The human mandible from the site of Banyoles was accidentally discovered in 1887 during limestone quarrying operations near the town of Banyoles in northeastern Spain (Hernandez-Pacheco and Obermaier, 1915; Julià et al., 1987; Figure 1, Figure 2). Historically, this mandible is one of the earliest fossil hominin discoveries on the Iberian Peninsula, and the specimen has a long history of study and analysis. Nevertheless, despite over a century of research, the taxonomic affinities of this fossil have remained elusive. Given the historical significance of the specimen, along with its fairly complete state of preservation and a clearer idea of its chronology, we undertook a reanalysis of the mandible from Banyoles (referred hereafter as Banyoles) relying on descriptive comparative morphology and quantitative three-dimensional geometric morphometrics (3D GM). We use these approaches to yield new insights into the taxonomic affinities of this mandible.

At the time of the mandible's discovery, the human fossil record consisted of only a few clearly recognizably nonmodern fossils, including the Neandertal child's skull from Engis in Belgium (1830; Schmerling, 1833–1834), the adult Neandertal cranium from Forbes' Quarry in Gibraltar (1848; Busk, 1865), the original Neandertal type specimen from Feldhofer Grotto (1856; Schaaffhausen, 1858), and the early Homo sapiens remains from Cro-Magnon (1868; Broca, 1868). Banyoles was found after Fraipont and Lohest (1887) demonstrated the antiquity of the Neandertal remains from Spy based on the presence of carbonaceous concretions on the bones similar to those found on extinct animal remains at this same site (Pirson et al., 2018). Thus, the discovery of Banyoles occurred at a time when the antiquity of human fossils was being increasingly accepted, along with the notion that humans evolved from an ape-like ancestor.

Banyoles was encased in a limestone block when it was first discovered, and the travertine matrix was subsequently removed by Pere Alsius, a local pharmacist and naturalist, who exposed most of the external surface of the mandible and published two short descriptions of the fossil (Alsius, 1907, 1915). These were followed by the first monographic treatment of the specimen by Hernandez-Pacheco and Obermaier (1915) who aligned Banyoles with the Neandertals. Other authors of the time also generally emphasized resemblances to the Neandertals (MacCurdy, 1915; Sergi, 1917; Giuffrida-Ruggeri, 1921; Hoyos-Sáinz, 1947) or European Middle Pleistocene specimens (Bonarelli, 1916). In contrast, Keith (1931) noted a rudimentary expression of the chin present in Banyoles that is not typically seen in Neandertals. It was not until the 1950s that Santiago Alcobé (1993) undertook the task of extracting the mandible from the remaining adhering matrix and cleaning it. Thus, before 1956, all studies of Banyoles were done while it was still partially encased in the limestone travertine matrix, concealing the posterior and internal aspects of the mandible.

De Lumley (de Lumley, 1971–1972) provided the first detailed study of Banyoles after its restoration by Alcobé, noting some characteristics shared with European Middle Pleistocene (now Chibanian; Suganuma et al., 2021) mandibles, such as a relatively low corpus height, large condylar dimensions, and nontruncation of the gonial margin. De Lumley (de Lumley, 1971–1972) argued that the symphyseal region is most similar to Middle Pleistocene mandibles in lacking a mentum osseum, but with some features resembling recent H. sapiens, including a fairly vertical symphysis and a very shallow alveolar depression and slight lateral tubercles on the external surface.

A monograph on Banyoles and its context was published following an international conference in 1987 (Maroto, 1993). Contributions from numerous scholars characterized the mandible as a late Middle Pleistocene form of archaic genus Homo (de Lumley, 1993; Roth et al., 1993; Rosas, 1993). Sánchez-Lopez (1993) argued for a late Neandertal classification, but also noted some slight chin structures on the mandibular symphysis as potentially more modern-like features. Most recently, Alcázar de Velasco et al. (2011) also argued for the presence of some morphological features Banyoles shares with modern H. sapiens, including some chin structures. Nevertheless, studies that have included Banyoles in comparative samples generally consider it to represent either a Middle Pleistocene European (MPE) specimen or a Neandertal (Rosas, 2001; Mounier et al., 2009; Roksandic et al., 2011). Thus, despite the long history of research, scholars remain divided regarding the taxonomic affinities of this fossil. Part of this confusion seems to be linked to the lack of a clear archaeological context and date.

The town of Banyoles lies on a floodplain adjacent to Lake Banyoles, which has accumulated large amounts of limestone travertine since the Neogene (Hernandez-Pacheco and Obermaier, 1915; Julià et al., 1987; Maroto, 1987). The mandible was found east of Lake Banyoles in the Pla de la Formiga open quarry, which extends into the Pla de la Mata. Although the exact findspot cannot currently be located because of the continued quarrying activities at the site, the mandible was discovered around four meters below the surface in a limestone travertine tuff in the Pla de la Formiga quarry (Julià et al., 1987).

Faunal and floral remains were found near the mandible and can perhaps provide insights into the paleoenvironment of Banyoles, but no formal analysis of the faunal remains has been done (Julià et al., 1987). Nevertheless, the Pla de la Mata travertine matrix does contain Middle-Late Pleistocene species, including several species of Equus (horse), Bos primigenius (wild auroch), Bison priscus (steppe bison), and Cervus elaphus (red deer; Maroto and Soler, 1993).

Several isotopic studies have attempted to date the mandible. An early study by Berger and Libby (1966) radiocarbon dated the site to 17 ± 1.0 ka suggesting the mandible was much younger and postdated the Neandertals, in contrast to stratigraphic estimates of a Middle-Late Pleistocene age (Hernandez-Pacheco and Obermaier, 1915; Bonarelli, 1916; Solé Sabarís, 1957; Bech, 1971). Later, Yokoyama et al. (1987) performed a U-series analysis of a travertine sample of unknown provenience from the site as well as directly on the bone of the mandible itself. Yokoyama et al. (1987) modeled for both closed and open system behaviors and noted that the Banyoles bone was much younger (16.2 ± 3.2 ka) than the travertine for both a closed (ca. 70 ka) and open system (ca. 110 ka).

Shortly thereafter, Julià and Bischoff (1991) performed U-series analyses of a travertine sample which was removed from the mandible during Alcobé's restoration in 1956 as well as several travertine samples from different stratigraphic layers from the Pla de la Mata and the nearby Les Pedreres sediments. Their analyses yielded a date of 45 ± 4 ka for the travertine sample, suggesting this was the age of the fossil. The good agreement between the dates for the mandible and those from Pla de la Mata (45–50 ka) suggested the mandible most likely derived from this site, and the site likely represented a closed system, with no evidence of the mandible having been reworked (Julià and Bischoff, 1991).

Most recently, Grün et al. (2006) conducted a joint U-series/electron spin resonance (ESR) analysis of enamel and dentine fragments from the right M3 of the mandible as well as on the travertine matrix adhering to the mandible. Considerable heterogeneity was seen between the ages derived from the dentine samples, and the mandible appears to have undergone at least two Uranium-accumulation stages (Grün et al., 2006). The travertine samples yielded an age estimate of 42.5 ± 4.1 ka, in good agreement with the results of Julià and Bischoff (1991), whereas the enamel/dentine samples yielded a somewhat older age estimate of 66.0 ± 7.0 ka. This discordance between the enamel/dentine age and the age of the adhering travertine may indicate the mandible was reworked, although Grün et al. (2006) do not favor this hypothesis. Even if it were reworked from the surface of an older travertine, it is unlikely to be more than a few thousand years older than the age of the adhering travertine (Grün et al., 2006). A consensus view, then, would consider the travertine date of Julià and Bischoff (1991) to likely represent a minimum age for the specimen, whereas the enamel/dentine date of Grün et al. (2006) would likely represent close to a maximum age for the specimen. Thus, the age of Banyoles can be considered to fall somewhere between 45 ± 4 ka and 66.0 ± 7.0 ka, and Grün et al. (2006) explicitly rule out the possibility of a late Middle Pleistocene age for the specimen.

The results of the joint U-series/ESR dating for the mandible from Banyoles place it either in the earlier part of Marine Isotope Stage (MIS) 3 or during MIS 4, when continental glaciation was at its most extensive in the Late Pleistocene. Based on the current fossil record, Europe was occupied solely by Neandertals during the first half of the Late Pleistocene, with H. sapiens fossils only appearing during MIS 3 (Ahern et al., 2013; Higham et al., 2014). Although many human fossils do not have a clear chronology, there are very few securely dated fossils from European sites during MIS 4, including perhaps the Neandertal skeleton from Régourdou cave in France (Vandermeersch and Trinkaus, 1995). In contrast, Neandertal remains recovered from sites in southwest Asia such as Amud, Kebara, and Ein Qashish likely fall between 50 and 70 ka (Schwarcz et al., 1989; Rink et al., 2001; Been et al., 2017), making them potentially contemporaneous with Banyoles, and a similar chronology has been reported for the fossil H. sapiens cranium from the site of Manot (Hershkovitz et al., 2015; Alex et al., 2017).

A notable increase in Mousterian occupation sites and Neandertal remains is documented in MIS 3 (Mellars, 1999; Van Andel et al., 2003; Smith et al., 2017), with numerous Neandertal fossils falling within the time range estimated for Banyoles (Higham et al., 2014). However, by around 40 ka, the Neandertals start to disappear from the archaeological and fossil records (Van Andel et al., 2003; Ahern et al., 2013; Smith, 2013; Higham et al., 2014), and the latter half of the Late Pleistocene saw the appearance of H. sapiens in Europe.

Although a potentially much earlier H. sapiens fossil has been reported from Apidima in Greece by 210 ka (Harvati et al., 2019), the earliest fossil evidence for H. sapiens from the Late Pleistocene was recently suggested at the Grotte Mandrin in France dating to between 56.8 and 51.7 ka (Slimak et al., 2022). Other, slightly younger European early H. sapiens fossils, include fragmentary specimens from Bacho Kiro cave (Bulgaria; Hublin et al., 2020), a partial cranium from Peştera cu Oase (Romania), some isolated teeth from the Grotta del Cavallo (Italy) and the cranium from Zlatý Kůň (Czechia) and the fragmentary maxilla from Kent's Cavern (England; Benazzi et al., 2011; Higham et al., 2011, 2014; Prüfer et al., 2021). All these early H. sapiens specimens likely overlap chronologically with the younger age estimate for Banyoles, and recent genetic evidence suggests that ancestral populations of H. sapiens potentially arrived much earlier than MIS 3 (Meyer et al., 2016; Posth et al., 2017; Petr et al., 2020).

Our understanding of the significance of Banyoles has been hampered by several factors. The lack of any archaeological context means behavioral inferences are difficult to draw for this individual, and the precise chronology of the specimen has been an open question for most of the past century. In addition, the extreme tooth wear in this individual also limits useful taxonomic information from the dentition. Fortunately, the chronology of the specimen has much improved, and, although still imprecise, clearly indicates a Late Pleistocene age, likely falling within MIS 3–4. Thus, a Middle Pleistocene age can be ruled out for Banyoles, and this means the most appropriate comparisons for the specimen are with other contemporary hominins in Late Pleistocene Europe and perhaps southwest Asia. As the long and complex history of the taxonomic placement of Banyoles demonstrates, new methodological approaches are necessary to shed further light on this enigmatic specimen.