4.1 Faunal analysis
The A9 taxonomy and taphonomy of the zooarchaeological assemblage were extensively published in previous studies (Cassoli and Tagliacozzo, 1991; Nannini, 2012; Fiore et al., 2016; Romandini et al., 2014, 2016). Results show a predominance of cervids, followed by bovids, caprids, suids, rodents (marmots), and carnivores. As authors argued during the archaeozoological analysis, the ungulates show a prevalence of fragments of metapodials (29.1%), zeugopods (24.2%), and stylopods (13%), followed by elements of extremities like phalanges and sesamoids (14.7%). Cranial remains are scarce and are predominantly represented by isolated teeth (11.5%) and parts of the hemimandible and jaw (4.4%), whereas the trunk and spinal column are nearly absent (1.8%) (Romandini et al., 2014). The number of identified specimens is low due to the high rates of anthropic bone breakage (over 92.5% of fragments are < 2 cm in size) and heat alteration (50%). Mechanisms of mobilisation that could have altered bones are ephemeral (2%). Among the post-depositional agents, the most common are root marks (38%), followed by micro-fissures and/or weathering cracks (0.8%), manganese oxide stains (18%) and exfoliations (2.2%). Some concretionated fragments distributed near the dripline limit (21.4%) and a few with trampling abrasions (15.9%) in the east part of the cave were documented. Anthropic traces observed on ungulate bone surfaces can be ascribed to different stages of the butchering process (Romandini et al., 2014). The frequency and patterning of rodent and carnivore marks suggest secondary access to ungulate remains; therefore, carnivores did not make a significant contribution to bone accumulation.
4.2 Refits
In unit A9, 79 bone refits were detected (73 mechanical and 6 anatomical), grouping 180 remains (6,1%) connected by 107 lines. Of these, 65 are composed of two elements, nine of three, three of four, one of five, and one of six (Table 1).
Many of the horizontal connections cover short distances (n° 79, min. 2, max. 100 cm), followed by medium (n° 12, min. 118, max. 294) and long-distance (n° 16, min. 303, max. 751 cm) (Fig. 2).
Overall refits concern medium-big fragments of the fore- and hindlimb, with signs of burning/calcination and others of an anthropic nature, followed by cranial elements (maxillaries, mandibles, and teeth) and a few remains of vertebrae or ribs. The longer distance refits are between the teeth of roe deer.
Taxonomic identification was possible for many of them, and refits have been documented as follows. Cervus elaphus has the largest number of refits (18, including cf. Cervus elaphus) affecting 41 bones (Fig. 3). The minimum distance between refitted elements is 7 cm and regards a fragment of mandible mechanically broken, whereas the maximum is 750 cm, and concern teeth (anatomical refit). The skeletal composition shows a balance between metapodials, zeugopodials and stylopodial, which offer higher nutritional resources in terms of meat and marrow, followed by cranial elements, in particular teeth (six refits) and mandibles.
Human activity is the main factor that caused the faunal scattering: nine refits show evidence of intentional breakage addressed to the marrow extraction. Fractures with curved/V-shaped delineations, oblique angles and smooth edges are the most common, followed by longitudinal and transversal delineations. In association with breakage patterns, cut marks were detected in 4.4% of the remains, impact points in 2.7%, and percussion cones (1.6%). Also burn damage is very common affecting 5.5% of fragments. Modifications caused by post-depositional agents show concretion as the most common agent affecting 9.6% of the fragments, followed by root etching (5.5%) and trampling (3.3%), which likely do not have caused movements, as their refits are very close.
Capreolus capreolus is represented by four refits involving 11 elements; three are anatomical and concern dental series, mandibles with teeth and thoracic vertebrae. The only mechanical refit regards a III phalanx, whose elements are burnt. Long and medium-distance connections have been documented among dental series (incisors) and vertebrae. Only one refit on short distance (129.6 cm) may have undergone a displacement due to non-anthropic agents: a mandible with teeth affected by trampling striae and smoothing edges caused by water.
Two refits have been attributed to Bos Primigenius (cf.): a metatarsus and a tibia, involving, respectively, three and two fragments. The refit on the metatarsal exhibits burning traces on all three fragments and carnivore marks on the percussion cone. There are no other post-depositional agents that could have altered the fragments. Despite this, the refit is local, reaching a maximum distance of 80.7 cm. The tibia refit, however, covers approximately 3 m. Despite the presence of trampling marks, the other modifications are of an anthropic nature. Both refits are located inside the structures A9 SVI.2 and A9I SXIII, the latter identified as occupation horizon (OH) based on micromorphological inspection (Marcazzan et al., 2022).
Two other refits have been uncertainly attributed to Megaloceros giganteus (cf.). They concern a tibia (25.8 cm) and a radius involving two fragments conjoined over a long distance (586.4 cm). The last case involves a retoucher, characterized by cut marks and impact points. Although trampling marks have been documented, a displacement of the fragments due to this agent is excluded (Fig. 4).
Of uncertain attribution is one refit (two elements) documented for bos/Bison, elk/giant deer, and elk (cf. Alces alces) (Table 1). Among other herbivores, one refit involving two elements belongs to chamois (Rupicapra rupicapra).
Other documented taxa include the marmot (Marmota marmota), which has one refit of two elements, and the weasel (Mustela nivalis), with a refit of three elements. Moreover, eight refits (involving 17 fragments) have been recognized among Ungulata and seven (14 fragments) for Cervidae. Connections between ungulates, mostly characterised by diaphyseal burnt elements are mostly local, reaching a maximum distance of 60 cm. The other two connections cover medium and long distances; some trampling striae, documented on both sets, may have caused the fragment movements. Even the elements attributed to cervids are characterised by the presence of tibiae, humeri, femurs, and metacarpals, with a balance between short- and medium-distance refits. Also, in this case, anthropic interventions and some phenomena related to trampling may have caused the scattering of the remains.
Unidentified fragments categorised in weight sizes documented 32 refits for a total of 75 elements. All refits are local, up to a maximum of 84 cm and 12.7% are affected by combustion. Only one conjoin has been documented within the A9 SXI.1 structure
The vertical projections identified eight refits, translated into 19 fragments, which show a mobilization between 1 and 23 cm. Identified taxa include large herbivores, such as auroch and elk, red deer, chamois, and two remains of big-size ungulates. Anatomical elements concern diaphyses of long bones, except one refit involving two incisors.
The orientation of connection lines concerning fragments moved by humans (over medium and long distances) indicates some directionalities towards north-south, south-north, west-east and east-west: these are movements from areas characterised by hearths/structures towards other shelter areas, for example, the east wall or the drip line. Although data are preliminary, and a sample selection was made for refits application, we note intentional movements towards the peripheral areas, compared to those characterised by hearths or other structures, where daily activities were carried out (Fig. 5).