM 176229 F L USNM — — NextEngine Pan troglodytes USNM 176227 M L USNM — — NextEngine Pan troglodytes USNM 220327 M L USNM — — NextEngine Pan troglodytes USNM 395820 M L USNM — — NextEngine Pan troglodytes UNSM 481804 M R USNM — — NextEngine urn:catalog:AMNH:Mammals:M- Pan troglodytes schweinfurthii AMNH 51376 M R MS M10175-14630 — 51376 urn:catalog:AMNH:Mammals:M- Pan troglodytes schweinfurthii AMNH 51393 M R MS M10242-14814 — 51393 urn:catalog:AMNH:Mammals:M- Pan troglodytes troglodytes AMNH 54330 M L MS M10240-14808 — 54330 Pan troglodytes troglodytes USNM 220064 F L USNM — — NextEngine http://n2t.net/ark:/65665/3dcfb7753- Pan troglodytes troglodytes USNM 220062 F L MS M56889-102175 — f4d7-4334-9b52-6f9f1b9ea03e http://n2t.net/ark:/65665/386ed1f25- Pan troglodytes troglodytes USNM 220063 F L MS M56483-101769 — 2f34-459d-91e5-d0111c2e0dc6 Pan troglodytes verus MPI-EVA 11778 F L MPI — — µCT (BIR ACTIS) Pan troglodytes verus MPI-EVA 13429 F L MPI — — µCT (BIR ACTIS) Pan troglodytes verus MPI-EVA 15001 F L MPI — — µCT (BIR ACTIS) Pan troglodytes verus AMNH 89406 M L AMNH — — NextEngine Papio anubis AMNH 52668 F L AMNH — — NextEngine 12 Papio anubis AMNH 120388 M L AMNH — — NextEngine Papio anubis neumanni USNM 384235 F L USNM — — NextEngine Papio anubis neumanni USNM 384229 M L USNM — — NextEngine Papio hamadryas ZMB 105450 M R ZMB — — µCT (BIR ACTIS) Papio hamadryas ZMB 65265 M L ZMB — — µCT (BIR ACTIS) Pongo abelii USNM 588109 F L USNM — — NextEngine Pongo abelii UNSM 143588 M L USNM — — NextEngine http://n2t.net/ark:/65665/33bd6f2f4- Pongo abelii USNM 143587 M L MS M56592-101878 — 8b1a-4ffd-966f-06506fd24428 http://n2t.net/ark:/65665/389dc210f- Pongo abelii USNM 143590 M L MS M56324-101610 — f5b3-4910-ae87-a26700227801 http://n2t.net/ark:/65665/329ae2628- Pongo abelii USNM 143593 M L MS M56494-101780 — 4c93-4da7-8e52-5f0c1e7bcc9e http://n2t.net/ark:/65665/3a893123e- Pongo abelii USNM 143594 M L MS M56426-101712 — 021c-4f9b-ab42-4b4050332c24 http://n2t.net/ark:/65665/3c26ea641- Pongo abelii USNM 143596 F L MS M56423-101709 — 6662-42df-9b0d-a288ade0d69c Pongo pygmaeus AMNH 200900 F L AMNH — — NextEngine Pongo pygmaeus USNM 142169 F L USNM — — NextEngine Pongo pygmaeus USNM 145302 F L USNM — — NextEngine Pongo pygmaeus USNM 153805 F R USNM — — NextEngine Pongo pygmaeus USNM 153822 F L USNM — — NextEngine Pongo pygmaeus ZMS 1982-0092 F R ZMS — — µCT (BIR ACTIS) Pongo pygmaeus USNM 145301 M L USNM — — NextEngine Pongo pygmaeus USNM 145305 M L USNM — — NextEngine Pongo pygmaeus USNM 153823 M L USNM — — NextEngine 13 Pongo pygmaeus ZMS 1909-0801 M L ZMS — — µCT (BIR ACTIS) Pongo pygmaeus ZMS 1966-0203 M R ZMS — — µCT (BIR ACTIS) Pongo pygmaeus ZMB 87092 ? L ZMB — — µCT (BIR ACTIS) Symphalangus syndactylus AMNH 106583 F L AMNH — — NextEngine Symphalangus syndactylus NMNH 271048 F L USNM — — NextEngine Symphalangus syndactylus AMNH 106581 M L AMNH — — NextEngine Symphalangus syndactylus PSU 105-1841 ? L MS M45351-82643 — — Symphalangus syndactylus UWBM 58721-1 ? R MS M69298-125011 — — Symphalangus syndactylus UWBM 82801-1 ? L MS M69299-125019 — — Symphalangus syndactylus ZMB 38573 ? R ZMB — — µCT (BIR ACTIS) Symphalangus syndactylus ZMB 38587 F L ZMB — — µCT (BIR ACTIS) Abbreviations: F = female; M = male; ? = unknown sex; L = left; R = right; AMNH = American Museum of Natural History, New York, USA; CMNH = Cleveland Museum of Natural History, Cleveland, USA; MPI-EVA = Max Planck Institute for Evolutionary Anthropology, Leipzig, Germany; MS = MorphoSource.org; PCM = Powell-Cotton Museum, Birchington, UK; RMCA = Royal Museum for Central Africa, Tervuren, Belgium; SBU = Stony Brook University, New York, USA; USNM = Smithsonian National Museum of Natural History, Washington D.C., USA; ZMB = Museum für Naturkunde – Leibniz Institute for Evolution and Biodiversity Science, Berlin, Germany; ZMS = Zoologische Staatssammlung Munchen, Munich, Germany. 14 SOM Table S3 Landmark (L) and semilandmark (SL) protocol for KNM-BG 40021.a L/SL no. Description Fovea capitis L1 Most medial point on fovea capitis outline L2 Most anterior point on fovea capitis outline L3 Most lateral point on fovea capitis outline SL1 Midpoint between L1 and L2 on fovea outline SL2 Midpoint between L2 and L3 on fovea outline Radial head L4 Most medial point on radial head outline L5 Most anterior point on radial head outline L6 Most lateral point on radial head outline SL3 Midpoint between L4 and L5 on radial head outline SL4 Midpoint between L5 and L6 on radial head outline L7 Most medial point on distal articular expansion of the radial head L8 Most anterior point on distal articular expansion of the radial head L9 Most lateral point on distal articular expansion of the radial head SL5 Midpoint between L7 and L8 on distal articular expansion of the radial head SL6 Midpoint between L8 and L9 on distal articular expansion of the radial head a True landmarks (type II) and semilandmarks follow the descriptions of Bookstein (1997) and O’Higgins (2000). 15 SOM Table S4 Number of correctly classified specimens (and percentages within parentheses) by the bgPCA for extant taxa without cross-validation. Taxon Cercopithecines Colobines Hominids Hylobatids Platyrrhines Cercopithecines 12 (75.0%) 0 (0%) 1 (6.3%) 0 (0%) 3 (18.8%) Colobines 0 (0%) 6 (75.0%) 0 (0%) 0 (0%) 2 (25.0%) Hominids 0 (0%) 0 (0%) 48 (81.4%) 9 (15.2%) 2 (3.4%) Hylobatids 0 (0%) 0 (0%) 0 (0%) 14 (100%) 0 (0%) Platyrrhines 1 (5.3%) 3 (15.8%) 2 (10.5%) 0 (0%) 13 (68.4%) bgPCA = between-group principal component analysis. 16 SOM Table S5 Number of correctly classified specimens (and percentages within parentheses) by the bgPCA for extant taxa with cross-validation. Taxon Cercopithecines Colobines Hominids Hylobatids Platyrrhines Cercopithecines 11 (68.7%) 1 (6.2%) 1 (6.25%) 0 (0%) 3 (18.7%) Colobines 0 (0%) 6 (75,0%) 0 (0%) 0 (0%) 2 (25,0%) Hominids 0 (0%) 0 (0%) 45 (76.3%) 11 (18.6%) 3 (5.1%) Hylobatids 0 (0%) 0 (0%) 2 (14.3%) 12 (85.7%) 0 (0%) Platyrrhines 2 (10.5%) 4 (21.0%) 3 (15.8%) 0 (0%) 10 (52.6%) bgPCA = between-group principal component analysis. 17 SOM References Adams, D.C., Collyer, M.L. 2016. On the comparison of the strength of morphological integration across morphometric datasets. Evolution 70, 2623–2631. Adams, D.C., Collyer, M.L., Kaliontzopoulou, A., 2020. Geomorph: Software for geometric morphometric analyses. R package version 3.2.1. https://cran.r- project.org/package=geomorph. Arias-Martorell, J., Almécija, S., Urciuoli, A., Nakatsukasa, M., Moyà-Solà, S., Alba, D.M., 2021. A proximal radius of Barberapithecus huerzeleri from Castell de Barberà: Implications for locomotor diversity among pliopithecoids. J. Hum. Evol. 157, 103032. Blomberg, S.P., Garland, T., Ives, A.R., 2003. Testing for phylogenetic signal in comparative data: behavioral traits are more labile. Evolution 57, 717–745. Bookstein, F.L., 1997. Landmark methods for forms without landmarks: Localizing group differences in outline shape. Med. Imaging Anal. 1, 225–243. Freckleton, R.P., Harvey, P.H., Pagel, M., 2002. Phylogenetic analysis and comparative data: A test and review of evidence. Am. Nat. 160, 712–726. Le Gros Clark, W.E., Thomas, D.P., 1951. Associated jaws and limb bones of Limnopithecus macinnesi. Fossil Mammals Afr. 3, 1–27. O’Higgins, P., 2000. The study of morphological variation in the hominid fossil record: biology, landmarks and geometry. J. Anat. 197, 103–120. Pagel, M., 1999. Inferring the historical patterns of biological evolution. Nature 401, 877– 884. Rose, M.D., Leakey, M.G., Leakey, R.E.F., Walker, A.C., 1992. Postcranial specimens of Simiolus enjiessi and other primitive catarrhines from the early Miocene of Lake Turkana, Kenya. J. Hum. Evol. 22, 171–237. Rossie, J.B., Gutierrez, M., Goble, E., 2012. Fossil forelimbs of Simiolus from Moruorot, Kenya. Am. J. Phys. Anthropol. 147 (S54), 252. Senut, B., 1989. Le Coude des Primates Hominoïdes. Anatomie, Fonction, Taxonomie, Évolution. Éditions du Centre National de la Recerche Scientifique, Paris. Sherwood, R.J., Ward, R.J., Hill, A., Duren, D.L., Brown, B., Downs, W., 2002. Preliminary description of the Equatorius africanus partial skeleton KNM-TH 28860 from Kipsaramon, Tugen Hills, Baringo District, Kenya. J. Hum. Evol. 42, 63–73. Walker, A.C., Pickford, M., 1983. New postcranial fossils of Proconsul africanus and Proconsul nyanzae. In: Ciochon, R.L., Corruccini, R.S. (Eds.), New Interpretations of Ape and Human Ancestry. Plenum Press, New York, pp. 325–351. 18 Ward, S., Brown, B., Hill, A., Kelley, J., Downs, W., 1999. Equatorius: A new hominoid genus from the middle Miocene of Kenya. Science 285, 1382–1386. Zapfe, H., 1958. The skeleton of Pliopithecus (Epipliopithecus) vindobonensis Zapfe and Hürzeler. Am. J. Phys. Anthropol. 16, 441–457. Zapfe, H., 1961. Die Primatenfunde aus der miozänen Spaltenfüllung von Neudorf an der March (Děvínská Nová Ves), Tschechoslowakei. Schweizer. Palaeontol. Abh. 78, 1–293. Zelditch, M. L., Swiderski, D. L., Sheets, H. D., 2012. Geometric morphometrics for biologists: A primer. New York, Academic Press. 19