Journal of Human Evolution 57 (2009) 248–259 Contents lists available at ScienceDirect Journal of Human Evolution journal homepage: www.elsevier.com/locate/jhevol Paleolithic hominin remains from Eshkaft-e Gavi (southern Zagros Mountains, Iran): description, affinities, and evidence for butchery Jeremiah E. Scott*, Curtis W. Marean School of Human Evolution and Social Change, Institute of Human Origins, Arizona State University, Tempe, AZ 85287-4101, USA article info abstract Article history: Eshkaft-e Gavi is a cave located in the southern Zagros Mountains of Iran and is one of the few Received 19 July 2008 archaeological sites in the region to preserve both Middle Paleolithic and Upper Paleolithic occupations. Accepted 25 May 2009 Excavation of the site in the 1970s yielded an assemblage of lithic and faunal remains, including ten hominin specimens: a mandibular molar, four cranial fragments, a clavicular diaphysis, the proximal half Keywords: of a metacarpal, a fragment of os coxa, the proximal diaphysis of a juvenile femur, and a patella. The Cannibalism bones derive from a minimum of four individuals, including two juveniles. Although many of these Epi-Paleolithic remains could be Epi-Paleolithic in age, one of the juvenile specimensdthe mandibular molardoccurs at Modern humans Upper Paleolithic the base of the cave’s Upper Paleolithic sequence. The remains are very fragmentary, but those that preserve diagnostic morphology indicate that they represent modern humans. The molar is taxonomi- cally diagnostic, thus confirming the association of the Aurignacian-like Baradostian Industry with modern humans. Four of the specimensda piece of frontal bone, the clavicle, the juvenile femur, and the patelladdisplay clear evidence for intentional butchery in the form of stone-tool cutmarks. These cut- marked specimens, along with a fragment of parietal bone, are also burned. Although this evidence is consistent with cannibalism, the small sample makes it difficult to say whether or not the individuals represented by the hominin remains were butchered and cooked for consumption. Nevertheless, the cutmarked Eshkaft-e Gavi specimens add to a growing sample of hominin remains extending back into the Plio-Pleistocene that display evidence of intentional defleshing. Ó 2009 Elsevier Ltd. All rights reserved. Introduction The character of the lithic sequence of the Zagros Mountains is now reasonably well known based on systematic descriptions of the The Zagros Mountains of Iran and Iraq were the focus of intensive materials from Warwasi and Bisitun (Dibble, 1984; Dibble and archaeological research in the middle twentieth century (Garrod, Holdaway, 1990, 1993; Olszewski, 1993a,b; Olszewski and Dibble, 1930; Coon, 1951; Braidwood and Howe, 1960; Braidwood et al., 1994), Kobeh (Lindly, 2005), and Kunji (Speth, 1971; Baumler and 1961; Young and Smith, 1966; Hole and Flannery, 1967; Solecki, Speth, 1993), but it is not well dated. The Middle Paleolithic in the 1963, 1971; Piperno, 1972, 1974; Rosenberg, 1985), after which Zagros is typically classified as the Zagros Mousterian, and to date, political events in Iran and instability in the Iraqi Zagros curtailed there is no dramatic variation found within it that justifies any research by scientists from the United States and western Europe. formal divisions. All radiocarbon dates for the Zagros Mousterian Field research is now once again commencing (Roustaei et al., 2002; have so far been infinite dates, and thus it is likely that the Zagros Heydari, 2007). While much of the initial effort targeted the tran- Mousterian is older than 35 ka. The succeeding Upper Paleolithic in sition to food production, Paleolithic studies benefited enormously. the Zagros is represented by what has been traditionally called the The highly dissected limestones of the Zagros (Oberlander, 1965)are Baradostian. Given its similarities with Aurignacian assemblages, riddled with caves and rockshelters, many of which were discovered Olszewski and Dibble (1994) proposed that the Baradostian be to be rich Paleolithic sites. The result is that these projects produced renamed the Zagros Aurignacian. a large sample of Paleolithic material from numerous sites. The only reasonably complete hominin remains found in asso- ciation with the Zagros Mousterian are the Neandertals at Shanidar in Iraq (Trinkaus, 1983). Although hominin fragments were said to have been found at the Iranian Middle Paleolithic sites of Bisitun * Corresponding author. E-mail addresses: [email protected] (J.E. Scott), [email protected] and Tamtama (Coon, 1951), only the Bisitun radius is hominin, most (C.W. Marean). likely representing a Neandertal (Trinkaus, 2006; Trinkaus and 0047-2484/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.jhevol.2009.05.012 J.E. Scott, C.W. Marean / Journal of Human Evolution 57 (2009) 248–259 249 Biglari, 2006). Upper Paleolithic human remains are more rare in shows that all but the tiniest fragments below 1 cm were captured this region. Hole and Flannery (1967) mentioned human remains by the sieving operation, and thus the sample is not strongly biased from Gar Arjeneh, but these are undescribed, and Trinkaus et al. to larger fragments and is unlikely to be missing identifiable frag- (2008) recently reported the recovery of an Upper Paleolithic ments through selective retention (Marean et al., 2004). human premolar at Wezmeh Cave, Iran. Here we describe ten The cave has a rough figure-eight shape, with each section of the Upper Paleolithic and Epi-Paleolithic hominin remains from the site eight being about 15 m  7m(Fig. 3). Roof collapse from blasting in of Eshkaft-e Gavi, Iran. In 1991, CWM initiated a study of several the roof rock occurred at the opening of the cave and in the middle of Zagros Paleolithic faunal assemblages, including the sites of Bisitun, cave near the junction of the two chambers, leaving piles of rubble in Eshkaft-e Gavi, Kobeh, Kunji, and Tamtama (Marean, 1998; Marean both places. Excavations were placed to avoid this rubble and any and Kim, 1998; Marean and Cleghorn, 2003; Cleghorn and Marean, disturbance resulting from modern activities, as well as to sample 2004). The analysis of the Eshkaft-e Gavi faunal assemblage both the front and rear chambers. The rear test excavation (Opera- resulted in the rediscovery of most of the hominin remains tion A, including Squares G, M, and N) was 3 m2 and was excavated to described here, the discovery of several new specimens, and the a depth of 1.5 m. The front test excavation (called Operation B, observation of evidence for butchery on some of those remains. including Squares C and D) was a 2 m2 excavation that reached a depth of 2 m. Excavations were carried out within 10-cm arbitrary Background to the site and excavations levels within natural stratigraphy. The excavators used a lot numbering system in which the smallest individual unit of excava- Eshkaft-e Gavi is a large cave on the Marv Dasht plain on the tion, usually a square/10-cm-level combination, was assigned lower Kur River Valley, about 14 km southwest of Persepolis, a unique ascending lot number, with each operation having its own southernmost Zagros (29 52 19.49 N, 52 44 53.85 E) (Fig. 1). It is set of numbers. The associations of the lots to the strata are indicated approximately 20 m above the plain and faces northeast, and the in Figs. 4 and 5 with shading and connecting lines. Kur River flows just 750 m to the northeast. The site was discovered A schematic of the sequence in Operation A was reconstructed by W. Sumner in 1969 during a survey, and M. Rosenberg investi- from the excavation notes and information provided by Rosenberg gated it in 1978 from Sumner’s lead. He found that substantial (Fig. 4). The excavators noted that the stratigraphy in Operation A disturbance had occurred from mechanical excavation of the talus was not always clear, and it was particularly challenging in the lower slope and partial collapse of the roof from blasting. Fortunately, the part of the sequence where light was very poor and the sediments interior cave preserved an intact deposit, and Rosenberg initiated were rocky and complex. Overall, Operation A had more dramatic excavations for two weeks in July of 1978 (Rosenberg, 1985) as part signs of disturbance than Operation B, most of it being concentrated of the Malyan Project, directed by W. Sumner and R. Dyson. Our near the top of the sequence, but with the lower part likely having description of the excavations and our drawings of the figures and some ancient rodent burrowing. Starting at the bottom, the deposits stratigraphy are from Rosenberg (1985) and unpublished notes have high quantities of limestone rock fall, some Middle Paleolithic made available to us by Rosenberg. artifacts, and fossil bone. The latter deposits are overlain by hard, All of the sediments were sieved, and all of the sieved material yellow clay that is cross-cut with possible rodent burrows that occur was retained. Recovery appears to have been excellent, as there is from roughly 120 cm to 80 cm below datum. Next is a concentration a sample of several thousand small animals (micromammals, of archaeological finds, including lithic artifacts and fauna. The lithic amphibians, reptiles, and birds) currently under study. A histogram artifacts are Upper Paleolithic in form, and probably Baradostian in (Fig. 2) of the large-animal faunal fragments shows that even very affinity. This is followed by a rocky, orange sediment (described as small specimens (less than 2 cm in size) were recovered and that ‘‘soil’’ by Rosenberg), and then finally, at the top, an organic sedi- the distribution is strongly skewed to small fragments. This pattern ment with significant amounts of dung and few archaeological finds. 48 54 60 Tamtama Ke-Aram 1 36 36 Khunik Cave Tal-i Iblis Hunter's Cave Eshkaft-i-Gavi 30 Ghar-i Kar Kobeh Bisitun and Warwasi Harsin Jahrom Kal-i Daoud Hulailan Sites Kunji, Houmian Gar Arjeneh, 54 60 Ghamari CAVE SITES OPEN AIR SITES Figure 1.