Introduction

The study of lithic raw materials constitutes a key tool for addressing fundamental aspects related to the social, economic, and technological organization of prehistoric communities1,2,3,4,5. The management of these resources is directly linked to the control of natural resources and the survival of human groups, including procurement, distribution, and consumption strategies6,7,8,9,10. In this regard, the characterization of raw materials allows for a deeper understanding of key issues, such as mobility and procurement strategies11,12,13, the duration and typology of human occupations12,14, and the social and commercial networks among prehistoric groups15,16,17. The precise identification of the geological sources of materials recovered from archaeological sites helps into the reconstruction of mobility routes and the interpretation of relationships between human groups and the territories they inhabited18. Likewise, the physical, petrographic, and geological properties of rocks make it possible to link apparently isolated events within a single site or among different locations across a broader regional landscape19,20,21,22.

Considering different raw materials used during Prehistory, non-detrital sedimentary siliceous rocks stand out due to their cryptocrystalline structure and high hardness, which make them especially suitable for the manufacture of lithic tools23,24,25,26,27. Although their outcrops are often discreet, they are widely distributed in both continental and marine sedimentary contexts. The specific geological context generates mineralogical and structural variations that allow for the establishment of differentiation criteria of the sources28. Some varieties of these rocks, due to their unique composition and limited distribution, can function as lithological markers or mobility tracers, enabling the identification of routes, intergroup contacts, and exchange networks3,29,30. In this sense, the concept of the “evolutionary chain” (chaîne évolutive) developed by Fernandes and Raynal31 which provides a comprehensive framework to address the geological, post-formational, and anthropogenic processes affecting flint, integrating natural and cultural factors throughout its life history.

In recent years, archaeometric studies on flint characterization have advanced significantly through the application of techniques such as cathodoluminescence (CL), polarized optical microscopy (POM), X-ray fluorescence (XRF), Raman spectroscopy, and laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS). On this regard, Legg et al.32 applied CL and petrography on flint and orthoquartzites from the Paleoindian site of Silver Mound in the North American Midwest to visually differentiate highly similar siliceous rocks used in lithic knapping, identifying microtextures and cements that provide information about their geological origin. In the lacustrine Neolithic of Zürich, Lombardo et al.33 characterized flint fragments with evidence of fire-use traces through a combined approach involving µ-XRF, µ-XRD, and Raman spectroscopy, revealing microscopic pyrite residues. In the lower Danube basin, Gurova et al.34 studied Mesolithic and Neolithic flint tools using petrographic analysis and LA-ICP-MS, comparing them with outcrops in northern Bulgaria, successfully establishing direct correlations between archaeological sites and specific geological sources. Sobkowiak-Tabaka et al.35 investigated glacial (“erratic”) flint from prehistoric settlements in northern Poland using SEM, EDS, and XRF to differentiate it from other local flint types based on Ca and Fe content. Similarly, Moreau et al.36 characterized Upper Paleolithic flint from the Gravettian site of Maisières-Canal and other sites in the Haine valley (Mons basin) using geochemical signatures obtained through LA-ICP-MS, revealing diversified source exploitation and lithic circulation networks between the Sambre and Rhine basins, or Hughes et al.37 used portable XRF (EDXRF) to analyze archaeological flint from Scandinavian Neolithic sites, distinguishing materials from Danish Cretaceous deposits from Swedish glacial flint. In the same line, Imbeaux et al.38 combined LA-ICP-MS/MS, compositional data analysis, and supervised machine learning to successfully discriminate flint and chert sources in Neolithic contexts, demonstrating the power of integrating advanced geochemical and statistical methods for provenance studies. Likewise, a series of geoarchaeometric investigations led by Sánchez de la Torre and colleagues in the Pyrenees and northeastern Iberia39,40,41,42 applied multi-analytical protocols including ED-XRF and LA-ICP-MS to trace long-distance mobility and reconstruct prehistoric raw material networks.

In the Iberian Peninsula, the systematic application of geoarchaeological and archaeometric methodologies has made it possible to document, in some detail, the dynamics of occupation and mobility patterns during the Paleolithic, particularly in the northern part of the peninsula3,10,18,43,44,45,46,47. However, in other areas, such as the west and especially the south of the peninsula, knowledge about the availability, management, and characterization of lithic raw materials remain fragmented. Despite relevant research in these territories17,29,48,49,50,51,52,53,54, the lack of systematic and comprehensive studies that consider the geological complexity of these zones continues to limit our understanding of the technological and territorial dynamics developed by prehistoric communities.

In this context, the present study aims to contribute with new geoarchaeological and archaeometric data to reduce the current gap in knowledge concerning flint characterization in southern Iberia. The study begins with a preliminary macroscopic and techno-typological analysis of lithic artifacts excavated at the sites of Cueva de Ardales and Sima de las Palomas (Málaga, Spain), followed by systematic sampling of the immediate geological surroundings to locate and characterize potential procurement sources, considering the terminology used for siliceous raw materials in archaeological contexts remains inconsistent, with ongoing debate regarding the distinction between “flint”, “chert”, and “radiolarite”55,56 and in this study, “flint” refers to cryptocrystalline silica typically associated with chalk or marl, whereas “radiolarite” is used for siliceous sedimentary rocks dominated by radiolarian microfossils55. The term “jasper”, often misapplied to radiolarite, is avoided here. The research is focused and structured around petrographic study, supported by cathodoluminescence analysis. Additionally, other analytical techniques such as X-ray diffraction (XRD), X-ray fluorescence (XRF), and scanning electron microscopy (SEM) have been employed. This multi-analytical study has made it possible to establish precise diagnostic criteria for source identification and has advanced the interpretation of provisioning patterns, mobility strategies, and the management of siliceous raw materials across the territory.

Archaeological context

The sites of Cueva de Ardales and Sima de las Palomas located in the province of Málaga (Spain) (Fig. 1a) represent two fundamental locations for the study of the Paleolithic Age in southern Iberia. Both sites have revealed occupation sequences ranging from the Middle Paleolithic to the Holocene, allowing for the analysis of the continuity of human populations and their mobility strategies in this territory57. Their historical importance lies in the combination of key elements such as Neanderthal occupation, the transition to the Upper Paleolithic, and the presence of rock art, which makes them essential references for the study of symbolic58,59 and social60 behavior in prehistoric societies.

Both sites are in a region of great geographical and ecological complexity, in a mountainous environment with strategic connections between the coast and the interior. Despite the steep topography, natural corridors have existed and facilitated the mobility of human groups, enabling the formation of exchange and communication networks over time61. Natural passes between Puerto de las Atalayas (Arroyo del Granado, Ardales) and Puerto de Málaga (Arroyo de las Cañas, Carratraca) favored north–south movement, while the Almargen–Guadalete valleys made it possible to connect the bays of Cádiz and Málaga, creating a culturally significant area of interaction in the prehistory of the region.

Fig. 1
figure 1

(a) Map of Spain showing the location of Málaga province. (b) Location of the archaeological sites: Ardales Cave in Ardales and Sima de las Palomas in Teba.

Full size image

Cueva de Ardales is located on the Cerro de La Calinoria, at 565 m a.s.l., approximately 50 km from the Mediterranean coast, and is connected to the basins of the Guadalhorce, Guadalteba, and Turón rivers (Fig. 1b). It was discovered in 1821 after an earthquake exposed its entrance, which had previously been sealed by colluvial deposits62, and it has been the subject of numerous investigations that have revealed continuous human occupation from the Middle Paleolithic to the Neolithic57 (Fig. 2a). Its significance lies in the presence of more than 1,000 engravings and cave paintings, some attributed to Neanderthals, reinforcing its role in the study of early symbolic behavior58,59. The cave is of karstic origin (dolomitic limestones) and features multiple chambers and galleries. The Sala de las Estrellas stands out for its stenciled hands and red abstract signs, while the Galería del Calvario contains the highest concentration of engravings and cave paintings at the site (Fig. 2b). In the Galerías Altas, Paleolithic rock art has been documented alongside a Neolithic and Chalcolithic necropolis60. Excavations conducted within the framework of a General Research Project (see Acknowledgements) have documented several phases of occupation. In Zone 3, Mode 3 knapped lithic tools associated with Neanderthal groups were found (51–53 ka BP). In Zone 5, levels from the Early Upper Paleolithic (33–45 ka BP) were identified. Zone 2 revealed Gravettian occupations (25,310–24,470 cal BP) and Solutrean levels (18,885 cal BP)57,60, contributing to a long history of occupation by different societies throughout much of the Paleolithic.

Sima de las Palomas, within the municipality of Teba, is located in the Tajo del Molino, in the basin of the Guadalteba River, a tributary of the Guadalhorce (Fig. 1b). Its setting represents a transitional zone between the Antequera Depression and the Ronda Plateau, with connections both toward the coast and inland63,64. More specifically, it is situated in the Las Palomas karst complex, between 470 and 430 m a.s.l. in the Peñarrubia mountains, within the Guadalteba region. The site forms part of a larger karstic system and is located 100 m below the summit of the limestone massif of the Sierra de Teba-Peñarrubia. The surrounding environment combines mountainous terrain and the alluvial plains of the Turón, Guadalteba, and Guadalhorce rivers, providing essential resources for prehistoric populations, including water, fauna, and lithic raw materials. The site includes a small rock shelter and a large cave over 100 m long and 20 m wide, known as Cueva de las Palomas. At its northern end, the cave has been cut by a large doline surrounded by a limestone rim that contains the sima, a shaft that, although now isolated from the underlying cave system, was possibly connected in the past, and both sima and cave are open toward the Tajo del Molino (Fig. 2c). Excavations have revealed continuous occupation since the Middle Paleolithic (Fig. 2d), with evidence dating back to more than 55 ka BP65. In its deepest layers (10/11PI), Levallois-technique knapped lithic products66,67, faunal remains, and combustion structures have been documented, indicating recurrent Neanderthal occupation until around 45–40 ka BP. Excavations reveal that, after a hiatus, occupation resumed during the Gravettian (28 ka BP, layer 6), continuing into the Solutrean (20–22 ka BP). However, a second hiatus of approximately 10,000 years is observed, with no evidence of occupation during the Magdalenian or the Final Upper Paleolithic, which may be due to erosion or abandonment of the site. During the Holocene, the cave was reused in the Middle Neolithic and the Bronze Age, with evidence of excavated structures and burials. Finally, during the Middle Ages (1223–1280 AD), a final occupation is documented, including ceramic fragments and a grinding stone. A total of 44 radiocarbon dates have been obtained58,60,64,65, confirming a long and complex history of human occupation, with significant changes in the site’s function and use over time.

Fig. 2
figure 2

General context and current views of the two archaeological sites. (A) Current entrance to Cueva de Ardales, now monumentalized for visitor access. (B) Interior view of the cave, likely similar in appearance to how it was during the Middle Paleolithic. (C) Landscape view of the surroundings of Sima de las Palomas (Tajo del Molino), showing indicating the karstic cave overlooking the río de la Venta (yellow arrow) and the main entrance to Sima (orange arrow). (D) Overhead photograph taken inside the Sima, showing members of the research team during excavation campaigns.

Full size image

Geological background

From a geological perspective, the study area is in a transitional region between three main structural domains of the Betic Cordillera: the Internal Zones, the External Zones, and the Campo de Gibraltar Complex, each with distinct lithological and structural characteristics68,69 (Fig. 3a).

The Campo de Gibraltar Complex, located in the western sector, is composed of Meso-Cenozoic sedimentary materials from the Betic Flysch, characterized by turbiditic facies with alternating sandstones and clays deposited in a deep marine environment70,71. This complex outcrops around the Guadalteba reservoirs and marks the western boundary of the study area. The Internal Zones, present locally in the southeastern sector of the study area, represent the most intensely deformed core of the orogen68,72. Metamorphic materials corresponding to the Nevado-Filábride, Alpujárride, and Maláguide complexes outcrop in this zone73,74. Within the context of this study, the Maláguide Complex is of particular interest; it crops out between Ardales and El Burgo and includes Paleozoic formations composed of schists and limestones containing flint75,76,77. However, the most relevant flint-bearing materials for this study are placed in the External Zones, where formations of the Middle Subbetic and Penibetic dominate. These are widely represented in the region of the Guadalteba Valley and surrounding sierras. These units are composed of Mesozoic limestones, marls, and dolomites, within which levels containing sedimentary flint nodules are interbedded69,72. Finally, the Quaternary fluvial terraces and alluvial deposits of the Guadalteba and Turón valleys have acted as secondary accumulators of siliceous materials (flint, radiolarites, sandstones, etc.), constituting key zones for raw material procurement during Prehistory68,78 (Fig. 3b).

Fig. 3
figure 3

(a) General geological map of the Betic Cordillera; (b) Geological map of the study area showing the main lithostratigraphic units, including the External Zones, Internal Zones, Campo de Gibraltar Complex, and post-orogenic materials. Superimposed on these units are fluvial terraces (Ríos de la Venta, Guadalteba, and Turón), located near the archaeological sites and characterized by a high concentration of flint pebbles.

Full size image

Materials and methods

Lithological identification of archaeological pieces

The study began with the analysis of lithic material from both archaeological sites, Cueva de Ardales and Sima de las Palomas. In an initial phase, a preliminary classification was conducted through macroscopic observation of all excavated pieces, allowing the identification and grouping of different varieties of flint and radiolarites, as well as other raw materials used to a lesser extent, such as compact sandstones and other lithologies. This analysis was based on characteristics visible to the naked eye and under a binocular magnifier, including color, texture, luster, type of fracture, and the presence of inclusions and/or fossil remains. The use of the binocular magnifier also facilitated the identification of structural details such as internal fabric, variations in the siliceous matrix, and sedimentary or biogenic structures that are not perceptible without optical magnification.

In parallel, an archaeological classification of the knapped lithic products was carried out using a techno-typological approach, following the analytical system of Laplace79,80, aimed at identifying the structural and functional organization81 of the artifacts according to different knapping schemes. The pieces were categorized into four main groups: Positive Bases (PB), Other Knapping Debris (OKD), First-Generation Negative Bases (1GNB), and Second-Generation Negative Bases (2GNB). Lithic artifacts were classified based on technological and morphological criteria. Positive Bases (PB) are flakes or fragments with a preserved positive platform, typically representing primary removals. Other Knapping Debris (OKD) includes fragments that cannot be directly attributed to specific flake types, often resulting from secondary knapping processes. First Generation Negative Bases (1GNB) are flakes with a negative platform from initial detachments during primary reduction. Second Generation Negative Bases (2GNB) correspond to flakes with a negative platform produced by subsequent removals, reflecting later stages of core reduction. During this process, each artifact was individually recorded, indicating its stratigraphic layer and corresponding chrono-stratigraphic unit55,58,65. A total of 1,373 lithic pieces were analyzed, of which 607 came from Cueva de Ardales and 766 from Sima de las Palomas.

Fieldwork: identification of source areas

The study area had previously been the subject of geoarchaeological investigations82,83,84,85,86,87. Following the initial characterization of the archaeological lithic material, fieldwork was carried out with the aim of identifying potential geological procurement sources. The survey was based on systematic sampling in the vicinity of the archaeological sites, focusing on the location and documentation of secondary outcrops of flint and other siliceous materials (Fig. 3b). Primary outcrops were excluded due to their low concentration of nodules, high hardness of the limestone matrix, and difficult accessibility. Therefore, the study focused on secondary deposits, such as alluvial and colluvial origin, where transport and sedimentation processes favor the accumulation of nodules with physical characteristics suitable for their exploitation.

The fieldwork campaigns were conducted primarily in the fluvial terraces of the De la Venta, Guadalteba, and Turón rivers (Fig. 3b), previously identified for their potential as raw material procurement areas. At each sampling point, geographic coordinates were recorded using GPS, and sedimentological and geological characteristics were documented with particular attention to the depositional context and post-depositional processes. Approximately 1,000 geological samples were collected—mainly flint and radiolarites—which were selected in situ by macroscopic comparison with the previously analyzed archaeological materials. All collected material was stored in the institutional lithotheque of the University of Cádiz (LitUCA)88 for subsequent petrographic and geochemical analyses, to evaluate their compatibility with the archaeological assemblages and contextualizing procurement strategies according to regional geological and geomorphological conditions.

Laboratory work

For the petrographic study, 150 thin sections of geological samples were prepared, of which 51 were materials collected from the three fluvial terraces near Cueva de Ardales and Sima de las Palomas, while the rest corresponded to samples from more distant areas, assigned to the Penibetic and Middle Subbetic zones of the Sierra de Cádiz and the Serranía de Ronda. The objective was to broaden knowledge of the flint available within a wider catchment range, considering the palaeostratigraphic continuity of the exposed formations (Fig. 3a). Thin section analysis is the most widely used and comprehensive microscopic technique for flint characterization, providing detailed information on its mineralogical composition, structure, and formative processes. Mineralogical and textural components were identified, establishing their formation and diagenetic conditions. For this purpose, an Olympus BH2 petrographic microscope equipped with a microphotography system was used. Additionally, cathodoluminescence images were obtained using a Technosyn Cold Cathodoluminescence Model 8200 MkII system, integrated into an optical microscope and microphotography system. This technique was applied to 50 thin sections selected for their representativeness, including 40 geological flint samples and 10 archaeological samples, operating at 15–18 kV, with occasional increases to 20–25 kV, and a beam current of 250–300 µA, optimal conditions for siliceous materials89. The microscopic study was complemented by the analysis of 8 representative geological samples from fluvial terraces, selected based on textural variability and degree of alteration, using a scanning electron microscopy (SEM), using a NOVA NANOSEM 450 equipment, operated in secondary electron (SE) and backscattered electron (BSE) modes at an accelerating voltage of 15 to 20 kV. Samples were cleaned with distilled water and coated with a thin layer of gold (Au) using a sputter coater to improve electrical conductivity. Microstructure, fossil content, porosity, fracturing, and the distribution of mineral inclusions in the siliceous matrix were evaluated. Energy-dispersive X-ray spectroscopy (EDS) analysis was used to identify elemental composition and to generate elemental maps showing spatial distribution.

For geochemical characterization, an energy-dispersive X-ray fluorescence (EDXRF) spectrometer, model AXS M-4 Tornado (Bruker), was used and operated under vacuum to optimize the detection of light elements. The following parameters were applied: voltage of 25–50 kV, current of 100–500 µA, and an integration time of 60–120 s per sample. This analysis enabled the identification and quantification of major and minor elements in twelve flint samples, providing insights into their geochemical variability and potential links to different geological sources or post-depositional alteration processes. Finally, mineral identification was performed by X-ray diffraction (XRD) in the same 12 samples, selected for their chromatic, textural, and alteration diversity. A Bruker D8 Advance A25 X-ray diffractometer was used, equipped with Cu-Kα radiation (λ = 1.5406 Å), operating at 40 kV and 30 mA. Diffractograms were recorded over a 2θ range of 5° to 70°, with a step size of 0.02° and an integration time of 1 s per step. Mineral phase identification was performed using the DIFFRACplus EVA software.

All equipment used in this study is available at the laboratories of the Department of Earth Sciences, the UGEA-PHAM unit, and the Scientific and Technological Research Central Services (SC-ICYT) of the University of Cádiz, located in Puerto Real, Cádiz.

Results

Categorization of archaeological Flint pieces by hand sample

A total of 1,373 observations were conducted on archaeological materials recovered from both caves, using direct visual inspection of hand samples and binocular magnification. These observations were complemented by comparison with geological specimens collected from nearby fluvial terraces. Based on this initial macroscopic analysis, petrographic information about the raw materials used could be inferred, allowing the identification of eight different varieties of flint present in both archaeological assemblages.

The first variety corresponds to homogeneous massive flint, which has a uniform fine-grained texture, color tones ranging from grey to beige, and lacks visible internal structures, making it easily distinguishable from other types (Fig. 4A). The second variety is a massive flint with inclusions, similar in texture to the first type but containing visible inclusions of ferric minerals and fossil remains, giving it a more heterogeneous texture and coloration (Fig. 4B). Oolitic flint is characterized by the presence of small ooids, resulting in a less homogeneous texture and generally lighter coloration, which facilitates macroscopic identification (Fig. 4C). The so-called “Azulejo-type” flint is easily recognized by its elongated spots filled with chalcedony on a beige massive matrix with black dendritic oxide patterns (Fig. 4D). On the other hand, the “Turón-type” flint has fine grains, homogeneous texture, and color tones ranging from beige to black, and is distinguished by the presence of bioturbation traces, mainly attributable to the genera Phycodes or Phicosiphon90 (Fig. 4E). Another variety is white flint with black microfossils, characterized by irregular bands rich in dark microfossil inclusions (some classified as incertae sedis), which contrast sharply with the surrounding light matrix (Fig. 4F). Additionally, two types of radiolarites were identified: red radiolarite, recognizable by its intense reddish coloration and homogeneous texture (Fig. 4G), and multicolored radiolarites, mainly in cream tones, with homogeneous texture and subtle chromatic variations (Fig. 4H). Finally, aside from these main types of flint and radiolarite, compact sandstones from the geological units of the Campo de Gibraltar were identified in smaller proportions, especially in Cueva de Ardales, along with occasional small fragments of limestone and calcined bone remains.

Fig. 4
figure 4

Each sample shows the hand sample (left image) and its detailed view of the same samples under a binocular magnifier (right image) from the archaeological sites representing each lithological variety (scale bar = 1 cm). Typologies: (a) Massive flint; (b) Massive flint with inclusions; (c) Oolitic flint; (d) Azulejo-type flint; (e) Turón-type flint; (f) White flint with black microfossils; (g) Red radiolarite; (h) Multicolored radiolarite.

Full size image

Relationship between archeological varieties and typologies

Once the flint varieties from the total lithic assemblage of both sites were defined based on macroscopic criteria, they were correlated with their techno-typological classification. For this purpose, each piece was associated with its corresponding chrono-stratigraphic unit, as determined by previously published absolute dates55. On the other hand, the OKD (that is, lithic fragments or by-products resulting from core reduction and flake production, consisting of chips and small waste pieces) were not included as formal artifacts, but were used as a source of information to establish this relationship.

Cueva de Ardales

A total of 607 lithic pieces from the archaeological sequence of Cueva de Ardales were analyzed (Fig. 5). The assemblage spans various cultural periods57, including the Neolithic, a disturbed level (SP) likely containing Solutrean materials alongside typically Neolithic industries, and levels corresponding to the Solutrean, Gravettian, and Middle Paleolithic. The most represented raw materials across the assemblage are red radiolarite (178 pieces), massive flint (133), and flint with inclusions (123, including porous flint), while white flint with microfossils (21), multicolored radiolarite (19), oolitic flint (12), “Azulejo”-type flint (10), compact sandstone (9), and other non-siliceous rocks (1) have a secondary presence.

Regarding the identified chronological periods, a total of 221 pieces were recorded for the Neolithic, with the predominant raw materials being red radiolarite (79), massive flint (75), and flint with inclusions (42). Other lithologies present in smaller proportions include multicolored radiolarite (8), “Azulejo”-type flint (7), white flint with microfossils (7), Turón-type flint (19), and oolitic flint (2). After excluding the OKD group, the assemblage is reduced to 70 pieces, primarily composed of flint with inclusions (24), followed by massive flint (9), red radiolarite (7), multicolored radiolarite (4), white flint with microfossils and “Azulejo”-type flint (3 pieces each), Turón-type flint (2), and oolitic flint (1).

In the disturbed level (SP), 227 pieces were recovered, characterized by a mixture of Solutrean and Neolithic elements. The dominant raw materials in this assemblage are massive flint (66), red radiolarite (57), and flint with inclusions (49). Less frequent varieties include white flint with microfossils (13), “Azulejo”-type and Turón-type flint (10 pieces each), porous flint and oolitic flint (8 pieces each), multicolored radiolarite (6), compact sandstone, and one non-siliceous rock fragment. After excluding OKD group, 124 pieces remain, of which 113 are classified as PB and 11 as 1GNB. The lithologies in this subset are represented by flint with inclusions (29), red radiolarite (18), massive flint (13), white flint with microfossils (10), multicolored radiolarite (3), “Azulejo”-type and oolitic flint (4 pieces each), along with one piece each of Turón-type flint, compact sandstone, and non-siliceous rock.

The Solutrean assemblage, composed of 38 pieces, is dominated by flint with inclusions (8), red radiolarite (6), and multicolored radiolarite (5), along with massive flint (4). In smaller numbers are oolitic and “Azulejo”-type flint (2 pieces each), and one piece each of white flint with microfossils and compact sandstone. After excluding OKD group, the total is reduced to 14 pieces, again highlighting flint with inclusions (6), multicolored radiolarite (4), massive flint (2), oolitic flint, and white flint with microfossils (1 piece each).

For the Gravettian, 11 pieces were identified, mainly distributed among red radiolarite, massive flint, and compact sandstone, with 3, 3, and 4 pieces respectively. One piece made from Turón-type flint was also documented. After excluding OKD group, only one PB piece remains, made of massive flint.

Finally, in the level corresponding to the Middle Paleolithic, a total of 110 pieces were recorded, with red radiolarite (40) and massive flint (21) being the predominant lithologies. Other raw materials identified include flint with inclusions (16), Turón-type flint (8), “Azulejo”-type flint (5), oolitic flint and compact sandstone (2 pieces each), and white flint with microfossils (1 piece). After excluding OKD group, the assemblage is considerably reduced to just 9 pieces, mainly composed of red radiolarite (5), flint with inclusions (2), massive flint, and “Azulejo”-type flint (1 piece each).

Fig. 5
figure 5

Comparative representation of the geological raw materials used, classified by techno-typological category, across the different chrono-stratigraphic occupation units of Cueva de Ardales. Diagrams are normalized to 100%. The S/P unit corresponds to a disturbed level (SP), where Solutrean materials are mixed with typically Neolithic industries.

Full size image

Sima de las Palomas

The lithic assemblage from Sima de las Palomas has been divided into two chronological periods: Holocene and Upper Paleolithic. A total of 766 pieces have been documented, of which 669 belong to the OKD group, while the remaining 97 pieces have been classified as 1GNB, 2GNB, or PB (Fig. 6).

During the Holocene, the assemblage consists of 124 pieces, clearly dominated by those belonging to the OKD group, with a total of 99 pieces. The most frequently used raw materials in this period are homogeneous massive flint (59 pieces), red radiolarite (33 pieces), and flint with inclusions (20 pieces). When excluding OKD group pieces, the Holocene assemblage is reduced to 25 pieces, divided into 1GNB (3 pieces) and PB (22 pieces). Within the 1GNB category, cores made of homogeneous massive flint, red radiolarite, and non-siliceous material were identified, with one piece each. In the PB category, homogeneous massive flint predominates (11 pieces), followed by flint with inclusions (5 pieces), along with grey radiolarite with inclusions (4 pieces), red radiolarite, and Turón-type flint (one piece each).

As for the Upper Paleolithic, the assemblage comprises a total of 642 pieces, with a strong predominance of the OKD group, totaling 570 pieces. The most represented raw materials in the complete assemblage are homogeneous massive flint (269 pieces), followed by red radiolarite (151 pieces), flint with inclusions (52 pieces), and multicolored radiolarites (47 pieces). After removing the OKD pieces, the Upper Paleolithic assemblage is reduced to 72 pieces, distributed among the categories 1GNB (4 pieces), 2GNB (9 pieces), and PB (59 pieces). In the 1GNB category, cores of homogeneous massive flint (2 pieces), flint with inclusions (1 piece), and green-black-white radiolarite (1 piece) were recorded. The 2GNB category is mainly represented by homogeneous massive flint (5 pieces), followed by flint with inclusions (2 pieces), red radiolarite, and green-black-white radiolarite (one piece each). Finally, in the PB category, homogeneous massive flint is the predominant raw material with 26 pieces, followed by flint with inclusions (8 pieces), oolitic flint and red radiolarite (7 pieces each), green-black-white radiolarite (6 pieces), “Azulejo”-type flint (4 pieces), and Turón-type flint (1 piece).

Fig. 6
figure 6

Comparative representation of the geological raw materials used, classified by typotechnological category, across the different chrono-stratigraphic occupation units of Sima de las Palomas. Diagrams are normalized to 100%.

Full size image

Characterization of geological samples

After the preliminary classification and geological observations of both sites and the samples, a complex multi-analytical study was conducted on the materials collected from fluvial terraces, to enable later comparison with some of the archaeological artifacts.

First, petrographic analysis allowed the identification of features not discernible in the initial macroscopic analysis of hand samples. All samples display a dominant matrix composed of microcrystalline quartz, accompanied by secondary minerals such as dolomite, calcite, and iron oxides. Variations were also observed in grain packing and grain-to-grain contacts, reflecting different formation and diagenetic processes. A notable presence of biogenic components and accessory minerals was also detected, allowing for a more precise sub-classification of the samples (Fig. 7) relative to the initial visual analysis (Fig. 4). The main petrographic characteristics identified are summarized in Table 1. More specifically, within the “Massive Flint” group, two sub-categories were clearly differentiated. The first, “Massive Flint with Radiolarians and Peloids,” displays a grain-supported packstone/grainstone structure89, dominated by well-sorted peloids and excellently preserved radiolarians, along with dispersed opaque minerals within the matrix (Fig. 7A.1). The second subcategory, “Light Grey Homogeneous Flint,” is characterized by a homogeneous matrix rich in dolomite crystals and iron oxides, lacking visible biogenic or allochemical components (Fig. 7A.2).

The “Massive Flint with Inclusions” category is represented by samples with a significant presence of poorly preserved fossil remains, primarily peloids and incertae sedis fragments, clearly affected by dissolution processes and extensive dolomitization (Fig. 7B).

Regarding the “Oolitic Flint,” two subcategories were distinguished. The “Oolitic Flint” subcategory has a clearly grain-supported packstone structure, with well-preserved ooids displaying concentric layering, as well as occasional inclusions of bivalves and foraminifera, showing variation in opacity and mineral distribution (Fig. 7C.1). The other, “Flint with Ooids and Peloids,” presents a packstone structure composed of peloids and ooids in a microcrystalline silica matrix, with well-preserved foraminifera and variability in density and translucency (Fig. 7C.2).

“Azulejo-Type Flint” is characterized by complete dolomitization or appears as a highly silicified limestone, rich in diverse fossil fauna including foraminifera, bivalves, and ostracods (Fig. 7D). “Turón Flint” presents an intermediate mudstone-type structure, with a high concentration of dolomite crystals and abundant organic matter (incertae sedis) (Fig. 7E). “White Flint with Black Inclusions” shows a matrix containing dispersed opaque microfossils, mainly peloids and incertae sedis fragments, associated with clear dissolution and dolomitization processes (Fig. 7F).

Within the “Red Radiolarite” group, three subcategories were defined: “Chocolate Radiolarite” with a grain-supported packstone structure dominated by very well-preserved radiolarians (Fig. 7G.1); “Red Radiolarites with Peloids” display an intermediate packstone–wackestone structure characterized by peloids, ooids, and well-preserved radiolarians with notable variability in density and the presence of opaque minerals (Fig. 7G.2); and “Red Flint with Spicules,” which stands out for its packstone structure with excellently preserved sponge spicules and a significant abundance of radiolarians (Fig. 7G.3).

Finally, the “Multicolored Radiolarites” are characterized by the “Cream–Greenish Radiolarites” subcategory showing a packstone structure with chromatic variations ranging from cream to greenish tones, and occasionally a subtle banding, with excellently preserved radiolarians, sponge spicules, and residual microfossils such as foraminifera, along with variability in the presence of opaque minerals within the matrix (Fig. 7H).

Fig. 7
figure 7

From right to left, for each set of images: hand sample (scale = 1 cm), POM image (scale = 1 mm), and MOP image (scale = 1 mm). Typologies: (A) Massive flint: (A.1) with radiolarians and peloids; (A.2) homogeneous grey; (B) Flint with inclusions; (C) Oolitic flint: (C.1) oolitic; (C.2) with ooids and peloids; (D) Azulejo-type flint; (E) Turón-type flint; (F) White flint with inclusions; (G) Red radiolarite: (G.1) chocolate-colored radiolarite; (G.2) with radiolarians and peloids; (G.3) with echinoid spicules; (H) Colored radiolarite. Rad: Radiolarian; Pel: Peloids; Biv: Bivalves; F: Foraminifera; Oo: Ooids; E. Sp.: Sea urchin spicule; I.S.: Incertae Sedis; Dol: Dolomite; MM: Metallic mineral; Qz: Quartz90).

Full size image
Table 1 Main petrographic characteristics used for the classification of subcategories defined through binocular magnifier and POM observations.
Full size table

Cathodoluminescence (CL) analyses conducted on flint samples revealed a set of mineralogical and textural features that complement the observations obtained through polarized optical microscopy (POM). The intensity and color of the emitted luminescence are related to the original mineralogical composition of the various components present in the flint (Fig. 8-A.1). CL allowed the identification of well-defined dolomite crystals (Fig. 8-A.2), as well as different degrees of preservation of biogenic components (Fig. 8-B). In the ooid layers, evidence of concentric growth was observed, providing information about the fossilization process (Fig. 8-C.1), as well as well-defined spheroidal structures within the ooids (Fig. 8-C.2). CL also revealed a significant number of foraminifera in the oolitic flint samples that were not visible under optical microscopy (Fig. 8-C.1). Additionally, previously undetected fossils were identified (Fig. 8-D), as well as microfractures filled with minerals of varying composition, mainly carbonates (Fig. 8-E). Specifically, dolomite crystals exhibit a strong orange luminescence. This same luminescence pattern, with hues ranging from orange to reddish, was recorded in biogenic elements that had a carbonate origin prior to the silicification process—such as peloids, bivalves, foraminifera, ooids, and incertae sedis—especially when operating at voltages between 15 and 20 kV (e.g., Fig. 8-F). In contrast, biogenic components with an originally siliceous structure (radiolarians and sponge spicules) do not exhibit luminescence under standard voltages (15–20 kV) and appear as opaque areas (Fig. 8-G.1). When the voltage is increased above 20 kV, these components acquire dark maroon tones, becoming visible (Fig. 8-G.1, Fig. 8-G.2, Fig. 8-G.3). Under the same conditions, microcrystalline quartz crystals display a bluish luminescence (Fig. 8-G.1). In the multicolored radiolarites (Fig. 8-H), in addition to well-defined dolomite crystals, sponge spicules and bivalve filaments were also observed, indicating greater biogenic diversity compared to the chocolate radiolarites, in which radiolarians predominate.

Fig. 8
figure 8

CL images with scale = 1 mm. Typologies: (A) Massive flint: (A.1) with radiolarians and peloids; (A.2) homogeneous grey; (B) Flint with inclusions; (C) Oolitic flint: (C.1) oolitic; (C.2) with ooids and peloids; (D) Azulejo-type flint; (E) Turón-type flint; (F) White flint with inclusions; (G) Red radiolarite: (G.1) chocolate-colored radiolarite; (G.2) with radiolarians and peloids; (G.3) with echinoid spicules; (H) Colored radiolarite.

Full size image

Scanning electron microscopy (SEM) analyses performed on freshly cut geological samples provided complementary information on their mineralogical and microstructural characteristics, supporting the observations obtained through POM and CL. In samples where dolomite crystals had previously been identified by CL, a high density of pores was observed (Fig. 9-A), along with signs of possible partial dissolution of both the crystals and associated fillings (Fig. 9-B), suggesting areas of localized mineral alteration. In samples where dolomite was not detected through POM and CL, SEM revealed pores sealed by quartz (Figs. 9-C.1 and 9-C.2), with no visible signs of carbonates. These observations correspond to samples that, under the aforementioned techniques, also displayed better preservation of biogenic components. Fossils were also identified on the surface of some sections, easily recognizable ooids and radiolarians (Fig. 9-D), whose morphological structures had already been clearly observed via POM. Additionally, areas of carbonate recrystallization were documented on the sample surface (Fig. 9-E). In these samples, prepared from fresh cuts, no cortex was observed in the hand sample. However, under SEM, recrystallized textures were identified, possibly associated with these processes in early stages.

Fig. 9
figure 9

Images obtained using scanning electron microscopy (SEM): (a) Pore likely resulting from a dissolution process; (b) Dolomite recrystallization; (c.1) Flint surface without dolomitization; (c.2) Magnified view of the surface showing a pore filled with quartz (Qtz); (d) Radiolarians; (e) Carbonate recrystallization.

Full size image

The obtained diffractograms (Fig. 10) show a homogeneous mineral composition, with a general predominance of quartz (Qz), indicated by intense and well-defined peaks. The samples exhibit variations in the relative intensity of the peaks and in the presence of secondary phases. Calcite (Cal) signals were detected in the samples corresponding to massive flint with radiolarians and peloids (sample 1), oolitic flint with oolitic texture (sample 4), Turón-type flint (sample 7), and red radiolarite with radiolarians and peloids (sample 10). Additionally, dolomite (Dol) was also identified in samples 1 and 7, though in lower proportions.

Fig. 10
figure 10

X-ray diffraction (XRD) patterns of the analyzed samples, numbered from bottom (1) to top (12): (1) massive flint with radiolarians and peloids, (2) homogeneous grey massive flint, (3) flint with inclusions, (4) oolitic flint with oolitic texture, (5) oolitic flint with ooids and peloids, (6) Azulejo-type flint, (7) Turón-type flint, (8) white flint with inclusions, (9) chocolate-colored red radiolarite, (10) red radiolarite with radiolarians and peloids, (11) red radiolarite with echinoid spicules, and (12) colored radiolarite. Major diffraction peaks corresponding to quartz (Qz), calcite (Cal), and dolomite (Dol) are indicated. The patterns are vertically offset to enhance visualization and avoid overlapping.

Full size image

Finally, the XRF analysis of the samples revealed a composition dominated by silica, with SiO2 contents ranging from 91.04% in oolitic flint with oolitic texture to 98.53% in oolitic flint with ooids and peloids. The highest CaO values were recorded in massive flint with radiolarians and peloids (7.07%) and in oolitic flint with oolitic texture (6.71%), while the lowest value corresponds to Turón-type flint (0.14%). Fe2O3 content ranges from 0.088% in red radiolarite with echinoid spicules to 2.52% in white flint with inclusions, the latter also exhibiting the highest K2O content (0.76%). MnO is present in all samples, with values ranging between 0.02% and 0.15%, the maximum found in chocolate-colored red radiolarite. NiO was detected in Azulejo-type flint (0.0094%) and in white flint with inclusions (0.0030%). Cobalt (Co) was identified in massive flint with radiolarians and peloids (0.0059%) and in flint with inclusions (0.0060%). Minor amounts of V2O₅, ZnO, TiO2, SrO, and Zr were also identified, with concentrations varying according to the sample (Table 2).

Table 2 Chemical composition of selected samples expressed as major oxides (in wt%), determined by X-ray fluorescence (XRF). Values have been normalized to 100%. Columns correspond to different types of Flint identified through POM and CL, while rows indicate the detected oxides. The analyzed samples are: (1) massive Flint with radiolarians and peloids, (2) homogeneous grey massive Flint, (3) Flint with inclusions, (4) oolitic Flint with oolitic texture, (5) oolitic Flint with ooids and peloids, (6) Azulejo-type Flint, (7) Turón-type Flint, (8) white Flint with inclusions, (9) chocolate-colored red radiolarite, (10) red radiolarite with radiolarians and peloids, (11) red radiolarite with echinoid spicules, and (12) colored radiolarite.
Full size table

Characterization of archaeological Flint

Based on a notable similarity between the geological and archaeological materials observed through macroscopic analysis and binocular magnification (Table 3), a specific analysis was conducted on selected archaeological pieces from Cueva de Ardales (Fig. 11) and Sima de las Palomas (Fig. 11) to determine whether the raw materials used in artifact production correspond precisely to those available in the immediate geological environment.

Table 3 On the left, lithic artifacts from both caves, belonging to an industry attributed to mode 3. On the right, samples collected from nearby fluvial terraces, showing notable macroscopic similarity.
Full size table

The comparison presented in Fig. 11 shows a high degree of correspondence between the archaeological artifacts and the reference geological samples. Thin section analysis revealed notable similarities in texture, biogenic content, internal fabric, and mineralogical composition. Likewise, matches were identified to the degree of silicification, as well as in the type and distribution of inclusions. Cathodoluminescence (CL) analysis revealed comparable patterns of zoning, luminescence intensity, and microstructural development of the silica. Overall, the observed features support a strong similarity between the archaeological materials and their corresponding geological sources.

Fig. 11
figure 11

In the upper section, the archaeological flint artifact under analysis is shown on the left, while the corresponding geological reference sample is displayed on the right. Both images are shown at a scale of 0.5 mm and correspond to POM-classified thin sections. In the lower section, the same order (left to right) is maintained for the same thin section observed under cathodoluminescence (CL). The comparisons presented are as follows: (a) SP artifacts compared with massive flint with radiolarians and peloids; (b) SP artifacts compared with chocolate-colored radiolarite; (c) SP artifacts compared with flint with inclusions; (d) SP artifacts compared with oolitic flint; (e) Ardales artifact compared with flint with inclusions; (f) Ardales artifact compared with Turonian-type flint.

Full size image

Discussion

Geoarchaeological interpretation

The lithic assemblage from Cueva de Ardales is characterized by high lithological diversity throughout the entire archaeological sequence, from the Middle Paleolithic to the Neolithic (Fig. 5). The most frequently represented raw materials are red radiolarite, massive flint, and flint with inclusions, which show consistent presence across all stratigraphic levels. During the Middle Paleolithic, the predominance of red radiolarite (40 pieces), followed by massive flint (21) and flint with inclusions (16), suggests a selection based on regional availability. Although the assemblage excluding the OKD group is limited (9 pieces), it maintains the same trend. In the Solutrean, and especially in the Neolithic, there is a noticeable expansion in the lithological spectrum, with frequent use of white flint with microfossils, “Azulejo”-type flint, oolitic flint, and multicolored radiolarites. This could reflect a more systematic management of raw materials and broader knowledge of the sources available in the surrounding area. Sampling conducted on the fluvial terraces near the site supports this interpretation (Fig. 12), as it reveals a wide variety of flint types, including those present in the archaeological assemblage. Notably, Turón-type flint and white flint with inclusions are absent from the environment around Sima de las Palomas but have been identified in secondary deposits in the Ardales area (Fig. 12). According to Lozano et al. (2010)88, Turón-type flint would originate from Numinoide units cropping out near Ardales (Fig. 3b), and these materials would have been transported by the Turón River and deposited in the immediate surroundings of the cave (Fig. 12). A similar process may explain the presence of white flint with inclusions, although its primary origin remains unidentified. Overall, the pattern observed at Cueva de Ardales indicates a broad and sustained procurement strategy, with access to multiple regional sources and diversified raw material management over time.

At Sima de las Palomas, a lower lithological diversity is observed, dominated by homogeneous massive flint and red radiolarite, both during the Upper Paleolithic and the Holocene (Fig. 6). This trend persists even when OKD pieces are excluded, with these lithologies dominating all techno-typological categories (1GNB, 2GNB, and PB). Although other varieties, such as flint with inclusions, oolitic flint, “Azulejo”-type flint, or multicolored radiolarites appear occasionally, their representation is minimal. This lower diversity is closely related to the local geological context, as the fluvial terraces of the Río de la Venta, near the site, primarily contain massive flint and red radiolarite, originated from the Middle Subbetic and Penibetic zones, as observed in thin sections of samples collected from more distant prospecting points. The absence of certain types (Turón-type flint or white flint with inclusions) despite intensive sampling confirms local geological limitations in the availability of these lithologies.

The correspondence between the archaeological raw materials and local geological sources reveals a recurrent and well-established strategy of nearby resource procurement. However, the occasional presence at Sima de las Palomas of pieces made from Turón-type flint and white flint with black inclusions (Fig. 6) suggests either sporadic contact and exchange with groups from other areas, such as Cueva de Ardales, or an occasional expansion of the procurement territory towards the south. This expansion may have been motivated not only by the search for raw materials, but also by the exploitation of subsistence resources, taking advantage of natural routes in the valleys now occupied by water reservoirs, which in the past would have functioned as ecological corridors for both human populations and game species (Fig. 12).

Fig. 12
figure 12

Geological map of the area indicating the studied procurement zones. Each zone shows the different flint types identified. Red radiolarites, massive flint, flint with inclusions, oolitic flint, Azulejo-type flint, and cream-colored radiolarite are present in all three terraces. Only the terrace near the río Turón, marked in green, contains Turonian-type flint and white flint with black inclusions. Black arrows indicate the potential procurement areas based on the results obtained. Possible movement routes of human groups associated with Sima de las Palomas or exchange zones between both populations—already active during the Paleolithic—are also shown in green.

Full size image

Archaeometric methodology

The preliminary visual analysis using a binocular magnifier and macroscopic examination enabled an initial lithological classification of the studied flints, providing a first approximation of their general characteristics. Although this approach is useful for fieldwork, it is insufficient to precisely determine the exact origin and depositional environment of the materials, making it essential to complement these results with advanced microscopic techniques.

The geological characterization of flint in archaeological contexts of southern Iberia has traditionally been limited to macroscopic approaches17,48,49,53,91,92 or to isolated microanalytical studies focused on specific raw material types29, without integrating systematic territorial analyses or mobility dynamics. In this context, the present work represents a decisive methodological advance by providing two fundamental contributions. First, it constitutes the first study to demonstrate the effectiveness of cathodoluminescence (CL) as a tool for the geochemical characterization of flint. Although relevant advances had been made, such as those in Olduvai93 or the studies by Legg et al.32 and Boszhardt et al.94 on quartzites in the Great Lakes region and Mississippi Valley, their application to flint had mainly focused on the study of microstructures. Traditionally, CL had been used to analyze diagenetic processes in carbonates and silicified rocks, as well as in archaeological ceramics95,96, but its potential for discriminating flint subtypes and establishing robust geological correlations within archaeological assemblages had not been explored. In this study, CL enabled the precise identification of variations in silicification processes, replacement and recrystallization patterns, and post-diagenetic alterations, many of which are undetectable by macroscopic or conventional microscopic techniques. Although CL can be performed directly on rock surfaces, optimal results are generally achieved on thin sections, allowing a more accurate evaluation of internal features. As CL provides information distinct from POM, it may serve as a valuable complementary approach rather than a standalone solution. However, as an innovative method, there is still a lack of comprehensive reference datasets for flint and related rocks from different regions, which currently limits its broader comparative potential. Nevertheless, the promising results obtained here underscore the significant potential of CL for future research on siliceous raw materials. Second, this research presents the first comprehensive multi-analytical approach applied to the petrological characterization of flint in southern Iberia, combining polarized optical microscopy (POM), cathodoluminescence (CL), scanning electron microscopy (SEM), and X-ray diffraction (XRD). This strategy has overcome the limitations of previous studies focused on morphological descriptions or isolated local analyses and provides a solid empirical foundation for understanding the availability, selection, and circulation of lithic raw materials in geologically complex settings. It is important to highlight that the geochemical analysis of siliceous materials, particularly radiolarites, presents specific challenges due to their high compositional similarity among sources from different regions. Therefore, robust multi-analytical characterization is strongly recommended to improve the reliability of provenance interpretations97,98,99. This approach complements the combined use of POM, CL, SEM, and XRD providing a more reliable framework for source discrimination.

The results obtained through X-ray diffraction (XRD) allow for the identification of significant mineralogical differences among the various flint types analyzed. The systematic presence of quartz (Qz) in all samples confirms the predominantly siliceous origin of these materials. However, the secondary detection of dolomite (Dol) and calcite (Cal) in specific varieties (Fig. 10) suggests particular diagenetic processes. These carbonate minerals may result from postgenetic inputs or be remnants of originally carbonate components that were partially silicified, leaving traces of calcite or dolomite within their structure. The X-ray fluorescence (XRF) analyses complement and reinforces these observations. The predominance of SiO2 supports the interpretation of a dominant siliceous matrix, previously identified through POM and CL. Variations in CaO content confirm the presence of dolomitized or calcitic phases in specific samples. Similarly, elevated concentrations of Fe2O3—particularly in “Azulejo”-type flint and multicolored radiolarite—are associated with oxidative processes and the presence of opaque minerals. In radiolarites, this iron content may explain their characteristic reddish tone, while in the case of the Azulejo type, it is linked to the dendritic stains observed macroscopically. Moreover, in this specific variety, high concentrations of Al2O3 and K2O may indicate silicate-derived detrital inputs. Trace elements such as Ni, Co, and V were also detected as minor elements. Their presence may be associated with specific redox conditions or adsorption processes on ferruginous and manganiferous phases during sedimentation or diagenesis. In this regard, a slight hydrothermal influence cannot be ruled out, related to the active geological conditions in the pelagic domain of the Subbetic during the Jurassic period100. While these geochemical results require further analysis for conclusive interpretation, they provide valuable insights into the geological provenance and formation processes of the different flint types. Lastly, the presence of carbonate phases has relevant technological implications, as it may affect physical properties such as fracture behavior and knapping quality. The recurrent selection of certain flint types, particularly red radiolarites at both sites (Figs. 5 and 6), may reflect empirical knowledge, on the part of past societies, of the technical properties of each variety.

In this context, the development of geological reference lithotheques containing systematically and multi-technically characterized samples becomes essential. Having a broad and well-documented lithotheque not only facilitates the comparison and accurate assignment of archaeological materials to their geological sources but also enables the detection of potential variability within a single site, the identification of post-depositional alteration processes, and ultimately, the refinement of models for raw material procurement and mobility in prehistoric societies. Therefore, the existence of well-structured lithotheques is a prerequisite for effectively applying these advanced characterization approaches and ensuring the replicability and robustness of provenance studies that support solid multi- and interdisciplinary archaeological research.

Conclusions

A clear pattern has been identified in the procurement and use of flint by prehistoric societies in the study area, revealing a marked preference for nearby sources, mainly fluvial terraces, due to their proximity to settlements and the ease of material collection. At the sites of Cueva de Ardales and Sima de las Palomas, the availability of raw materials in the immediate surroundings favored intensive exploitation. Specifically, the prehistoric communities of Sima de las Palomas primarily obtained flint from the Río de la Venta–Cerro del Olivar and the terraces of the Guadalteba River, while in Cueva de Ardales, procurement focused mainly on the Turón River terraces and possibly also those of the Guadalteba.

In Ardales, the confluence of materials from the Numinoide, Penibetic, and Subbetic units crossed by the Turón River and its tributaries facilitated access to a wider diversity of siliceous materials. In contrast, Sima de las Palomas is located in a more geologically homogeneous setting, restricting the range of available raw materials. No significant relationship was observed between raw material type, artifact typology, and chronology, indicating continuous and long-term use of these terraces from the Paleolithic to the Neolithic. These procurement patterns suggest that the area now occupied by the Guadalteba, Guadalhorce, and Conde del Guadalhorce reservoirs functioned as a strategic point of connection between the human groups inhabiting both caves during the Paleolithic and Neolithic, allowing us to infer a scenario of shared procurement routes and exchange networks. This would have facilitated the circulation of both raw materials and technical knowledge, providing new insights into the social and technological behavior of these prehistoric societies.

This study has demonstrated that the combination of polarized optical microscopy (POM) and cathodoluminescence (CL) constitutes a novel and highly effective methodological approach for the comprehensive characterization of flint in complex archaeological and geological contexts. POM enabled precise petrographic classification, revealing in detail the mineralogical composition, textures, and fossil structures essential for interpreting the original depositional environments of these raw materials, particularly in light of the geological complexity of the Betic Cordillera. It was complemented by CL, which revealed internal structures and components invisible under traditional optical techniques, especially those of carbonate origin or related to recrystallization and dolomitization processes. The previously untapped potential of CL to differentiate phases of silicification and specific diagenetic stages significantly enhances interpretative resolution, thereby providing a robust and comprehensive framework for pinpointing provenance and understanding the geological evolution of flint used by prehistoric communities in southern Iberia.

Furthermore, chemical analyses and scanning electron microscopy (SEM) have proven to be effective tools for examining postsdiagenetic processes. These techniques allowed for the observation of surface structures and post-depositional recrystallization phases in flint. However, due to the high compositional homogeneity of this material, they do not permit direct characterization. Therefore, the data obtained should be considered complementary to the observations made using optical microscopy (OM), cathodoluminescence (CL), and binocular magnification, as their combined use offers a more comprehensive perspective on the alteration and transformation processes of siliceous raw materials over time.

Lastly, the development of complete, well-documented, and georeferenced lithotheques is essential for building a robust reference database that enhances accuracy in the identification and classification of flint in archaeological and geoarchaeological research.