The Sierra de Atapuerca is a large karstic hill formation in northern Spain that contains several sites which have yielded archaeological artifacts and fossilized remains ranging from the Early Pleistocene to the Holocene. Of all the sites in Atapuerca, the Sima de los Huesos (SH) is the richest in hominin fossils, containing remains from a minimum of 29 individuals dated to around 430 ka, ranging from around 10 years of age to over 35 years of age at the time of death and including both sexes (Bermúdez de Castro et al., 2004, 2020).

The taxonomy of the SH hominins is a matter of debate as their skeletal remains seem to contain primitive traits, as well as derived traits consistent with Homo neanderthalensis (Arsuaga et al., 1993, 1997, 2014, 2015; Martínez and Arsuaga, 1997; Rightmire, 1998; Stringer, 2012). This led scholars, such as Hublin (2009), to alternate between identifying the SH hominins as Neandertals or as members of Homo heidelbergensis. Previously, Arsuaga et al. (2014) have suggested including European Early and Middle Pleistocene hominin fossils with some Neandertal traits in H. heidelbergensis. Nevertheless, they argued that the case of the SH hominins was more difficult, and, given the differences from the type specimen of H. heidelbergensis, the Mauer mandible, the SH hominins should not be placed in this taxon. In addition, the lack of some derived Neandertal features in the cranium (e.g., suprainiac fossa, reduced mastoid process, elongated cranial shape; Arsuaga et al., 2014) and postcranial skeleton (e.g., shortened distal limb segments; Carretero et al., 2012; Arsuaga et al., 2015) also indicates the SH hominins are not fully Neandertal and should not simply be classified as H. neanderthalensis.

In recent years, researchers have managed to sequence the nuclear and mitochondrial genome of the SH hominins (Meyer et al., 2014, 2016). The results of the mitochondrial genome study unexpectedly indicated that the SH hominins were more closely related to the Denisovans, a group thought to represent an eastern sister taxon of the Neandertals, than they were to the Neandertals or modern humans. In contrast, the nuclear genome study concluded the SH hominins were closer to the Neandertals (Meyer et al., 2014). Thus, the anatomical and nuclear genetic evidence suggests the SH hominins likely represent a population close to the evolutionary origin of the Neandertal lineage (Martínez et al., 2008; Meyer et al., 2014).

Previous studies have demonstrated that the mammalian cochlea and semicircular canals contain strong phylogenetic signals (Billet et al., 2015; Braga et al., 2015; Ekdale, 2016; Mennecart and Costeur, 2016; Costeur et al., 2018; Pfaff et al., 2019; Schwab et al., 2019). The bony labyrinth has also been shown to contain phylogenetic information in primates, including humans and fossil hominins (Spoor et al., 2003; Lebrun et al., 2010; Braga et al., 2015; Quam et al., 2016; Morimoto et al., 2020). As the main sensory organs of the vestibular system, the relative sizes of the semicircular canals in primates have been related to the form of locomotion (Spoor et al., 1994; Spoor, 2003; Rook et al., 2004), while the cochlear length and volume have been argued to be useful in estimations of hearing parameters in primates (Kirk and Gosselin-Ildari, 2009; Coleman and Boyer, 2012).

The hominin bony labyrinth has received increasing attention over the past 20 years, particularly in studies that focus on comparative anatomy and taxonomy (Spoor, 2003; Rook et al., 2004; Kirk and Gosselin-Ildari, 2009; Quam et al., 2016; Conde-Valverde et al., 2018, 2019; Beaudet, 2019; Beaudet et al., 2019). Studies focusing on semicircular canal morphology have concluded that differences in canal dimensions reflect differences in locomotor behavior (e.g., Spoor et al., 1994, 2007; Spoor, 2003). Though the cochlea has received less attention compared to the semicircular canals, recent studies have shown taxonomic differences reflected in cochlear morphology (Braga et al., 2015). Conde-Valverde et al. (2019) have shown that interspecific differences in the proportional lengths of the cochlear turns and the cross-sectional shape of the basal turn of the cochlea demonstrate a mosaic pattern of evolution in the ear of the genus Homo.

The SH hominin bony labyrinth is characterized by a derived pattern of canal proportions shared with Neandertals, including relatively small anterior and posterior canals and a relatively large lateral canal (Quam et al., 2016). One point of departure from Neandertals is the absence of a low placement of the posterior canal in the SH hominins, which are less derived in this regard (Quam et al., 2016). Regarding the cochlea, the SH hominins are characterized by some primitive features resembling chimpanzees, including a proportionally long basal turn, the small size and round shape of the cross-section of the basal turn, a smaller cochlear volume and a low cochlear thickness. Most of these features are also seen in Indonesian Homo erectus and early members of the genus Homo (Urciuoli et al., 2022). The SH hominin cochlea also shows some derived features resembling Neandertals and Homo sapiens, including their long cochlear length (for their body size) and a Neandertal-like proportionally short third turn (Conde-Valverde et al., 2019).

Gunz et al. (2012) reconstructed the bony labyrinths of 50 modern humans and 30 chimpanzees and digitized a series of landmarks and semilandmarks around the semicircular canals, common crus, and cochlea in order to compare the mean shapes of each structure between groups. The protocol used by Gunz et al. (2012) was also able to detect subtle differences between subspecies of chimpanzees, demonstrating that three-dimensional geometric morphometric (3DGM) analyses are able to distinguish between closely related taxa. Beaudet et al. (2019) conducted a similar study on early hominin individuals representing the genera Australopithecus and Paranthropus, and early Homo, comparing them to modern humans and chimpanzees. That study concluded that interspecific differences were largely related to the relative size and shape of each semicircular canal. The genus Paranthropus and most specimens of Australopithecus tended to cluster more closely with the chimpanzees, suggesting plesiomorphy in the relative size and shape of the semicircular canals in these taxa. The cochlear analysis indicated a taxonomic separation based on how tightly coiled the cochlea is, with chimpanzees showing a more tightly coiled cochlea, while the cochlea was less tightly coiled in most of the H. sapiens, Australopithecus, and Paranthropus individuals. Urciuoli et al. (2020) carried out a deformation analysis to compare shape differences in the semicircular canals in a diverse sample of extant monkeys, apes, and humans, as well as fossil hominins, and concluded that there are morphological differences between the extant taxa and fossil hominins which reflect possible phylogenetic relationships between them, as well as major differences between extant taxonomic groups at the genus level.

Three-dimensional geometric morphometric data can also be used to test for morphological integration and modularity between skeletal structures. Morphological integration describes strong nonrandom covariation between different characteristics. In contrast, the concept of morphological modularity arranges sets of characteristics into units called modules, which are relatively independent of one another (Klingenberg, 2008). Morphological integration may influence evolution by augmenting the degree to which interrelated traits are able to change over time, while modularity can cause different characteristic modules to evolve independently of one another and at differing rates (Klingenberg, 2005; Bastir, 2008).

The mammalian ear is divided into three anatomical units representing the outer, middle, and inner ear. Although all three units are involved in the sense of audition, each contains different structures that contribute to the process of translating the mechanical energy of sound waves into nervous impulses for the brain to interpret. Specifically, the structures of the outer and middle ear are related to transmitting sound energy from the surrounding environment to the inner ear, where the cochlea facilitates the perception of sound by the brain. Within the inner ear, the semicircular canals and the macular organs of the vestibule do not have a physiological role in the sense of hearing, but are instead related to proprioception, the sense of balance and equilibrium. Thus, while the semicircular canals are part of the inner ear, they form a different functional module than the cochlea. In this sense, it is likely that each part of the bony labyrinth may be considered a distinct module, subject to different selective pressures related to their function.

While the bony labyrinth in the SH hominins has been analyzed previously, the semicircular canals (Quam et al., 2016) and cochlea (Conde-Valverde et al., 2019) were considered largely separately. Nevertheless, based on linear measurements, Conde-Valverde et al. (2019) concluded that evolutionary changes in the outer, middle, and inner ear likely occurred independently of one another. Here we have applied a 3DGM approach to the bony labyrinth of the SH hominins both to further assess their taxonomic affinities and to examine potential covariation and modularity between the semicircular canals and the cochlea.