Beyond 2,000 meters, first evidence of intense prehistoric occupation in the Pyrenees

Abstract

Cave 338 is a high-altitude prehistoric site located at 2,235 m a. s.l. in the eastern Pyrenees (Queralbs, Girona, NE Iberian Peninsula). Excavated between 2021 and 2023, the site preserves an exceptional and well-stratified archaeological sequence documenting recurrent episodes of human occupation spanning from at least the early 5th millennium cal BC to the late 1st millennium cal BC. Radiocarbon dates indicate that these occupations occurred during several discrete phases separated by intervals of reduced or absent activity. The cave currently represents the highest-altitude prehistoric cave site with sustained occupation currently documented in the Pyrenees. The archaeological record reveals a dense succession of combustion features, abundant faunal and ceramic remains, and an extraordinary assemblage of green mineral fragments, most likely malachite, repeatedly introduced into the cave and processed in situ. This evidence indicates the systematic exploitation of copper-rich minerals in a high-mountain environment from the Late Neolithic to the Bronze Age, providing an unprecedented record for the Pyrenean range and one of the earliest high-altitude contexts of mineral exploitation documented in Europe. The organization of space, the density of combustion features and the nature of the associated activities indicate that Cave 338 was not a marginal or sporadically used shelter, but rather a repeatedly occupied logistical site integrated within structured seasonal mobility systems. These findings challenge prevailing interpretative models that characterize prehistoric occupations above 2,000 m.a.s.l. as ephemeral and low-intensity. Instead, Cave 338 demonstrates that alpine environments could play a central role in long-term prehistoric land-use strategies, particularly in relation to the exploitation of mineral resources. As such, the site provides a key reference framework for understanding high-mountain occupation, resource exploitation and mobility dynamics in the Pyrenees during later prehistory.

1 Introduction

1.1 Settling high mountain environments: prehistoric archaeology in the Iberian Pyrenees

Over the last two decades, prehistoric archaeology in the Pyrenees has undergone a significant conceptual and methodological transformation. This renewal has been particularly notable on the southern, Spanish side of the mountain range, reshaping our understanding of long-term human occupation in the Pyrenean mountains.

At an empirical level, these advances have led to a much better documented and chronologically robust picture of prehistoric settlement dynamics. Clear and well-dated evidence now demonstrates human presence in the eastern Pyrenees at least since the Upper Paleolithic and the Late Glacial Maximum. Sites such as Montlleó (Prats i Sansor, Cerdanya Valley) and Bauma dels Fadrins (Queralbs, Freser Valley) document occupations dated between approximately 21-15.5 ka cal BC and 11-10.1 ka cal BC, respectively, providing compelling evidence for the integration of mountain environments into long-distance mobility networks during the Late Pleistocene (Sánchez de la Torre et al., 2024; Rafart-Vidal et al., 2025; Díez-Canseco et al., 2024). In parallel, research carried out in recent years has revealed a marked increase in the number of sites dating to the onset of the Holocene, with Epipalaeolithic and Mesolithic occupations documented in several Pyrenean valleys. Alongside classic reference sequences such as Bauma Margineda (Guilaine and Martzluff, 1995) and Sota Palou (Carbonell, 1985; Tornero et al., 2022a), new stratigraphically excavated sites and more fragmentary archaeological evidence point to a proliferation of occupations during the Preboreal and Boreal, indicating a sustained and diversified use of mountain landscapes from the beginning of the Holocene onwards (Orengo et al., 2014; Gassiot et al., 2024, 2017; Díez-Canseco et al., 2022). Furthermore, research conducted over the last two decades has significantly refined our understanding of the earliest Neolithic occupations in the Pyrenees, now attested in many more Pyrenean valleys and newly excavated sites. This work has also improved our understanding of how these groups developed agropastoral production economies and incorporated vertical and altitudinal mobility into their organizational systems (Tejedor-Rodríguez et al., 2021; Antolín et al., 2018; Sierra et al., 2021; Díez-Canseco et al., 2023; Gassiot et al., 2015).

A central aspect of this transformation has been the explicit extension of archaeological research into high-mountain environments located above 2,000 m a.s.l., areas for which archaeological information was previously scarce or virtually non-existent. Large-scale and long-term research programmes, most notably those developed by the GIAP (Grup d'Investigació en Arqueologia del Paisatge; hosted in ICAC) and the GAAM (Grup d'Arqueologia de l'Alta Muntanya; hosted inUAB–CSIC), have implemented systematic survey strategies combined with diagnostic test pits to explore alpine and subalpine zones (Gassiot et al., 2012, 2024; Gassiot, 2016; Palet et al., 2007, 2012, 2017; Palet M. et al., 2019; but see also previous works in (Rendu 2003). These efforts have resulted in the documentation of hundreds of archaeological points of interest. The overwhelming majority correspond to modern or historical occupations, often represented by well-preserved surface structures. Only a very small proportion of the documented points above 2,000 m a.s.l. can be confidently attributed to prehistoric periods.

Despite their limited number, these prehistoric occurrences demonstrate that alpine environments were frequented from very early chronologies. However, the nature of these occupations has, until now, been interpreted predominantly as brief, low-intensity and episodic. Most are characterized by sparse archaeological assemblages, limited stratigraphic development and short-life use of space, and are commonly interpreted as seasonal, strategic, or transitory occupations linked to mobility across the mountain range. This interpretative model partly reflects the methodological approaches employed, which have prioritized extensive, quantitative documentation through surface survey and diagnostic testing over intensive, qualitative excavation strategies. It is also conditioned by the substantial logistical and environmental challenges involved in conducting high-resolution prehistoric excavations in isolated high-altitude contexts. As a result, only a very limited number of prehistoric sites in alpine zones have been excavated stratigraphically and in extension.

Against this background, Cave 338, excavated between 2021 and 2023 and located in the eastern Pyrenees at over 2,000 m a.s.l., offers a markedly different perspective on prehistoric occupation in high-mountain environments. Unlike previously documented sites at similar altitudes, Cave 338 presents evidence for an intensive and repeated prehistoric use of space, reflected in the density of archaeological remains, the diversity of documented activities and the development of well-preserved stratigraphic sequences. The characteristics of this site indicate a form of high-altitude occupation that goes beyond brief and non-intense or purely transitory use, challenging dominant interpretative models currently applied to prehistoric sites above 2,000 m in the Pyrenees. The data derived from the excavation and analysis of this new archaeological site therefore constitute a critical contribution to the reassessment of prehistoric land use in alpine environments and provide a new framework for understanding the role of high mountain zones within long-term settlement and mobility systems.

1.2 Prehistoric archaeology in the Freser Valley

The Freser Valley constitutes one of the main natural corridors of the Eastern Pyrenees, both from a geomorphological and a geographical perspective. Extending from the high mountain massifs of Puigmal and Balandrau to its confluence with the Ter River at Ripoll, the valley articulates a continuous altitudinal gradient that connects alpine environments, mid-mountain sectors and lowland areas within a relatively confined spatial framework. This configuration has long been recognized as a key factor facilitating human movement and access across the eastern Pyrenees, particularly by providing a natural route of penetration from the plains toward high-altitude zones and intermontane passes (Carbonell et al., 1985). The combination of accessible slopes, transversal tributaries, shelter-rich karstic formations, and diverse ecological niches makes the Freser basin especially favorable for prehistoric mobility, seasonal circulation and repeated occupation, a role that was already emphasized in early synthetic studies of the valley as a structurally coherent unit for understanding long-term human presence in this sector of the Pyrenees (Carbonell et al., 1985).

The prehistoric archaeology of the Freser Valley has been shaped by a markedly discontinuous research trajectory. The earliest archaeological investigations in the area date to the first decades of the twentieth century and are closely linked to the pioneering development of prehistoric research in the Pyrenees. Within the framework of the Institut d'Estudis Catalans, Josep Colominas carried out the first explorations of caves and funerary contexts, providing the earliest formal archaeological references for prehistoric occupation in the valley (Bosch Gimpera, 1919, 1923; Colominas Roca, 1927-1931), albeit with limited stratigraphic control and within the methodological constraints of the period.

Among the sites documented during this initial phase, the Cova de Rialb, located in the upper Freser basin, yielded prehistoric materials that were broadly attributed to the Neolithic or Late Neolithic. In addition, Colominas' work at the Dolmen of Pardinella (Gombrèn) represents a key historiographic milestone closely connected to the Freser Valley. The monument, discovered shortly before 1925 and partially dismantled prior to any archaeological intervention, was visited, and documented by Colominas around 1931, who published a plan and photograph and reported the recovery of ceramic fragments and human remains. Although interpreted at the time as a simple dolmen, later reassessments have identified the structure as a Neolithic cist with tumulus, most likely dating to the Middle Neolithic and related to the Solsonià group (Carreras and Tarrús, 2017).

After this pioneering phase, prehistoric research in the Freser Valley remained largely dormant for several decades, coinciding with the military uprising of 1936, the Spanish Civil War and the subsequent dictatorship of Francisco Franco. From this long and poorly documented period, the only known archaeological intervention is a small stratigraphic test pit excavated by G. López Pérez at the Tut de Fustanyà during the summer of 1960 (Ripoll Perelló, 1960–1961).

A second, clearly identifiable research cycle began in the early 1970s under the direction of Eudald Carbonell and collaborators (Ramon Busquets and Cisco Coll), associated with the Grup Grober Xaialsa. This phase marked a major methodological and conceptual renewal in the study of prehistoric occupation in the basin and introduced a systematic territorial approach explicitly framed at the scale of the hydrographic basin. Early work combined targeted excavation with extensive survey strategies and focused on the upper Freser Valley, particularly in the Torrent d'Estremera area, where surveys and excavations carried out between 1971 and 1974 led to the identification of numerous cave sites, funerary contexts and surface stations distributed along this corridor (Carbonell, 1978; Carbonell et al., 1978, 1985; Toledo and Pons, 1982).

Within this framework, the Roc de les Orenetes site was excavated in 1972 and 1973 and identified as a cave necropolis with multiple inhumations, preliminarily dated to the Bronze Age (Carbonell et al., 1985; Toledo and Pons, 1982). At the same time, renewed excavations were carried out at the Tut de Fustanyà. Although the archaeological record was heavily affected by earlier disturbances and systematic looting, severely limiting stratigraphic resolution and interpretative potential (Carbonell et al., 1976, 1978), these works provided the first evidence for Paleolithic occupations along the Freser corridor and reinforced the interpretation of the valley as a long-term axis of prehistoric mobility.

Research subsequently expanded toward the mid and lower sectors of the basin, culminating in the excavation of the open-air site of Sota Palou (Campdevànol). Excavated between 1978 and 1982, Sota Palou preserved a stratified Mesolithic occupation embedded within fluvial deposits of the Freser River and represented a major methodological breakthrough. Its excavation and subsequent monographic publication marked a turning point not only for the archaeology of the Freser Valley but also for Catalan prehistoric research, as it constituted the first open-air Mesolithic site in Catalonia excavated and published according to modern methodological standards (Carbonell and Mora, 1983; Carbonell et al., 1985).

Despite these advances, prehistoric research in the Freser Valley entered a prolonged phase of relative inactivity after the early 1980s, lasting until very recent years. A first exception to this hiatus is the excavated cist of the Coll de Fontalba (Queralbs, Ripollès), located at approximately 2,070 m a.s.l. in a high-mountain pass linking the upper Freser Valley with the southern slopes of the Puigmal massif. The monument was excavated in August and December 2015 under the direction of Esther Medina, within the framework of a preventive archaeological intervention promoted by the Servei d'Arqueologia i Paleontologia of the Generalitat de Catalunya (Medina, 2016a,b). Unfortunately, the site was found in a very advanced state of deterioration, attributed to the combined effects of high-altitude erosion, sedimentary dynamics and possible ancient disturbance or looting, and the excavation yielded very limited archaeological remains. Nevertheless, the architectural characteristics of the monument and its regional parallels support its attribution to the Late Neolithic-Chalcolithic. Its location in a high-mountain pass suggests that funerary practices in the Freser Valley extended beyond habitation areas into elevated zones linked to circulation routes and seasonal movement, contributing to a more nuanced understanding of prehistoric land use in the eastern Pyrenees.

A second exception is represented by the work carried out by Josep Maria Palet and his team in the Natural Park of the Headwaters of the Ter and Freser Rivers between 2010 and 2017. This research aimed to document archaeological evidence in high-altitude environments within the framework of the GIAP-ICAC. Although the project recorded hundreds of archaeological points of interest, mainly corresponding to historical periods, prehistoric evidence was limited to a small number of sites, probably reflecting the focus of interest of the project. In addition to Cave 338, first tested by Palet's team (Palet et al., 2016), two further sites with prehistoric occupation were identified through small test pits: Aigols Podrits II (c. 2,470 m a.s.l.) and Coma de Vaca I (c. 2,110 m a.s.l.). In both cases, open-air occupations located in tributary alpine basins of the Freser River, basal prehistoric phases were documented beneath later, predominantly historical, deposits. Although the available information from these prehistoric occupations is limited, radiocarbon dates obtained from charcoal samples in the basal levels indicate occupations dating to the late fourth millennium cal BC at Aigols Podrits II (structure 346, layer 211; Palet et al., 2017) and to the mid-third millennium cal BC at Coma de Vaca I (structure E1, layer 114; Palet et al., 2014). Despite their limited extent, these data provide important evidence for prehistoric occupation in very high-altitude zones of the Freser Valley and further highlight the significance of its altitudinal transect.

In this context, the ARRELS project (2018-2025), led by Carlos Tornero, has played a decisive role in reactivating prehistoric research in the Freser Valley after more than three decades of relative inactivity. Conceived as a long-term, basin-scale research programme, ARRELS has combined renewed fieldwork with the systematic re-evaluation of museum collections and the identification of new archaeological sequences (Tornero and Carbonell, 2018, 2022; Tornero and Díez-Canseco, 2026). The project has enabled the re-study and contextualization of classic reference sites such as Roc de les Orenetes, Tut de Fustanyà and Sota Palou, while integrating modern excavation and recording methodologies with large-scale radiocarbon dating programmes, bioarchaeological analyses, isotopic studies and ancient DNA. At Roc de les Orenetes (Queralbs), the reopening of the site after more than four decades has resulted in a high-resolution bio archaeological framework for Late Chalcolithic and Early Bronze Age funerary practices at high altitude, including new data on the preservation and composition of the collective burial, palaeo pathological evidence and interpersonal violence, as well as broader insights into diet and subsistence strategies (Ramírez-Pedraza et al., 2022; Moreno Ibáñez et al., 2024a,b; Hernando et al., 2024). In parallel, ARRELS survey and excavation have significantly expanded the prehistoric settlement record through the excavation of previously unknown or poorly documented sites, most notably Bauma dels Fadrins and Cave 338, which have yielded well-preserved stratigraphic sequences spanning from the Late Upper Paleolithic to later prehistoric periods (Díez-Canseco et al., 2020, 2022, 2023, 2024; Rovira et al., 2024). In addition, recent interventions at the Cova del Cingle del Cullell have further enriched the prehistoric funerary record of the Freser Valley, documenting new evidence integrated within the same territorial research framework (Soler et al., 2024). Finally, the systematic revision of archival documentation and unpublished excavation records from the 1970s and 1980s, particularly those associated with the Grup Xaialsa, has enabled the historiographic reassessment of earlier research phases in the valley and their integration into current research questions (Busquets, 2020; Fabrés, 2021). Through this combined strategy, ARRELS has re-established the Freser Valley as a key reference area for exploring long-term human occupation, mobility, and high-altitude land use in the eastern Pyrenees, while laying the foundations for future research within a coherent territorial and diachronic framework.

This study presents a first description of the prehistoric deposits documented at the Cave 338 site (Queralbs, Girona). Earlier diagnostic test pits and soundings had already revealed the site's singular importance and high scientific value, given its location, the excellent preservation of its archaeological deposits and their antiquity. The site is unique for several reasons: (i) it represents the highest-altitude prehistoric cave occupation documented in the Pyrenees; (ii) a green mineral, most likely the copper-rich malachite, was intensively exploited and it is documented across multiple combustion pits, a practice unparalleled in these region and prehistoric chronologies; and (iii) the state of preservation of the archaeological remains is exceptional, to the extent that the Government of Catalonia authorized the protection and closure of the site to the general public until at least the completion of the ongoing research.

2 The Cave 338

2.1 Location and general setting

Cave 338 is located in the Eastern Pyrenees in the Núria Valley within the Natural Park of the Ter and Freser Headwaters (Parc Natural de les Capçaleres del Ter i del Freser), in the municipality of Queralbs (Ripollès, Girona) (Figure 1). The site lies at an altitude of 2,235 meters above sea level, halfway between the Monastery of Núria Valley and the summit of Puigmal, the highest peak in the area. The cave is in the Coma de l'Embut area, on the north-eastern slope (Figure 2). The site lies approximately 150 meters from the shaft known as Forat de l'Embut, but on the right bank of the river, about 25 meters above the riverbed, and facing north. Its coordinates are X: 4,29,234, Y: 46,94,287 (UTM ETRS89, zone 31N) (Tornero et al., 2022b) (Figure 3).

Figure 1

Figure 2

Figure 3

The Cave 338, also known as Cova del Catau de l'Os and Cova del Forat de l'Embut, was first mentioned in the speleological journal Cavernes, published by the Grup Espeleològic de Badalona (GEB) in 1964 (Carreras, 1964). In 2006, it was topographically described by members of the GEB (de Vallès, 2009). It is a karstic cavity of over 100 m² and structured around two small caves separated by approximately 6 meters. The larger one is about 5 meters wide, gradually narrowing into a gallery extending for 28 meters. The smaller cave consists of a narrow passage, currently difficult to access (Figure 3).

2.2 Background

Between 2010 and 2011, the Landscape Archaeology Research Group (GIAP) of the Catalan Institute of Classical Archaeology (ICAC-CERCA) identified surface archaeological remains at the site, as part of its survey and diagnostic test-pitting programme carried out in the Vall de Núria area (Palet and Orengo, 2013).

In 2012, the GIAP conducted an archaeological test pit measuring 0.70 × 1.00 m and 50 cm deep at the entrance of the main cavity. This excavation revealed a homogeneous stratigraphic sequence composed of clayey silts, gravel, small pebbles, and blocks of varying size. Moreover, the discovery of charcoal, handmade ceramic production debris, faunal remains, and chunks of a rock with predominant green patches (most likely the copper carbonate malachite) suggested multiple episodes of use of the cavity, some of which were associated with prehistoric periods. Specifically, two radiocarbon dates were obtained from stratigraphic units 103 and 112, which identified an early phase of use of the cavity dated to the late 4th millennium BC (3360-3098 cal. BC), and a later phase dated to around the 11th-12th centuries AD (1028–1184 cal. AD), respectively (Palet et al., 2014, 2016, 2017; Palet J. M. et al., 2019). Dr. Lídia Colominas identified sheep/goat (Ovis/Capra) bone remains in the strata associated with the most recent period, as well as highly fragmented remains in the lower levels, probably related to meat consumption. On the other hand, a preliminary study of the anthracological sample recovered throughout the test pit sequence revealed the almost exclusive presence of Pinus uncinata, which may be associated with the nearby presence of mountain pine (black pine) forest (cited in Palet et al., 2017).

In 2017, a new excavation campaign was carried out, jointly organized, and directed by the ICAC and the Catalan Institute of Human Paleoecology and Social Evolution (IPHES) led by Dr J. M. Palet and Dr Carlos Tornero. During this intervention, three new test pits were excavated at different points within the cavity. The excavation work made it possible to identify strata associated with habitation areas and pit-like structures. The archaeological remains recovered, including anthracological remains, faunal remains with evidence of processing and consumption, and handmade ceramic fragments; indicate that the cavity was used at least since the Late Prehistoric period. The new radiocarbon dates obtained from stratified materials recovered in the new test pits also placed the occupations of the cave between the late 5th millennium BC (4260-4045 cal. BC) and the mid 3rd millennium BC (2873-2599 cal. BC), defining a long sequence of prehistoric occupation (Palet et al., 2018; Palet M. et al., 2019). In addition to the archaeological remains, one of the test pits yielded palaeontological remains of brown bear (Ursus arctos), indicating the preservation of a sedimentary deposit in the cave dating back to at least the early Holocene (Tornero et al., 2022b).

3 Methods

3.1 Archaeological field works: 2021 and 2023 campaigns

Between 2021 and 2023, new archaeological fieldwork was carried out at Cave 338. The work was conducted within the framework of the archaeological research project ARRELS: Arrels prehistòriques de la transhumància a l'Alt Ripollès (Prehistoric Roots of Transhumance in the Alt Ripollès county; period 2018-2021, CLT009/18/00048; and period 2022-2025, CLT009/22/00060), under the direction of Dr Carlos Tornero and Dr Eudald Carbonell, and with authorization from the Archaeology Service of the Department of Culture of the Government of Catalonia (Generalitat de Catalunya, Spain). The planning of these new research activities aimed to: (i) obtain new data on Prehistoric settlement in the Eastern Pyrenees, and to characterize the types of economic activities carried out by the communities under study; and (ii) understand the occupational dynamics of Cave 338, including how the space was used and structured, as well as the diachronic changes that took place within it.

Access to the cave is from the Vall de Núria (ca 1970 m a.s.l.) and access to this valley is not possible by motor vehicle, only via the rack railway. The approach from the Monastery of Vall de Núria to Cave 338 takes approximately 45 min along a steep slope. These conditions influenced the excavation, concentrating the excavation days, extending the working hours at the site, and requiring a highly qualified excavation team with considerable experience in extreme fieldwork conditions. Despite the significant difficulties involved in accessing the cave, the work carried out made it possible to implement a new, detailed recording system and a highly rigorous excavation methodology. This included water sieving of the sediments, flotation of the charred fraction, and the systematic collection of sediment samples for various analyses. These tasks had to be carried out at the Monastery of Vall de Núria, used as a base camp, after manually transporting all the excavated sediment from Cave 338.

These new fieldworks made it possible to begin the stratigraphic excavation in extension of the significant prehistoric occupation levels previously identified. Topographic and planimetric surveys of the interior of the cavity were carried out, allowing the planning of future archaeological work. The excavation of the cavity began over an area of 6 m², located at the frontal sector or entrance of the cave. Due to the archaeological significance of the site, and thanks to the to the support of the Archaeology and Paleontology Service of the Government of Catalonia and the Natural Park of the Ter and Freser Headwaters, a metal gate was installed at the entrance of the cave to protect and to help ensure the protection of the site, and the preservation of the deposits, and facilitate future research work, following the corresponding authorization from the Department of Culture of the Government of Catalonia (Generalitat de Catalunya). At the same time, the installation of the minimum necessary infrastructure was carried out (lighting, access improvement, walkways), along with the detailed planimetric and graphic recording of the spaces and walls of the cavity, in preparation for future years of work.

3.2 System for archaeological recording and documentation

The methods and techniques used in carrying out the fieldwork were consistent with the objectives of the project and the excavation. The diachronic perspective in documenting the archaeological deposits of the site was as important as establishing the spatial and contemporaneous relationships between the artifacts and ecofacts recovered during the excavation process. In this regard, the work was carried out with a view to representing the occupation surfaces and activity areas associated with the archaeological remains and structures documented during the excavation.

The excavation process and the recording of the remains were conducted using a three-dimensional Cartesian coordinate system (coordinates x, y and z) and a robotic Total Station theodolite (Trimble 5,503). The depth (z) was measured relative to an arbitrary zero point established on the cave's entrance wall, located approximately 1 m above the surface. This allows to proceed with a cumulative three-dimensional record of both the dispersions and associations among the archaeological remains and across the stratigraphic sequence, using topographic equipment and a digital recording system. The stratigraphy, the archaeological structures, and the position of all archaeological items recovered during excavation was recorded using this system. Further, the excavation work was monitored considering both the dynamics of natural sedimentary deposition and the potential anthropic modifications of the deposits. Thus, each of the items recovered during the excavation could be associated with a specific Stratigraphic Unit (here, UE) of origin and subsequently integrated into its corresponding Archaeological Phase (here, NA).

The archaeological remains were collected by hand and individually recorded once their documentation was completed. The recovery of smaller archaeological remains, as well as of the charred fraction, was ensured through the implementation of water sieving of all the sediment from the excavated substrate while the flotation of the layers richest in macro-botanical material, primarily those forming the fills of the anthropic structures. Subsequently, cleaning, inventory, and classification of all the remains were carried out. The main inventory categories, with their abbreviations to facilitate reference, were charcoal (CB), ceramics (CE), faunal remains (F), lithics (LIT), and malachite (MQ). The excavation process was recorded and documented through planimetry (at scales 1:10, 1:20 and 1:50) and using the Total Station. Additionally, a topographic survey of the cave was conducted, and a detailed plan with contour lines of the cavity was produced using the software Q-GIS (v. 3.16). Additionally, all work areas and their stratigraphic units were photographed and inventoried using a table designed in Excel, which includes the photo number, corresponding area, and stratigraphic unit.

3.3 Ecofactual studies: archaeozoology, anthropological and archaeobotanical analysis

Faunal remains recovered were identified both anatomically and taxonomically using the reference collections of the Archaeozoological laboratory of the Autonomous University of Barcelona (Spain), in addition to several osteological manuals. Whenever the species identification or the precise anatomical identification was not possible due to the fragmentation of the assemblage, the remains were classified according to size categories (large, medium, and small-sized species). Classification of the remains by size was based on the cortical thickness and the overall dimensions of the bone fragment. The quantification of the assemblages was based on the Number of Identified Specimens (NISP) and the Minimum Number of Individuals (MNI) when needed. The latter was calculated considering the estimated age, anatomical representation, and laterality. The estimation of age at death was based on epiphyseal fusion of long bones according to the methodologies proposed by (Barone 1976) and on observations of tooth eruption and wear stages following (Grant 1982) and (Payne 1985). The taphonomic analysis was performed using a USB digital microscope (40x-1000x magnification). A few human remains were also recovered during the excavation procedures. They were identified based on (White et al. 2012), and their age at death was estimated according to (AlQahtani et al. 2010).

Charcoal fragments were recovered during excavation with a combination of two parallel strategies. On the one hand, all charcoal fragments visible during excavation were recovered, registered, and coordinated. On the other hand, all excavated soil was sampled, and water sieved with 2 mm and 0.25 mm meshes, and charcoal remains were recovered by sorting the resulting residue. In the case of pits, additionally, all soil was processed by bucket flotation with the same meshes. In this sense, the methodology implemented in the field guarantees the recovery of all archaeobotanical macroremains. For this first study at the site, only hand-picked charcoal fragments were studied. Taxonomical identification was performed following the standard procedures of the anthracological discipline (Chabal, 1997; Chabal et al., 1999); this is hand breaking the fragments to obtain fresh cuts of the three anatomical planes: transversal (or cross section), longitudinal tangential and longitudinal radial. Wood anatomy was then observed under a reflected-light microscope with dark/bright fields and with objectives of 50x, 100x, 200x and 500x magnifications (Motic Panthera). Taxonomical identification after wood anatomy analysis was assisted with the use of wood anatomy atlases (Schweingruber, 1990) and the wood reference collection of the ArqueoUIB research team (Universitat de les Illes Balears, Spain). For this prospective study a total of 34 fragments were identified to assess the state of preservation of the materials and their taxonomical richness. Moreover, taxonomical information was valuable at this stage to select the charcoal fragments to be radiocarbon dated.

3.4 Artifactual studies: pottery, lithic and mineralogical analysis

The morphological and typological classification of the ceramic vessels was carried out following the framework established by (Balfet et al. 1983). Further, the ceramic analysis was subsequently based on the principles of the chaîne opératoire, understood as a sequence of operations comprising four main phases that may vary among social groups: (i) acquisition and preparation of raw materials; (ii) vessel shaping; (iii) surface finishing and treatment; and (iv) firing techniques (Gelbert, 2003; Livingstone-Smith, 2007; Roux, 2019; Longacre and Skibo, 1994; Stark et al., 2000). Within this framework, the present study focused on the second and third phases (vessel shaping and surface finishing) through the identification of traces left by technical gestures on vessel surfaces. The combination of these traces enables the determination of the manufacturing techniques used to shape the vessels, thereby allowing the reconstruction of this segment of the chaîne opératoire (Roux, 2019).

To study the lithic knapping evidence, a systematic and comprehensive recovery strategy was implemented during fieldwork to ensure the collection of all knapping-related remains. Regardless of the raw material, all lithic fragments were either spatially coordinated in situ or recovered during the water-sieving of the sediment. Given the apparent scarcity of lithic reduction products observed during excavation, the laboratory methodology consisted of a systematic macroscopic review of all collected mineral fractions. This process aimed to distinguish natural clasts (geofacts) from anthropogenic products by identifying diagnostic fracture attributes (e.g., bulbs of percussion, striking platforms, dorsal ridges). For the identified assemblage, the analysis focused on raw material identification and technological attribution following to the principles of the Logical Analytical System (Carbonell and Mora, 1983; Rodríguez, 2004).

A small macrolithic assemblage was also recovered, coming both from the excavation itself and from the surface of the cave. Both technological and functional aspects will be examined with the aim of determining the provenance of the raw materials, identifying anthropogenic modifications to the pieces (including manufacturing and use-wear traces), and assessing their spatial distribution within the cave. These questions will be addressed using a standardized inventory system for recording technological attributes (Risch, 1995; Delgado-Raack and Risch, 2008) and for documenting qualitative characteristics of wear and residues (Adams et al., 2009; Dubreuil et al., 2015), both of which were specifically designed for this type of lithic material.

Finally, rock fragments with patches of a predominantly green mineral were also found during the excavation procedures. These elements were documented in several stratigraphic units both as larger fragments during excavation and as smaller fragments (< 0.4 cm) recovered in large numbers during the sediment sieving. These rocks were preliminary macroscopically examined.

3.5 Radiocarbon dates

Several charred woods remains were selected for radiocarbon dating. The selected charcoal samples were hand-collected, and their positions recorded in the coordinated system during the excavation process. All samples represent the formation moment of the stratigraphic units in which they were recovered, and their stratigraphic relationship, recovery location, and archaeological inventory have been verified before radiocarbon dating. All the charcoal remains dated were previously analyzed anthracologically to determine the corresponding species. The radiocarbon (¹⁴C) measurements were carried out at the CIRAM laboratories in Martillac (France), under the supervision of Patrick Rossetti, Director of the Archaeology Department and Head of Archaeology and Radiocarbon Dating. Radiocarbon determinations were calibrated and Bayesian-modeled in OxCal v4.4 (Bronk Ramsey, 2009) using the IntCal20 atmospheric calibration curve (Reimer et al., 2020). The model was structured as a stratigraphic Sequence of site phasess (NA 2-4), with determinations from each phase grouped within a period and bounded by start/end boundaries to estimate the timing of each occupational unit. To assess internal consistency and accommodate potential aberrant determinations, we applied the ‘General Outlier Model' with an a priori outlier probability of 0.05 per date. Phases durations and gaps between successive phases were derived from the posterior distributions using Difference/Interval queries. In OxCal, Difference and Interval are posterior query functions based on the model's MCMC samples: Difference returns the posterior distribution of the time difference between two model parameters (e.g., start/end boundaries, to estimate durations), while Interval returns the posterior distribution of the gap/hiatus between successive events or groups in a Sequence (Bronk Ramsey, 2009, n.d.).

4 Results

4.1 Description of the findings, stratigraphy, and spatial distribution of the archaeological record

The archaeological work was concentrated within an area of 6 m², corresponding to grid's squares meters F49, F50, G49, G50, H49, and H50. These archaeological investigations allowed for the documentation of a stratigraphic sequence with continuous human presence. In total, the detailed recording system documented 43 stratigraphic units, registering the relationships among them. Further, the archaeological works enable the identification of combustion features as well and archaeological remains in each of the excavated strata. Between the 2021 and 2023 field missions more than a thousand new remains were recovered by hand during the excavation process, including ceramic fragments, knapped lithic industry, anthracological and faunal remains, and green rock remains. Additional small fragments, mostly less than 1 cm, were recovered during the cleaning and water-sieving of the excavated sediments. The maximum archaeological depth reached in units F49 and F50 was approximately 0.60 m (z = −1.65). The stratigraphic sequence of the site has not been fully excavated in depth across the current area of excavation; however, in units F49 and F50, where greater depths were attained, the quantity of archaeological material had become notably scarce by the time excavation work was ended in 2023. Detailed counts of archaeological materials per Stratigraphic Unit are provided in Supplementary Table 1.

Across the entire excavated area, a broadly consistent pattern of sedimentary infilling has been identified. An initial stratigraphic package composed of very shallow, light-colored deposits, strongly affected by ongoing surface-soil erosion and trampling, gives way to a second stratigraphic package formed by more compact, darker layers in which discrete, well-defined, superimposed strata are documented. At the base of this second package, the first pit-like features and/or layers with substantial accumulations of charred remains are recorded. This is followed by a third, thicker, stratigraphic package dominated by a light clay matrix with reddish-brown hues, into which multiple combustion structures are repeatedly cut. In some cases, these features intersect one another, reflecting successive reuse of the same activity area, sufficiently separated in time to allow for the formation and preservation of multiple distinct pits in section. Finally, the clay matrix becomes more compact, with a higher proportion of gravels and small calcareous concretions, defining the fourth and latest sedimentary phase documented.

In total, 23 pit-like combustion structures (EC) were documented, numbered from EC1 to EC21. Within these pit-like combustion structures, very dark-colored strata were documented, characterized by a high concentration of charred plant remains, including notably centimetric charcoal fragments, as well as chunks of rocks with green-mineral patches and, to a lesser extent, faunal remains. These pit-like features are no longer documented in the excavation area beyond a depth of approximately z = −1.55 m. The characteristics of these pit-like features, as well as the type and quantity of archaeological remains documented within them, are described below in their corresponding archaeological phases. Figures 4A, B show two north-south sections of the excavated area unit, illustrating the spatial distribution of the pit structures from units F49-F50 and G49-G50 squares and their relation to the main stratigraphic phases.

Figure 4

4.2 Description of the archaeozoological, anthropological and archaeobotanical remains

A total of 945 faunal remains have been analyzed. Among these, 93 specimens were identified to anatomical element and taxonomic category (NISP). Sheep/goat represents the most frequently documented taxonomic category (n = 59). To a lesser extent, based on a smaller number of remains, other taxa have also been identified, as Sus sp. (n = 5), Canis sp. (n = 4), brown bear (Ursus arctos, n = 1) and Leporidae (n = 11). From a taphonomic perspective, several natural agents have been recorded, with staining by manganese dioxide being the most frequent (21.9% of the remains). Traces of roots (4.3%) and tooth marks caused by carnivores and rodents (0.7%) were also recorded, albeit in smaller percentages. Furthermore, anthropogenic agents have likewise been documented, including burned remains (5.2%), cut marks (0.7%), and intentional fracturing (0.2%).

Within the faunal assemblage, two remains have also been documented and classified as bone or shell artifacts. These consist of a Glycimeris sp. specimen (Inv.No. C338-23-2-535) and a lower incisor of a brown bear (Ursus arctos) (Inv.No. C338-23-7-H49). In both cases, the elements exhibit anthropic perforations at their distal ends and were likely intended to function as pendants or personal ornamentation.

Only two fossil remains belong to Homo sp., a manual phalanx, and an upper deciduous canine.

A total of 34 charcoal fragments recovered by hand were selected for detailed anthracological analysis. The analysis aimed to obtain an initial taxonomic assessment of the identified woody taxa and to select suitable specimens for subsequent radiocarbon dating. The botanical macro remains recovered by flotation and water sieving are still under study and will be made available for further analysis.

The fragments analyzed belongs to three different taxa: Alnus/Betula sp. (n = 1), Pinus t. sylvestris/nigra (n = 28) and Salix sp. (n = 1). Moreover, 1 fragment was determined as Pinus sp. and another one as cf. Pinus sp. because it was not possible to attribute the fragment to specific pine type; and 2 are classed as indeterminable conifers, as they presented wood anatomy alterations that made not possible to attribute a genus or family to those fragments.

4.3 Description of the artifactual and the green mineral remains

The ceramic assemblage under study consists of 333 fragments. The collection exhibits a high degree of fragmentation, with clear evidence of intense post-depositional processes, particularly in fragments recovered from the more superficial stratigraphic units. This fragmentation has hindered the reconstruction of vessel shapes and typologies. Of the total ceramic remains, 294 undiagnostic body sherds, 29 rim fragments, two neck fragments, and eight base fragments were identified. All ceramic remains was made using non-wheel-made pottery, as forming techniques are considered muscle energy techniques, specifically, coil technique; except for four fragments (recovered on surface phases) that have marks on their surfaces and very refined pastes corresponding to rotary energy techniques (kinetic rotary energy).

Thelithic assemblage is quantitatively scarce, confirming the initial field observations. After an exhaustive review of the recovered materials, only three lithic elements presenting clear diagnostic attributes of anthropogenic fracture were identified. This small set consists of two flint and one quartz flakes, corresponding to debitage products devoid of cortical surfaces. No cores or formal retouched tools were documented. The morphological attributes of these flakes are consistent with waste products derived from the resharpening or maintenance of tool edges, rather than primary reduction sequences.

The macrolithic assemblage comprises ten specimens, half of which must be considered artifacts. These include two hammerstones, a knapping by-product, a pick, and a fragment of a grinding tool. The remaining items show no modifications beyond having been transported into the cave as cobbles potentially intended for use as raw material. The represented lithologies are predominantly of local origin (slate, quartz, granite, and marble-like limestone), with a smaller proportion of exogenous materials (basalt and flint).

Over 170 chunks of rock fragments with patches of a predominantly green mineral were recovered by hand during the excavation. They had different sizes, from < 5 cm to 0.5 cm. Additionally, a high quantity of small fragments (< 0.5 cm to 0.1 cm) were recovered during water sieved. These rocks stand out for their vivid colors. Some fragments also have visible smaller patches of a blue mineral and other darker siliceous-looking inclusions visible macroscopically. In terms of the rock fragments with eye-catching green mineral patches, it is proposed that they correspond to malachite, a copper carbonate, according to preliminary descriptions (Palet M. et al., 2019b). Parallelly, the co-occurrent blue patches within these rocks could likely be other copper-rich carbonate (e.g., azurite) or silicate (e.g., chrysocolla). The on-going chemical and microstructural analyses will clarify the specific mineralogy. Crucially, these green and blue minerals are not naturally present in the cave, and the smaller, predominantly green fragments have an angular morphology indicating that they were processed. This evidence together with their abundant presence across archaeological layers suggests that larger rock chunks (predominantly green but with blue and dark mineral patches too) were intentionally brought into the cavity to be crushed. Further analyses will explore potential subsequent stages of mineral processing to understand the role of these minerals within this prehistoric community and their relationship with the excavated pits.

4.4 Radiocarbon dates measurements

To date, 12 radiocarbon measurements have been carried out on charred wood remains. Radiocarbon dates are presented in Table 1. Figure 5 sum up all the radiocarbon dates, after being calibrated and processed by Bayesian-modeled analysis.

Figure 5

The Difference queries indicate that NA4 and NA3 are both long-lasting phases, with modeled durations of 520-1,290 years for NA4 and 1092-1515 years for NA3 (68.3% posterior ranges; 95.4%: 344-2,388 and 1,001-2,002 years, respectively). The Interval queries suggest that the NA4-NA3 transition most likely includes a centuries-long hiatus (237-765 years, 68.3%), although near-continuity cannot be excluded at 95.4% (0-823 years). By contrast, the gap between NA3 and NA2 is more substantial and better supported, with a most probable hiatus of 971-1522 years (68.3%; 95.4%: 279-1572 years), consistent with a marked break between the end of NA3 and the onset of NA2.

5 Discussion

5.1 Site's occupation sequence

Archaeological Phase 1 (NA 1)

In all excavation units, the archaeological stratigraphic sequence begins with very thin, outward-dipping layers characterized by low compaction and high organic content, indicative of limited post-depositional consolidation. These deposits form part of the present-day cave floor and contain abundant loose clasts of variable size, derived from roof and wall spalling and accumulated through gravity-driven processes. The layers display light brown coloration and frequent plant intrusions. This initial stratigraphic horizon has been designated Archaeological Phase (NA 1). This is present throughout the entire excavated area, in UEs 0, 1, 3, 4, and 5, with a variable depth between 5 and 15 centimeters.

The chronological attribution of this first phase is difficult to determine. From an interpretive standpoint, this phase appears to consist of several occupation episodes that are relatively recent in age. This is supported by the presence of archaeological artifacts within these deposits that are attributable to different origins and/or to historical or recent periods. Macrolithic tools recovered from the immediate surroundings of the cave, i.e., a pick made of flint and a grinding tool made of basalt, exhibit specific characteristics that are absent from those managed inside the cave. On the one hand, unlike the items found within the cave, the raw materials represented among the surface finds do not originate from the local vicinity but from exogenous geological sources located several tens of kilometers away. On the other hand, the only artifact showing frictional wear and likely associated with food preparation (grinding tool) was recovered from the NA 1 surface layer. Most of the ceramic remains recovered come from UE 0 and UE 1. This assemblage is highly diverse and is dominated by hand-made vessel fragments (n = 29), although a few wheel-thrown sherds were also documented (n = 4). Most fragments are undiagnostic body sherds, although bases (n = 1), neck fragments (n = 1), and rim fragments from different vessels (n = 5) were also identified. These latter elements allow for the recognition of a wide range of vessel forms, including short and well-defined necks, globular bodies, rounded bases, and spherical containers. All these forms are known from later prehistoric contexts in the north-eastern Iberian Peninsula, although a more precise chronological attribution is difficult to establish. In addition to ceramic fragments attributable to Late Prehistoric periods and isolated historical episodes, several objects associated with later, historical, or modern contexts have been documented, including a bronze spiral fragment and others. These finds, together with the presence of wheel-thrown ceramics, also indicate a later and more recent use of the cavity.

The nature and intensity of these more recent occupations of the cavity, currently grouped within NA1, are difficult to assess; however, they appear to represent sporadic uses, temporally spaced yet relatively frequent, which overlie the earlier prehistoric occupations. It is likely that the characteristics of these more recent activities within the cavity also contributed to the preservation of the underlying sedimentary deposits, for instance, because they did not produce extensive ground alteration or significant disruption of earlier sedimentary layers. A more precise chronological attribution of the remains has not been feasible.

Archaeological Phase 2 (NA 2)

The NA 2 has been defined based on a set of well-delimited stratigraphic units, exhibiting consistent preservation and integrity, and located stratigraphically beneath the superficial layers that constitute phase 1. At the base of this occupation, a combustion structure (EC), called EC3, was also documented. The homogeneous radiocarbon dates obtained from different UEs within the phase further confirm the consistency and integrity of this archaeological horizon.

NA 2 has been documented throughout the entire excavation area and displays a variable thickness of 12-22 centimeters. The principal strata exhibit heterogeneous dip patterns: toward the exterior of the cavity along the eastern margin of the excavation, and toward the interior along the western margin. This variability in layer distribution, inclination, and thickness (as well as in the boundaries and lateral continuity of individual strata) reflects the influence of roof- and wall-collapse blocks that detached and shifted across the cave floor. These collapse events locally altered microtopography and, consequently, the sedimentation dynamics and depositional geometry of the level. During excavation, numerous fallen blocks were recorded, some exceeding 50 centimeters in length, distributed across much of the excavation surface, particularly in areas where the current ceiling height is greatest.

In addition to the laminated strata UE 2, 6, 7, 8, 9/9N, 10/10N, 13, 20 and 21, a well-defined combustion structure (EC 3) was documented, measuring about 25 in length and 22 in width, filled with a dark layer (UE 16). Within this last pit, charcoal remains, and fragments of green mineral rocks were also recovered.

The radiocarbon dates obtained from EC 3 (infill UE 16) and from stratigraphic unit UE 21 are very similar (Sample CIRAM-13164, from UE 16 dates 1,109-908 cal BC; Sample CIRAM-13165, from UE 21 dates 1010-890 cal BC). Both formations, as well as the stratigraphically associated UEs, can be considered homogeneous components of the same horizon. Other UEs have a more tentative attribution to this chronological horizon, at least until more precise dates are obtained for each of them, particularly those overlying the dated units and located at the top of the level (UEs 2, 8, 9/9N and 10/10N), although all remain stratigraphically below NA 1.

The NA 2 context yielded 574 faunal remains, resulting in a NISP of 33, most of which corresponded to caprine (n = 16), followed by small carnivores (n = 5), suids (n = 4), leporidae (n = 4), birds (n = 3) and Canis sp. (n = 1). Anthropogenic modifications caused by burning are detected on the 4% of the remains, whilst the 0.7% presents cut marks. The latter were recorded in UEs 7 (n = 2), 2 (n = 1) and 6 (n = 1). The remains presenting signs of combustion were mainly located in UE 7 (n = 25) and, to a lesser extent, in UE 2 (n = 3), UEs 16 and 9 (n = 2), and UEs 13 and 6 (n = 1). Furthermore, two remains from UE 7 display a greenish colouration.

Seven analyzed hand-picked charcoal fragments belong to NA 2 contexts (5 from UE 16; 2 from UE 21). All of them identified as Pinus t. sylvestris/nigra. This species accounts for the most important part of the assemblage studied, which is also the most common case in other archaeological sites in the Pyrenees with similar chronologies at lower or mid altitudes (Euba, 2008; Obea et al., 2021; Obea, 2024). However, anthracological data from north-eastern Iberia shows that since the Late Pleistocene, mountain pines (Pinus t. sylvestris/nigra) dominated forested landscapes, including coastal and low-altitude areas. With the climatic shift at the onset of the Holocene, these pines persisted only in a relict and spatially limited manner during the Early Holocene, progressively declining as Mediterranean forests expanded and human impact increased, eventually becoming restricted to high-altitude areas from the Middle Holocene onwards (Allué et al., 2017). According to this, we can assume that the Pinus t. sylvestris/nigra charcoal fragments analyzed corresponds to the residues of the use of firewood collected in the local forests of this species, which persisted dominating the high-altitude forests of the Eastern Pyrenees until the beginning of the first millennium cal BC.

In NA 2, a total of 161 ceramic fragments were recovered from stratigraphic units 2, 7, 9 and 21. Most of these correspond to undiagnostic body sherds (n = 152), although rim fragments (n = 5) and vessel bases (n = 4) were also identified. The great majority of fragments are plain, lacking any form of decoration, and notably, no handles or attachment elements are present. Only one group of ceramic vessels were identified within the assemblage: vessels with necks and flat bases (Figure 6A). Regarding the technological data, manufacturing processes observed display production techniques such as coils by drawing, with beveled joining of the clay coils, as evidenced by the numerous, sometimes elongated, finger impressions (Figure 7). Noteworthy is the combination of preforming techniques, in which most vessels exhibit a beating technique used to achieve the final shaping of the container.

Figure 6

Figure 7

No knapped lithic remains were recovered from NA 2, but some pieces associated with macrolithic tools were found: two cobbles, made of slate (Inv.No. 248) and granite (Inv.No. 630) respectively, both in UE 7. Neither shows evidence of use as a tool nor of any intentional modification for such a purpose. Green mineral fragments were recovered in NA2, totalling 56 specimens. They were recovered in the most extensive stratigraphic units (i.e., UE 6 n = 24; UE 7 n = 13), but also within the EC 3 basin, in its UE 16 fill (n = 3).

Finally, the two personal ornamentation remains found in Cave 338 are also stratigraphically associated with the NA 2 horizon (Figura 8B). In the first case, it is an elongated pendant made from the ventral part of a Glycimeris sp. shell (36,71 mm length, 11,48 mm width, and 1,94 mm thickness). It is perforated centrally at the apical area but shows a fracture that makes it difficult description of its original morphology. The surface and edges are polished, giving rounded shapes, especially in the basal area. It exhibits a pronounced curvature. Although a concretion patina covers the entire surface, the shell's ribs can still be clearly observed. Pendants of this lozenge-shaped form, made from Glycymeris sp., have been found in Catalonia at Can Roqueta-II (Sabadell, Barcelona), in the complex burial pit or hypogeum E-459 and niche C. They were associated (n = 2) with an inhumation of 11 individuals (8 subadults and 2 adults) (Oliva, 2004). Similar examples have also been reported in the Pyrenean chamber tomb of Santes Masses (Pinell) (Castany et al., 1994). In both cases, however, this type of pendant is associated with chronologies approximately corresponding to the beginning of the 2nd millennium BC. The second object is a lower third canine of a brown bear (Ursus arctos). It preserves its entire original morphology, including the enamel crown and the root (length 17.74 mm; width 5.72 mm; thickness 5.41 mm). It exhibits a perforation at the distal part of the root, which also shows signs of polish. In this case, no similar finds associated with chronologies of the 2nd or 1st millennium BC are known in the study area.

Archaeological Phase 3 (NA 3)

To date, Archaeological Phase 3 (NA 3) constitutes the most substantial and archaeologically meaningful occupation documented at Cave 338. It is represented by a sedimentary package with internally homogeneous characteristics (UE 11), within which numerous combustion features (pit structures) have been recorded. These features are dispersed across the entire 6 m² excavation area and occur in vertically superposed positions. The relative integrity of the sedimentary context and the good preservation of the associated archaeological material enable the isolation and characterization of this occupation as a discrete depositional and behavioral episode, clearly distinguishable from the overlying archaeological phase.

Between 2021 and 2023, a total of 22 combustion structures associated with NA 3 were documented. All of them are superimposed and/or cut into UE 11: a clay matrix with yellowish to orange hues, reaching a maximum thickness of 30 to 35 cm and extending across the entire excavation area. The pits show a homogeneous distribution throughout the portion of NA 3 excavated to date, and in several cases, superposition between structures was observed (e.g., EC 12c over EC 12b and both over EC 12a). All combustion structures display similar characteristics in both shape and depth. In plan view, the pits exhibit circular to sub-elliptical outlines with irregular boundaries. Their depths vary, ranging between 10 and 22 cm at their deepest point. The internal volume of the pits is irregular, although all of them show an approximately conical morphology, with a single deepest point.

Given their state of preservation, it is likely that in some cases neither the original volume of the pits nor their complete sedimentary fill has been fully preserved. Continuous erosive processes affecting the clay matrix of UE 11, particularly those associated with seasonal freeze, thaw cycles, must have altered their initial characteristics. Nevertheless, the fact that these features were constructed inside the cavity, although positioned close to the entrance, likely contributed significantly to their preservation to the present day.

The occupation dynamics during this period show clear continuities, expressed in both the organization and use of the space. This is reflected in the construction and recurrent use of pit-shaped combustion structures over and within the UE 11. Further, infill deposits of combustion structures are characterized by dark colors with abundant charred remains and ash, as well as the presence of green mineral fragments (likely malachite) within their fills.

The chronology of NA 3 is constrained to an interval between 3,464 cal BC to 2,532 cal BC at 68% (Bayesian-modeled). This temporal range is based on eight radiocarbon dates obtained from charcoal samples that were hand-recovered from the infill of structures EC 7, EC 12a, EC 12b, EC 12c, EC 14, EC 16, EC 18 and EC 19. While EC7 (Sample CIRAM-13167, from infill UE 22 dates 2580-2451 cal BC) represents a higher and final position within the phase; the EC19 (Sample CIRAM-16466, from infill UE 37 dates 3640-3526 cal BC) represent the early episode. Several radiocarbon dates from infills on EC 12a, EC 12b, EC 12c, EC 16 and EC 18 suggest a continuous use of the cave during several centuries.

A total of 268 faunal remains associated with the NA 3 context were recovered, with a NISP of 16. The assemblage is primarily composed of sheep/goat remains (n = 12) and also includes Canis sp. and Leporidae (n = 1 in both cases). From a taphonomic perspective, 8.6% of the remains exhibit signs of combustion and 0.3% cut marks. The burnt remains mainly come from UE 11, except for three remains recovered in infill UEs 27, 29, and 32. As for cut marks, only one was found on a remain from UE 11. Four greenish-colored remains come from UE 11 (n = 2) and from UEs 15 and 28 (n = 1 each).

The two human remains recovered in Cave 338 belong to UE 11 and NA 3. The two skeletal remains could belong from the same individual. The age of the individual is estimated in phase II of childhood, approximately 11.5 years old.

In NA 3, the analyzed charcoal remains (n = 21) are again dominated by Pinus sylvestris/nigra. All samples belong from infill soils of pit-combustion structures (UE 27 n = 2; UE 28 n = 1; UE 30 n = 1; UE 32 n = 3; UE 33 n = 3; UE 35 n = 4; UE 37 n = 4). However, a quantitative approach to the anthracological results is not currently appropriate due the small number of fragments analyzed that will have to be increased by studying the fragments recovered through flotation. In any case, a single fragment of Alnus/Betula sp. was identified in one of the structures (UE 35). It was not possible to differentiate between both genus due to their very similar wood anatomy and the size of the fragment. In the current vegetation of north-eastern Iberia Alnus is common in riparian forests north of the Llobregat river, and Betula commonly growth in mixed forests with Fagaceae or with Pinus sylvestris L. (Bolòs et al., 2005). As both cases could be possible, its presence could reflect that this species was only sporadically collected for fuel use either from the same areas where Pinus t. sylvestris/nigra firewood was gathered, or from riparian forests. In any case, the recurrence and predominance of Pinus t. sylvestris/nigra in the charcoal assemblage studied so far is coherent with the only parallel known for this region, chronology and altitude. At the Aigols Podrits site, a total of 100 charcoal fragments from a combustion structure dated to the late fourth millennium cal BC were analyzed. Of these, 96 were identified as Pinus t. sylvestris/nigra (published under the synonym Pinus mugo/uncinata), and 4 as Pinus cf. sylvestris/nigra (published under the synonym Pinus cf. mugo/uncinata) (Pescini and Fossati, 2021). The anthracological data from both sites, then, suggest that Pinus t. sylvestris/nigra forests accounted for the most relevant forests in the region, from which the prehistoric groups obtained the most important part of their energetic resources.

A total of 55 ceramic fragments were recovered from UEs 21/11, 7/11, 11, 34 and 36 linked to NA 3. Most of these correspond to undiagnostic body sherds (n = 49), although rim fragments (n = 5), and vessel bases (n = 1) were also identified. All ceramic remains show similar aesthetic and external attributes than those recorded in NA 2: all fragments are plain; no decoration is recorded and handles or attached elements are not present. However, and in clear contrast to NA 2, the ceramic vessels identified from NA 3 all correspond to hemispherical forms (Figure 6B). Here, as well as in NA 2, although the number of potteries remains and ceramic vessels is still low, similar production techniques such as coils by drawing, with beveled joining of the clay coils, as evidenced by the numerous finger impression, are also present in this assemblage, suggesting similar traditions in manufacturing processes (Figure 7).

The three lithic records at Cave 338 belongs to NA 3. Two flakes (one in flint, a second in quartz) were recovered in UE 11; the third flake (flint) was recovered from infill UE 30 of pit EC 14. They provide significant insight into the logistical strategies of the groups occupying the cave. The complete absence of cores and primary reduction waste indicates that stone tool production was not a localized activity. The raw materials identified, specifically flint, are likely exogenous to the immediate geological setting of the cave. It potentially implies that the inhabitants accessed the high-mountain zone with a “personal gear” of finished tools manufactured elsewhere, which were curated and maintained during their stay. The dichotomy between the presence of processed faunal remains (indicating the use of cutting edges) and the absence of the tools themselves supports the interpretation of Cave 338 as a seasonal occupation. In this context, the toolkits were likely transported away by the groups upon their departure, leaving behind only the small fragments as evidence of their maintenance.

The macrolithic remains from NA 3, comprising six specimens from UE 11, 21 and 23, contain the majority of the site's items, although three of these are unmodified cobbles made of slate (Inv.No. C338-11-769), quartz (Inv.No. C338-11-745), and marble (Inv.No. C338-11-1004). The remaining three include a marble flake obtained by knapping (Inv.No. C338-11-1003) and two slate hammerstones (Inv.No. C338-23-11-727 and Inv.No. C338-23-11-1002) bearing early-stage percussion traces (Figure 6B). An important question to explore is whether these tools were functionally associated with the tasks carried out in the combustion pits. Finally, green mineral fragments (likely malachite) were recovered by hand in NA 3 (n = 108), both in UE 11 (n = 43) and in the infill deposits of combustion pits (Figure 8A). Indeed, green mineral fragments were recovered by hand in all combustion pits, although in some of them the number of remains is higher (for example: in EC 2, n = 9; EC 6, n = 9; or EC 18, n = 8). The smaller, numerous fragments recovered from the sediment sieving, still under study, will likely considerably increase the number of these elements documented in this level.

Figure 8

Archaeological Phase 4 (NA 4)

Below the deepest pits excavated to date and attributed to NA 3, starting from a depth of approximately z = −1.55 m, the characteristics of the dominant clay matrix become more compact, with increased carbonate content and a mixed presence of small gravels, giving rise to UE 11b. To date, this fourth archaeological phase has only been documented in units F49 and F50 where deeper layers were reached. No archaeological remains were recovered in these units, except for charcoal fragments. These correspond to centimetric charcoal pieces hand-collected in situ, forming spatially coherent dispersions consistent with a single episode of cave use. Again, Pinus t. sylvestris/nigra account for the most important number of analyzed fragments (n = 6). However, a fragment of Salix sp. has also been identified. This taxon grows in riparian forests (Bolòs et al., 2005), which could again suggest the occasional use of riparian woods as fuel in the site.

At present, it is not possible to precisely characterize the types of occupations represented in NA 4. The radiocarbon dates obtained to date indicate a chronological horizon between 5,000-4,832 cal BC and 4,500-4,360 cal BC (Sample CIRAM-16468, and CIRAM-16467, respectively). Based on the initial test excavations carried out in the cavity (Palet M. et al., 2019), charcoal remains were also documented at comparable depths; in that case, the charcoal was associated with a thick-walled, hand-made ceramic base fragment. In those earlier excavations, charcoal dated from contexts identical to the NA4 horizon described here yielded a chronology of 4100-4000 cal BC (Palet M. et al., 2019), supporting the existence of an earlier prehistoric use of the cavity prior to NA 3.

5.2 Prehistoric occupationss of cave 338 in the frame of the Eastern Pyrenees

Cave 338 constitutes a new prehistoric record for the Eastern Pyrenees. To date, three reliable archaeological phasess associated with different periods of prehistory (NA 2, 3, and 4) have been documented. The cave shows intensive use during the period 3464-2532 cal BC (NA 3), and probably thereafter in different phases during Late Prehistory (NA 2) and Protohistoric or Early Historic periods (NA 1). The still scarce evidence relating to NA 4 indicates that the cave was already occupied from the beginning of the 5th millennium cal BC at least.

The discovery of this new archaeological site in a high-altitude environment above 2,000 m.a.s.l., with evidence of intensive and sustained occupation over several millennia, calls for a reassessment of several interpretations previously proposed in the existing literature. Firstly, these data significantly modify the geographical framework of high-mountain prehistoric archaeology, at least for the Eastern Pyrenees, and potentially for the Pyrenean range. To date, prehistoric occupations higher than 2000 m a.s.l. have generally been characterized by sparse and poorly preserved archaeological records, mainly due to limited sedimentary accumulation and increased exposure of deposits to erosional processes and are typically associated with small rock shelters or crevices among fallen blocks (Orengo et al., 2014; Gassiot et al., 2024, 2017; Palet et al., 2014, 2017; Martzluff, 2024). In contrast, the Cave 338 presented here corresponds to a medium-sized cave containing a thick deposit and a well-preserved archaeological sequence generated by repeated and prolonged occupations. This evidence indicates a more complex pattern of high-altitude landscape use and suggests that the earliest significant human presence in alpine environments occurred earlier than previously hypothesized. In summary, Cave 338 provides compelling evidence for sustained human interest in Pyrenean high-mountain landscapes for over 7,000 years.

Secondly, due to its location (2,235 m.a.s.l., high-mountain environment), Cave 338 suggests a markedly seasonal pattern of use, with occupations constrained by the harsh conditions of the colder seasons during the annual cycle and by the accessibility and availability of resources. Evidence supporting this interpretation includes medium-sized combustion features containing abundant, crushed green mineral fragments (likely malachite), which were likely exploited and possibly processed during short episodes and in a highly localized manner in the immediate vicinity of the cave. The limited presence of faunal and botanical remains consumed directly on site, together with the predominance of vessels lacking a culinary function recovered from NA 3, further supports the seasonal and intermittent nature of these occupations. Consistent with this interpretation is also the near absence of lithic industry, represented only by small retouch flakes associated with the maintenance of tools. Nonetheless, this seasonal and episodic use of the site must have been embedded within a broader and carefully structured repetitive or repeated mobility system, in which the exploitation of high-mountain resources was planned and relied on detailed knowledge of the environment, access routes, and associated risks.

Finally, its historical dimension. Cave 338 reveals the existence of complex and well-structured patterns of occupation linked to the use of high-mountain environments from the mid and late 4th millennium cal BC onwards. The repeated intense use of the cave (at least between ~3400 and ~900 cal BC) indicates the accumulation of knowledge regarding the site, its topographic setting, its surrounding landscape, and its resources, which must have been transmitted across generations. Evidence for this continuity includes, once confirmed the mineralogical characterization, the repeated exploitation of malachite from the late 4th millennium to the early 1st millennium cal BC, as well as the persistence of ceramic manufacturing traditions. Future research will aim to define technological practices and raw material procurement strategies to identify patterns of continuity or discontinuity among the groups occupying the site during this long chronological interval. Moreover, the discovery of human remains, although so far only isolated, could suggest the use of the cavity as a funerary deposit, demonstrating the importance of the cave's location for prehistoric communities. This practice is attested in the study area for the same chronological periods in the caves of Roc de les Orenetes (1,836 m.a.s.l.) (Carbonell et al., 1985; Moreno Ibáñez et al., 2024a,b) and Cova del Cingle del Cullell (1,596 m.a.s.l.,) (Soler et al., 2024, 2025), both located only a few kilometers from Cave 338, although at a much lower altitude. Far from representing testimonial or opportunistic occupations, Cave 338 demonstrates deliberate planning and organization in the use of Pyrenean high-altitude environments during periods in which this type of archaeological evidence had previously been documented almost exclusively in valley floors and lowland areas.

6 Conclusion

The location of the cavity, together with the characteristics of the archaeological assemblage, makes the Cova 338 site unique within the entire Pyrenean region. Other prehistoric sites documented at similar elevations in different areas of the Pyrenean range are in small rock shelters or in open-air structures with stone foundations, typically yielding scarce and poorly preserved archaeological records due to limited sediment deposition and increased exposure of their weather deposits to the elements. The excellent preservation observed at Cave 338, together with the condition of the recovered materials, enables the recovery of a unique and exceptional archaeological record at this elevation and for these chronologies.

The available data for NA 3, currently the best-represented phase, indicate that the cave was occupied on a seasonal basis, with activities involving subsistence (i.e., food consumption, pit structures), tools maintenance tasks as well as the selective exploitation of specific resources that may be exclusively available in the area, particularly the rock with eye-catching green mineral patches which most likely are malachite. The dominant presence of hemispherical vessels associated with the exploitation of this resource suggests that their use may be related to this activity. Ongoing studies on use-wear and the uses of macro-lithic tools will potentially make it possible to determine whether they are also related to its processing. Further, these activities may also have relied on the use of other locally available resources, such as firewood and possibly game. In addition to these functions, we cannot overlook the discovery of human remains in NA 3, at least those of a childhood, and the possibility that the cave may also have functioned as a funerary deposit during this phase. Future excavation work in NA 3 will need to determine whether any disruptions or clear continuity occurred during the period from ca ~3,400 to ~2,500 cal BC; however, the data currently available show no evidence of discontinuity or change throughout a phase in which occupation is closely associated with combustion features containing abundant, crushed green mineral fragments.

Excavations are expected to continue over the coming years. The extension of the excavation area in the cave entrance area will allow a better understanding of the temporal dynamics of occupation, a more accurate characterization of NA 4, and the potential identification of more discrete events within NA1 and NA 2. These future research efforts will also contribute to a deeper palaeoecological understanding of high-mountain environments in this region during the period of cave use (through pollen, anthracological, and carpological analyses), as well as to a more detailed characterization of strategies for the management and exploitation of animal resources. Finally, and as a matter of urgency, the ongoing work will enable a more precise characterization of the rock with green mineral patches which we hypothesize are malachite. Its characterization will allow approaching the processing stage(s) occurring at the cave as well as determining the final purpose of this activity. To this end, systematic surveys in the vicinity of the cave are also required to locate and identify the source area from which the mineral was obtained.

Overall, these data will contribute to a clearer understanding of the importance and significance of Cave 338 in the current state of research on high-mountain prehistoric occupations in the Eastern Pyrenees and the Pyrenean range.

References

• 1

  AdamsJ.DelgadoS.DubreuilL.HamonC.PlissonH.RischR. (2009). “Functional analysis of macro-lithic artefacts,” in Costa, Non-flint Raw Material Use in Prehistory: Old Prejudices and New Directions, eds. F. Sternke, L. Eigeland, and L. [Oxford: Archaeopress (BAR International Series 1939)], 43–66.

  • Google Scholar

• 2

  AlluéE.Picornell-GelabertL.DauraJ.SanzM. (2017). ‘Reconstruction of the palaeoenvironment and anthropogenic activity from the Upper Pleistocene/Holocene anthracological records of the NE Iberian Peninsula (Barcelona, Spain)', Quat. Int.431, 1–18. doi: 10.1016/j.quaint.2016.10.024

  • CrossRef
  • Google Scholar

• 3

  AlQahtaniS. J.HectorM. P.LiversidgeH. M. (2010). Brief communication: the London atlas of human tooth development and eruption. Am. J. Phys. Anthropol. 142, 481–490. doi: 10.1002/ajpa.21258

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 4

  AntolínF.NavarreteV.SañaM.ViñertaA.GassiotE. (2018). Herders in the mountains and farmers in the plains? A comparative evaluation of the archaeobiological record from Neolithic sites in the eastern Iberian Pyrenees and the southern lower lands. Quat. Int.484, 75–93. doi: 10.1016/j.quaint.2017.05.056

  • CrossRef
  • Google Scholar

• 5

  BalfetH.Fauvet-BerthelotM. F.MonzónS. (1983). Pour la Normalisation de la Description des Poteries. Paris: Éditions du CNRS.

  • Google Scholar

• 6

  BaroneR. (1976). Anatomie Comparée des Mammifères Domestiques. Vol. 1. Ostéologie. Paris: Vigot Frères.

  • Google Scholar

• 7

  BolòsO.VigoJ.MasallesR. M.NinotJ. M. (2005). Flora Manual dels Països Catalans. 3rd edn. Barcelona: Edicions Pòrtic.

  • Google Scholar

• 8

  Bosch GimperaP. (1919). Prehistòria Catalana: Edats de la Pedra i dels Metalls, Colonització grega i Etnografia. Barcelona: Editorial Catalana.

  • Google Scholar

• 9

  Bosch GimperaP. (1923). Resultats de l'exploració de coves de Catalunya (1915–1920). Anu. Inst. Estud. Cat. VI. pp. 473–481.

  • Google Scholar

• 10

  Bronk RamseyC. (2009). Bayesian analysis of radiocarbon dates. Radiocarbon51, 337–360. doi: 10.1017/S0033822200033865

  • CrossRef
  • Google Scholar

• 11

  Bronk RamseyC. (n.d.). “Analysis operations and models (Queries: Difference, Interval),” in OxCal Online Manual. Oxford Radiocarbon Accelerator Unit, University of Oxford. Available online at: https://c14.arch.ox.ac.uk/oxcalhelp/hlp_analysis_oper.html/ (Accessed March 4, 2026).

  • Google Scholar

• 12

  BusquetsR. (2020). ‘Diari d'excavació del Roc de les Orenetes (1973)', Ann. Centre Estud. Com. Rip. 30, 13–37.

  • Google Scholar

• 13

  CarbonellE. (1978). L'arqueologia de les rodalies de Queralbs (Freser mitj‘a). Muntanya697, 131–134.

  • Google Scholar

• 14

  CarbonellE. (dir.). (1985). Sota Palou (Campdevànol): un Centre D'intervenció Prehistòrica Postglaciar a l'aire lliure. Girona: Centre d'Investigacions Arqueològiques de la Diputació de Girona (Sèrie Monogràfica, 5).

  • Google Scholar

• 15

  CarbonellE.CulíN.BusquetsR. (1976). El Paleolític inferior i mitjà a la conca del Freser. Cypsela1976, 23–27. Available online at: https://raco.cat/index.php/Cypsela/article/view/120177

  • Google Scholar

• 16

  CarbonellE.CulíN.BusquetsR. (1978). El Paleolític inferior i mitjà a la conca del Freser. Ann. Centre Estud. Com. Rip. 1. 23–27.

  • Google Scholar

• 17

  CarbonellE.CulíN.BusquetsR. (1985). ‘Els pobladors prehistòrics de la vall del Freser', in Queralbs. Bar. Centre Excurs. Cat. Editorial Montblanc-Martín21–32.

  • Google Scholar

• 18

  CarbonellE.MoraR. (1983). Sota Palou (Campdevànol): Excavacions Arqueològiques. Girona: Diputació de Girona.

  • Google Scholar

• 19

  CarrerasA. (1964). Los fenómenos cársticos del Valle de Nuria (Pirineos catalanes). Cavernas, no. 3, Revista del Grup d'Espeleologia de Badalona.

  • Google Scholar

• 20

  CarrerasE.TarrúsJ. (2017). ‘El Quintà de les Planes (Sant Jaume de Frontanyà, Berguedà) i Pardinella (Gombrèn, Ripollès) i les cistes neolítiques amb túmul del Prepirineu i altiplans de Catalunya', Ann. Inst. Estud. Gir. 8, 129–144.

  • Google Scholar

• 21

  CastanyJ.Guerrero i SalaL.Estany i MorrosI. (1994). La Cambra pirinenca de Santes Masses: un sepulcre col·lectiu del Bronze antic. Barcelona: Generalitat de Catalunya.

  • Google Scholar

• 22

  ChabalL. (1997). Forêts et sociétés en Languedoc (Néolithique final–Antiquité tardive). Paris: CNRS/Éditions de la Maison des Sciences de l'Homme. doi: 10.4000/books.editionsmsh.43380

  • CrossRef
  • Google Scholar

• 23

  ChabalL.FabreL.TerralJ. F.Théry-ParisotI. (1999). “L'anthracologie,” in La Botanique. Paris: Errance, A. Ferdière (Paris: Éditions Errance), 43–104.

  • Google Scholar

• 24

  Colominas RocaJ. (1927-1931). Sepulcre de Pardinella (Gombrèn, Ripollès). Ann. Inst. Estud. Cat. 8:11.

  • Google Scholar

• 25

  de VallèsJ. (ed.). (2009). Catàleg espeleològic de Catalunya, Vol. 3.Barcelona: Espeleo Club de Gràcia and Federació Catalana d'Espeleologia, 381.

  • Google Scholar

• 26

  Delgado-RaackS.RischR. (2008). “Lithic perspectives on metallurgy: an example from Copper and Bronze Age south-east Iberia,” in Prehistoric Technology, 40 Years Later, eds. L. Longo, and N. Skakun Oxford: Archaeopress (BAR International Series 1783), 235–251.

  • Google Scholar

• 27

  Díez-CansecoC.Ramírez PedrazaI.RoviraM.CarbonellE.TorneroC. (2022). “Intervencions 2020–2021 a la Bauma dels Fadrins (Queralbs, Ripollès): un nou jaciment amb seqüència mesolítica-neolítica a la província de Girona,” in Setzenes Jornades d'Arqueologia de les Comarques de Girona, eds. J. Burch, R. Buxó i Capdevila, J. Frigola, M. Fuertes, and M. Mataró Pladelasala (Girona: Documenta Universitaria), 40–43.

  • Google Scholar

• 28

  Díez-CansecoC.Ramírez-PedrazaI.CarbonellE.TorneroC. (2020). “Nous treballs de recerca al Torrent d'Estremera i a la Bauma dels Fadrins (Queralbs, Ripollès),” in Quinzenes Jornades d'Arqueologia de les Comarques de Girona. (Catalonia: Castelló d'Empúries), 43–46.

  • Google Scholar

• 29

  Díez-CansecoC.Ramírez-PedrazaI.ExpósitoI.MoralesJ. I.Picornell-GelabertL.CarbonellE.et al. (2023). “La Bauma dels Fadrins (Queralbs, Girona): nuevos datos para el estudio de las primeras comunidades campesinas en los Pirineos orientales,” in Actas del VII Congreso del Neolítico en la Península Ibérica, ed. D. García Rivero (Sevilla: Editorial Universidad de Sevilla), 41–59.

  • Google Scholar

• 30

  Díez-CansecoC.RoviraM.Del ValleH.CarbonellE.TorneroC. (2024). “Intervencions 2022–2023 a la Bauma dels Fadrins (Queralbs, Ripollès): noves dades sobre la seqüència d'ocupació de l'abric,” in Dissetenes Jornades d'Arqueologia de les Comarques de Girona. (Girona: Documenta Universitaria), 31–34.

  • Google Scholar

• 31

  DubreuilL.SavageD.Delgado-RaackS.PlissonH.StephensonB.de la TorreI. (2015). “Current analytical frameworks for studies of use-wear on ground stone tools,” in Use-Wear and Residue Analysis in Archaeology. Manuals in Archaeological Method, Theory and Technique, eds. J. Marreiros, J. Gibaja Bao, and N. Ferreira Bicho (Cham: Springer). doi: 10.1007/978-3-319-08257-8_7

  • CrossRef
  • Google Scholar

• 32

  EubaI. (2008). Análisis Antracológico de Estructuras Altimontanas en el Valle de la Vansa-Sierra del Cadí (Alt Urgell) y en el valle del Madriu. Andorra: Explotación de recursos forestales del Neolítico a época moderna. (Unpublished PhD). Universitat Rovira i Virgili.

  • Google Scholar

• 33

  FabrésA. (2021). Grober Xaialsa. Recerca prehistòrica a la Vall del Freser. Ann. Centre Estud. Com. Rip.2020, 2–23.

  • Google Scholar

• 34

  GassiotE. (ed.). (2016). Arqueología del Pastoralismo en el Parque Nacional d'Aigüestortes i Estany de Sant Maurici: Montañas humanizadas. Madrid: Organismo Autónomo Parques Nacionales.

  • Google Scholar

• 35

  GassiotE.MazzuccoN.Clemente-ConteI.Rodríguez-AntónD.ObeaL.QuesadaM.et al. (2017). “The beginning of high mountain occupations in the Pyrenees: human settlements and mobility from 18,000 cal BC to 2000 cal BC,” in High Mountain Conservation in a Changing World, eds. J. Catalan, J. Ninot, and M. Aniz, M. (Cham: Springer).

  • Google Scholar

• 36

  GassiotE.MazzuccoN.ObeaL.TarifaN.AntolínF.ClopX.et al. (2015). La Cova del Sardo de Boí i l'explotació de l'alta muntanya als Pirineus occidentals en època neolítica. Trib. Arqueol. 2012–2013, 199–218.

  • Google Scholar

• 37

  GassiotE.Rodríguez AntónD.Pèlachs MañosaA.Pérez ObiolR.Julià BruguésR.Bal-SerinM.-C.et al. (2024). ‘La alta montaña durante la Prehistoria: 10 años de investigación en el Pirineo catalán occidental', Trab. Prehis. 71, 261–281. doi: 10.3989/tp.2014.12134

  • CrossRef
  • Google Scholar

• 38

  GassiotE.Rodríguez-AntónD.BurjachsF.AntolínF.BallesterosA. (2012). Poblamiento, explotación y entorno natural de los estadios alpinos y subalpinos del Pirineo central durante la primera mitad del Holoceno. Cuatern. Geomorfol. 26, 29–45.

  • Google Scholar

• 39

  GelbertA. (2003). Traditions céramiques et emprunts techniques dans la vallée du fleuve Sénégal. Paris: Éditions de la Maison des sciences de l'Homme.

  • Google Scholar

• 40

  GrantA. (1982). “The use of tooth wear as a guide to the age of domestic ungulates,” in Ageing and Sexing Animal Bones from Archaeological Sites, eds. B. Wilson, C. Grigson, and S. Payne (Oxford: BAR British Series), 109, 91–108.

  • Google Scholar

• 41

  GuilaineJ.MartzluffM. (1995). La Bauma de la Margineda (Andorre). Toulouse: CNRS.

  • Google Scholar

• 42

  HernandoR.Moreno-IbáñezM. A.CarbonellE.DauraJ.Díez-CansecoC.EdoM.et al. (2024). Eating through time: understanding dietary practices across late prehistory in the northeastern Iberian Peninsula. Am. J. Biol. Anthropol. 184:e24950. doi: 10.1002/ajpa.24950

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 43

  Livingstone-SmithA. (2007). Chaîne opératoire de la poterie: références ethnographiques, analyses et reconstitution. Tervuren: Musée royal de l'Afrique centrale.

  • Google Scholar

• 44

  LongacreW. A.SkiboJ. M. (eds.) (1994). Kalinga Ethnoarchaeology: Expanding Archaeological Method and Theory. Washington, DC: Smithsonian Institution Press.

  • Google Scholar

• 45

  MartzluffM. (2024). Les plus anciens peuplements préhistoriques au cœur des Pyrénées. Élém. pour une Hist. Rech. Archéo.66, 7–28.

  • Google Scholar

• 46

  MedinaE. (2016a). ‘La cista del Coll de fontalba (Queralbs, Ripollès)', in XIII Jornades d'arqueologia de les comarques de Girona. Banyoles, pp. 85–88. (Girona: Departament de Cultura, Generalitat de Catalunya).

  • Google Scholar

• 47

  MedinaE. (2016b). Memòria de la intervenció arqueològica a la cista del Coll de Fontalba (Queralbs, Ripollès). Servei d'Arqueologia i Paleontologia, Direcció General del Patrimoni Cultural, Generalitat de Catalunya. Unpublished report.

  • Google Scholar

• 48

  Moreno IbáñezM. A.Del ValleH.Ramírez PedrazaI.SolerR.CarbonellE.TorneroC. (2024a). “L'enterrament col·lectiu del Roc de les Orenetes: últims avenços i estat de preservació del conjunt ossi,” in Dissetenes Jornades d'Arqueologia de les Comarques de Girona. Girona: Museu d'Arqueologia de Catalunya.

  • Google Scholar

• 49

  Moreno IbáñezM. A.SaladiéP.Ramírez-PedrazaI.Díez-CansecoC.CarbonellE.TorneroC. (2024b). Death in the high mountains: evidence of interpersonal violence during the late chalcolithic and early bronze age at roc de les orenetes (Eastern Pyrenees, Spain). Am. J. Biol. Anthropol. 184:e24909. doi: 10.1002/ajpa.24909

  • CrossRef
  • Google Scholar

• 50

  ObeaL. (2024). La recol·lecció i ús de la llenya al Pirineu Central des del Mesolític fins l'Alta Edat Mitjana (9è mil·lenni cal ANE – s. XIII). Aportacions des de l'antracologia. (Unpublished PhD). Universitat Autònoma de Barcelona. Available online at: http://hdl.handle.net/10803/693119 (Accessed February 15, 2026).

  • Google Scholar

• 51

  ObeaL.Celma MartinezM.Piqué HuertaR.Gassiot BallbèE.Martin SeijoM.Salvador BaigesG.et al. (2021). Firewood-gathering strategies in high mountain areas of the Parc Nacional d'Aigüestortes i Estany de Sant Maurici (Central Pyrenees) during Prehistory. Quat. Int. 593, 129–143. doi: 10.1016/j.quaint.2020.11.044

  • CrossRef
  • Google Scholar

• 52

  OlivaM. (2004). Els ornaments personals de la primera meitat del segon mil·lenni ANE del jaciment de Can Roqueta-II (est), Sabadell. Cypsela15, 229–249.

  • Google Scholar

• 53

  OrengoH.PaletJ. M.EjarqueA.MirasY. (2014). Shifting occupation dynamics in the Pyrenees: spatial and temporal patterns of high mountain land use. Quat. Int. 353. 140–152. doi: 10.1016/j.quaint.2014.01.035

  • CrossRef
  • Google Scholar

• 54

  PaletJ. M.ColominasL.PolonioT. (2018). Informe de les intervencions arqueològiques a Vall de Núria (Queralbs, Ripollès) - 2017. ICAC (Unpublished report).

  • Google Scholar

• 55

  PaletJ. M.EjarqueA.MirasY.RieraS.EubaI.OrengoH. (2007). Formes d'ocupació d'alta muntanya. Tribun. Arqueol.2006, 229–253.

  • Google Scholar

• 56

  PaletJ. M.GarciaA.ColominasL.PolonioT. (2019). “Memòria de les intervencions arqueològiques a Coma de Vaca i Coma del Fresser i a Vall de Núria (Queralbs, Ripollès),” in ICAC/Servei d'Arqueologia i Paleontologia, Generalitat de Catalunya. Unpublished report.

  • Google Scholar

• 57

  PaletJ. M.GarciaA.OrengoH. A.PolonioT. (2016). ‘Ocupacions ramaderes altimontanes a les capçaleres del Ter (Vall de Núria i Coma de Vaca, Queralbs): resultats de les intervencions arqueològiques 2010–2015′, in XIII Jornades d'Arqueologia de les Comarques de Girona. Banyoles: Universitat de Girona, pp. 67–75.

  • Google Scholar

• 58

  PaletJ. M.GarciaA.OrengoH. A.PolonioT. (2017). Els espais altimontans pirenaics orientals a l'Antiguitat: 10 anys d'estudis en arqueologia del paisatge del GIAP-ICAC. Treb. Arquel. 21, pp. 77–97. doi: 10.5565/rev/tda.59

  • CrossRef
  • Google Scholar

• 59

  PaletJ. M.JuliàR.RieraS.EjarqueA.OrengoH.MirasY.et al. (2012). ‘Landscape systems and human land-use interactions…', E-topoi, 3. pp. 305–310.

  • Google Scholar

• 60

  PaletJ. M.OrengoH. (2013). Memòria final de les prospeccions arqueològiques a Vall de Núria- Coma de Vaca i Coma de Vaca-Coma del Fresser (Queralbs, Ripollès): Bienni 2010-2011. ICAC (Unpublished report).

  • Google Scholar

• 61

  PaletJ. M.OrengoH.GarciaA. (2014). Memòria final de les intervencions arqueològiques (excavació i sondejos) a Coma de Vaca - Coma del Freser (Queralbs, Ripollès): Bienni 2012-2013. ICAC (Unpublished report).

  • Google Scholar

• 62

  PaletM.OlmosP.GarciaA.PolonioT.OrengoH. (2019). “Occupation et anthropisation des espaces de haute montagne dans les vallées de Nuria et de Coma de Vaca (Gerona, Espagne): résultats des recherches archéologiques et patrimoniales,” in La Conquête de la Montagne: des Premières Occupations Humaines à L'anthropisation du Milieu, eds. M. Deschamps, S. Costamagno, P. Milcent, M. Pétillon (Aubervilliers: Éditions du Comité des travaux historiques et scientifiques), 2019. doi: 10.4000/books.cths.7007

  • CrossRef
  • Google Scholar

• 63

  PayneS. (1985). Morphological distinctions between the mandibular teeth of young sheep, Ovis, and goats, Capra. J. Archaeol. Sci. 12, 139–147. doi: 10.1016/0305-4403(85)90058-5

  • CrossRef
  • Google Scholar

• 64

  PesciniV.FossatiM. (2021). Resultati delle analisi antracologiche della valle di Coma de Vaca, Coma del Freser e Val di Núria. Campagne 2016 e 2017. Relazione preliminare (Unpublished report).

  • Google Scholar

• 65

  Rafart-VidalE.Rubio-CampilloX.Mangado LlachX.González-OlivaresC.Sánchez de la TorreM. (2025). Spatial analysis of long-term mobility dynamics among hunter-gatherer groups in mountain areas: the case of Montlleó (Eastern Pyrenees). Archaeol. Anthropol. Sci. 17:59. doi: 10.1007/s12520-025-02351-y

  • CrossRef
  • Google Scholar

• 66

  Ramírez-PedrazaI.Díez-CansecoC.Moreno-IbáñezM. Á.SolerR.CarbonellE.TorneroC. (2022). “Intervencions al Roc de les Orenetes (Queralbs, Ripollès): campanyes 2020–2021,” in Setzenes Jornades d'Arqueologia de les Comarques de Girona, eds. M. Burch, J. M. Nolla, and J. Tremoleda (Girona: Generalitat de Catalunya), 74–77.

  • Google Scholar

• 67

  ReimerP. J.AustinW. E. N.BardE.BaylissA.BlackwellP. G.Bronk RamseyC.et al. (2020). The IntCal20 Northern Hemisphere radiocarbon age calibration curve (0–55 cal kBP). Radiocarbon62, 725–757. doi: 10.1017/RDC.2020.41

  • CrossRef
  • Google Scholar

• 68

  RenduC. (2003). La Montagne d'Enveig, une Estive Pyrénéenne Dans la Longue Durée. Canet: Trabucaire.

  • Google Scholar

• 69

  Ripoll PerellóE. (1960–1961). Notas de Arqueología de Cataluña y Baleares. Ampurias22, 353–354.

  • Google Scholar

• 70

  RischR. (1995). Recursos naturales, medios de producción y explotación social: un análisis económico de la industria lítica de Fuente Álamo (Almería), 2250-1400 a.C. Madrider Beiträge 21. Mainz: Philipp von Zabern.

  • Google Scholar

• 71

  RodríguezX. P. (2004). Technical Systems of Lithic Production in the Lower and Middle Pleistocene of the Iberian Peninsula. Oxford: BAR Publishing. doi: 10.30861/9781841713922

  • CrossRef
  • Google Scholar

• 72

  RouxV. (2019). Ceramics and Society: A Technological Approach to Archaeological Assemblages. Cham: Springer Nature. doi: 10.1007/978-3-030-03973-8

  • CrossRef
  • Google Scholar

• 73

  RoviraM.Díez-CansecoC.SolerR.TorneroC. (2024). “La cova 338, treballs d'excavació de l'any 2023 (Queralbs, Ripollès),” in Dissetenes Jornades d'Arqueologia de les Comarques de Girona, eds. J. Burch, F. Codina, R. Buxó, J. Frigola, M. Fuertes, M. Mataró, and E. Cortal (Girona: Museu d'Arqueologia de Catalunya, Universitat de Girona, Generalitat de Catalunya i Documenta Universitaria), 43-46.

  • Google Scholar

• 74

  Sánchez de la TorreM.Rafart-VidalE.González-OlivaresC.GratrezB.Mangado LlachX. (2024). Long-distance movements during the Last Glacial Maximum in the Pyrenean Mountain Range. J. Archaeol. Sci. Rep. Int. Dev. Res. Cent. Can.61:4952881. doi: 10.2139/ssrn.4952881

  • CrossRef
  • Google Scholar

• 75

  SchweingruberF. H. (1990). Microscopic Wood Anatomy: Structural Variability of Stems and Twigs in Recent and Subfossil Woods from Central Europe. Birmensdorf: Swiss Federal Institute for Forest; Snow and Landscape Research (WSL).

  • Google Scholar

• 76

  SierraA.BalasseM.RivalsF.FiorilloD.UtrillaP.SañaM. (2021). Sheep husbandry in the early Neolithic of the Pyrenees: New data on feeding and reproduction in the cave of Chaves. J. Archaeol. Sci. Rep.37:102935. doi: 10.1016/j.jasrep.2021.102935

  • CrossRef
  • Google Scholar

• 77

  SolerR.Moreno-IbáñezM. A.TrujilloJ. A.CarbonellE.TorneroC. (2024). ‘Intervenció a la Cova del Cingle del Cullell (Queralbs)', in Dissetenes Jornades d'Arqueologia de les Comarques de Girona. Girona: Documenta Universitaria.

  • Google Scholar

• 78

  SolerR. M.CalvoS.Díez-CansecoC.MorenoM.TorneroC. (2025). Informe-memòria de les intervencions arqueològiques a la Cova del Cingle del Cullell (Queralbs, Girona). IPHES-CERCA (Unpublished report).

  • Google Scholar

• 79

  StarkM. T.BishopR. L.MiskaE. (2000). Ceramic technology and social boundaries: cultural practices in Kalinga clay selection and use. J. Archaeol. Method Theor. 7, 295–332. doi: 10.1023/A:1026518922642

  • CrossRef
  • Google Scholar

• 80

  Tejedor-RodríguezC.Moreno-GarcíaM.TorneroC.HoffmannA.García-Martínez de LagránÍ.Arcusa-MagallónH.et al. (2021). ‘Investigating Neolithic caprine husbandry in the Central Pyrenees: insights from a multi-proxy study at Els Trocs cave (Bisaurri, Spain)', PLoS ONE 16:e0244139. doi: 10.1371/journal.pone.0244139

  • CrossRef
  • Google Scholar

• 81

  ToledoA.PonsE. (1982). Estat de la qüestió de l'Edat del Bronze a les comarques de la Garrotxa i del Ripollès. Ausa10, 165–186.

  • Google Scholar

• 82

  TorneroC.CarbonellE. (2018). Arrels prehistòriques de la transhumància a l'Alt Ripollès. Projecte arqueològic 2018-2021. Departament de Cultura, IPHES-CERCA.

  • Google Scholar

• 83

  TorneroC.CarbonellE. (2022). Arrels prehistòriques de la transhumància a l'Alt Ripollès. Projecte arqueològic 2022-2025. Departament de Cultura/IPHES-CERCA.

  • Google Scholar

• 84

  TorneroC.Díez-CansecoC. (2026). Arrels prehistòriques de la transhumància a l'Alt Ripollès. Projecte arqueològic 2026-2029. Departament de Cultura/IPHES-CERCA.

  • Google Scholar

• 85

  TorneroC.Díez-CansecoC.Ramírez PedrazaI.Del ValleH.Picornell GelabertL.SolerR.et al. (2022a). ‘Sota Palou (Campdevànol, Ripollès): Intervencions a Nous Sectors del Jaciment', in Setzenes Jornades d'Arqueologia de les Comarques de Girona. Girona: Documenta Universitaria, 36–39.

  • Google Scholar

• 86

  TorneroC.Díez-CansecoC.Ramírez-PedrazaI.RoviraM.CarbonellE. (2022b). Nous treballs d'excavació a la Cova 338 (Forat de l'Embut, Queralbs, Ripollès)', in Setzenes Jornades d'Arqueologia de les Comarques de Girona. Girona: Documenta Universitaria, 55–58.

  • Google Scholar

• 87

  WhiteT. D.BlackM. T.FolkensP. A. (2012). Human Osteology, 3rd Edn. Amsterdam: Academic Press. doi: 10.1016/B978-0-12-374134-9.50010-6

  • CrossRef
  • Google Scholar