300 words) 50 51 StW 573, from Sterkfontein Member 2, dated ca 3.67 Ma, is by far the most complete 52 skeleton of an australopith to date. Joint morphology is in many cases closely matched in 53 available elements of Australopithecus anamensis (eg. proximal and distal tibial and humeral 54 joint-surfaces) and there are also close similarities to features of the scapula, in particular , of 55 KSD-VP-1/1 A. afarensis from Woranso-Mille. The closest similarities are, however, to the 56 partial skeleton of StW 431 from Sterkfontein Member 4. When considered together, both 57 StW 573 and StW 431 express an hip joint morphology quite distinct from that of A. 58 africanus Sts14, and a proximal femur of a presumed A. africanus from Jacovec Cavern at 59 Sterkfontein, StW 598. This, and other evidence presented herein, suggests there are two 60 pelvic girdle morphs at Sterkfontein, supporting Clarke (2013) in his recognition of a second 61 3 species, A. prometheus, containing StW 573 and StW 431. StW 573 is the first hominid 62 skeleton where limb proportions are known unequivocally. It demonstrates that some early 63 hominins, at the time of formation of the Laetoli footprints (3.6 Ma), were large-bodied. with 64 hindlimbs longer than forelimbs. Modelling studies on extant primates indicate that the 65 intermembral index (IMI) of StW 573, low for a non-human great ape, would have 66 substantially enhanced economy of bipedal walking over medium-to-long distances, but that 67 it was still too high for effective walking while load-carrying. It would, however, have 68 somewhat reduced the economy of horizontal climbing, but made Gorilla-like embracing of 69 large tree-trunks less possible. Consideration of both ethnographic evidence from modern 70 indigenous arboreal foragers and modern degeneracy theory cautions against prescriptive 71 interpretations of hand-and foot-function, by confirming that both human-like upright 72 bipedalism and functional capabilities of the hand and foot can be effective in short-distance 73 arboreal locomotion. 74 75 1. Introduction 76 77While it is now largely accepted that there was no phase of terrestrial knucklewalking in 78 hominin evolution (see eg., Dainton and Macho, 1999; Dainton, 2001; Clarke, 2002; Kivell 79 and Schmitt, 2009), and that australopiths show adaptations to both terrestrial bipedalism and 80 arboreal locomotion, there is still no firm consensus on whether the 'arboreal' features of 81 australopith postcrania would have been the subject of positive selection or were selectively 82 neutral anachronisms (Ward 2002(Ward , 2013. The view that the two activities must be 83 substantially mechanically incompatible is still current (see eg., Kappelman et al., 2016). The 84 two alternative paradigms date from extended debates (see eg., Latimer, 1991 versus Stern 85 and Susman, 1991) concerning the significance of the AL-288-1 'Lucy' skeleton of 86 Australopithecus afarensis in 1974. Although some one-third complete, this partial skeleton, 87 and other more recently discovered...