Within the southern Levant, the period between 55,000 and 35,000 years ago was marked by significant biocultural and techno-typological changes, including the replacement of Neanderthals by anatomically modern humans and the transition from Middle-to-Upper Paleolithic technologies and material culture. To more fully understand the inter-relationship between these events—and their potential relationships with environmental change—we need more data points on both sides of the transition. While the region preserves a rich record from the late Middle Paleolithic (MP), our knowledge of the earliest stages of the Upper Paleolithic (UP)—known as the Early Upper Paleolithic (EUP, ∼45–30 ka)—remains sparse. With in situ deposits dating to 45–39 ka cal BP (Stutz et al., 2015), Mughr el-Hamamah (MHM, Jordan) helps to fill this gap, offering a key opportunity to expand our understanding of EUP lifeways in the southern Levant.

Faunal remains have an important role to play in these discussions; not only do they provide information on human subsistence behavior, but they also speak to paleoenvironmental conditions and landscape use. The zooarchaeological record is particularly relevant given that several hypotheses regarding the MP–UP transition are rooted in the idea that UP populations exploited a wider range of resources than their MP counterparts. The mechanisms proposed to underlie this dietary expansion are varied, ranging from the development of new forms of hunting technology (e.g., Shea and Sisk, 2010), to the development of a gender-based division of labor (e.g., Kuhn and Stiner, 2006), to shifts in social relationships that deemphasized big game hunting (O'Connell, 2006). Evidence for diet-breadth expansion can take a number of forms, including: 1) among the ungulate remains, a focus on smaller species and/or juveniles (e.g., Broughton and Grayson, 1993; Speth and Clark, 2006); 2) increased processing intensity, reflected in the exploitation of marrow from low yield elements and/or bone grease rendering (e.g., Binford, 1978; Stiner, 2003; Munro and Bar-Oz, 2005; Bar-Oz and Munro, 2007); and 3) small game use, with particular attention given to the exploitation of agile and difficult to capture species such as hares and birds (e.g., Binford, 1968; Flannery, 1969; Stiner et al., 1999; Stiner et al., 2000; Stiner and Munro, 2002).

In the well-studied Mt. Carmel/Western Galilee area, Speth and Clark (2006) identified a cycle of increasing predation impact on the largest ungulates (aurochs and red deer) and intensification in the hunting of medium and smaller ungulates (reflected by increasing frequencies of juveniles) within the late MP at Kebara, patterns they attribute to overhunting. Within the same region, research spanning from the Levantine Aurignacian to the Natufian has documented increasing human predation pressure on ungulates; not only do gazelle come to dominate the record by the late Epipaleolithic, typically accounting for more than 80% of the ungulate fauna, but intensification in the use of gazelle is evidenced in the exploitation of juveniles and the processing of even the smallest marrow bearing bones (Bar-Oz, 2004; Munro, 2004, 2009; Munro and Bar-Oz, 2005; Bar-Oz and Munro, 2007; Stutz et al., 2009; Yeshurun et al., 2014; Munro et al., 2018). While some have proposed that there is a robust, long-term feedback between human predation pressure on ungulate populations and subsistence intensification in the Levant—a process that crosscuts the MP–UP transition (e.g., Speth, 2013)— the limited nature of data from the EUP makes it difficult to assess the relative degree of intensification in ungulate exploitation during that period.

Using data from across the Mediterranean Basin, Stiner and colleagues propose a marked shift in small game use around the time of the MP–UP transition (∼50–40 ka), with a shift in focus from slow moving, slow growing game such as tortoises and shellfish, to agile, fast maturing animals like hares and birds (e.g., Stiner et al., 2000; Stiner and Munro, 2002; Stiner, 2005, 2006). Morin et al. (2019) suggest that this pattern may not be so clear-cut across the entire region, with leporids incorporated into both MP and UP diets in the northwestern Mediterranean. Within the northern Levant, the site of Üçağızlı fits within Stiner's framework, with evidence for dietary expansion occurring as early as 50–42 ka (Stiner, 2009). However, our understanding of the nature and timing of changes in small game use within the southern Levant remains incomplete. Not only is there a rarity of well-preserved faunal assemblages from the earliest millennia of the EUP, but small game has often been incompletely studied at late MP sites in the region. The available data suggest that the focus on agile, fast maturing small game may have emerged after 40 ka, with birds increasing in frequency in the Levantine Aurignacian and small mammals becoming a focus of human subsistence during the Epipaleolithic (e.g., Rabinovich, 2003; Munro, 2009; Yeshurun et al., 2021).

To date, the largest and best documented southern Levantine faunal assemblages from the earliest stages of the UP derive from Manot Cave (Yeshurun et al., 2021) and Kebara Cave (Speth, 2019); however, each of these assemblages has limitations. While Manot preserves fauna from the Early Ahmarian (∼46–42 ka) and the Levantine Aurigancian (∼38–34 ka), the published sample from the Ahmarian is small—the number of identified specimens, or NISP, is 476. Yeshurun et al. (2021) report that relative to Ahmarian, the Aurigancian (NISP = 1066) shows a greater degree of intensification, including the culling of more juveniles and an increased exploitation of small game (especially birds). That said, in comparing the EUP fauna from Manot to broader data from the southern Levant, Yeshurun et al. (2021:16) conclude “if the breadth of the taxonomic spectrum and the proportion of sub-adults and fawns are used as a marker of intensification, then the EUP societies were somewhat more intensified than the preceding Middle Paleolithic (Stiner et al., 1999; Stiner, 2005), but still maintained similar foraging lifeways.”

The EUP at Kebara also includes material associated with the Ahmarian and the Aurignacian, dating as early as ∼47 ka cal BP and 39–36 ka cal BP, respectively (dates from Bar-Yosef et al., 1996; Rebollo et al., 2011). However, because the boundaries between the stratigraphic units associated with the Ahmarian and Aurginacian were unclear, Speth (2019) treats the EUP fauna as one unit for analytical purposes. Furthermore, his analysis is limited to the ungulate fauna, meaning he cannot address small game use. Despite utilizing a combined EUP sample (total NISP = 6461), Speth (2019) finds little difference in ungulate exploitation strategies between the MP and UP at Kebara, and the changes he does document are gradual in nature. However, Yeshurun (2020) expresses some caution about these results. The surface modification data from Kebara (Speth and Tchernov, 2007; Speth, 2019) suggest a greater degree of hyena activity in the UP, which may have elevated the representation of fallow deer relative to gazelle during that period. Per Yeshurun (2020), this could create an appearance of more continuity in animal exploitation across the MP-UP transition than is warranted.

Although somewhat later in time (37.5–34.0 ka cal BP), the EUP fauna from Nahal Rahaf 2 (NR2) is also noteworthy (Marom et al., 2022; Shemer et al., 2023). Associated with the Arkov-Divshon culture, the site, located in the Judean Desert, provides rare insight into human subsistence strategies in the arid southern Levant. The faunal data (NISP = 987) suggest a focus on caprines (identified as Capra cf. Capra ibex) and gazelle, with caprines increasing in frequency throughout the sequence. Marom et al. (2022) propose that the greater focus on caprines may relate to a shift toward drier conditions. Small game (primarily hares) and birds are both rare, each accounting for 2.2% of the total NISP, although Marom et al. (2022) note that the avian assemblage is diverse for its size, including pigeon, partridge, and goose.

Other southern Levantine sites that preserve fauna from the EUP either lack deposits pre-dating the Aurignacian or lack appropriately comparable data. Hayonim Cave, best known for its extensive MP deposits, also preserves material from the UP and the Epipaleolithic (see Rabinovich, 1998; Stiner, 2005 for detailed faunal analyses). However, the earliest UP deposits at the site are associated with the Aurignacian, with 14C dates of ∼29–27 ka (calibrated to ca. 34–31 ka; Belfer-Cohen and Bar-Yosef, 2015). Sefunim Cave also has deposits spanning from the MP through the Epipaleolithic, including extensive Levantine Aurignacian deposits (Shimelmitz et al., 2018). Underlying the Aurignacian is a ‘transitional' MP–UP layer that has not yet been assigned to an archaeological culture; however, that layer was recently assigned an OSL age of 51.0 ± 3.0 ka, suggesting that it pre-dates the onset of the EUP (Slon et al., 2022). Ksar-Akil, which sits at the boundary of the southern and central Levant, preserves a 23-m-long sequence that includes EUP deposits associated with the Ahmarian and the Aurignacian (Douka et al., 2013; Bosch et al., 2015a). However, the EUP layers were excavated in the 1930s and 1940s, using techniques that “were not up to present-day standards” (Bosch et al., 2015a: Supplementary Information). An early report on the fauna did not include taphonomic analysis and provided aggregate taxonomic data that crosscut archaeological cultures (Hooijer, 1961). The EUP fauna was recently reanalyzed; however, publications from that analysis have focused on the molluscan remains (e.g., Bosch et al., 2015a; Bosch et al., 2015b; Bosch et al., 2018, but see the Supplementary Information for Bosch et al., 2015a for a summary discussion of the vertebrate fauna).

With an EUP occupational horizon dating to 45–39 ka cal BP and a large, well-preserved faunal assemblage, MHM thus provides a critical opportunity for new insights into animal exploitation strategies during the EUP. The assemblage also provides the first substantial sample from this time period from the Jordan Valley, joining the late MP site of Amud Cave (ca. 70–55 ka; Rabinovich and Hovers, 2004) as a key source of data on human hunting behavior in the broader Lisan Basin. The goals of this paper are twofold. First, we present the results of taxonomic and taphonomic research at the site, the latter of which is critical in understanding the depositional history of the assemblage. Second, we assess evidence for subsistence intensification, focusing especially on the exploitation of gazelle and the use of small game. The results reported here are part of a broader study integrating zooarchaeological and stable isotope data; later publications will focus more directly on paleoenvironmental reconstruction and landscape use.

Mughr el-Hamamah is located in the western portion of the Ajlun Governate, Jordan (Figure 1). The site comprises five caves that formed on the bank of a minor wadi between the Wadis Rajib (to the south) and Kufrinja (to the north). Within this complex, Cave 2 is the largest extant karstic cavity and is the only cave to exhibit significant sedimentary infilling. Test excavations carried out in 2010 revealed an in situ EUP layer dating to 45–39 ka cal BP (Layer B, Stutz et al., 2015). Excavations were undertaken using a 1 × 1 m grid, divided into 50 × 50 cm sub-squares (Figure 2). During the initial season, the team used arbitrary 5 cm horizontal levels and piece plotting was not employed. However, several combustion features were uncovered and these were excavated as separate features. All sediments were dry screened using 2 mm mesh (Stutz et al., 2015). Although the total area excavated was small, with less than 5 m2 of intact deposits, a large lithic and faunal assemblage was recovered. Lithic data are reported in Shea et al. (2019) and described below; this paper focuses on the fauna recovered during the 2010 field season. Excavations at MHM resumed in 2017; a key goal of the 2017 field season was the recovery of a representative sample of botanical remains, the analysis of which is ongoing.

Preservation of Layer B (the intact EUP deposits) is variable across the site. Units B5 and C5 were characterized by excellent organic preservation. In those units, Layer B was ∼40–70 cm in depth and was directly overlying the bedrock (Stutz et al., 2015:160, Fig. 4). However, bone was largely absent in unit B2, despite the presence of stratified hearths and botanical remains. A multiproxy analysis indicates that the lack of bone in unit B2 is most parsimoniously explained by a diagenetic bone dissolution zone (Stutz et al., 2015; Alonso-Eguiluz et al., n.d.). While bone was preserved in portions of units B3 and B4, bioturbation in those units resulted in the mixture of recent/historic material (Layer A) with the EUP deposits. Units in the F-trench, where Layer B was more ephemeral, had limited, poorly preserved fauna. Given this, we limited our study to faunal materials associated with the EUP from units B5 and C5.

Environmental context Mughr el-Hamamah is located just east of the Jordan Valley on the western slopes of the Transjordanian Plateau, at an elevation of ∼80 m above mean sea level (amsl). This positions the site in an ecotonal environment. The inhabitants of MHM would have had access to the Jordan Valley, with its fertile alluvial fans, elevated temperatures and abundance of freshwater (200 m below msl), and to the summits of the Transjordanian mountain ridge (1100 m amsl). Modern precipitation originates primarily from the Mediterranean and is distributed seasonally; between the cooler months of October and April, rainfall ranges from ∼100 mm/yr in the Jordan Valley to >500 mm/yr at higher elevations near Ajloun (Ababsa, 2013).

The flora of the region is diverse. The Jordan Valley has an open savanoid-like landscape that includes thermophilous trees such as Tamarix (Tamarix sp.) and Christ's thorn jujube (Ziziphus spina-christi) and expansive annual and perennial vegetation coverage (Danin, 2004). On the slopes of the Transjordanian Plateau, the vegetation transitions to semi-steppe Mediterranean scrubland with dwarf shrubs such as white wormwood (Artemisia herba-alba) and thorny burnet (Sarcopoterium spinosum) and then progresses into sclerophyllous evergreen Kermes oak (Quercus coccifera L.) forests at an elevation of ∼700 m amsl (Palmer, 2013). At present, MHM is less than 5 km from the modern tree line. However, analyses of the microfauna and wood charcoal suggest that the tree line was closer to the cave during the EUP (Stutz et al., 2015; Alonso-Eguiluz et al., n.d.).

The EUP occupation of MHM (∼45–39 ka cal BP) is bracketed by two cold and arid Heinrich events (H5 at ∼45 ka, and H4 at ∼38 ka; Bartov et al., 2003; Hemming, 2004). Regional paleoclimate archives suggest that the period between the H5 and H4 events was marked by higher water availability. This is indicated by evidence for a rise in the level of Lake Lisan (the predecessor of the Dead Sea; Torfstein et al., 2013) and by evidence for continuous speleothem formation in the Mediterranean Judean Hills and Eastern Samaria (Bar-Matthews et al., 1999; Vaks et al., 2003). The apparent increase in water availability is not necessarily a product of higher precipitation, but may reflect cooler temperatures and less evaporative conditions that enabled the formation of speleothems in semi-steppe locations (i.e., Eastern Samaria) where climatic conditions have impeded speleothem growth since the beginning of the Holocene (Vaks et al., 2003; Rohling, 2013). Modeling suggests that Lake Lisan reached an elevation of ∼220 m below msl during the Late Pleistocene (Torfstein et al., 2013), placing it within ∼3 km of the MHM. The proximity of Lake Lisan suggests that the EUP inhabitants of MHM also had access to a diverse range of wetland resources.

Lithic analysis Shea et al. (2019) provide an initial report on the stone tool assemblage recovered from the intact EUP deposits (Layer B) at MHM. The assemblage comprises >4000 artifacts and is dominated by flint, most of which is consistent with that available in gravels found in the dry riverbeds near the site. Characteristically Upper Paleolithic cores (prismatic blade cores) and retouched pieces are present throughout Layer B. Although only 14 artifacts have been identified as points, the assemblage is typologically diverse, including El Wad points, a Levallois point, an Umm el Tlel point, and ‘Emireh-like’ points. In comparison to material from the broader Levantine region, Shea et al. (2019) propose that the MHM assemblage is mostly closely related to the Early Ahmarian; however, it does exhibit some key differences from other Early Ahmarian assemblages, including a dominance of endscrapers and a rarity of El Wad/Ksar Akil points.

Given that other Early Ahmarian sites are in close proximity to MHM (defined as being within two days’ walk), Shea et al. (2019) argue that it is important to tease out the behavioral factors that may account for the differences in the lithic assemblages at MHM as compared to other Early Ahmarian sites. While bipolar flake production is often considered to be an adaptation to raw material scarcity, flint is abundant in the MHM region. As such, Shea et al. (2019) propose that the abundance of scaled pieces and bipolar cores at MHM could reflect higher occupational intensity relative to other Early Ahmarian sites. Given the apparent availability of food, fuel, and other raw materials in the vicinity of the site, they argue that foraging may have largely focused on the local area—this would have increased the incentive to recycle tool materials, which could account for the apparent focus on bipolar production. Shea et al. (2019:16) go on to propose that bipolar percussion was “a key part of EUP economic intensification strategies”.

The range of artifacts/ecofacts identified at MHM suggests that it was a place at which a wide range of technological and subsistence tasks were carried out. As compared to late MP assemblages in the region, the lithic assemblage at MHM is consistent with a greater degree of task diversity and on-site work intensity, which Shea et al. (2019:17) see as consistent with a potentially “substantial” increase in labor complexity in the EUP. Shea et al. (2019:17) ultimately hypothesize that “… the apparent rise in task diversity at the site involved intensification … on multiple fronts, from sustained hunting across a larger catchment to more frequent logistical plant food foraging trips, with work parties also remaining in or close to camp, carrying out raw material provisioning, production, and processing tasks.” The incorporation of additional datasets—like the faunal remains reported here—will be key in assessing the degree to which the MHM record is more broadly consistent with increasing intensification.