ARTICLE RESEARCH b 1.00 Figure 3 Box-and-whisker plots of 0.50 1.80 0.95 pedalelement comparative ratios in 0.45 工王 cercopithecines,colobines 1.60 T 0.90 0.40 0.85 chimpanzees,gorillas,humans and 1.40 T 0.80 fossil specimens.(See 0.35 Z9S MIS S6S MIS 工 0.75 Supplementary Table 6 for 0.30 0.70 taxonomic composition.)Whisker 0.25 0.65 lines indicate maximum and 0.80 0.60 minimum values.a,Base height 0.20 (PDP)of the first metatarsal to its 0.60 0.55 0.15 0.50 length(L).b,Base height of the first nae a s metatarsal to base height of the rcopithecinae P.troglodytes Colobinae piens H. Colobinae Cercopithecinae G.gorilla P.troglodytes H.sapiens second metatarsal.c,Hallucal length to the second metatarsal length d,Hallucal length to the fourth metatarsal length.e,Base height of 1.00 e the second metatarsal to its length. 0.95 125 f,Length of the second metatarsal to 0.45 0.90 1.20 the length of the fourth metatarsal. 0.85 0.40 0.35 1.15 Measurements of the South African 0.80 and Miocene hominoids were taken 1.10 0.75 0.30 from refs 10 and 17,respectively. 1.05 0.70 0.25 0.65 0.20 T直 68 MIS 1.00 0.60 0.95 白 0.55 0.15 y 0.90 0.50 0.10 0.85 Cercopithecinae P.troglodytes Colobinae G.gorilla H.sapiens P.troglodytes Colobinae Cercopithecinae G.gorilla H.sapiens Cercopithecinae P.troglodytes Colobinae H.sapiens Its dorsoproximal articular margin is continuous and it does not hallucal segment is relatively short,falling within the ranges of exhibit the 'nonsubchondral isthmus'described in Ar.ramidus'. the African apes(Fig.3c,d)and outside the range for anatomically A simple ratio comparing the length of the first metatarsal to the modern humans(Supplementary Table 2).However,its tall hallucal lengths of the second and fourth metatarsals demonstrates that the base,relative to the shorter bases of the associated metatarsals,indi- cates that the BRT foot had a transverse arch more developed than in 3.0 apes and falls in this ratio at the higher range for anatomically modern humans (Fig.3a). ★ The hallucal proximal phalanx is essentially complete and,when 品 combined with its associated metatarsal,further confirms that the 2.5 hallucal ray is relatively short.A ratio formed between the combined ◆★ lengths of the first metatarsal and its associated proximal phalanx (MT1 PP1)and the same elements from the second ray (MT2 PP2)demonstrates that anatomically modern humans with their 2.0 elongated halluces are notably distinct.BRT-VP-2/73c falls within the ranges of apes and monkeys,indicating that the foot had a rela- 9 tively short,abductable great toe(Supplementary Fig.3).This ratio also confirms that the BRT-VP-2/73 hallucal ray was not used during 1.5 a human-like toe-off in the terminal phase of the gait cycle.However, △Cercopithecine the degree of its proximal joint canting(97)is lower than in the OColobine second ray (100),which is a condition seen in humans,whereas 公出 Gorilla the opposite is the case in chimpanzeests(Supplementary Fig.4a). 1.0 ▣Pan The second ray is represented by a metatarsal (BRT-VP-2/73b),a q4 0 大Human proximal phalanx(BRT-VP-2/73e)and an intermediate phalanx(BRT- ◆BRT-VP-2/73 VP-2/73h).BRT-VP-2/73b is a well-preserved second metatarsal.The 0 proximal base is triangular in outline.In lateral view,the base is 0.5 slightly rounded in profile (distally directed concavity)and the 0.5 1.0 1.5 2.0 2.5 3.0 3.5 shaft is longitudinally curved (Fig.2c).Relative to the bone's overall P℃1：41.2% length the dorsoplantar basal height is compressed,falling below the average for Pan,Gorilla,anatomically modern humans(Fig.3e),and Figure 4 Principal component analysis(PCA)of metatarsal ratios.Both the single reported Ar.ramidus sample (see supplementary figure 4 PCI and PC2 for 11 metatarsal ratios(descriptions of the ratios are provided in in ref.11).The dorsum of the BRT-VP-2/73 base does not exhibit Supplementary Table 1)discriminate anatomically modern humans and apes the two 'chondral invaginations'described for Ar.ramidus!. from monkeys on the one hand and chimpanzees from anatomically modern humans and gorillas on the other.BRT-VP-2/73 falls in the human/gorilla Torsion along the shaft results in the long axis of the articular head cluster.Both components are heavily influenced by ratios 6,9 and 10,which are being oriented about 23 medially from the dorsoplantar axis of the all associated exclusively with dimensions of the hallux(see Supplementary base.This torsion towards the hallux is on average less than that seen in Information for further discussion). Pan and Gorilla,but significantly more than that seen in anatomically 29 MARCH 2012 VOL 483 NATURE 567 2012 Macmillan Publishers Limited.All rights reservedIts dorsoproximal articular margin is continuous and it does not exhibit the ‘nonsubchondral isthmus’ described in Ar. ramidus11. A simple ratio comparing the length of the first metatarsal to the lengths of the second and fourth metatarsals demonstrates that the hallucal segment is relatively short, falling within the ranges of the African apes (Fig. 3c, d) and outside the range for anatomically modern humans (Supplementary Table 2). However, its tall hallucal base, relative to the shorter bases of the associated metatarsals, indi￾cates that the BRT foot had a transverse arch more developed than in apes and falls in this ratio at the higher range for anatomically modern humans (Fig. 3a). The hallucal proximal phalanx is essentially complete and, when combined with its associated metatarsal, further confirms that the hallucal ray is relatively short. A ratio formed between the combined lengths of the first metatarsal and its associated proximal phalanx (MT1 1 PP1) and the same elements from the second ray (MT2 1 PP2) demonstrates that anatomically modern humans with their elongated halluces are notably distinct. BRT-VP-2/73c falls within the ranges of apes and monkeys, indicating that the foot had a rela￾tively short, abductable great toe (Supplementary Fig. 3). This ratio also confirms that the BRT-VP-2/73 hallucal ray was not used during a human-like toe-off in the terminal phase of the gait cycle. However, the degree of its proximal joint canting (97u) is lower than in the second ray (100u), which is a condition seen in humans, whereas the opposite is the case in chimpanzees15 (Supplementary Fig. 4a). The second ray is represented by a metatarsal (BRT-VP-2/73b), a proximal phalanx (BRT-VP-2/73e) and an intermediate phalanx (BRT￾VP-2/73h). BRT-VP-2/73b is a well-preserved second metatarsal. The proximal base is triangular in outline. In lateral view, the base is slightly rounded in profile (distally directed concavity) and the shaft is longitudinally curved (Fig. 2c). Relative to the bone’s overall length the dorsoplantar basal height is compressed, falling below the average for Pan, Gorilla, anatomically modern humans (Fig. 3e), and the single reported Ar. ramidus sample (see supplementary figure 4 in ref. 11). The dorsum of the BRT-VP-2/73 base does not exhibit the two ‘chondral invaginations’ described for Ar. ramidus11. Torsion along the shaft results in the long axis of the articular head being oriented about 23u medially from the dorsoplantar axis of the base. This torsion towards the hallux is on average less than that seen in Pan and Gorilla, but significantly more than that seen in anatomically BRT-VP-2/73 StW 562 StW 595 SKx 5017 BRT-VP-2/73 KNM-RU 2036 BRT-VP-2/73 KNM-RU 2036 BRT-VP-2/73 KNM-RU 2036 BRT-VP-2/73 StW 89 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 BRT-VP-2/73 Cercopithecinae Colobinae P. troglodytes G. gorilla H. sapiens MT1PDP/MT1L 1.80 1.60 1.40 1.20 1.00 0.80 0.60 Cercopithecinae Colobinae P. troglodytes G. gorilla H. sapiens MT1PDP/MT2PDP x 100 1.00 0.95 0.90 0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 Cercopithecinae Colobinae P. troglodytes G. gorilla H. sapiens 1.00 0.95 0.90 0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 MT1L/MT2L 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.45 1.20 1.15 1.10 1.05 1.00 0.95 0.90 0.85 MT2L/MT4L MT2PDP/MT2L MT1L/MT4L Cercopithecinae Colobinae P. troglodytes G. gorilla H. sapiens Cercopithecinae Colobinae P. troglodytes G. gorilla H. sapiens Cercopithecinae Colobinae P. troglodytes G. gorilla H. sapiens 1.25 a bc d e f Figure 3 | Box-and-whisker plots of pedal element comparative ratios in cercopithecines, colobines, chimpanzees, gorillas, humans and fossil specimens. (See Supplementary Table 6 for taxonomic composition.) Whisker lines indicate maximum and minimum values. a, Base height (PDP) of the first metatarsal to its length (L). b, Base height of the first metatarsal to base height of the second metatarsal. c, Hallucal length to the second metatarsal length. d, Hallucal length to the fourth metatarsal length. e, Base height of the second metatarsal to its length. f, Length of the second metatarsal to the length of the fourth metatarsal. Measurements of the South African and Miocene hominoids were taken from refs 10 and 17, respectively. PC 1: 41.2% 0.5 1.0 1.5 2.0 2.5 3.0 3.5 PC 2: 36.2% 3.0 2.5 2.0 1.5 1.0 0.5 BRT-VP-2/73 Human Pan Gorilla Colobine Cercopithecine Figure 4 | Principal component analysis (PCA) of metatarsal ratios. Both PC1 and PC2 for 11 metatarsal ratios (descriptions of the ratios are provided in Supplementary Table 1) discriminate anatomically modern humans and apes from monkeys on the one hand and chimpanzees from anatomically modern humans and gorillas on the other. BRT-VP-2/73 falls in the human/gorilla cluster. Both components are heavily influenced by ratios 6, 9 and 10, which are all associated exclusively with dimensions of the hallux (see Supplementary Information for further discussion). ARTICLE RESEARCH 29 MARCH 2012 | VOL 483 | NATURE | 567 ©2012 Macmillan Publishers Limited. All rights reserved