Life and death at Dmanisi, Georgia: Taphonomic signals from the fossil mammals

Early hominins with bipedal adaptations were confined to Africa for at least two million years—possibly twice that—before any fossil evidence of them in Eurasia. Currently, the majority of evidence indicates early Homo erectus was able to expand its biogeographic distribution to more northerly latitudes, whereas earlier hominins, the australopiths, did not. Dmanisi in Georgia has long been an anchor for minimum dates in Eurasia, with occupation going back to 1.80 Ma (Ferring et al., 2011, Ferring et al., in press, but there are many unanswered questions about when and why hominins expanded their geographic range into Eurasia. An ongoing issue is that ‘earliest’ sites may date to later than the first expansion, and some sites have contested dates and evidence (Dennell and Roebroeks, 2005). The issue of moving dates makes testing hypotheses difficult, as does the evidence for multiple dispersal events and local dying-out of populations (Bar-Yosef and Belmaker, 2011).

In the last five years, new sites have been found and dated, and other sites have been redated. Hominin occupation still appears to be much earlier in Asia than in Europe. In China at Shangchen in the southern Loess Plateau, there is a series of archaeological occurrences, some with fauna, in a loess-paleosol sequences going back to at least 2.12 Ma using paleomagnetism (Zhu et al., 2018), and the Palaeolithic site of Xihoudu 100 km east of Shangchen has recently been dated to 2.43 Ma (Shen et al., 2020). Other sites in China likely older than 1.5 Ma include Yuanmou incisors (1.7 Ma, Zhu et al., 2008) and the Gongwangling cranium (1.63 Ma, Zhu et al., 2015), while archaeological sites in the Nihewan basin may or may not date as far back as 1.66 Ma (Dennell, 2015). In India, probable tool cut marks have been found on fauna from the Mosol Formation in the Siwaliks that may be 2.6 Ma (Malassé et al., 2016). The evidence of occupation of this region has been long suggested, but the erosional context of the early archaeological sites in India and Pakistan makes these sites and dates uncertain (Dennell, 2021a). Meanwhile, the first Javanese H. erectus fossils have been redated to be a bit younger, about 1.45–1.30 Ma (Matsu'ra et al., 2020). Furthermore, some interpretations of Homo floresiensis post-cranial morphology suggest retention of plesiomorphies that may have evolved from an australopithlike post-cranium rather than from Homo, implying even earlier dates for occupation of Eurasia (Morwood and Jungers, 2009; Ferring et al., 2011; Dennell et al., 2014). However, new analyses of dental morphology suggest H. floresiensis and Homo luzonensis have more in common with H. erectus rather than earlier H. habilis/H. rudolfensis, suggesting these islands in southeast Asia were initially colonized by H. erectus (Zanolli et al., 2022). ‘Ubeidiya in Israel, at about 1.5 Ma, with hominins, fauna, and tools demonstrates occupation in the Levant at this time (Barash et al., 2022).

In Europe, the earliest sites are in the southern regions and so far date to after 1.5 Ma, such as at Barranco León and Fuente Nueva-3, Orce, Spain (1.4 Ma, Espigares et al., 2019, but refer to the studies by Muttoni et al., 2013, Alvarez et al., 2015, for discussions of these dates), and at Trinchera Elephante, Spain (1.2 Ma, Carbonell et al., 2008). Alto de las Picarazas may be this age as well (Gabarda et al., 2016). In Iberia, the early human occupation is associated with increased temperature and precipitation (Blain et al., 2021). Pirro-Nord, Italy, also represents an early occupation in the Mediterranean area, with cut-marked bones (1.7–1.3 Ma, Cheheb et al., 2019).

Some important hypotheses of hominin range expansion emphasize aspects of the environment such as climate/environmental change and greening of the Sahara and Arabia (e.g., Vrba, 1995; Groucutt et al., 2021), differences in migration routes (e.g., O'Regan et al., 2006; Dennell, 2021b), and/or co-occurrence with other dispersing mammals (e.g., Martínez-Navarro, 2004). With these environmental hypotheses, changes in hominin behavior and biology are also brought to the fore, such as increased developmental and behavioral plasticity (Antón et al., 2016), meat-eating (Foley, 2001), longevity and cooperation (Tappen, 2009), habitual terrestrial bipedalism (Pontzer et al., 2010), and increases in body and brain size (Antón et al., 2002). Some hypotheses regarding biogeographical spread, like those evoking substantial increases in brain and body size, adoption of fire, new stone tool technologies, or the ‘green light’ corridors of the opening for major interchanges of African and Eurasian faunal species with major climatic events, have not been compelling using the evidence from the Dmanisi. This is because most of the fauna is Eurasian, brain and body sizes of the hominins are smaller than initially expected, there has yet to be found evidence for fire, and there are no Acheulian bifaces so far (Tappen et al., 2007; Tappen, 2009). Another consequential hypothesis emphasizes that increased meat-eating allowed hominins to spread to the higher latitudes. The meat-eating hypothesis makes sense based upon many lines of reasoning as our largely fruit- and leaf-eating ancestors had to add substantially more animal products to cope with a lack of fresh fruits and fresh vegetation during winter seasons (e.g., Shipman and Walker, 1989; Foley, 2001; Stiner, 2002; Antón and Swisher, 2004; Foley and Gamble, 2009; Antón et al., 2014a). Recent paleoenvironmental evidence from Dmanisi, however, suggests a warm, semi-arid Mediterranean climate with long dry seasons in the summer (Blain et al., 2022). These dry seasons too may have required the addition of meat in the Dmanisi paleoenvironment. Meat and marrow consumption is first documented in the Early Pleistocene in Africa, just prior to when hominin sites are found in Eurasia (Domínguez-Rodrigo et al., 2005; Ferraro et al., 2013). Carnivorous animals tend to range far and have wide geographic distributions and can disperse quickly and widely. Furthermore, the trends toward body and brain size increases and the reduction of cheek tooth size over time in H. erectus imply higher quality diets (Antón et al., 2002, 2014), although the Dmanisi hominins are not very derived in these traits as compared to early Homo in Africa (Lordkipanidze et al., 2013; Rightmire et al., 2018). Importantly, H. erectus at Dmanisi is recognized as an obligatory terrestrial biped which may also be the key to its spread (Pontzer et al., 2010). At and before the time period in question, many hominins with bipedal features also retain arboreal adaptations.

As an exploration of the primary taphonomic data from Dmanisi, we contribute to the understanding of the site's formation and to the understanding of Early Pleistocene hominin carnivory and hominin–carnivore interactions at a time and place with few other sites.

Dmanisi in Georgia remains among the earliest sites in Eurasia with thousands of vertebrate fossils, numerous hominin fossil specimens, and stone tools and a well-established stratigraphic sequence. The site shows evidence of hominin occupations from 1.80 to 1.76 Ma. (Dzaparidze et al., 1989, Gabunia et al., 2000; Vekua et al., 2002; Lordkipanidze et al., 2006, 2013; Ferring et al., 2011; Mgeladze et al., 2011; Rightmire et al., 2018). It has been excavated for more than 30 years in several different excavation blocks (Fig. 1).

Here we address questions of site formation and behavior based on information gleaned from mammalian taphonomy from the medium and large mammals at Dmanisi (>5 kg). We focus on presenting the bone surface modifications, breakage, and skeletal element frequencies with a greatly expanded sample leading to substantial revisions to previous interpretations of the site. Preliminary systematic taphonomic studies of the fauna have been published (Lordkipanidze et al., 2007; Tappen et al., 2007). At the time of these publications, the sample size of specimens examined for taphonomic information was about a third of what we present here. These earlier papers suggested much lower rates of large carnivore activity than we have now, and we present this evidence. A few probable stone tool cut marks were reported earlier (Lordkipanidze et al., 2007; Tappen et al., 2007), but given recent discussions about the identifiability of cut marks and our increased sample sizes, we take a closer look at these marks and discuss the expanded evidence. We address how firm the evidence is for hominin meat-eating at the site and discuss the taxa utilized and anatomical locations of cut marks. We report on the taphonomic features of the fossil assemblages in different stratigraphic units and facies to present evidence of site formation processes as well as clues to the relationship of H. erectus and the predators at the site. Spatial analysis of some of the taphonomic variables in Block 2 related to these questions have been published (Coil et al., 2020), but here we provide the primary taphonomic information and include frequencies of mammal families and of taphonomic variables from the main excavation areas going back to 1983, but focusing on Block 2N, M5, and M6.

Due to the long history of excavations, the site's location underneath a medieval city, and geological testing around the promontory, the Dmanisi fossil assemblages come from several excavation areas. The main excavations in the 1980s were in Buildings 11 and 14 (Dzaparidze et al., 1989). In Block 1, a large expanse above the calcrete (‘kerki’) and 9 m2 below it was excavated in the 1990's; among the bones recovered were a mandible and two crania of early H. erectus (Gabunia and Vekua, 1995; Gabunia et al., 2000). Block 2S (South) was excavated in 2000–2001 (Jöris, 2008). We have excavated Block 2N (north of the 59 m gridline) since 2001, where dense concentrations of fauna and more H. erectus fossils were found (Vekua et al., 2002; Lordkipanidze et al., 2005, 2006, 2007, 2013). Nineteen areas expanding outward from these central excavation blocks have been tested, each test/excavation locality designated with an ‘M,’ standing for the Medieval Room in which the test was located. The ‘M’ excavations range from small test pits of 1 m2 to larger excavation areas, such as in M5 M6, M11, and M17 (Ferring et al., in press). Each excavation block or test unit has exhibited differences in sedimentary facies and density of the faunal and stone tool assemblages (Ferring et al., in press), but all are Early Pleistocene, late Calabrian Stage. We present pooled taphonomic data collected from these units, but we focus most intensely on the B1 stratum in Blocks 2N, M6, and M5 because these excavation blocks have large samples that were excavated using the A–B stratigraphic scheme that was developed since the year 2000, rather than the I–VI scheme used before.

The stratigraphy of the site is described (Ferring et al., 2011, Ferring et al., in press; Coil et al., 2020), but a brief review is warranted, since we compare faunal signals from differing stratigraphic units and facies. The Mashavera Basalt underlies the bone- and stone tool–bearing sediments and is dated to 1.80 Ma in M5 (Ferring et al., in press). A series of volcanic ashfalls fell soon after the lava hardened. The lowest is the A1 ‘Black Ash’. The A strata have normal polarity and correlate with the Olduvai subchron between 1.80 and < 1.78 Ma (Ferring et al., in press). A1 contains stone tools in block M5, and some fossil vertebrates across excavation blocks. This ashfall sets the date for the earliest hominin occupation at Dmanisi, one of the earliest in all of Eurasia (Ferring et al., 2011).

Ferring et al. (2011; Ferring et al., in press determined that the degree of weathering of the basalt and ashfalls is low, and the B1 ash was deposited on top of A soon after the A ashfalls. The B1 ashfall has little soil development, a reversed paleomagnetic signature, and is found stratigraphically below a basalt at the nearby site of Orozmani that is dated to 1.76 Ma (Gabunia et al., 2000; Messager et al., 2010; Nomade et al., 2016). This correlation indicates the B1 ashfall dates to just after the Olduvai normal subchron that ended at about 1.78 Ma, and before 1.76 Ma (Gabunia et al., 2000; Vekua et al., 2002; Ferring et al., 2011, Ferring et al., in press). Most of the fauna and the hominins from Dmanisi come from these B1 ashfall deposits. Of the large excavation blocks excavated since 2000, close to 85% of the fossils come from B1 (Table 1, Table 2). B2 and B3 ashfalls then buried B1, and like A and B1, they contain fossils and stone tools of Oldowan character, but the number and density of fossils is lower than in B1. B2 and B3 have more soil formation features and were subaerial longer than B1.

In several excavation areas, notably in Block 1 and Block 2N and in several of the ‘M’ excavations, there are geological piping and/or gully features cutting through A ashfalls that are filled with fauna and B1 ashes. These features are of great interest because they contain information about the integrity of Block 2N and Block 1 and so about the nature of the associations of the fossil mammals with the hominin fossils and stone tools. These pipe and gully features in Block 2N are described in detail in Coil et al. (2020). In Block 2N, these features were divided into substrata B1z, B1x, and B1y (Fig. 2). B1a is the primary in situ ashfall of the B1 phase. Erosion into the A ashes by piping and gullying then occurred and filled with sediment and bones referred to as B1y and then B1x. B1y and B1x were covered in B1z gully fill, and B1x is near the breach leading to B1y. The number of bones in the B1xyz pipe and gully fill features is four to seven times higher than in B1a and much more than in the A stratum, B2, or B3 for that matter (Table 3). Thus, it seems that the superabundance of bones in these fills provides evidence that the bones are not simply derived from B1a or the adjacent A sediments slumping in. The deposits are not fluvial in nature. We explore if these assemblages differed from each other in taphonomic characteristics to see if they should be interpreted as a single analytic unit or separated. Some field observations are important to note: they are provided in Supplementary Online Material (SOM) S1 and Figure S1.

留言 (0)

沒有登入
gif