Humeral torsion does not dictate shoulder position, but does inuence throwing speed Neil T. Roach a, b, * , Brian G. Richmond a a Division of Anthropology, American Museum of Natural History, Central Park West at 79th Street, New York, NY 10024, USA b Center for the Advanced Study of Human Paleobiology, The George Washington University, 800 22nd Street NW, Washington, DC 20052, USA article info Article history: Received 24 March 2015 Accepted 19 May 2015 Available online 19 June 2015 1. Introduction A debate has emerged in the last few years over the shape and position of the shoulder in early Homo. That the shoulder joint underwent changes approximately 2 million years ago is not in dispute. A number of newly discovered and relatively complete scapulae show that the orientation of the glenohumeral joint shifted caudally from the more cranial orientation seen in the apes and earlier hominins (Walker and Leakey, 1993; Larson et al., 2007; Lordkipanidze et al., 2007; Haile-Selassie et al., 2010; Green and Alemseged, 2012; Churchill et al., 2013). However, just how modern human-like this caudally rotated shoulder complex is remains less clear. Larson (2007, 2009) has proposed that early Homo possessed a novel, transitional shoulder morphology in which the shoulder joint faced anteriorly. We have proposed that Homo erectus had an essentially modern human-like shoulder complex with a laterally oriented glenohumeral joint (Roach et al., 2013; Roach and Richmond, 2015). Why does this debate matter? These differing reconstructions of the shoulder have important functional impli- cations for a number of crucial behavioral shifts hypothesized to occur at or near the origins of our genus (e.g., reduced climbing behavior, intensication of tool manufacture and use, endurance running, and high speed throwing). Much of this debate has hinged on the length of the clavicle. As the only bony strut attaching the shoulder complex to the torso, clavicle length may be an important factor in determining shoulder position. Along these lines, Larson (2007, 2009) has argued that relatively short clavicles found in early Homo suggest that their shoulders had a more anterior orientation. We tested this idea, but found that the relative clavicle length measure Larson used as data support for her model does not accurately predict shoulder position (Roach and Richmond, 2015). We further concluded that the data better support the presence of modern human-like, laterally facing shoulders dating back to H. erectus or earlier. Larson (2015) does not agree with our conclusions or with the way we present her research. Our paper addressed the role of clavicle length in determining shoulder position and throwing performance (Roach and Richmond, 2015). Larson's reply (2015), Humeral torsion and throwing prociency in early human evolution,does not address clavicle length and does not dispute either the data we present or our methodology. Instead, Larson refocuses her anterior shoulder model on humeral torsion (Fig. 1). She argues that 1) we misrep- resent her anterior shoulder hypothesis, 2) very low humeral tor- sion presents an unresolved problem for our reconstruction of the H. erectus shoulder as modern human-like, 3) low torsion is asso- ciated with higher injury risk in throwers, and 4) very low torsion does not support the capacity for high-speed throwing dating back to H. erectus. We disagree with Larson on all of these points and in this paper present data that strongly support our position. 2. A distinction without a difference We disagree that we misrepresented Larson's work by stating that she proposed an anterior facing shoulder joint was necessary DOI of original article: http://dx.doi.org/10.1016/j.jhevol.2015.03.003. * Corresponding author. E-mail address: nroach@amnh.org (N.T. Roach). Contents lists available at ScienceDirect Journal of Human Evolution journal homepage: www.elsevier.com/locate/jhevol http://dx.doi.org/10.1016/j.jhevol.2015.05.011 0047-2484/© 2015 Elsevier Ltd. All rights reserved. Journal of Human Evolution 85 (2015) 206e211