The Ein Gev 1 remains were discovered in situ during excavations at the site of Ein Gev I in 1964, within a burial pit immediately below Layer 3 (Arensburg and Bar-Yosef, 1973 and Bar-Yosef, 1970). The site is on the eastern shore of the Sea of Galilee, ∼ 150 m bsl, on the talus slope of the flanks of the Golan Heights (32° 46′ N, 35° 39′ E). The body had been laid on its right side with the legs strongly flexed at the knees and partially flexed at the hip, such that the left hand and femur were in contact. The grave was within a hut structure dug into the hillside and adjacent to a concentration of lithic and faunal remains (Arensburg and Bar-Yosef, 1973, Bar-Yosef, 1970 and Stekelis and Bar-Yosef, 1965). Layer 3 yielded a rich Kebaran lithic assemblage (Bar-Yosef, 1970 and Stekelis and Bar-Yosef, 1965) and a diverse faunal assemblage dominated by Gazella, Dama mesopotamica and Capra aegagrus (Davis, 1974 and Marom and Bar-Oz, 2008). Charcoal adjacent to the burial in Layer 3 provided a radiocarbon date of 15,700 ± 415 ¹⁴C BP (GrN-5576) or 18,910 ± 465 cal BP (calibrated using www.calpal.de v.1.5). This date places the remains about four millennia younger than the Ohalo 2 burial (Hershkovitz et al., 1995) and probably 7-8 millennia after the ‘Atlitian’ Nahal Ein Gev 1 individual (Arensburg, 1977, Belfer-Cohen and Goring-Morris, 2014 and Belfer-Cohen et al., 2004).

The burial was removed in situ as a block, and subsequently excavated in 1968 by B. Arensburg. The human remains were originally described as “almost complete but seriously damaged and covered by a thick sandy calcareous incrustation” and that “the only complete bones of the postcranial skeleton were a left femur, a left ulna and some segments of the hand and the foot. All the other long bones were present, but lacked epiphyses” (Arensburg and Bar-Yosef, 1973: 202). The original publication provided little further postcranial information, but a limited set of osteometrics (mostly diaphyseal diameters) were provided subsequently (Arensburg, 1977). My analysis of the remains in 1995, as part of a comparative analysis of Southwest Asian late Pleistocene appendicular remains, revealed the potential to obtain substantial additional paleobiological data from the Ein Gev 1 limb bones, which are provided here.

The humeri consist of an essentially complete right diaphysis and distal epiphysis; it retains the distal lesser tubercle, thereby permitting accurate length estimation (Supplementary data, Table S1; Fig. 1). The left humerus is less complete but retains much of the shaft and a partial cubital area. The left ulna provides direct measurement lengths, and both ulnae and radii provide diaphyseal diameters (Fig. 2; Supplementary data, Table S1). The left os coxae preserves a damaged antero-inferior ilium with the femoral head and neck crushed into the acetabulum; the acetabular rim is nonetheless sufficiently complete to provide a height measurement of ∼ 54.5 mm. The left femur retains the intact diaphysis from the lesser trochanter to the distal condyles with minor damage to the distal epiphysis (Fig. 3); it is sufficiently complete to estimate its original lengths (Supplementary data, Table S4). The right femur is eroded anteriorly, twisted proximally and compressed anteroposteriorly distally, and provides little additional data. The tibiae retain a largely complete, but heavily encrusted, left diaphysis and the proximal two-thirds of the right diaphysis (Fig. 4), with the tibial tuberosities preserved for orientation and alignment. The distal right tibia is cemented to its talus and distal fibula. These long bones are joined by scattered and variably complete manual and pedal remains, some of which are cemented to adjacent bones at their articulations (Supplementary data, Tables S3 and S6 to S10).

The only observed pathological lesion on the Ein Gev 1 limb remains is an osteoarthritic degeneration of the left metatarsal 5 proximal epiphysis, which has osseous changes around the cuboid facet. The lesion may be related to a probable healed fracture of the lateral half of the facet. The lesions are minor and would not have affected the individual's gait.

The Ein Gev 1 skeleton has been described as female and 30–40 years of age (Arensburg and Bar-Yosef, 1973), apparently on the basis of the craniofacial remains. It is here considered as a prime-age adult, probably female, and its sex is of concern only with respect to overall body size.

The Ein Gev 1 remains are compared to Upper Paleolithic sensu lato human remains from western Eurasia and North Africa. Note that the Southwest Asian assemblages from ∼ 25 to ∼ 13 ka BP are normally referred to as Epipaleolithic, which is broadly equivalent to the Late Upper Paleolithic in European terminology; penecontemporaneous northern African assemblages are usually subsumed within the Later Stone Age. For convenience, despite its eurocentrism (no more so than “Levant” or “Near East”), the term Upper Paleolithic is employed here to include these circum-Mediterranean Marine Isotope Stage (MIS) 3 and MIS 2 assemblages and samples of their early modern human remains.

Despite the growing sample of Southwest Asian Epipaleolithic human remains (see 1.0), only a few provide postcranial comparative data. The comparative sample therefore includes Southwest Asian human remains from the sites of Mataha, Neve David, Ohalo II and Kharaneh IV, plus the modestly older Nahal Ein Gev I (SWA). Additional data are from the Later Stone Age (mostly late MIS 2) remains from the Nile Valley (NEA) and Northwest Africa (NWA), and Upper Paleolithic remains from Europe. Although Ein Gev 1 is dated to after the LGM, the usual Early/Mid-to-Late Upper Paleolithic boundary (e.g., Holt and Formicola, 2008), and changes have been noted in crania, body proportions and stature in Europe and Northeast Africa across the LGM (e.g., Brewster et al., 2014, Crevecoeur, 2008, Holliday, 1997 and Holt and Formicola, 2008), reflections of appendicular hypertrophy changed little (Sparacello et al., 2017 and Trinkaus, 2015). The European Upper Paleolithic sample is nonetheless separated into earlier (EUP) and later (LUP) Upper Paleolithic samples; the NEA sample includes limited upper limb data from the MIS 3 Nazlet Khater 2, but it is otherwise dominated by the large late MIS 2 Jebel Sahaba and Wadi Halfa samples. The NWA sample includes the MIS 2 Later Stone Age Afalou and Taforalt remains.

The sexes are separate for body size assessments but are otherwise pooled for body proportion and robustness comparisons; the degree of sexual dimorphism in the latter features is unclear (Sparacello et al., 2017, Trinkaus, 2015 and Trinkaus, 2018); previous analyses noting a sexual difference in robustness are biased through the use of ratios.

The complete list of sites providing specimens for which comparative data are available is in Supplementary data, Table S14. The majority of the European and northern African data are available in Shackelford (2005), Crevecoeur (2008), Sparacello et al. (2017) and Trinkaus and Ruff (2012). Linear osteometrics are available through these sources and/or the primary descriptions of the remains. The cross-sectional geometric parameters for the comparative southwest Asian Upper Paleolithic humeri, femora and tibiae, only some of which have been published, are in Supplementary data, Table S13.

The body size and proportions of Ein Gev 1 are assessed using its (albeit estimated) femoral bicondylar length and head diameter (Fig. 5), the former reflecting stature and the latter reflecting body mass (Feldesman et al., 1990 and Ruff et al., 2018). Following Ruff et al. (2018), the femoral head diameter provides a mean body mass estimate of 63.1 kg.

The femoral length of Ein Gev 1 is modest for an Upper Paleolithic human, falling among the smaller of the European LUP specimens and below the ranges of the European EUP and especially the pooled North African sample (Fig. 5A). It is well below the high value for Ohalo 2, yet above the low values for Mataha F-81 and Nahal Ein Gev 1. Its femoral head diameter, however, is moderate for an Upper Paleolithic human (Fig. 5B), among the smaller of the EUP specimens and the LUP males, moderately large for a North African female, but well within the ranges of variation of the comparative samples. Incorporation of the error ranges for the Ein Gev 1 dimensions (Supplementary data, Table S5) would have little effect on these comparative assessments.

The combination of the articular and length dimensions (or “mass-to-stature” proportions) (Fig. 5C) provides a largely complete (and highly significant; Supplementary data, Table S12) separation of the more linear (relatively low mass-to-stature, from narrow trunks and long limbs, as reflected in femoral head to length proportions) Northeast African and earlier European samples from the Northwest Africa and later European samples (see also Holliday, 1997 and Holliday, 2015). Ein Gev 1, along with Mataha F-81 and (to a lesser extent) Ohalo 2, falls with the less linear circum-Mediterranean samples, it is substantially different only from the NEA sample. These stockier (higher mass-to-stature; “mid-latitude”) body proportions, which are evident in terminal Pleistocene/early Holocene Natufian remains (Holliday, 2015), were therefore present close to the LGM in southwest Asia.

It is also possible to estimate the brachial (radio-humeral) index for Ein Gev 1; the mean length estimates for the humerus and radius (Supplementary data, Tables S1 and S3) provide an index of ∼ 71.5, using the mean length estimates minus and plus two standard errors of the humerus and radius respectively raises the brachial index to ∼ 76.7. These values bracket those of 73.2 for Ohalo 2 and 76.6 for Mataha F-81. They are similar to the male and female average values of 72.9 and 74.8 for the Northwest African Taforalt sample (Ferembach, 1962), and modestly lower than the European EUP and LUP means of 77.9 (± 2.3, n = 21) and 78.5 (± 2.7, n = 15). Although recent human brachial indices vary with climate (Holliday, 1999 and Trinkaus, 1981), there is little patterning among them in these Upper Paleolithic samples.

As with most other late Pleistocene humans (Sparacello et al., 2017) and relative to most Holocene humans (Sládek et al., 2016 and Trinkaus et al., 1994), the western Eurasian and Northwest African Upper Paleolithic humeri exhibit an elevated level of diaphyseal asymmetry (Fig. 6); the Northeast African sample has less pronounced asymmetry, producing the modest but minimally significant differences across the samples (Supplementary data, Table S12). The Ein Gev 1 humeral asymmetry is unexceptional at the mid-distal (35%) level, moderately low but still above the EUP and NEA medians at the midshaft (50%) level. The Ohalo 2, Kharaneh 1 and Mataha F-81 humeri are clearly asymmetrical at midshaft, but they are unexceptional more distally.

In this context, the Kharaneh 1 and Mataha F-81 dominant humeral midshafts fall among the majority of the Upper Paleolithic humeri in robustness (Fig. 7A), whereas Ohalo 2 is among the more gracile (predominantly EUP) of them (see also Trinkaus, 2018). The same pattern is evident for the non-dominant humeri (Fig. 7B). The Ein Gev 1 humeri are average, moderately gracile on the right (dominant) side and in the middle of the overall distribution on the left. Although the comparative samples are significantly different with respect to humeral hypertrophy, Ein Gev 1 is unexceptional for any of these samples (Supplementary data, Table S12).

Because midshaft humeral hypertrophy is commonly reflected in an increase in the antero-lateral to postero-medial (or maximum) dimension, the relative maximum versus minimum second moments of area should also reflect general humeral hypertrophy (Fig. 7C and D). There is more scatter in these proportions in the non-dominant side, with little separation of the comparative samples (Supplementary data, Table S12). The comparative SWA humeri fall well within the comparative distributions, but the Ein Gev 1 humeri are both relatively round.

Lower limb hypertrophy can be assessed for the femoral midshaft, along with indirect reflections from femoral and tibial diaphyseal proportions (Fig. 8). In the comparison of femoral midshaft polar moment of area to body mass times length, there are modest differences across the comparative samples, with the LUP and NWA samples being generally more robust. Ein Gev 1, Mataha F-81 and Ohalo 2 all fall in the middle of the overall distribution (Fig. 8A).

There is also little difference in femoral midshaft second moment of area proportions across the comparative samples (Fig. 8C). Ein Gev 1, however, has a pronounced pilaster through midshaft (Fig. 3), which places it (along with Ohalo 2) among the Upper Paleolithic femora with the larger anteroposterior second moments of area relative to their medio-lateral ones (Fig. 8C). It contrasts primarily with the LUP sample. There are more differences across the proximal femur and tibial midshaft, with especially the NEA sample having less prominent femoral gluteal buttresses and less tibial anteroposterior expansion. The Ein Gev 1 proximal femur is unexceptional in these proportions (Fig. 8B). Its tibial midshaft proportions contrast primarily with the NEA sample, but they are similar to those of the other Southwest Asian Epipaleolithic specimens (Fig. 8D). It is possible that the tibial 50% section is modestly too proximal (see 2.3), but any reduction in its I_max would place it more in the middle of the overall distribution.

The distal humeri of Ein Gev 1 are unexceptional for a late Pleistocene human (Fig. 1). They have moderately prominent medial epicondyles and medial flanges to the trochlea. A comparison of the summed medial and lateral pillar thicknesses versus olecranon breadth, a feature which varies through Pleistocene humans (Trinkaus, 2012), provides an index of 97.8 for the Ein Gev 1. Comparative Upper Paleolithic samples provide a wide range of values (European EUP: 107.3 ± 20.9, n = 21; Northeast Africa: 88.3 ± 14.6, n = 45; Northwest Africa: 86.8 ± 15.6, n = 37), such that Ein Gev 1 and Ohalo 2 (89.7) are in the middle of the variation, whereas Nahal Ein Gev 1, with an index of 69.4, is among the specimens with narrower pillars.

The largely complete left ulna and the right radius are notable for their modest degrees of dorsal or lateral curvature and their small interosseous crests (Fig. 2). Their proximal portions are too eroded to provide trochlear notch or radial tuberosity orientations. The right hamate bone exhibits a relatively pronounced hamulus, similar in projection (Ein Gev 1: 9.8 mm; Ohalo 2: 9.3 and 9.1 mm) and length (Ein Gev 1: 9.6 mm; Ohalo 2: 11.2 and 10.0 mm) to those of the larger Ohalo 2 (cf. Trinkaus et al., 2014; Supplementary data, Table S4). The tuberosity of the left trapezium exhibits little projection (1.5 mm), similar to those of Ohalo 2 (2.8 and 1.8 mm). The metacarpal 3 has minimal projection of the styloid process, extending only 0.8 mm from the mid-capitate facet, a pattern which characterizes Ohalo 2 (1.0 and 0.6 mm) and most Pleistocene human third metacarpals (Trinkaus et al., 2014 and Ward et al., 2014).

The right and left pedal remains are variably eroded and cemented to each other (and to the distal tibia/fibula), but they provide discrete trait observations in addition to morphometrics (Supplementary data, Tables S7 to S11). The right talus exhibits anterior extensions of medial malleolar facet and lateral trochlea, along with a lateral squatting facet, indicating habitual hyperdorsiflexion of the talo-crural articulation. The anterior and medial talo-calcaneal facets are fully fused, but there is no evidence of a sulcus tali facet. Naviculo-cuboid facets are present, and the cuboid bone has a distinct facet on the proximal margin of the peroneus longus sulcus. The hallux would have exhibited normal hallux valgus, as reflected in its metatarsal 1 horizontal head angle of 9°. The proximal phalanx 2 has modest diaphyseal diameters (5.9 and 6.0 mm), similar to those from Ohalo 2 and other Upper Paleolithic sites (Trinkaus and Patel, 2016); reflecting habitual use of footwear (Trinkaus, 2005).