The upper limb of Paranthropus boisei from Ileret, Kenya

Journal of Human Evolution 141 (2020) 102727 Contents lists available at ScienceDirect Journal of Human Evolution journal homepage: www.elsevier.com/locate/jhevol The upper limb of Paranthropus boisei from Ileret, Kenya B.G. Richmond a, b, *, D.J. Green c, d, e, f, *, M.R. Lague g, H. Chirchir h, i, A.K. Behrensmeyer j, R. Bobe k, l, M.K. Bamford e, N.L. Griffin m, P. Gunz b, E. Mbua n, S.R. Merritt o, B. Pobiner i, P. Kiura p, M. Kibunjia q, J.W.K. Harris r, D.R. Braun f a Division of Anthropology, American Museum of Natural History, USA Department of Human Evolution, Max Planck Institute for Evolutionary Anthropology, Germany c Department of Anatomy, Campbell University School of Osteopathic Medicine, USA d Department of Anatomy, Midwestern University, USA e Evolutionary Studies Institute, University of the Witwatersrand, South Africa f Center for the Advanced Study of Human Paleobiology, The George Washington University, USA g Biology Program, Stockton University, USA h Department of Biological Sciences, Marshall University, USA i Human Origins Program, National Museum of Natural History, Smithsonian Institution, USA j Department of Paleobiology, National Museum of Natural History, Smithsonian Institution, USA k University of Oxford, Oxford, UK l Gorongosa National Park, Sofala, Mozambique m Department of Anatomy and Cell Biology, Temple University, USA n Department of Earth Sciences, National Museums of Kenya, Kenya o Department of Anthropology, University of Alabama at Birmingham, USA p Sites & Monuments, National Museums of Kenya, Kenya q National Museums of Kenya, Kenya r Department of Anthropology, Rutgers University, USA b a r t i c l e i n f o a b s t r a c t Article history: Received 14 October 2017 Accepted 17 December 2019 Available online xxx Paranthropus boisei was first described in 1959 based on fossils from the Olduvai Gorge and now includes many fossils from Ethiopia to Malawi. Knowledge about its postcranial anatomy has remained elusive because, until recently, no postcranial remains could be reliably attributed to this taxon. Here, we report the first associated hand and upper limb skeleton (KNM-ER 47000) of P. boisei from 1.51 to 1.53 Ma sediments at Ileret, Kenya. While the fossils show a combination of primitive and derived traits, the overall anatomy is characterized by primitive traits that resemble those found in Australopithecus, including an oblique scapular spine, relatively long and curved ulna, lack of third metacarpal styloid process, gracile thumb metacarpal, and curved manual phalanges. Very thick cortical bone throughout the upper limb shows that P. boisei had great upper limb strength, supporting hypotheses that this species spent time climbing trees, although probably to a lesser extent than earlier australopiths. Hand anatomy shows that P. boisei, like earlier australopiths, was capable of the manual dexterity needed to create and use stone tools, but lacked the robust thumb of Homo erectus, which arguably reflects adaptations to the intensification of precision grips and tool use. KNM-ER 47000 provides conclusive evidence that early Pleistocene hominins diverged in postcranial and craniodental anatomy, supporting hypotheses of competitive displacement among these contemporaneous hominins. © 2019 Elsevier Ltd. All rights reserved. Keywords: Human evolution Hominins Postcrania Australopithecus Paranthropus Early Homo 1. Introduction Over six decades of discoveries and analyses have made Paranthropus boisei one of the best documented early hominin taxa in * Corresponding authors. (D.J. Green), (B.G. Richmond) E-mail addresses: [email protected] (B.G. campbell.edu (D.J. Green). https://doi.org/10.1016/j.jhevol.2019.102727 0047-2484/© 2019 Elsevier Ltd. All rights reserved. Richmond), dgreen@ terms of its craniodental anatomy, diet, and biogeography. It is well known for a number of remarkably derived characteristics, including a structurally strong facial skeleton, large and anteriorly positioned attachments for chewing muscles, distinctive zygomatic shape, huge postcanine and small front teeth, and thicker enamel than found in any hominin or modern primate (Tobias, 1967; Rak, 1983; Smith et al., 2015). Its postcranial anatomy, however, has remained largely unknown, hampering our understanding of many 2 B.G. Richmond et al. / Journal of Human Evolution 141 (2020) 102727 important aspects of the paleobiology of this member of the human family tree. Domínguez-Rodrigo et al. (2013) reported their breakthrough discovery of postcranial fossils associated with premolars reliably attributable to P. boisei. The associated remains (OH 80) include shaft fragments of a tibia and femur, a proximal radius, and a distal humerus fragment. OH 80 provides compelling evidence that the postcranial skeleton of P. boisei is robust compared with that of other hominins but leaves open many questions about the postcranial anatomy and paleobiology of this species. Here, we report the discovery of the first upper limb skeleton, KNM-ER 47000 (Fig. 1), attributable to P. boisei. It provides evidence of scapular morphology, hand and upper limb proportions, upper limb robusticity, and locomotor and manual dexterity capabilities. We provide basic morphometric and comparative analyses with living hominoids and fossils attributed to Australopithecus, Paranthropus, and early Homo from eastern and South Africa. 2. Discovery and stratigraphic and paleoecological context KNM-ER 47000 was found along the slope of an exposure in area 1A of the Koobi Fora Formation (Fig. 2), at the site FwJj14E where hominin footprints were later discovered as a result of our work to document the discovery of KNM-ER 47000 (Bennett et al., 2009; Figure 1. Right upper limbs of a modern human (left), chimpanzee (center), and KNM-ER 47000 (right), which preserves lateral portions of the scapula, the distal portion of the humerus, most of the ulna, and most of metacarpals (MCs) 1-3 and proximal phalanges 2-4. KNM-ER 47000 has primitive traits including a gracile thumb MC, lack of MC 3 styloid process, curved phalanges with prominent flexor sheaths, a long and curved ulna, a humerus with thick cortical bone and a prominent brachioradialis flange, and obliquely oriented scapular spine. Derived traits include a relatively long thumb, short manual phalanges, and a lateral scapular glenoid orientation. Scale bar at right is 10 cm. 3 B.G. Richmond et al. / Journal of Human Evolution 141 (2020) 102727 Figure 2. KNM-ER 47000 was recovered at site FwJj14E in northwestern Kenya (a) in area 1A (red outlined area indicated by the arrow, just south of the Il Eriet River) of the Koobi Fora Formation near the northeastern shores of Lake Turkana. (b) Hominin (red stars) and identifiable nonhominin (black circles) fossils were recovered over seven years of survey and excavations. The central rectangular area and other areas outlined in blue show the excavated portions of the surface deposits and the geological trench. This site also preserves layers of previously published hominin and other animal footprints (Bennett et al., 2009; Dingwall et al., 2013; Hatala et al., 2017), including the lower level (black outline) and upper level (green outline) excavations. Each square in the grid system delineated on the borders of the map is 10  10 m. Dingwall et al., 2013; Hatala et al., 2016, 2017). Hillary Sale found the first element, the right third metacarpal (MC), on July 12, 2004, near the end of the field season. Intensive survey led to the discovery of several additional elements, including phalanges, and humeral and ulnar fragments (Table 1). Systematic crawling and surface scraping in 2005 led to the discovery of more elements, including the MC 1, MC 4, and more fragments of the ulna and scapula. Five years of intensive excavation at the bottom of the gully of reworked sediments derived from the primary outcrops yielded more than a thousand mammalian fossil fragments, including portions of KNM-ER 47000's scapula, humerus, and ulna. During this time, excavations of the primary sediment yielded a few sparsely scattered mammalian fossils in situ but no fragments of KNM-ER 47000. All elements and fragments belonging to KNM-ER 47000 were recovered within a constrained area (Fig. 3). The geochronological age is well constrained by three volcanic tuffs preserved at the site and identified as the Lower Ileret (1.53 ± 0.01 Ma), Ileret (1.52 ± 0.01 Ma), and Northern Ileret (1.51e1.52 ± 0.01 Ma) tuffs (Fig. 3; Bennett et al., 2009). The local stratigraphic section consists of about 10 m of interbedded sands, silts, and volcaniclastics from ~1 m below the Lower Ileret Tuff to the top of the Northern Ileret Tuff. Well-preserved surface fossils representing a wide range of taxa occur on the slopes and along the base of these deposits and are clearly derived from them as there are no other local sources, and in situ fossils also are present in the Ileret Tuff Complex. The concentration of the associated hominin upper limb fossils near the base of the local outcrop implies that these were derived from previously eroded sediments in the lower part of the exposures, most likely from a source stratum below the Ileret Tuff (1.52 ± 0.01 Ma) and above the Lower Ileret Tuff (1.53 ± 0.01 Ma). This interval includes paleosols, silt-sand couplets with vertebrate footprints, and sands representing small-scale channel deposits. The environment has been reconstructed as a deltaic plain (Roach et al., 2016), with a fluctuating base level resulting in alternating wetter and drier conditions over the ~20 kyr time span encompassed by the dates of the tuffs. At this time (Okote Mb at Ileret), the mammalian fauna and the stable isotope data (Patterson et al., 2017) indicate habitats that included both lake margin grasslands and more varied vegetation such as tall riverine forest, wooded grassland, and bushland (Bamford, 2017), in a Table 1 Elements, preservation details, and date of discovery for the associated KNM-ER 47000 right arm skeleton. Accession number Element KNM-ER 47000A KNM-ER 47000B KNM-ER 47000C Scapula Humerus Ulna KNM-ER 47000D Ulna KNM-ER KNM-ER KNM-ER KNM-ER KNM-ER KNM-ER Metacarpal 1 Metacarpal 3 Metacarpal 4 Proximal phalanx 3 Proximal phalanx 4 Proximal phalanx 5 47000E 47000F 47000G 47000H 47000I 47000J Preservation Date of discovery Glenoid, acromion, axillary border Distal midshaft, trochlea, medial epicondyle Proximal ~two-thirds of the shaft; lacks the olecranon process Distal ~one-third of the shaft, lacks the distal articular surface Complete, lacks head Complete Shaft missing both articular ends Proximal end and ~two-thirds of the shaft Complete Complete, lacks lateral trochlea 7/2004; 7/2005 7/2004; 7/2005; 7/2008 7/2004; 7/2007 7/2004 6/2005 7/2004 7/2004 7/2004 6/2005 7/2004 4 B.G. Richmond et al. / Journal of Human Evolution 141 (2020) 102727 Figure 3. Stratigraphic section (left) shown with the image (right) of the outcrop highlighting the positions of the three major tuffs. All hominin fossils were found on the surface or in secondarily deposited sediment below the Ileret Tuff (1.52 ± 0.01 Ma). Some of the remains were found above the Lower Ileret Tuff (1.53 ± 0.01 Ma) indicating that the bones must have been buried above it. The large excavation area is visible on the right side of the outcrop and gully; the lower footprint level (Bennett et al., 2009; Dingwall et al., 2013; Hatala et al., 2017) is exposed on the left side of the image. All fragments of KNM-ER 47000 were located on the surface of the lower portion of the outcrop or secondarily buried in sediment eroded from the drainage that extends up the slope from the excavation site toward the right margin of the picture. fluvio-deltaic landscape with cycles of emergence and paleosol formation alternating with periods of sedimentation that preserved both the skeletal remains and footprints of the local inhabitants. The five most abundant species identified from area 1A (in the order of decreasing abundance: Kobus kob, Metridiochoerus compactus, Kolpochoerus limnetes, Theropithecus oswaldi, and Kobus sigmoidalis) have dental enamel with stable carbon isotopic signatures, consistent with a predominantly C4 vegetation diet (Patterson et al., 2017, 2019). 3. Preservation and morphology Most of KNM-ER 47000's elements are well preserved, with undamaged and undistorted articular and shaft anatomy. The close spatial association of recovered fragments along with refitting pieces that display similar sediment abrasion and trampling damage indicates recent erosion from nearby exposures. In addition, dry bone fractures suggest this specimen was exposed for 1e3 years before burial (Behrensmeyer, 1978). The humerus has a tooth mark on the anterior distal shaft produced by a large carnivore, possibly a crocodile based on the shape of the indentation (Njau and Blumenschine, 2006) (Fig. 4). Multiple small, irregular, linear grooves along the MC 4 and ulna shaft indicate early postmortem bone modification by an unknown(possibly invertebrate) agent. These provide some evidence of the agents of carcass modification consistent with predation and/or scavenging and exposure on a floodplain surface before final burial. The fact that only right limb elements were recovered during the extensive excavation and lateral survey (Figs. 2, 3), together with the presence of a large bite mark, raises the possibility that the right upper limb was separated from the rest of the individual by predation or scavenging by a crocodile or large terrestrial carnivore. 3.1. Scapula (KNM-ER 47000A) Preservation KNM-ER 47000's scapula preserves the glenoid fossa, portions of the scapular spine and neck, most of the axillary border, and the proximal portion of the acromion process (Fig. 1). One fragment (four pieces fit together) preserves the complete glenoid fossa, partial acromion, and 44.4 mm of the axillary border; the maximum preserved length is 67.5 mm. The second fragment (three pieces fit together, including small, fragile pieces of the infraspinous fossa) preserves much of the remaining axillary border, with a maximum preserved length of 74.1 mm. The two fragments do not have obvious conjoining surfaces. However, based on the morphology and the shape of the broken borders, these two fragments have little, if any, length missing between them and likely conjoin. When conjoined, the maximum preserved length is 135.3 mm. The acromion and axillary border pieces are light in color, similar to that of the first MC (47000E), whereas the glenoid cavity is much darker, more like the third MC (47000F) (Richmond et al., 2016). This suggests that the individual pieces of the scapula were broken postdepositionally and likely subject to different diagenetic conditions before their recovery. Regardless, the edges of the acromion and glenoid cavity are intact, and the excellent fit warranted their reattachment. The maximum preserved length of the acromion process is 54.3 mm. At the lateral break, the acromion is 16.3 mm superoinferiorly and 16.8 mm anteroposteriorly. The spinoglenoid notch is intact, as is the scapular neck, the base of the spine, and the beginning of the acromion. Roughly 16.8 mm B.G. Richmond et al. / Journal of Human Evolution 141 (2020) 102727 5 Figure 4. (A) The anterior distal shaft of the KNM-ER 47000 humerus has an indentation interpreted as the tooth mark of a large carnivore, which most closely resembles damage caused by crocodile feeding. (B) The MC 4 has a series of indeterminate striations that do not appear to have been the result of sediment abrasion and may represent damage from invertebrate scavengers. of the spine is preserved medial to the spinoglenoid notch. A prominent infraglenoid tubercle is preserved, as is much of the axillary border to the lateral extent of a prominent teres major muscle attachment site. The inferior angle is largely absent, as is the vertebral border. On the ventral surface, the full extent of the ventral bar is present, until it tapers at about the level of the teres major attachment. Very little of the scapula blade superior to the axillary border/ventral bar is preserved; about 15.7 mm of the subscapularis fossa is preserved superior to the ventral bar, and ~20.5 mm of the infraspinous fossa is preserved at the midpoint level of the axillary border. Surface preservation is excellent, with no signs of abrasion or weathering. Trabecular bone is exposed at the medial and lateral broken ends of the acromion and at the broken ends of the bar of the axillary border. A more complete account of the preservation and morphology of KNM-ER 47000A may be found in the study by Green et al. (2018). Morphology The glenoid fossa is 35.9 mm superoinferiorly and 26.5 mm anteroposteriorly, with a minimum postglenoid anteroposterior (AP) depth of 12.0 mm (Table 2). Thus, it is round in shape and not particularly tall relative to its AP width. The glenoid superoinferior curvature has a maximum depth of 3.0 mm along the posterior margin, but reaches a greater depth anterior and inferior to the center of the glenoid fossa. The supraglenoid tubercle is well developed. There is a visible, although not pronounced, notch on the anterosuperior border where the subscapularis tendon would pass en route to the lesser tubercle of the humerus. The scapular neck is broad; the minimum AP thickness is 12.0 mm, and the superoinferior distance between the spinoglenoid notch and the infraglenoid tubercle is 21.0 mm. Similarly, the scapular neck is 12.0 mm thick anteroposteriorly at the infraglenoid tubercle. The acromion is markedly thick, about 10.7 mm superoinferiorly at the base of the spine and 12.4 mm at the level of the glenoid fossa (see the study by Green et al., 2018 for further discussion). The minimum breadth of the acromion is 11.1 mm, and its maximum length is 17.2 mm. The base of the coracoid process is preserved, and there is no indication of an unfused epiphysis. As such, it is reasonable to assume this to be an adult individual. Muscle attachment scars are prominent, including the attachment for the long head of triceps brachii, the center of which is approximately 25.8 mm from the inferior margin of the glenoid fossa. The axillary sulcus has a minimum length (i.e., if the two fragments conjoin with no gap) of 106.5 mm, from the inferior glenoid margin to the projecting teres major flange. The sulcus is very well developed. Midway along its length, the sulcus has a width of 11.5 mm (from the axillary border to the most anteriorly projecting point on the bar). At this point, the bar is 7.5-mm thick anteroposteriorly. Inferiorly, the sulcus ends with a teres major flange projecting 10.5 mm from the medial teres major scar. 3.2. Humerus (KNM-ER 47000B) Preservation The humerus was recovered in two diaphyseal segments and a distal articular fragment that comprise most of the distal half of the bone (Fig. 1). Although much of the trochlea and medial epicondyle are intact, the specimen is missing the entire capitulum, the lateral epicondyle, and a distal piece of the lateral pillar. In addition, much of the posterior trochlear surface is missing, as is the distolateral quadrant of the olecranon fossa. The middle portion of the brachioradialis flange is missing some bone along the lateral margin. The anterior margin of the trochlea is slightly abraded. There is a large tooth mark (4.5  3.5 mm) with a drag mark (3.7-mm long and 0.5-mm wide) on the anterior shaft, proximal to the trochlea (Fig. 4). This mark is similar to a crocodile tooth score; a more complete account of the preservation and morphology of KNM-ER 47000B may be found in the study by Lague et al. (2019b). Morphology The preserved length of the humerus is 161 mm. The overall anatomy of the specimen suggests that the proximal fracture is located either at or very close to the midshaft. Only the most distal end of the deltopectoral crest is preserved, and a nutrient foramen is present about 3 mm below the proximal fracture surface on the anteromedial side. Proximally, the maximum width of the diaphysis (anterolateral to posteromedial) is 26.2 mm, and the minimum width (anteromedial to posterolateral) is 18.7 mm, whereas the true AP and mediolateral (ML) breadths are closer to 25.7 mm and 20.3 mm, respectively (Table 2). The cortical bone is 6 B.G. Richmond et al. / Journal of Human Evolution 141 (2020) 102727 Table 2 Standard morphometric data for the associated KNM-ER 47000 right arm skeleton. Accession number Element KNM-ER 47000A Scapula KNM-ER 47000B Humerus KNM-ER 47000C & D Ulna KNM-ER 47000E Metacarpal 1 KNM-ER 47000F Metacarpal 3 KNM-ER 47000G Metacarpal 4 KNM-ER 47000H Proximal phalanx 3 KNM-ER 47000I Proximal phalanx 4 KNM-ER 47000J Proximal phalanx 5 Measure mm Superoinferior (SI) glenoid height Anteroposterior (AP) glenoid width Length (estimated) Medial trochlear ridge to the nutrient foramen Midshaft (~50%) AP midshaft width Midshaft (~50%) mediolateral (ML) midshaft width Distal (~75% total length) AP shaft depth Distal (~75% total length) ML shaft breadth AP trochlear ridge depth Proximodistal (PD) trochlear ridge height PD trochlear notch height ML trochlear notch to trochlear ridge breadth ML trochlear notch to medial epicondyle breadth ML trochlear ridge to medial epicondyle breadth AP depth of medial epicondyle Length (estimated) AP depth distal to radial notch ML breadth distal to radial notch AP depth of radial notch PD height of radial notch Length (estimated) Dorsopalmar (DP) base (articular) breadth Radioulnar (RU) base (articular) breadth DP base (total) breadth RU base (total) breadth DP midshaft breadth RU midshaft breadth Maximum length Interarticular length DP base breadth RU base (dorsal) breadth DP head breadth RU head breadth (palmar) RU head breadth (dorsal) DP midshaft breadth RU midshaft breadth Length (estimated) DP midshaft breadth RU midshaft breadth Maximum length (estimated) Interarticular length (estimated) DP base (articular) breadth RU base (articular) breadth DP base (total) breadth RU base (total) breadth DP midshaft breadth RU midshaft breadth DP 75% shaft breadth RU 75% shaft breadth Maximum length (estimated) Interarticular length (estimated) DP base (articular) breadth RU base (articular) breadth DP base (total) breadth RU base (total) breadth DP 25% shaft breadth RU 25% shaft breadth DP midshaft breadth RU midshaft breadth DP 75% shaft breadth RU 75% shaft breadth DP head (articular) breadth DP head (trochlear groove) breadth RU head breadth (palmar) RU head breadth (dorsal) Maximum length (estimated) Interarticular length (estimated) DP base (articular) breadth RU base (articular) breadth DP base (total) breadth RU base (total) breadth DP 25% shaft breadth RU 25% shaft breadth DP midshaft breadth 35.9 26.5 300.0 152.0 25.7 20.3 18.8 30.8 24.1 22.3 12.5 11.7 29.1 19.5 12.7 282.0 18.4 20.2 16.0 8.7 46.0 12.9 13.0 14.6 13.3 9.4 10.0 64.1 63.2 15.0 12.1 14.2 14.6 9.2 9.0 8.3 54.1 7.3 6.0 45.0 42.0 10.3 13.7 11.4 15.3 6.3 12.0 7.8 11.7 40.8 38.1 10.1 11.2 10.9 13.4 6.1 9.9 6.4 11.2 7.8 10.3 7.3 5.7 10.2 6.7 33.5 32.0 8.6 10.9 9.1 12.0 6.5 9.4 5.3 7 B.G. Richmond et al. / Journal of Human Evolution 141 (2020) 102727 Table 2 (continued ) Accession number Element Measure RU midshaft breadth DP 75% shaft breadth RU 75% shaft breadth DP head (articular) breadth DP head (trochlear groove) breadth RU head breadth (palmar) RU head breadth (dorsal) mm 9.0 4.8 7.8 5.9 5.0 8.1 5.5 See text for dimensions of fragmentary regions and nontraditional morphometric data. thick, with the walls measuring 6.1 mm (medial), 6.3 mm (lateral), 9.2 mm (anterior), and 6.2 mm (posterior). Distal to the proximal fracture, the shaft becomes much broader mediolaterally than anteroposteriorly, with a very prominent brachioradialis flange arising from the lateral side. At a point about halfway along the total length of the preserved specimen (i.e., just proximal to the point where a fragment of the brachioradialis crest is missing), the shaft measures 18.8 mm anteroposteriorly and 30.8 mm mediolaterally (Table 2). The medial epicondyle is large (19.5-mm proximodistal [PD] height, 12.7-mm AP depth) and projects about 19.5 mm beyond the medial border of the trochlea. The olecranon fossa lacks its distolateral border and has a preserved minimum ML width of 22.0 mm and PD height of 17.8 mm. (The estimated width is 24 mm, and the height is 18 mm.) Enough of the olecranon fossa remains to detect the presence of a large supratrochlear foramen (septal aperture) that measures 6.1-mm high and is missing only its lateral border. The minimum widths of the pillars on the medial and lateral side of the olecranon fossa measure 11.8 mm and 16.9 mm, respectively. Anteriorly, the trochlea is mostly intact, save for the bone missing from the most proximal and distal extensions of the lateral trochlear crest. The medial trochlear crest is entirely preserved with some minor proximal abrasions on its anterior surface. The medial border of the trochlea is roughly circular in the sagittal plane, with 24.1 mm preserved anteroposteriorly and an anterior trochlear PD height of 22.3 mm. The lateral side of the trochlea, however, was fractured in a vertical plane that runs obliquely from the approximate position of the lateral trochlear crest (anteriorly) to a point just shy of the medial trochlear crest. Everything lateral to this fracture is missing, including much of the posterior trochlear surface, the capitulum and capitular fossa, the lateral epicondyle, the distolateral quadrant of the olecranon fossa, and the distal half of the lateral pillar. The trochlear notch (central sulcus) has a minimum PD height (to the septal aperture) of 12.5 mm. The ML width from the trochlear notch to the medial trochlear ridge is 11.7 mm. Figure 5. Cross-sectional views of the ulnar shafts at the distal end of the proximal shaft fragment and at the proximal and distal ends of the distal shaft fragment (which is just proximal to the unpreserved distal end of the bone). Note the thickness of the cortical bone. The 5-cm scale bar is associated with the photograph of the entire bone; the 2-cm scale bar is associated with the three sections. 8 B.G. Richmond et al. / Journal of Human Evolution 141 (2020) 102727 3.3. Ulna (KNM-ER 47000C/D) Preservation This specimen was recovered in four major and a number of small pieces that comprise most of a right ulna, with a 165-mm-long proximal fragment (KNM-ER 47000C) and an unattached distal fragment (KNM-ER 47000D) that has a maximum length of 58.3 mm. The olecranon and most of the trochlea are missing. The distolateral portion of the trochlear surface is preserved including the distal end of the keel, and the radial notch is intact. The proximal piece (~59-mm long) of the long proximal fragment is abraded and weathered, whereas the more distal portions have smooth, well-preserved bone surfaces. The entire proximal fragment is dark in color. The distal fragment is dark, with the exception of light patches covering the proximal third of the surface. The proximal and distal diaphyseal shaft fragments do not refit, and notable differences in the cross-sectional shape indicates that a substantial portion of the distal third of the midshaft is unrecovered (Fig. 5). The cortical bone of the distal portion of the distal shaft fragment is thinner relative to the proximal end and preserves incipient trabecular bone, indicating that it is near the distal articular end. However, the presence of cortical bone indicates that a portion of the distal shaft remains unrecovered, in addition to the distal articular surface. These facts support the reconstruction of a relatively long ulnar length (see the following section for further discussion). Morphology In the side view, the ulna is longitudinally curved and posteriorly convex. In the anterior view, the proximal shaft curves slightly medially. At the distal end of the trochlea, the articular surface extends from the summit of the distalmost keel 18.0 mm laterally. The radial notch measures 8.7 mm superoinferiorly; the preserved AP dimension is 14.8 mm and can be estimated at about 16 mm before the abrasion damage (Table 2). In the coronal plane, the notch is set at an angle of about 137 relative to the lateral portion of the distal trochlear surface (i.e., the angle between the trochlear surface and radial notch). The radial notch angle (see the study by Richmond et al., 1998) relative to the sagittal plane is about 42 . The notch is semicircular in shape and mildly concave. Immediately distal to the radial notch, the shaft is 18.4 mm anteroposteriorly and 20.2 mm mediolaterally; these values are approximate because the shaft tapers distally, and some of the anterior surface bone appears to be abraded. A strongly projecting supinator crest extends distally from the posterior margin of the radial notch. Beginning about 34 mm distal to the notch, a distinct and very prominent interosseous crest rises anteromedially from the shaft and extends to the distal end of the shaft at the level of the pronator quadratus attachment. The crest is so pronounced that along much of the shaft, the cross section has a teardrop shape with a mildly concave lateral border. At the distal end of the lateral trochlea, a narrow (4.4 mm mediolaterally) crest runs along the shaft until it blends into the shaft around 35 mm distal to the trochlear margin. Medial to this crest, the brachialis insertion forms a narrow, deep pit measuring 4.5 mm mediolaterally and about 12.5 mm proximodistally. The center of the insertion lies about 28 mm distal to the distal margin of the trochlear keel. A nutrient foramen emerges from the anteromedial shaft about 89 mm from the distal margin of the trochlear keel. The pronator quadratus crest is smooth, pronounced, and projects medially from the shaft. The distal end of the preserved bone expands slightly, suggesting that it was broken not far from its head, but far enough that there is no preservation of the groove for the extensor carpi ulnaris tendon. The exposed cortical bone at this end measures 2.2 mm anteriorly, 2.7 mm posteriorly, 2.3 mm medially, and 1.8 mm laterally. The exposed cortical bone at the distal end of the proximal fragment (roughly 67% of bone length; an estimated 190 mm from the proximal end) measures 4.1 mm anteriorly, 4.2 mm posteriorly, 4.0 mm medially, and 4.2 mm laterally; the cross section is thickest anteromedially, measuring 4.5 mm at the Figure 6. Various views of the metacarpals: KNM-ER 47000E (MC 1) (left), KNM-ER 47000F (MC 3) (center), and KNM-ER 47000G (MC 4) (right). From top to bottom, all bones are shown in dorsal, palmar, ulnar, and radial views. At the bottom, a proximal view of the base (or the most proximal preserved surface) is shown on the left, whereas the distal view is to the right. MC, metacarpal. B.G. Richmond et al. / Journal of Human Evolution 141 (2020) 102727 interosseous crest. The external dimensions at this level are 14.8 mm anteroposteriorly and 13.8 mm mediolaterally, with a maximum and minimum width of 15.4 mm and 13.3 mm, respectively . The exposed cortical bone at the proximal end of the distal fragment (roughly 78% of bone length; an estimated 220 mm from the proximal end) measures 3.4 mm anteriorly, 3.4 mm posteriorly, 3.5 mm medially, and 3.5 mm laterally. The external dimensions are 11.2 mm anteroposteriorly and 12.9 mm mediolaterally. The medullary cavity is small (4.4 mm anteroposteriorly and 5.9 mm mediolaterally). Hand. The hand partial skeleton includes MCs 1, 3, and 4 (KNM-ER 47000E, F, and G, respectively), and proximal phalanges 3, 4, and 5 (KNM-ER 47000H, I, and J, respectively) (Fig. 1). 3.4. Right first MC (KNM-ER 47000E) Preservation This specimen preserves the base and most of the shaft, but is lacking the distalmost shaft and head. The preserved length is 36.7 mm (from the dorsal base to distalmost shaft). The distal break is mostly transverse, exposing the cortical bone thickness (Fig. 6). The surface shows excellent preservation with no evidence of weathering or abrasion. Morphology The shaft is gracile and curves longitudinally (palmarly concave). At the break, the external dimensions are 8.7 mm dorsopalmarly and 10.5 mm radioulnarly, and the cortical bone (approximately 80% of bone length) measures 2.8 mm palmarly, 2.0 mm posteriorly, 3.2 mm radially, and 3.7 mm ulnarly. The midshaft cross section is slightly wider than deep. The radial side of the shaft bears a faint ridge along the proximal end of the dorsal margin, corresponding to part of the attachment of the opponens pollicis muscle. Proximal to this, the base of the radial shaft has a tubercle in the region of the attachment of the abductor pollicis longus muscle. On the ulnar side of the proximal shaft, a rugosity runs along the dorsal margin that corresponds to the origin of the first dorsal interosseous muscle. At the base, there is a tubercle at the palmar-ulnar margin (possibly the attachment for the deep head of flexor pollicis brevis or “anterior oblique ligament,” as reported for A.L. 333-58 in the study by Bush et al., (1982)). The dorsal surface has a moderately developed ridge, possibly corresponding to the attachment of the dorsoradial ligament. The base has a saddle-shaped trapezium facet. This facet is strongly convex radioulnarly, and extends slightly farther on the ulnar side; this radioulnar (RU) convexity is markedly more curved than the curvature of the dorsopalmar concavity. The palmar aspect of the base has a beak-like projection, with a margin that has a flattened semicircular shape radioulnarly. We estimate that the original length was about 46 mm, based on anatomical comparisons with first MCs of humans, chimpanzees, and australopiths (Table 2). 3.5. Right third MC (KNM-ER 47000F) Preservation This specimen is a very well-preserved, complete adult right third MC (Fig. 6). The surface preservation is almost perfect, with the exception of very slight abrasion on the palmar MC 2 facet and on the dorsoulnar corner and palmoulnar margin of the base. The base and shaft have a very dark, almost black, color, and the head is mottled with patches of lighter color. Morphology This bone is moderately robust with a strong, palmarly concave longitudinal curvature. From a palmar view, the shaft is straight. The head is slightly supinated relative to the base and is asymmetric. The articular surface is convex radioulnarly and widens palmarly to slightly more than 1.5 times the dorsal width. From a palmar or dorsal view, the distal apex of the joint curvature 9 lies in the ulnar half of the head, with a flatter radial half. This asymmetry continues palmarly, whereas the joint surface flattens dorsally. The palmar margin of the head projects 2.5 mm palmarly from the shaft and separates into ulnar (6.7 mm radioulnarly) and radial (3.8 radioulnarly) articular projections. The articular surface does not extend onto the dorsum of the shaft. The fossae for the collateral ligaments are marked. The fossa on the ulnar side is slightly deeper and contains two nutrient foramina. The midshaft is slightly deeper than wide (Table 2). Faint ridges show the dorsal extent of the dorsal interosseous muscles, running from the dorsal tubercles behind the head proximally where they join near the dorsal midshaft. The base lacks a styloid process and has facets for the capitate, MC 2, and MC 4. The capitate facet is wider dorsally than palmarly, and the ulnar portion is evenly convex dorsopalmarly. The dorsoradial portion of the facet projects ~1.0 mm proximally. The dorsal and radial facets margins have pronounced invaginations. The MC 2 articulation has two facets: a large, slightly convex dorsal one measuring 7.5 mm dorsopalmarly and 6.1 mm proximodistally and a small palmar one measuring 2.3 mm proximodistally and about 5.0 mm dorsopalmarly (estimate due to slight abrasion damage). Between the facets is a recessed nonarticular area. The MC 4 articulation consists of a single dorsal facet that is roughly trapezoidal in shape and mildly concave proximodistally. The facet measures 6.5 mm dorsopalmarly and 6.0 mm proximodistally, with a maximum length of 7.9 mm (proximopalmar to distodorsal) and minimum of 4.9 mm (proximodorsal to distopalmar). The dorsal nonarticular surface has a marked area corresponding to the attachment site of the extensor carpi radialis brevis muscle. The ridge is interrupted by a deep depression slightly ulnar to the middle. 3.6. Right fourth MC (KNM-ER 47000G) Preservation This specimen is the shaft of a fourth MC lacking the base and head. The surface preservation is very good. Extensive thin, subparallel striations of unclear origin are visible along the dorsoradial surface. These marks do not resemble classic sediment abrasion or rodent gnawing and may have been the result of invertebrate activity. The color is dark, almost black, with the exception of the light-colored exposed trabecular bone at the distal end (Fig. 6). Morphology This MC is gracile and has slight longitudinal curvature in the side view. In the palmar view, the shaft is straight. The broken end of the distal shaft, near the base of the head, exposes the cortical bone and some trabecular bone on the palmar side. The distal end measures 8.5 mm dorsopalmarly and 8.3 mm radioulnarly. The cortical thickness measures 2.0 mm palmarly, 2.0 mm radially, and 2.0 mm ulnarly. (The dorsal cortex is broken at this level.) The midshaft is ovoid, slightly larger dorsopalmarly than radioulnarly. Along the dorsal surface, faint ridges extend proximally from the epicondyle bases and converge near the base, at about one-fourth of the length of the shaft. The base is broad (8.4mm preserved breadth) relative to the narrow shaft; at it is narrowest on the proximal end, the shaft measures 6.9 mm dorsopalmarly and 6.2 mm radioulnarly. The dorsal aspect of the base preserves a small portion of the dorsoulnar projection. We estimate that the original length was about 54 mm, based on the anatomical comparisons with fourth MCs of humans, chimpanzees, and australopiths (Table 2). 3.7. Right third proximal phalanx (KNM-ER 47000H) Preservation This specimen preserves the proximal half of an adult proximal manual phalanx that is identified as belonging to the right 10 B.G. Richmond et al. / Journal of Human Evolution 141 (2020) 102727 third digit. The break is roughly transverse and angled slightly proximally. Thin laminae are missing from the dorsal surface of the shaft. The remaining surface is very well preserved. The bone is dark in color with mottled light patches on the surface (Fig. 7). Morphology This phalanx is robust and exhibits notable curvature. The flexor ridges are well developed, projecting farther palmarly than the midline keel of the shaft, making the palmar shaft slightly concave radioulnarly. The basal palmar tubercles are pronounced, and the ulnar one is broader (6.6 mm radioulnarly) than the radial one (5.2 mm radioulnarly). Continuous with the tubercles, the radial side of the base corresponding to the attachment of the dorsal interosseous muscle projects to the side more prominently than the ulnar side. There is a distinct tubercle near the dorsoradial margin of the base that is typical of hominin third proximal phalanges. This tubercle, in addition to its large size relative to the other phalanges, provides evidence that the bone belongs to the right third digit. The MC joint is ovoid and concave, with greater RU curvature. The radial, dorsal margin of the joint extends farther proximally (~0.8 mm) than does the ulnar margin. The broken end of KNM-ER 47000H exposes cortical bone that is so thick, it almost appears to lack a medullary cavity altogether (<1 mm diameter; Fig. 7). 3.8. Right fourth proximal phalanx (KNM-ER 47000I) Preservation This specimen is a complete adult manual proximal phalanx, identified as most likely belonging to the right fourth digit. The surface preservation is excellent, apart from some root etching marks and minimal abrasion of the dorsal margin of the radial condyle and proximal, palmar margin of the ulnar condyle. The fossil has a light color (Fig. 7). Morphology The phalanx is robust with a palmarly concave longitudinal curvature. The head is asymmetric, the radial condyle wider and projecting further distally than the ulnar condyle; in the palmar view, the radial condyle is 5.6 mm radioulnarly and the ulnar condyle is 4.7 mm radioulnarly. In the palmar view, a ridge runs from the base of the ulnar condyle proximally and radially to the flexor ridge. The flexor ridges run most of the length of the shaft and are very well developed. They project palmarly from the shaft, creating a concave RU profile and a crescent-moon cross section. At the base, the tubercles are prominent, and the ulnar tubercle is slightly wider than its radial counterpart. The ulnar side of the base also projects slightly more than the radial side owing mainly to a dorsopalmar ridge corresponding to the attachment site of the dorsal interosseous muscle. The MC joint has an oval facet with a sharp dorsal margin. 3.9. Right fifth proximal phalanx (KNM-ER 47000J) Preservation This specimen is a complete right adult manual proximal phalanx. The bone is intact from the proximal end to the head, where an oblique break separates a dorsoradial head fragment from the rest of the bone. The surface preservation is excellent (Fig. 7). Morphology The phalanx is robust with a slight, palmarly concave longitudinal curvature. The head is slightly asymmetric. The condyles project distally about the same extent, and the radial condyle is slightly wider and more convex radioulnarly. In the palmar view, the ulnar condyle has a thin (2.1 mm radioulnarly) joint surface that extends 2.8 mm proximally. The flexor ridges are pronounced and Figure 7. Various views of the manual phalanges: KNM-ER 47000H (PP 3) (left), KNMER 47000I (PP 4) (center), and KNM-ER 47000J (PP 5) (right). From top to bottom, all bones are shown in dorsal, palmar, radial, proximal, and distal views. PP, proximal phalanx. B.G. Richmond et al. / Journal of Human Evolution 141 (2020) 102727 11 run virtually the entire length of the shaft. From the sides of the shaft, they project almost as far palmarly as the apex of the shaft's midline surface, giving the cross section a semicircular shape. However, shallow grooves separate the ridges from the midline keel of the shaft. At the base, the basal tubercles are well developed. The ulnar tubercle is larger and connects distally to the midline keel via a ridge. This ridge is absent on the radial side, making the palmar aspect of the base asymmetric. The ulnar side of the base features a tubercle corresponding to the attachment of the abductor digiti minimi muscle. The MC articular surface is an ovoid joint, broader and more strongly concave radioulnarly, and slightly dorsally canted relative to the shaft. surface scans of original specimens collected by J.M. Plavcan and C.V. Ward. Our extant comparative sample is the same as that in the study by Lague (2015) and consists of 65 adult humeri representing four hominid species. Before making shape comparisons, we used generalized orthogonal least-squares Procrustes analysis (Gower, 1975) to superimpose raw landmark configurations into the same shape space (via tpsRelw software; Rohlf, 2010). Semilandmarks were allowed to slide along the diaphyseal outline using the criterion of minimized bending energy (Bookstein, 1997). Procrustes distances (Dp) were calculated as a measure of overall shape dissimilarity (Rohlf, 1999; Zelditch et al., 2012). 4. Materials and methods The two diaphyseal fractures on KNM-ER 47000B provide excellent views of cortical bone at sections that are very close (if not identical) to those typically used for analyzing the cross-sectional geometric properties of fossil and modern hominin humeri (i.e., 35% and 50% of humerus length from the distal end). A detailed analysis of the cross-sectional geometric properties of KNM-ER 47000B may be found in the study by Lague et al. (2019a). Here, we present an alternative means of assessing the relative bending strength of KNM-ER 47000B that uses distal articular breadth and therefore does not require estimates of body mass or humerus length. Periosteal and endosteal contours of the ~35% section were digitally traced on a photographic image to avoid incorporation of noncortical material in defining the edge of the medullary cavity (see also figure 1 from the study by Lague et al., 2019a). The resulting binary image (with cortical bone represented in black) was imported into ImageJ software (Schneider et al., 2012) to calculate cross-sectional properties using the MomentMacroJ plugin (version 1.4B; available at http://www.hopkinsmedicine.org/ fae/mmacro.html). The image was size calibrated by matching distances in pixels to known distances (in mm) recorded from the fossil. Diaphyseal strength was assessed by the section modulus (Z; mm3), which is calculated by dividing the polar second moment area (J; mm4) by the maximum radius. The polar section modulus of KNM-ER 47000B was assessed relative to distal humeral articular breadth (HDARTML) using regression equations published by Ruff et al. (2016). In particular, we calculated the relative deviation (in standard error units) of KNM-ER 47000B from published regression equations (Zp against HDARTML) for both modern humans and chimpanzees. In this way, the humeral strength of KNM-ER 47000 could be compared with that expected for Homo and Pan (based on HDARTML) and with that of other fossil hominin humeri examined by Ruff et al. (2016). Although the distal articular surface of KNM-ER 47000B is incomplete, we estimated HDARTML to be 41.6 mm from linear regressions of HDARTML against trochlear width for a sample of 65 extant hominids (Homo, Pan, Pongo, Gorilla; Lague, 2015) and 15 fossil hominins (R ¼ 0.96, P < 0.001). Our analysis of humeral strength includes a range of reasonable values for the HDARTML of KNM-ER 47000B (40e43 mm). 4.1. Scapular shape Two scapular shape variables were considered to assess glenohumeral joint (ventral bar/glenoid angle) and scapular spine orientation (axillary border/spine angle). These two variables enable comparison of KNM-ER 47000 with representatives of Australopithecus afarensis (A.L. 288-1 [first order cast from the Cleveland Museum of Natural Sciences] and KSD-VP-1/1 [data from the study by Melillo, 2016]), Australopithecus africanus (Sts 7 and StW 162 [original fossils]); Australopithecus sediba (MH2 [threedimensional printed cast provided by S.E. Churchill]), Homo ergaster or early African Homo erectus (KNM-WT 15000 [actual fossil]), and Homo floresiensis (LB6/4; data from the study by Larson et al., (2009)). All of the fossil specimens in this study are adults with the exception of KNM-WT 15000. Because this fossil is closest to KNM-ER 47000A within our sample in both space and geological time, this specimen was included together with a comparative sample of extant individuals representing similar developmental ages. This enabled more appropriate comparisons between the fossil and extant samples. Comparative data were collected on scapulae from adult and adolescent modern humans (Homo sapiens [n ¼ 160]) and great apes (Pan troglodytes [n ¼ 209], Gorilla gorilla [n ¼ 207], and Pongo pygmaeus [n ¼ 86]) using an Immersion MicroScribe G2 digitizer. Angle calculations were performed using R (Ihaka and Gentleman, 1996) by defining a plane on the scapula using three of the four points (which are roughly coplanar) that define the two lines forming the angle of intersection (Green, 2013; Green and Alemseged, 2012). Given that KNM-ER 47000A and several of the comparative fossil specimens are fragmentary, some modifications from the methods outlined in the study by Green (2013) were required; see the study by Green et al. (2018) for a more complete discussion. 4.2. Distal humeral diaphysis shape Previous research has demonstrated that a transverse section through the distal humeral diaphysis (at about 19% of humerus length from the distal end) is particularly informative for taxonomic identification among Lower Pleistocene hominins (Lague, 2015; Senut, 1981; Susman et al., 2001). KNM-ER 47000B was added to the multivariate analysis presented in the study by Lague (2015). For each specimen, 60 two-dimensional coordinates were collected from the surface outline of a transverse section through the distal diaphysis (located at ~19%), including two “type 2” landmarks (i.e., the most medial and lateral points of the section) and 29 sliding semilandmarks on both the anterior and posterior surfaces. For the majority of fossils (including KNM-ER 47000B), the cross-sectional profiles were collected from three-dimensional 4.3. Diaphyseal strength of the humerus 4.4. Humerus and ulna length Two of us (B.G.R. and M.R.L.) independently estimated humerus length by visual inspection against humeri of modern apes, humans, and early Pleistocene humeri (especially KNM-ER 739, which is larger but closely resembles the morphology of KNM-ER 47000B). These visual comparisons independently yielded similar length estimates of 300 mm. Geometric morphometric scaling down of KNM-ER 739 to a best fit with KNM-ER 47000 was used as an alternative way to estimate its length. The maximum length of KNM-ER 739 (which is missing only its head) was estimated by 12 B.G. Richmond et al. / Journal of Human Evolution 141 (2020) 102727 adapting the methods described by Gunz et al. (2009) to reconstruct missing anatomy. A thin plate spline interpolation yielded multiple reconstructions based on different reference specimens (i.e., 41 specimens of extant H. sapiens, P. troglodytes, and Pongo sp.). Our average thin plate splineebased maximum length estimate for KNM-ER 739 is 329 mm, which is nearly identical to McHenry's (1973) estimate of 328 mm. Scaling down KNM-ER 739 to best fit KNM-ER 47000 produced an estimate of 300 ± 20 mm for KNM-ER 47000, bolstering the previous estimate based on visual comparisons. By visually comparing the ulnar fragments to intact modern human, chimpanzee, and some fossil hominin ulnae (e.g., OH 36, A.L. 438, Omo L90-19), the original length was estimated to be between 290 and 300 mm. Geometric morphometric methods best fitting modern ape and human ulnae to the proximal main fragment produced an estimate of 282 ± 5 mm (Table 2). Figure 8. Standard box and whisker plots of (A) ventral bar/glenoid and (B) axillary border/spine angles across extant taxa and fossil individuals. Given that the KNM-WT 15000 fossil is a subadult, fossils were compared with like-aged (i.e., “adolescent” and adult) extant specimens to limit the potential for ontogenetic influences on shape. The Australopithecus fossils have more cranially oriented glenoid fossae than those of modern Homo, KNM-ER 47000A, and KNM-WT 15000, all of which display more laterally oriented joints. With the exception of Sts 7 and KNM-ER 47000A, most fossil taxa have spine orientations that are more horizontal, falling either within the modern Homo or Pongo range. Boxes encompass the middle 50% of the data range, the central line represents the median value, and the whiskers extend from the lower 10% to the upper 90% of the samples. B.G. Richmond et al. / Journal of Human Evolution 141 (2020) 102727 13 Figure 9. Comparison of KNM-ER 47000 with extant and fossil humeri. KNM-ER 6020 and KNM-ER 47000 are depicted with estimated regions of missing bone (black lines). Apart from KNM-WT 15000, all the depicted fossil humeri are hypothesized to belong to P. boisei (Lague, 2015; Lague et al., 2019b). 4.5. Relative size and shape of the humerus and hand Robusticity of the first MC was assessed by comparing ML breadth of the midshaft relative to the ML breadth of the humerus at both midshaft (~50% of total length) and at 25% of total length (from the distal end). Midshaft MC 3 breadth also served as a comparative measure of MC 1 breadth. In addition, intrinsic hand dimensions were assessed by comparing total MC 1 length and MC Figure 10. Analysis of distal diaphyseal shape variation of the humerus (~20% section) based on Procrustes-aligned coordinate data (see the study by Lague, 2015 for details). (A) Procrustes distances between KNM-ER 47000B and other fossil hominin humeri. The horizontal dashed lines represent median distances associated with australopiths (lower) and fossil Homo (upper), respectively. KNM-ER 47000B is far more similar, on average, to specimens of the former group than to specimens of the latter group. (B) Standard box plots representing distributions of Procrustes distances calculated between individual and extant hominid specimens and species means. Samples are from the study by Lague (2015) and include H. sapiens (n ¼ 21), P. troglodytes (n ¼ 21), G. gorilla (n ¼ 12), and P. pygmaeus (n ¼ 11). The 65 “intraspecific” values were calculated as the distance of each extant individual to their own species mean. For each specimen, Procrustes distances were also calculated to each of the three contraspecific species means; the resulting 195 distance values comprise the “intergeneric” distribution. Symbols: diamonds ¼ specimens associated with taxonomically diagnostic craniodental material; squares ¼ specimens identified with conventional taxonomic allocation (e.g., based on provenience); circles ¼ specimens identified by Lague (2015). 14 B.G. Richmond et al. / Journal of Human Evolution 141 (2020) 102727 3 length. In addition to KNM-WT 15000, representatives of A. sediba, A. afarensis, A. africanus, H. naledi, and H. neanderthalensis were included in the comparisons. Comparative data were collected from a sample of modern humans (H. sapiens [n ¼ 51 for MC comparison and n ¼ 51 for humerus]) and great apes (P. troglodytes [n ¼ 43 and n ¼ 62, respectively], Gorilla sp. [n ¼ 56 and n ¼ 67, respectively], and Pongo sp. [n ¼ 33 and n ¼ 20, respectively]). KNM-WT 15000, MH2 (A. sediba), and StW 418 (A. africanus) were measured on original specimens by B.G.R. and D.J.G. A. afarensis data are from the study by Bush et al. (1982). H. naledi data are from the study by Kivell et al. (2015), and Neanderthal data are from the study by Trinkaus (1983) and T.L. Kivell and S.E. Churchill (pers. comm.). Finally, phalangeal curvature was assessed by comparing the included angle of manual proximal phalanges (as defined by Richmond, 1998) relative to phalangeal length. In addition to KNM-ER 47000I and J, Orrorin tugenensis (BAR 3490 00) and seven A. afarensis specimens (A.L. 2881; A.L. 333-33, -57, -93, 333w-4, 333x-13a, -19) comprised the comparative fossil data set. All fossil and comparative extant data (H. sapiens [n ¼ 42], Pan [n ¼ 17], Gorilla sp. [n ¼ 43], and Pongo sp. [n ¼ 48]) were collected by B.G.R. 5. Results 5.1. Scapula Both KNM-ER 47000A and KNM-WT 15000 have ventral bar/ glenoid angles that approach the modern human mean and fall in the upper ranges observed for nonhuman hominids. This morphology is indicative of a more derived, laterally oriented shoulder joint than typically occurs in apes and contrasts with the more cranial (ape-like) glenoid orientation seen in Australopithecus (Fig. 8). At the same time, the KNM-ER 47000A scapula possesses an obliquely oriented spine that suggests a narrow, long infraspinous region reminiscent of modern apes. Raw and comparative metrics for KNM-ER 47000A may be found in the study by Green et al. (2018). 5.2. Humerus The overall anatomy of KNM-ER 47000B strongly resembles the morphology observed for multiple fossil specimens attributed to P. boisei, including KNM-ER 739, KNM-ER 1504, and KNM-ER 6020 Figure 11. Diaphyseal strength of fossil hominin humeri (35% section) relative to that observed in modern humans (A) and chimpanzees (B), along with transverse fracture surfaces (C) observed on KNM-ER 47000B at approximately 50% and 35% of humerus length (measured from the distal end). Values on the chart represent relative deviations (RDs, in standard error units) from published regression equations (Ruff et al., 2016) for diaphyseal strength (polar section modulus; Zp) against distal humeral articular breadth (HDARTML). RD values for fossils other than KNM-ER 47000B (red dot: Zp ¼ 1678, HDARTML ¼ 41.6 mm) are from the study by Ruff et al. (2016). Dotted lines above and below the red dot represent RD values associated with a range of reasonable estimates for the HDARTML (40e43 mm) of KNM-ER 47000. Fossil specimens with RD values closer to zero have Zp closer to the predicted value for the given regression (Homo or Pan); specimens within the shaded areas are within the 95% confidence interval of the given regression line. KNM-ER 47000B is similar to australopiths (green circles), and different from fossil Homo (blue hexagons), in having high relative humeral strength comparable with that of chimpanzees. B.G. Richmond et al. / Journal of Human Evolution 141 (2020) 102727 (Fig. 9). These specimens share a unique elbow morphology that includes a highly projecting medial epicondyle, a relatively shallow olecranon fossa, relatively wide medial and lateral pillars, and a deep trochlear central sulcus (Lague and Jungers, 1996; Lague et al., 2019b). KNM-ER 47000B is particularly similar in its preserved diaphyseal and elbow morphology to KNM-ER 739, a larger and more complete 1.5 Ma specimen from Ileret. The outline of the ~19% diaphyseal section on KNM-ER 47000B is more similar to that of australopiths (Paranthropus, Australopithecus) than to that of fossil Homo specimens (H. habilis, H. erectus, H. naledi) (Fig. 10; Lague et al., 2019b). Diaphyseal shape differences between KNM-ER 47000B and australopith humeri are comparable with the relatively small differences observed within extant hominid species. In contrast, shape differences between KNM-ER 47000B and fossil Homo specimens are more consistent with observed intergeneric differences. Among fossil humeri associated with taxonomically diagnostic craniodental material, KNM-ER 47000B is most similar to OH 80 (P. boisei), TM 1517 (P. robustus), and A.L. 288-1 (A. afarensis). KNM-ER 47000B is clearly characterized by very thick cortical bone, as also reported for the P. boisei humerus from the Olduvai Gorge (OH 80-10) (Domínguez-Rodrigo et al., 2013). The more distal of the two diaphyseal fractures on KNM-ER 47000B is at approximately the same section level (35%) as the proximal fracture surface of OH 80-10, and the two specimens share thick cortical bone and an unusually small medullary cavity (Lague et al., 2019a). The diaphyseal strength (polar section modulus; Zp ¼ 1678 mm3) of KNM-ER 47000B at this section can be assessed using published regression equations for Zp against distal articular breadth (Ruff et al., 2016). KNM-ER 47000B is similar to specimens attributed to Australopithecus (A.L. 288-1, StW 431) and P. boisei (KNM-ER 739) in having humeral strength (relative to articular size) that is high compared with that of modern humans and similar to that of 15 chimpanzees (Fig. 11). The cortical robusticity of this specimen reflects extreme bending/torsional strength that is consistent with habitual use of the upper limbs for arboreal climbing (Lague et al., 2019a). A complete account of raw and comparative metrics for KNM-ER 47000B may be found in the study by Lague et al. (2019a, b). 5.3. Ulna The ulnar shaft bears a strongly projecting supinator crest. A distinct and very prominent interosseous crest rises anteromedially from the shaft and extends to the distal end of the shaft at the level of the pronator quadratus attachment. The crest is so pronounced that along much of the shaft, the cross section has a teardrop shape with a mildly concave lateral border (Fig. 5). The broken ends expose very thick cortical bone, providing additional evidence of great upper limb strength (Ruff et al., 2006). The ulna shows a substantial longitudinal curvature, similar to that of great apes and OH 36, the robust ulna from the Olduvai Gorge (Fig. 12). The curvature similarity is noteworthy in that OH 36 has been hypothesized to represent P. boisei (Aiello et al., 1999; Lague et al., 2019b; McHenry et al., 2007). Using an estimated ulnar length of 282 mm, the ulna:humerus proportions (94% [92e100%, considering the range of morphometric and anatomical ulna estimates, holding humerus estimate at 300 mm]) are comparable with those of A. afarensis (94e95% [A.L. 288-1: ulna ¼ 222e226 mm, humerus ¼ 237 mm; Johanson et al., 1982; Kimbel et al., 1994; Nachman et al., 2018]) and A. sediba (91%; [UW88-62: ulna ¼ 246 mm and UW88-57: humerus ¼ 270 mm, lengths measured by B.G.R.]). These estimates suggest brachial proportions shorter than those typical for modern Pan, longer than those typical for modern humans, and comparable with those of earlier australopiths (Churchill et al., 2013). Figure 12. Comparison of KNM-ER 47000 with extant and fossil ulnae. KNM-ER 47000 exhibits a similar diaphyseal curvature as OH 36 (Upper Bed II, Olduvai Gorge), which most likely also represents P. boisei (Aiello et al., 1999; Lague et al., 2019b; McHenry et al., 2007). 16 B.G. Richmond et al. / Journal of Human Evolution 141 (2020) 102727 5.4. Hand The KNM-ER 47000E MC 1 is primitive in its gracility relative to humeral shaft dimensions, more similar to that seen in A. sediba, and distinct from the relatively broad MC 1 of (possibly) H. erectus and more derived members of Homo (Fig. 13). In the palmar view, the shaft is parallel sided like A.L. 333w-39 unlike SKX5020 from Swartkrans (attributed to Homo) and H. naledi (Kivell et al., 2015; Richmond et al., 2016). The beak-like projection on the basal palmar surface of the saddle joint (as opposed to the “beak-like” projection described in some hominin MC 1 heads) is more rounded in comparison with the fairly pointed projections in A.L. 333-58 and A.L.333w-39. The MC 1 trapezium facet is much more curved than that in modern humans and more closely resembles the condition in several other early hominin taxa (e.g., SKX5020, SK84, Stw418, A.L. 333w-39). The absence of a styloid process on the KNM-ER 47000F MC 3 is a primitive feature shared with Australopithecus. Both MCs 3 and 4 of KNM-ER 47000 have a dorsally convex longitudinal curvature, similar to that seen in MCs attributed to A. afarensis (e.g., A.L.333-16, A.L.333-56; Bush et al., 1982). The MC 1 shaft is also gracile with respect to the midshaft breadth of the MC 3 shaft. In this particular comparison, there is Figure 13. Mediolateral midshaft metacarpal 1 breadth compared with (A) midshaft humeral ML breadth and (B) ML breadth at 25% of total humerus length (from the distal end) as a measure of relative first metacarpal robusticity (all units are millimeters). KNM-ER 47000 and MH2 have primitive, ape-like humerus and metacarpal proportions, with a wide
n C1 and Shanidar (Shanidar 3 humerus distal humerus shaft and pronounced brachioradialis flange and narrow thumb metacarpal. In contrast, KNM-WT 15000 and the Tabu combined with the Shanidar 4 MC 1) Neanderthals show a derived, humanlike anatomy combining a robust thumb with a relatively gracile humerus. These fossils show that early Pleistocene hominins differed substantially in postcranial and craniodental anatomy. ML, mediolateral; MC, metacarpal. B.G. Richmond et al. / Journal of Human Evolution 141 (2020) 102727 more overlap between the Homo and Gorilla samples; KNM-ER 47000, along with A. afarensis and A. africanus, falls among this area of overlap. In contrast, and despite having absolutely narrower MCs than the modern sample, A. sediba and H. naledi appear to lie along the modern human trend line (as does the Shanidar Neanderthal) (Fig. 14). Estimated MC 1:MC 3 length proportions of the KNM-ER 47000 are comparable with those of modern humans (Fig. 14) and are consistent with more humanlike thumb:hand length proportions in A. africanus, A. sediba, and primitive Homo taxa, where the relevant fossils have been recovered (Green and Gordon, 2008; Kivell et al., 2011, 2015; Richmond et al., 2016). 17 KNM-ER 47000 also has short fingers relative to MC lengths (MC 3 and MC 1), similar to the proportions in A. sediba and modern humans. The KNM-ER 47000I and J (4th and 5th) proximal phalanges show a curvature (included angle) that is on par with O. tugenensis and A. afarensis, greater than that seen in modern humans, and near the area of overlap between Pan and Gorilla (Figs. 7 and 15). The phalanges display extremely thick cortical bone, to the point that the midshaft has almost no medullary cavity. The cortical bone is thick even compared with suspensory hominoids such as siamangs (Richmond, 2007), indicating that the phalanx was capable of Figure 14. Comparative intrinsic hand proportions. (A) KNM-ER 47000 has a long thumb relative to the hand, like modern humans and unlike great apes. Estimated MC 1:MC 3 proportions are comparable with those of other early hominin taxa including representatives of A. afarensis (A.L. 333w-39 [MC 1]; A.L. 333w-16 [MC 3]), A. africanus (StW 418 [MC
n C1). (B) The midshaft breadth comparison of MC 1 and MC 3 also 1]; StW 64 and 68 [MC 3]), A. sediba (MH2), H. naledi (Hand 1), and H. neanderthalensis (Shanidar 3 and Tabu demonstrates the relatively gracile nature of the KNM-ER 47000, A. afarensis, and A. africanus MC 1s compared with modern humans, the Shanidar Neanderthal, A. sediba, and H. naledi. All units are millimeters. MC, metacarpal. 18 B.G. Richmond et al. / Journal of Human Evolution 141 (2020) 102727 withstanding considerable loading. Compared with hominin proximal phalanges from Swartkrans (e.g., SK22431, 22741, 22511 5018), KNM-ER 47000 phalanges have more pronounced flexor ridges that project farther palmarly and extend farther proximally and distally along the shaft. In total, this makes the shaft wider mediolaterally with respect to the articular ends. In this regard, the KNM-ER 47000 proximal phalanges are more similar to proximal phalanges from Hadar (e.g., A.L.333-63, A.L. 333w-4). The KNM-ER 47000 phalanges are also similar to the Hadar phalanges in degree of longitudinal curvature and shape (Fig. 15), but KNM-ER 47000 phalanges are wider mediolaterally relative to their length. 6. Discussion The Okote Member (1.4e1.6 Ma) near Ileret has yielded taxonomically identifiable craniodental remains of three hominin taxa: H. erectus, H. habilis, and P. boisei (Constantino and Wood, 2007; Spoor et al., 2007). The northern areas of the Koobi Fora Formation include some of the most intensively prospected and studied paleoanthropological regions in Africa, with a half-century of active research and sampling, making it unlikely that any other hominin taxon lived there at the time and has remained undetected. This article presents a description of the KNM-ER 47000 forearm and hand with additional commentary on the shoulder and arm (Fig. 16); the latter two elements are more thoroughly considered in the studies by Green et al. (2018) and Lague et al. (2019a, b), respectively. Comparisons with fossils reliably attributed to H. erectus, H. habilis, and P. boisei provide strong evidence that KNM-ER 47000 can be attributed to P. boisei. These taxa differ from one another in the diaphyseal shape of the distal humerus (Lague, 2015), one of the few anatomical regions preserved in OH 80, the only fossil to date with postcrania associated with uncontested P. boisei craniodental anatomy (Domínguez-Rodrigo et al., 2013). As noted previously, KNM-ER 47000's distal humerus shape matches that of OH 80 and other humeri now attributable to P. boisei, as well as humeri of P. robustus and Australopithecus. Comparisons among KNM-ER 47000 and broadly sympatric, contemporaneous fossils are vital with regard to taxonomic considerations. As demonstrated previously, there are striking differences between KNM-ER 47000 and KNM-WT 15000, in particular, in scapular, humeral, ulnar, and hand anatomy, reinforcing our contention that these fossils do not represent the same taxon. Figure 15. Comparative phalangeal curvature (phalangeal length is reported in millimeters). The included angle for both KNM-ER 47000I and J is comparable with that observed for A. afarensis and O. tugenensis proximal phalanges. These fossils all exhibit curvature values similar to those of Gorilla and Pan and greater curvature than that observed in modern Homo. KNM-ER 47000I is represented by the empty triangle; KNM-ER 47000J is the gray-shaded triangle next to O. tugenensis. Furthermore, the KNM-ER 47000B humerus is larger than that of KNM-ER 3735 and OH 62, both of which are attributed to H. habilis. KNM-ER 47000B is nearly double the size of OH 62 in the cortical, medullary, and total subperiosteal area at the midshaft (~50%) (Lague et al., 2019a). It is also larger than KNM-ER 3735 in most measurements of the distal humerus and in overall distal humeral size (Lague et al., 2019b). At the same time, KNM-ER 47000B is on the smaller end of the range of humerus fossils previously attributed to P. boisei (Lague et al., 2019b). The KNM-ER 47000 ulna is also smaller than OH 36 (Fig. 12), another purported representative of P. boisei (Aiello et al., 1999; McHenry et al., 2007). Such variation might be consistent with high levels of size dimorphism observed in P. boisei craniodental fossils (Silverman et al., 2001). KNM-ER 47000 has a narrow, gracile thumb MC and broad distal humerus shaft. In contrast, KNM-WT 15000 has a derived, modern humanlike broad, robust thumb MC coupled with a narrow, gracile humerus shaft (Fig. 13). Some researchers (Walker and Leakey, 1993) have been hesitant about the attribution of specimens “BU” and “BV” to the KNM-WT 15000 skeleton as right and left MC 1s, respectively, based on their lack of epiphyses and provenance. If these bones do not belong to KNM-WT 15000 (or if they are not MCs), there would be no MC 1s currently known for H. erectus. Nonetheless, there remains the stark difference in humeral size and shape between KNM-ER 47000 and KNM-WT 15000, and the KNMER 47000E MC 1 is gracile compared with modern human MC 1 morphology. The gracile thumb coupled with the lack of an MC 3 styloid process adds evidence to the conclusion that KNM-ER 47000 should not be attributed to H. erectus (see also the studies by Green et al., 2018; Lague et al., 2019a, b). Although more projecting than that of KNM-ER 47000, the styloid processes of H. naledi MC 3s are described as absent with respect to the typical presentation in more derived members of the genus Homo (Kivell et al., 2015). This evidence challenges the notion that a prominent styloid process evolved as a suite of integrated carpometacarpal features early in the evolution of the genus Homo. Alternatively, a prominent MC3 styloid process may characterize the adaptive suite of more derived members of the genus Homo, possibly including H. erectus (Ward et al., 2014). As such, the morphology of the KNM-WT 51260 MC (Ward et al., 2014) further supports the notion that KNM-ER 47000 is not a member of the genus Homo, at least as it pertains to contemporaneous, eastern African representatives. Hand proportions show that P. boisei, like earlier australopith species, had the capacity for fine precision grips (Kivell et al., 2011; Richmond et al., 2016). The H. naledi material further demonstrates that the lack of an MC 3 styloid process coupled with curved phalanges does not appear to have precluded stone tool use in P. boisei (Kivell et al., 2015). This raises interesting questions about whether or not the tools found at the many archaeological sites containing P. boisei fossils were used, transported, or even made by P. boisei. This is not to suggest that Homo did not make and use stone toolsdit is incontrovertible that H. erectus made and used stone tools, since both persisted after the extinction of P. boisei. The question remains whether or not P. boisei made and used them as well and, if so, to what degree and capacity and in what contexts? To be sure, P. boisei lacked derived adaptations seen in later representatives of the genus Homo, including the lack of thumb robusticity. Robust thumb anatomy is not required to make and use stone tools, however, because they appear in the archaeological record long before thumb robusticity evolved (Braun et al., 2019; Harmand et al., 2015; McPherron et al., 2010; Semaw et al., 1997). Recent experimental results have been somewhat equivocal with regard to the magnitude of loads experienced by the thumb during tool making, thus B.G. Richmond et al. / Journal of Human Evolution 141 (2020) 102727 19 Figure 16. Summary of the principal features of the preserved upper limb elements associated with the KNM-ER 47000 remains. MC, metacarpal. confounding the relationship between these activities and the evolution of the robust modern human thumb. Williams et al. (2012) showed that skilled knappers did not experience high mechanical loads on the thumb, whereas Rolian et al. (2011) and Williams-Hatala et al. (2018) did find higher thumb loading in novice tool makers. The thumb applies higher forces and experiences elevated loads when using precision pinch grips during stone tool use. As such, robust thumbs are capable of generating these high precision grip forces and withstanding greater stresses (Williams et al., 2012). In addition to evolving greater robusticity, the thumb evolved a larger and more diverse musculature (Diogo et al., 2012), evolved a derived wrist structure (Tocheri et al., 2008), and became hyperopposable in later Homo (Richmond et al., 2016). The origin of thumb robusticity in H. erectus occurred long after the advent of the earliest stone tools. Meanwhile, contemporary hominins such as P. boisei retained more primitive anatomy overall, supporting the hypothesis that the intensification of making and using toolsdrather than the origin of these behaviorsddrove much of the evolution of derived modern human hand anatomy. Multiple sources of evidence show that the Paranthropus upper limb was adapted to strong loads. Very thick cortical bone throughout the upper limb skeleton, features of the elbow joint, and notably curved phalanges provide compelling evidence that P. boisei habitually used its upper limbs with far greater mechanical activity than typically seen in modern humans today. Like its australopith predecessors (Green et al., 2007; Ruff et al., 2016; Susman et al., 1985), P. boisei most likely climbed trees as part of its typical activity regime. A more laterally oriented glenohumeral joint implies that arboreal locomotion may have been a smaller part of the positional repertoire of P. boisei than has been suggested for taxa like A. afarensis, but apart from this, KNM-ER 47000 exhibits a suite of features that are consistent with arboreal climbing. The discovery of the upper limb and hand of P. boisei provides conclusive evidence of divergence among early Pleistocene hominins in postcranial and craniodental anatomy. Such divergence could be the product of competitive displacement in adaptive niche space between contemporary, potentially cohabiting, species that are similar in so many respects. Much remains to be learned about the postcranial anatomy of P. boisei, various species of earliest Homo, when and why they went extinct, and how and why H. erectus persisted and thrived. Acknowledgments This research would not have been possible without the support and assistance of the Ileret Community. Research was facilitated by a research permit issued to the collaborative project that included the National Museums of Kenya and colleagues at The George Washington University issued by the National Council of Science and Technology. We are also grateful for the continued enthusiasm and support of the students of the Koobi Fora Field School. We thank Hannah Carter-Menn and Kathryn C. Braun, in particular, for their roles in organizing the fieldwork between 2005 and 2010, during which many of the fossil specimens were recovered. We are grateful to Sarah Elton, Tracy Kivell, and an anonymous reviewer for their comments on an earlier draft of this manuscript. We thank J. Michael Plavcan, Carol Ward, Manuel Dominguez-Rodrigo, Fabio Di Vincenzo, Steve Churchill, Tracy Kivell, Stephanie Melillo, Susan Larson, and William Kimbel who provided casts, three-dimensional surface scans, and data of fossil specimens for comparison with the KNM-ER 47000 specimens. This work was supported by the National Science Foundation (BCS-0647557, -0924476, -1128170, -1515054; DGE-9987590, -0801634), the Humboldt Foundation, the George Washington University Facilitating Fund, and grants from the Wenner-Gren Foundation and the Leakey Foundation (to J.M. Plavcan and C.V. Ward). 20 B.G. Richmond et al. / Journal of Human Evolution 141 (2020) 102727 References Aiello, L.C., Wood, B.A., Key, C., Lewis, M., 1999. Morphological and taxonomic affinities of the Olduvai Ulna (OH36). American Journal of Physical Anthropology 109, 89e110. Bamford, M.K., 2017. Pleistocene fossil woods from the Okote Member, site FwJj 14 in the Ileret region, Koobi Fora Formation, northern Kenya. Journal of Human Evolution 112, 134e147. Behrensmeyer, A.K., 1978. Taphonomic and ecologic information from bone weathering. Paleobiology 4, 150e162. Bennett, M.R., Harris, J.W., Richmond, B.G., Braun, D.R., Mbua, E., Kiura, P., Olago, D., Kibunjia, M., Omuombo, C., Behrensmeyer, A.K., Huddart, D., Gonzalez, S., 2009. Early hominin foot morphology based on 1.5-million-year-old footprints from Ileret, Kenya. Science 323, 1197e1201. Bookstein, F.L., 1997. Morphometric tools for landmark data: geometry and biology. Cambridge University Press. Braun, D.R., Aldeias, V., Archer, W., Arrowsmith, J.R., Baraki, N., Campisano, C.J., Deino, A.L., DiMaggio, E.N., Dupont-Nivet, G., Engda, B., Feary, D.A., Garello, D.I., Kerfelew, Z., McPherron, S.P., Patterson, D.B., Reeves, J.S., Thompson, J.C., Reed, K.E., 2019. Earliest known Oldowan artifacts at >2.58 Ma from Ledi-Geraru, Ethiopia, highlight early technological diversity. Proceedings of the National Academy of Sciences of the United States of America 116, 11712e11717. Bush, M.E., Lovejoy, C.O., Johanson, D.C., Coppens, Y., 1982. Hominid carpal, metacarpal, and phalangeal bones recovered from the Hadar Formation: 1974-1977 collections. American Journal of Physical Anthropology 57, 651e677. Churchill, S.E., Holliday, T.W., Carlson, K.J., Jashashvili, T., Macias, M.E., Mathews, S., Sparling, T.L., Schmid, P., Ruiter, D.J. de, Berger, L.R., 2013. The upper limb of Australopithecus sediba. Science 340, 1233477. Constantino, P., Wood, B., 2007. The evolution of Zinjanthropus boisei. Evolutionary Anthropology 16, 49e62. Dingwall, H.L., Hatala, K.G., Wunderlich, R.E., Richmond, B.G., 2013. Hominin stature, body mass, and walking speed estimates based on 1.5 million-year-old fossil footprints at Ileret, Kenya. Journal of Human Evolution 64, 556e568. Diogo, R., Richmond, B.G., Wood, B., 2012. Evolution and homologies of primate and modern human hand and forearm muscles, with notes on thumb movements and tool use. Journal of Human Evolution 63, 64e78. Domínguez-Rodrigo, M., Pickering, T.R., Baquedano, E., Mabulla, A., Mark, D.F., rezMusiba, C., Bunn, H.T., Uribelarrea, D., Smith, V., Diez-Martin, F., Pe lez, A., S Gonza anchez, P., Santonja, M., Barboni, D., Gidna, A., Ashley, G., Yravedra, J., Heaton, J.L., Arriaza, M.C., 2013. First Partial skeleton of a 1.34million-year-old Paranthropus boisei from Bed II, Olduvai Gorge, Tanzania. PLoS ONE 8 e80347. Gower, J.C., 1975. Generalized procrustes analysis. Psychometrika 40, 33e51. Green, D.J., Gordon, A.D., Richmond, B.G., 2007. Limb-size proportions in Australopithecus afarensis and Australopithecus africanus. Journal of Human Evolution 52, 187e200. Green, D.J., Gordon, A.D., 2008. Metacarpal proportions in Australopithecus africanus. Journal of Human Evolution 54, 705e719. Green, D.J., Alemseged, Z., 2012. Australopithecus afarensis scapular ontogeny, function, and the role of climbing in human evolution. Science 338, 514e517. Green, D.J., 2013. Ontogeny of the hominoid scapula: the influence of locomotion on morphology. American Journal of Physical Anthropology 152, 239e260. Green, D.J., Chirchir, H., Mbua, E., Harris, J.W.K., Braun, D.R., Griffin, N.L., Richmond, B.G., 2018. Scapular anatomy of Paranthropus boisei from Ileret, Kenya. Journal of Human Evolution 125, 181e192. Gunz, P., Mitteroecker, P., Neubauer, S., Weber, G.W., Bookstein, F.L., 2009. Principles for the virtual reconstruction of hominin crania. Journal of Human Evolution 57, 48e62. €s, X., Harmand, S., Lewis, J.E., Feibel, C.S., Lepre, C.J., Prat, S., Lenoble, A., Boe ment, S., Daver, G., Brugal, J., Quinn, R.L., Brenet, M., Arroyo, A., Taylor, N., Cle Leakey, L., Mortlock, R.A., Wright, J.D., Lokorodi, S., Kirwa, C., Kent, D.V., Roche, H., 2015. 3.3-million-year-old stone tools from Lomekwi 3, West Turkana, Kenya. Nature 521, 310e315. Hatala, K.G., Roach, N.T., Ostrofsky, K.R., Wunderlich, R.E., Dingwall, H.L., Villmoare, B.A., Green, D.J., Harris, J.W.K., Braun, D.R., Richmond, B.G., 2016. Footprints reveal direct evidence of group behavior and locomotion in Homo erectus. Scientific Reports 6, 28766. Hatala, K.G., Roach, N.T., Ostrofsky, K.R., Wunderlich, R.E., Dingwall, H.L., Villmoare, B.A., Green, D.J., Braun, D.R., Harris, J.W.K., Behrensmeyer, A.K., Richmond, B.G., 2017. Hominin track assemblages from Okote Member deposits near Ileret, Kenya, and their implications for understanding fossil hominin paleobiology at 1.5 Ma. Journal of Human Evolution 112, 93e104. Ihaka, R., Gentleman, R., 1996. R: a language for data analysis and graphics. Journal of Computational and Graphical Statistics 5, 299e314. Johanson, D.C., Lovejoy, C.O., Kimbel, W.H., White, T.D., Ward, S.C., Bush, M.E., Latimer, B.M., Coppens, Y., 1982. Morphology of the Pliocene partial hominid skeleton (A.L. 288-1) from the Hadar Formation, Ethiopia. American Journal of Physical Anthropology 57, 403e451. Kimbel, W.K., Johanson, D.C., Rak, Y., 1994. The first skull and other new discoveries of Australopithecus afarensis at Hadar, Ethiopia. Nature 368, 449e451. Kivell, T.L., Kibii, J.M., Churchill, S.E., Schmid, P., Berger, L.R., 2011. Australopithecus sediba hand demonstrates mosaic evolution of locomotor and manipulative abilities. Science 333, 1411e1417. Kivell, T.L., Deane, A.S., Tocheri, M.W., Orr, C.M., Schmid, P., Hawks, J., Berger, L.R., Churchill, S.E., 2015. The hand of Homo naledi. Nature Communications 6, 8431. Lague, M.R., Jungers, W.L., 1996. Morphometric variation in Plio-Pleistocene hominid distal humeri. American Journal of Physical Anthropology 101, 401e427. Lague, M.R., 2015. Taxonomic identification of Lower Pleistocene fossil hominins based on distal humeral diaphyseal cross-sectional shape. PeerJ 3 e1084. Lague, M.R., Chirchir, H., Green, D.J., Mbua, E., Harris, J.W.K., Braun, D.R., Griffin, N.L., Richmond, B.G., 2019a. Cross-sectional properties of the humeral diaphysis of Paranthropus boisei: implications for upper limb function. Journal of Human Evolution 126, 51e70. Lague, M.R., Chirchir, H., Green, D.J., Mbua, E., Harris, J.W.K., Braun, D.R., Griffin, N.L., Richmond, B.G., 2019b. Humeral anatomy of the KNM-ER 47000 upper limb skeleton from Ileret, Kenya: implications for taxonomic identification. Journal of Human Evolution 126, 24e38. Larson, S.G., Jungers, W.L., Tocheri, M.W., Orr, C.M., Morwood, M.J., Sutikna, T., Awe, R.D., Djubiantono, T., 2009. Descriptions of the upper limb skeleton of Homo floresiensis. Journal of Human Evolution 57, 555e570. McHenry, H.M., 1973. Early hominid humerus from East Rudolf, Kenya. Science 180, 739e741. McHenry, H.M., Brown, C.C., McHenry, L.J., 2007. Fossil hominin ulnae and the forelimb of Paranthropus. American Journal of Physical Anthropology 134, 209e218. McPherron, S.P., Alemseged, Z., Marean, C.W., Wynn, J.G., Reed, D., Geraads, D., arat, H.A., 2010. Evidence for stone-tool-assisted consumption of Bobe, R., Be animal tissues before 3.39 million years ago at Dikika, Ethiopia. Nature 466, 857e860. Melillo, S.M., 2016. The shoulder girdle of KSD-VP-1/1. In: Haile-Selassie, Y., Su, D.F. (Eds.), The postcranial anatomy of Australopithecus afarensis: new insights from KSD-VP-1/1. Springer, New York, pp. 113e141. Nachman, B., Davis, C.A., Akins, W., Feseha, M., Ketcham, R., Ruff, C., Todd, L., Witzel, A., Kappelman, J., 2018. A new reconstruction of the forearm of A.L. 2881 (“Lucy”) and functional interpretations. PaleoAnthropology A21eA22. Njau, J.K., Blumenschine, R.J., 2006. A diagnosis of crocodile feeding traces on larger mammal bone, with fossil examples from the Plio-Pleistocene Olduvai Basin, Tanzania. Journal of Human Evolution 50, 142e162. Patterson, D.B., Braun, D.R., Behrensmeyer, A.K., Lehmann, S.B., Merritt, S.R., Reeves, J.S., Wood, B.A., Bobe, R., 2017. Landscape scale heterogeneity in the East Turkana ecosystem during the Okote Member (1.56e1.38 Ma). Journal of Human Evolution 112, 148e161. Patterson, D.B., Braun, D.R., Allen, K., Barr, W.A., Behrensmeyer, A.K., Biernat, M., Lehmann, S.B., Maddox, T., Manthi, F.K., Merritt, S.R., Morris, S.E., O'Brian, K., Reeves, J.S., Wood, B.A., Bobe, R., 2019. Comparative isotopic evidence from East Turkana supports a dietary shift within the genus Homo. Nature Ecology and Evolution 3, 1048e1056. Rak, Y., 1983. The Australopithecine Face. Academic Press, New York. Richmond, B.G., 1998. Ontogeny and biomechanics of phalangeal form in primates (Ph.D. Dissertation). SUNY at Stony Brook. Richmond, B.G., Fleagle, J.G., Kappelman, J., Swisher, C.C., 1998. First hominoid from the Miocene of Ethiopia and the evolution of the catarrhine elbow. American Journal of Physical Anthropology 105, 257e277. Richmond, B.G., 2007. Biomechanics of phalangeal curvature. Journal of Human Evolution 53, 678e690. Richmond, B.G., Roach, N.T., Ostrofsky, K.R., 2016. Evolution of the early hominin hand. In: Kivell, T.L., Lemelin, P., Richmond, B.G., Schmitt, D. (Eds.), The Evolution of the Primate Hand: Developments in Primatology: Progress and Prospects. Springer Press, pp. 515e543. Roach, N.T., Hatala, K.G., Ostrofsky, K.R., Villmoare, B., Reeves, J.S., Du, A., Braun, D.R., Harris, J.W.K., Behrensmeyer, A.K., Richmond, B.G., 2016. Pleistocene footprints show intensive use of lake margin habitats by Homo erectus groups. Scientific Reports 6, 26374. Rohlf, F.J., 1999. Shape statistics: procrustes superimpositions and tangent spaces. Journal of Classification 16, 197e223. Rohlf, F.J., 2010. TpsRelw. Stony Brook University, Department of Ecology and Evolution. Rolian, C., Lieberman, D.E., Zermeno, J.P., 2011. Hand biomechanics during simulated stone tool use. Journal of Human Evolution 61, 25e41. Ruff, C., Holt, B., Trinkaus, E., 2006. Who's afraid of the big bad Wolff?: “Wolff's law” and bone functional adaptation. American Journal of Physical Anthropology 129, 484e498. Ruff, C.B., Burgess, M.L., Ketcham, R.A., Kappelman, J., 2016. Limb bone structural proportions and locomotor behavior in A.L. 288-1 (“Lucy”). PLoS ONE 11 e0166095. Schneider, C.A., Rasband, W.S., Eliceiri, K.W., 2012. NIH Image to ImageJ: 25 years of image analysis. Nature Methods 9, 671e675. Semaw, S., Renne, P., Harris, J.W.K., Feibel, C.S., Bernor, R.L., Fesseha, N., Mowbray, K., 1997. 2.5-million-year-old stone tools from Gona, Ethiopia. Nature 385, 333e336. Senut, B., 1981. Humeral outlines in some hominoid primates and in Plio-Pleistocene hominids. American Journal of Physical Anthropology 56, 275e283. Silverman, N.J., Richmond, B.G., Wood, B.A., 2001. Testing the taxonomic integrity of Paranthropus boisei sensu stricto. American Journal of Physical Anthropology 115, 167e178. Smith, A.L., Benazzi, S., Ledogar, J.A., Tamvada, K., Pryor Smith, L.C., Weber, G.W., Spencer, M.A., Lucas, P.W., Michael, S., Shekeban, A., Al-Fadhalah, K., Almusallam, A.S., Dechow, P.C., Grosse, I.R., Ross, C.F., Madden, R.H., B.G. Richmond et al. / Journal of Human Evolution 141 (2020) 102727 Richmond, B.G., Wright, B.W., Wang, Q., Byron, C., Slice, D.E., Wood, S., Dzialo, C., Berthaume, M.A., van Casteren, A., Strait, D.S., 2015. The feeding biomechanics and dietary ecology of Paranthropus boisei. The Anatomical Record 298, 145e167. Spoor, F., Leakey, M.G., Gathogo, P.N., Brown, F.H., Anton, S.C., McDougall, I., Kiarie, C., Manthi, F.K., Leakey, L.N., 2007. Implications of new early Homo fossils from Ileret, east of Lake Turkana, Kenya. Nature 448, 688e691. Susman, R.L., Stern, J.T., Jungers, W.L., 1985. Locomotor adaptations in the Hadar hominids. In: Delson, E. (Ed.), Ancestors: The Hard Evidence. Alan R .Liss, Inc., New York, pp. 184e192. Susman, R.L., de Ruiter, D., Brain, C., 2001. Recently identified postcranial remains of Paranthropus and early Homo from Swartkrans Cave, South Africa. Journal of Human Evolution 41, 607e629. Tobias, P.V., 1967. Olduvai Gorge. In: The Cranium and Maxillary Dentition of Australopithecus (Zinjanthropus) boisei, vol. 2. Cambridge University Press, Cambridge. Tocheri, M.W., Orr, C.M., Jacofsky, M.C., Marzke, M.W., 2008. The evolutionary history of the hominin hand since the last common ancestor of Pan and Homo. Journal of Anatomy 212, 544e562. 21 Trinkaus, E., 1983. The Shanidar Neandertals. Academic Press, New York. Walker, A.C., Leakey, R.E.F., 1993. The postcranial bones. In: Walker, A.C., Leakey, R.E.F. (Eds.), The Nariokotome Homo erectus skeleton. Harvard University Press, Cambridge, pp. 95e160. Ward, C.V., Tocheri, M.W., Plavcan, J.M., Brown, F.H., Manthi, F.K., 2014. Early Pleistocene third metacarpal from Kenya and the evolution of modern humanlike hand morphology. Proceedings of the National Academy of Sciences of the United States of America 111, 121e124. Williams, E.M., Gordon, A.D., Richmond, B.G., 2012. Hand pressure distribution during Oldowan stone tool production. Journal of Human Evolution 62, 520e532. Williams-Hatala, E.M., Hatala, K.G., Gordon, M., Key, A., Kasper, M., Kivell, T.L., 2018. The manual pressures of stone tool behaviors and their implications for the evolution of the human hand. Journal of Human Evolution 119, 14e26. Zelditch, M.L., Swiderski, D.L., Sheets, H.D., 2012. Geometric morphometrics for biologists: a primer. Academic Press.