LETTER LETTER Reply to Melillo: Woranso-Mille is consistent with an australopithecine shoulder intermediate between African apes and Homo

In reference to our recent paper (1), Melillo (2) said, Melillo offers that “...the magnitude of the continued, but diminished, importance of makes three claims: (i) the adult Woranso- shapedifferencebetweenthetwofossilsdoes arboreality in australopithecines. Mille (KSD-VP-1/1) scapula, which we not exceed the level observable in living spe- did not consider in our analyses, suggests cies” (2). Consequently, whereas including a Nathan M. Younga,1, Terence D. Australopithecus afarensis shoulders may be more complete Woranso-Mille would inevita- Capellinib,c, Neil T. Roachb, more derived than we report; (ii)ourrecon- bly increase the morphospace occupied by and Zeresenay Alemsegedd structions indicate homoplasy, consistent Dikika, the australopithecine shoulder is still aDepartment of Orthopaedic Surgery, with an “ape convergence” model; and (iii) less derived than Homo and modern humans. University of California, San Francisco, CA Australopithecus shoulder shape is best ex- Second, support for an African ape model 94110; bDepartment of Human Evolutionary “ ” plained by committed terrestriality and tool is not contingent upon the absence of ho- Biology, Harvard University, Cambridge, use and not a trade-off with arboreal efficiency. moplasy. As we discuss in our paper, both MA 02138; cBroad Institute of MIT and None of these claims alter the findings of our models predict that chimpanzee/gibbon and Harvard University, Cambridge, MA 02142; paper (1), that the australopithecine shoulder is human/orangutan similarities in spine orien- d intermediate between African apes and Homo, tation convergently evolved (1). Instead, we and Department of Anthropology, and that the hominin shoulder underwent based our conclusion on the well-established California Academy of Sciences, San a slow, sustained evolutionary transformation principle that the simplest explanation is Francisco, CA 94118 from an African ape-like last common ances- preferred. Because the ape convergence tor to modern humans. model posits living hominoids evolved from 1 Young NM, Capellini TD, Roach NT, Alemseged Z (2015) Fossil First, the Dikika juvenile is currently the a more primitive shared ancestral morpho- hominin shoulders support an African ape-like last common ancestor best-preserved example of A. afarensis shoul- type, similar blade shapes would have had of humans and chimpanzees. Proc Natl Acad Sci USA 112(38): derbladeanatomy(3),anditsadultshape to independently evolve five times: once 11829–11834. 2 Melillo SM (2015) An alternative interpretation of the can be modeled using conservative assump- each between gorillas, chimpanzees/bono- Australopithecus scapula. Proc Natl Acad Sci USA, 10.1073/ tions about growth (1, 4). As Melillo notes (2), bos, and humans, and twice between gibbons pnas.1520902112. 3 Green DJ, Alemseged Z (2012) Australopithecus afarensis scapular blade shape has a strong phylogenetic signal and orangutans. In contrast, the African ape ontogeny, function, and the role of climbing in human evolution. (4), yet the Woranso-Mille scapula is missing model predicts one event to evolve the African Science 338(6106):514–517. critical portions of the inferior angle and ape blade shape from a more primitive one 4 Young NM (2008) A comparison of the ontogeny of shape variation in the anthropoid scapula: Functional and phylogenetic supraspinous fossa (5), limiting its utility for shared by Asian apes, a much more parsimo- signal. Am J Phys Anthropol 136(3):247–264. ancestral state reconstruction. Although diffi- nious explanation. 5 Haile-Selassie Y, et al. (2010) An early Australopithecus afarensis cult to discern from the two published pic- Finally, we explicitly argue for a scenario in postcranium from Woranso-Mille, Ethiopia. Proc Natl Acad Sci USA 107(27):12121–12126. tures, spine orientation in Woranso-Mille which both reduced reliance on arboreality 6 Young NM (2006) Function, ontogeny and canalization of shape may be somewhat more lateralized than and selection on more lateralized activities, variance in the primate scapula. J Anat 209(5):623–636. Dikika, but it also overlaps with African apes such as tool-use in australopithecines and in a number of other metrics (5). Such differ- throwing in later Homo, served as sustained Author contributions: N.M.Y., T.D.C., N.T.R., and Z.A. wrote ences between Dikika and Woranso-Mille are selective forces acting on shoulder shape the paper. not surprising, because hominoid scapulae are (1). We do not dispute the importance of The authors declare no conflict of interest. known to be significantly more variable in terrestrial bipedality, but simply note that 1To whom correspondence should be addressed. Email: nathan.m. shape compared with other primates (6). That the slow pace of scapular evolution supports [email protected].

www.pnas.org/cgi/doi/10.1073/pnas.1521824112 PNAS Early Edition | 1of1