Submerged bridge constructed at least 5600 years ago indicates early human arrival in Mallorca, Spain

The study investigates the timing of the earliest human arrival on Mallorca, a long-debated question due to sparse and contested archaeological evidence. Previous claims based on radiocarbon-dated materials from caves (e.g., Cova de Moleta and Cova de Canet) have been questioned due to issues of sample integrity, context, and lack of clear association with human activity. A more recent consensus places initial settlement between 4600 and 4200 calibrated years B.P., with a synthesis of arrival models (early ~9000 cal B.P., intermediate ~7600 cal B.P., late ~5000 cal B.P.) favoring the late model. Yet, chronological constraints relying on the last occurrence of Myotragus balearicus and earliest domesticates provide only a narrow and potentially incomplete window. The purpose of this study is to provide an independent geochronological constraint on human presence by dating a submerged stone bridge within Genovesa Cave and correlating it with sea-level indicators preserved as phreatic overgrowths on speleothems (POS) and a distinct coloration band. By tying the structure’s elevation to a high-resolution Holocene relative sea-level curve for Mallorca and U-series ages on POS, the authors aim to bracket the construction time of the bridge and thereby constrain the minimum age of human presence on the island.

• Earlier radiocarbon evidence suggesting human presence as early as 7000–9000 cal B.P. from Cova de Moleta and Cova de Canet has been deemed unreliable due to poor preservation, uncertain stratigraphic context, and lack of clear human association (e.g., possible misidentification of bones, controversial charcoal layers).
• Reassessments and new radiocarbon datasets from charcoal, ash, and bones led to a consensus for earliest settlement between 4600 and 4200 cal B.P.
• A synthesis of arrival models for the Balearic Islands includes Early (~9000 cal B.P.), Intermediate (~7600 cal B.P.), and Late (~5000 cal B.P.) phases; only the Late model has been considered plausible in local literature.
• The latest pre-study consensus (~4400 cal B.P.) hinges on a 350-year window (p > 90%) between the last documented Myotragus bone (4581–4417 cal B.P.) and the first dated sheep bone (4417–4231 cal B.P.), but it is uncertain whether these represent true last and first occurrences.
• The study builds upon speleothem-based sea-level reconstructions in Mallorca and integrates POS proxies with archaeological observations in coastal caves where cave water levels track sea level due to negligible hydraulic gradients.

• Study site: Genovesa Cave (Cova de'n Bessó; 39°31′32″ N, 3°19′2″ E) in eastern Mallorca, containing ceramic sherds and stone constructions, including a submerged stone bridge (8.62 m long, 0.5 m high) currently submerged by 1.05 ± 0.1 m relative to preindustrial sea level. Nearby Drac Cave (39°32′9″ N, 3°19′49″ E) provided comparative samples.
• Conceptual basis: In coastal karst caves with negligible hydraulic gradients, cave lake levels track sea level. During sea-level stillstands, calcite/aragonite phreatic overgrowths on speleothems (POS) precipitate at the water table; stalactites formed in air become submerged as sea level rises. A distinct light-colored “bathtub ring” (thin calcite crust) on the bridge’s upper part marks a former stillstand.
• Underwater mapping and depth measurements: Used Suunto SK-8 compass, pre-knotted diving line, rollable tape (±1 cm), and Suunto Gekko depth-meter (±1%, 0.1 m resolution) to locate and measure sample depths (0.75–2 m below present sea level). Bridge position and coloration band elevations mapped with Suunto Tandem 360PC compass/clinometer (0.33°/0.25° accuracy) and Bosch GLM-50 laser (1.5 mm accuracy). Hydraulic gradient measured using Leica TC 405 total station; head change 0.037 m over 412 m implied gradient 9 × 10⁻⁵ m/m (even lower at Drac), confirming water table ≈ sea level. Lake surface elevation precision better than 2 mm; bridge and coloration band depths measured by diver using millimeter tape (±5 mm).
• Sampling and U-series dating: 34 U-series ages (28 POS, 6 stalactites) from Genovesa and Drac; some previously published. POS and stalactite subsamples prepared via acid dissolution, spiked with ²²⁹Th–²³³U–²³⁶U, cleaned using anion exchange chromatography. U and Th measured on Thermo Neptune Plus MC-ICP-MS with Aridus II; Faraday cups and SEM used as appropriate; gain calibration with CRM-112 and in-house ²³⁰Th–²²⁹Th standard. Ages calculated using Cheng et al. (2013) decay constants; isochron methods (Rosholt A type) implemented in IsoplotR.
• Sea-level framework: Relative sea level (RSL) interpreted using a high-resolution Late Holocene RSL curve for Mallorca (based on a Glacial Isostatic Adjustment model using ICE-6G (VM5) ice history with upper mantle viscosity 1.3 × 10²⁰ Pa s). Depths of POS and the bridge’s coloration band compared against this RSL curve to bracket construction timing.
• Data curation: All samples are archived at the University of the Balearic Islands Repository of Phreatic Overgrowth on Speleothems; full dataset available via Zenodo.

• The bridge’s upper surface lies 1.05 ± 0.1 m below preindustrial sea level. A ~15 cm wide, light-colored “bathtub ring” coloration is present along the bridge’s upper part at the same depth as POS (~1.1 ± 0.1 m below preindustrial sea level).
• U-series POS ages at ~1.10 ± 0.1 m below preindustrial sea level: • GE-D6 isochron age: 5479 ± 120 yr B.P. • GE-D7 weighted average: 5510 ± 549 yr B.P. • DR-D23 isochron age: 5824 ± 140 yr B.P. These indicate a sea-level stillstand between 5820 ± 140 and 5479 ± 120 yr B.P.; maximum span 5964–5359 yr B.P.
• Additional POS-defined stillstands: • At 0 ± 0.04 m: 2720 ± 11 to 296 ± 18 yr B.P. • At ~0.25 m below preindustrial sea level: 3703 ± 14 to 3368 ± 8 yr B.P.
• RSL curve analysis shows that at 3500 cal B.P. sea level was ~0.25 ± 0.1 m below preindustrial, submerging the 0.5 m-high bridge by ~0.8 m, making construction then unlikely; at ~4400 cal B.P., sea level was ~0.35 ± 0.1 m below preindustrial, implying ~0.7 m submergence—also unlikely for construction.
• The top of the bridge would have been at or near water level no earlier than ~5600 yr B.P., and the coloration band could only form when the top was at the air–water interface during the ~5964–5359 yr B.P. stillstand.
• Therefore, bridge construction is constrained between ~6000 and ~5600 yr B.P.: likely initiated when lake depth was ~0.25 m (near ~6000 yr B.P.) and completed before ~5600 yr B.P., after which the top became submerged.
• This provides evidence for human presence on Mallorca by at least 5600 yr B.P., potentially approaching ~6000 yr B.P., earlier than the prevailing ~4400 cal B.P. estimate.

By integrating U-series ages of POS and the elevation of a distinctive coloration band on a submerged bridge with a high-resolution Holocene RSL curve for Mallorca, the study constrains the bridge’s construction to ~6000–5600 yr B.P. This directly implies human presence on Mallorca earlier than the commonly cited ~4400 cal B.P. date based on Myotragus and earliest domesticates. The logic rests on the physical impossibility or improbability of constructing a dry-use bridge while it was submerged by 0.7–0.8 m at ~4400–3500 cal B.P., and on the requirement that the coloration band formed at the air–water interface during a brief sea-level stillstand recorded by POS at ~1.1 m below preindustrial sea level. The results address the research question by providing an independent, sea-level–calibrated archaeological proxy for minimum human presence, reducing reliance on contested radiocarbon contexts. The significance lies in pushing back the earliest substantiated human activity on Mallorca to at least 5600 yr B.P., refining regional colonization models for the western Mediterranean and illustrating the value of integrating coastal cave sea-level proxies with archaeological structures.

The study combines new U-series ages of phreatic overgrowths on speleothems in Genovesa and Drac caves with the observed coloration band on a submerged stone bridge and a robust Holocene relative sea-level curve for Mallorca to constrain the bridge’s construction to between ~6000 and ~5600 yr B.P. A brief sea-level stillstand at ~1.1 m below preindustrial sea level (ca. 5964–5359 yr B.P.) allowed both POS formation and the development of the bridge’s “bathtub ring,” implying the bridge was already in place. These findings provide strong evidence for human presence on Mallorca by at least 5600 yr B.P., possibly as early as ~6000 yr B.P., predating widely accepted estimates. Future research could target direct dating of associated archaeological materials in Genovesa Cave, expand POS-based sea-level reconstructions across additional sites, refine GIA models and local RSL uncertainties, and explore functional and cultural contexts of such cave constructions.

• The construction date is inferred indirectly from sea-level proxies (POS ages, RSL curve) and the bridge’s elevation and coloration band; there are no direct radiocarbon dates from materials unequivocally associated with the bridge’s construction event.
• The assembly date lacks a robust time-stratigraphic context or written records, introducing uncertainty.
• The POS stillstand at ~1.1 m below preindustrial sea level spans <600 years, and POS morphology is atypical/smaller, potentially reflecting limited duration and affecting preservation/recognition.
• Conclusions rely on the accuracy of the regional RSL curve and GIA model parameters as well as depth measurements; while uncertainties are quantified, residual model/measurement uncertainties remain.
• Evidence is from a limited number of sites (Genovesa and nearby Drac), and broader regional replication would strengthen generalizability.