In Southern Patagonia (SP), human populations were characterized by unique craniofacial traits that set them apart from other Native American groups (Cocilovo Reference Cocilovo1981; González-José et al. Reference González-José, García-Moro, Dahinten and Hernández2002). Overall, skulls from this region exhibit large frontal and basal dimensions, large orbits and nasal cavities, highly pronounced mastoid and glabellar regions, and a marked occipital ridge (Bernal et al. Reference Bernal, Pérez and González2005, Reference Bernal, Pérez and González2006; Hernández et al. Reference Hernández, Lalueza and García-Moro1997; Lahr Reference Lahr1995; Pérez and Monteiro Reference Pérez and Monteiro2009). These cranial morphometric differences among SP populations are thought to have evolved due to a combination of factors, including low population density, high levels of gene flow among SP human groups, and geographic isolation from external sources of variability (González-José et al. Reference González-José, García-Moro, Dahinten and Hernández2002).
SP populations have played a significant role in various theories regarding the peopling of the Americas; however, their position within these frameworks remains ambiguous and under debate (González-José et al. Reference González-José, García-Moro, Dahinten and Hernández2002; Lahr Reference Lahr1995; Neves and Pucciarelli Reference Neves and Pucciarelli1990, Reference Neves and Pucciarelli1991; Pérez and Monteiro Reference Pérez and Monteiro2009; Pérez et al. Reference Pérez, Lema, Diniz-Filho, Bernal, González, Gobbo and Pucciarelli2011). Several hypotheses have been proposed to explain these traits observed in SP populations, including phylogenetic, biomechanical, climatic and environmental, or multicausal.
The phylogenetic hypothesis stems from the observation of ancestral craniofacial characteristics linking individuals from SP to populations of African and Oceanian ancestry. Shared traits with the first inhabitants of the continent have led to the characterization of these populations as “primitive” or “ancestral,” closely related to Paleoamerican ones (Neves and Pucciarelli Reference Neves and Pucciarelli1990, Reference Neves and Pucciarelli1991; Neves et al. Reference Neves, Powells and Ozolins1999). The biomechanical hypothesis was first proposed by Lahr (Reference Lahr1995), who observed cranial similarities between SP and Inuit populations and associated them with biomechanical masticatory stress, typically correlated with craniofacial robustness and dental wear (Bernal et al. Reference Bernal, Pérez and González2005, Reference Bernal, Pérez and González2006; Hernández et al. Reference Hernández, Lalueza and García-Moro1997; Pérez-Pérez and Lalueza-Fox Reference Pérez-Pérez and Lalueza-Fox1992). However, despite the remarkable craniofacial robustness observed in SP populations, dental wear rates are lower than those reported for individuals from Patagonia, Cuyo, and northwestern Argentina (Bernal et al. Reference Bernal, Pérez and González2005, Reference Bernal, Pérez and González2006). Climate and environmental hypotheses seek to explain the similarities shared by SP and Inuit populations by emphasizing the cold environments common to both. Despite being genetically distant (Chiaroni et al. Reference Chiaroni, Underhill and Cavalli-Sforza2009; de la Fuente et al. Reference de la Fuente, Galimany, Kemp, Judd, Reyes and Moraga2015; de Saint Pierre, Bravi, et al. Reference de Saint Pierre, Bravi, Motti, Fuku, Tanaka, Llop, Bonatto and Moraga2012; de Saint Pierre, Gandini, et al. Reference de Saint Pierre, Gandini, Perego, Bodner, Gómez-Carballa, Corach and Angerhorfer2012), these populations exhibit notable convergences in traits related to cold adaptation. These include cranial morphology (D’Angelo del Campo et al. Reference D’Angelo del Campo, Suby, García-Laborde and Guichón2017; Galland and Friess Reference Galland and Friess2016; Hernández et al. Reference Hernández, Lalueza and García-Moro1997; Lahr Reference Lahr1995; Lalueza, Hernández, and García-Moro Reference Lalueza, Hernández and García-Moro1997), nasal cavity (Noback et al. Reference Noback, Harvati and Spoor2011), os coxae (Betti Reference Betti2014; Betti et al. Reference Betti, von Cramon-Tabaudel, Manica and Lycett2014), pelvic canal (Betti and Manica Reference Bookstein1991; Kurki Reference Kurki2013), and ribs (García-Martínez et al. Reference García-Martínez, Nalla, Ferreira, Bastir, D’Angelo del Campo and Guichón2018), as well as other aspects such as the frequency of oral (Pérez-Pérez and Lalueza-Fox Reference Pérez-Pérez and Lalueza-Fox1992) and spinal pathologies (D’Angelo del Campo, Pastor, et al. Reference D’Angelo del Campo, Pastor, Medialdea, Caballero Grijalba, García Laborde, Salemme, Martín, González Martín, Seldes and Guichón2023; D’Angelo del Campo et al. Reference D’Angelo del Campo, Curti, López, García Laborde, Valenzuela, Motti, Martucci, Palacio, González Martín and Guichón2020, Reference D’Angelo del Campo, González Martín and Guichón2023). They also share metabolic characteristics, stature, body mass and proportions, and overall robustness (Hernández et al. Reference Hernández, Lalueza and García-Moro1997; Leonard Reference Leonard2018; Pearson and Millones Reference Pearson and Millones2005; Pérez and Monteiro Reference Pérez and Monteiro2009). A combination of phenotypic plasticity and genetic adaptations to low temperatures may influence development and growth, resulting in acclimatization evident in both cranial and postcranial anatomical structures (Bernal et al. Reference Bernal, Pérez and González2006; Pérez and Monteiro Reference Pérez and Monteiro2009; Pérez et al. Reference Pérez, Lema, Diniz-Filho, Bernal, González, Gobbo and Pucciarelli2011). Finally, the multicausal hypothesis proposes a combination of the previously described mechanisms. However, due to the lack of systematic evaluation, establishing specific causal factors remains challenging, as each model presents its own strengths and weaknesses (Bernal et al. Reference Bernal, Pérez and González2005, Reference Bernal, Pérez and González2006). In this context, as González-José (Reference González-José2003) noted, biomechanical and climatic/environmental factors alone are insufficient to explain the craniofacial differences observed among SP groups.
Despite the morphological affinity observed among Native populations in SP, finer morphological differences exist among these human groups (Cocilovo Reference Cocilovo1981; Hernández Reference Hernández1992; Hernández et al. Reference Hernández, Lalueza and García-Moro1997; Lalueza et al. Reference Lalueza, Hernández and García-Moro1996; Lalueza, Hernández, and García-Moro Reference Lalueza, Hernández and García-Moro1997; Pérez et al. Reference Pérez, Lema, Diniz-Filho, Bernal, González, Gobbo and Pucciarelli2011; Turbón et al. Reference Turbón, Arenas and Cuadras2017; Varela et al. Reference Varela, Cocilovo and Guichón1997). For instance, several authors (Cocilovo Reference Cocilovo1981; Hernández et al. Reference Hernández, Lalueza and García-Moro1997; Turbón et al. Reference Turbón, Arenas and Cuadras2017) have reported variations in skull shape between marine hunter-gatherers (Yaghan and Kawésqar) and terrestrial groups (Selk’nam and Aonikenk). These differences involve aspects such as prognathism, frontal and nasal regions, and overall cranial size. Based on these variations, Hernández (Reference Hernández1992) described the Kawésqar population as occupying an undefined position based on the craniometric analyses, rather than overlapping with or falling midway between Selk’nam and Yaghan groups. Subsequently, Lalueza and others (Reference Lalueza, Hernández and García-Moro1996) suggested a closer affinity between Selk’nam and Kawésqar, while Varela and others (Reference Varela, Cocilovo and Guichón1997) identified both differences and areas of overlap among groups. Thus, it is evident that cranial shape variation exists among SP populations; however, current evidence does not clearly define their geographic distribution patterns or address these differences from an explicit ecogeographical perspective.
From a genetic perspective, SP populations exhibit low genetic diversity, characterized predominantly by mitochondrial haplogroups C and D (Lalueza, Pérez-Pérez, et al. Reference Lalueza, Pérez-Pérez, Prats, Cornudella and Turbón1997). Individuals from the Beagle Channel region show the lowest genetic diversity and the highest frequency of the C1 lineage (Crespo et al. Reference Crespo, Lanata, Cardozo, Avena and Dejean2018; de la Fuente et al. Reference de la Fuente, Galimany, Kemp, Judd, Reyes and Moraga2015). Y-chromosome studies similarly indicate low diversity in the region (García-Bour et al. Reference García-Bour, Pérez-Pérez, Álvarez, Fernández, María López-Parra, Arroyo-Pardo and Turbón2004). During the Late Holocene, a genetic cline extended from the continental mainland to the western archipelagos, passing through the Isla Grande de Tierra del Fuego. This pattern suggests sustained admixture among SP human groups, with gene flow occurring across ecogeographical regions (Nakatsuka et al. Reference Nakatsuka, Luisi, Motti, Salemme, Santiago, D’Angelo del Campo and Vecchi2020).
Southern Patagonia Environmental and Cultural Characteristics
SP is in the meridional region of the American continent and comprises both continental and insular territories situated south of the 50th parallel of south latitude. Notably, with the exception of Antarctica, SP is the only continental territory located below this parallel. Like the rest of Patagonia, SP is crossed from north to south by the Andes Mountains, which shift to a west–east orientation in Tierra del Fuego. The Andes divide South America into two differentiated slopes: the Pacific and the Atlantic. The Strait of Magellan, the only natural passage between both oceans and located within SP, separates the mainland from the Tierra del Fuego Archipelago (Coronato et al. Reference Coronato, Mazzoni, Vázquez and Coronato2017).
The environment of SP is shaped by its unique location within the South Westerly Wind Belt (SWW), where the Andes Mountains generate pronounced contrasts between the Pacific (windward) and Atlantic (leeward) slopes. Variability in the SWW primarily influences precipitation, and consequently humidity, and, to a lesser extent, temperature. These climatic dynamics directly shape the biomass on both sides of the Andes, resulting in distinct climatic zones that range from hyper-humid oceanic environments in the western archipelagos to dry, windy steppe conditions in the east (Coronato et al. Reference Coronato, Mazzoni, Vázquez and Coronato2017; Garreaud et al. Reference Garreaud, López, Minvielle and Rojas2013; Kilian and Lamy Reference Kilian and Lamy2012). These zones differ markedly in vegetation and resource availability, which in turn has shaped patterns of human settlement and mobility (Coronato et al. Reference Coronato, Mazzoni, Vázquez and Coronato2017; for more details see Supplementary Text 1).
During the Magellan expedition in AD 1520, five ethnographic human groups were identified in SP: the Selk’nam, Haush, Aónikenk, Kawésqar, and Yaghan. The Yaghan and Kawésqar were maritime hunter-gatherers who inhabited the Beagle Channel region, the Austral Fuegian Archipelago, and the southwestern archipelagos (Borrero Reference Borrero2001; Chapman Reference Chapman2010; Gusinde Reference Gusinde1982). The Aónikenk occupied the southern mainland, while the Selk’nam inhabited the northern portion of Isla Grande de Tierra del Fuego; both groups practiced a terrestrial hunter-gatherer subsistence strategy (Borrero Reference Borrero2001; Chapman Reference Chapman2010; Gusinde Reference Gusinde1951, Reference Gusinde1982). The Haush resided in the Mitre Peninsula, located in the southeastern sector of Isla Grande de Tierra del Fuego, and relied on a mixed subsistence strategy incorporating both terrestrial and marine resources (Vidal Reference Vidal, Zangrando, Vázquez and Tessone2011). Nevertheless, there is no consensus regarding their origins: some ethnographers consider the Haush as a distinct group, while others regard them as a Selk’nam lineage (Borrero Reference Borrero2001; Vidal Reference Vidal, Zangrando, Vázquez and Tessone2011; see Supplementary Text 1 for further details).
The study of SP has traditionally relied on ethnographic classifications to interpret past human groups, despite increasing recognition of their limitations. Recent research has emphasized the problematic nature of applying these ethnographic categories, largely derived from postcontact records, to ancient populations, as these may not accurately reflect earlier cultural realities. As several scholars have noted, the use of such classifications in archaeological and bioanthropological contexts risks oversimplifying the complex and dynamic human history of the region (Borrero Reference Borrero2001; Borrero and Martin Reference Borrero, Morello and Román2020; Borrero et al. Reference Borrero, Martin, Wardle, Hitchcock, Schmader and Yu2023; D’Angelo del Campo, González Martín, and Guichón Reference D’Angelo del Campo, Medialdea, García Laborde, Salemme, Santiago, Martín, González Martín and Guichón2021; D’Angelo del Campo et al. Reference D’Angelo del Campo, García Laborde, Valenzuela, Motti, Martucci, Palacio and Guichón2018; Guichón Reference Guichón, Briones and Lanata2002; Nakatsuka et al. Reference Nakatsuka, Luisi, Motti, Salemme, Santiago, D’Angelo del Campo and Vecchi2020; see Supplementary Text 1). In response to these concerns, Borrero and others (Reference Borrero2001) have proposed an ecogeographic framework based on environmental and archaeological factors rather than ethnic boundaries.
Climate/Environmental Hypothesis
The potential influence cold climates on skull size and shape continues to be debated in bioanthropology (Beals et al. Reference Beals, Smith, Dodd, Angel, Armstrong, Blumenberg and Girgis1984; D’Angelo del Campo et al. Reference D’Angelo del Campo, Suby, García-Laborde and Guichón2017; Harvati and Weaver Reference Harvati and Weaver2006; Hubbe et al. Reference Hubbe, Hanihara and Harvati2009; Menéndez Reference Menéndez2018; Stansfield et al. Reference Stibel2023). Individuals from colder environments have often been characterized by larger and more robust skulls (Beals et al. Reference Beals, Smith, Dodd, Angel, Armstrong, Blumenberg and Girgis1984; D’Angelo del Campo et al. Reference D’Angelo del Campo, Suby, García-Laborde and Guichón2017; Hubbe et al. Reference Hubbe, Hanihara and Harvati2009; Menéndez Reference Menéndez2018; Pérez and Montiero Reference Pérez and Monteiro2009), as well as by larger brains (Stibel Reference Suby and Guichón2010). Despite ongoing discussion regarding the relative influence of climate versus population history on different cranial regions, there is general agreement that certain aspects of facial shape are correlated with climatic variables (Harvati and Weaver Reference Harvati and Weaver2006; Hubbe et al. Reference Hubbe, Hanihara and Harvati2009; Menéndez Reference Menéndez2018). These include the nasal cavity, both external and internal nasal structures, and the maxillary sinuses, particularly in individuals adapted to cold and dry environments (D’Angelo del Campo et al. Reference D’Angelo del Campo, Suby, García-Laborde and Guichón2017; Noback et al. Reference Noback, Harvati and Spoor2011; Stansfield et al. Reference Stibel2023). It is important to note that these morphological differences tend to be more pronounced in populations living under conditions of extreme cold (Harvati and Weaver Reference Harvati and Weaver2006).
Despite continued debate at both local and global scales regarding the climatic influence on craniofacial morphology, the similarities observed between southern and northern populations may suggest a possible climatic influence, as proposed by the climate/environmental hypothesis. At the local level, particularly in SP, this hypothesis has been deemed insufficient to fully explain the craniofacial variation observed (González-José Reference González-José2003). However, this conclusion was reached more than two decades ago and was based on classical morphological approaches. Since then, the topic has not been revisited using alternative methodologies. To date, no study has examined craniofacial size and shape variation in SP populations using a geometric morphometric framework (D’Angelo del Campo, González Martín, and Guichón Reference D’Angelo del Campo, Medialdea, García Laborde, Salemme, Santiago, Martín, González Martín and Guichón2023). This is a gap that the present study seeks to fill.
The objective of this study is to analyze variation in skull size and shape among SP groups during the Late Holocene. This region encompasses a broad range of climatic and environmental conditions, with notable differences between the Atlantic and the Pacific slopes. Specifically, we aim to assess craniofacial size and shape variation using 3D geometric morphometrics in individuals from SP, and to explore how these patterns covary with ecogeographical region, subsistence strategy, and sex.
Materials and Methods
Sample
The initial dataset included 173 adult skulls from various American and European collections. However, following rigorous selection, the final sample comprised 54 well-preserved skulls (18 females, 31 males, and five individuals of undetermined sex). The majority (76%) of the analyzed sample dates to the Late Holocene (4200 years BP; sensu Walker et al. Reference Walker, Berkelhammer, Björk, Cwynar, Fisher, Long, Lowe, Newnham, Rasmussen and Weiss2012) (Supplementary Table 1). Detailed bioanthropological information for each individual (Table 1), as well as excavation site information, were sourced from the BIBPA database (Base de Información Bioantropológica de Patagonia Austral; D’Angelo del Campo et al. Reference D’Angelo del Campo, Gökdoğan Aktepe, Bastir and García-Martínez2025). Skulls were excluded from the study if they exhibited poor preservation, pathological alterations (e.g., trauma), or evidence of cultural modification.
Individuals Analyzed from Southern Patagonia.
a Ecogeographical regions: Continent (C), North of Isla Grande de Tierra del Fuego (NIG), Mitre Peninsula (PM), Beagle Channel and Austral Fuegian Archipelago (BC), Última Esperanza Mountains, internal seas, and Brunswikc Peninsula (UEM).
b Subsistence economy: terrestrial (THG) and maritime (MGH) hunter-gatherers.
c Abbreviations: M, male; F, female; U, undetermined sex.
d Institution: CADIC: Centro Austral de Investigaciones Científicas (Ushuaia, Argentina); IMHICIHU: Instituto Multidisciplinario de Historia y Ciencias Humanas (Buenos Aires, Argentina); IP: Instituto de la Patagonia (Punta Arenas, Chile); ME: Museo Etnográfico J.M. Ambrosetti (Buenos Aires,Argentina); MFM: Museo del Fin del Mundo (Ushuaia, Argentina); LEEH: Laboratorio de Ecología Evolutiva Humana (Necochea, Argentina); MNHNC: Museo Nacional de Historia Natural de Chile (Santiago de Chile, Chile); LFM: Reservorio Puerto Santa Cruz Comunidad Lof Fem Mapu (Puerto Santa Cruz, Argentina); NHM: Naturhistorisches Museum (Viena, Austria).
e Salesian Mission “Nuestra Señora de La Candelaria” (Río Grande, Argentina).
f Restored to Provincia de Tierra del Fuego, Antártida e Islas del Atlántico Sur for the restitution to native Selk’nam community Rafaela Ishton.
Ecogeographic and Subsistence Strategy
The present study proposes a revised ecogeographic division of SP, updating earlier models such as proposed by Borrero and others (Reference Borrero2001) to incorporate recent archaeological and bioanthropological findings. This updated model reduces the number of zones and refines regional boundaries based on ecological, archaeological, and climatic evidence. The revised classification is designed to facilitate more robust analytical comparisons, while recognizing the inherent uncertainties in reconstructing SP’s deep human past (see Supplementary Text 1).
Cranial samples were assigned to five ecogeographic regions (Table 1; Figure 1) according to the following ecogeographic subdivision: Beagle Channel and Austral Fuegian Archipelago (BC; n = 16), Continent (C; n = 3), North of Isla Grande de Tierra del Fuego (NIG; n = 29), Mitre Peninsula (MP; n = 2), and Última Esperanza Mountains, internal seas, and Brunswick Peninsula (UEM; n = 4). Ecogeographic regions associated with maritime hunter-gatherers include BC and UEM, while C, NIG, and MP correspond to areas inhabited by terrestrial hunter-gatherers.
Southern Patagonia map showing the five geographical regions: Última Esperanza Mountain, internal seas, and Brunswick Peninsula (UEM); North of Isla Grande de Tierra del Fuego (NIG); Beagle Channel and Austral Fuegian Archipelago (BC); Mitre Peninsula (MP); and Continent (C). Color coding is in the legend. (Color online)
Data Acquisition
Morphometric skull data were acquired following the workflows proposed by Bastir and others (Reference Bastir, García-Martínez, Torres-Tamayo, Palancar, Fernández Pérez, Riesco-López, Osborne-Márquez, Ávila and López-Gallo2019), comprising three main stages: digitizing, post-processing, and landmark registration. Skulls were digitized in 3D using three different methods (Table 1): (1) Computed Tomography (CT) scan: samples at LEEH (Laboratorio de Ecología Evolutiva Humana) and LFM (Reservorio Lof Fem Mapu) were scanned using a Philips Brilliance CT 64 clinical machine (Instituto Radiológico, Mar del Plata, Argentina) with a slice thickness of 0.9 mm and a interslice gap of 0.45 mm; NHM samples were digitized with a Philips Brilliance CT 64 clinical machine (Allgemeines Krankenhaus-Wien, Vienna, Austria), with a slice thickness of 0.67 mm and a interslice gap of 0.33 mm; (2) Photogrammetry: this was conducted manually using PhotoModeler 6.0 (www.photomodeler.com), following a protocol of capturing 13 images of each skull from different positions. A 16-megapixel Canon PowerShot SX520 compact camera with a 42× ultra-wide-angle zoom mounted on a Manfrotto tripod was used for image acquisition. Skulls were placed inside a lightbox at a fixed distance of 40 cm to ensure consistent lighting and photographic parameters; (3) Surface scan: a high-resolution Next Engine surface scanner (67,000 dots per inch and a precision of 0.127 mm) was employed for surface digitization. Subsequently, postprocessing of CT and surface scan data, including cleaning, smoothing, and aligning, was conducted using the 3D Slicer imaging computing platform (https://www.slicer.org; Fedorov et al. Reference Fedorov, Beichel, Kalpathy-Cramer, Finet, Fillion-Robin, Pujol and Christian2012). Photogrammetric data required no postprocessing, as landmarks were manually placed on each photograph and exported directly. Finally, for landmark recording, 3D configurations of homologous landmarks (neurocranium and splanchnocranium) were designed to capture the entire skull shape (Table 2; Figure 2). Surface models from CT and Next Engine scans were imported into Viewbox software version 4.0 (https://www.dhal.com), while photogrammetric landmark configurations were imported from Photomodeler 6.0 as a .txt.
Landmark Template.
Landmark template in four views: frontal, left lateral, posterior, and basal. (Color online)
Geometric Morphometric and Statistical Analyses
Geometric morphometric (GM) methods were employed for the quantitative characterization, analysis, visualization, and comparison of biological form (Bookstein Reference Betti and Manica2018; Slice Reference Stansfield, Mitteroecker, Vasilyev, Vasilyev and Butaric2021; Zelditch et al. Reference Zelditch, Swiderski and Sheets2012). GM methods are based on the analysis of landmark configurations used to capture the shape of an organism or specific structure of an organism. In the present study, cranial morphogeometric variability was assessed using 35 landmarks. Landmark configurations were superimposed using Generalized Procrustes Analysis (GPA), which aligns configurations to a common origin, scales them to unit centroid size (CS), and rotates them to achieve the best-fit criterion (Zelditch et al. Reference Zelditch, Swiderski and Sheets2012). This procedure removes the effects of position, orientation, and scale, allowing subsequent analyses to focus on shape variation. Following superimposition, size was quantified using CS, defined as the square root of the summed squared distances of each landmark from the configuration centroid. A consensus shape configuration, representing the mean shape of all individuals, was calculated and used as reference.
To evaluate the effects of ecogeographic regions, subsistence strategy, and sex on cranial shape (Procrustes coordinates) and to establish a statistical framework for partitioning variance among these factors a Procrustes ANOVA (p < 0.05) was performed, excluding five individuals of undetermined sex from the sex-related analysis (Klingenberg and MacIntyre Reference Klingenberg, Barluenga and Meyer2002; Klingenberg et al. Reference Klingenberg and McIntyre1998). This approach allows the quantification of relative amounts of variation at different levels (Klingenberg et al. Reference Klingenberg and McIntyre1998). Analyses were conducted in MorphoJ (Klingenberg Reference Klingenberg2011) using permutation-based significance testing (10,000 iterations). Given the limited sample size, results were interpreted cautiously, emphasizing the magnitude and structure of shape variation rather than relying solely on p-value significance. The same effect for those factors was explored for size (CS) component by a univariate ANOVA without interactions among factors, also employing permutation-based significance testing (10,000 iterations). These analyses were additionally applied to evaluate differences between the ecogeographic regions associated with maritime hunter-gatherers (BC and UEM) and terrestrial hunter-gatherers (C, NIG, and MP). Measurement error had already been assessed previously (D’Angelo del Campo, Medialdea, et al. Reference D’Angelo del Campo, Medialdea, García Laborde, García-Martínez, Bastir, González-José, González Martín and Guichón2023), thereby safeguarding the reliability of interpretations, which were subsequently evaluated through a Principal Component Analysis (PCA) of the shape variables to visualize and interpret the major axes of shape variation (Zelditch et al. Reference Zelditch, Swiderski and Sheets2012). Variations among subjects, stored in the sample variance-covariance matrix, were analyzed to obtain a set of principal vectors describing relevant trends in the shape variation (Adams et al. Reference Adams, Rohlf and Slice2013; Bookstein Reference Betti and Manica2018). The PCA facilitates the exploration of how specimens are distributed in morphospace and the potential clustering of groups, as well as visualization of associated shape variations. Scatterplots of the first two principal components (PCs) were generated, incorporating 90% confidence ellipses to visualize group distributions by ecogeographic region and subsistence strategy. A second PCA was conducted to assess the effect of sex, including only individuals with known sex (n = 49). All statistical analyses were performed in MorphoJ (Klingenberg Reference Klingenberg2011). To visualize shape changes associated with extreme PC scores, landmark configurations corresponding to the extremes of each axis were warped and rendered using the EVAN Toolkit (http://www.evan-society.org/).
Finally, because data from multiple digitizing techniques were combined, intermethod error was evaluated to assess the comparability of datasets. This error was quantified in a previous study by D’Angelo del Campo, Medialdea, nd others (Reference D’Angelo del Campo, Medialdea, García Laborde, García-Martínez, Bastir, González-José, González Martín and Guichón2023), who analyzed eight skulls, also included in the present study, digitized using CT, photogrammetry, and surface scanner. Using Procrustes ANOVA, PCA, and a UPGMA cluster analysis, no statistically significant differences among digitizing methods were found. Consequently, 3D files derived from different sources were deemed comparable and suitable for combined analysis (D’Angelo del Campo, Medialdea, et al. Reference D’Angelo del Campo, Medialdea, García Laborde, García-Martínez, Bastir, González-José, González Martín and Guichón2023).
Results
The Procrustes ANOVA (Table 3) revealed significant shape differences between the five ecogeographic regions (F = 1.75; p < 0.001), as well as between terrestrial and marine subsistence strategies (F = 2.47; p < 0.001). In addition, significant differences in CS were detected among ecogeographic regions (F = 3.49; p = 0.014) and between subsistence strategies (F = 12.05; p < 0.001). No significant shape differences were observed between males and females (F = 0.92; p = 0.699); however, significant differences in size were found (F = 16.71; p < 0.001; Figure 3).
Procrustes ANOVA Test for Shape (Procrustes Coordinates) and Size (Log Centroid Size) in the Sample by Means of Ecogeographical Regions, Subsistence Strategy, and Sex.
Notes: Procrustes sums of squares (SS), Procrustes mean squares (MS), degrees of freedom (df). Goodall’s F statistic (F).
a Individuals whose sex was indeterminate were excluded.
b Significance level: p < .05.
c Significance level: p < .0001.
Box-plot distributions of centroid size (CS) considering sex, ecogeographic regions, and subsistence strategy as factors. Sex: F, Female; M, Male. Ecogeographic regions: BC, Beagle Channel and Austral Fuegian Archipelago; C, Continent; MP, Mitre Peninsula; NIG, North of Isla Grande de Tierra del Fuego; UEM, Última Esperanza Mountains, internal seas. Subsistence strategy: MHG, Marine hunter-gatherers; THG, Terrestrial hunter-gatherers.
Subsequently, differences in cranial size and shape associated with ecogeographic regions and subsistence behaviors were tested separately for males and female (Figure 4; Supplementary Tables 2 and 3). In males, significant shape differences were observed across ecogeographic region (F = 1.54; p < 0.001) and subsistence strategy (F = 1.46; p = 0.002). In contrast, females only presented significant differences by subsistence strategy (F = 1.45; p = 0.003). Regarding ecogeographical regions (Supplementary Table 4), terrestrial hunter-gatherers showed significant shape differences (F = 1.78; p < 0.001), whereas maritime hunter-gatherers did not (F = 1.19; p = 0.100). Within the male sample, significant morphometric differences were detected for both maritime (F = 1.49; p = 0.002) and terrestrial hunter-gatherers (F = 1.8; p < 0.001). No comparable differences were observed within the female subsample. Overall, only shape exhibited significant sexual dimorphism across ecogeographic regions and subsistence strategies.
Box-plot distributions of centroid size (CS) considering ecogeographic regions and subsistence strategy as factors depending on sex. Ecogeographic regions: BC, Beagle Channel and Austral Fuegian Archipelago; C, Continent; MP, Mitre Peninsula; NIG, North of Isla Grande de Tierra del Fuego; UEM, Última Esperanza Mountains, internal seas. Subsistence strategy: MHG, Marine hunter-gatherers; THG, Terrestrial hunter-gatherers.
PCA extracted a total of 53 PCs (Supplementary Table 5). The first two PCs accounted for 20.73% of the total variance (PC1 = 12.05%, PC2 = 8.68%). Shape variation along these dimensions reflected ecogeographic differentiation. Relative to the consensus shape, individuals with strongly negative PC1 scores were characterized by an expanded neurocranium with an obtuse cranial base angle, increased anteroposterior and superoinferior dimensions, and a reduced splanchnocranium. These individuals also displayed lower and more laterally positioned orbits, a higher palatal and nasal cavity, and a lateral reduction of the zygomatic arches. Conversely, individuals with positive PC1 scores resembled a reduced neurocranium with a smaller cranial base angle, an expanded splanchnocranium with marked prognathism, higher and more medially orbits, a notable lower palate, and pronounced, elevated zygomatic arches. Along PC2, negative scores were associated with posterior projection of the frontal, parietal, and occipital bones. The splanchnocranium was characterized by a squared facial outline, lower and more centrally positioned orbits, a lower nasal cavity, and low, laterally projecting zygomatic arches. Positive PC2 scores reflected a reduction of the neurocranium in an anteroinferior direction and a splanchnocranium with a conic and rhomboidal shape, higher and more laterally positioned orbits, a higher nasal cavity, and smaller zygomatic bones positioned laterally and superiorly, with elevated zygomatic arches. In summary, individuals with positive scores were characterized by increased robustness, marked prognathism, and a narrower neurocranium, whereas negative scores were associated with gracility, a reduced splanchnocranium, and an expanded neurocranium.
Regarding geographical origin, the distribution of shapes along PC1 differentiated individuals from the Pacific slope (negative PC1 and PC2 scores; Figure 5) from those from the Atlantic slope (positive PC1 and PC2 scores; Figure 5), indicating a longitudinal west–east gradient in cranial shape variation. A latitudinal gradient was also evident along PC1, with negative forms predominating in continental regions (C and UEM) and positive scores in archipelagic regions (NIG, BC, and MP). Additionally, a north–south gradient placed higher positive scores in southern, higher latitude areas. Differences between maritime and terrestrial hunter-gatherers were observed along PC2 (Figure 5), with maritime hunter-gatherer configurations tending toward negative scores and terrestrial hunter-gatherer ones toward positive scores.
PC1 versus PC2 plot in shape space of skull. Surface warps show morphological variation associated with different parts of the morphospace. (Color online)
Figure 5 further showed that the morphospace defined by positive PC1 and negative PC2 scores was dominated by terrestrial hunter-gatherers from the NIG region. In contrast, morphologies with negative PC1 and positive PC2 scores were predominantly associated with maritime hunter-gatherers from the western UEM areas of the continent and western archipelagos along the Pacific slope of the Andes.
In the PCA restricted to individuals of known sex, 48 PCs were extracted (Supplementary Table 6). PC1 and PC2 accounted for 20.737% of the total variance (PC1 = 12.054%, PC2 = 8.684%). The sex-specific scatterplot revealed substantial overlap between males and females along both PC1 and PC2 (Figure 6).
PC1 versus PC2 plot in shape space of skull. Surface warps show morphological variation associated with different parts of the morphospace. Only individuals with sex determined were included. (Color online)
Discussion
This study employed geometric morphometric methods to investigate variation within a unique sample of Late Holocene hunter-gatherer populations from SP. Previous morphological studies in the region have predominantly relied on ethnographic group classification as proxies to infer geographic origins (D’Angelo del Campo, González Martín, and Guichón Reference D’Angelo del Campo, Medialdea, García Laborde, Salemme, Santiago, Martín, González Martín and Guichón2021). However, several studies have argued that the use of ethnographic categories may be inappropriate for studying SP populations (Borrero Reference Borrero2001; Borrero and Martin Reference Borrero, Martin, Wardle, Hitchcock, Schmader and Yu2023; Borrero et al. Reference Borrero, Morello and Román2020; Guichón Reference Guichón, Briones and Lanata2002; see Supplementary Text 1):
We cannot know how the individuals from ∼1500 to 500 BP self-identified, if language differentiation reflected the languages in historical periods, if the cultures of the past were similar enough to those of historical periods to have a meaningful degree of continuity, or if there were further important subdivisions beyond the groupings recognized by Europeans at the time of contact [Nakatsuka et al. Reference Nakatsuka, Luisi, Motti, Salemme, Santiago, D’Angelo del Campo and Vecchi2020:4].
This issue is particularly significant, as the present study represents the first analysis of craniofacial morphological variation across ecogeographical regions of SP.
Previous studies (Cocilovo Reference Cocilovo1981; Hernández Reference Hernández1992; Hernández et al. Reference Hernández, Lalueza and García-Moro1997; Lalueza et al. Reference Lalueza, Hernández and García-Moro1996; Lalueza, Hernández, and García-Moro Reference Lalueza, Hernández and García-Moro1997; Pérez et al. Reference Pérez, Lema, Diniz-Filho, Bernal, González, Gobbo and Pucciarelli2011; Turbón et al. Reference Turbón, Arenas and Cuadras2017; Varela et al. Reference Varela, Cocilovo and Guichón1997) reported differences between ethnic groups throughout SP, although with inconclusive results. The findings of the present study partially align with previous research (Cocilovo Reference Cocilovo1981; Hernández et al. Reference Hernández, Lalueza and García-Moro1997; Lalueza et al. Reference Lalueza, Hernández and García-Moro1996; Lalueza, Hernández, and García-Moro Reference Lalueza, Hernández and García-Moro1997; Pérez et al. Reference Pérez, Lema, Diniz-Filho, Bernal, González, Gobbo and Pucciarelli2011; Turbón et al. Reference Turbón, Arenas and Cuadras2017; Varela et al. Reference Varela, Cocilovo and Guichón1997), as overlap between individuals from different ecogeographic regions was observed when plotting PC1 against PC2. However, statistically significant differences emerged between these regions when considering all shape variables, including comparisons between maritime and terrestrial hunter-gatherers. For the first time, variations in craniofacial shape and size among human groups from SP are differentiated based on specific ecogeographic characteristics of the region.
The distribution of shapes along PC1 appears to reflect existing morphometric variability from the west (negative scores) to the east (positive scores) of SP, corresponding to different ecogeographical regions. The most gracile forms are associated with the hyper-humid, temperate, leafy, and resource-rich Pacific slope, whereas the most robust forms are linked to the dry, windy, arid, and resource-scarce Atlantic slope. Notably, previous studies (D’Angelo del Campo et al. Reference D’Angelo del Campo, Suby, García-Laborde and Guichón2017; Harvati and Weaver Reference Harvati and Weaver2006; Stansfield et al. Reference Stibel2023) have reported that the most robust forms occur in the coldest and driest regions worldwide, a pattern also observed in SP. A similar longitudinal pattern was described by Alfonso-Durruty and others (Reference Alfonso-Durruty, Giles, Román and Morello2017), who reported larger, more robust, and heavier long bones in the eastern region of SP, particularly pronounced in males. The authors attributed these differences to humidity, suggesting that this environmental factor may act as a selective force; humidity, along with temperature, has been widely considered to correlate with variations in general body morphology (Ruff Reference Ruff1994). However, in SP, geomorphological, climatic, and environmental studies (Garreaud et al. Reference Garreaud, López, Minvielle and Rojas2013; Kilian and Lamy Reference Kilian and Lamy2012) indicate that wind, specifically the SWW, and the Andes Mountains play a key role in shaping regional biomes and environmental variation. These factors influence precipitation intensity and frequency as well as, to a lesser extent, temperature and humidity. As D’Angelo del Campo, Romero, and others (Reference D’Angelo del Campo, Romero, Salega and Guichón2023:662) have argued for SP, “If an environmental factor could influence bone morphology, it would probably be the wind, and not the humidity.” Nonetheless, elucidating the precise mechanisms by which humidity or wind may act as selective forces shaping cranial or other skeletal morphology lies beyond the scope of this study.
The results of this study have revealed latitudinal differences in craniofacial shape (Figure 5). Human groups in the north and west of the Strait of Magellan exhibit more gracile shapes, whereas those in the southeast display the most robust morphologies. A previous craniometric study identified latitude as the primary factor differentiating human groups in SP (Lalueza et al. Reference Lalueza, Hernández and García-Moro1996). While the present work also identifies significant size and shape differences related to latitude, this relationship is not consistent across SP. As discussed above, ecogeographic and climatological variation in SP is more strongly influenced by the Pacific and Atlantic slopes than by latitude. Consequently, the coldest ecogeographic regions, where the most robust shapes and largest size were observed, are located on the Atlantic slope, rather than in the southern regions predicted by a latitudinal model. Thus, cranial variations appear to follow Bergmann’s rule (Bergman Reference Bergmann1847) without being strictly influenced by latitude. In this case, continental shapes (UEM and C) exhibit more gracile forms and smaller sizes than CB, whereas individuals from NIG and MP show the most robust shapes and largest sizes, exceeding those observed in BC in southern SP. Similarly, D’Angelo del Campo, Romero, and others (Reference D’Angelo del Campo, Romero, Salega and Guichón2023) have reported that, in postcranial bones from SP, the expected relationship between latitude and decreasing temperature is not consistently observed. This phenomenon is clearly evident in our cranial data.
Supporters of the climatic/environmental hypothesis argue that environmental influences may explain part of the morphological diversity observed in SP, highlighting notable phenotypic plasticity in response to environmental and climatic variations (Hernández et al. Reference Hernández, Lalueza and García-Moro1997; Lalueza, Hernández, and García-Moro Reference Lalueza, Hernández and García-Moro1997; Pérez and Monteiro Reference Pérez and Monteiro2009; Pérez et al. Reference Pérez, Lema, Diniz-Filho, Bernal, González, Gobbo and Pucciarelli2011). Considering our findings, environmental and climatic factors appear to exert a significant influence on the skull morphologies observed across SP. These results lend support to the environmental and climate hypothesis in SP without requiring comparisons with Inuit populations. Previous reluctance toward accepting this hypothesis stemmed from the challenge of explaining the extensive craniofacial diversification existing in SP, underscoring the need for further evidence regarding the role of climate in the region (Bernal et al. Reference Bernal, Pérez and González2006; Pérez and Monteiro Reference Pérez and Monteiro2009), an aspect addressed by the study. Nevertheless, it remains difficult to rule out genetic adaptations associated with such morphogeometric patterns. An illustrative example of environmental and climatic adaptations is observed in high-altitude populations. Andean, Tibetan, and Ethiopian populations are adapted to chronic hypoxic conditions and have gradually developed distinctive biological adaptations over time, including specific genetic traits associated with high-altitude adaptation (Frisancho Reference Frisancho2013; Sharma et al. Reference Slice and Slice2005). Among Andean highlanders, morphological and anatomical variation is also evident. Until a few decades ago, researchers proposed two entirely distinct pathways for high-altitude adaptation: a primarily genetic route in Tibetan and Ethiopian populations, and a morphological one in Andean populations (Beall Reference Beall2007; Frisancho Reference Frisancho2013). Regardless of the existence of genetic adaptations, discerning the extent of influence of environmental and climatic versus genetic factors (extrinsic versus intrinsic factors) remains unattainable without further experimental investigation.
Genetic studies conducted on SP populations have revealed low levels of genetic diversity among the inhabitants of this region (García-Bour et al. Reference García-Bour, Pérez-Pérez, Álvarez, Fernández, María López-Parra, Arroyo-Pardo and Turbón2004; Lalueza, Pérez-Pérez, et al. Reference Lalueza, Pérez-Pérez, Prats, Cornudella and Turbón1997), possibly influenced by migration histories, environmental adaptation, founder effects, and other forms of genetic drift (Crespo et al. Reference Crespo, Lanata, Cardozo, Avena and Dejean2018; de la Fuente et al. Reference de la Fuente, Galimany, Kemp, Judd, Reyes and Moraga2015). Nevertheless, recent research has identified distinct genetic human groups with significant differences (Balentine et al. Reference Balentine, Alfonso-Durruty, Reynolds, Vilar, Morello, Román and Springs2023; Crespo et al. Reference Crespo, Lanata, Cardozo, Avena and Dejean2018; Nakatsuka et al. Reference Nakatsuka, Luisi, Motti, Salemme, Santiago, D’Angelo del Campo and Vecchi2020). Notably, a cline extending from the Continent to the UEM, particularly across the MP between the NIG and BC regions, has been reported (Nakatsuka et al. Reference Nakatsuka, Luisi, Motti, Salemme, Santiago, D’Angelo del Campo and Vecchi2020). This genetic pattern suggests active admixture among SP human groups during the last millennia of the Late Holocene, with gene flow occurring across ecogeographic regions (Nakatsuka et al. Reference Nakatsuka, Luisi, Motti, Salemme, Santiago, D’Angelo del Campo and Vecchi2020), a phenomenon also noted by ethnographers (Bridges Reference Bridges1886; Gusinde Reference Gusinde1982). In contrast, the phenotypic expression, as revealed by morphometric analysis, displays a pattern that differs from that observed in genetic studies. Consequently, the genetic and the morphological data do not fully align, an outcome also reported by D’Angelo del Campo, Romero, and others (Reference D’Angelo del Campo, Romero, Salega and Guichón2023) in their analysis of humeri from SP individuals. This discrepancy may be partly explained by environmental or climatic factors.
Regarding sexual dimorphism, the overlapping ranges between males and females observed in this study have also been previously reported in SP populations (Hernández Reference Hernández1992; Varela et al. Reference Varela, Cocilovo and Guichón1997). This apparent homogeneity may reflect either a genuinely low level of sexual dimorphism, as documented in other populations globally (Bigoni et al. Reference Bigoni, Velemínská and Bružek2010; Green and Curnoe Reference Green and Curnoe2009), or methodological limitations related to the landmark template employed, which was not specifically designed to capture sexually dimorphic traits. In particular, features such as the mastoid process, nuchal crest, or superciliary arches, traditionally associated with sexual dimorphism (Buikstra and Ubelaker Reference Buikstra and Ubelaker1994; Ferembach et al. Reference Ferembach, Schwidetzky and Stloukal1980; White et al. Reference White, Black and Folkens2012), were not included. Future research should delve deeper into this topic. Furthermore, the absence of significant differences among females across ecogeographical regions is especially intriguing. One plausible explanation could be related to attacks and raids between human groups in SP aimed at abducting women. Various ethnographers have documented these incidents, highlighting Selk’nam raids on Yaghan settlements for this purpose (Bridges Reference Bridges1886; Gusinde Reference Gusinde1982). Such exchanges would have reduced consanguinity and diluted population-level differences, likely affecting females more strongly, as they were the individuals most frequently moving between groups.
Finally, several limitations should be considered in future analyses. Increasing the sample size, particularly from MP and C, to better represent variation and enhance the statistical power of the analyses would be advisable. However, it must be noted that only 1,251 Holocene human remains are currently available from SP, distributed across nearly 10,000 years of human occupation in a vast area (D’Angelo del Campo et al. Reference D’Angelo del Campo, Gökdoğan Aktepe, Bastir and García-Martínez2025). Furthermore, contextual and bioanthropological information for many of these individuals is scarce and heterogeneous (D’Angelo del Campo, González Martín, and Guichón Reference D’Angelo del Campo, Medialdea, García Laborde, Salemme, Santiago, Martín, González Martín and Guichón2021; D’Angelo del Campo et al. Reference D’Angelo del Campo, García Laborde, Valenzuela, Motti, Martucci, Palacio and Guichón2018, Reference D’Angelo del Campo, Gökdoğan Aktepe, Bastir and García-Martínez2025), often due to nonsystematic excavations or donations to regional institutions, situations where archaeological and bioanthropological contextual data are lost (Suby and Guichón Reference Sharma, Varshney and Sethy2022). Efforts to address this information gap include stable oxygen isotope (δ18O) analysis, which can help assess geographic origin and paleomobility. The initial study on this topic in SP commenced a few years ago and contributed to identifying the continental origin of some individuals exhumed in the cemetery of the Salesian Mission “Nuestra Señora de La Candelaria” in NIG (Motti et al. Reference Motti, Winingear, Valenzuela, Nieves-Colón, Harkins, García Laborde, Bravi, Guichón and Stone2020). Further research in this direction remains essential.
Conclusion
The skulls analyzed from SP exhibited statistically significant differences associated with ecogeographic region and subsistence strategy. These differences followed a clear longitudinal pattern. Individuals inhabiting the Pacific slope, classified as maritime hunter-gatherers, displayed more gracile and a lighter cranial morphology, characterized by a reduced splanchnocranium, an expanded neurocranium, and smaller overall size. In contrast, individuals from the Atlantic slope, classified as terrestrial hunter-gatherers, exhibited marked robustness, pronounced prognathism, reduced neurocranial dimensions, and larger overall size.
Despite these west–east differences, a latitudinal pattern was also evident. Individuals from continental areas (UEM and C) exhibited more gracile cranial shapes compared to those from island regions (NIG, BC, MP). However, this trend was reversed in the Austral Fuegian Archipelago, where the heaviest, most robust, and largest cranial forms were observed in NIG and MP rather than in the southernmost regions of BC. NIG and MP, situated on the Atlantic slope, experienced harsher climatic conditions than BC. Accordingly, these populations conformed to Bergmann’s rule, with the most robust and largest forms occurring in the coldest environments, independent of latitude.
No evidence of sexual dimorphism was detected in the analyzed parameters. Further analyses will require increasing the currently small sample sizes, especially for UEM and MP. This can be achieved by improving the individual-level metadata and by conducting additional chronological and stable isotope analyses, which will enhance our understanding of individual mobility patterns and population origins.
Acknowledgments
We extend our sincere gratitude to museums and research institutions in Argentina (Museo del Fin del Mundo, Ushuaia; CADIC-CONICET, Ushuaia; Museo Etnográfico Juan B. Ambrosetti, Buenos Aires) and Chile (Instituto de la Patagonia, Punta Arenas) for generously granting us permission to study the human skeletons in their collections. We also wish to acknowledge the Lof Fem Mapu Mapuche-Tehuelche community of Puerto Santa Cruz for their collaboration. Additionally, we express our appreciation to the Instituto Radiológico de Mar del Plata and SYNTHESIS funds from Vienna for their invaluable assistance and permission to conduct CT scans for our scientific research. Finally, we want to thank Luis A. Borrero and Ricardo A. Guichón for their valuable contributions.
Funding Statement
This project was supported by CONICET – Proyectos de Investigación Plurianuales (PIP), 2021–2023. Grant/Award number: PIP-11220200102594CO “Cazadores recolectores de la costa patagónica meridional. Arqueología ambiental, ecología humana y vinculaciones regionales.” FONCyT–PICT, Ministerio de Ciencia, Tecnología e Innovación (Argentina) 2011–2014. Grant/Award number: PICT 0575 “Escenarios paleopatológicos y epidemiológicos pre y post contacto interétnico en la Patagonia Austral y Tierra del Fuego la Isla Grande, Tierra del Fuego.” CONICET postdoctoral grant to MDDdC. MB and DGM are funded by PID2020-115854GB-I00/MCIN/AEI/10.13039/501100011033, Ministerio de Innovación y Ciencia (Spain) and the European Union.
Data Availability Statement
Table 1 and Supplementary Table 1 show the institutions where individuals are located, except for the SMLC individuals, which as noted, were returned in December 2019 to the authorities of the province of Tierra del Fuego, Antártida e Islas del Atlántico Sur.
Competing Interests
The authors declare that they have no conflicts of interest.
Supplementary Material
The supplementary material for this article can be found at https://doi.org10.1017/laq.2026.10183.
Supplementary Text 1. Climate, landscape units, Southern Patagonia hunter-gatherers, and Southern Patagonia regions (includes Supplementary Figure 1. Southern Patagonian climates, Supplementary Figure 2. Southern Patagonian landscape units, and Supplementary Figure 3. Map of Southern Patagonia showing the six geographical regions proposed by Borrero et al. [Reference Borrero2001]).
Supplementary Table 1. Southern Patagonia Initial Sample.
Supplementary Table 2. Procrustes ANOVA Test by Ecogeographical Region and Subsistence Strategy.
Supplementary Table 3. Procrustes ANOVA Test by Means of Sex by Ecogeographical Region and Subsistence Strategy.
Supplementary Table 4. Procrustes ANOVA Test by Means of Subsistence Strategy.
Supplementary Table 5. Eigenvalue of each PC.
Supplementary Table 6. Eigenvalue of each PC for Sex Sample.
