S C I E N C E A D VA N C E S | R E S E A R C H A R T I C L E A N T H RO PO LO GY Copyright © 2023 The Authors, some The earliest basketry in southern Europe: Hunter- rights reserved; exclusive licensee gatherer and farmer plant-based technology in Cueva American Association for the Advancement de los Murciélagos (Albuñol) of Science. No claim to original U.S. Government Works. Distributed Francisco Martínez-Sevilla1†*, Maria Herrero-Otal2†, María Martín-Seijo3, Jonathan Santana4, under a Creative José A. Lozano Rodríguez1,5, Ruth Maicas Ramos6, Miriam Cubas1, Anna Homs7, Commons Attribution Rafael M. Martínez Sánchez8, Ingrid Bertin2,9, Rosa Barroso Bermejo1, Primitiva Bueno Ramírez1, NonCommercial Rodrigo de Balbín Behrmann1, Antoni Palomo Pérez2, Antonio M. Álvarez-Valero10, License 4.0 (CC BY-NC). Leonor Peña-Chocarro11, Mercedes Murillo-Barroso12, Eva Fernández-Domínguez13, Manuel Altamirano García14, Rubén Pardo Martínez15, Mercedes Iriarte Cela16, Javier L. Carrasco Rus12, Carmen Alfaro Giner17, Raquel Piqué Huerta2 Plant material culture can offer unique insights into the ways of life of prehistoric societies; however, its perish- able nature has prevented a thorough understanding of its diverse and complex uses. Sites with exceptional preservation of organic materials provide a unique opportunity for further research. The burial site of Cueva de los Murciélagos in southern Iberia, uncovered during 19th-century mining activities, contained the best-pre- served hunter-gatherer basketry in southern Europe, together with other unique organic artifacts associated with the first farming communities, such as sandals and a wooden hammer. We present 14 14C dates for the perishable artifacts (N = 76), situating the assemblage between the Early and Middle Holocene (c. 7500 to 4200 cal BCE). Our integrated analysis includes raw material determination and technological and chrono-cul- tural contextualization of this unique and important set of materials. INTRODUCTION early Pleistocene (3, 4). The earliest direct evidence derives from the Perishable organic raw materials are widely available and versatile, Middle Pleistocene (5–8), but examples are scarce and widely and have played a crucial role in human history. Wood, plant fibers, dispersed. and skins have been used for the manufacture of multiple artifacts Most current knowledge of past societies is built from analysis of essential to the daily life of past populations. The poor preservation durable materials. Information relating to vegetal crafts in prehisto- of such objects means that they can rarely be studied in detail, even ry is limited (1) because of the lack of recovered remains, but, like though they are key to understanding human adaptations to envi- nonperishable objects, plant-based artifacts allow us to study eth- ronments and past technological and ecological knowledge, and nicity (9–11), frontiers of culture and identity (12–15), trade net- provide insight into the diversity of needs for which these materials working (16, 17), and human-environmental relationships have provided a solution. For this reason, sites with well-preserved including economy, adaptation, and subsistence (18–21). The perishable materials present a unique opportunity to study past so- craft of basketry is considered a particularly useful indicator for de- cieties (1, 2). Plant-based goods and tools were among the first ar- fining technological and cultural traditions (2). tifacts in history, as suggested by the use-wear of lithic tools dated to The degree of preservation of perishable evidence depends on the contexts in which the materials are deposited or preserved. In southern Europe, good preservation in archaeological contexts 1 Departamento de Historia y Filosofía, Área de Prehistoria, Universidad de Alcalá, dated to Mesolithic and Neolithic periods is extremely rare and re- Alcalá de Henares, Spain. 2Departament de Prehistòria, Universitat Autònoma de stricted to a few sites where waterlogging, charring, or desiccation Barcelona, Bellaterra, Spain. 3Departamento de Ciencias Históricas, Universidad de Cantabria, Santander, Spain. 4Departamento de Ciencias Históricas, Universidad de occurs. The sites of La Marmotta (c. 5840 to 5010 cal BCE) in west Las Palmas de Gran Canaria, Las Palmas de Gran Canaria, Spain. 5Centro Oceano- Italy (22) and La Draga (c. 5300 to 4700 cal BCE) in northeast Spain gráfico de Canarias (COC), Instituto Español de Oceanografía (IEO-CSIC), Tenerife, (21, 23) are both lakeshore settlements of Early Neolithic date, and Spain. 6Departamento de Prehistoria, Museo Arqueológico Nacional, Madrid, Spain. 7Independent researcher. 8Departamento de Historia, Universidad de are well known for the exceptional preservation, due to waterlog- Córdoba, Córdoba, Spain. 9Cultures et Environnements Préhistoire, Antiquité, ging, of numerous organic items such as wooden artifacts, Moyen Âge, Université Côte d’Azur, Nice, France. 10Departamento de Geología, cordage, and basketry. In Iberia, plant-based materials from the Me- Universidad de Salamanca, Salamanca, Spain. 11Departamento de Arqueología y solithic-Neolithic are mostly preserved by charring and impression Procesos sociales, CSIC (Consejo Superior de Investigaciones Científicas), Madrid, Spain. 12Departamento de Prehistoria y Arqueología, Universidad de Granada, in fired clay, for example, at the Coves de Santa Maira (c. 11200 to Granada, Spain. 13Archaeology Department, Durham University, Durham, UK. 8200 cal BCE) in Alicante (24) and at Coves del Fem (4941 to 4545 14 UNED (Universidad a Distancia), Córdoba, Spain. 15Departamento de Análisis cal BCE) in Tarragona (25, 26). Material preserved by desiccation is Geográfico Regional y Geografía Física, Universidad de Granada, Granada, Spain. 16 Departamento de Nanotecnología, Centro Tecnologico y de Investigación SDLE, concentrated in southeastern Iberia, which has a predominantly Madrid, Spain. 17Departament de Prehistòria, Arqueologia i Història Antiga, Univer- sub-arid climate. Most of such evidence has been recovered from sitat de València, València, Spain. Chalcolithic and Bronze Age sites, with Neolithic examples restrict- *Corresponding author. Email: f.martinezs@uah.es ed to two textile fragments from Peñacalera Cave (c. 3462 to 3163 cal †These authors contributed equally to this work. Martínez-Sevilla et al., Sci. Adv. 9, eadi3055 (2023) 27 September 2023 1 of 18 S C I E N C E A D VA N C E S | R E S E A R C H A R T I C L E BCE) in Córdoba (27), and the outstanding examples of basketwork blades and flakes of flint, quartz, polished axe head, bone awls, as and wooden objects from Cueva de los Murciélagos in Granada (28, well as diverse ornaments such as perforated shells, wild boar teeth, 29) that are the focus of this article. stone and shell bracelets (34, 35), and a unique gold diadem. The Cueva de los Murciélagos is one of the best-known sites in south- pottery assemblage comprises globular vessels with impressed, ern Europe for its exceptional preservation of organic materials by incised, and almagra (red-painted) decorations (36). All the archae- desiccation. The cave is a burial site identified in the 19th century by ological remains are now deposited in the Museo Arqueológico Na- mining activities that uncovered partially mummified corpses ac- cional (Madrid), Museo Nacional de Antropología (Madrid), and companied by baskets, wooden tools, and other goods. Numerous Museo Arqueológico y Etnográfico (Granada). critiques were published regarding the authenticity of the materials Although the archaeological organic materials from Cueva de los and their chronology (30, 31). Questions persisted until the 1970s, Murciélagos have been cited countless times as the most well-pre- when the first radiocarbon analysis was carried out on a sample of served set of archaeological fiber-based materials in southern vegetal fiber and another from wood, both yielding Neolithic dates Europe, just three publications focus primarily on these materials (5200 to 4850 cal BCE) (32). After the death of Manuel de Góngora (28, 29, 37). Alfaro (28, 29) made the first systematic study of the (1884), the first investigator and owner of the remains, the archae- fiber-based materials from the site regarding the technology of ological materials became part of the first collections of the Museo their production. Her publications reveal the main typologies of Arqueológico Nacional of Madrid. basketry, cordage, and sandals, with a catalog describing each Here, we investigate the enduring transmission of Early-Middle object, adapting the terminology and technical variations. She spec- Holocene plant-based technologies by analyzing the perishable ar- ifies that the entire set of fiber-based objects was made from esparto tifacts of Cueva de los Murciélagos. We present a robust chronolog- grass (Stipa tenacissima) and explains the differences in the appear- ical assessment of the artifacts and other archaeological remains to ance of the raw material as due to the processing of the fiber through contextualize our findings and provide a chrono-cultural sequence crushing the esparto—as is still traditionally done today. In a sub- of the site. Our study also included the first geological characteriza- sequent publication, Cacho et al. (37) detected geometric decora- tion of the cave formation, explaining the preservation of the per- tions on seven of the baskets using spectrophotometry analysis. ishable materials. We analyzed the technological features and the Two sandals, a basket fragment, a piece of wood, and one undeter- raw material of plant-based artifacts, including a unique set of mined object were dated by conventional radiocarbon methods, baskets, sandals, and wooden objects (table S1). Our research re- providing a chronology between 5200 and 4850 cal BCE (37), vealed that the plant-based artifacts of Cueva de los Murciélagos placing the site in the Early Neolithic of the region, and therefore were produced in both the Mesolithic and Neolithic periods, pro- representing an outstanding example of the earliest basketry in viding evidence of the ways in which these technologies evolved Europe. Although another wood sample provided an earlier chro- through time. We also present the earliest and most diverse set of nology (6450 to 6030 cal BCE) (37), this was interpreted as due to plant-based footwear documented in the prehistory of Europe. the “old wood effect” and excluded from the interpretation. The possibility that this earlier date could be related to the use of the Site background and archaeological record cave was previously postulated by some of the authors contributing Cueva de los Murciélagos [Albuñol, Granada, Spain; UTM to the present article (38). Our current study is the first to investigate (ETRS89) 483323.00-4073500.00] is a karstic cave, 7 km from the these materials using an interdisciplinary approach combining geo- Mediterranean coastline of southern Spain and 2 km from the archaeology, radiocarbon dating and Bayesian modeling, raw mate- village of Albuñol (Fig. 1A). The cave was first accessed in 1831 rial identification, and analysis of technological features. by the owner of the surrounding lands who collected the abundant bat guano in the main chamber for fertilizer. The shelter was also used to keep goats until the identification of a vein of galena led to RESULTS its exploitation by a mining company in 1857. The mining activities Site geology and environmental conditions and the removal of blocks to access the mineral led to the discovery Cueva de los Murciélagos is a karstic cave (Fig. 1 and the Supple- of a gallery that housed several partially mummified corpses accom- mentary Materials). The cave has a lenticular entrance, yielding panied by baskets, wooden tools, and other archaeological remains. direct access to the main chamber (Fig. 1C). The entrance is 15 m The activity of the miners resulted in plant-based objects being wide, oriented to the east; the cave is 60 m long and 30 m wide, with burned and scattered around the outside of the cave, while other a drop of about 48 m between the highest and the deepest parts (fig. baskets and objects were distributed among the Albuñol villagers. S1). It is located at the beginning of the lowest section of the Angos- Ten years after the discovery and subsequent looting, the cave was turas gorge, on its right bank, 450 m above sea level and about 70 m visited by the archaeologist Manuel de Góngora y Martínez who from the base of the gorge, on gray limestone that is highly fractured collected the testimonies of the miners regarding the artifacts, gath- by Alpine tectonics (Fig. 1B). Impermeable rocks (chalcoschist and ered the archaeological remains, and published accounts of them phyllite) are found very close to the base of the stratigraphic (33). He describes the remains of 68 individuals distributed in dif- column, so the cave is subaerial with a few small globular speleo- ferent areas. Most of these human remains have yet to be relocated. thems (between 0.5 and 4 cm) (fig. S2). This limited development Góngora’s publication associates the basketry and the wooden and growth of the speleothems is due to periods of low rainfall and/ objects with the burials but without giving an exact location for or high temperature, giving the cave practically zero humidity. each artifact. We know only that they were recovered from the floor surface in the inner part of the cave but have no information The set of plant-based tools about the association of the objects or their original positions. The The perishable artifacts comprise 76 individual objects (table S1) nonperishable materials from the site include ceramic sherds, including 10 wooden items (one of these being a composite Martínez-Sevilla et al., Sci. Adv. 9, eadi3055 (2023) 27 September 2023 2 of 18 S C I E N C E A D VA N C E S | R E S E A R C H A R T I C L E Fig. 1. The site of Cueva de los Murciélagos. (A) Location of Cueva de los Murciélagos in southeast Spain (Andalucía, Albuñol). (B) Elevation profile [east (E) to west (W)] of Angosturas gorge showing the situation of cave. (C) View from the north toward the Angosturas gorge and cave entrance. (D) Plan of the cave made with the 3D model. object involving wood, reed, and vegetal fibers), 1 made of reed, and section of the leaves as protuberances on the abaxial face that 65 fiber-based artifacts. Radiocarbon analyses were performed on a contain the main isolated vascular bundles. Secondary vascular selection of 14 objects (Table 1); raw material and technological bundles are also visible all along the leaves, and these are surround- aspects were studied in detail for each of them. Analysis revealed ed by a thick layer of sclerenchyma tissue. The adaxial epidermis is that a variety of plants were used to provide wood, reeds, and fibers. glabrous, with alternated short cells known as rondels and long The wooden assemblage is heterogeneous, composed of worked sinuate cells, and the absence of stomatal structures (Fig. 2G). wood, fragments of torches, and other wooden remains with no ev- The abaxial epidermis is characterized by a high concentration of idence of crafting (now under study). The present article focuses on silica hairs called tricomes (Fig. 2H). The parenchyma tissue is the two wooden items that were dated by radiocarbon analysis: a visible in Fig. 2I. pointed object (479) and a mallet (475). The pointed object (479) It was possible to distinguish differences in the processing of the was fashioned from the branch of strawberry tree (Arbutus fibers used in some objects with the naked eye, as previously indi- unedo) (Fig. 2, A to C), a shrub that produces a dense, hard cated by Alfaro (29). While physical treatment, such as crushing, is wood. The mallet (475) was produced by longitudinally splitting easy to identify by observation or by histological analysis, other an olive tree (Olea europaea) (Fig. 2, D to F) trunk, selected to types of processing such as retting and fermentation are less have a lateral branch present at the required angle (Supplementary visible. Our methodology was able to differentiate between raw, un- Materials). treated esparto and crushed esparto that had been physically pro- Systematic analysis of the fibers used as raw material confirmed cessed. Of the whole set of materials studied, 50.77% (33 of 65) of the use of esparto grass (S. tenacissima) for producing all types of the objects indicate use of crushed esparto leaves, 41.54% (27 of 65) baskets and cordage. Microscopic analysis showed microanatomical show raw esparto, and a further 7.69% (5 of 65) contain a characteristics present on modern samples of esparto. These were combination. absent on other Poaceae family, esparto-type plants (Stipa gigantea The treatment of the fibers differs depending on the crafting and Lygeum spartum) present near the site and used for compari- techniques represented across the assemblage. For basketry, the son. Esparto leaves are naturally flat in shape, but in arid environ- esparto was either raw or crushed. Raw esparto was used for ments, they convolute to minimize their size and limit water objects made from “twined” (both simple and diagonal) and transpiration (39). This convolution is visible in the transverse “braided” basketry. Items fashioned using the “pseudobraided”/ Martínez-Sevilla et al., Sci. Adv. 9, eadi3055 (2023) 27 September 2023 3 of 18 Table 1. Results of radiocarbon dating from organic-based material from Cueva de Los Murciélagos. Laboratory Object Object type Material Species Sample Analyzed Method Pretreatment Date BP δ13C Reference code code weight (mg) o/oo (mg) Beta-628427 594 Pseudobraided/ Vegetal S. 64.4 2.8 AMS-standard Acid/alkali/acid: cellulose 6150 ± −21.0 This study cofín basket fibers tenacissima extraction: solvent 30 BP extraction Beta-628426 626 Linked rings Vegetal S. 22.5 3.0 AMS-standard Acid/alkali/acid: cellulose 8400 ± −21.4 This study fibers tenacissima extraction: solvent 30 BP extraction Beta-627342 617 Diagonal Vegetal S. 45.1 3.0 AMS-standard Acid/alkali/acid 6210 ± −21.1 This study twined basket fibers tenacissima 30 BP Beta-627341 615 Coiled basket Vegetal S. 16.5 3.0 AMS-standard Acid/alkali/acid 5580 ± −20.0 This study Martínez-Sevilla et al., Sci. Adv. 9, eadi3055 (2023) fibers tenacissima 30 BP Beta-627340 479 Pointed stick Wood A. unedo 1.8 0.59 AMS-micro- Acid/alkali/acid 6170 ± −22.4 This study sample 30 BP analysis Beta-627338 475 Mallet Wood O. europaea 8.0 1.7 AMS-standard Acid/alkali/acid 5660 ± −21.9 This study 30 BP 27 September 2023 Beta-627337 623 Pseudobraided/ Vegetal S. 19.7 2.8 AMS-standard Acid/alkali/acid 5570 ± −22.8 This study cofín basket fibers tenacissima 30 BP S C I E N C E A D VA N C E S | R E S E A R C H A R T I C L E Beta-627336 624 Pseudobraided/ Vegetal S. 7.1 2.4 AMS-standard Acid/alkali/acid 5550 ± −21.2 This study cofín basket fibers tenacissima 30 BP Beta-627335 625 Braided basket Vegetal S. 18.5 3.4 AMS-standard Acid/alkali/acid 5550 ± −20.3 This study fibers tenacissima 30 BP Beta-627334 581 Simple Vegetal S. 46.6 2.5 AMS-standard Acid/alkali/acid 8300 ± −21.8 This study twined basket fibers tenacissima 30 BP Beta-627333 580 Simple Vegetal S. 73.0 3.3 AMS-standard Acid/alkali/acid 8320 ± −19.6 This study twined basket fibers tenacissima 30 BP Beta-627332 579 Simple Vegetal S. 48.5 2.8 AMS-standard Acid/alkali/acid 8350 ± −20.4 This study twined basket fibers tenacissima 30 BP Beta-627331 611 Simple Vegetal S. 14.4 2.7 AMS-standard Acid/alkali/acid 5640 ± −23.7 This study twined basket fibers tenacissima 30 BP Beta-627330 603 Central core sandal Vegetal S. 66.3 3.2 AMS-standard Acid/alkali/acid 5630 ± −23.5 This study fibers tenacissima 30 BP CSIC-1132 616 Coiled basket Vegetal S. 5861 ± (37) fibers tenacissima 48 BP CSIC-1134 609 Central core sandal Vegetal S. 5900 ± (37) fibers tenacissima 38 BP CSIC-1133 598 Central core sandal Vegetal S. 6086 ± (37) fibers tenacissima 45 BP CSIC-246 – – Vegetal – 5400 ± (32) fibers 80 BP CSIC-247 – – Wood – 7440 ± (32) 100 BP 4 of 18 S C I E N C E A D VA N C E S | R E S E A R C H A R T I C L E Fig. 2. Raw material scanning electron microscopy images. (A to C) 479 A. unedo wood (cross and radial sections). (D to F) 475 O. europaea wood (cross, tangential, and radial sections). (G) 582 S. tenacissima adaxial epidermal view. (H and I) 580 S. tenacissima abaxial epidermal view and parenchymatic tissue. “cofín” technique were also produced primarily with raw esparto, comprising smashed fibers and fractures are clearly visible on except for a single object (619) made with crushed esparto. Raw both head edges, and polished surfaces are present on the handle. esparto was also used for the bundles found in “coiled” basketry, Two different basketry techniques are present: two-dimensional while the stitches show opened fibers, indicating that the leaves (2D), flat objects like mats, and 3D “baskets” (although baskets can had been crushed. For the cords, both twisted and braided, the be made with the same variety of techniques as mats). A standard- fibers are crushed, and both types of sandals, “simple” and ized nomenclature of basketry production is still needed, although it “central core,” are also made with crushed esparto. Selection of has been discussed by many authors. Terminology is also affected by leaves by size is also apparent. Analysis of the width of raw linguistic issues (40). Alfaro (29) identifies several basketry tech- esparto leaves (it was not possible to measure crushed leaves) re- niques in the Cueva de los Murciélagos assemblage: simple (N = vealed that those used for coiled basketry (1.263 to 1.456 mm) 10) or diagonal twining (N = 7), coiling (N = 11), pseudobraided/ and for a braided bag (1.513 mm) were wider than the leaves used cofín basketry (N = 6), and a particular type of basket constructed in twined basketry (0.801 to 1.162 mm). For the pseudobraided/ from knots on the base and called braided basketry (N = 1). Partic- cofín technique, the width varied (0.745 to 1.258 mm). ular techniques are associated with the production of specific basket Both of the wooden objects showed use-wear marks, ending shapes and sizes. A detailed technological description of the dated their life cycle within the cave, discarded or intentionally deposited. objects is available in the Supplementary Materials. The pointed object preserves the diameter of the branch from which Cords are also represented: a single piece (591b) of twisted cord the bark and the lateral branches were removed. Only the edge of the with an S2z pattern, and a few three-strand braids (N = 5). Most of tool is shaped to obtain a point. Smashed fibers and polished sur- the cords seem to be part of other objects such as baskets, but faces on the pointed edge are identified as use-wear, but no frac- mainly sandals, which have been fragmented by postdepositional tures, notches, or erosions—indicative of use as a digging stick— processes. are clearly visible to the naked eye. This object is therefore classified Sandals form an important part of the inventory. Alfaro (29) dis- only as a “pointed stick.” The head and handle of the mallet is tinguished two types: simple (N = 2) and central core (N = 20). Al- roughly shaped, likely with cutting tools. Use-wear marks though no evidence of “laces” is conserved for the simple type, for Martínez-Sevilla et al., Sci. Adv. 9, eadi3055 (2023) 27 September 2023 5 of 18 S C I E N C E A D VA N C E S | R E S E A R C H A R T I C L E the central core type, a small group of fibers emanating from the ending around 7480 to 7150 cal BCE (95% HPD). After a hiatus, base of the sole may have been placed between the first and there is a second phase of intensity in around 5380 to 5070 cal second toes. These fibers are also connected to a braid fixed to BCE (95% HPD), lasting until 4390 to 4050 cal BCE (95% HPD). the middle of the sandal, which could be tied around the ankle. A An intermediate peak is observed between these two phases that single “ring” (626) and a nondetermined object comprising knotted proceeds from the unidentified desiccated wood sample CSIC-247 fibers were not included in our analysis. [7440 ± 100 before present (BP)] (table S2 and fig. S3). This radio- carbon measurement is likely affected by inbuilt age issues and the Chronology and Bayesian modeling of plant-based tools wide-ranging standard deviation. We also carried out a model with The kernel density estimation (KDE) output from 19 radiocarbon KDE_plots of the radiocarbon dates of the two main phases. The dates (14 determined as part of the current research) indicates that estimates are analogous to those observed in the first KDE model the human deposition of plant-based artifacts in the site began (table S2 and Fig. 3A). Our results therefore suggest that the around 7950 to 7360 cal BCE [95% highest probability density objects from Cueva de los Murciélagos were likely placed during (HPD)] and ended in approximately 4370 to 3740 cal BCE (95% two main chronological phases, separated by around 2000 years. HPD) (table S2 and Fig. 3). The KDE model also reveals a first The Bayesian single model (41) included all the radiocarbon phase starting around 7960 to 7680 cal BCE (95% HPD) and dates from the site, which yielded a good agreement (Amodel = 98.5; Aoverall = 98.4). The earliest evidence of human deposition of artifacts started in 7986 to 7391 cal BCE (95% HPD) and ended in 4373 to 3740 cal BCE (95% HPD) (fig. S3 and table S3). The modeled chronology indicates a period of human use of around 3110 to 4040 years (95% HPD) (table S3). These results are analo- gous to those obtained from KDE analyses. The results from the Outlier Charcoal model (see Materials and Methods) were also similar to those from the Bayesian single model (table S4 and fig. S4). A Bayesian model of two sequential phases was created to con- sider the two main clusters of radiocarbon dates but excluding sample CSIC-247 (7440 ± 100 BP) (Fig. 3 and Table 2). The first phase comprises the four oldest radiocarbon dates, while the second phase contains the remaining measurements. The radiocar- bon dates of phase 1 pass the test of contemporaneity {T0 = 6.3 [T (5% = 7.8)]} (42), which means that the artifacts were probably de- posited during a short period of time. The result of the model indi- cates a good agreement (Amodel = 96.6; Aoverall = 97.3) (Fig. 3 and Table 2). All the radiocarbon dates also show good agreement (A = >73%). Phase 1 starts in around 7680 to 7360 cal BCE (95% HPD) and ends around 7480 to 7100 (95% HPD). This phase was relatively short at 0 to 530 years (95% HPD) (Table 2). The chronological gap between the end of phase 1 and the start of phase 2 was estimated at 2340 to 1780 years using OxCal’s Dif- ference command (95% HPD). Our results indicate a significant hiatus between phases 1 and 2 of around 2000 years (Table 2). Phase 2 begins in 5420 to 5070 cal BCE (95% HPD) and ends in 4390 to 4070 cal BCE (95% HPD). The analysis also indicates that the artifacts related to this phase were deposited in the cave over a long period {T0 = 897.189 [(5% 22.4)]}, estimated at 740 to 1260 years (95% HPD). The results of the Charcoal Outlier model were also similar to those of the Bayesian two-phase model. We also tested a Bayesian model of three sequential phases including, as an additional phase, the desiccated sample CSIC-247 (7440 ± 100 BP), yielding high agreement indices (>95) in both the multiphase and outlier models (figs. S5 and S6 and tables S4 and S5). However, this model is not discussed here as the wide boundaries of this phase overlap the two main phases previously analyzed. Fig. 3. Probability distribution of dates from Cueva de los Murciélagos. (A) The attribution of the directly radiocarbon-dated objects to the KDE plots and Bayesian chronological ranges of the overall distribution of the dated events of Cueva de los Murciélagos and within each phase (phase 1, Meso- identified phases on the site is as follows. The oldest artifacts, cor- lithic; phase 2, Neolithic). (B) Multiphase Bayesian chronological ranges for the es- responding to phase 1 of the use of the cave, are 597 (Fig. 4A), 580 timated start and end of each phase and the modeled ranges of each radiocarbon (Fig. 4B), 581 (Fig. 4C), and 626 (Fig. 4D) together with the piece of date (except outlier). OxCal v4.4.4 Bronk Ramsey (69); r:5 Atmospheric data from wood previously dated in the 1970s, the reference for which is Reimer et al. (68). unknown. The objects corresponding to phase 2 are 617 Martínez-Sevilla et al., Sci. Adv. 9, eadi3055 (2023) 27 September 2023 6 of 18 S C I E N C E A D VA N C E S | R E S E A R C H A R T I C L E Table 2. Modeled dates and ranges for the estimated start, interval, and end for each phase of Cueva de Los Murciélagos. The estimated difference between phases 1 and 3 is also indicated. Name Unmodeled Modeled Indices (BCE/AD) (BCE/AD) Amodel 96.3 Aoverall 97 68.3% 95.4% 68.3% 95.4% From To From To From To From To Acomb A L P C Sequence Cueva de los Murciélagos Boundary start 1 −7540 −7387 −7673 −7359 98 Phase 1 R_Date −7531 −7384 −7576 −7357 −7504 −7370 −7531 −7353 73.3 99.8 Beta-628426 R_Date −7491 −7358 −7516 −7336 −7481 −7361 −7501 −7346 107.5 99.8 Beta-627332 R_Date −7469 −7343 −7507 −7196 −7469 −7361 −7487 −7332 109.2 99.8 Beta-627333 R_Date −7463 −7206 −7480 −7192 −7464 −7356 −7486 −7317 108 99.8 Beta-627334 Interval 1 0 185 0 522 99.1 Boundary end 1 −7425 −7294 −7478 −7094 99.5 Boundary start 2 −5271 −5120 −5422 −5074 99.7 Phase 2 R_Date −5215 −5072 −5297 −5050 −5156 −5067 −5281 −5041 104.5 99.9 Beta-627342 R_Date −5208 −5054 −5214 −5015 −5173 −5046 −5210 −5013 100.1 99.9 Beta-627340 R_Date −5208 −5033 −5209 −5005 −5187 −5013 −5206 −5001 98.2 99.9 Beta-628427 R_Date −5201 −4936 −5208 −4847 −5056 −4910 −5206 −4846 101.7 99.9 CSIC-1133 R_Date −4827 −4720 −4886 −4690 −4826 −4720 −4886 −4691 99.8 99.9 CSIC-1134 R_Date −4795 −4681 −4843 −4555 −4795 −4680 −4844 −4556 99.9 99.8 CSIC-1132 R_Date −4536 −4453 −4581 −4369 −4535 −4453 −4552 −4370 99.8 99.9 Beta-627338 R_Date −4533 −4406 −4542 −4367 −4533 −4406 −4542 −4367 99.6 99.9 Beta-627331 R_Date −4498 −4371 −4537 −4365 −4499 −4371 −4537 −4366 99.2 99.9 Beta-627330 R_Date −4446 −4364 −4486 −4350 −4446 −4365 −4484 −4350 99.8 99.9 Beta-627341 R_Date −4443 −4360 −4454 −4348 −4443 −4361 −4454 −4348 99.8 99.9 Beta-627337 R_Date −4443 −4350 −4447 −4346 −4443 −4351 −4447 −4346 99.4 99.9 Beta-627336 R_Date −4443 −4350 −4447 −4346 −4443 −4351 −4447 −4346 99.3 99.9 Beta-627335 R_Date −4342 −4068 −4440 −3996 −4438 −4262 −4453 −4173 93.6 99.8 CSIC-246 Interval 2 817 1046 744 1258 99.6 Boundary end 2 −4352 −4219 −4391 −4071 99.3 continued on next page Martínez-Sevilla et al., Sci. Adv. 9, eadi3055 (2023) 27 September 2023 7 of 18 S C I E N C E A D VA N C E S | R E S E A R C H A R T I C L E Name Unmodeled Modeled Indices (BCE/AD) (BCE/AD) Amodel 96.3 Aoverall 97 Difference phase −2257 −2034 −2340 −1783 99.4 1 to phase 2 Fig. 4. Mesolithic organic-based artifacts. (A) Basket 579 (Beta-627332: 8350 ± 30 BP). (B) Basket 580 (Beta-627333: 8320 ± 30 BP). (C) Basket 581 (Beta-627334: 8300 ± 30 BP). (D) 626 Linked rings (Beta-628426: 8400 ± 30 BP). (Fig. 5A), 594 (Fig. 5B), 479 (Fig. 5C), and, in chronological order, DISCUSSION 475 (Fig. 6A), 611a (Fig. 6B), and 603 (Fig. 6C). The dates obtained The unique conditions for the preservation of organic material in in the 1990s correspond to 616a (Fig. 7A), 598 (Fig. 7B), and 609 the cave are related to the null humidity resulting from the geolog- (Fig. 7C). The last are 623 (Fig. 8A), 625 (Fig. 8B), 615a (Fig. 8C), ical character of the cave. Moreover, a dry wind current is generated and 624 (Fig. 8D). The most recent radiocarbon date obtained is by the prevailing climate in the area, and the north-south direction that for a nonidentified object dated in the 1970s. A detailed de- and narrow and deep morphology of the Angosturas gorge channel scription of each of the dated items is available in the Supplemen- the wind toward the cave, through the narrow upper entrance. The tary Materials. wind cools as it travels through the cave, increasing in speed; it is cold as it exits through another narrow entrance located in the lower part of the shelter. The lack of prevailing humidity in the Martínez-Sevilla et al., Sci. Adv. 9, eadi3055 (2023) 27 September 2023 8 of 18 S C I E N C E A D VA N C E S | R E S E A R C H A R T I C L E Fig. 5. Neolithic organic-based artifacts. (A) Basket fragment 617 (Beta-627342: 6210 ± 30 BP). (B) Basket 594 (Beta-628427: 6150 ± 30 BP). (C) Digging stick 479 (Beta- 627340: 6170 ± 30 BP). area and the circulation of wind in the cave as it cools and dries based objects, both occurring between the Early and Middle Holo- prevent the proliferation of bacteria, increasing the amount and di- cene (c. 7500 to 4200 cal BCE), but each linked to periods having versity of preserved perishable material at the site. Unfortunately, very different economic and social systems: The first phase of dep- the original contexts of the artifacts are unknown, so establishing osition is related to early Holocene hunter-gatherer populations, associations is difficult. Even so, the characteristics of the fragments, and the second phase is associated with Middle Holocene in terms of the techniques used to make them, their measurements, farmers. Other nonperishable materials from the site, like bone and the raw materials and fiber processing used, suggest a refitting tools, lithics, and pottery sherds, confirm that the cave was used reg- hypothesis for some of the pieces (table S1). According to Góngora ularly by Early and Middle Neolithic populations, but beyond the (33), the materials suffered substantial fragmentation or charring organic-based objects, no other durable material can now be when the lead workers burned them to keep their boiler working. related to the Mesolithic use (phase 1) of the site. Our results also In summary, despite the mining activity, this assemblage repre- highlight that direct radiocarbon dates from plant-based objects are sents one of the oldest and best-preserved collections of hunter- essential for understanding the depositional sequence of perishable gatherer basketry in southern Europe. The sandals, baskets, and materials in nonstratified archaeological contexts such as caves (43). wooden artifacts with Neolithic chronologies constitute a unique This information also provides a chronological framework within sample of organic artifacts absent in other archaeological sites of which to integrate these objects, allowing more general archaeolog- early farmer communities. Previous studies assumed that all ical discussions on the significance of plant-based technologies in organic materials from the cave were Neolithic, based on three con- human evolution. Furthermore, this contribution indicates differ- ventional radiocarbon dates (37). Our study, however, offers a fine- ences between raw materials, techniques, and object typologies grained chronological framework based on accelerator mass spec- for each depositional phase in Cueva de los Murciélagos. trometry (AMS) radiocarbon dating and Bayesian modeling that clearly indicates two distinct main phases of deposition of plant- Martínez-Sevilla et al., Sci. Adv. 9, eadi3055 (2023) 27 September 2023 9 of 18 S C I E N C E A D VA N C E S | R E S E A R C H A R T I C L E Fig. 6. Neolithic organic-based artifacts. (A) Mallet 475 (Beta-627338: 5660 ± 30 BP). (B) Sandal 603 (Beta-627330: 5630 ± 30 BP). (C) Sandal 611 (Beta-627331: 5640 ± 30 BP). The Mesolithic fiber-based objects dated (593). Our study therefore offers direct evidence of the The three dated objects (579, 580, and 581) corresponding to the oldest fiber-based objects made by hunter-gatherer populations in earlier occupation phase of the cave (phase 1) are 3D twined Iberia and in Europe. Elsewhere, fiber artifacts are associated with baskets made with raw esparto. They are well preserved compared hunter-gatherer societies at various archaeological sites in the Great with the other fiber-based materials, being almost complete and Basin (United States), where a notable number of objects made decorated with geometric motifs made with dyed fibers (Fig. 2A); using a wide range of basketry techniques have been recorded, in- some incorporate human hairs or pigments (Fig. 2B). Other cluding plaited, twined, and coiled basketry (13, 14, 44). baskets recovered at the cave (582, 583, 585, 586, 586a, 586b, 592, Baskets 584 and 593 present variations of the twining technique, and 593) share similar morphologies, techniques, and raw material, specifically regarding the warp, which gives a twill weave (584) and a suggesting that they are also related to phase 1. diagonal twining (593) appearance to the baskets. The attribution of Simple twining is the only basketry technique clearly represented these examples to the Mesolithic phase is, however, unclear, but in the Mesolithic phase. It is not present in later periods and may some of their characteristics, such as raw material measurements thus be the oldest method represented in Cueva de los Murciélagos. and preparation, are similar to those of the Mesolithic baskets. The same technique has also been identified in basketry imprints on Only radiocarbon dating will confirm this hypothesis. clay fragments from the Coves de Santa Maira site (Alacant, Spain, Four linked rings (626), whose function is unknown, also have 12,900 to 10,200 cal BCE). Here, however, the imprints show space hunter-gatherer chronologies. Góngora (33) suggested that the between the weft elements (open twining). In the dated baskets from rings were part of an ornamental necklace, although no further in- Cueva de los Murciélagos, there is no space (closed twining), al- formation is available. Five rings were preserved in the 19th century, though there is one example of open twining that has not been but only four now survive. No other objects or techniques are Martínez-Sevilla et al., Sci. Adv. 9, eadi3055 (2023) 27 September 2023 10 of 18 S C I E N C E A D VA N C E S | R E S E A R C H A R T I C L E Fig. 7. Neolithic organic-based artifacts dated in previous work with detail of damage. The white dashed line shows the portion of the object now lost. (A) Basket fragment 616a (CSIC-1132: 5861 ± 48 BP). (B) Sandal 598 (CSIC-1133: 6086 ± 45 BP). (C) Sandal 609 (CSIC-1134: 5900 ± 38 BP) (32, 37). represented in phase 1. Only raw esparto was used for the twined made of crushed esparto. The third item, 624, is a 2D object that Mesolithic baskets and the linked rings. No evidence of physical exhibits both closed and open pseudobraiding in different areas. processing of the fibers was identified in the earlier chronologies Object 617 is the only example of diagonal twining documented of the cave. Although the twining technique and its variations can in this phase. Other undated objects (622, 627, 627a, and 627b) be achieved with crushed esparto, processing was not used during present similar techniques and raw material, which suggest that this phase. they may also be attributed to this phase. These are all fragmented, and no specific object shape can be identified. The esparto used in The Neolithic perishable objects these pieces is raw, as used for the simple twining baskets in phase 1. Phase 2 of Cueva de los Murciélagos encompasses a wider diversity This group of plant-based objects provides the only evidence of di- of objects, techniques, and raw material processing. The plant-based agonal twining, pseudobraided/cofín, and braided baskets in the artifacts include cords, 2D and 3D baskets, and sandals made with prehistory of the Iberian Peninsula and, indeed, that of Europe. several techniques. The modeled dates of three pieces (594, 623, and No earlier parallels are documented, indicating that these tech- 624) place the pseudobraided or cofín technique in this phase. niques may originate exclusively from south Iberia. These items were selected for radiocarbon analysis because of dif- The unique braided basket or bag (625) also corresponds to the ferences in their morphology and the treatment of raw material, but Neolithic phase. In this object, the esparto leaves are, again, used in this technique is also present in three other objects (618, 619, and their raw state. The coiled basketry technique is represented in this 621). Object 594 is an almost complete bag. It shows space between phase by two artifacts: one (616a) dated in the 20th century (37), the pseudobraids, as is also observed in the second object, 623, a and one (615) dated within the present study. With regard to the formless fragment. Fragment 623 is the only cofín object, which is processing of the esparto, it is important to differentiate between Martínez-Sevilla et al., Sci. Adv. 9, eadi3055 (2023) 27 September 2023 11 of 18 S C I E N C E A D VA N C E S | R E S E A R C H A R T I C L E Fig. 8. Neolithic organic-based artifacts. (A) Basket/mat fragment 623 (Beta-627337: 5570 ± 30 BP). (B) Basket 625 (Beta-627335: 5550 ± 30 BP). (C) Basket fragment 615 (Beta-627341: 5580 ± 30 BP). (D) Basket/mat fragment 624 (Beta-627336: 5550 ± 30 BP). the active and passive elements of the coiled baskets. The stitches seems to be the most widely used in southern Europe during the (active) are always made using crushed esparto (except for object Neolithic. 620), while bundles (passive) can comprise either raw or crushed Twenty-two artifacts corresponding to two typologies of fiber- esparto, although the degree of crushing differs across the objects. based sandals were found in the cave. Our radiocarbon analysis of This technique is represented in nine further pieces (587, 588, 589, a simple sandal (611) and a central core sandal (603) agrees with the 590, 616b, 620, 627c, 627d, and 627e). Contemporaneous examples previous analysis of two central core shoes (598 and 609) (37), con- have been documented in western Mediterranean sites, including La firming that they date to the Neolithic period. No chronological dif- Marmotta (Lazio, Roma, c. 5840 to 5010 BCE) (22), La Draga ferences are visible between the two forms, although the central core (Girona, 5207 to 4862 cal BCE), and Coves del Fem (Tarragona; type is most common. Other examples of shoes have been recovered 6065 to 4545 cal BCE) (25, 26). In La Draga, eight coiled baskets in Europe, including one from a pit in the Areni-1 Cave (Armenia, were identified, made of fibers from various monocot families 3627 to 3377 cal BCE). This was made of leather and a large number (Poaceae, Cyperaceae, and Typhaceae) and lime tree (Tilia sp.). A of grasses (45). Sandals made with lime bast (Tilia sp.) and ramie coiled basket made of Cyperaceae fibers, found covering a pit and (Boehmeria nivea) were recovered at the Allensbach site related to storage purposes, was found at Coves del Fem (Tarragona; (Germany, c. 3000 cal BCE), being very similar to examples found 6065 to 4545 cal BCE) (25, 26). The coiling technique therefore in Sipplingen (Germany; 2900 cal BCE) (46). The footwear associ- ated with the Ötzi (Italy, 3350 cal BCE) present a sock-like structure Martínez-Sevilla et al., Sci. Adv. 9, eadi3055 (2023) 27 September 2023 12 of 18 S C I E N C E A D VA N C E S | R E S E A R C H A R T I C L E made of lime/linden bast and covered with leather (47, 48), and fibers are extracted by combing with a comb to produce similar to a shoe found at the Sutz-Rütte site (Switzerland; 2750 “combed” esparto. cal BCE (49). Given the possible variations in esparto processing, although the The various sandals mentioned above are completely different in appearance of the leaves in phase 1 (Mesolithic) indicates that they technological terms from those found at Cueva de los Murciélagos are not crushed, this does not mean that they did not undergo prep- in that, although they used grasses in their structure, other materials aration. Although no microscopic differences are detectable to in- including leather, lime, and ramie bast were also used. Further dicate whether the esparto is retted/cooked, rehydration of the studies should be undertaken regarding this aspect of the fiber- fibers would be needed to make them malleable and useful for craft- based collection, which potentially contains both direct and indirect ing. None of the objects in the cave could have been produced using information about the population that wore the sandals. This would dried fibers, so raw material would have been prepared by hydration be of interest given the loss of the human remains extracted from the before the items were fashioned. Further, the esparto leaves used in cave in the 19th century. The esparto used in sandals is always the Mesolithic objects present a smaller diameter (0.801 to 1.162 crushed, and this may be directly related to their function as they mm) than those used in the Neolithic phase (1.263 to 1.513 mm), needed to be flexible and comfortable to wear. This sandal set there- except for the material used in some specific pseudobraided arti- fore represents the earliest and widest-ranging assemblage of prehis- facts. This could indicate the selection of younger esparto leaves toric footwear, both in the Iberian Peninsula and in Europe, for Mesolithic crafting. It is important to highlight that the physical unparalleled at other latitudes. treatment of crushing the leaves is evidenced only in Neolithic ma- Cord elements are also recorded in the cave, and these are either terials, suggesting that this practice was not used in earlier periods. twisted or braided. Most seem to be detached fragments from basket A degree of humidity is still needed when using crushed esparto, but handles or cords used to tie sandals to the feet. The Neolithic site of to a lesser extent. Histological differences in the cross section of the La Draga has yielded examples with an extended S and Z twist, and leaves are visible in some of the samples. These may be related to the braided cords made of lime tree and nettle (Urtica sp.) (21), and a part of the leaf where the sample originated (apical-basal)—which is small twisted strand made of an undetermined monocot family was difficult to determine during sampling from archaeological objects also recovered from Coves del Fem (26). One twisted cord example —or to its growth stage. Our results therefore indicate an extended from Cueva de los Murciélagos (591b) is especially striking. The knowledge of the plant resources within the local environment and vegetal material used to produce this cord is heavily crushed. No a high level of understanding and expertise among the last hunter- specific association can be made with braids, although these are gatherer populations and, most likely, their continuity toward the also made with crushed esparto. The twist (S2z), measurements, first farming societies. and processing of the fibers are similar to those of the bowstring The wooden remains studied are also unusual within the archae- associated with Ötzi (50) and with a possible bowstring roll from ological record in southern Europe. The raw material, size, and La Draga (21). The similarity of the twisted cord to these examples morphology of the pointed object (479) are similar to those of Neo- suggests that it could also be a bowstring; several arrow shafts recov- lithic digging sticks, such as the set from La Draga (53); however, ered from the same context in the cave confirm that archery was some technical features are absent, including the fire treatment of practiced. the point. In the case of the wooden mallet (475), a parallel with the same morphology but different size has been identified in a Technological features Neolithic context at Meare Heath (Somerset, UK) (54). The same The esparto grass shows evidence of several processing methods, approach, using wood from the trunk and a branch as a handle, which produce distinctive working characteristics and which influ- has been identified at several other Neolithic sites, where it has ence the typology and the aesthetics of the final object. Esparto ar- been used for crafting adze or axe handles and spoons (55, 56). It tisans traditionally collect the leaves in the summer months. They has been suggested that the mallet could have been involved in the are then dried for 20 to 30 days (51, 52) to produce “raw” esparto or processing of vegetal fibers. It is notable that this artifact is contem- en rama, which can be used for crafting after rehydrating by water poraneous with the evidence for the use of crushed esparto for craft- submersion for approximately 24 hours. The nomenclature en rama ing purposes in the Neolithic phase of Cueva de los Murciélagos. refers to the fact that the leaves are complete and have not suffered physical damage. The dried, raw esparto can be further processed Typologies and function of the plant-based artifacts for specific applications by retting in stagnant water for 20 to 40 This assemblage presents an unprecedented diversity of basket ty- days to produce an anaerobic fermentation resulting in the decay pologies and techniques compared to other Mesolithic and Neolith- of organic materials such as lignin. The organic degradation is ic sites in the Mediterranean region. This high variability may be halted by drying the material. The timing of this process is depen- related to the unusual environmental conditions of the cave and dent on environmental conditions such as temperature and humid- its impact on the conservation of the plant-based artifacts. Other ity. There are no standardized parameters because it is a traditional Mediterranean Early Neolithic sites preserve these artifacts under process that varies depending on the artisan and their experience. waterlogged conditions, as at La Marmotta (22) and La Draga The material resulting from this processing is known as “cooked” or (57), or by carbonization as at the Coves del Fem (26). The desiccat- “retted” esparto and is traditionally acknowledged to be more resis- ing conditions of Cueva de los Murciélagos likely favored greater tant and flexible than raw esparto. The retted esparto can be rehy- preservation of organic materials, yielding a wider variability of pre- drated in the same way as the raw esparto (to produce cooked en served artifacts. The nature of the human activities in the cave rama esparto) or can be crushed physically using a wooden mallet during both Mesolithic and Neolithic periods may also explain to break the leaves and extract the fibers, creating “crushed” esparto. this distinctive assemblage. In contrast to the preserved organic In later periods—and linked to industrial processes—the thinnest objects from La Marmotta, La Draga, and Coves del Fem, which Martínez-Sevilla et al., Sci. Adv. 9, eadi3055 (2023) 27 September 2023 13 of 18 S C I E N C E A D VA N C E S | R E S E A R C H A R T I C L E were recovered from domestic contexts (so basket fragments prob- MATERIALS AND METHODS ably correspond to discarded elements), we know from 19th- Cave geomorphology and 3D documentation century accounts that the objects found in Cueva de los Murciélagos To study the geomorphology of the cave, in situ data on the various were associated with burials (33). Unfortunately, only a few of these speleothems present were recorded. The morphological character- human remains have been preserved in the museum and they have istics and the dimensions of these formations, which are created by not yet been dated. We cannot, therefore, determine whether all the chemical precipitation (dripstone, existence or not of flowstone, plant-based objects were related to burial practices as there is no complex shapes, etc.), were described. A complete topographical contextual information for the finds; they may have been the model of the entire cave was constructed using two different laser result of single or multiple events of deposition over time. scanners: a portable NavVis VLX manual and a fixed portable The most prevalent functional category of the plant-based FARO Laser Scanner Focus3D. A virtual 3D visit to the cave is avail- objects is the “container.” Our results highlight the diversity of able at: Cueva_Murciélagos - FARO WebShare items in this category in terms of shape, size, raw material, and pres- (websharecloud.com). ervation. They include objects made of carved wood and 3D basketry, both rigid and flexible. Bayesian modeling of the radiocar- Raw material identification and technological analysis bon dates indicates that the oldest basket elements are the semi- An inventory of the whole set of perishable objects was produced, rigid, simple-twined baskets, with a very closed weft, and others many of which were also included in previous studies (28, 29). The with similar shapes but different sizes. Some contain rare materials objects were described during observation in the museum, and the such as hair and mineral pigment, and this, together with their ex- documentation was enhanced with an extensive photographic traordinary state of preservation—mainly complete—suggests that record, created using a Nikon D5000 camera with a NIKKOR AF- they were introduced to the cave for a unique purpose, which is con- S DX 18-55 mm VR lens. More detailed images were captured using sistent with their use in funerary practices during the Mesolithic a portable digital microscope (Dino-Lite Edge Digital Microscope (phase 1). AM7915MZT) with a magnification of ×20 to ×220. The volume Our results suggest differences in the functions of Mesolithic and of the baskets studied in this work has been calculated with the Neolithic plant-based objects. The Neolithic containers are more free software Blender 3.5.1 using the revolution 3D model of diverse than those from the Mesolithic in terms of shape, size, the baskets. raw material preparation, and state of preservation. The Neolithic Wooden objects were taxonomically identified following stan- basketry objects are largely fragmented, preventing their identifica- dard methods by observing the three anatomical sections of tion as either 2D or 3D artifacts. The flexibility and diversity in wood: cross, tangential, and radial (58). Each object was observed shape and size also suggest a greater range of functions compared under the portable digital microscope. Where possible, a tiny with the Mesolithic basketry. Flexibility is a valuable attribute for wood sample was obtained from cracks or broken areas for transport, and the more open or closed the weft and warp, and further analysis. These samples were observed under a scanning the presence of active and passive elements (depending on the tech- electron microscope (ZEISS EVO LS 15, RIAIDT-USC). Diagnostic nique) can be related to the type of material to be contained (2). 2D features were compared with atlases of wood anatomy (59–63). basketry often corresponds to mats. The description of the finds In tandem with taxonomic identification, dendrological features provided by Góngora (33) indicates that the individuals buried in from the wooden objects, such as plant part (trunk, twig, root) and this cave wore textiles made of esparto grass, as well as hats and tree-ring curvature, were also registered using qualitative categories sandals. The only remains in both museum collections that can (strong, moderate, weak) (64). Other features related to taphonomic be identified as clothing are the sandals, all the dated examples aspects were also documented, such as evidence of biodeterioration having Neolithic chronologies. Some sandals had clear use marks, (xylophagous galleries, fungal hyphae) (65, 66). In addition to while others were apparently never used, suggesting that while some aspects related directly to their raw material, objects were described individuals were buried with their daily clothing, others had specific in morphological terms, and technical aspects of their crafting were clothing prepared. recorded, along with other features related to their life cycle (use, In summary, the technology and finishing of these baskets and repair, reuse) (67). Last, they were classified according to their func- tools from Cueva de los Murciélagos open up groundbreaking per- tion or their morphology when function was unclear. spectives on the complexity of Early-Middle Holocene populations Fiber-based objects were sampled using tweezers and a blade in Europe. They provide insights into the perishable material tech- cleaned with ethanol 96% between each different object to avoid nologies used on plants and into the knowledge involved in the ac- cross-contamination for further studies. Although sampling is a de- quisition and processing of these plants. Only a small sample of structive process, very small fragments are needed for identification basketry from the cave has been dated, and although these objects analysis. Because of the fragility of these objects, small parts break are representative of the typology of organic objects recovered, not off during restoration and storage processes and, in most cases, it all the organic remains can be ascribed to these periods. Further ra- was these fragments that were sampled to avoid further damage to diocarbon analyses are needed to corroborate these initial the objects. In just one case, 592, sampling was not possible because hypotheses. the object was completely consolidated, and sampling could have Plant material culture offers unique insights into the life of pre- been damaging. Identifications were performed under ×50 to historic societies. Lack of preservation has meant that perishable ×500 magnification using a transmitted light bright-dark field materials have not previously been extensively considered during Olympus BX51 microscope coupled with an Olympus DP26 archaeological research. It is vital that, where they do survive, camera and linked to Olympus cellSens software. Some samples these materials are the focus of detailed study to further examine were placed on stubs using double-sided carbon tape and were the role of such technologies during prehistory. Martínez-Sevilla et al., Sci. Adv. 9, eadi3055 (2023) 27 September 2023 14 of 18 S C I E N C E A D VA N C E S | R E S E A R C H A R T I C L E coated with a 15-nm layer of gold using the Emitech K550 Sputter example of the twisted form, this having an S2z pattern. Most of the Coater Unit to be observed under a Zeiss Merlin field-emission cords seem to be components of other objects such as baskets and scanning electron microscope. Fiber taxonomic identification was sandals, which have been fragmented by postdepositional processes. based on the anatomical and histological description of the archae- Sandals represent an important part on the inventory of the ological samples and by comparison with modern species collected fiber-based materials from the cave. Alfaro (28, 29) distinguished in the immediate area of the site. Alfaro (29) identified the raw ma- two types: simple and central core. Simple sandals consist of a terial as “esparto grass” but provided no other details. Because of the braid wrapped around itself, which shapes the sandal sole and variety of Gramineae known as esparto, the archaeological samples makes it wider. Central core sandals are based on a bunch of were compared with a reference collection of similar plants also wrapped fibers that forms a central core, which is surrounded by used in fiber crafting activities, and which also grow naturally fiber bundles, increasing the width of the sole. The front of the close to the site. These included species of different genera such sole is always wider than the back because it includes one extra as Stipa (S. tenacissima, S. gigantea) and Lygeum (L. spartum). bundle. Although no evidence of tying elements was conserved Other characteristics of the raw material were also recorded, in- for the simple sandals, in the central core type, a small group of cluding the size of the leaves and their integrity, to identify the cri- fibers originating from the bottom of the sole may have been teria used to select the fibers. The diameter of the leaves was placed between the first and second toes. This element is connected measured to check for differences in the raw material selection in to a braid in the middle of the sandal, which could have been fixed relation to the technique used and processing method applied, to the instep and tied around the ankle. Other objects (e.g., 626) are and to the chronology of the materials. not described here because they are singular elements, the typolo- A standardized nomenclature for basketry production is still gies of which have not been determined. needed. Terminology has been much discussed and used differently by many authors throughout history; it is also affected by linguistic Artifact selection and sampling issues (40). The vocabulary used by Adovasio (2) is applied here to Selection of objects for radiocarbon dating was based on criteria in- describe the functional relationship between passive (warp) and cluding sample availability, raw material, technology, and typology. active (weft) elements of fiber-based materials, but some techniques A set of 14 artifacts representing the complete typology of the as- are not covered, so the work of other authors was also considered semblage was selected for dating. Of the wooden artifacts, the (28, 52). The basketry techniques were therefore classified as simple hammer (475) and the digging stick (479) were chosen because of or diagonal twining, coiling (2), cofín (52), or pseudobraided (29), their unique presence in the set but were also prioritized as they and a specific type of basket with knots on the base called braided by were partially broken with areas available for sampling. Obtaining Alfaro (29). Descriptions of the techniques are presented below samples from other wooden materials was not possible because of using this nomenclature. Each object exhibited individual charac- preservation issues (472, 473, 474, 1140, 1141, 1165, 1138/532, and teristics not detailed here. Detailed accounts of the dated objects 1139/532). For the vegetal fiber–based materials, the dating selec- can be found in the Supplementary Materials. tion covered the basketry techniques of twining (579, 580, 581, In our study, we refer to weft twining that consists of vertical, and 617), coiling (615), cofin basketry (594, 623, and 624), the passive warps linked by an active weft, which forms a twisted cord special basketry technique with knots on the base (625), and the around them. In closed twining, no space is visible between the weft four linked rings (626). Last, the two types of sandals were strands. The newly added fibers of the weft are visible on the inner chosen: the simple circular string (611) and the central core type surface. Simple twining presents a single warp around the whole (603). Only one piece (592) in the fiber-based set could not be basket, while in diagonal twining [also called rhomboidal twining sampled because of complete consolidation. by Alfaro (29)], a pair of warps alternates in each round of the basket production, creating a diagonal pattern on the finished Radiocarbon dating object. The coiling technique involves a passive, wrapped bundle Our analysis considered 18 radiocarbon dates: 14 made in the or foundation wrapped by active stitches or a winder, forming a framework of this research and 4 from published results (32, 38). spiral. There are variations in the typology of the sewing element. Radiocarbon dates for the present study were obtained from Beta The cofín (52) or pseudobraided (29) basketry technique involves Analytic Laboratory (Miami, USA). Sampling was carefully under- braiding three bundles of fibers, one of which is changed in each taken to preserve the total integrity of the objects, taking small fibers round, forming the warp of the basket. The technique may be or fragments from areas of breakage. The weight of each sample was “closed” or “open,” depending on the distance between the around 0.5 mg (Table 2). At the level of taxon, the material was braided rows. This method has been documented only in Iberia found to belong to short-lived terrestrial species: 12 were plant ma- (various chronologies) and does not appear in basketry atlases. terial (S. tenacissima), and 2 samples were wood (A. unedo and O. The final basketry technique documented in the Cueva de los Mur- europaea). Thirteen samples were dated by AMS standard, and one ciélagos is braided (28). This rare type is not described in basketry was dated by AMS-Microsample (Beta-627340) due to the small size handbooks. Five knotted fiber bundles on the base of the basket of the wood sample. The samples were pretreated with acid/alkali/ form the warp. They pass several times toward a braid (weft) acid: The sample was gently crushed and then dispersed in deion- started from one of these five knotted bundles, which increases in ized water; it was then washed with hot HCl acid to eliminate car- the number of braids and makes the basket larger at the top. It is bonates and then treated with an alkali wash (NaOH) to remove finished with a wider braid, wrapping the leaves from the warp secondary organic acids. This was followed by a final acid rinse to and the weft, and has a braided handle. neutralize the solution before drying. For two of the samples (Beta- Cords are also represented within the materials from Cueva de 628427 and Beta-628426), it was necessary to carry out a comple- los Murciélagos. Braids are most common form, with only a single mentary cellulose extraction pretreatment to remove solvents Martínez-Sevilla et al., Sci. Adv. 9, eadi3055 (2023) 27 September 2023 15 of 18 S C I E N C E A D VA N C E S | R E S E A R C H A R T I C L E resulting from restoration. This followed the full acid/alkali/acid REFERENCES AND NOTES pretreatment and involved bathing the samples in sodium chlorite 1. L. Hurcombe, Perishable Material Culture in Prehistory: Investigating the Missing Majority (Routledge, 2014). (NaClO2) under controlled conditions (pH 3 and temperature at 2. J. M. Adovasio, Basketry Technology: A Guide to Identification and Analysis (Routledge, 2016). 70°C) to eliminate all components except wood cellulose. 3. L. Keeley, N. Toth, Microwear polishes on early stone tools from Koobi Fora, Kenya. Nature Our analysis also considered five radiocarbon measurements ob- 293, 464–465 (1981). tained using standard radiometric dating at Consejo Superior de In- 4. P. Bello-Alonso, J. Rios-Garaizar, J. Panera, S. Rubio-Jara, A. Pérez-González, R. Rojas, vestigaciones Científicas (Madrid, Spain) (37). Four samples E. Baquedano, A. Mabulla, M. Domínguez-Rodrigo, M. Santonja, The first comprehensive correspond to short-lived terrestrial species, and one to an uniden- micro use-wear analysis of an early Acheulean assemblage (Thiongo Korongo, Olduvai Gorge, Tanzania). Quat. Sci. Rev. 263, 106980 (2021). tified desiccated wood artifact (Table 2). All the radiocarbon mea- 5. N. J. Conard, J. Serangeli, G. Bigga, V. Rots, A 300,000-year-old throwing stick from Schö- surements were calibrated with the internationally agreed IntCal20 ningen, northern Germany, documents the evolution of human hunting. Nat. Ecol. Evol. 4, atmospheric calibration curve (68). 690–693 (2020). 6. B. L. Hardy, M.-H. Moncel, C. Kerfant, M. Lebon, L. Bellot-Gurlet, N. Mélard, Direct evidence Bayesian chronology of Neanderthal fibre technology and its cognitive and behavioral implications. Sci. Rep. 10, The radiocarbon measurements were subjected to Bayesian analyses 4889 (2020). to obtain reliable estimations of the start, end, and duration of dep- 7. B. Aranguren, A. Revedin, N. Amico, F. Cavulli, G. Giachi, S. Grimaldi, N. Macchioni, osition of the artifacts in Cueva de los Murciélagos. Our analyses F. Santaniello, Wooden tools and fire technology in the early Neanderthal site of Poggetti Vecchi (Italy). Proc. Natl. Acad. Sci. U.S.A. 115, 2054–2059 (2018). were performed using the OxCal online software version 4.4 (69). 8. J. Rios-Garaizar, O. López-Bultó, E. Iriarte, C. Pérez-Garrido, R. Piqué, A. Aranburu, The two-sigma probability interval (95.4%) was used when discuss- M. J. Iriarte-Chiapusso, I. Ortega-Cordellat, L. Bourguignon, D. Garate, I. Libano, A middle ing the 14C measurements, and the one-sigma probability interval Palaeolithic wooden digging stick from Aranbaltza III, Spain. PLOS ONE 13, (68.2%) was added in the tables and figures, as recommended by e0195044 (2018). Millard (70). The degree of contemporaneity between the different 9. J. M. Adovasio, Artifacts and ethnicity: Basketry as an indicator of territoriality and popu- radiocarbon measurements was tested through the chi-square test lation movements in the prehistoric great basin, in Anthropology of the Desert West: Essays in Honor of Jesse D. Jennings, C. Condie, D. D. Fowler, Eds. (Utah Anthropological Press, (42), which assesses the degree of overlap between the probability 1986), pp. 44–84. ranges of each of the dates. 10. D. R. Croes, Prehistoric ethnicity on the northwest coast of North America: An evaluation of We first applied nonparametric statistical methods based on style in basketry and lithics. J. Anthropol. Archaeol. 8, 101–130 (1989). KDE (71) as an exploratory mode to characterize the potential 11. E. A. Goldberg, Recent analyses of fibre perishables from promontory Caves, UT. COMPASS phases of the human deposition activities in the cave through 2, 36–53 (2018). time. KDE methods allow the identification of discontinuities in 12. P. Barker, Woven sandals as boundary markers between the great basin and southwest culture areas, in Past, Present, and Future Issues in Great Basin Archaeology: Papers in Honor the probability distribution that may relate to several chronological of Don D. Fowler, B. Hockett, Ed. (Bureau of Land Management Cultural Resource Series, No. phases (71). These methods are widely implemented when there is 20, 2009), pp. 125–145. no formal prior data, reducing the noise from the calibration pro- 13. A. Camp, Catlow twine basketry in the western Great Basin: Use and reuse in the ar- cedure (71). We used this method as the analyzed samples originate chaeological record. Quat. Int. 466, 318–323 (2018). from historical excavations, and there is a lack of information re- 14. T. J. Connolly, J. P. Barker, “Basketry chronology of the early Holocene in the Northern Great Basin” in: Early and Middle Holocene Archaeology of the Northern Great Basin, D. L. Jenkins, garding the stratigraphic relationships between them (71). We T. J. Connolly, C. M. Aikens, Eds. (University of Oregon Anthropological Papers 62, 2004), used the KDE_Plot Oxcal tool (71, 72). pp. 241–250. The uncalibrated radiocarbon dates were also modeled using 15. T. J. Connolly, E. A. Kallenbach, P. Barker, S. J. McCabe, Netting in the northern and western single uniform phase and multiphase models (72). This approach Great Basin. J. California Great Basin Anthropol. 37, 143–167 (2017). combines the radiocarbon dates in a uniform distribution model 16. C. S. Fowler, E. M. Hattori, Exploring prehistoric trade in western great basin textiles, in based on the hypothesis that all the dated events have the same like- Perspectives on Prehistoric Trade and Exchange in California and the Great Basin, R. E. Hughes, Ed. (University of Utah Press, 2011), pp. 201–220. lihood of occurring at any time at the start and end of the phase. The 17. C. S. Fowler, E. M. Hattori, Prehistoric textile trade and exchange in the western Great Basin: model then calibrates the radiocarbon dates based on prior infor- Outland coiling and catlow twining, in Meeting at the Margins: Prehistoric Cultural Inter- mation from other early measurements of the chronological actions in the Intermountain West, D. Rhode, Ed. (University of Utah Press, 2012), pp. 92–102. phase. The model was developed using OxCal tools (Sequence, 18. J. M. Adovasio, R. L. Andrews, J. S. Illingworth, Netting, net hunting, and human adaptation Phase, Boundary, Interval, and Difference commands). in the eastern great basin, in Past, Present, and Future Issues in Great Basin Archaeology: Papers in Honor of Don D. Fowler, B. Hockett, Ed. (Bureau of Land Management Cultural We also modeled the radiocarbon dates with the Charcoal Resource Series, No. 20, 2009), pp. 84–102. Outlier model (41, 73). This model is designed to lessen the 19. P. R. Geib, E. A. Jolie, The role of basketry in early Holocene small seed exploitation: Im- impact of potential inbuilt age issues, as three measurements corre- plications of a ca. 9,000 year-old basket from Cowboy Cave, Utah. Am. Antiq. 73, spond to undetermined and long-lived desiccated wood samples 83–102 (2008). (Table 2). The Charcoal Outlier method provides younger age esti- 20. S. Noshiro, Y. Sasaki, K. Kobayashi, M. Suzuki, I. Nishida, Material selection and weaving mates, as the correct age of the modeled events is more recent than techniques for the oldest basketry in Japan found at the Higashimyou site, Saga prefecture. J. Archaeol. Sci. 23, 12–24 (2019). the nonmodeled calibrated radiocarbon dates (73). A prior 5% 21. R. Piqué, S. Romero, A. Palomo, J. Tarrús, X. Terradas, I. Bogdanovic, The production and use probability of being outliers was assigned to these three samples. of cordage at the early Neolithic site of La Draga (Banyoles, Spain). Quat. Int. 468, In the text, the modeled dates are rounded to the nearest half- 262–270 (2018). decade since the modeled results vary from run to run. 22. M. Mineo, N. Mazzuco, M. Rottoli, G. Remolins, L. Caruso-Fermé, J. F. Gibaja, Textiles, basketry and cordage from the Early Neolithic settlement of La Marmotta, Lazio. Antiquity Supplementary Materials 97, 314–330 (2023). This PDF file includes: 23. A. Bosch, J. Chinchilla, J. Tarrús, Els objectes de fusta del poblat neolític de La Draga. Exca- Supplementary Text vacions de 1995–2005 (Centre Arqueologia Subaquàtica de Catalunya-Museu d’Arqueolo- Figs. S1 to S6 gia de Catalunya, 2006). Tables S1 to S5 Martínez-Sevilla et al., Sci. Adv. 9, eadi3055 (2023) 27 September 2023 16 of 18 S C I E N C E A D VA N C E S | R E S E A R C H A R T I C L E 24. J. E. A. Tortosa, G. Pérez-Jordà, Y. C. Marco, J. R. Seguí Seguí, J. F. J. Pardo, C. Miret i Estruch, 48. A. Hafner, K. Glauser, P. Curdy, D. von Rütte, C. Rérolle, M. Bader, Schnidejoch und Löt- C. C. V. Cebrián, Cordage, basketry and containers at the Pleistocene–Holocene boundary schenpass: Naturraum, Entdeckung und Feldarbeiten, in Schnidejoch und Lötschenpass, in southwest Europe. Evidence from Coves de Santa Maira (Valencian region, Spain). Veg. A. Hafner, Ed. (Archäologische Forschungen in den Berner Alpen 1, 2015), pp. 69–129. Hist. Archaeobot. 29, 581–594 (2020). 49. A. Hafner, R. Stapfer, J. Francuz, Die Bucht von Sutz-Lattrigen, in 4000 Jahre Pfahlbauten. 25. M. Herrero-Otal, S. Romero-Brugués, R. Piqué Huerta, Plants used in basketry production Archäologischen Landesmuseum Baden-Württemberg und dem Ladesamt für Denkmalpflege during the Early Neolithic in the north-eastern Iberian Peninsula. Veg. Hist. Archaeobot. 30, im Regierungspräsidium Stuttgart (Thorbecke, 2016). 729–742 (2021). 50. J. Junkmanns, J. Klügl, W. Schoch, G. Di Pietro, A. Hafner, A. (2019) Neolithic and Bronze 26. S. Romero-Brugués, M. Herrero-Otal, R. Piqué, R. Rosillo, X. Terradas, O. López-Bultó, Age archery equipment from Alpine Ice-Patches: A review on components, construction A. Berrocal-Barberà, A. Palomo, Los implementos elaborados con fibras vegetales del techniques and functionality. Journal of Neolithic. Archaeology 21, 283–314 (2019). neolítico antiguo de Coves del Fem, Ulldemolins (Tarragona). Munibe Antropologia-Ar- 51. B. Kuoni, Cestería Tradicional Ibérica (Ediciones del Serbal, 1981). keologia 72, 43–56 (2021). 52. J. Fajardo, A. Verde, D. Rivera, C. Obón, S. Leopold, Traditional craft techniques of Esparto 27. M. Gleba, M. D. Bretones-García, C. Cimarelli, J. C. Vera-Rodríguez, R. M. Martínez-Sánchez, Grass (Stipa tenacissima L.) in Spain. Econ. Bot. 69, 370–376 (2015). Multidisciplinary investigation reveals the earliest textiles and cinnabar-coloured cloth in 53. O. López-Bultó, R. Piqué, F. Antolín, J. A. Barceló, A. Palomo, I. Clemente, Digging sticks and Iberian Peninsula. Sci. Rep. 11, 21918 (2021). agriculture development at the ancient Neolithic site of la Draga (Banyoles, Spain). 28. C. Alfaro, Estudio de los materiales de cestería procedentes de la Cueva de los Murciélagos J. Archaeol. Sci. Rep. 30, 102193 (2020). (Albuñol, Granada). Trabajos de Prehistoria 37, 109–139 (1980). 54. J. M. Coles, F. A. Hibbert, A Neolithic wooden mallet from the Somerset levels. Antiquity 46, 29. C. Alfaro, Tejido y cestería en la Península Ibérica. Historia de su técnica e industrias desde la 52–54 (1972). prehistoria hasta la romanización (Bibliotheca Praehistorica Hispana 21, CSIC Instituto 55. X. Terradas, R. Piqué, A. Palomo, F. Antolín, O. López, J. Revelles, R. Buxó, Farming practices Español de Prehistoria, 1984). in the Early Neolithic according to agricultural tools: Evidence from La Draga site (North- 30. M. Gómez Moreno Martínez, La cerámica primitiva ibérica, in Homenagem a Martins Sar- eastern Iberia), in Times of Neolithic Transition along the Western Mediterranean, O. García- mento no centenario do seu Nascimento (Ediçao da Sociedade Martins Sarmento, 1933), Puchol, D. C. Salazar-García, Eds. (Springer International Publishing, 2017), pp. 199–220. pp. 125–136. 56. A. Bocquet, M. Noël, The Neolithic or Wood Age. Endeavour 9, 34–41 (1985). 31. M. Gómez Moreno Martínez, La cerámica primitiva ibérica, in Misceláneas. Historia del Arte y 57. S. Romero-Brugués, R. Piqué Huerta, M. Herrero-Otal, The basketry at the early neolithic Arqueología. Primera Serie. La antigüedad (Consejo Superior de Investigaciones Científicas, site of La Draga (Banyoles, Spain). J. Archaeol. Sci. Rep. 35, 102692 (2021). 1949), pp. 104. 58. C. R. Cartwright, The principles, procedures and pitfalls in identifying archaeological and 32. P. López, La problemática cronológica del Neolítico Peninsular, in C-14 y Prehistoria de la historical wood samples. Ann. Bot. 116, 1–13 (2015). Península Ibérica, M. Almagro-Gorbea, M. Fernández-Miranda, Eds. (Fundación Juan March - 59. F. H. Schweingruber, Anatomy of European Woods. An Atlas for the Identification of European Serie Universitaria, 1978), pp. 45–56. Trees, Shrubs and Dwarf Shrubs (Paul Haupt, 1990). 33. M. de Góngora, Antigüedades Prehistóricas de Andalucía (Imprenta a cargo de 60. Ü. Akkemik, B. Yaman, Wood Anatomy of Eastern Mediterranean Species (Kessel Publishing C. Moro, 1868). House, 2012). 34. P. López, Estudio de la cerámica, industria ósea y lítica de la Cueva de los Murciélagos de 61. R. Gale, D. Cutler, Plants in Archaeology. Identification Manual of Vegetative Plant Materials Albuñol (Granada). Trabajos de Prehistoria 37, 163–180 (1980). Used in Europe and the Southern Mediterranean to c. 1500 (Kew: Westbury and Royal Botanic 35. F. Martínez-Sevilla, Brazaletes de piedra neolíticos en la península ibérica (VI-V milenio a. C.): Gardens, 2000). Tecnología, funcionalidad y circulación (British Archaeological Reports, International Series 62. J.G. Hather, The Identification of the Northern European Woods. A Guide for Archaeologists 2913, 2018). and Conservators (Archetype Publications, 2000). 36. J. Carrasco Rus, J. A. Pachón Romero, Algunas cuestiones sobre el registro arqueológico de 63. A. Crivellaro, F. H. Schweingruber, Atlas of Wood, Bark and Pith Anatomy of Eastern Medi- la Cueva de los murciélagos de Albuñol (Granada) en el contexto Neolítico andaluz y sus terranean Trees and Shrubs: With a Special Focus on Cyprus (Springer Science & Business posibles relaciones con las representaciones esquemáticas. Cuad. Prehist. Arqueol. Univ. Media, 2013). Granada 19, 227–287 (2009). 64. D. Marguerie, J. Y. Hunot, Charcoal analysis and dendrology: Data from archaeological sites 37. C. Cacho, C. Papí, A. Sánchez-Barriga, F. Alonso, La cestería decorada de la Cueva de los in north-western France. J. Archaeol. Sci. 34, 1417–1433 (2007). Murciélagos (Albuñol, Granada). Complutum 6, 105–122 (1996). 65. R. A. Blanchette, A review of microbial deterioration found in archaeological wood from 38. J. Carrasco Rus, F. Martínez-Sevilla, Las cronologías absolutas del Neolítico Antiguo en el different environments. Int. Biodeter. Biodegr. 46, 189–204 (2000). sur de la península ibérica Nuevas dataciones. Archivo Prehist. Levant 30, 57–80 (2014). 66. M. Toriti, A. Durand, F. Fohrer, Traces of Common Xylophagous Insects in Wood (Springer 39. S. S. Aliscioni, Anatomía ecológica de algunas especies del género Paspalum (Poaceae, International Publishing, 2021). Panicoideae, Paniceae). Darwiniana 38, 187–207 (2000). 67. R. Sands, Life beyond life: Repair, reuse, and recycle—The many lives of wooden objects 40. W. Wendrich, The World According to Basketry. An Ethnoarcheological Interpretation of and the mutability of trees. Archaeometry 64, 168–186 (2022). Basketry Production in Egypt (UCLA—Cotsen Digital Archaeology series, Leiden Universi- 68. P. J. Reimer, W. E. N. Austin, E. Bard, A. Bayliss, P. G. Blackwell, C. B. Ramsey, M. Butzin, ty, 1999). H. Cheng, R. L. Edwards, M. Friedrich, P. M. Grootes, T. P. Guilderson, I. Hajdas, T. J. Heaton, 41. C. Bronk Ramsey, Dealing with outliers and offsets in radiocarbon dating. Radiocarbon 51, A. G. Hogg, K. A. Hughen, B. Kromer, S. W. Manning, R. Muscheler, J. G. Palmer, C. Pearson, 1023–1045 (2009). J. van der Plicht, R. W. Reimer, D. A. Richards, E. M. Scott, J. R. Southon, C. S. M. Turney, 42. G. K. Ward, S. R. Wilson, Procedures for comparing and combining radiocarbon age de- L. Wacker, F. Adolphi, U. Büntgen, M. Capano, S. M. Fahrni, A. Fogtmann-Schulz, R. Friedrich, terminations: A critique. Archaeometry 20, 19–31 (1978). P. Köhler, S. Kudsk, F. Miyake, J. Olsen, F. Reinig, M. Sakamoto, A. Sookdeo, S. Talamo, The 43. C. Earwood, Radiocarbon dating of late prehistoric wooden vessels. J. Ir. Archaeol. 5, IntCal20 Northern Hemisphere radiocarbon age calibration curve (0–55 cal kBP). Radio- 27–44 (1989). carbon 62, 725–757 (2020). 44. M. M. Coe, Basketry, cordage, and perishable artifact manufacture at Bonneville Estates 69. C. Bronk Ramsey. OxCal v.4.4.4 (2021); https://c14.arch.ox.ac.uk/oxcal.html [accessed Rockshelter: Diachronic technological variation. J. Ir. Archaeol. 64, 101325 (2021). February 2023]. 45. R. Pinhasi, B. Gasparian, G. Areshian, D. Zardaryan, A. Smith, G. Bar-Oz, T. Highman, First 70. A. R. Millard, Conventions for reporting radiocarbon determinations. Radiocarbon 56, direct evidence of chalcolithic footwear from the near eastern highlands. PLOS ONE 6, 555–559 (2014). e10984–e12010 (2010). 71. C. Bronk Ramsey, Methods for summarizing radiocarbon datasets. Radiocarbon 59, 46. J. Banck-Burgess, Mehra als nur Leder. Regionen und Kulturen – die Pfahlbauten der 1809–1833 (2017). Jungsteinzeit, in 4000 Jahre Pfahlbauten. Archäologischen Landesmuseum Baden-Würt- 72. C. Bronk Ramsey, Bayesian analysis of radiocarbon dates. Radiocarbon 51, 337–360 (2009). temberg und dem Ladesamt für Denkmalpflege im Regierungspräsidium Stuttgart (Thor- 73. M. W. Dee, C. Bronk Ramsey, High-precision Bayesian modeling of samples susceptible to becke, 2016). inbuilt age. Radiocarbon 56, 83–94 (2014). 47. R. Goedecker-Ciolek, Zur Herstellungstechnik von Kleidung und Aüstrustungsgegen- sta¨nden, in Die Gletschermumie vom Ende der Steinzeit aus den Ötztaler Alpen, M. Egg, Acknowledgments: We thank the Museo Arqueológico Nacional (Madrid) and the curator, R. Goedecker-Ciolek, W. Groenman-van Waateringe, K. Spindler, Eds. (Jahrbuch des E. Galán Domingo, and Museo Arqueológico y Etnográfico (Granada) and its director, M. Ramos Römisch-Germanisches Zentralmuseums, 1993), pp. 100–113. Lizana, for permission to study the material and undertake sampling. We thank the landowners of Cueva de los Murciélagos, D. Lupiáñez Soto, M. I. Lupiáñez Soto, and P. Lupiáñez Soto, for allowing us to carry out fieldwork at the cave. We thank the fieldwork team: A. Peralta, B. Ramos Martínez-Sevilla et al., Sci. Adv. 9, eadi3055 (2023) 27 September 2023 17 of 18 S C I E N C E A D VA N C E S | R E S E A R C H A R T I C L E Rodríguez, L. M. Zahonero Gómez, N. Acosta Gómez, and F. Fernández Castaño for their J.S., J.A.L.R., and R.P.H. Writing—review and editing: F.M.-S., M.H.-O., M.M.-S., J.S., J.A.L.R., M.C., commitment and S. Fernández Martín, the Historical Heritage Conservator, for support and A.H., R.M.M.S., I.B., R.B.B., P.B.R., A.P.P., A.M.A.-V., L.P.-C., M.M.-B., E.F.-D., M.A.G., R.P.M., M.I.C., advice. We also thank K. Sharpe for editing the work and G. Jiménez Gisbert for calculating the J.L.C.R., C.A.G., and R.P.H. Competing interests: The authors declare that they have no volume of the baskets. We thank the Excmo. Ayuntamiento de Albuñol for their support for our competing interests. Data and materials availability: All data needed to evaluate the project and the fieldwork, especially M. M. J. Sánchez Sánchez and culture technician, conclusions in the paper are present in the paper and/or the Supplementary Materials. All E. Sánchez García. Funding: This work was supported by the project “De los museos al necessary permits were obtained from the Museo Arqueológico Nacional (Madrid) and Museo territorio: actualizando el estudio de la Cueva de los Murciélagos de Albuñol (Granada)” Arqueológico y Etnográfico (Granada) for the artifact sampling and the described study. The (MUTERMUR) (Referencia CM/JIN/2021-009) financed by the program of Young Researchers of fieldwork in Cueva de los Murciélagos had all the required permits from the Delegación Comunidad de Madrid (directed by F.M.-S. as the principal investigator). J.S. is funded by Ramon Territorial de Cultura (Granada), local authorities, and landowners; the work was directed by y Cajal fellowship (RYC2019-028346) by the Spanish Ministry of Science and Innovation (MCIN). F.M.-S. The identification numbers (IDs) of the artifacts analyzed here are provided in Table 1. M.C. is funded by a Ramón y Cajal fellowship (RYC2019-026697-I) and by the Spanish Ministry of The materials are housed at the Museo Arqueológico Nacional (Madrid) and Museo Science and Innovation (MCIN). R.P.H., M.H.-O., and A.P.P. are members of the research group Arqueologico y Etnográfico (Granada). The objects are available to any researcher to be 2021-SGR 00190 funded by the AGAUR. R.P.H. is an ICREA academia researcher. M.M.-S. was inspected with the permission of both institutions. funded by the Beatriz Galindo program as Junior Distinguished Researcher (BG20/00076). Author contributions: Conceptualization: F.M.-S., M.H.-O., M.C., and R.P.H. Supervision: F.M.-S., Submitted 7 May 2023 M.H.-O., R.M.R., M.C., and R.P.H. Selection of archaeological samples: M.H.-O., M.M.-S., R.M.R., Accepted 18 August 2023 A.H., I.B., P.B.R., R.d.B.B., and R.P.H. Methodology: F.M.-S., M.H.-O., M.M.-S., J.S., J.A.L.R., A.M.A.-V., Published 27 September 2023 M.I.C., and R.P.H. Funding acquisition: F.M.-S. Writing—original draft: F.M.-S., M.H.-O., M.M.-S., 10.1126/sciadv.adi3055 Martínez-Sevilla et al., Sci. Adv. 9, eadi3055 (2023) 27 September 2023 18 of 18