Appendix S1 - First Partial Skeleton of a 1.34-Million-Year-Old Paranthropus boisei from Bed II, Olduvai Gorge, Tanzania

Supporting information. Figure S1. CT-scan image of AP section of the hominin distal humerus diaphysis (OH 80-10). Although the strong density limits the quality of the image, it clearly shows that the specimen's cortex is thick and its medullary cavity is narrow and full of trabecular bone. Figure S2. Differences between distal humeri attributed to Homo and to Paranthropus from Swartkrans (1) (upper half) and OH 80-10 (lower half). Note the elongated section distally in OH 80-10, above the biepicondylar line, which extends to the red line. Figure S3. Canonical Variable Analysis (CVA) of australopithecine (Australopithecus+Paranthopus) and Homo proximal radial specimens. CVA increases the inter-group sample differences compared to PCA and MDS. A, CVA including SK2045, which is tentatively identified as Homo sp by morphology, in contrast with SK8b (which is associated with dental remains and, thus, more securely identified). However, SK2045 is spatially placed closer to other australopithecine specimens (see main text, Figure 2), indicating that the morphology of its articular surface is ambiguous. B, CVA of the same hominin radial sample removing SK2045. Differences between Homo and australopithecines are increased. OH80-11 seems thus even closer to the latter. The average amount of predictive error in each variable is reduced from 14% in A to 10% in B. Figure S4. Dimensions (AP = anterior-posterior; ML = medial-lateral) of the proximal articular head of hominin radii of different hominins. Measurements for modern humans are from Senut [3] and for Neanderthal and Atapuerca specimens are from Carretero [4]. Abbreviations: Krp, Krapina; Ferr, Ferrasie; SKX, Swartkrans; STW, Sterkfontein; AT, Atapuerca. Bars for modern humans and Neanderthals show maxima and minima. Figure S5. Multiple correspondence analysis (MCA) of OH 80-12 and the femora of Homo erectus. Abbreviations: HTF = hyper-trochanteric fossa; LEGT+lateral extension of the gluteal tuberosity; LAP = linea aspera position. Homo erectus data from [5]. Figure S6. Comparison of OH 80-12 (right) and three representative femora of Homo erectus [5], illustrating different expressions of taxonomically diagnostic criteria. Note OH 80-12's lack of a hyper-trochanteric fossa and its more medially placed linea aspera, which is also more robust than those of H. erectus. Figure S7. CT-scan image of the AP section of OH 80-12 (left) and of diaphysis section (center, right), illustrating the robust thickness of the diaphysis and thickness distribution (second moment of area) according to orientation: A = anterior; P = posterior; L = lateral; M = medial. The femur section is oriented distally (upper) to proximally (lower). The diaphysis section (center,right) is taken at the mid-diaphysis, coinciding with the natural fracture as seen on the left scan. A diaphysis section of a the KNM- ER 1808 Homo erectus femur is overlaid (lower right) to show differences. Note the proportionally smaller medullary cavity and more thicker cortex of OH 80-12. Figure S8. OH 80-13 (tibial shaft) compared to a modern human tibia. Image by E. Organista and JL Heaton. Figure S9. Stratigraphic sequence of Upper Bed II, Olduvai Gorge, and location of the BK site (right), with isometric reconstruction showing level 4 and a map with the distribution of materials and location of each of the excavated hominin fossils. Figure S10. Map of Olduvai Gorge showing the localities where the tuff IID samples were collected. Figure S11. Photographs of Tuff IID in each of the localities and stratigraphic position of the tuff at each of them. Figure S11. Photographs of Tuff IID in each of the localities and stratigraphic position of the tuff at each of them. Figure S12. Ar/Ar results for the BK sample. Figure S13. Ar/Ar results for the RHC sample. Figure S14. Ar/Ar results for the MCK sample. Figure S15. Ar/Ar results for the JK sample. Figure S16. Compositions of mineral phases in the Olduvai tuff samples. (a) Ab-rich anorthoclase and plagioclase feldspar compositions. (b) Augite and sodic augite compositions. (c) Spinel and rhombohedral (Ti-rich) Fe-Ti oxide compositions. (d) Hornblende compositions. Loading scores (Table S1). Loading scores (Table S2). Loading scores (Table S3). Loading scores (Table S4). Loading scores (Table S5). Table S6. Ar/Ar results are summarised in the table below and ages are calculated using two different sets of standard ages and decay constants. Table S7. Mineralogy and composition of some phases in the Olduvai Gorge (Tanzania) samples. (PDF)