Journal of Archaeological Science: Reports 29 (2020) 102120

Contents lists available at ScienceDirect

Journal of Archaeological Science: Reports

journal homepage: www.elsevier.com/locate/jasrep

The cave of Atxurra: A new major Magdalenian rock art sanctuary in Northern T ⁎ Diego Garatea, , Olivia Riverob, Joseba Ríos-Garaizarc, Martin Arriolabengoad, Mª Ángeles Medina-Alcaidee, Juan Francisco Ruiz-Lópezf, Iñaki Intxaurbea, Sergio Salazarb, Iñaki Libanog a Instituto Internacional de Investigaciones Prehistóricas de Cantabria (IIIPC). Universidad de Cantabria. 39005 Santander, Spain b Dpto. Prehistoria, Historia Antigua y Arqueología, Universidad de Salamanca, 37008 Salamanca, Spain c Archaeology Program, Centro Nacional de Investigación sobre la Evolución Humana (CENIEH), Paseo Sierra de Atapuerca 3, 09002 Burgos Spain d Dpto. Mineralogía y Petrología, Euskal Herriko Unibertsitatea/Universidad del País Vasco, 48940 , Spain e Dpto. Geografía, Prehistoria y Arqueología, Euskal Herriko Unibertsitatea/Universidad del País Vasco, 01006 Vitoria-Gasteiz, Spain f Dpto. de Historia. Universidad de Castilla – La Mancha. 16001 Cuenca, Spain g Edestiaurre Arkeologia Elkartea, 48650 Barrika, Spain

ARTICLE INFO ABSTRACT

Keywords: The discovery of Palaeolithic parietal art in the cave of Atxurra took place within an archaeological surveying Cave art project that has been carried out over the last decade in the eastern Cantabrian region. As a consequence of this Archaeological context project, the number of caves with parietal art known in this region has tripled. The case of Atxurra Cave is a Engravings remarkable contribution because of the number of parietal representations and the presence of a related external Magdalenian and internal archaeological context. Through this contribution, we present the main data derived from the cave Accessibility study project that is currently underway, as well as its implications for the reformulation of symbolic interaction during the late Upper Palaeolithic.

1. Introduction (Garate et al. 2016). The decorated sectors appear at a distance of be- tween 186 m and 366 m from the prehistoric access, the entrance of In recent years, research on Palaeolithic parietal art in the north of Atxurra, in the upper level of the system, and they are located mainly the Iberian Peninsula has been reactivated to a remarkable extent, above side-ledges reached by more or less dangerous climbs. Because of especially in the eastern part of the region. It is an area where the this important discovery, a project for the integral study of the cave is density of decorated caves was long thought to be significantly lower underway, to be carried out between 2016 and 2020. than in bordering regions (Central and Western Cantabrian Region, The aim of this paper is to share the current state of the study in Northern Pyrenees and Périgord), despite its geostrategic position be- Atxurra Cave, involving different research lines, as an important site for tween them, on the natural connection of the Iberian Peninsula with the understanding symbolic contacts between populations in Western rest of the Europe. Europe during the Late Palaeolithic. As a result of the new research, the number of known parietal as- semblages has tripled (Fig. 1), from a scant ten to more than thirty, 2. Materials and methods including important Magdalenian sanctuaries such as Aitzbitarte IV, Armintxe and Atxurra (Garate, 2018). The latter site is the main object 2.1. Atxurra Cave of study in this publication. An archaeological excavation project to assess the potential of the Atxurra Cave is in the municipal district of (Basque remaining deposit was restarted in 2014 under the direction of D. Country, Spain) on the right bank of the River Lea, at 70 m a.s.l. and Garate. During the second excavation season in 2015, D. Garate and I. 6 km from the modern coastline. It is located in an area with numerous Intxaurbe commenced examining the walls inside the cave and found archaeological sites, including such important Late Upper Palaeolithic several decorated panels with dozens of Magdalenian-style engravings deposits as Lumentxa, Santa Catalina, Armintxe and Laminak II, among

⁎ Corresponding Author. E-mail address: [email protected] (D. Garate). https://doi.org/10.1016/j.jasrep.2019.102120 Received 2 July 2019; Received in revised form 18 November 2019; Accepted 22 November 2019 2352-409X/ © 2019 Elsevier Ltd. All rights reserved. D. Garate, et al. Journal of Archaeological Science: Reports 29 (2020) 102120

Fig. 1. Decorated caves in Eastern Cantabrian Spain and the Western Pyrenees (base: https://maps-for-free.com): black discovered during the 20th century; red found in the 21st century. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) others (Garate, 2014). could be identified or were larger than 3 cm were located with X, Y, Z The archaeological deposit in Atxurra Cave was discovered by J.M. coordinates using a Leica© TS02 Power-7 laser total station. As each Barandiaran in 1929 and excavated in 1934 and 1935. The results were remain was collected, it was assigned a consecutive record number published very briefly much later (Barandiarán, 1961) and the remains according to the square and spit. have only been reviewed partially since then (Barandiaran Maestu, 1967; Fernández Eraso, 1985; Mc Coullough, 1971; Utrilla, 1976, 2.2.2. Parietal surveying 1981). Currently all the stored materials from Barandiaran’s collection This consists of a detailed inspection of all potential art surfaces: are under re-study (Rios-Garaizar et al., 2019). floors, walls and ceilings. Different LED illumination systems were The cave of Atxurra formed in Aptian-Albian reef limestones (Lower employed to identify paintings and engravings and the cave was sec- Cretaceous) outcropping in the province of , near the coast in the torized in order to organise the survey. The rock art record was made at eastern sector of Northern Spain. The entrance is small, 2 × 1.5 m in different levels. Firstly, panel sketches were drawn to obtain a first size, and triangular. The cave consists of two subhorizontal passages reading of parietal motifs. Secondly, individual technical datasheets (cave levels) situated at different heights. The lower one called Armiña, were used to note down all the graphic motif information (identifica- and the upper one called Atxurra. The entrances are located on the right tion, measurements, techniques) and the surface characteristics (con- bank of the Zulueta stream, a tributary of the River Lea, which flows stitution, shape, taphonomic alterations). Finally, 2D and 3D digital into the Bay of Biscay. There are two points of union between the two methods were used to record the rock art panels and the graphic units. levels, one of them a few metres from the upper level occupation site in Atxurra and the other at the end of the lower level passage in Armiña, 2.2.3. Digital recording methods under the point where the decorated sector begins. As mentioned For the 3D reconstruction of parietal motifs, the remote sensing above, the decorated sectors are found between 186 m and 366 m from recording technique selected was close-range photogrammetry because the prehistoric access; after this point, the cave continues for around of its flexibility in reconstructing 3D panels at different scales, the high 400 m, after a blockage recently removed, with an absence of any accuracy of the results and the production of fast deliverables as a low- Palaeolithic remains (drawings or any other evidence) (Fig. 2). cost solution. Agisoft PhotoScan Pro 1.4.5 was the software selected to process the images. However, the difficulties involved in documenting 2.2. Archaeological survey method such fine engravings over large surfaces, as in the panel on the Ledge of the Horses (11 m long) required a specific documentation methodology The project has aimed to define the occupations that took place in to overcome the limitations of this technique (Rivero et al., 2019a) the cave, document and analyse the rock art, and determine the nature (Fig. 3). This took place in three phases. First a general model of the of activities related to the artistic work. This would enable a more decorated wall was obtained with uniform frontal lighting. This al- global understanding of the different uses and visits made to the in- lowed precise restitution of the wall. The second model concentrated on terior of the cave by prehistoric groups. the documentation of the graphic evidence. This used raking light to highlight the engravings, allowing their reliable restitution. Lastly, both 2.2.1. Habitat excavation models were fused using the topographic coordinates. In this way, the In order to define more precisely the known archaeological se- final product integrated the graphic restitution of the engraved re- quence, the section was cleaned up and a test pit with a surface area of presentations with the model of the wall. 2m2 was excavated. This was performed by squares and arbitrary spits adapted to the stratigraphy, in a method which had proved effective at 2.2.4. Inner archaeological context such other sites as Arlanpe (Rios-Garaizar et al., 2013). The archae- Detailed and interdisciplinary analysis of the palaeo-speleological ological remains (lithic objects and identifiable faunal remains) that remains on the preserved palaeo-floors on the high ledges in the cave

2 D. Garate, et al. Journal of Archaeological Science: Reports 29 (2020) 102120

Fig. 2. Topographic plan of Atxurra Cave with the area decorated enhanced (Gim-Geomatics).

merit attention because they might have some direct or indirect re- lationship with the human use of a cave (Clottes, 1993), or as traces of the internal frequentation of bears. The methodology for the study consisted of several steps (Garate et al., 2015, Medina-Alcaide et al., 2018b). The data gathering included the division of the cave space into sectors according to its geomorphology, the detailed inspection of surfaces (working with portable microscopes Dino-lite® 10x-460x), the cataloguing of remains, their geolocation, their photographic doc- umentation, sampling the superficial remains, conducting archae- ological tests in singular sites and the “flotation” of extracted sedi- ments. Additionally, a photogrammetric survey of some palaeo-floors was made with a GIGAPAN EpicPRO. Taking into account the diversity of the archaeological remains, the examination of the inner archaeological context must use an inter- disciplinary approach to reach an overall understanding of the specific activities that may have been carried out in each sector, apart from graphic activity. We have included anthracology (wood charcoal), mi- cromorphology (palaeo-floors), use-wear analysis (flint tools), radio- carbon dating (organic material), etc.

3. Results

We have divided the current results according to the three areas of archaeological study of the cave. Although the project is at an advanced stage, all the results are based on partial data. However, they allow a first characterization of the archaeological content of the cave:

3.1. Entrance occupation site

The excavation in the entrance of Atxurra (2014–2015) covered a Fig. 3. Digital documentation of the decorated walls in the high sectors of surface area of 1.5 m2 affecting the northern section left by Atxurra Cave (D. Garate). Barandiarán. Here a 1.5 m sequence with VIII stratigraphic units has been defined (Fig. 4). The lowermost archaeological occupations oc- provides valuable information about the “use-life” of the cave during curred during the Noaillan Gravettian (Unit VI). Later occupations are the past. Preferentially, we focused on remains stemming from human attributed to the Lower Magdalenian with microlithic triangles (Units V activities. We also recorded remains derived from animal activity. They and IV), including a sub unit with intense occupation and hearths (Unit IVb). Finally, Late Magdalenian occupations (Unit II) include a 6 cm

3 D. Garate, et al. Journal of Archaeological Science: Reports 29 (2020) 102120

Fig. 4. Section N of the 2014–2015 excavation (J. Rios-Garaizar). I(a,b) - Holocene; II(a,b,c,d) - Upper Magdalenian; III- Middle Magdalenian; IV(a,b) - Lower Magdalenian; V - Lower Magdalenian; VI- Gravettian; VII - Paleontological; VIII - Sterile. thick carbonous layer (IIb) that evidences the presence of hearths. during the late Magdalenian. That would have been the present en- The review of Barandiaran’s excavation materials (Rios-Garaizar trance of Atxurra Cave as the excavation in the first chamber in Armiña et al., 2019) and the evidence collected in the 2014–15 section recovery Cave has shown that the entrance of that cave would have been prac- excavation suggest that the entrance was intensely occupied during the tically blocked or very difficult in the Upper Palaeolithic (Rios-Garaizar Lower Magdalenian. During the Late Magdalenian, there is also evi- et al., submitted). dence of continuous occupation, but less intense than in the Lower The sectors with parietal art and archaeological remains end near a Magdalenian. narrowing of the passage caused by flowstone. However, a further section of the cave about 400 m long can be reached through a small fl 3.2. Armiña ephemeral occupation gap 4 m above the cave oor. The absence of archaeological remains in this passage and the conclusions of the geomorphological study indicate In 2016 and 2017 several tests and a 6 m2 excavation were per- that this gap would have been blocked, or it would have been extremely ffi formed in the main hall of Armiña. At the top of the stratigraphic se- di cult to squeeze through it, until it was opened up at some point quence were several sterile units (Ia-Ic, II) that sealed Unit III, con- between the 1930s and 60s (Arriolabengoa et al., submitted). ffi taining a few transported bones, flint fragments and charcoal Access to the decorated part of the cave is quite di cult from fragments, and Unit IV where an ephemeral occupation dating to the Atxurra entrance. After the entrance hall with the archaeological de- Late Magdalenian (14,722–14,059 cal BP) has been identified. This posit, the inner part of the cave is entered through a small window. A occupation is characterized by the presence of a small hearth, an ochre ramp then connects with Armiña Cave on a lower level. By going round stain and few lithic and faunal remains. This ephemeral occupation has the side of the ramp along a steep ledge, the passage continues with a been interpreted as an Inner Archaeological Context (IAC) with a non- low irregular roof until it decreases in height at the start of a crawl utilitarian purpose, despite its proximity to the current entrance to the several metres long. The geomorphologic study of the cave stresses that cave, which was probably sealed during the occupation (Rios-Garaizar this is the part of the cave that has changed most since the Upper fl et al., submitted). Palaeolithic. The oor is formed by a series of phases of detritic and lithochemical sedimentation up to 30 cm thick, and therefore this passage would have been relatively more comfortable in the Upper 3.3. Rock art ensemble Palaeolithic (Arriolabengoa et al., submitted). The cave continues in a larger passage with hardly any side-pas- From 1881 until the discovery of the Palaeolithic rock art, the cave sages, with an irregular floor with boulders and changes in height, has been open to all kinds of visitors. Thousands of people have entered before reaching the last connection with the end of the lower passage, ffi during this time and more than 800 gra ti have been documented, Armiña Cave. The passage then increases in size, particularly in height. including historical ones from the 19th century but also recent ones Hibernation pits dug by Ursus spelaeus show that the present floor in during those 134 years. Because of this vandalism, the archaeological this area has not changed since prehistoric groups visited it. However, heritage of the cave has been partially damaged. Fortunately, most of the lithology of the walls in this area (stratified limestone with flint the rock art is in hidden sectors, away from the route followed by those nodules) favours collapses, as shown by the accumulations of clasts in visitors. However, some of the decorated walls had been erased, hin- some places. dering a correct reading of the Palaeolithic traces. The decorated panels start in this area and over the next 180 m of Currently 19 decorated places are known with depictions of dif- the cave. The passage is narrower, with a variable height of 3 to 4 m, ferent animals, especially bison, horses and ibex, but also hinds, aur- with small side-ledges and hanged chambers. The parietal decoration ochs and a composite animal with bear and reindeer features. All of tends to be in those places, and therefore is not visible from the usual fi them are nely incised on the limestone walls and, in some cases, also route through the cave but only reached by some quite difficult and fi painted in black. To date, 113 animals have been identi ed, corre- occasionally dangerous climbs. A study of the state of the cave when it sponding to a Late Magdalenian style. was frequented and decorated by late Magdalenian groups has de- fi Evidence of Palaeolithic rock art has been identi ed in Atxurra Cave termined the accessibility of the decorated parts of the cave at that time over a distance of 180 m, beginning 186 m from the entrance in use

4 D. Garate, et al. Journal of Archaeological Science: Reports 29 (2020) 102120

Fig. 6. Frieze with engraved bison and a black horse in the Hall of the Bison in Atxurra Cave (O. Rivero and M. Cobo).

Fig. 5. Horse represented by a deep engravings in a calcite flowstone full of bear claw marks in the Alcove of the Claw Marks in Atxurra Cave (D. Garate).

(Intxaurbe et al., submitted). The results show that the artists’ choices of panel were directly affected by the difficulty of access. They chose the places that were hardest to reach and discarded much easier ones. In this way, the areas that are most difficult to access display the largest concentrations of graphic activity, even though the cave offers many more potential spaces which are less demanding. Of the 19 decorated areas mentioned above, most contain a mod- erate number of representations (up to a maximum of 30 graphic units, generally stigmas or non-figurative lines and at most ten figurative units or animals). In addition, red stains are found in two sectors. However, three areas of the cave accumulate large numbers of animal figures, with dense concentrations of engravings superimposed on one another over several metres.

Fig. 7. 3D graphic restitution of the central part of Panel J.II in the Ledge of the 3.3.1. Alcove of the claw marks Horses, detailing a pendant engraved with ibex figures, two of them in frontal On the right-hand wall, 216 m from the entrance, at least two dozen view. (O. Rivero; J.F. Ruiz-López; I. Intxaurbe; S. Salazar and D. Garate). animal figures appear on calcite flowstone covered by bear claw marks and in an adjacent suspended passage. These mainly represent bison, twelve ibex, two of them in frontal perspective (Fig. 7). The ledge can ibex and horses, all of them engraved (Fig. 5). They are extremely be reached by climbs of moderate difficulty at either end, although ffi fi fl di cult to access, as these gures are located 5 m above the oor of the access is easiest from the side furthest from the entrance. passage, at the top of a practically vertical wall. They can be reached by climbing up a series of exposed ledges and steps. The geomorphologic study has concluded that the artists must have followed a similar route, 3.4. Inner archaeological context as the speleothems separating this area from a proximate sector (Hall of the Bison) already existed in the Upper Palaeolithic (Arriolabengoa More than 700 archaeological remains have been documented in et al., submitted). different sectors in the inner context of Atxurra Cave. These remains have their origin in the different activities carried out by the palaeo- 3.3.2. Hall of the bison groups inside the cave, such as artificial lighting (wood charcoal and Located on the right-hand wall as the continuation of the previous other combustion residues) and the execution of the Palaeolithic art, for sector, 221 m from the entrance, this is a semi-circular space with a low example, flint tools found at the foot of some decorated walls. Other roof. Several panels are found on the left of the chamber, with en- types of evidence have been also recorded, such as possible stigmas of gravings and remains of black paint, representing at least twelve bison the extraction of some of the natural flint nodules that are present in the (Fig. 6), horses, and more exceptionally, a stag in frontal perspective. cave, and probably ancient anatomical prints on the clay floor (Fig. 8). Despite its proximity to the previous sector, access is easier up a gentler These prints are usually preserved in high zones of the passage, and are climb from the main passage. patinated, but as the cave was heavily frequented in the last century, it is difficult to ensure their Palaeolithic chronology. ff 3.3.3. Ledge of the horses These remains are distributed in fourteen di erent sectors, espe- Also on the right-hand wall, at the end of the decorated part of the cially in the high ledges of the decorated sector in the upper gallery, cave, 330 m from the Palaeolithic entrance, this is on a natural plat- between 143 and 365 m from the entrance. Seven of them only contain form, initially 4 m above the floor of the passage although this de- archaeological remains, without Palaeolithic rock art. In contrast, creases to about 2.5 m at the end. The ledge is about 1.5 m wide, al- twelve spaces only contain rock art and no archaeological remains. lowing access to a wall about 12 m long. About 84 graphic units have These sectors are located in the lower part of the gallery, where pre- been identified in two panels. They include about 50 animal engrav- servation is conditioned by continuous human presence during the last fl ings, many of them superimposed and overlapping. Among the most century and periodic oods. Just like the rock art, the archaeological ff outstanding figures are two horses, easily viewed as they were drawn remains are not distributed homogeneously in the di erent sectors and by scraping, several large bison, a discreet group of hinds and about most of them are located in the three areas already cited:

5 D. Garate, et al. Journal of Archaeological Science: Reports 29 (2020) 102120

Fig. 8. Different archaeological remains in the inner context. A. Wood charcoal. B. Flint tool. C. Bear bone. D. Probable prehistoric hand print (M. Medina-Alcaide).

3.4.1. “Alcove of the claw marks” interior archaeological context remains. This discovery provides very We have found about thirty wood charcoal fragments, up to 1 cm relevant information about artistic activities on the Iberian side of the long in their maximum axis, on the surface level, in the clay floor. These Pyrenees during the Upper Magdalenian, and about other palaeo-spe- remains are linked to the underground lighting activities. The scattered leological incursions. and random distribution of charcoal remains, together with the absence Firstly, the new study of the historic excavation area (1934–35) has of other combustion residues (ash, rubefaction), indicates that their verified the existence of a very long occupation sequence, from the origin would be a mobile lighting system such as a wooden torch. early Gravettian to the Final Upper Magdalenian. In this way, it would However, black marks derived from the use of torches have not been be the longest Upper Palaeolithic occupation of the eight known to date recorded on the walls and the ceiling of the sector, despite the narrow in the basin of the River Lea. dimensions. The anthracological analysis indicates a preference for Considering the current state of the research of the rock art, and angiosperm wood and especially of deciduous oak for lighting in this without finishing the surveying work in some decorated areas, the en- sector. In addition, in the first section of this space there is a cave bear semble located in the cave consists of engraved animal figures (and hibernation pit with superimposed charred remains. This indicates that some painted in black), engraved non-figurative lines and some stains the prehistoric frequentation occurred after the temporary presence of in red and black. Thematically, the ensemble is similar to other recent this animal. Also, the Palaeolithic engravings are superposed on bear Magdalenian sites in Cantabrian Spain, the Pyrenees and Périgord re- claw marks on flowstone in this sector. gion (Sauvet and Wlodarczyk, 2001), with a predominance of figures of bison and horses. However, the large number of ibex figures, at the 3.4.2. “Hall of the bison” same level as the other two principal themes, is noteworthy. This relates We have found six bone fragments, mainly of cave bears, and nine the site with two recent Magdalenian ensembles in the west of Can- scattered wood charcoal remains, up to 1 cm long in the maximum axis. tabrian Spain: El Bosque (Fortea, 1996) and Covarón (Arias and Pérez, fi Charred wood of deciduous oak and juniper has been identified in the 1994), which include numerous ibex gures (in these cases painted in wood charcoal analysis. A sandstone pebble with signs of abrasion, black), but also with Niaux (Clottes, 2010) and Trois-Frères (Bégouën possible anthropic impacts and combustion residues has also been and Breuil, 1958) in the Pyrenees, with tens of engraved and painted fi identified. This has been provisionally interpreted as a portable lamp, ibex gures. through an interdisciplinary analysis. The small dimensions of this hall Furthermore, Atxurra Cave contains some compositions found in explain the preference for a small and controlled flame like that pro- other recent Magdalenian caves, like the slender hinds in a line (in the duced by this type of lighting system. great panel in Sector J) (Fig. 9A and B), present in Sovilla (central Cantabrian Spain) (González Sáinz et al., 1993)(Fig. 9C) and Ker de Massat (central/eastern Pyrenean region) (Barrière, 1990)(Fig. 9D). In “ ” 3.4.3. Ledge of the horses the latter cave, a large panel displays the same composition as Atxurra’s In this ledge the palaeo-surface used by the artist who engraved the main panel in Sector J, relating both ensembles very closely: firstly, walls has been extraordinarily well-preserved. Many charcoal remains there are some hinds in line, surrounded by horses and hairy bison; all (probably residue of wood torches) and some flint tools were identified “ ” 2 of them presided by an ibex composition (one of them in frontal view) on the surface, and during the excavation of 7.85 m in 2016 hundreds in the central/upper zone. of charcoal remains and more lithic tools, chips and fragments were The stylistic conventions in the animal figures allow their correla- fi fi recovered. Also ve areas of reddened clay have been identi ed, and tion with others located in different regional contexts (Rivero and the presence in some of them of charcoal and ashes suggest that they Sauvet, 2014). The bison’s horns and legs perspective in two planes is fi fi were xed res used probably for lighting. The anthracological analysis usual in recent Magdalenian chronologies (ca. uncal 14,500–11,700 BP) indicates that the combustion material used for lighting this area comes in Cantabrian Spain, Pyrenees and Périgord region (Fortea et al., 2004). from deciduous oak and juniper wood, with a marginal addition of The horse figures also display some particular Recent Magdalenian willow wood and Scots pine. In addition, the preliminary analysis of the conventions, like small lines representing the hair in the hindquarters, abandoned lithic tools (burins, retouched blades) suggests that they as in a portable figure from Linar Cave (Central Cantabrian Spain) (De were used for engraving the limestone wall. las Heras et al., 2008) and one from El Polvorín Cave (Eastern Can- tabrian Spain) (Ruiz Idarraga, 2009) or the ventral M quartering, re- 4. Discussion presenting different tones of the pelage (Vázquez, 2017). Other con- ventions seen in the animal figures in Atxurra are the indications of Although Atxurra Cave has been vandalized over a long period of pelage and internal anatomical details, like ears, eyes, etc. Equally time; it has incredibly preserved a unique rock art ensemble formed by noteworthy is the abundance of “wounded” animals –with projectile- around one hundred animal engravings and an enormous amount of

6 D. Garate, et al. Journal of Archaeological Science: Reports 29 (2020) 102120

Fig. 9. A and B- Engraved hinds in the composition on the Ledge of the Horses (O. Rivero, D. Garate, I. Intxaurbe, S. Salazar). C- Engraved hinds in the main panel in Sovilla Cave (González Sáinz et al., 1993) D- Engraved hinds in one of the main panel in the cave of Ker de Massat (Barrière, 1990). like linear motives and / or inverted “V” motifs inside them -, especially and two bison, but much deteriorated (like all the painted motifs in the bison (one with almost twenty projectiles in it), but also ibex. Similarly, cave). the presence of animals in frontal view - two ibex and a deer - is re- Atxurra Cave also conserves an exceptional inner archaeological levant, as it is a convention with a very precise chronological attribu- context closely related to the rock art because of its location. The huge tion in portable art (Barandiaran Maestu et al., 2013; Rivero et al., size of the ensemble (with more than 700 remains) provides significant 2014; Utrilla, 1987). Therefore, taken together, the figures display re- information about the prehistoric activities that took place in the deep markable internal coherence that can be attributed, without reserva- zones of the cave, particularly the artistic production, since flint tools tion, to the final moments of the Upper Palaeolithic and in particular to used to make the engravings have been discovered. In addition, nu- the cultural period of the Upper Magdalenian (ca. uncal. 13,000–11,700 merous remains are related to the illumination in the cave. The use of BP) (Fig. 10). three different prehistoric lighting systems has been identified: wooden Technically, engraving is almost exclusive, in different variants torches (scattered charcoal), a portable lamp and fixed fires. Juniper -incised, scraped, etc.-, sometimes combined with black paint, very and oak wood was the main wood fuel collected for this activity. In the poorly preserved. In the larger animal figures, multiple engraved thin nearest regional context, these types of wood have been found in the lines usually form the outline, and a single line represents internal deep zones of other caves with Palaeolithic rock art, such as Morgota anatomical details (ears, eyes, etc.). Besides, some figures have scraped Cave (Medina-Alcaide, 2015) or Ondaro (Ruiz-Alonso, 2018). The zones to depict internal quartering, and some hind figures have striated combination of oak and juniper wood is suitable for the lasting pro- zones in the head and chest zone. This last convection is usual in duction of the flame. The former has a high calorific value and this Cantabrian Lower Magdalenian figures (Utrilla, 1979), but also in other produces long-term combustion. The latter intensifies the combustion chronological contexts, as in our case (Fig. 9A) (Rivero et al., 2019b). In due to its resin content, it is consumed slowly, and, unlike other con- some parts of the cave, rubbing spots are observed in red, possibly re- ifers, it produces a small amount of smoke (Perlès, 1977). Both species lated to prehistoric underground progression (Medina-Alcaide et al., were available in the immediate surroundings of Atxurra Cave when the 2018a) but in no case do they form figurative motifs. Finally, there are engravings were drawn (Iriarte, 2011, Uzquiano et al., 2016). some depictions in black (without engraved lines), at least two horses

7 D. Garate, et al. Journal of Archaeological Science: Reports 29 (2020) 102120

Fig. 10. Distribution of caves with ibex and deer images in frontal view in parietal and / or portable art (I. Intxaurbe, base; https://www.gmrt.org/GMRTMapTool/. Sea level is represented about 60 m lower than today (Lambeck et al., 2014)): 1. La Paloma; 2. Sofoxo; 3. Tito Bustillo; 4. La Güelga; 5. Cueto de la Mina; 6. Llonín; 7. El Pendo; 8. Morín; 9. El Polvorín; 10. La Chora; 11: El Otero; 12. El Valle; 13. Bolinkoba; 14. Santimamiñe; 15. Lumentxa; 16. Atxurra; 17. Agarre; 18. Urtiaga; 19. Ekain; 20. Aitzbitarte IV; 21. Torre; 22. Abauntz; 23. Berroberria; 24. Isturitz; 25. St Michel (Arudy); 26. Lortet; 27. Gourdan; 28. Mas d́Azil; 29. Le Portel; 30. Ker de Massat; 31. La Vache; 32. Niaux; 33. Belvis; 34. Laugerie-Basse; 35. La Madeleine; 36. La Mairie de Teyjat; 37. Montgaudier.

5. Conclusions Upper Palaeolithic art, and especially to investigate the activities of Upper Magdalenian societies not related with daily subsistence, like the Although the study of Atxurra Cave has not concluded, it can be exploration and frequentation of difficult and dangerous zones in a affirmed that it is one of the most important sanctuaries of engravings cave, and the production of a large ensemble of rock art inside. (in addition to some paintings in black and a few red stains) in eastern Cantabrian Spain, because of its size (with at least 113 animal figures, Acknowledgements mostly bison, ibex, horses, but also some hinds, a deer, and a composite animal with features of reindeer and bear), its stylistic features, and the The authors wish to thank the Cultural Heritage Service of the archaeological context associated with it. Diputación Foral de Bizkaia for funding the four-year multidisciplinary Some of the depictions, especially the ibex and the deer in frontal study project (2016–2020) “Study of rock art in Atxurra Cave” directed view, indicate a chronology in the Upper Magdalenian for this syn- by Dr Diego Garate. The present study has been carried out within the chronic ensemble, contemporaneous with a level found in the habitat framework of this project. We would also like to thank the University of site at the entrance, and the ephemeral occupation of Armiña (inter- Xían Jiantong for analyzing and carrying out the dating of the uranium preted as an inner archaeological context). This convention (animals in series of samples, as well as Gim-Geomatics for the 3D model of the frontal view) links this rock art ensemble with others located above all cave and the following researchers have collaborated in fieldworks: M. in Cantabrian Spain, but also in the Pyrenees and Périgord region. Cobo, J. Cobos, I. de la Fuente, S. Gálvez, M. García-Bustos, P. García- The location of most of the art is peculiar, because of its difficult Bustos, X. Gezuraga, A. Mateo, J. Moreno, R.M. Ruiz, J. Tapia, A.J. access. Some areas are almost 2 m over the path (the ledges), small Torres and J. Vizcaíno. In addition, some results in this paper have been areas (very narrow galleries) and zones elevated in very high parts of partially funded by the research project of the Spanish Ministry the cave (e.g. Alcove of the Claw Marks). This demonstrated the ability “Learning and developing of artistic skills in anatomically modern hu- of Palaeolithic groups to enter deep inside the caves and it indicates the mans: a multidisciplinary approach” HAR2017-87739-P, led by Olivia criteria of selection of decorated spaces at that time (in the case of Rivero. Atxurra, they chose the most complicated areas, which challenges some traditional rock art models). References Despite the frequentation of the cave and vandalism (especially the large graffiti on all the walls), the cave has preserved some zones with a Arias, P., Pérez, C., 1994. Las pinturas rupestres paleolíticas de El Covarón. Zephyrus, large number of archaeological remains on the floor, because of their XLVI, pp. 37–76. Arriolabengoa, M., Intxaurbe, I., , P., Aranburu, A., Ríos-Garaizar, J., Medina- location (on high platforms, above the most usual path). Over 700 ar- Alcaide, M.A., Rivero, O., Hai Cheng, Beyong, Líbano, I., & Garate, D. (submitted). chaeological remains consist mostly of scattered charcoal fragments, Geomorphology for paleocaving: landscape evolution and palaeoenvironmental but also lithic industry in flint, fixed fire remains, a sandstone lamp, and changes in Atxurra-Armiña cave (northern Iberian Peninsula). Quaternary Science Reviews. probable anatomical prints in the patinated clay. These provide valu- Barandiarán, J.M., 1961. Excavaciones Arqueológicas en Vizcaya: Silibranka, Atxurra y able information about the illumination or the objects used to make the Goikolau, Revista de la Excelentísima Diputación de Vizcaya, No 17. In: De art. Barandiarán, J.M. (Ed.), Obras Completas de José Miguel de Barandiarán XV. La Gran – All in all, Atxurra provides an excellent opportunity for the study of Enciclopedia Vasca, Bilbao, pp. 199 219.

8 D. Garate, et al. Journal of Archaeological Science: Reports 29 (2020) 102120

Barandiaran Maestu, I., 1967. El Paleomesolítico del Pirineo Occidental. Universidad de Medina-Alcaide, M.A., Garate, D., Ruiz-Redondo, A., Sanchidrián, J.L., 2018b. Beyond Zaragoza, Zaragoza. art: the internal archaeological context in Paleolithic decorated caves. J. Anthropol. Barandiaran Maestu, I., Cava, A., Gundín, E., 2013. La cabra alerta: marcador gráfico del Archaeol. 49, 114–128. Magdaleniense cantábrico avanzado. In M. de la Rasilla (ed.) F. Javier Fortea Pérez: Perlès, C., 1977. Préhistoire du feu. Masson, Paris. Universitatis Ovetensis Magister: estudios en homenaje, 263–286. Ediciones de la Rios-Garaizar, J., Garate, D., Gómez-Olivencia, A., 2013. La cueva de Arlanpe (, Universidad de Oviedo. Oviedo. Bizkaia). Ocupaciones humanas desde el Paleolítico Medio Antiguo hasta la Barrière, C., 1990. L'art pariétal du Ker de Massat. Presses Univ. du Mirail, Toulouse. Prehistoria Reciente. In: Ríos-Garaizar, J., Garate, D., Gómez-Olivencia, A. (Coords.): ́ Bégouën, L., Breuil, H., 1958. Les Cavernes du Volp. Trois Frères - Tuc dAudoubert à Kobie (serie BAI 3): La cueva de Arlanpe (Lemoa): Ocupaciones humanas desde el Montesquieu-Avantès (Ariège). Travaux de ĺInstitut de Paléontologie Humaine. París. Paleolítico Medio Antiguo hasta la Prehistoria Reciente, 277–294. Diputación Foral Clottes, J., 1993. Contexte archéologique interne. In: G.R.A.P.P. (eds.), L’art pariétal de Bizkaia, Bilbao. Paléolithique. Techniques et méthodes d’étude, 49–58. CTHS, Paris. Rios-Garaizar, J., San Emeterio, A., Arriolabengoa, M., Aranbarri, J., Rofes, J., Marin- Clottes, J., 2010. Les cavernes de Niaux: art préhistorique en Ariège-Pyrénées. Errance, Arroyo, A.B., Rivero, O., Intxaurbe, I., Arranz-Otaegi, A., Salazar, S., Medina Alcaide, Paris. M.A., Garate, D. (2019). Sporadic occupation in Armiña cave (Northern Spain) during De las Heras, C., Montes, R., Lasheras, J.A., Rasines, P., Fatás, P., 2008. Dos rodetes the Upper Magdalenian: what for? J. Archaeol. Sci. Rep. paleolíticos procedentes de las cuevas del Linar y Las Aguas, Alfoz de Lloredo Rios-Garaizar, J., San Emeterio, A., Larrea, M., Marín-Arroyo, A.B., Agudo, L., Cubas, M., (Cantabria). Veleia 24–25, 161–174. & Garate, D. (submitted). La secuencia prehistórica de la cueva de Atxurra (Berriatua, Fernández Eraso, J., 1985. Las culturas del Tardiglaciar en Vizcaya. UPV-EHU, Bilbao. Bizkaia): evaluación de las excavaciones de J.M. Barandiarán Ayerbe (1934–1935). Fortea, J., 1996. La grotte de El Bosque (Carreña de Cabrales, Asturies, Espagne). INORA, Munibe (Arkeologia-Antropologia). n°13, pp. 3–5. Rivero, O., Sauvet, G., 2014. Defining Magdalenian cultural groups in Franco-Cantabria Fortea, J., Fritz, C., García, M., Sanchidrian, J.L., Saucet, G., Tosello, G., 2004. L'art by the formal analysis of portable artworks. Antiquity 88 (339), 64–80. pariétal paléolithique à l'épreuve du style et du carbone-14. In La spiritualité. Actes Rivero, O., Fano, M., Garate, D., 2014. Relationships between recent Magdalenian so- du colloque de la commission 8, UISPP (Paléolithique supérieur) (Vol. 106, pp. cieties in Cantabrian Spain, through the technical and formal analysis of frontal re- 163–176). Eraul Liège. presentations of Ibex. Otte, M., Le Brun F. (Coord.): Modes de contacts et de Garate, D., 2014. Neandertales y Cromañones. Los primeros pobladores de Bizkaia. Guías déplacements au Paléolithique eurasiatique. ERAUL 140, 589–601. del Arkeologi Museoa. 2. Diputación Foral de Bizkaia, Bizkaikoa y Arkeologi Museoa, Rivero, O., Garate, D., Salazar, S., Intxaurbe, I., 2019. The Cantabrian Lower Magdalenian Bilbao. striated hinds on scapulae: Towards a new definition of a graphic morphotype. Quat. Garate, D., 2018. New insights into the study of Palaeolithic rock art: dismantling the Int. 506, 69–79. ‘Basque Country Void’. J. Anthropol. Res. 74 (2), 168–200. Rivero, O., Ruiz-López, J.F., Intxaurbe, I., Salazar, S., Garate, D., 2019 On the limits of 3D Garate, D., Medina-Alcaide, M.A., Ruiz-Redondo, A., Sanchidrián, J.L., 2015. Around art: capture: a new method to approach the photogrammetric recording of Palaeolithic the internal archaeological context of decorated caves. Arkeos 37, 127–130. thin incised engravings in Atxurra Cave (Northern Spain). Digital Appl. Archaeol. Garate, D., Rivero, O., Rios-Garaizar, J., Intxaurbe, I., 2016. La grotte d’Atxurra: un Cult. Heritage, in press. nouveau sanctuaire majeur du Magdalénien au Pays Basque. Int. Newsl. Rock Art Ruiz Idarraga, R., 2009. Fragmento de rodete decorado con el grabado de una figura 76, 1–4. incompleta de caballo del yacimiento de El Polvorín. Arte mueble en Bizkaia. Una González Sáinz, C., Montes Barquín, R., Muñoz Fernández, E., 1993. La Cueva de Sovilla mirada al arte decorativo de la Edad de Piedra Los Cuadernos del Arkeologi 1, 38–39. (San Felices de Buelna, Cantabria). Zephyrus 46, 7–36. Ruiz-Alonso, M., 2018. Maderas carbonizadas en la cueva de Ondaro (, Bizkaia). Intxaurbe, I., Arriolabengoa, M., Medina-Alcaide, M.A., Rivero, O., Ríos-Garaizar, J., Kobie, Paleoantropología 35, 57–66. Salazar, S., Líbano, L., Garate, D. (submitted): Accessibility to Palaeolithic rock-art in Sauvet, G., Wlodarczyk, A., 2001. El arte parietal, espejo de las sociedades paleolíticas. underground spaces: proposal of methodology for the study of human transit in Zephyrvs 53, 215–238. Atxurra cave (Northern Spain). J. Archaeol. Sci. Utrilla, P., 1976. El Magdaleniense inicial en el País Vasco peninsular. Munibe 4, Iriarte, M.J., 2011. Polen y vegetación en la secuencia estratigráfica de Santimamiñe 245–275. (, Bizkaia). La Cueva de Santimamiñe: revisión y actualización Utrilla, P., 1979. Acerca de la posicion estratigrafica de los cérvidos y otros animales de (2004–2006). In: López Quintana, J.C. (Ed.), Kobie (serie BAI 1): La cueva de trazo multiple en el Paleolitico Superior espanol. (A propos de la position strati- Santimamiñe: revisión y actualización (2004–2006), 321–341. Bilbao. graphique des cervidés et autres animaux à tracé multiple dans le Paléolithique Lambeck, K., Rouby, H., Purcell, A., Sun, Y., Sambridge, M., 2014. Sea level and global ice supérieur espagnol). Caesaraugusta Zaragoza 49–50, 65–72. volumes from the Last Glacial Maximum to the Holocene. Proc. Natl. Acad. Sci. 111 Utrilla, P., 1981. El Magdaleniense inferior y medio en la costa cantábrica. Ministerio de (43), 15296–15303. Cultura, Dirección General de Bellas Artes, Archivos y Bibliotecas, Madrid. Mc Coullough, C., 1971. Perigordian facies in the Upper Palaeolithic of Cantabria. Utrilla, P. 1990. Bases objectives de la chronologie de l'art mobilier paléolithique sur la Doctoral Thesis. University of Pennsylvania. Côte Cantabrique. L’ art des objets au Paléolithique, 1: L’art mobilier et son contexte. Medina-Alcaide, M.A., 2015. Los restos vegetales carbonizados del contexto arqueológico Colloque de Foix-Mas d'Azil 11–16 Novembre, 87–10 1987 Ministère de la Cultures interno de Morgota koba (Kortezubi Bizkaia). In: López Quintana, J.C. (Coord): Kobie Clamecy. (serie BAI 5): La cueva de Morgota (Kortezubi, Bizkaia): evaluación arqueológica y Uzquiano, P., Ruiz-Zapata, M., Gil-García, M., Fernández, S., Carrión, J.S., 2016. Late estudio del conjunto parietal paleolítico. 105–119. Diputación Foral de Bizkaia, quaternary developments of Mediterranean oaks in the Atlantic domain of the Iberian Bilbao. Peninsula: the case of the Cantabrian region (N Spain). Quat. Sci. Rev. 153, 63–77. Medina-Alcaide, M.Á., Garate, D., Sanchidrián, J.L., 2018a. Painted in red: in search of Vázquez, M.C., 2017. Despieces ventrales ‘en M’ en las representaciones peninsulares de alternative explanations for European Palaeolithic cave art. Q. Int. 491, 65–77. équidos superopaleolíticos. Sautuola 19, 471–500.

9